Repositorio Institucional de Documentos

2021 81 Ernesto Anoz Carbonell Molecular and dynamic mechanisms of prokaryotic and eukaryotic flavoenzymes: insights into their implication in human metabolism and health Director/es Medina Trullenque, Milagros Aínsa Claver, José Antonio © Universidad de Zaragoza Servicio de Publicaciones ISSN 2254-7606 Ernesto Anoz Carbonell MOLECULAR AND DYNAMIC MECHANISMS OF PROKARYOTIC AND EUKARYOTIC FLAVOENZYMES: INSIGHTS INTO THEIR IMPLICATION IN HUMAN METABOLISM AND HEALTH Director/es Medina Trullenque, Milagros Aínsa Claver, José Antonio Tesis Doctoral Autor 2021 Repositorio de la Universidad de Zaragoza – Zaguan http://zaguan.unizar.es UNIVERSIDAD DE ZARAGOZA Escuela de Doctorado Programa de Doctorado en Bioquímica y Biología Molecular Repositorio de la Universidad de Zaragoza – Zaguan http://zaguan.unizar.es Tesis Doctoral Molecular and dynamic mechanisms of prokaryotic and eukaryotic flavoenzymes: insights into their implication in human metabolism and health Autor Ernesto Anoz Carbonell Director/es Milagros Medina Trullenque José Antonio Aínsa Claver UNIVERSIDAD DE ZARAGOZA Facultad de Ciencias 2020 Dña. MILAGROS MEDINA TRULLENQUE, Catedrática del Departamento de Bioquímica y Biología Molecular y Celular de la Universidad de Zaragoza, D. JOSÉ ANTONIO AÍNSA CLAVER, Profesor titular del Departamento de Microbiología, Pediatría, Radiología y Salud Pública de la Universidad de Zaragoza, CERTIFICAN: Que la Tesis Doctoral “Molecular and dynamic mechanisms of prokaryotic and eukaryotic flavoenzymes: insights into their implication in human metabolism and health” ha sido realizada por el graduado ERNESTO ANOZ CARBONELL en el Departamento de Bioquímica y Biología Molecular y Celular (Facultad de Ciencias), y en el Departamento de Microbiología, Pediatría, Radiología y Salud Pública (Facultad de Medicina), y en el Instituto de Biocomputación y Física de Sistemas Complejos (BIFI), de la Universidad de Zaragoza bajo su dirección, y que reúne, a su juicio, las condiciones requeridas para optar al Grado de Doctor por la Universidad de Zaragoza y a la mención de Doctorado Internacional. Zaragoza, Octubre de 2020 Fdo.: Milagros Medina Trullenque Fdo.: José Antonio Aínsa Claver Index Ernesto Anoz Carbonell – Doctoral Thesis 4 − Publication III. Anoz-Carbonell E, Timson DJ, Pey AL, Medina M. 2020. The catalytic cycle of the antioxidant and cancer-associated human NQO1 enzyme: hydride transfer, conformational dynamics and functional cooperativity. Antioxidants, 9(9):E772. JCR Impact Factor 2019: 5.014. Rank: Q1 (56/297) Biochemistry and Molecular Biology; Q1 (7/61) Medicinal Chemistry; D1 (10/139) Food Science & Technology. − Publication IV. Sebastián M, Anoz-Carbonell E, Gracia B, Cossio P, Aínsa JA, Lans I, Medina M. 2018. Discovery of antimicrobial compounds targeting bacterial type FAD synthetases. Journal of Enzyme Inhibition and Medicinal Chemistry, 33:1, 241-254. JCR Impact Factor 2018: 4.027. Rank: Q1 (10/61) Medicinal Chemistry; Q2 (84/299) Biochemistry and Molecular Biology. − Publication V. Lans I, Anoz-Carbonell E, Palacio-Rodríguez K, Aínsa JA, Medina M, Cossio P. 2020. In silico discovery and biological validation of ligands FAD synthase, a promising new antimicrobial target. PLOS Computational Biology, 16(8):e1007898. JCR Impact Factor 2019: 4.700. Rank: Q1 (9/77) Biochemical Research Methods; Q1 (6/59) Mathematical & Computational Biology. Resumen Resumen 7 Las flavoenzimas y flavoproteínas son biomoléculas versátiles y diversas que están implicadas en el metabolismo energético y otros procesos celulares como la transducción de señales, la síntesis de nucleótidos, el plegamiento de proteínas o la defensa frente al estrés oxidativo. Estas proteínas tienen como cofactores los derivados de riboflavina (RF, vitamina B2), mononucleótido de flavina (FMN) y/o dinucleótido de flavina y adenina (FAD), que les confieren propiedades únicas y versátiles. Todos los organismos contienen flavoproteínas y flavoenzimas clave, y muchas de ellas se han convertido en interesantes dianas terapéuticas o herramientas biotecnológicas. En esta tesis, se ha indagado en los mecanismos moleculares de flavoenzimas y flavoproteínas con funciones metabólicas clave en procariotas y eucariotas, como las enzimas humanas RF quinasa (Publicación I), FAD sintetasa (FADS) (Publicación II) o NAD(P)H:quinona oxidorreductasa 1 (Publicación III), o las FADS procariotas (Publicaciones IV y V). La caracterización detallada de estas enzimas contribuye a la mejor comprensión de sus patologías asociadas, y sienta las bases de nuevas estrategias terapéuticas y del diseño de compuestos dirigidos a estas dianas. Por ejemplo, aquí presentamos una primera aproximación a la búsqueda de inhibidores de las FADS procariotas que puede contribuir a su explotación farmacológica como potencial agentes antimicrobianos (Publicaciones IV y V). Esta Tesis Doctoral, presentada en la modalidad de compendio de publicaciones, incluye las siguientes publicaciones: − Publicación I. Anoz-Carbonell E, Ribero M, Polo V, Velázquez-Campoy A, Medina M. 2020. Human riboflavin kinase: species-specific traits in the biosynthesis of the FMN cofactor. The FASEB Journal, 34:10871–10886. JCR Impact Factor 2019: 4.966. Rank: Q1 (57/297) Biochemistry and Molecular Biology; D1 (9/93) Biology; Q2 (58/195) Cell Biology. − Publicación II. Leone P, Galluccio M, Quarta S, Anoz-Carbonell E, Medina M, Indiveri C, Barile M. 2019. Mutation of aspartate 238 in FAD synthase isoform 6 increases the specific activity by weakening the FAD binding. International Journal of Molecular Sciences, 20(24):6203. JCR Impact Factor 2019: 4.556. Rank: Q1 (74/297) Biochemistry and Molecular Biology; Q2 (48/177) Chemistry (multidisciplinary). Ernesto Anoz Carbonell – Doctoral Thesis 8 − Publicación III. Anoz-Carbonell E, Timson DJ, Pey AL, Medina M. 2020. The catalytic cycle of the antioxidant and cancer-associated human NQO1 enzyme: hydride transfer, conformational dynamics and functional cooperativity. Antioxidants, 9(9):E772. JCR Impact Factor 2019: 5.014. Rank: Q1 (56/297) Biochemistry and Molecular Biology; Q1 (7/61) Medicinal Chemistry; D1 (10/139) Food Science & Technology. − Publicación IV. Sebastián M, Anoz-Carbonell E, Gracia B, Cossio P, Aínsa JA, Lans I, Medina M. 2018. Discovery of antimicrobial compounds targeting bacterial type FAD synthetases. Journal of Enzyme Inhibition and Medicinal Chemistry, 33:1, 241-254. JCR Impact Factor 2018: 4.027. Rank: Q1 (10/61) Medicinal Chemistry; Q2 (84/299) Biochemistry and Molecular Biology. − Publicación V. Lans I, Anoz-Carbonell E, Palacio-Rodríguez K, Aínsa JA, Medina M, Cossio P. 2020. In silico discovery and biological validation of ligands FAD synthase, a promising new antimicrobial target. PLOS Computational Biology, 16(8):e1007898. JCR Impact Factor 2019: 4.700. Rank: Q1 (9/77) Biochemical Research Methods; Q1 (6/59) Mathematical & Computational Biology. Introduction Introduction 11 I. FLAVINS Biological cofactors extend the catalytic repertoire of proteins by increasing the limited chemical space of the aminoacyl side chains of peptides. Some of these cofactors catalyze a small but nevertheless essential set of biological reactions, such as biotin or thiamine pyrophosphate. In contrast, other cofactors are far more versatile contributing to very different chemical reactions, as for example flavin and pyridoxine-derived cofactors (Leys and Scrutton, 2016). Flavin cofactors are a family of heterocyclic compounds with the basic structure of 7,8-dimethyl-10-alkylisoalloxazine (FIG I.1), also referred as lumichrome. The term is derived from the Latin word flavus, due to their characteristic yellow color in the oxidized state. Figure I.1. Chemical structure and atomic numbering of the 7,8-dimethilisoalloxazine ring. R determines the nature of the flavin: R=CH3, lumiflavin; R=ribityl, riboflavin (RF); R=ribityl-5′-phosphate, flavin mononucleotide (FMN); R=ribityl-(9adenosyl)-pyrophosphate, flavin adenine dinucleotide (FAD). Another non-technical but widely extended terminology is the designation of the two faces of the isoalloxazine ring as re and si faces. If the isoalloxazine ring is oriented with the N(10) in the lower part of the molecule, the re face is the one in which the benzene is located at the left side, as shown in FIG I.2. Figure I.2. Re and Si faces of the isoalloxazine ring of FMN. Ernesto Anoz Carbonell – Doctoral Thesis 12 In general, the flavins with higher biological relevance are riboflavin (RF) and their active derivatives flavin mononucleotide (FMN, also termed riboflavin 5'- phosphate) and flavin adenine dinucleotide (FAD) (FIG I.3), as well as the inactive products of their photodegradation, lumiflavin and lumichrome. Figure I.3. Chemical structure of oxidized isoalloxazine ring and its most biologically relevant derivatives (in red): RF (with a ribityl chain attached to the N(10)), FMN (formed through the addition of a phosphate group to the 5’ carbon of the RF ribityl chain) and FAD (formed by condensation of FMN with a molecule of AMP). Despite flavins are usually referred as nucleotides, this term is not strictly correct since the bond stablished between the ribityl chain and the isoalloxazine is a carbon-nitrogen bond, susceptible to hydrolysis, and not a glycosidic bond. Therefore, the nomenclature of FMN and FAD would be incorrect, but it is traditionally accepted. Flavins are not only riboflavin-derivatives Flavins are ubiquitous in the three-domains of life (Archaea, Bacteria and Eukarya) (Macheroux et al., 2011). In general, RF is not used in any enzyme as a cofactor, but it seems to be the preferred storage form in some organisms (e.g. archaea dodecins, riboflavin-binding protein in chicken eggs) (Grininger et al., Introduction 13 2009; Monaco, 1997). Furthermore, all organisms use RF as the precursor of FMN and FAD. On the contrary, FMN and FAD are used as cofactors of a wide plethora of flavoproteins and flavoenzymes (Fraaije and Mattevi, 2000). In addition, some other flavins with modifications in the ribityl chain and/or in the isoalloxazine ring, have been found in nature, but in general they are restricted to few species or genus. Examples of flavins modified at the ribityl chain would include schizoflavins (2,3,4-trihydroxy-4-carboxybutyl)isoalloxazine and 7,8-dimethyll0-(2,3,4-trihydroxy-4-formylbutyl)isoalloxazine) from Schizophillum commune (Tachibana and Murakami, 1975) and the glycosylated or phosphorylated flavin derivatives (apart from FMN). Some examples of flavins with modifications at the isoalloxazine moiety are roseoflavin (8-dimethylaminoriboflavin) produced from Streptomyces davawensis (Lee et al., 2009), 8-hydroxy-5-deazaflavin (constituent of cofactor F420) produced by some Actinobacteria (Daniels et al., 1985), or nekoflavin (8α-hydroxyriboflavin) and 7α-hydroxyriboflavin produced in some mammals (Matsui, 1965). Furthermore, flavin cofactors covalently linked to their corresponding flavoproteins could be also included within this later group (Piano et al., 2017). Recently, a prenylated FMN has been found as cofactor of yeast ferulic acid decarboxylases Fdc1 and prokaryotic UbiD (de)carboxylases (Payne et al., 2015), but only when co-expressed with the corresponding flavin prenyltransferases Pad10 or UbiX, respectively. This unexpected discovery opens the door to new flavin-derived cofactors broadening the chemical space and catalytic repertoire of flavoproteins. Additionally, it has been extensively reported the presence of 6or 8-hydroxyflavins when isolating and purifying some flavoproteins, such as electrontransferring flavoprotein of Peptostreptococcus elsdenii, pig liver glycolate oxidase, the D-aspartate oxidase from Bos taurus or Octopus vulgaris, cellobiose dehydrogenase from Humicola insolens or the human apoptosis factor (Igarashi et al., 1999; Marshall et al., 2005; Mayhew and Ludwig, 1975; Muller and Edmondson, 2018; Negri et al., 1987; Tedeschi et al., 1994). The biological role of these derivatives is still unknown, and they are probably generated as result of the oxidation of flavoproteins during the protein production and/or purification processes (Bertagnolli and Hager, 1993; Gong et al., 2007; Negri et al., 1987). This is further supported by the fact that proteins loaded with these cofactors are less active than with FMN or FAD, and there is not a characterized biosynthetic route for the modified cofactor. Ernesto Anoz Carbonell – Doctoral Thesis 20 In addition to de novo biosynthesis, yeast species are capable of RF uptake. For example, to date, five RF transporters have been described in the yeast Saccharomyces cerevisiae: plasma membrane transporter MCH5 (flavin import from the extracellular media), flavin-antiporter Flx1p (exchange of mitochondrial-matrix FMN by cytosolic FAD) and three putative FLC transporters (transport of FAD to the endoplasmic reticulum). Nonetheless, higher organisms have lost the ability to synthesize RF and they need to acquire it from the diet or to a lesser extent from its intestinal microflora’s production. For instance, in the case of human nutrition, the daily recommended allowance for RF is 1.1-1.3 mg/day, which is mainly supplied by the consumption of vegetables, milk, eggs and cereals. In mammals, flavins from diet are transformed into RF by non-specific hydrolases and absorbed at the small and large intestine via carrier-mediated processes. Recently, five flavin transporters have been discovered: mitochondrial riboflavin transporter (mRF), mithocondrial FAD transporter and three plasma membrane transporters (hRFTV1, hRFTV2 and hRFTV3) (Spaan et al., 2005; Yonezawa and Inui, 2013). These proteins have different subcellular and tissue-specific expression profiles, as well as functional and kinetic properties. To date, data on RF transport mechanisms in archaea, fungi and plants, are still quite limited on the literature. Riboflavin supply pathways are tightly regulated The transcriptional organization of RF biosynthesis genes is highly variable among bacteria species. In the simplest cases, all rib genes are encoded by a single operon, as observed in Bacillus or Mycobacteria genomes (García-Angulo, 2017). This seems to allow the coordinated expression of all rib genes. However, genes can also be dispersed along the chromosome, such as in E. coli. Hence, each gene can be independently regulated to satisfy cellular requirements. Additionally, many bacteria have redundant duplications or incomplete RF biosynthetic genes. Both the de novo biosynthesis and the uptake of riboflavin is energetically costly for the microorganisms, consequently both processes are tightly regulated. As in other anabolic pathways, the first step of the pathway (GTP hydrolysis) is rate-limiting and could be inhibited by the product RF, at least in B. subtilis (Hümbelin et al., 1999). Also, it has been reported the feedback inhibition of 6,7dimethyl-8-ribityllumazine synthase from Pichia guilliermondii by RF Introduction 21 (Logvinenko et al., 1973). Furthermore, apart from the regulation at the protein level, RF synthesis is regulated at the transcriptional and translational level. The most widely distributed regulatory factor is the FMN riboswitch (also termed RFN element), a sequence in the 5’-UTR that adopts a characteristic metabolitedependent secondary structure (García-Angulo, 2017). FMN binding to the riboswitch avoids the transcription and/or translation by forming a terminator and by hiding the ribosome binding site (Winkler et al., 2002). Therefore, this mechanisms couple RF needs with the expression of the genes required to produce it, avoiding waste of metabolic energy when intracellular flavin levels are sufficient. In some species (from bacteria to yeast and plants), a link between iron and RF metabolisms has been observed, since iron starvation induces rib gene expression and the replacement of iron-dependent enzymes by flavindependent enzymes (Crossley et al., 2007; Fedorovich et al., 1999; Sepúlveda Cisternas et al., 2018; Welkie and Miller, 1960). Nonetheless, regulatory mechanisms are neither present in all organisms or genes (García-Angulo, 2017; Vitreschak et al., 2002), or remain still uncharacterized. III. BIOSYNTHESIS OF THE FLAVIN COFACTORS Independently on the endogenous or exogenous origin of RF, all the organisms transform RF into FMN and FAD through two sequential reactions. Firstly, a riboflavin kinase (RFK, ATP:riboflavin 5´phosphotransferse, EC 2.7.1.26) phosphorylates RF into FMN by transferring the -phosphate group from ATP to the C5´ ribityl chain (Eqn. 1.1; reaction (10) in FIG I.6). After the phosphorylation step, an ATP:FMN adenylyltransferase (FMNAT, EC 2.7.7.2) activity converts FMN into FAD (Eqn. 1.2; reaction (11) in FIG I.6). RF + ATP → FMN + ADP (Eqn. 1.1) FMN + ATP → FAD + PPi (Eqn. 1.2) Although all organisms synthesize FMN and FAD from RF, they utilize different enzymes for that purpose. Ernesto Anoz Carbonell – Doctoral Thesis 22 In most of prokaryotes, both enzymatic reactions RFK and FMNAT are catalyzed by a single bifunctional enzyme, the type-I FAD synthase (FADS) (FIG I.7) (Herguedas et al., 2010; Krupa et al., 2003; Sebastián et al., 2017a; Wang et al., 2003, 2004). Its C-terminus RFK module is homologous to monofunctional eukaryotic RFKs. However, its N-terminus FMNAT module belongs to the nucleotidyl-transferase superfamily and differs from mammalian eukaryotic FMNATs. In addition to these type-I FADS bifunctional enzymes, some bacterial species own monofunctional enzymes with RFK and/or FMNAT activities (Solovieva et al., 1999; Yruela et al., 2010), and bifunctional enzymes different from type-I FADS. These latter enzymes show a N-terminus FMNAT module and a C-terminal module of unknown function, and are designated type-II FADS (Sebastián et al., 2019; Yruela et al., 2010). Figure I.7. Distribution of the activities related to FMN and FAD synthesis in the subcellular locations of different organisms. Enzymes in solid color have been purified and characterized biochemically. Enzymes shown with transparency have been predicted by bioinformatical analysis even though they have not been found experimentally. RFK (riboflavin kinase), FMNAT (FMN adenylyltransferase), FADasa (FAD hydrolase), FADpp (FAD pyrophosphorilase), FHy (FMN hydrolase). Figure modified from (Medina, 2012). Introduction 23 Contrary to prokaryotes, two independent monofunctional enzymes catalyzing RFK and FMNAT activities are present in yeast, animals and, in general, non-photosynthetic eukaryotes (FIG I.7) (Barile et al., 2000; Kasai et al., 1990; Merrill Jr. and McCormick, 1980; Santos et al., 2000). Monofunctional RFKs are ATP-dependent, and sequence analyses revealed a highly conserved PTAN motif and a glutamic residue that acts as catalytic base (Bauer et al., 2003b; Karthikeyan et al., 2003a, 2003b). Likewise, eukaryotic FMNATs are currently classified as member of the 3'-phosphoadenosine 5'-phosphosulfate (PAPS) reductaselike family, belonging to the adenine nucleotide α-hydroxilase-like superfamily (Huerta et al., 2009; Leulliot et al., 2010; Santos et al., 2000). Both proteins can present different isoforms with different sub-cellular location (mitochondria, nucleus, cytosol…) (Brizio et al., 2006; Giancaspero et al., 2013; Leone et al., 2018; Miccolis et al., 2012; Torchetti et al., 2010). In addition to the enzymes involved in FMN and FAD biosynthesis, eukaryotic cells harbor enzymes that hydrolyze FMN (FHy) in the cytosol and the intermembrane space (FIG I.7) (Barile et al., 1997; Fischer and Bacher, 2006; Fuchs et al., 1992; Granjeiro et al., 1997; Lee and Ford, 1997; Sandoval et al., 2008). Moreover, in S. cerevisiae, FAD pyrophosphatase (FADase) (EC 3.6.1.18) and FHy (EC 3.1.3.102) have also been reported in the mitochondrial inner membrane and in the matrix, respectively (Pallotta, 2011). In plants, the situation is even more intricate, since additionally to the monofunctional RFKs and FMNATs, two different bifunctional enzymes can be found (FIG I.7) (Yruela et al., 2010). One of them presents RFK activity in its Cterminus module, and hydrolase activity in its N-terminus one (Sandoval and Roje, 2005). The other bifunctional enzyme shows a N-terminus module similar to prokaryotic FMNATs, and a shorter C-terminal of unknown function (shorter than eukaryotic RFKs and without the PTAN motif) (Sandoval et al., 2008). These plant bifunctional enzymes have been named as FADSs-type II and plant-like FADSs, respectively (Yruela et al., 2010). The monofunctional RFKs and FMNATs have been found in multiple sub-cellular localizations (mainly mitochondria and cytosol but also in chloroplast) (Mitsuda et al., 1970). However, the location of bifunctional enzymes is still not clear, but they might be located in mitochondrion and/or chloroplast, since they have location sequences for these organelles (Giancaspero et al., 2009; Sandoval and Roje, 2005; Sandoval et al., 2008). Ernesto Anoz Carbonell – Doctoral Thesis 24 Archaea possess two monofunctional proteins that are phylogenetically different from the eukaryotic counterparts. The archaeal RFKs use exclusively CTP as substrate (instead of ATP in eukaryotic RFKs) (Ammelburg et al., 2007). The FMNATs belong to the nucleotidyl-transferase superfamily, although they do not show some of its characteristic motifs (Mashhadi et al., 2008), and either ATP or CTP can be used as substrates. Eukaryotic riboflavin kinases Monofunctional RFKs have been purified and characterized from yeasts (Santos et al., 2000) and diverse mammal tissues (Kasai et al., 1990; Kashchenko and Shavlovsky, 1976; McCormick et al., 1997; Schrecker and Kornberg, 1950; Yamada et al., 1990). Additionally, the crystallographic structures of the RFKs of Schizosaccharomyces pombe (SpRFK) and Homo sapiens (HsRFK) have been resolved (Bauer et al., 2003b; Karthikeyan et al., 2003a, 2003b). In general, RFKs are cytosolic enzymes of 18-24 kDa (Merrill and McCormick, 1980; Yamada et al., 1990) and their expression levels are particularly high in brain, placenta, bladder, liver and intestine of higher organisms. However, mitochondrial isoforms have also been identified in yeast and other eukaryotic organisms, including plants and animals (Barile et al., 2000; Giancaspero et al., 2009; Mitsuda et al., 1970; Sandoval et al., 2008; Santos et al., 2000). RFKs catalyze the phosphorylation of RF into FMN, using exclusively ATP as phosphate donor (Merrill and McCormick, 1980; Yamada et al., 1990). However, some RFKs show moderate activity using other nucleotides as substrates (as GTP) (Kashchenko and Shavlovsky, 1976). Apart from RF, eukaryotic RFKs can also phosphorylate a wide variety of RF-analogs (Kashchenko and Shavlovsky, 1976; Kearney, 1952; McCormick and Butler, 1962), being the determinants of specificity the length of 4-6 C-atoms of the N(10) substituent, the substitution in positions 7 and 8, and the absence of substituents in N(3) and N(5) (McCormick and Butler, 1962). The RFK activity also requires of the presence of a divalent cation, such as Zn2+ or Mg2+ (Nakano and McCormick, 1991). Nowadays, the tridimensional structures of the RFKs of S. pombe (PDB entries 1N05, 1N06, 1N07 and 1N08) (Bauer et al., 2003b) and H. sapiens (PDB entries 1NB0, 1NB9, 1P4M and 1Q9S) (Karthikeyan et al., 2003a, 2003b) are available, both free and/or in complex with their ligands (substrates and products of the RFK reaction). Both RFKs fold into a six-stranded antiparallel β-barrel core with Introduction 25 Greek key topology and some segments of α-helix located mainly at the Cterminal end (FIG I.8). The adenine nucleotide is located between a glycine-rich loop, a short 310-helix, and a reverse turn leading to a short β-strand that contains the most conserved PTAN sequence motif (Bauer et al., 2003b; Karthikeyan et al., 2003a). This motif is implicated in the coordination of the divalent cation and the binding of the triphosphate moiety of ATP. In contrast, the flavin binds in a pocket between the surface of the β-barrel and an α-helix, with the ribityl moiety extended to the nucleotide binding site. This folding represented a new fold in the kinase family with a novel nucleotide binding motif never previously described (Cheek et al., 2005; Karthikeyan et al., 2003a). Figure I.8. Structural properties of HsRFK. (A) Crystallographic structures of HsRFK in complex with FMN and ADP showing the FlapII loop in either open (PDB ID 1P4M, in purple) or closed conformation (PDB ID 1Q9S, in blue). FMN and ADP ligands are shown as CPK colored sticks colored with carbons in orange and yellow, respectively. (B) Topology of HsRFK, with α-helices as dark rectangles and β-strands as light arrows. Numbers indicate residue positions. Despite the overall structural identity among the RFK available structures, differences are found in the conformation of the loops constituting the adenine and flavin binding sites, termed FlapI and FlapII (FIG I.8A) (Karthikeyan et al., 2003a, 2003b). These variations regarding the apo-RFK and in complex with different combinations of ligands (ADP, RF:ADP and FMN:ADP) point to drastic conformational changes upon substrate binding and catalysis. A B Ernesto Anoz Carbonell – Doctoral Thesis 26 The structure of the apo-form has only been resolved for SpRFK (PDB entry 1N05). However, high sequence identity between eukaryotic RFKs suggest that SpRFK structure is a representative model of the whole family (Bauer et al., 2003b; Karthikeyan et al., 2003b). In this structure, FlapI collapses onto the β-barrel and blocks the adenine binding site, hampering ATP binding; whereas FlapII is disordered (not observed in the crystallographic structure), leaving the flavin pocket accessible to RF. Hence, this structure suggests that the RFK activity occurs through a sequential ordered bi-bi mechanism in which the preferred pathway is for RF to bind to the enzyme first followed by ATP. Binding of RF might induce a conformational change both in FlapII, that encloses the flavin into the protein, and in FlapI, opening the adenine binding site to ATP, as observed in the crystallographic structures of RFK in complex with ADP and RF/FMN. After the catalysis, the FMN is blocked by FlapII, whereas ADP is completely solvent exposed, favoring its release. Finally, both protein loops might acquire the initial conformation, in which the flavin binding site is open for FMN product dissociation. Eukaryotic FMN adenylyltransferases FMN adenylyltransferases (FMNAT, but also referred as FAD synthases, FADS) have been purified and characterized from yeast (Huerta et al., 2009; Wu et al., 1995), rat liver (Gomes and McCormick, 1983; McCormick et al., 1997), and Homo sapiens (Brizio et al., 2006; Galluccio et al., 2007). The crystallographic structure has been resolved for the FADS from the yeasts Saccharomyces cerevisiae (ScFMNAT) and Candida glabrata (CgFMNAT) (Huerta et al., 2009; Leulliot et al., 2010), but not for the human protein (hFADS). In yeast and fungi, FMNATs are monofunctional enzymes with a PAPS reductase-like domain, which has conserved motifs different from those of nucleotidyl-transferases (FMNAT modules of bifunctional prokaryotic FADS). Nonetheless, an additional domain is found at the N-terminus of FMNATs of higher eukaryotes, including H. sapiens. This module resembles molybdopterinbinding (MPTb) domain and harbors FAD pyrophosphatase activity (Cialabrini et al., 2013; Giancaspero et al., 2015a). Therefore, these latter enzymes are bifunctional FMNAT/FADases that can both synthesize and hydrolyze FAD in response to the physiological state of the cell. Introduction 27 Eukaryotic FMNATs are enzymes of 30-40 kDa (monofunctional fungi FMNATs) or 50-60 kDa (bifunctional FMNAT/FADases), and their subcellular location is still controversial since both FMNAT and FADase activities are detected in almost all cellular compartments (Bafunno et al., 2004; Giancaspero et al., 2013; Torchetti et al., 2010, 2011). Furthermore, these enzymes are constitutively expressed in all tissues, but their expression levels are particularly high in brain, placenta, bladder, liver and gastrointestinal tract of higher organisms (Expression Atlas, http://www.ebi.ac.uk/gxa) (Papatheodorou et al., 2018). FMNATs catalyze the adenylylation of FMN into FAD, using exclusively ATP as adenylyl donor (Bowers-Komro et al., 1989; Schrecker and Kornberg, 1950). Hence, no enzymatic activity is detected when using AMP, ADP, dATP, TTP, UTP, CTP or GTP as nucleotide donors. Nonetheless, although these enzymes show higher enzymatic activity with FMN as substrate, a wide variety of FMNanalogs can be adenylated. The determinants of specificity are the absence of substitutions in the pyrimidine portion of the isoalloxazine ring (CO and NH groups at positions 2 and 3), and the length of 5-6 C-atoms of the N(10) substituent (Bowers-Komro et al., 1989). Additionally, the FMNAT activity has a strict requirement of Mg2+, although other divalent cations as Co2+, Mn2+, Ca2+ and Zn2+ can be used with less efficiency (Bowers-Komro et al., 1989; McCormick et al., 1997; Torchetti et al., 2011). The FMNAT activity follows an sequential ordered bi-bi mechanism in which ATP binds first to the enzyme followed by FMN, and product pyrophosphate (PPi) is released first followed by the release of FAD (Giancaspero et al., 2015a; Huerta et al., 2009; Oka and McCormick, 1987; Yamada et al., 1990). Both products of the forward reaction (FAD and PPi), as well as nucleotides other than ATP, act as inhibitors of the FMNAT activity (Bowers-Komro et al., 1989; Torchetti et al., 2011). Additionally, FAD release may represent the rate-limiting step of the whole catalytic cycle and is consistent with the proposed role of eukaryotic FMNATs as FAD chaperones: these enzymes do not only synthesize FAD but also deliver it to the recipient apo-flavoproteins through protein-protein interactions (Giancaspero et al., 2015b; Torchetti et al., 2011). Structural information of eukaryotic FMNATs is limited to ScFMNAT in complex with FAD (PDB entry 2WSI) (Leulliot et al., 2010), and CgFMNAT (PDB entries 3FWK, 3G6K, 3G59 and 3G5A), free and in complex with the substrates or products of the FMNAT reaction (Huerta et al., 2013). In both FMNATs, the Ernesto Anoz Carbonell – Doctoral Thesis 28 PAPS reductase-like domain has an α/β fold with a twisted six-stranded β-sheet sandwiched by α-helices, organized into a non-canonical Rossmann-fold (Dym and Eisenberg, 2001; Rossmann et al., 1974). The ATP nucleotide binds in a crevice formed at the β-sheet, with its βγ-phosphate tail positioned in an anion-binding pocket near the N-terminus of α3 (Huerta et al., 2009). The flavin binding site is located on the same side of the central β-sheet than the ATP, in a groove created by the loop connecting β5 to β6, helix α5 and the β-sheet. Within this site, the adenylyl (of ATP or FAD) and isoalloxazine moieties (of FMN or FAD) are packed against each other and face inwards, and the phosphoribityl moiety faces to the solvent (Huerta et al., 2009; Leulliot et al., 2010). Therefore, FAD adopts a bent conformation, different from those observed in other FMNor FAD-flavoproteins (Dym and Eisenberg, 2001), emphasizing this flavinbinding mode as a characteristic of eukaryotic FMNATs. Furthermore, the Mg2+ cation in the ternary complexes stabilize the phosphate groups of the substrates and products, positioning them in an optimal configuration for the adenylyl transfer reaction. Figure I.9. (A) Crystallographic structure of CgFMNAT in complex with the products FAD, PPi and Mg2+ (PDB ID 3G6K). Secondary structure elements are labeled, together with structural motifs involved in substrate binding and catalysis. (B) Comparison of CgFMNAT with the closely related APS reductase of P. aeruginosa. Equivalent structural motifs are colored similarly. The characteristic PP-loop motifs are shown in red. Regions that deviate from the typical Rossmann-fold topology are shown in magenta. FAD and APS are shown as CPK colored sticks colored with carbons in orange and yellow, respectively. Figure from (Huerta et al., 2009) A B Introduction 29 Altogether, the crystallographic structures are consistent with the proposed catalytic mechanism for eukaryotic FMNATs (Giancaspero et al., 2015b; Huerta et al., 2009; Leulliot et al., 2010). ATP binds preferentially to the apo-enzyme, as its binding site would be partially blocked by FMN otherwise. Additionally, the presence of ATP may induce the binding and proper allocation of FMN through the interaction of its isoalloxazine ring with the adenosine moiety of ATP. Although the phosphoribityl tails of FMN is highly flexible and can adopt multiple conformations (as observed in the crystallographic structure of the substrate tertiary complex), this part of the FMN substrate would move close the α-phosphate of ATP during the adenylylation reaction. The cleavage of the αβphosphodiester bond is facilitated by the coordination of the Mg2+ ion. After the catalysis, subtle structural rearrangements are observed: the PPi group is in the same position as the βγ-phosphates of the ATP interacting with the same set of protein residues and the Mg2+ cation. Dissociation of PPi induces a conformational change in the β4-β5 loop, allowing release of the FAD reaction product. Human FAD synthase In Homo Sapiens, hFADS is encoded by the gene FLAD1, located on chromosome 1 at 1q21.3. Up to six isoforms have been identified as a result of alternative splicing of the same gene (FIG I.10) (UniProtKB database, https://www.uniprot.org - Q8NFF5). The three longest isoforms (hFADS1, 2 and 3) contain both FMNAT and FADase domains (Brizio et al., 2006; Galluccio et al., 2007; Torchetti et al., 2011). However, two isoforms (hFADS3 and 4) lack the PAPS reductase module, meanwhile the last isoform discovered (hFADS6) lacks the FADase module (Leone et al., 2018; Olsen et al., 2016). These isoforms also differ in subcellular location as well as in catalytic efficiencies and ion requirements. For example, hFADS1 has an additional N-terminal putative mitochondrial targeting peptide, and therefore localizes in the mitochondria (Torchetti et al., 2010); whereas hFADS2 is a cytosolic isoform. In addition, FMNAT activity has also been detected in the cellular nucleus (Giancaspero et al., 2013), but the nuclear isoform has not been characterized to date. Altogether, subcellular distribution of hFADS isoforms, together with flavin transporters, might contribute to create different flavin cofactor pools, constituting a flavin network presumably involved in the regulation of metabolism and homeostasis (Barile et al., 2016; Giancaspero et al., 2015b). Ernesto Anoz Carbonell – Doctoral Thesis 36 The CaFADS C-terminal module (residues 187-338) folds in a globular domain formed by a β-barrel with six antiparallel β-strands (β1cβ6c), a long -helix (α1c) and seven loops connecting them (FIG I.12) (Herguedas et al., 2010). This domain displays sequence and structural homology to monofunctional RFKs and other FADS RFK-modules (Sebastián et al., 2017a; Wang et al., 2003), but CaFADS owns a 12 residues insertion in L3c (Leu232-Val246) that is only observed in corynebacteria and mycobacteria (Frago et al., 2008; Herguedas et al., 2010). Other differences include (1) the conformation of the loops L1c or FlapI (V193-T208) and L4c or FlapII (P258-E263), both involved in the stabilization of nucleotides and flavins (Herguedas et al., 2015; Karthikeyan et al., 2003b); (2) the position of the P207 and T208 residues at the 207-PTAN-210 motif (Frago et al., 2008), and (3) the presence of a 15 residues loop (L7c, A324-S338) in two different conformations (FIG I.13), that is absent in eukaryotic RFKs, as HsRFK or SpRFK. Both RFK and FMNAT modules interact with each other through H-bonds and the formation of a hydrophobic core that provides rigidity to their interface. Loop L2c, which is implicated in these inter-domain interactions, is an α-helix in eukaryotic RFKs (Herguedas et al., 2010). Figure I.13. Detailed cartoon representation of the RFK (A) and FMNAT (B) modules of CaFADS (PDB ID 2X0K). RFK and FMNAT modules are represented in orange and green, respectively, with their loops in purple. Secondary structure elements are labeled. In (A), the alternative conformation of loop L7c is shown in light blue. Figure adapted from (Herguedas et al., 2010). Introduction 37 Conformational rearrangements upon CaFADS RFK catalytic cycle CaFADS crystallographic structures are available for the full-length protein free of ligands (apo form, PDB ID ), as well as for the RFK module in complex with adenine and flavin nucleotide ligands (holo forms) (FIG I.14A) (Herguedas et al., 2015): binary RFK:ADP-Mg2+ complex and ternary RFK:FMN:ADP-Mg2+ complex (PDB entries 5A88 and 5A89, respectively). This structural information, together with computational analysis and kinetic and binding data, suggest that CaFADS suffers conformational changes during ligand binding and catalysis. The structure of the RFK module in the apo form of the CaFADS significantly differs from those found for the binary RFK:ADP-Mg2+ and the ternary RFK:FMN:ADP-Mg2+ complexes (Herguedas et al., 2015). Both complexes exhibit important structural rearrangements that affect the conformation of the PTAN motif, as well as most of the protein loops with the only exception of L2c. In the apo-protein structure, the adenine nucleotide binding site of the RFK module is closed by the interaction of L1c-FlapI, L3c and L5c loops. However, in the protein-ligand, the adenine nucleotide cavity is opened by the displacement of L1c to α1c, while L3c and L5c move to the opposite direction (FIG I.14C). Molecular dynamics (MD) simulations of the free RFK module show that these loops establish an open-closed equilibrium for the adenine nucleotide binding site, in which the opening is favored by the presence of either flavin or adenine nucleotides (Herguedas et al., 2015). Regarding the flavin binding site of the RFK module, L4c-FlapII is disarranged in the binary complex and as a consequence the flavin binding cavity is open and wide (FIG I.14B). However, FMN binding induces a change in the size, shape and solvent accessibility of the cavity (Herguedas et al., 2015). The consensus PTAN motif, implicated in the stabilization of ATP phosphates and Mg2+ cofactor, also experiences different conformational changes upon ligand binding. In the free structure, the side chain of T208 is occupying the position where the adenine ring is in the binary and ternary complexes (FIG I.14E). Therefore, the binding of the adenine nucleotides should necessarily induce a conformational change that results in the opening of their binding site. MD simulations show that the presence of either the flavin or the adenine nucleotides is not sufficient to induce the Thr208 conformational change. Moreover, these simulations suggest that the coordination of its side chain with ADP/ATP through the Mg2+ ion is the most probable determinant for the achievement of the catalytically competent geometry of the PTAN motif. Ernesto Anoz Carbonell – Doctoral Thesis 38 Figure I.14. Superposition of free CaFADS RFK domain (apo form, in light blue), the binary complex with ADP (in light pink) and the tertiary complex with ADP and FMN (in brown) (PDB entries 2X0K, 5A88 and 5A89, respectively). (A) Overall view of the structural alignment, with the main differences highlighted with squares. (B) Zoom in the flavin binding site, showing the displacement of loops L4c-FlapII and L6c. (C) Detail of the adenine nucleotide binding site, showing the conformational changes of loops L1c, L3c and L5c. (D) Zoom into C-terminus of the RFK module showing the displacement of L7c. (E) Conformation of the PTAN motif in the three models. The conformational change of Pro207 and Thr208 can be appreciated. Conformational changes are indicated with yellow arrows. Figure from (Herguedas et al., 2015). Introduction 39 CaFADS can stabilize quaternary assemblies in the form of a dimer of trimers The structural analysis of the crystal structure of CaFADS predicted an A3B3 hexameric oligomer (being A and B the two chains of the asymmetric unit of the CaFADS crystal) (FIG I.15), constituted by a dimer of trimers (A1-A2-A3 and B1-B2B3, one on top of the other) (Herguedas et al., 2010; Serrano et al., 2015). Formation of these in silico predicted assemblies was imaged at the single-molecule level through atomic force microscopy techniques (AFM) and detected in vivo in C. ammoniagenes protein extracts (Marcuello et al., 2013). Figure I.15. Oligomeric state of CaFADS. (A) Space-filling representation of the dimer of trimers. Left, view through the 3-fold crystallographic axis. Right, lateral view of the hexamer. (B) Surface electrostatic potential of the CaFADS hexamer along the interaction surfaces of the two trimers (negative, positive and uncharged residues are shown in red, blue and white respectively). Arrows indicate the positions with electrostatic contacts between trimers for the formation of the hexamer. Figure from (Herguedas et al., 2010). Ernesto Anoz Carbonell – Doctoral Thesis 40 Trimers are stabilized through contacts of different nature between 20 side chains (18 H-bonds, 18 hydrophobic contacts and 3 salt bridges) (FIGS I.16A) (Herguedas et al., 2010; Serrano et al., 2015). The interface between trimers shows electrostatic complementarity (FIG I.15B). Moreover, 14 H-bonds, 4 salt bridges and 12 hydrophobic contacts strengthen the hexamer (Serrano et al., 2017). Additionally, sulphate ions play an crucial role the by stablishing additional Hbonds and salt bridges (Herguedas et al., 2010). Within each trimer, the protomers are organized in a head-to-tail configuration (FIGS I.15A and I.16B); thereby approaching the FMNAT module of one protomer to the RFK module of the neighboring protomer. In this disposition, both actives sites come in close contact which modifies both the active site environment and its solvent and ligand accessibility. The RFK cavity appears partially closed by α2n and L4n of the neighboring protomer (FMNAT module) (FIG I.16D), whereas the FMNAT is also partially closed by residues at α1c and L6c (RFK module) (FIG I.16C). CaFADS oligomerization is a dynamic process occurring upon ligand binding and catalysis (Herguedas et al., 2010; Marcuello et al., 2013; Serrano et al., 2017). Some of the substrates and products of the RFK (RF, ATP:Mg2+ and FMN:ADP:Mg2+) and FMNAT activities (FAD) induce the oligomerization of CaFADS, yielding different quaternary assemblies (Marcuello et al., 2013). Hence, the assembly of the organized dimer of trimers might occur/take place through a series of oligomerization intermediates, such as dimers, trimers and amorphous hexamers. The oligomerization process is tightly regulated by the substrates and products of CaFADS, what points to be a mechanism to regulate the biosynthesis of flavin cofactors and thus maintain the flavin and flavoproteome homeostasis. Initially, it was proposed that the dimer of trimers was implicated in the channeling of the FMN product from the RFK active site to the FMNAT one, to be subsequently transformed into FAD (Herguedas et al., 2010; Marcuello et al., 2013). However, MD simulations suggest that the RFK module prevents catalysis at the transferase site by modulating negatively the FMN binding to the FMNAT module (Lans et al., 2018). Nevertheless, it remains for future investigations to describe whether this might be a general behavior for prokaryotic FADSs or, on the contrary, a particular feature for the FADSs from Corynebacterium and, also probably, Mycobacterium species. Introduction 41 Figure I.16. (A) Detail of the contacts between the RFK and FMNAT modules of different protomers within a trimer, including the hydrophobic residues, H-bonds, and salt bridges involved in trimer stabilization. (B) Cartoon representation of one of the trimers conforming the hexamer. RFK and FMNAT modules are shown in orange and green, respectively, and the molecular surface of each monomer is displayed in different colors. Flavin ligands are represented as yellow spheres. (C) Detail of the active-site surface at the FMNAT module showing how α1c and L1c of the RFK module of the neighboring protomer close the binding site. (D) Detail of the active site of the RFK module. α2n and L4n of the FMNAT module of the neighboring protomer close the binding site. FMN and FAD are represented as violet and yellow sticks respectively. Figure from (Herguedas et al., 2010) Ernesto Anoz Carbonell – Doctoral Thesis 42 Prokaryotic FADS as potential antimicrobial targets The rise of antimicrobial resistance (AMR) threatens the effective prevention and treatment of an ever-increasing range of infections caused by microorganisms. Hence, the World Health Organization pointed AMR as a public health problem of growing concern (World Health Organization, 2014). Examples of microorganisms included in the list of antibiotic-resistant pathogens for which is urgent to develop new antimicrobial drugs include methicillinresistant Staphylococcus aureus, Streptococcus pneumoniae resistant to both penicillin and macrolides, vancomycin-resistant Enterococcus faecalis, and multidrug and extensively drug-resistant strains of Mycobacterium tuberculosis (World Health Organization, 2014, 2017). The only long-term solution to overcome this global problem is the continuous research and development of new antimicrobials, focusing on the discovery of new druggable targets. Although there are approximately 450 essential proteins conserved along prokaryotes (Christen et al., 2011; Hutchison et al., 2016), most of the commercially available antibiotics are directed to a limited set of antimicrobial targets: the cell wall biogenesis, the prokaryotic ribosome or the DNA replication and transcription (Lewis, 2013). Hence, the rest of essential cellular processes remain nearly unexplored as antimicrobial targets. Until relatively recently, flavin metabolism had not been considered a potential target of antimicrobials. First compounds directed to this pathway were compounds inhibiting lumazine synthase and/or riboflavin synthase (Cushman et al., 1999; Morgunova et al., 2005; Ritsert et al., 1995). Additionally, in 2009 and 2015, roseoflavin and ribocyl C were described as the first compounds acting indirectly on riboflavin biosynthesis through the binding to the FMN riboswitch and repression of the expression of the rib operon (Howe et al., 2015; Lee et al., 2009). Furthermore, the prokaryotic FADS was also proposed as an antimicrobial target (Serrano et al., 2013b) and afterwards inhibitory compounds endowed with antimicrobial activity and targeting FADS were further discovered (Sebastián et al., 2018b). Prokaryotic bifunctional FADSs are key proteins in flavin homeostasis and in the maintenance of flavoprotein and flavoenzyme function (Frago et al., 2008). Hence, the inhibition of their enzymatic activities leads into FMN and FAD deficiency, and consequently into the accumulation of the apo forms of numerous flavoproteins, which are unable to carry out their expected functions in the cellular metabolism and other essential processes (Serrano et al., 2013b). Indeed, Introduction 43 FADS are highly conserved among prokaryotes and have been described as essential for numerous microorganisms including Bacillus subtilis (Kobayashi et al., 2003), Escherichia coli (Goodall et al., 2018), M. tuberculosis (Griffin et al., 2011; Minato et al., 2019; Sassetti et al., 2003), S. pneumoniae (Liu et al., 2017), etc. This crucial role in cellular metabolism, together with the significant biochemical and structural differences with their eukaryotic counterparts (monofunctional RFKs and FMNATs) (Karthikeyan et al., 2003a; Miccolis et al., 2014), converts the FADS in an attractive potential drug target for the development of antimicrobials (Serrano et al., 2013b). Thus, halting the production of FMN and FAD by inhibiting the FADSs would prevent, from the very beginning, all pathways that involve flavoproteins and flavoenzymes. In principle, riboflavin biosynthesis inhibitors (targeting both RF synthase or lumazine synthase) would be expected to trigger similar effects than FADS inhibitors. However, RF biosynthesis is conditionally essential since some microorganisms (depending also on growth conditions) have other mechanisms, in addition to the de novo synthesis, to obtain RF (García-Angulo, 2017; GutiérrezPreciado et al., 2015; Matern et al., 2016). By contrast, in most bacteria (except for prokaryotes harboring monofunctional RFKs and/or FMNATs), the only pathway for FMN and FAD biosynthesis occurs through bifunctional FADS. Druggability of prokaryotic FADSs is also supported by their strict functionstructure relationship. Mutations affecting the active site (most common binding site of inhibitory compounds) would probably alter the enzymatic activity. Therefore, FADS inhibitors might be less prone to generate resistances, since mutations conferring resistance also might compromise the protein function and thus the cellular viability. Additionally, the availability of structures of several bacterial FADSs facilitates the design of both inhibitory drugs and activity assays (Herguedas et al., 2010; Sebastián et al., 2018b; Wang et al., 2003). In addition, the biochemical characterization of several prokaryotic FADS (including TmFADS CaFADS, SpFADS, Lm1FADS and Lm2FADS) envisages significant species-specific traits in their enzymatic activity and regulatory strategies (Herguedas et al., 2015; Sebastián et al., 2017c, 2017a, 2019; Wang et al., 2004). Such differences might provide with a framework to design speciesspecific antimicrobial compounds. Ernesto Anoz Carbonell – Doctoral Thesis 44 IV. FLAVOPROTEINS Flavin cofactors show unique spectroscopic and redox properties, and thus flavoproteins, proteins harboring flavin as cofactors, have marveled generations of enzymologist along the time. Among the redox cofactors, flavins can participate in both one-electron and two-electron transfer processes. Other cofactors usually catalyze exclusively one- (iron-sulfur clusters, heme groups, quinones…) or two-electron transfer processes (almost exclusively nicotinamide nucleotides). Therefore, flavoenzymes are indispensable mediators between these processes, as in the case of the well-known mitochondrial and chloroplast electron-transport chains, or beta-oxidation of fatty acids. Additionally, flavins are capable of reacting with oxygen (dioxygen activation processes). This catalytic versatility encompasses many established reaction classes, including dehydrogenation, oxidation, monooxygenation, epoxidation, Baeyer-Villiger oxidation, decarboxylation, halogenation and reduction (Leys and Scrutton, 2016; Massey, 2000; Walsh and Wencewicz, 2013). Furthermore, the complexity of flavin-catalyzed reactions is further increased when they join forces with other redox-active cofactors, such as iron–sulfur clusters ([2Fe–2S], [3Fe–4S] and⁄or [4Fe–4S]), heme, molybdopterin, or thiaminediphosphate. However, the role of flavins in flavoproteins is not only limited to redox processes, as ∼ 10% of flavin-dependent enzymes catalyze non-redox reactions (Macheroux et al., 2011). Hence, the flavin cofactor is also widely used as signaling and sensing molecule in biological processes (FIG I.17). Examples would include signal transduction upon apoptosis (Natarajan and Becker, 2012; Susin et al., 1999), embryonic development (Murty and Adiga, 1982), chromatin remodeling (Forneris et al., 2005), nucleotide synthesis (Myllykallio et al., 2002), tRNA methylation (Nishimasu et al., 2009), protein folding (Gross et al., 2004), defense against oxidative stress (Natarajan and Becker, 2012), among others. Flavoproteins are also implicated in detoxification of aromatic compounds (xenobiotic metabolism) (Dagley, 1987) and in light-dependent processes as luciferase light-emission (Meighen, 1991), plant phototropism (Briggs et al., 2001) or DNA reparation (Jorns et al., 1987), in which flavin is capable of receive photons. Introduction 45 Figure I.17. Structure of the isoalloxazine ring of flavin cofactors and some of the biological processes they are involved in. Figure from (Joosten and van Berkel, 2007) In a recent study, Macheroux et al. (2011) analyzed FMN and FAD-dependent proteins in 22 genomes, including archaea, eubacteria, protozoa and eukaryotic organisms. They reported 374 classes of flavin-dependent proteins (276 fully classified and 98 with no or partial classification and without enzymatic activity), entailing approximately the 2% of all the proteins. As expected for a redox-active cofactor, most of the 276 flavoenzymes are oxidoreductases (91%), whereas other enzymatic functions are underrepresented, such as transferases (4.3%), lyases (2.9%), isomerases (1.4%) or ligases (0.4%) (Macheroux et al., 2011), although the exact percentages varies among organisms. The majority of flavoenzymes bind FAD (75%) rather than FMN (25%). Moreover, some eukaryotic flavoproteomes are biased towards FAD-dependent enzymes, as in the case of human proteome (84% of FAD-dependent proteins) (Lienhart et al., 2013; Wegrzyn et al., 2019). Additionally, in most of the cases the cofactor is noncovalently bound, although covalent attachment is also observed (approximately 10% of flavoproteins), especially for FAD-flavoproteins. FMN is preferentially bound by both the 6and/or 8α-position of the isoalloxazine ring with a nucleophilic side chain of the protein (Singer and McIntire, 1984). Covalent binding might increase protein stability, ensure cofactor retention and⁄or induce a more positive redox Ernesto Anoz Carbonell – Doctoral Thesis 52 This catalytic cycle is inhibited by dicoumarol (and other coumarin-based molecules), a hydroxycoumarin with strong anticoagulant activity that acts as potent competitive inhibitor by blocking the NAD(P)H access to the active site and partially occupying its binding site (FIG I.18B and I.18C) (Nolan et al., 2007). However, inhibitor binding causes only subtle structural rearrangements in the conformation of NQO1 (primarily located at the surface of the catalytic site) (Asher et al., 2006). In addition, other selective inhibitors of NQO1 have been further identified as anti-cancer drugs due to the association of its overexpression with cancer progression and development (Nolan et al., 2007, 2009, 2010; Scott et al., 2011). Functions of NQO1 NQO1 catalyzes different reactions with antioxidant, detoxification and metabolic roles, such as the two-electron reduction of quinones to their hydroquinone form, thus avoiding the formation of highly reactive and cytotoxic semiquinones (Anusevičius et al., 2002; Hosoda et al., 1974; Lind et al., 1982), reduction of coenzyme Q10 and vitamin E to their reduced state (antioxidant form) (Beyer et al., 1996; Siegel et al., 1997), scavenging of superoxide anions (Siegel et al., 2004), reduction of catecholamines and vitamin K (Ingram et al., 2013) and modulation of the NADH/NAD+ redox balance (Ross and Siegel, 2018; Siegel et al., 2018). Moreover, this chemoprotective role is further supported by its activity towards a wide variety of reactive species (quinones, naphthoquinones, quinone-imines, azo and nitro compounds, hexavalent chromium compounds …) (Misevičiene et al., 2006; Newsome et al., 2007; Šarlauskas et al., 1997; Talalay and Dinkova-Kostova, 2004) and its induction under a variety of cellular stress responses including oxidative stress (Prochaska et al., 1992; Ross and Siegel, 2018). Additionally, NQO1 is also implicated in the bioreductive activation of several compounds, such as the antitumoral drugs mitomycin C and β-lapachone (Pink et al., 2000; Siegel et al., 1992). Although NQO1 has been traditionally considered as a cellular reductase, new non-enzymatic roles have emerged recently where this enzyme binds to and regulates the stability of many biologically relevant molecules. For example, NQO1 physically interacts to 20S proteasome (Moscovitz et al., 2012) and protects some intrinsically-disordered proteins from degradation such as p33, p53, p63, p73, c-Fos, C/EBP, PGC-1, Hif-1α or ornithine decarboxylase (Adamovich et al., 2013; Asher et al., 2001). However, it is still uncertain if this protection from Introduction 53 degradosome degradation is through the association with the proteins undergoing degradation and/or the direct interaction with the 20S proteasome. Additionally, it has been recently found that NQO1 can bind to several mRNAs including SERPINA1 mRNA (encodes the serine protease inhibitor α-1antitrypsin) increasing its translation (Di Francesco et al., 2016). These interactions seem to be modulated by NAD(P)H/NAD(P)+ ratio, although the enzymatic activity of NQO1 may not be required (Asher et al., 2005; MedinaCarmona et al., 2017). Altogether, this suggest that the intracellular pyridine nucleotide balance regulates these interactions with its protein and RNA partners, pointing to the role of NQO1 as a redox molecular switch to modulate the downstream cellular functions (Siegel et al., 2018). Polymorphism in NQO1 To date, more than 250 single-nucleotide polymorphisms (SNPs) in the NQO1 gene have been described (according to the dbSNP database, http://www.ncbi.nlm.nih.gov/SNP), including the two most prevalent variants in the human population, NQO1*2 and NQO1*3. The NQO1*2 polymorphism (rs1800566) is a missense SNP which replaces the cytosine 609 of the cDNA for a thymine (c.C609T), resulting in a proline-to-serine substitution at residue 187 (p.P187S). Its allelic frequency is higher in Asian populations (approximately 50%) compared with Caucasians (approximately 25%) (Gaedigk et al., 1998). Additionally, this polymorphism is associated with higher cancer risk (particularly, gastrointestinal, liver, colorectal, lung, breast and thyroid cancers) (Fagerholm et al., 2008; Lee et al., 2013; Long et al., 2002; Peng et al., 2014; Stoehr et al., 2012; Xu et al., 2001), increased susceptibility to poisoning by benzene and other carcinogenic agents (Bauer et al., 2003a; Nebert et al., 2002; Rothman et al., 1997), and reduced NQO1-dependent activation of antitumoral prodrugs (Traver et al., 1992). The NQO1 P187 residue is located in a loop close to the surface of the protein at the N-terminal domain (FIG I.18C), far away from the FADand NAD(P)Hbinding sites. However, the p.P187S substitution produces a dramatic reduction on both NQO1 enzymatic activity and FAD binding affinity (Lienhart et al., 2014; Medina-Carmona et al., 2017). In addition, this polymorphism has reduced intracellular stability of the protein and is promptly ubiquitinated and degraded by the 20S/60S proteasome (Moscovitz et al., 2012; Siegel et al., 2001). Although Ernesto Anoz Carbonell – Doctoral Thesis 54 the overall crystallographic structure of p.P187S is identical to the wild-type NQO1 structure, this polymorphism causes the presence of unfolded states in solution and a significant increase in local dynamics of two functionally and structurally distant sites: the N-terminal domain associated with enzyme inactivation, and the C-terminal domain linked with its enhanced proteasomal degradation (Lienhart et al., 2014; Medina-Carmona et al., 2017). The NQO1*3 polymorphism (rs1131341) represents a cytosine for thymine change at position 465 of the NQO1 cDNA (c.C465T), and results in an arginineto-tryptophan substitution at residue 139 (p.R139W). It has an estimated global allele frequency of 2%, reaching up to 7% in Spanish Iberian population (Gaedigk et al., 1998). This polymorphism has also been associated with cancer development (specifically childhood acute lymphoblastic leukaemia and urinary bladder neoplasm) (Sanyal et al., 2007; Stanulla et al., 2007) and reduced responsiveness to several antitumoral drugs (Hu et al., 1996; Pan et al., 1995), although this NQO1 variant is less well studied than p.P187S. These deleterious effects might be result of an enhanced alternative splicing of the NQO1 premRNA causing skipping of exon 4, and therefore a decrease of the expression of the active full-length enzyme (Gasdaska et al., 1995; Pan et al., 2002, 1995). NQO1 R139 residue forms part of a solvent-exposed loop located in the Nterminal domain of the enzyme, without interacting with the active site residues. The p.R139W variant adopts the same tertiary structure than the wild-type as observed by X-ray crystallography and nuclear magnetic resonance spectroscopy (Lienhart et al., 2017). However, this substitution causes subtle differences in enzyme kinetic properties (higher Km values and/or lower kcat values for quinone substrates), reduced in vitro and in vivo stability of the protein, and minor stabilization of protein partners such as p53 (Lienhart et al., 2017; Medina Carmona et al., 2016; Megarity and Timson, 2019; Pey et al., 2014) Objectives Objectives 57 This thesis is focused on two interrelated topics, both associated with flavin cofactors and their crucial biological role in different organisms. The first aims to further delve into the molecular mechanisms of eukaryotic flavoenzymes of biological interest (HsRFK, hFADS and hNQO1), associated with flavin and redox intracellular homeostasis. The second objective deals with the identification and evaluation of potential inhibitors of the FMNAT activity of prokaryotic FADSs, backing our proposal of these enzymes as potential antimicrobial targets. With these general aims, in this study we tackle the following specific goals: − To explore specific features of the mechanisms underlying the regulation of human RFK activity, in comparison with previous structure-function studies of RFK modules of prokaryotic FADSs. − To propose a plausible catalytic cycle for HsRFK, based on the kinetic and thermodynamic data and in the context of the available crystallographic structures. − To elucidate the mechanism underlying the higher enzymatic activity of the D238E hFADS “supermutant” in comparison with the WT protein. − To confirm the catalytic mechanism proposed for human FAD synthase through its kinetic and thermodynamic characterization. − To provide an integrated perspective on the hNQO1 catalytic mechanism by studying both the reductive and oxidative half-reactions in the presteady state. − To evaluate the contribution of quantum tunneling, conformational dynamics and reorganizations in the hydride transfer process from the NADPH coenzyme to hNQO1. − To characterize the effect of hit molecules envisaged to interact with CaFADS by either high-throughput-screening (HTS) or in silico screening on the RFK and FMNAT activities of CaFADS, and to determine the binding affinities of the best hits inhibiting its FMNAT activity. − To evaluate the antimicrobial effect of HTS hits against several bacterial species, as well as the cytotoxic effect in eukaryotic cell lines. Publication I Human riboflavin kinase: species-specific traits in the biosynthesis of the FMN cofactor Ernesto Anoz Carbonell – Doctoral Thesis 60 Ernesto Anoz Carbonell contributions: Experimental work: protein purification, spectroscopic characterization, steady-state enzymatic activity measurements, pre-steady-state kinetics through stopped-flow spectroscopy, and isothermal titration calorimetry experiments. Data analysis. Manuscript writing, review and editing The FASEB Journal. 2020;34:10871–10886. | 10871 wileyonlinelibrary.com/journal/fsb2 Received: 9 March 2020 | Revised: 26 May 2020 | Accepted: 5 June 2020 DOI: 10.1096/fj.202000566R RESEARCH ARTICLE Human riboflavin kinase: Species-specific traits in the biosynthesis of the FMN cofactor ErnestoAnoz-Carbonell1,2 | MaribelRivero1 | VictorPolo2,3 | AdriánVelázquez-Campoy1,2,4,5,6 | MilagrosMedina1,2 © 2020 Federation of American Societies for Experimental Biology Abbreviations: ANP, adenine nucleotide (ATP or ADP); Ca, Corynebacterium ammoniagenes; CD, circular dichroism; FADS, FAD synthase; FLV, flavinic nucleotide (FMN or FAD); FMNAT, ATP:FMN adenylyltransferase; GST, glutathione S-transferase; HPLC, high-performance liquid chromatography; Hs, Homo sapiens; ITC, isothermal titration calorimetry; Ka, association constant; kcat, catalytic constant; Kd, dissociation constant; Ki, Inhibition constant; KM, Michaelis constant; kobs, observed rate constant; koff, dissociation rate constant; kon, association rate constant; PIPES, 1,4-piperazine diethane sulfonic acid; rpm, revolutions per minute; RF, riboflavin, Vitamin B2; RFK, ATP:riboflavin kinase; SDS-PAGE, sodium dodecyl sulfatepolyacrylamide gel electrophoresis; Sp, Schizosaccharomyces pombe; Spn, Streptococcus pneumonia; α, heterotropic interaction constant; ε, extinction coefficient. 1Departamento de Bioquímica y Biología Molecular y Celular, Facultad de Ciencias, Universidad de Zaragoza, Zaragoza, Spain 2Instituto de Biocomputación y Física de Sistemas Complejos (GBsC-CSIC and BIFI-IQFR Joint Units), Universidad de Zaragoza, Zaragoza, Spain 3Departamento de Química Física, Universidad de Zaragoza, Zaragoza, Spain 4Fundación ARAID, Diputación General de Aragón, Zaragoza, Spain 5Aragon Institute for Health Research (IIS Aragon), Zaragoza, Spain 6Biomedical Research Networking Centre for Liver and Digestive Diseases (CIBERehd), Madrid, Spain Correspondence Milagros Medina, Departamento de Bioquímica y Biología Molecular y Celular, Facultad de Ciencias, Pedro Cerbuna 12, Universidad de Zaragoza, 50009-Zaragoza, Spain. Email: [email protected] Funding information Spanish Ministry of Economy, Industry and Competitiveness, Grant/Award Number: BIO2016-75183-P AEI/FEDER; Spanish Ministry of Science and Innovation, Grant/ Award Number: PID2019-103901GB-I00 AEI/FEDER; Government of AragonFEDER, Grant/Award Number: E35_20R Abstract Human riboflavin kinase (HsRFK) catalyzes vitamin B2 (riboflavin) phosphorylation to flavin mononucleotide (FMN), obligatory step in flavin cofactor synthesis. HsRFK expression is related to protection from oxidative stress, amyloid-β toxicity, and some malignant cancers progression. Its downregulation alters expression profiles of clock-controlled metabolic-genes and destroys flavins protection on stroke treatments, while its activity reduction links to protein-energy malnutrition and thyroid hormones decrease. We explored specific features of the mechanisms underlying the regulation of HsRFK activity, showing that both reaction products regulate it through competitive inhibition. Fast-kinetic studies show that despite HsRFK binds faster and preferably the reaction substrates, the complex holding both products is kinetically most stable. An intricate ligand binding landscape with all combinations of substrates/products competing with the catalytic complex and exhibiting moderate cooperativity is also presented. These data might contribute to better understanding the molecular bases of pathologies coursing with aberrant HsRFK availability, and envisage that interaction with its client-apoproteins might favor FMN release. Finally, HsRFK parameters differ from those of the so far evaluated bacterial counterparts, reinforcing the idea of species-specific mechanisms in RFK catalysis. These observations support HsRFK as potential therapeutic target because of its key functions, while also envisage bacterial RFK modules as potential antimicrobial targets. KEYWORDS calorimetry, kinetics limiting step, ligand binding and cooperativity, pre-steady-state kinetics, product inhibition, riboflavin kinase, therapeutic target 10878 | ANOZ-CARBONELL Et AL. uncompetitive for ADP in CaFADS. FMN is a more potent inhibitor in CaFADS ( KRF M / KFMN I =7.1) than in SpnFADS or HsRFK (0.8 and 1, respectively); and also, RF is a very strong inhibitor of the RFK activity in CaFADS but not in the other enzymes.19,20,27 Therefore, FMN biosynthesis in Homo sapiens will be regulated by the products of the RFK reaction. In addition, and considering that transport across membranes appears favored for the RF substrate over FMN and FAD, intracellular compartmentalization with different substrate/ product concentrations and/or the presence of different isoforms might also apply in the in vivo HsRFK regulation.11,28,44 3.3 | Binding of HsRFK substrates is the kinetically preferred process We used stopped-flow spectrophotometry to kinetically differentiate individual processes during the HsRFK reaction. This technique allowed us to detect small changes in the dielectric environment of the flavin isoalloxazine upon its binding and/or dissociation to the protein, and/or as consequence of conformational changes in its environment,19,20 while RF transformation into FMN is not observed due to the same fluorescence spectra and quantum yields of both flavins. When we mixed HsRFK with either RF or FMN, we only detected very slow and linear fluorescence decays consistent with previously reported flavin photobleaching.19,20 These observations suggested that the apo-form of HsRFK is not able to bind RF or FMN (herein both referred as FLV) ligand, or at least to internalize its isoalloxazine ring in the expected catalytically competent enclosed conformation. A fast and intense exponential decay in fluorescence in the 2seconds time frame was, however, detected when mixing HsRFK with all combinations of ATP or ADP (denoted herein as ANP) and FLV ligands (Figure3A). We related the fluorescence decay to FLV binding and/or internalization in the protein matrix by FlapII displacement (Figure3A), concluding that the ANP presence/binding prones HsRFK to bind and internalize the FLV ligand. Noticeably, no subsequent recover of fluorescence was observed for the assayed FIGURE 3 Pre-steady-state stopped-flow kinetics of the binding of RF and FMN to HsRFK in the presence of adenine nucleotides. (A) Normalized evolution of kinetic changes in fluorescence upon mixing HsRFK (0.2µM) with all possible FLV-ANP ligand combinations (0.125 and 250µM, respectively). (B) Example of the fittings of kinetic traces (in this case, corresponding to mixtures of HsRFK with RF-ATP), and residuals of the fitting of the 1µM RF-250µM ATP data to a biexponential function. Evolution of (C) kobs1 and (D) kobs2 as a function of the FLV concentrations. Insets show schemes representing the corresponding processes | 10879 ANOZ-CARBONELL Et AL. combinations of ligands, differing this behavior from the one reported when similarly evaluating the RFK modules of CaFADS and SpnFADS. For these two enzymes, the initial flavin fluorescence decay was followed by fluorescence recover related to an ATP-induced conformational change that re-opens the flavin binding site making the isoalloxazine accessible to the solvent after the reaction has taken place.19,20 Kinetic traces corresponding to mixtures of HsRFK with the RF-ADP ligands fitted to a single exponential decay, whereas two and up to three independent processes were identified when, respectively, evaluating the binding kinetics in the FLV-ATP and FMN-ADP combinations (Figure3B). Noticeably, while the amplitude of the first process (A1) dominates the fluorescence decay for RF binding, A1 and A2 became similar when assaying the binding kinetics of the FMN product (no shown). Therefore, we identified the initial process, generally accounting for most of the amplitude decay, as FLV binding/internalization in HsRFK by FlapII displacement (up to ~14Å reorganization of amino acids comprising the loop), similarly to that reported for CaFADS and SpnFADS.19,20 The succeeding fluorescence decays have to relate to subsequent conformational changes in HsRFK loops further contributing to additional changes in the isoalloxazine environment after the initial binding. Considering these extra processes are not observed when mixing the protein with RFADP, they appear related to the extra phosphates of ATP and FMN, over ADP and RF, respectively, influencing some conformational flexibility. kobs1 values showed a linear dependence on the FLV concentration (Figure3C) that permitted to determine kon and koff for flavin binding and, as a consequence the process dissociation constant (Kd) (Table2). These data indicated that binding processes containing the RF substrate are the fastest (Table2A). In addition, the largest amplitude in fluorescence decay is observed for RF and ATP (Figure3A). Therefore, HsRFK binds preferably the substrates of the RFK reaction, RF and ATP, over other combinations of substrates/products, similarly to SpnFADS but contrary to CaFADS. Noticeably, binding of the FMN product in presence of ANP (particularly ATP) were the least favored processes from the kinetic point of view (smaller kon), though the combination of FMN-ADP products showed an amplitude comparable to that of substrates and a considerably lower koff. As consequence, the complex of HsRFK with the products of its activity exhibits the smaller Kd and appears, therefore, as the most stable one. On their side, when detected, kobs2 showed a saturation profile on the FLV concentration (Figure3D). When fitting these data to an induced fit model representing changes in the protein conformation induced by binding of the ligand,32 we were able to determine an equilibrium reorganization constant as well as the kinetic constant to achieve the final state (Table2B). Noticeably, this process presented very particular features when evaluating the FMN-ADP products; it was considerably slower (kr), A2 was comparable to A1, and the product of its reorganization was the most stable (Kreorg). Therefore, binding of the products of the RFK activity to the enzyme occurs through stabilization of a transient intermediate in the isoalloxazine internationalization by FlapII displacement, while these intermediate is hardly populated, or not at all, in the binding of other ligand combinations. Noticeably, this FMN-ADP combination is the only one for which a third considerably slower process (80-100min−1 at the FMN concentrations assayed), likely independent of the flavin concentration, is observed when binding to HsRFK. Interestingly, kon and kr values for the binding of RF-ATP substrates are substantially faster than the kcat, while koff for the FMN-ADP products is situated in its range (Tables1 and 2). Collectively, these observations indicate that kinetics of products release limits the HsRFK catalytic activity. These data point to differences in the regulation of the catalytic activity of HsRFK when compared to CaFADS and SpnFADS. In HsRFK, as in SpnFADS, the binding of substrates of the RFK reaction -RF and ATPis the kinetically favored process,19 whereas in CaFADS binding of any other combination of ligands is faster.20 Moreover, while in CaFADS and SpnFADS the ATP substrate activates FLV ligand internalization as well as the subsequent cavity reopening to make this ligand again solvent accessible, such TABLE 2 Pre-steady-state kinetic parameters for the binding and dissociation of flavins to HsRFK in the presence of adenine nucleotides. Experiments were performed at 25°C in 20mM of PIPES, pH 7.0, 0.3mM of MgCl2. (n=5, mean±SEM) in and stopped-flow equipment Ligands combination kobs1 (flavin binding) kobs2 (conformational rearrangement) kon (min−1µM−1)koff (min−1)Kd (µM) ΔG (kcal mol−1)kr (min−1)Kreorg (µM) ΔGreorg (kcal mol−1) RF-ATP 1760±50 700±40 0.40±0.02 −8.6±0.1 530±80 0.38±0.08 −8.8±0.1 RF-ADP 1670±90 610±10 0.36±0.03 −8.7±0.1 –a –a –a FMN-ATP 420±40 110±10 0.26±0.04 −8.9±0.1 127±14 0.83±0.41 −8.3±0.3 FMN-ADP 1140±30 59±8 0.05±0.01 −9.9±0.1 12±2 0.16±0.1 −9.2±0.4 aProcess not observed for this combination of ligands. 10880 | ANOZ-CARBONELL Et AL. final exposure is not detected with HsRFK. In addition, this enzyme is the only from the three for which koff for the FMNADP ligands is considerably slower than kon (nearly 20-fold), making the binding of the products stronger and envisaging different mechanisms for the FMN product release from the RFK site among these three proteins. FIGURE 4 Scheme of the conformational and ligand binding spaces along the HsRFK cycle. The diagram summarizes the different HsRFK species envisaged considering kinetic, inhibition, binding, and structural data available. All presented structures correspond to different crystal structures for HsRFK (PDB IDs are indicated), with the only exception of the apo-form that has been produced by MD simulations of the 1NB0 pdb after removing the ADP:Mg2+ ligand. Crystal structures are represented by B-factor of backbone atoms, with higher radius of ribbons and warmer colors indicating higher fluctuations, being all snapshots normalized according the color code. Those states lacking structural data are represented by circles (nonproductive states are shown in red, alternative paths in violet and competent state in blue start). Processes leading to the formation of sub-stoichiometric complexes (HsRFK:RF) are highlighted with blue arrows | 10881 ANOZ-CARBONELL Et AL. 3.4 | Thermodynamics modulates the ligand binding landscape of HsRFK Subsequently, we performed ITC experiments to assess if the kinetically detected processes were significant in reaching the thermodynamic equilibrium. Binary and ternary interactions of HsRFK with ANP and/or FLV ligands were analyzed at pH 7.0 and 25°C both in absence and in presence of 0.3mM of Mg2+. The corresponding determined thermodynamic parameters are summarized in Table SP1, while some examples of the experimental thermodynamic dissections are displayed in Figure SP5. Direct titrations allowed the determination of the intrinsic binding parameters of the interaction of HsRFK with substrates and products of the RFK reaction. For ANP ligands, KdANP values were in the low micromolar range and the stoichiometry of the interaction, around 0.6, was consistent with a unique ANP-binding site, with occupancy below unity being probably associated to protein conformational heterogeneity. This agrees with low B-factors of the bound ADP-Mg2+ in the available crystallographic structures (Figure4). The presence of Mg2+ resulted in the reduction of the favorable enthalpic contribution to binding, as well as of the unfavorable entropic one (Figure SP6). Nonetheless, ΔG remained mostly insensitive to the cation through entropy/enthalpy compensation. A similar situation was previously reported for the RFK module of CaFADS.33 On the contrary, FLV (RF and FMN) ligands were hardly able to directly bind HsRFK. No interaction heat was detected for the protein titration with FMN, suggesting either lack of interaction, very slow binding or interaction occurring without appreciable exchange of heat. In titrations with the RF substrate, data allowed for estimation of a KRF d value in the micromolar range, but the low interaction stoichiometry observed envisages very low occupancy (N around 0.16). Mg2+ further hindered RF binding due to a higher increase in the entropic contribution to the binding than in the enthalpic one (2.32and 1.37-fold, respectively) (Table SP2, Figure SP6). Titrations of ANP:HsRFK or FLV:HsRFK binary mixtures with, respectively, FLV or ANP permitted to further unravel the complete thermodynamic landscape of ligand binding. Figure5 summarizes all possible binary and tertiary interactions of the enzyme with substrates and products in absence (Figure5A) and presence of Mg2+ (Figure5B), including the fraction of binding-competent protein (N) in each case as the thickness of the arrows. Titrations involving both RF and ATP in the presence of the divalent cation (catalytic conditions) were not measured, since the heat of the catalytic reaction masked the interaction heat. As shown in the figure, pathways leading to non-competent tertiary complexes compete with formation of the HsRFK:ATP:RF catalytically complex (orange pathways in Figure5). Nevertheless, there was no thermodynamically preferred binding pathway, both in terms of final complex stability and production probability (N, interaction stoichiometry). The presence of Mg2+ slightly increased the fraction of protein prone to interact and, consequently, the probability of a particular path to occur, but, contrary to CaFADS and SpnFADS,19,20 hardly modulated the binding landscape (Figure5A,B, SP5, Table SP1). The only exception to this behavior was observed for the FMNADP products combination, where the cation presence made enthalpic as well as entropic contributions to the HsRFK binding favorable. Therefore, differences in conformation of ternary HsRFK complexes as a consequence of the cation FIGURE 5 Gibbs free energy flow for the interaction of HsRFK with substrates and products. Diagrams summarize the thermodynamics of the interaction of HsRFK with different combination of its ligands as obtained by ITC (Table SP1) at 25°C (A) in 20mM of PIPES, pH 7.0, 0.3mM of MgCl2, and (B) in 20mM of PIPES, pH 7.0. HsRFK is represented as blue spheres, RF and FMN as orange and yellow hexagons, and ATP and ADP as green and blue triangles, respectively. The length of the arrows is proportional to the ΔG for the interaction (value in kcal mol−1 are indicated in numbers), and its thickness is representative to the fraction of HsRFK binding the titrating ligand. Processes not directly observed by ITC (interaction of HsRFK with FMN) are shown as dotted arrows. Paths leading to the formation of the tertiary catalytic complex (HsRFK with ATP and RF substrates) are highlighted with orange arrows. NM indicates processes that could not be measured in the presence of Mg2+ since the reaction heat would mask the interaction heat 10882 | ANOZ-CARBONELL Et AL. presence are only predicted for the formation of the ternary complex containing the FMN and ADP products, contrary to that reported for CaFADS and SpnFADS.19,20 The energy diagram in the absence of MgCl2 (Figure5B) showed two alternative pathways leading to HsRFK:ATP:RF “pseudo-reactive” complexes, which in addition are among the most probable. Therefore, the HsRFK behavior is more similar to SpnFADS than to CaFADS, enzyme that favors all the other nonproductive CaFADS:ANP:FLV complexes against the CaFADS:ATP:RF one.19,20 3.5 | ANP and FLV ligands cooperate in their binding to HsRFK Although direct FLV binding to the free protein was hardly observed by ITC or stopped-flow spectrophotometry, the presence of FLV increased HsRFK affinity for ANP ligands, particularly when the cation is present (compare ΔG values for the titrations of free HsRFK and binary mixtures, Table SP1, Figure5A and SP6). These observations suggest that, as in the bacterial RFK modules, (a) FLV ligands have a slowbinding mode to HsRFK that permits to indirectly estimate their binding parameters 34,45,46 and (b) ANP and FLV show cooperativity in their binding. To evaluate cooperativity, we titrated HsRFK:FLV binary mixtures with ANP and fitted the resulting thermograms to a model for heterotrophic interactions applying, as explained in Materials and Methods section, two complementary methodologies.34,36,46 Our data, summarized in Table3, show that Mg2+ modulates ligand binding cooperativity to HsRFK. In its presence, FLV and ANP ligands show positive cooperativity (α>1). When Mg2+ is absent, RF-ADP and FMN-ADP binding cooperativity increases (up to 4-fold and 3-fold, respectively), while binding cooperativity becomes slightly negative for the RF-ATP substrates combination (α<1). In general, the magnitudes of the cooperativity constants for ligand binding to HsRFK are moderated and in the range of those for SpnFADS, while those for CaFADS are considerably larger, particularly in the presence of MgCl2. Cooperativity in RF and ATP substrates binding to HsRFK and SpnFADS is slightly negative and positive, respectively. Noticeably, in the case of CaFADS its sign and magnitude are highly influenced by the RF substrate concentration that in this case also acts as inhibitor.19,20 Thus, differences are also found in the cooperation of substrates and products binding to RFK enzymes from different organisms. 4 | DISCUSSION 4.1 | Conformational landscape in the HsRFK catalytic cycle To date, there is no an available 3D crystal structure of HsRFK in the absence of any ligand, either ANP or FLV, or both. To gain insight into such conformation, we generated a model of apo-HsRFK by removing Mg2+:ADP from the HsRFK-Mg2+:ADP crystal structure (PDB ID 1NB0). This model was minimized and relaxed by MD simulations (5 replicas) (Figure6 and SP7). Trajectories for Cα root mean square deviation (RMSD), energy, solvent accessible surface (SAS) and radius of gyration (GyR) indicate that apo-HsRFK keeps overall folding along simulations (Figure SP7). The most remarkable fact was the transient breaking of the Lys20-Asp88 salt bridge, with the consequent displacement of FlapI and loop 5 (L5c), and the observation of a dynamic opening/closing of the ADP/ATP-binding cavity (Figure6). On the contrary, the conformation of the active site, formed by the consensus PTAN motif and Glu78, retained conformations similar to those observed in the crystal binaryand ternaryHsRFK complexes (Figure6C and SP8). Such conformations resemble those in the apo-forms of SpRFK and RFK module of SpnFADS (Figure1 and SP7). This leaves apo-CaFADS as the only RFK showing a different PTAN conformation due to its Thr exhibiting considerably different Φ and ψ values (Figure SP8A). In conclusion, our MD data show that FlapI and L5c adopt different conformations in apo-HsRFK with respect to binary HsRFK:ADP complexes.2 Additionally, our simulations predict an open Flavin binding site for apo-HsRFK, while crystal structures indicate that closed conformations must be populated in ternary complexes due FlapII displacement toward this cavity (Figure1 and SP2).17 Therefore, it is accepted that HsRFK must undergo a series of sequential conformational changes during the catalytic cycle (Figure4). An ordered bi-bi mechanism for mammalian RFKs was previously proposed, in which RF binding was followed by TABLE 3 Cooperativity coefficients for the binding of the different combinations of FLV and ANP ligands to HsRFK in presence and absence of Mg2+. Experiments were performed at 25°C in 20mM of PIPES, pH 7.0, both in absence and presence of 0.3mM of MgCl2. (n=5, mean±SEM) [MgCl2] Ligands α N Δh kcal/mol 0.3mM RF-ATP N.M.a N.M.a N.M.a RF-ADP 2.2±0.2 0.62±0.01 −1.3±0.2 FMN-ATP 1.4±0.1 0.49±0.01 6.2±1.0 FMN-ADP 2.0±0.2 0.50±0.01 −10±0.3 0mM RF-ATP 0.87±0.20 0.68±0.01 5.3±0.4 RF-ADP 8.3±1.2 0.46±0.01 1.8±0.6 FMN-ATP 0.96±0.10 0.61±0.01 2.7±0.2 FMN-ADP 6.9±1.2 0.56±0.01 −5.9±0.6 aN.M., not measured. When mixing RF and ATP in presence of Mg2+, the catalytic reaction heat conceals the interaction heat. | 10883 ANOZ-CARBONELL Et AL. FIGURE 6 The apo-HsRFK structural model. (A) Cartoon overlapping and (B) surfaces around the ADP/ATP binding site of apo-HsRFK structural models for the starting structure (violet) and the snapshot after 5ns of MD (brown). L5c and FlapI are, respectively, colored in salmon and olive. Side chains for Lys20 and Asp88 are shown in sticks. (C) Stick representation of the conformation of the PTAN motif and the catalytic residue Glu78 in the starting structure (violet) and the snapshot after 5ns of MD (brown). Panels A-C show snapshots of replica 1, and initial distances among selected atoms are shown as black dashed black lines, while corresponding distances at the end of the simulation are shown as grey dashed lines. (D) Trajectories for the evolution of the relative distances between residues Lys20 and Asp88, as well as among Thr26 at the PTAN motif and Ser19 and Glu78 along 5ns MD simulations of apo-HsRFK. Data are shown for the 5 replicates. Simulations carried out at 300K (A) (C) (D) (B) 10884 | ANOZ-CARBONELL Et AL. ATP binding, being ADP and FMN subsequently released after catalysis.1,2,47 However, our ITC and stopped-flow experiments demonstrate that FLV (RF and FMN) ligands hardly interact with HsRFK in the absence of ANP ligands. The interaction of RF detected by ITC—a minority pathway in the whole interaction landscape—might be associated to the slow binding of a few molecules, probably associated to RF molecules recognizing some motives of the large flavin binding site in the open conformation expected in apo-HsRFK (Figure4) (as observed in SpRFK17). Nonetheless, RF recognition in this state appears scarce, slow and unable to trigger by itself the structural reorganization of FlapII. This is confirmed by the absence of changes indicative of internalization of the isoalloxazine ring in our stoppedflow experiments as well as in the crystallographic structures (differences in FlapI and FlapII disposition relative to HsRFK:ADP:FMN complex, Figures 1, 4 and SP2). Nonetheless, the larger magnitudes for the entropic and enthalpic contributions for ATP binding to apo-HsRFK when compared to the HsRFK-RF mixture (Table SP1 and Figure SP6) point to structural rearrangements associated to the RF presence in its binding cavity that favors ANP binding. Nonetheless, even if RF favors the initial binding of ANP, the accommodation of the ANP in the cavity and the establishment of new FlapI-ANP interactions probably elapses the initial nonproductive RF interaction mode to a new one. Thus, conformational changes in FlapI have also an effect in FlapII conformation (see FlapI and FlapII in HsRFK:ADP:FMN crystal structures, Figure 1 and SP2), further contributing to place RF in an arrangement compatible with catalysis.1,2 Our transient kinetic experiments support such mechanism, since FlapII displacements can be indirectly perceived by the changes in flavin fluorescence. Thus, for mixtures of HsRFK with FMN-ANP, we observed a slower reorganization process consistent with additional isoalloxazine burial into the binding pocket by FlapII reorganization. However, in the case of RF-ATP mixes, k2 is relatively faster and shows lower amplitude, probably only reflecting the protein dynamics during the catalytic turnover. Noticeably, our stopped-flow data also envisage that after catalysis the flavin binding site of HsRFK remains blocked by FlapII (Figure4), making FMN release the rate-limiting step in FMN production and envisaging that this reaction process might be controlled by factors different from the protein itself. In this context, it is worth to note that HsFADS, besides its FMNAT activity, also operates as a FAD chaperone for flavin delivery to its client apoproteins.29,48 Our data envisage that a similar mechanism might apply in HsRFK for FMN transfer to HsFADS as well as to FMN dependent client apoproteins, with direct protein-protein interaction favoring FMN release from HsRFK. Such tight regulation agrees with FMN, as well as FAD, being crucial cofactors in a pletora of enzymes devoted to manage cell bioenergetics. 4.2 | Different organisms, different regulatory strategies Our previous hypothesis of a species-specific inhibition and activity modulation of the RFK activity in bifunctional FADS is here reinforced, as well as extended to monofunctional proteins, by the HsRFK data. Despite the overall structural similarity among eukaryotic RFK enzymes and prokaryotic RFK-modules (RMSD of core Cα positions are only 1.2, 1.6, and 1Å when comparing HsRFK with SpRFK, CaFADS, and SpnFADS, respectively), differences among species occur in the conformation of several structural elements, including the FlapI and FlapII loops, and, particularly, the catalytic PTAN motif (Figure SP8A).18 In this structural context, activity and binding studies reflect some common regulatory mechanisms, as well as highly relevant differences among HsRFK and the RFK-modules of CaFADS and SpnFADS. These variations modulate ligand binding and, consequently, catalytic cycles, resulting in a variety of species-specific mechanisms regulating the biosynthesis of flavin cofactors. Thus, the inhibition of the RFK activity by the products of the reaction seems to be common for all of them, although inhibition potency and mechanism varies with the enzyme. For example, CaFADS is more strongly inhibited by both reaction products as deduced from the KIFMN/KMRF and KIADP/KMATP ratios (0.2 vs 1 and 1, and 0.42 vs 1.16 and 11.2, for HsRFK and SpnFADS, respectively).19,20 Kinetic data for binding are consistent with these differences, since binding of the RFK reaction products is kinetically preferred in CaFADS, while HsRFK and SpnFADS bind the substrates faster. Thermodynamics also shows that pathways leading to the catalytic RFK complex are un-favored respect to those leading to “pseudo-reactive” complexes in CaFADS,20 whereas these differences either do not exist or favor formation of the catalytic complex in HsRFK and SpnFADS, respectively (Table SP1).19 Differences are also observed in heterotropic ligand binding cooperativity. Cooperativity seems to be determinant in regulating RFK activity in CaFADS, while in general the HsRFK behavior is more modest and resembles that of SpnFADS. Thus, our data indicate that HsRFK, CaFADS, and SpnFADS achieve the catalytic RFK:RF:ATP complex through mechanisms exhibiting relevant differences. HsRFK and SpnFADS follow a random sequential binding of the RFK substrates, while substrates binding to CaFADS is concerted. Differences in the substrates cooperative behavior and magnitude might relate also to the different conformations of the PTAN motif among species (Figure SP8A), which points to specific conformational changes during the RFK activity. In CaFADS, the occupation of the ANP binding site by RF—when it is in excess—might prevent the ligand-induced conformational change of this motif, which is necessary for ATP binding.18 This structural rearrangement is not expected to be necessary neither for HsRFK (Figure1D,E) nor for SpnFADS.19 Therefore, the absence of | 10885 ANOZ-CARBONELL Et AL. inhibition by RF in HsRFK and SpnFADS might be associated to the minimal rearrangement of PTAN motif during the catalytic cycle of these enzymes. For all three proteins release of FMN and ADP products appears to be the reaction limiting step, but clear differences are also envisaged in the conformation of such ternary complex in solution. Thus, our stopped-flow analyses suggest that FMN is not accessible to the solvent when the HsRFK:FMN:ADP complex is in solution, while it is accessible when similarly evaluating the RFK modules of CaFADS and SpnFADS. This might implicate different modes for the transfer of the newly synthesized FMN to the client proteins. In this context, we must also consider that while eukaryotic RFKs are relatively small, monofunctional and monomeric proteins, their bacterial counterparts have an additional FMNAT module that duplicates its size.16 Moreover, these bifunctional enzymes can stabilize quaternary assemblies with direct interaction of RFK and FMNAT ligand binding cavities, which potentially will contribute to FMN release from the RFK module by direct transfer to the FMNAT module as well as to regulate FlapII conformation and flavin accessibility to the solvent.16,25 In conclusion, we present here an integrated thermodynamic and kinetic description of the catalytic mechanism of HsRFK that might contribute to the better understanding of the molecular bases of certain pathologies coursing with changes in the expression or catalytic efficiency of this protein. We also report key thermodynamic, kinetic and structural differences of the regulation of the HsRFK catalytic cycles relative to bacterial modules performing the same activity. To date, antimicrobials only targeting the FMNAT activity of prokaryotic FADSs have been investigated,49 probably because of the overall sequence and structural similarity of their RFK module to their eukaryotic counterparts envisaged specificity compromise and as a consequence deleterious effects to the host.50 However, the here presented data also foresee that the bacterial RFK activity might be consider a potential antimicrobial target for some bacterial pathogens. ACKNOWLEDGMENTS Authors would like to acknowledge the use of Servicios Generales de Apoyo a la Investigación-SAI, Universidad de Zaragoza. CONFLICT OF INTEREST The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results. AUTHOR CONTRIBUTIONS Conceptualization, M. Medina; methodology, E. AnozCarbonell, M. Rivero, V. Polo, A. Velazquez-Camoy and M. Medina; formal analysis, E. Anoz-Carbonell, M. Rivero, V. Polo and M. Medina; investigation, E. Anoz-Carbonell and M. Rivero; data curation, E. Anoz-Carbonell and M. Medina; writing—original draft preparation, E. Anoz-Carbonell and M. Medina; writing—review and editing, E. Anoz-Carbonell and M. Medina; project administration, M. Medina; funding acquisition, M. Medina. REFERENCES 1. Karthikeyan S, Zhou Q, Mseeh F, Grishin NV, Osterman AL, Zhang H. Crystal structure of human riboflavin kinase reveals a beta barrel fold and a novel active site arch. Structure. 2003;11:265-273. 2. Karthikeyan S, Zhou Q, Osterman AL, Zhang H. Ligand binding-induced conformational changes in riboflavin kinase: structural basis for the ordered mechanism. Biochemistry. 2003;42:12532-12538. 3. Yazdanpanah B, Wiegmann K, Tchikov V, et al. Riboflavin kinase couples TNF receptor 1 to NADPH oxidase. Nature. 2009;460:1159-1163. 4. Hirano G, Izumi H, Yasuniwa Y, et al. Involvement of riboflavin kinase expression in cellular sensitivity against cisplatin. Int J Oncol. 2011;38:893-902. 5. Chen X, Ji B, Hao X, et al. FMN reduces Amyloid-β toxicity in yeast by regulating redox status and cellular metabolism. Nat Commun. 2020;11:867. 6. Zou YX, Zhang XH, Su FY, Liu X. Importance of riboflavin kinase in the pathogenesis of stroke. CNS Neurosci Ther. 2012;18:834-840. 7. Hirano A, Braas D, Fu YH, Ptáček LJ. FAD regulates CRYPTOCHROME protein stability and circadian clock in mice. Cell Rep. 2017;19:255-266. 8. Capo-chichi CD, Guéant JL, Lefebvre E, et al. Riboflavin and riboflavin-derived cofactors in adolescent girls with anorexia nervosa. Am J Clin Nutr. 1999;69:672-678. 9. Capo-Chichi CD, Feillet F, Guéant JL, et al. Concentrations of riboflavin and related organic acids in children with protein-energy malnutrition. Am J Clin Nutr. 2000;71:978-986. 10. Lee SS, McCormick DB. Thyroid hormone regulation of flavocoenzyme biosynthesis. Arch Biochem Biophys. 1985;237:197-201. 11. Barile M, Giancaspero TA, Brizio C, et al. Biosynthesis of flavin cofactors in man: implications in health and disease. Curr Pharm Des. 2013;19:2649-2675. 12. Barile M, Brizio C, Valenti D, De Virgilio C, Passarella S. The riboflavin/FAD cycle in rat liver mitochondria. Eur J Biochem. 2000;267:4888-4900. 13. Lienhart WD, Gudipati V, Macheroux P. The human flavoproteome. Arch Biochem Biophys. 2013;535:150-162. 14. Patel MV, Chandra TS. Metabolic engineering of Ashbya gossypii for enhanced FAD production through promoter replacement of FMN1 gene. Enzyme Microb Technol. 2020;133:109455. 15. Yruela I, Arilla-Luna S, Medina M, Contreras-Moreira B. Evolutionary divergence of chloroplasts FAD synthetase proteins. BMC Evol Biol. 2010;10:311. 16. Herguedas B, Martinez-Julvez M, Frago S, Medina M, Hermoso JA. Oligomeric state in the crystal structure of modular FAD synthetase provides insights into its sequential catalysis in prokaryotes. J Mol Biol. 2010;400:218-230. 17. Bauer S, Kemter K, Bacher A, Huber R, Fischer M, Steinbacher S. Crystal structure of Schizosaccharomyces pombe riboflavin kinase reveals a novel ATP and riboflavin-binding fold. J Mol Biol. 2003;326:1463-1473. 10886 | ANOZ-CARBONELL Et AL. 18. Herguedas B, Lans I, Sebastián M, Hermoso JA, Martínez-Júlvez M, Medina M. Structural insights into the synthesis of FMN in prokaryotic organisms. Acta Crystallogr D Biol Crystallogr. 2015;71:2526-2542. 19. Sebastián M, Velázquez-Campoy A, Medina M. The RFK catalytic cycle of the pathogen Streptococcus pneumoniae shows species-specific features in prokaryotic FMN synthesis. J Enzyme Inhib Med Chem. 2018;33:842-849. 20. Sebastián M, Serrano A, Velázquez-Campoy A, Medina M. Kinetics and thermodynamics of the protein-ligand interactions in the riboflavin kinase activity of the FAD synthetase from Corynebacterium ammoniagenes. Sci Rep. 2017;7:7281. 21. Walsh CT, Wencewicz TA. Flavoenzymes: versatile catalysts in biosynthetic pathways. Nat Prod Rep. 2013;30:175-200. 22. Sebastián M, Lira-Navarrete E, Serrano A, et al. The FAD synthetase from the human pathogen Streptococcus pneumoniae: a bifunctional enzyme exhibiting activity-dependent redox requirements. Sci Rep. 2017;7:7609. 23. Matern A, Pedrolli D, Großhennig S, Johansson J, Mack M. Uptake and metabolism of antibiotics roseoflavin and 8-demethyl-8-aminoriboflavin in riboflavin-auxotrophic Listeria monocytogenes. J Bacteriol. 2016;198:3233-3243. 24. Marcuello C, Arilla-Luna S, Medina M, Lostao A. Detection of a quaternary organization into dimer of trimers of Corynebacterium ammoniagenes FAD synthetase at the single-molecule level and at the in cell level. Biochim Biophys Acta. 2013;1834:665-676. 25. Lans I, Seco J, Serrano A, et al. The dimer-of-trimers assembly prevents catalysis at the transferase site of prokaryotic FAD synthase. Biophys J. 2018;115:988-995. 26. Solovieva IM, Tarasov KV, Perumov DA. Main physicochemical features of monofunctional flavokinase from Bacillus subtilis. Biochemistry (Mosc). 2003;68:177-181. 27. Sebastián M, Arilla-Luna S, Bellalou J, Yruela I, Medina M. The biosynthesis of flavin cofactors in Listeria monocytogenes. J Mol Biol. 2019;431:2762-2776. 28. Barile M, Giancaspero TA, Leone P, Galluccio M, Indiveri C. Riboflavin transport and metabolism in humans. J Inherit Metab Dis. 2016;39:545-557. 29. Giancaspero TA, Colella M, Brizio C, et al. Remaining challenges in cellular flavin cofactor homeostasis and flavoprotein biogenesis. Front Chem. 2015;3:30. 30. Serrano A, Sebastián M, Arilla-Luna S, et al. The trimer interface in the quaternary structure of the bifunctional prokaryotic FAD synthetase from Corynebacterium ammoniagenes. Sci Rep. 2017;7:404. 31. Serrano A, Frago S, Herguedas B, Martinez-Julvez M, VelazquezCampoy A, Medina M. Key residues at the riboflavin kinase catalytic site of the bifunctional riboflavin kinase/FMN adenylyltransferase from Corynebacterium ammoniagenes. Cell Biochem Biophys. 2013;65:57-68. 32. Vogt AD, Di Cera E. Conformational selection or induced fit? A critical appraisal of the kinetic mechanism. Biochemistry. 2012;51:5894-5902. 33. Frago S, Velázquez-Campoy A, Medina M. The puzzle of ligand binding to Corynebacterium ammoniagenes FAD synthetase. J Biol Chem. 2009;284:6610-6619. 34. Martinez-Julvez M, Abian O, Vega S, Medina M, VelazquezCampoy A. Studying the allosteric energy cycle by isothermal titration calorimetry. Methods Mol Biol. 2012;796:53-70. 35. Velázquez-Campoy A, Goñi G, Peregrina JR, Medina M. Exact analysis of heterotropic interactions in proteins: Characterization of cooperative ligand binding by isothermal titration calorimetry. Biophys J. 2006;91:1887-1904. 36. Vega S, Abian O, Velazquez-Campoy A. A unified framework based on the binding polynomial for characterizing biological systems by isothermal titration calorimetry. Methods. 2015;76:99-115. 37. Olsson MHM, Søndergaard CR, Rostkowski M, Jensen JH. PROPKA3: consistent treatment of internal and surface residues in empirical pKa predictions. J Chem Theory Comput. 2011;7:525-537. 38. Delano WL. PyMOL: an open-source molecular graphics tool. CCP4 Newsletter Pro Crys. 2002;40:82-92. 39. Abraham MJ, Murtola T, Schulz R, et al. GROMACS: High performance molecular simulations through multi-level parallelism from laptops to supercomputers. SoftwareX. 2015;1-2:19-25. 40. Duan Y, Wu C, Chowdhury S, et al. A point-charge force field for molecular mechanics simulations of proteins based on condensed-phase quantum mechanical calculations. J Comput Chem. 2003;24:1999-2012. 41. Humphrey W, Dalke A, Schulten K. VMD: visual molecular dynamics. J Mol Graph. 1996;14:33-38, 27-38. 42. Lee SS, McCormick DB. Effect of riboflavin status on hepatic activities of flavin-metabolizing enzymes in rats. J Nutr. 1983;113:2274-2279. 43. Pedrolli DB, Nakanishi S, Barile M, et al. The antibiotics roseoflavin and 8-demethyl-8-amino-riboflavin from Streptomyces davawensis are metabolized by human flavokinase and human FAD synthetase. Biochem Pharmacol. 2011;82:1853-1859. 44. Jin C, Yao Y, Yonezawa A, et al. Riboflavin transporters RFVT/ SLC52A mediate translocation of riboflavin, rather than FMN or FAD, across plasma membrane. Biol Pharm Bull. 2017;40:1990-1995. 45. Bollen YJ, Westphal AH, Lindhoud S, van Berkel WJ, van Mierlo CP. Distant residues mediate picomolar binding affinity of a protein cofactor. Nat Commun. 2012;3:1010. 46. Martínez-Júlvez M, Medina M, Velázquez-Campoy A. Binding thermodynamics of ferredoxin:NADP+ reductase: two different protein substrates and one energetics. Biophys J. 2009;96:4966-4975. 47. Yamada Y, Merrill AH Jr, McCormick DB. Probable reaction mechanisms of flavokinase and FAD synthetase from rat liver. Arch Biochem Biophys. 1990;278:125-130. 48. Torchetti EM, Bonomi F, Galluccio M, et al. Human FAD synthase (isoform 2): a component of the machinery that delivers FAD to apo-flavoproteins. FEBS J. 2011;278:4434-4449. 49. Sebastián M, Anoz-Carbonell E, Gracia B, et al. Discovery of antimicrobial compounds targeting bacterial type FAD synthetases. J Enzyme Inhib Med Chem. 2018;33:241-254. 50. Serrano A, Ferreira P, Martínez-Júlvez M, Medina M. The prokaryotic FAD synthetase family: a potential drug target. Curr Pharm Des. 2013;19:2637-2648. SUPPORTING INFORMATION Additional Supporting Information may be found online in the Supporting Information section. How to cite this article: Anoz-Carbonell E, Rivero M, Polo V, Velázquez-Campoy A, Medina M. Human riboflavin kinase: Species-specific traits in the biosynthesis of the FMN cofactor. The FASEB Journal. 2020;34:10871–10886. https://doi. org/10.1096/fj.20200 0566R 1  Supplementary Material Human Riboflavin Kinase: species-specific traits in the biosynthesis of the essential FMN cofactor Ernesto Anoz-Carbonella,b, Maribel Rivero a, Victor Polob,c, Adrián Velázquez-Campoya,b,d,e,f and Milagros Medinaa,b* a Departamento de Bioquímica y Biología Molecular y Celular, Facultad de Ciencias, Universidad de Zaragoza, Spain. b Instituto de Biocomputación y Física de Sistemas Complejos (GBsC-CSIC and BIFIIQFR Joint Units), Universidad de Zaragoza, Spain. c Departamento de Química Física, Universidad de Zaragoza, 50009 Zaragoza, Spain. d Fundación ARAID, Diputación General de Aragón, Spain. e Aragon Institute for Health Research (IIS Aragon), Zaragoza, Spain. f Biomedical Research Networking Centre for Liver and Digestive Diseases (CIBERehd), Madrid, Spain. Abbreviated title for the running head: The biosynthesis of FMN in human cells  8  SUPPLEMENTARY TABLES Table SP1. Thermodynamic parameters for ligand (ANP or FLV) binding to free HsRFK and pre-formed binary HsRFK-FLV and HsRFK-ANP mixtures. Experiments carried out at 25ºC in 20 mM PIPES, pH 7.0, 0.3 mM MgCl2 (upper table) and 20 mM PIPES, pH 7.0 (lower table). Errors in ΔG, ΔH and –TΔS were estimated in ±0.3 kcal/mol and those in Kd in ±15%, taken in general larger than the standard deviation between replicates (n=3) and the numerical error after fitting analysis. 0.3 mM MgCl2 Pre-bound ligand Titrating ligand Kd (µM) ΔG (kcal mol-1) ΔH (kcal mol-1) -TΔS (kcal mol-1) - ATP 0.16 -9.1 -13.7 4.6 - ADP 0.18 -9.1 -16.0 6.9 - RF 0.52 -8.5 -21.5 13 - FMN N.D.a N.D.a N.D.a N.D.a ATP RF N.M.b N.M.b N.M.b N.M.b ATP FMN 0.05 -9.8 -22.2 12.4 ADP RF 0.003 -11.4 -28.9 17.5 ADP FMN 0.01 -10.6 -20.2 -8.3 RF ATP N.M.b N.M.b N.M.b N.M.b RF ADP 0.07 -9.7 -18.1 8.4 FMN ATP 0.15 -9.3 -16.0 6.7 FMN ADP 0.10 -9.6 -28.9 19.3 0.0 mM MgCl2 Pre-bound ligand Titrating ligand Kd (µM) ΔG (kcal mol-1) ΔH (kcal mol-1) -TΔS (kcal mol-1) - ATP 0.07 -9.6 -15.8 6.2 - ADP 1.37 -7.9 -20.1 -5.7 - RF 0.03 -10.1 -15.7 5.6 - FMN N.D.a N.D.a N.D.a N.D.a ATP RF 0.29 -8.8 -9.0 0.1 ATP FMN 0.16 -9.9 -23.3 13.4 ADP RF 0.04 -10.1 -27.0 16.9 ADP FMN 0.16 -9.9 -15.8 5.9 RF ATP 0.07 -9.7 -9.6 -0.1 RF ADP 0.18 -9.2 -18.8 9.6 FMN ATP 0.07 -9.8 -12.9 3.1 FMN ADP 0.04 -10.1 -24.1 14 a N.D. Not detected. No heat of interaction detected for this titration. b N.M. Not measured. Combinations including RF and ATP substrates in presence of Mg2+ lead to the catalytic reaction, which mask the interaction heats. SUPPLEMENTARYREFERENCES 1.Szklarczyk,D.,Gable,A.L.,Lyon,D.,Junge,A.,Wyder,S.,Huerta‐Cepas,J.,Simonovic,M.,Doncheva,N.T., Morris,J.H.,Bork,P.,Jensen,L.J.,andMering,C.V.(2019)STRINGv11:protein‐proteinassociation networkswithincreasedcoverage,supportingfunctionaldiscoveryingenome‐wideexperimental datasets.NucleicAcidsRes47,D607‐D613 2.Robert,X.,andGouet,P.(2014)DecipheringkeyfeaturesinproteinstructureswiththenewENDscript server.NucleicAcidsRes42,W320‐324 Publication II Mutation of aspartate 238 in FAD synthase isoform 6 increases the specific activity by weakening the FAD binding Ernesto Anoz Carbonell – Doctoral Thesis 86 Ernesto Anoz Carbonell contributions: Experimental work: pre-steady-state kinetics through stopped-flow spectroscopy, and isothermal titration calorimetry experiments. Data analysis. Manuscript writing, review and editing International Journal of Molecular Sciences Article Mutation of Aspartate 238 in FAD Synthase Isoform 6 Increases the Specific Activity by Weakening the FAD Binding Piero Leone 1,2, Michele Galluccio 2, Stefano Quarta 1, Ernesto Anoz-Carbonell 3, Milagros Medina 3, Cesare Indiveri 2and Maria Barile 1,* 1Department of Biosciences, Biotechnology and Biopharmaceutics, University of Bari, via Orabona 4–, 70126 Bari, Italy; [email protected] (P.L.); [email protected] (S.Q.) 2Department of Biology, Ecology and Earth Sciences (DiBEST), Unit of Biochemistry and Molecular Biotechnology, University of Calabria, via P. Bucci 4c, 87036 Arcavacata di Rende, Italy; [email protected] (M.G.); [email protected] (C.I.) 3Departamento de Bioquímica y Biología Molecular y Celular, Facultad de Ciencias, Instituto de BiocomputaciónyFísica de Sistemas Complejos (GBsC-CSIC and BIFI-IQFR Joint Units), Universidad de Zaragoza, 50009 Zaragoza, Spain; [email protected] (E.A.-C.); [email protected] (M.M.) *Correspondence: [email protected] Received: 5 November 2019; Accepted: 6 December 2019; Published: 9 December 2019   Abstract: FAD synthase (FADS, or FMN:ATP adenylyl transferase) coded by the FLAD1 gene is the last enzyme in the pathway of FAD synthesis. The mitochondrial isoform 1 and the cytosolic isoform 2 are characterized by the following two domains: the C-terminal PAPS domain (FADSy) performing FAD synthesis and pyrophosphorolysis; the N-terminal molybdopterin-binding domain (FADHy) performing a Co ++ /K + -dependent FAD hydrolysis. Mutations in FLAD1 gene are responsible for riboflavin responsive and non-responsive multiple acyl-CoA dehydrogenases and combined respiratory chain deficiency. In patients harboring frameshift mutations, a shorter isoform (hFADS6) containing the sole FADSy domain is produced representing an emergency protein. With the aim to ameliorate its function we planned to obtain an engineered more efficient hFADS6. Thus, the D238A mutant, resembling the D181A FMNAT “supermutant” of C. glabrata, was overproduced and purified. Kinetic analysis of this enzyme highlighted a general increase of K m , while the k cat was two-fold higher than that of WT. The data suggest that the FAD synthesis rate can be increased. Additional modifications could be performed to further improve the synthesis of FAD. These results correlate with previous data produced in our laboratory, and point towards the following proposals (i) FAD release is the rate limiting step of the catalytic cycle and (ii) ATP and FMN binding sites are synergistically connected. Keywords: FAD synthase; FMN adenylyl transferase; FADS isoform 6; supermutant 1. Introduction The riboflavin (Rf) derived FMN and FAD cofactors play a pivotal role in cell economy ensuring the functionality of the flavoproteome, mainly localized in mitochondria [ 1 , 2 ]. Consistent with the crucial role of flavins and flavoenzymes in cell life, several diseases, including neuromuscular and neurological disorders are linked to flavin-dependent enzyme deficiency or impairment in Rf homeostasis in humans and experimental animals. These disorders, in some cases, can be cured with high doses of Rf, as the two Rf-responsive (RR) disorders Brown–Vialetto van Laere syndrome (BVVLS) [ 3 , 4 ] and RR-multiple acyl-CoA dehydrogenase deficiency (RR-MADD) [ 5 – 7 ]. In mammalian cells, Rf is taken up via translocators (SLC52A1-3, also named RFVT1-3) [ 8 , 9 ] and converted into FMN Int. J. Mol. Sci. 2019,20, 6203; doi:10.3390/ijms20246203 www.mdpi.com/journal/ijms Int. J. Mol. Sci. 2019,20, 6203 2 of 17 and FAD via the sequential action of Rf kinase (RFK, EC 2.7.1.26) and FAD synthase or FMN-ATP adenylyl transferase (FADS, EC 2.7.7.2). The only gene identified for coding functional FAD synthases in humans is FLAD1 gene (GenBank, A. N. DQ458779, [ 10 ]) localized on chromosome 1, which is orthologue of flad-1 in C. elegans, [ 11 ], Fad1 in S. cerevisiae [ 12 ] and FMNAT in C. glabrata [ 13 ]. The structures of FADSs from yeast, but not that of the human orthologue, have been solved [13,14]. The FLAD1 gene in humans generates different alternative transcript variants, with unknown differential expression profile, producing protein isoforms, only in part characterized, which have different subcellular localization [15] and domain organizations. The most abundant variant in all the tissues and cells tested so far, is isoform 2 (NM_201398.3 in NCBI GenBank), which corresponds to a cytosolic enzyme of 490 amino acids [ 15 ]. The FADS or FMN-AT module of this protein is localized in the C-terminus of the protein; it contains a phosphoadenosine 5-phosphosulfate (PAPS) reductase domain and it is fused with a molybdopterin binding resembling (MPTb) domain located in the N-terminus [16,17]. FAD synthesis catalyzed by hFADS2 follows a bi-bi ordered kinetics with ATP entering prior to FMN and pyrophosphate released before FAD [ 17 ]. This enzyme contains 10 cysteines, some of which are relevant for catalysis; two of them are stably reduced (C139 and C241, one for each protein domain), four are implicated in stable disulfide bridges (C399 to C402, C303 to C312, both in the PAPS domain), and the other four are forming redox sensitive disulfides (C39 to C50; C440 to C464) [18]. Following the discovery that hFADS2 is a bifunctional enzyme, with the N-terminal domain working as a Co 2+ -dependent FAD hydrolase, the two domains of the protein were renamed as FADHy and FADSy, functionally corresponding to E.C. 3.6.1.18 and E.C. 2.7.7.2, respectively [ 19 , 20 ]. FADHy is also present in the other three isoforms of the protein reported in the NCBI GenBank (FLAD1 isoform 1, NM_025207.5; FLAD1 isoform 3, NM_001184891.2; and FLAD1 isoform 4, NM_001184892.2). Nonetheless, it is indeed absent in the FAD1p yeast counterparts and, more interestingly, in another human isoform not yet annotated as FADS, but reported as CRA_d in NCBI and known as hFADS6. The isoform hFADS6 is a 320-residue long protein, containing the sole FADSy domain, whose corresponding transcript was recently described in the frame of studying FLAD1 mutations leading to RR-MADD [ 7 ]. The relevance of this novel isoform lies in its ability to ensure FAD supply to patients carrying frameshift mutations in exon 2 of the FLAD1 gene and, for this reason, it has been named an “emergency protein” for MADD patients. When produced in E. coli and purified at homogeneity, hFADS6 behaves as a yellow monomer, able to tightly, but not covalently, bind FAD. Recombinant hFADS6 is more stable than hFADS2 and is able to perform FAD synthesis starting from FMN and ATP. As expected, it is not able to perform FAD hydrolysis [ 21 ]. The molecular features of this novel natural form well correlate with those of a previously produced artificial construct, lacking the first 231 residues of hFADS2, which per se can fold and catalyze the FAD synthesis reaction [ 22 ]. Therefore, this novel isoform of FAD forming enzymes is, in our opinion, a good model to address remaining challenges in the catalytic behavior of the FAD synthesis reaction in humans, as compared with the our deeper understanding of the yeast orthologues [ 13 , 14 ], from which structure the human protein was modeled [ 21 ]. The more striking point concerning the catalytic cycle of FAD forming enzyme concerns the observation that the turn-over number of the reaction, as catalyzed by hFADS2, is quite low ( 0.069 ±0.011 s−1 ), with FAD release being the limiting step of the over-all reaction [ 23 ]. This apparently sounds strange for a protein which is expected to be devoted to FAD delivery. We postulate that redox events or protein–protein interaction in a sort of chaperoning process may promote cofactor delivery to cognate apo-flavoprotein [17,23]. The aim of this work was to confirm the proposed mechanism by proving that lowering the FAD affinity toward the catalytic site (i.e., facilitating FAD release) results in increasing the turn-over number of the FAD forming reaction. To obtain these results we took into account studies on Candida glabrata FMN-AT (CgFMN-AT) [ 24 ] that identified a cryptic residue, precisely D181, whose mutation to Ala resulted in an increase of the V max , and therefore it was named “super-mutant”. The orthologous Int. J. Mol. Sci. 2019,20, 6203 3 of 17 residue in the human enzyme is D238. We investigate, here, the effects of the D238A mutation on hFADS6 steady-state activity and binding capability to confirm our hypothesis that a weak FAD binding to the active site could led to an increased rate of synthesis. The possible final goal of this study is to open a perspective towards increasing the emergency enzyme activity in patients suffering for FLAD1 mutations. 2. Results 2.1. Homology Model of D238A-hFADS6 The amino acid sequences of WT and D238A mutant hFADS6 were aligned by Clustal Omega software with the FMN-AT protein Q6FNA9 of C. glabrata. The alignment presented in Figure 1 highlights not only a high percentage identity (about 32%) between the human and the yeast proteins, but also confirms that D238 of hFADS6 corresponds to D181 of the yeast orthologue and that it is part of the flavin binding motif. Int. J. Mol. Sci. 2019, 20, x FOR PEER REVIEW 3 of 17 mutation on hFADS6 steady-state activity and binding capability to confirm our hypothesis that a weak FAD binding to the active site could led to an increased rate of synthesis. The possible final goal of this study is to open a perspective towards increasing the emergency enzyme activity in patients suffering for FLAD1 mutations. 2. Results 2.1. Homology Model of D238A-hFADS6 The amino acid sequences of WT and D238A mutant hFADS6 were aligned by Clustal Omega software with the FMN-AT protein Q6FNA9 of C. glabrata. The alignment presented in Figure 1 highlights not only a high percentage identity (about 32%) between the human and the yeast proteins, but also confirms that D238 of hFADS6 corresponds to D181 of the yeast orthologue and that it is part of the flavin binding motif. Figure 1. Sequence alignment of CgFMN-AT (Q6FNA9), hFADS6, and hFADS6D238A. The protein sequences were aligned by Clustal Omega software. D181 of the yeast protein, D238 of the hFADS6, and A238 in the mutant are shadowed in grey. Amino acids forming the flavin binding motif are highlighted by boxes. Due to the lack of suitable templates for modeling the N-terminus (amino acids 1 to 108) of hFADS6, an ab initio strategy was previously adopted to obtain its three-dimensional (3D) structural model [21]. FAD was then inserted in the active site according to the structure of CgFMN-AT (3G6K) as described in Materials and Methods. Figure 2 clearly shows a reduced steric hindrance upon substitution of D238 with A. However, the major reason for varied FAD binding and release kinetics should be the loss of the dipole interaction between the Asp-carboxylate group and the N(3)H-FAD, which most affects FAD binding. Figure 1. Sequence alignment of CgFMN-AT (Q6FNA9), hFADS6, and hFADS6 D238A . The protein sequences were aligned by Clustal Omega software. D181 of the yeast protein, D238 of the hFADS6, and A238 in the mutant are shadowed in grey. Amino acids forming the flavin binding motif are highlighted by boxes. Due to the lack of suitable templates for modeling the N-terminus (amino acids 1 to 108) of hFADS6, an ab initio strategy was previously adopted to obtain its three-dimensional (3D) structural model [ 21 ]. FAD was then inserted in the active site according to the structure of CgFMN-AT (3G6K) as described in Materials and Methods. Figure 2clearly shows a reduced steric hindrance upon substitution of D238 with A. However, the major reason for varied FAD binding and release kinetics Int. J. Mol. Sci. 2019,20, 6203 4 of 17 should be the loss of the dipole interaction between the Asp-carboxylate group and the N(3)H-FAD, which most affects FAD binding. Int. J. Mol. Sci. 2019, 20, x FOR PEER REVIEW 4 of 17 (b) (a) (c) Figure 2. Modeling and FAD docking of hFADS6: (a) Ribbon representation of hFADS6 modeled as described in [21]. FAD transferred in the active site from PDB ID: 3G6K and represented as yellow licorice colored by heteroatom type. (b and c) Zoom into the binding site of the isoalloxazine moiety of FAD. Amino acids involved in isoalloxazine binding are colored in pink. D238 in the WT (b) and A238 in the mutant (c) are rendered as green licorice colored by heteroatom type. 2.2. Cloning, Expression, and Purification of the D238A-hFADS6 Isoform The D238A mutant of the hFADS6 was constructed as described in Materials and Methods. The protein overexpressed in E. coli showed similar electrophoretic mobility to that of the WT [21] when purified by Ni-chelating chromatography, i.e., an apparent molecular mass of about 35 kDa. This value was compatible with the theoretical mass derived from the tagged sequence of the expressed polypeptide (38.222 kDa, Figure 3). Figure 3. Protein fractions obtained by Ni 2+ -chelating chromatography were separated by SDS–PAGE on 12% polyacrylamide gel and stained with Coomassie blue. Lane 1, insoluble fraction of IPTGinduced cell lysate (6 μg); lane 2, soluble fraction of IPTG-induced cell lysate (28 μg); lane 3, first flowR246 F244 W241 R246 F244 W241 D238 A238 Figure 2. Modeling and FAD docking of hFADS6: ( a ) Ribbon representation of hFADS6 modeled as described in [ 21 ]. FAD transferred in the active site from PDB ID: 3G6K and represented as yellow licorice colored by heteroatom type. ( b , c ) Zoom into the binding site of the isoalloxazine moiety of FAD. Amino acids involved in isoalloxazine binding are colored in pink. D238 in the WT ( b ) and A238 in the mutant (c) are rendered as green licorice colored by heteroatom type. 2.2. Cloning, Expression, and Purification of the D238A-hFADS6 Isoform The D238A mutant of the hFADS6 was constructed as described in Materials and Methods. The protein overexpressed in E. coli showed similar electrophoretic mobility to that of the WT [ 21 ] when purified by Ni-chelating chromatography, i.e., an apparent molecular mass of about 35 kDa. This value was compatible with the theoretical mass derived from the tagged sequence of the expressed polypeptide (38.222 kDa, Figure 3). Int. J. Mol. Sci. 2019,20, 6203 5 of 17 Int. J. Mol. Sci. 2019, 20, x FOR PEER REVIEW 4 of 17 (b) (a) (c) Figure 2. Modeling and FAD docking of hFADS6: (a) Ribbon representation of hFADS6 modeled as described in [21]. FAD transferred in the active site from PDB ID: 3G6K and represented as yellow licorice colored by heteroatom type. (b and c) Zoom into the binding site of the isoalloxazine moiety of FAD. Amino acids involved in isoalloxazine binding are colored in pink. D238 in the WT (b) and A238 in the mutant (c) are rendered as green licorice colored by heteroatom type. 2.2. Cloning, Expression, and Purification of the D238A-hFADS6 Isoform The D238A mutant of the hFADS6 was constructed as described in Materials and Methods. The protein overexpressed in E. coli showed similar electrophoretic mobility to that of the WT [21] when purified by Ni-chelating chromatography, i.e., an apparent molecular mass of about 35 kDa. This value was compatible with the theoretical mass derived from the tagged sequence of the expressed polypeptide (38.222 kDa, Figure 3). Figure 3. Protein fractions obtained by Ni2+-chelating chromatography were separated by SDS–PAGE on 12% polyacrylamide gel and stained with Coomassie blue. Lane 1, insoluble fraction of IPTGR246 F244 W241 R246 F244 W241 D238 244 A238 244 Figure 3. Protein fractions obtained by Ni 2+ -chelating chromatography were separated by SDS–PAGE on 12% polyacrylamide gel and stained with Coomassie blue. Lane 1, insoluble fraction of IPTG-induced cell lysate (6 µ g); lane 2, soluble fraction of IPTG-induced cell lysate (28 µ g); lane 3, first flow-through fraction (19 µ g), lane 4, proteins eluted with 50 mM imidazole (13 µ g), lane 5, first fraction of proteins eluted with 100 mM imidazole (2 µ g), lane 6, second fraction of proteins eluted with 100 mM imidazole (3 µ g), lane 7, third fraction of proteins eluted with 100 mM imidazole (2 µ g); lane 8, first fraction of proteins eluted with 250 mM imidazole (5 µ g); lane 9, second fraction of proteins eluted with 250 mM imidazole (1 µg); and lane 10, molecular mass markers. The overexpressed protein was recovered in the soluble fraction of the cell lysate and was used for spectrophotometric analysis and compared with the WT protein. The spectrum of the most abundant purified mutant protein fraction (fraction 8), whose purity is higher than 98%, is shown in Figure 4. An absorbance peak at 274 nm was observed, but surprisingly, no other peaks were present at higher absorption values. As the control, the WT hFADS6 spectrum was shown, presenting two additional minor peaks, at 350 and 450 nm, typical of oxidized FAD (Figure 4). From the absorbance value at 280 and 450 nm a ratio FAD/hFADS6 of 0.43 was calculated, as reported in [ 21 ], and thus indicating the presence of some apoprotein. The spectra of fractions 5, 6, 7, and 9 showed the same features as fraction 8. The spectral features of the D238A mutant hFADS6 indicated that this protein does not stabilize the strong binding of flavins. Int. J. Mol. Sci. 2019, 20, x FOR PEER REVIEW 5 of 17 induced cell lysate (6 μg); lane 2, soluble fraction of IPTG-induced cell lysate (28 μg); lane 3, first flowthrough fraction (19 µg), lane 4, proteins eluted with 50 mM imidazole (13 µg), lane 5, first fraction of proteins eluted with 100 mM imidazole (2 µg), lane 6, second fraction of proteins eluted with 100 mM imidazole (3 µg), lane 7, third fraction of proteins eluted with 100 mM imidazole (2 μg); lane 8, first fraction of proteins eluted with 250 mM imidazole (5 μg); lane 9, second fraction of proteins eluted with 250 mM imidazole (1 μg); and lane 10, molecular mass markers. The overexpressed protein was recovered in the soluble fraction of the cell lysate and was used for spectrophotometric analysis and compared with the WT protein. The spectrum of the most abundant purified mutant protein fraction (fraction 8), whose purity is higher than 98%, is shown in Figure 4. An absorbance peak at 274 nm was observed, but surprisingly, no other peaks were present at higher absorption values. As the control, the WT hFADS6 spectrum was shown, presenting two additional minor peaks, at 350 and 450 nm, typical of oxidized FAD (Figure 4). From the absorbance value at 280 and 450 nm a ratio FAD/hFADS6 of 0.43 was calculated, as reported in [21], and thus indicating the presence of some apoprotein. The spectra of fractions 5, 6, 7, and 9 showed the same features as fraction 8. The spectral features of the D238A mutant hFADS6 indicated that this protein does not stabilize the strong binding of flavins. Figure 4. Absorption spectra of hFADS6 and D238A-hFADS6 purified to homogeneity. The spectra of WT hFADS6 (8.6 μM, grey dotted line) and of D238A-hFADS6 (9.7 µM, black line) were recorded in 40 mM HEPES/Na, 5 mM β-mercaptoethanol, pH 7.4. The protein concentration was measured as indicated in [21]. 2.3. Kinetics of the D238A hFADS6 The mutant protein was characterized in terms of kinetics to uncover possible variations of interactions with the substrates or the effectors and inhibitors. Figure 5 shows the response of the mutant to Mg2+, which is a known effector of hFADSs, including the hFADS6 isoform [21,23]. The presence of Mg2+ is also essential for activity in the mutant. However, a different behavior was observed. The AC50 (half maximum activation constant) of Mg2+ is quite higher in the mutant than in the WT. Its value 3.5 ± 0.9 mM is twenty-fold that of the WT. Moreover, a higher value of Vmax is measured in the case of the mutant. Another distinctive feature is the inhibition observed at higher Mg2+ concentrations, which was not present in the WT. The effect of Hg2+ was also studied because this heavy metal typically inhibits the WT protein due to interaction with the Cys residues [18,21,23]. Differently from the case of Mg2+, the effect of Hg2+ on the mutant was very similar to that on the WT enzyme. Figure 4. Absorption spectra of hFADS6 and D238A-hFADS6 purified to homogeneity. The spectra of WT hFADS6 (8.6 µ M, grey dotted line) and of D238A-hFADS6 (9.7 µ M, black line) were recorded in 40 mM HEPES/Na, 5 mM β -mercaptoethanol, pH 7.4. The protein concentration was measured as indicated in [21]. Int. J. Mol. Sci. 2019,20, 6203 6 of 17 2.3. Kinetics of the D238A hFADS6 The mutant protein was characterized in terms of kinetics to uncover possible variations of interactions with the substrates or the effectors and inhibitors. Figure 5shows the response of the mutant to Mg 2+ , which is a known effector of hFADSs, including the hFADS6 isoform [ 21 , 23 ]. The presence of Mg 2+ is also essential for activity in the mutant. However, a different behavior was observed. The AC 50 (half maximum activation constant) of Mg 2+ is quite higher in the mutant than in the WT. Its value 3.5 ± 0.9 mM is twenty-fold that of the WT. Moreover, a higher value of V max is measured in the case of the mutant. Another distinctive feature is the inhibition observed at higher Mg 2+ concentrations, which was not present in the WT. The effect of Hg 2+ was also studied because this heavy metal typically inhibits the WT protein due to interaction with the Cys residues [ 18 , 21 , 23 ]. Differently from the case of Mg 2+ , the effect of Hg 2+ on the mutant was very similar to that on the WT enzyme. Int. J. Mol. Sci. 2019, 20, x FOR PEER REVIEW 6 of 17 (a) (b) Figure 5. Fluorimetric evidence of FAD synthesis. The FAD synthesis reaction was started by the addition of purified recombinant proteins (hFADS6 open circle or D238A-hFADS6 closed circle) and measured by the initial rate of fluorescence decrease (λ excitation = 450 nm, λ emission = 520 nm). (a) Dependence on the MgCl2 concentration. FAD synthesis rate, catalyzed by purified hFADS6 (open circle, 10 µg, 0.26 nmoL) or D238A-hFADS6 (closed circle, 2.3 µg, 0.06 nmol), was fluorimetrically measured at 37 °C in 2 mL of 50 mM Tris/HCl pH 7.5, in the presence of 100 µM ATP, 2 µM or 4 µM FMN, respectively, and of the given MgCl2 concentrations. (b) Inactivation by the mercurial reagent HgCl2. FAD synthesis rate, catalyzed by purified hFADS6 (open circle, 10 µg, 0.26 nmoL) or D238AhFADS6 (closed circle, 3.2 µg, 0.08 nmol), was fluorimetrically measured at 37 °C in 2 mL of 50 mM Tris/HCl pH 7.5, in the presence of 2 µM or 3 µM FMN, respectively, 100 µM ATP, 5 mM MgCl2 and of the given HgCl2 concentrations. The values of V0 are reported as nmol min−1 mg−1 (a) and as percentages of the maximum rate (b) arbitrarily set to 100%. Data points are fitted according to the Michaelis–Menten equation (a) and according to the IC50 equation (b) with Grafit 3.0 software. Then, the dependence of the FAD synthesis rate on the concentrations of the main substrates FMN and ATP was studied in the presence of saturating concentration of the counter-substrate, i.e., ATP and FMN, respectively. As shown in Figure 6a, dependences on the ATP concentration for the WT and the mutant were quite different. Both the Km and the Vmax increased in the mutant. The Km for ATP increased from 6.9 ± 0.5 µM (WT) to 44 ± 4 µM (mutant). The Vmax increased from 79 ± 1 (WT) to 145 ± 5 nmol min−1 mg−1 protein. The derived kcat of the mutant was 0.093 ± 0.003 s−1 (5.6 ± 0.2 min−1) which is about double than that of the WT 0.050 ± 0.001 s−1 (3.0 ± 0.1 min−1) [21]. To graphically highlight the difference in Km, the same data were expressed as percentage of the Vmax (Figure 6a’). A similar behavior was observed when measuring the dependence of the reaction rate on FMN concentrations. Again, the Km for FMN and the Vmax increased in the mutant (Figure 6b). The Km for FMN increased from 0.13 ± 0.01 µM (WT) to 1.3 ± 0.3 µM (mutant). The Vmax increased from 74 ± 1 (WT) to 162 ± 12 nmol min−1 mg−1 protein (mutant). The derived kcat of the mutant was 0.103 ± 0.007 s−1 (6.2 ± 0.5 min−1) which is about twice that of the WT 0.047 ± 0.001 s−1 (2.9 ± 0.1 min−1) [21], very similar to that obtained for the ATP kinetics. In this case, the same data were represented as percentage of the Vmax (Figure 6b’). Table 1 summarizes the kinetic comparison between WT and D238A hFADS6. Figure 5. Fluorimetric evidence of FAD synthesis. The FAD synthesis reaction was started by the addition of purified recombinant proteins (hFADS6 open circle or D238A-hFADS6 closed circle) and measured by the initial rate of fluorescence decrease ( λ excitation =450 nm, λ emission =520 nm). ( a ) Dependence on the MgCl 2 concentration. FAD synthesis rate, catalyzed by purified hFADS6 (open circle, 10 µ g, 0.26 nmoL) or D238A-hFADS6 (closed circle, 2.3 µ g, 0.06 nmol), was fluorimetrically measured at 37 ◦ C in 2 mL of 50 mM Tris/HCl pH 7.5, in the presence of 100 µ M ATP, 2 µ M or 4 µ M FMN, respectively, and of the given MgCl 2 concentrations. ( b ) Inactivation by the mercurial reagent HgCl 2 . FAD synthesis rate, catalyzed by purified hFADS6 (open circle, 10 µ g, 0.26 nmoL) or D238A-hFADS6 (closed circle, 3.2 µ g, 0.08 nmol), was fluorimetrically measured at 37 ◦ C in 2 mL of 50 mM Tris/HCl pH 7.5, in the presence of 2 µ M or 3 µ M FMN, respectively, 100 µ M ATP, 5 mM MgCl 2 and of the given HgCl 2 concentrations. The values of V 0 are reported as nmol min −1 mg −1 ( a ) and as percentages of the maximum rate ( b ) arbitrarily set to 100%. Data points are fitted according to the Michaelis–Menten equation (a) and according to the IC50 equation (b) with Grafit 3.0 software. Then, the dependence of the FAD synthesis rate on the concentrations of the main substrates FMN and ATP was studied in the presence of saturating concentration of the counter-substrate, i.e., ATP and FMN, respectively. As shown in Figure 6a, dependences on the ATP concentration for the WT and the mutant were quite different. Both the K m and the V max increased in the mutant. The K m for ATP increased from 6.9 ± 0.5 µ M (WT) to 44 ± 4 µ M (mutant). The V max increased from 79 ±1 (WT) to 145 ±5 nmol min−1mg−1protein . The derived k cat of the mutant was 0.093 ±0.003 s−1 ( 5.6 ±0.2 min−1 ) which is about double than that of the WT 0.050 ± 0.001 s −1 (3.0 ± 0.1 min −1 ) [ 21 ]. To graphically highlight the difference in K m , the same data were expressed as percentage of the V max (Figure 6a’). A similar behavior was observed when measuring the dependence of the reaction rate on FMN concentrations. Again, the K m for FMN and the V max increased in the mutant (Figure 6b). The K m Int. J. Mol. Sci. 2019,20, 6203 7 of 17 for FMN increased from 0.13 ± 0.01 µ M (WT) to 1.3 ± 0.3 µ M (mutant). The V max increased from 74 ± 1 (WT) to 162 ± 12 nmol min −1 mg −1 protein (mutant). The derived k cat of the mutant was 0.103 ±0.007 s−1 (6.2 ± 0.5 min −1 ) which is about twice that of the WT 0.047 ± 0.001 s −1(2.9 ±0.1 min−1) [21] , very similar to that obtained for the ATP kinetics. In this case, the same data were represented as percentage of the V max (Figure 6b’). Table 1summarizes the kinetic comparison between WT and D238A hFADS6. Int. J. Mol. Sci. 2019, 20, x FOR PEER REVIEW 7 of 17 (a) (a’) (b) (b’) Figure 6. Fluorimetric evidence of FAD synthesis. The FAD synthesis reaction was started by the addition of purified recombinant proteins (hFADS6 open circle or D238A-hFADS6 closed circle) and measured by the initial rate of fluorescence decrease (λ excitation = 450 nm and λ emission = 520 nm). V0 was expressed as nmol min−1 mg−1 (a and b) and as a percentage of the Vmax value (a’ and b’) set arbitrarily to 100%. Data points are fitted according to the Michaelis–Menten equation with Grafit 3.0 software. (a) ATP concentration dependence. FAD synthesis rate, catalyzed by purified hFADS6 (open circle, 10 µg, 0.26 nmoL) or D238A-hFADS6 (closed circle, 4.2 µg, 0.11 nmoL), was fluorimetrically measured at 37 °C in 2 mL of 50 mM Tris⁄HCl pH 7.5, in the presence of 5 mM MgCl2, 2 µM or 3 µM FMN, respectively, and of the given ATP concentrations. (b) FMN concentration dependence. FAD synthesis rate catalyzed by purified hFADS6 (open circle, 10 µg, 0.26 nmoL) or D238A-hFADS6 (closed circle, 3.9 µg, 0.10 nmoL) was fluorimetrically measured at 37 °C in 2 mL of 50 mM Tris⁄HCl pH 7.5, in the presence of 5 mM MgCl2, 100 µM ATP, and of the given FMN concentrations. The reverse reaction, i.e., pyrophosphorolysis was also revealed by using recombinant FADS enzymes [21,23]. The kinetics of this reverse reaction was measured for the mutant in comparison with the WT. In the reverse reaction, the Km for NaPPi (Figure 7a) and for FAD (Figure 7b) increased in the mutant, while the Vmax remained very similar. Even though an increase of Km for FAD (0.045 ± 0.008 µM) is clearly evident in the mutant, the Km value for the WT could not be accurately measured due to both instrumental limitations and to the presence of tightly bound FAD in the active site of the WT protein [21,23]. Therefore, to measure a reliable Km of FAD for the WT, the apo-enzyme was prepared autocatalytically, i.e., by incubating the purified protein at 37 °C for 10 min in the presence of MgCl2 and NaPPi and in the absence of externally added FAD, allowing the reverse reaction to occur using the endogenous bound FAD. At this stage, after endogenous FAD conversion to FMN, external FAD was added, and the reverse reaction started (Figure 7c). Under this condition, in which the reaction rate could be measured, being much lower than in the case of the holo-enzyme and, hence, an accurate value of Km could be calculated. As shown in Figure 7d, the Km for FAD of the WT Figure 6. Fluorimetric evidence of FAD synthesis. The FAD synthesis reaction was started by the addition of purified recombinant proteins (hFADS6 open circle or D238A-hFADS6 closed circle) and measured by the initial rate of fluorescence decrease ( λ excitation =450 nm and λ emission =520 nm). V 0 was expressed as nmol min −1 mg −1 ( a , b ) and as a percentage of the V max value ( a’ , b’ ) set arbitrarily to 100%. Data points are fitted according to the Michaelis–Menten equation with Grafit 3.0 software. ( a ) ATP concentration dependence. FAD synthesis rate, catalyzed by purified hFADS6 (open circle, 10 µ g, 0.26 nmoL) or D238A-hFADS6 (closed circle, 4.2 µ g, 0.11 nmoL), was fluorimetrically measured at 37 ◦ C in 2 mL of 50 mM Tris ⁄ HCl pH 7.5, in the presence of 5 mM MgCl 2 , 2 µ M or 3 µ M FMN, respectively, and of the given ATP concentrations. ( b ) FMN concentration dependence. FAD synthesis rate catalyzed by purified hFADS6 (open circle, 10 µ g, 0.26 nmoL) or D238A-hFADS6 (closed circle, 3.9 µ g, 0.10 nmoL) was fluorimetrically measured at 37 ◦ C in 2 mL of 50 mM Tris ⁄ HCl pH 7.5, in the presence of 5 mM MgCl2, 100 µM ATP, and of the given FMN concentrations. Int. J. Mol. Sci. 2019,20, 6203 14 of 17 4.6. Measurements of Enzyme Catalyzed Rates for FAD Synthesis and FAD Pyrophosphorolysis The rate of FAD synthesis and FAD cleavage were measured in continuo as in [ 23 ], by exploiting the different fluorescence properties of FAD with respect to FMN. Fluorescence time courses ( λ excitation at 450 nm and λ emission at 520 nm) were followed at 37 ◦ C in a FP-8300 Jasco spectrofluorometer. In each experiment, FMN and FAD fluorescence were calibrated by using standard solutions whose concentrations (used in µ M range) were calculated by using ε450 of 12.2 mM −1 cm −1 for FMN and 11.3 mM −1 cm −1 for FAD. Under the experimental condition used here, the specific relative FAD fluorescence coefficient ( φFAD ) proved to be about ten times lower than those of FMN ( φFMN ) [ 30 , 31 ]. For FAD synthesis rate measurements, purified protein fractions (2 to 10 µ g, 0.06 to 0.26 nmoL protein as monomer, unless otherwise indicated) were incubated in 50 mM Tris ⁄ HCl, pH 7.5, containing 5 mM MgCl 2 , 3 µ M FMN, 100 µ M ATP, and additional reagents as appropriate. The rate of FAD synthesis, expressed as nmoL FAD min −1 mg protein −1 , was calculated from the rate of fluorescence decrease, measured as the tangent to the initial part of the experimental curve by applying the following equation: V0=[(∆F450/520/∆φ450/520)×Vf]/(t ×m) (1) where ∆ F is expressed in fluorescence arbitrary units, ∆φ = φFMN −φFAD is expressed as µ M −1 , Vf is expressed in mL, t is time expressed in min, and m is the mass of protein in mg. The rate of FAD pyrophosphorolysis catalyzed by 6His-D238A-hFADS6 (5 to 10 µ g, 0.13 to 0.26 nmoL protein as monomer) was measured in 50 mM Tris ⁄ HCl, pH 7.5, containing 5 mM MgCl 2 in the presence of 5 mM MgCl 2 , 1 mM NaPPi (sodium pyrophosphate), and 0.5 µ M FAD, unless otherwise indicated. The rate of FAD cleavage was expressed as nmol FAD min −1 mg protein-1, and was calculated from the rate of fluorescence increase, measured as the tangent to the initial part of the experimental curve, as previously described. In the case of using WT hFADS6, which is a FAD binding protein, the measurement of Km for FAD might not be accurate. To overcome this limitation, the apo-form of the protein was obtained by a rapid procedure in the cuvette simply by allowing autocatalytic endogenous bound FAD conversion to FMN by adding only 1 mM NaPPi. The reaction was followed for about 10 min until endogenous bound FAD is completely converted to FMN, i.e., the fluorescence does not change. At this stage, external FAD was added at different concentrations and the dependence of the reaction rate was studied. 4.7. Kinetics for the Binding of Flavinic Ligands to WT hFADS6 and hFADS6-D238A Proteins Pre-steady-state kinetic experiments were performed using stopped-flow spectroscopy on an Applied Photophysics SX17.MV spectrophotometer using the Pro-Data SX software (Applied Photophysics Ltd. Leatherhead, Surrey, UK). The fluorescence of flavins was measured in a continuous assay with an excitation wavelength of 445 nm, while the emission was recovered using a >530 nm cut-offfilter and the voltage was set to 350 V. Then, 100 nM hFADS6 samples were mixed with samples that contained increasing concentrations of FMN or FAD (which varied in the range 0.025–1 µ M), in the absence and in the presence of saturating concentrations of their respective co-substrates ATP or PPi (250 µ M). All the indicated concentrations are final ones in the stopped-flow observation cell. Measurements were carried out at 25 ◦ C in 50 mM HEPES/NaOH, 10 mM MgCl 2 , pH 7.0, 5 mM β -mercaptoethanol. At least three reproducible kinetic traces of changes in fluorescence were recorded for each condition and fitted to exponential equations (usually one or two exponentials were used) y=XAie−kobs,it(2) where A i and k obs,i represent the amplitude and the observed rate constant, respectively, for each of the processes (i) that contribute to the overall time-dependent fluorescence change for each experimental condition. Processes whose k obs showed a linear dependence on flavin concentration were fit to a Int. J. Mol. Sci. 2019,20, 6203 15 of 17 one-step model that accounts for the kinetic equilibrium of the formation and dissociation of the protein-flavin complex, whose kinetics can be represented by kobs =kon[FLV]+koff(3) where kon and koffare the kinetic constants for complex formation and dissociation, respectively. 4.8. Isothermal Titration Calorimetry (ITC) Measurements were carried out using an Auto-ITC200 microcalorimeter (MicroCal LLC, Northampton, MA, USA) thermostated at 25 ◦ C. Ligand (100 µ M FMN or FAD) and proteins (~15 µ M) were dissolved in 50 mM HEPES/NaOH, 10 mM MgCl 2 , pH 7.0, 2 mM β -mercaptoethanol, and degassed prior to titration. Up to 19 injections of 2 µ L of ligand were added to the sample cell (~0.2 mL) containing the enzyme and then mixed via the rotating (1000 rpm) stirrer syringe. Since either no binding or slow binding was detected, data fitting to obtain thermodynamic binding parameters was not reliable. 4.9. Electrophoretic Analysis Proteins were separated by SDS-PAGE on 12% total polyacrylamide gels, according to Laemmli [ 32 ]. Quantitative evaluation of Coomassie blue-stained protein bands was carried out using the Chemidoc imaging system and the Quantity One software (Bio-Rad), as described previously [33]. Supplementary Materials: Supplementary materials can be found at http://www.mdpi.com/1422-0067/20/24/ 6203/s1. Author Contributions: M.G. and C.I. were involved in hFADS6 D238A cloning, expression, and modeling; P.L., S.Q., and M.B. were involved in protein purification and characterization; M.M. was involved in conceptualization and supervision of data on stopped-flow analysis; E.A.-C. was involved in data curation, formal analysis, and methodology investigation. All authors discussed the results and contributed to writing the manuscript. M.B. coordinated and supervised the work and the manuscript writing. Funding: This work was supported by “Fondi di Ateneo” the Universit à degli Studi di Bari (to M.B.); “Effetto di mutazioni di FLAD1 e di alterazioni dell’omeostasi delle flavine sullo stato redox e sulla biogenesi mitocondriale: uno studio integrato su fibroblasti umani” the Universit à degli Studi di Bari (to M.B.); “Fondi di Ateneo” the Universit à della Calabria (to C.I. and M.G.); the Spanish Agencia Estatal de Investigaci ó n and Fondo Europeo de Desarrollo Regional BIO2016-75183-P AEI/FEDER, UE (to M.M.), and the Government of Arag ó n-FEDER Grupo de Referencia Biología Estructural E35_17R (to M.M.). Acknowledgments: The helpful collaboration of Maria Tolomeo (University of Bari) who participated in preparation of cellular cultures, and the technical assistance of Vito Giannoccaro (University of Bari) are gratefully acknowledged. Authors would like to acknowledge the use of Servicio General de Apoyo a la Investigaci ó n-SAI, Universidad de Zaragoza. Conflicts of Interest: The authors declare no conflict of interest. Abbreviations FADS FAD synthase FMNAT FMN adenylyl transferase hFADS6 human FAD synthase isoform 6 Rf riboflavin FMN flavin mono nucleotide FAD flavin adenine dinucleotide PAPS phosphoadenosine 5-phosphosulfate FADSy FAD synthase domain MPTb molybdopterin binding FADHy FAD hydrolase domain BVVLS Brown–Vialetto-van Laere syndrome RR-MADD PMSF riboflavin responsive multiple acyl-CoA dehydrogenase deficiency phenylmethyl sulfonyl fluoride Int. J. Mol. Sci. 2019,20, 6203 16 of 17 References 1. Barile, M.; Giancaspero, T.A.; Leone, P.; Galluccio, M.; Indiveri, C. Riboflavin transport and metabolism in humans. J. Inherit. Metab. Dis. 2016,39, 545–557. [CrossRef] [PubMed] 2. Lienhart, W.D.; Gudipati, V.; Macheroux, P. The human flavoproteome. Arch. Biochem. Biophys. 2013 ,535, 150–162. [CrossRef] [PubMed] 3. Green, P.; Wiseman, M.; Crow, Y.J.; Houlden, H.; Riphagen, S.; Lin, J.P.; Raymond, F.L.; Childs, A.M.; Sheridan, E.; Edwards, S.; et al. Brown-Vialetto-Van Laere syndrome, a ponto-bulbar palsy with deafness, is caused by mutations in c20orf54. Am. J. Hum. Genet. 2010,86, 485–489. [CrossRef] [PubMed] 4. Jaeger, B.; Bosch, A.M. Clinical presentation and outcome of riboflavin transporter deficiency: mini review after five years of experience. J. Inherit. Metab. Dis. 2016,39, 559–564. [CrossRef] [PubMed] 5. Ryder, B.; Tolomeo, M.; Nochi, Z.; Colella, M.; Barile, M.; Olsen, R.K.; Inbar-Feigenberg, M. A Novel Truncating FLAD1 Variant, Causing Multiple Acyl-CoA Dehydrogenase Deficiency (MADD) in an 8-Year-Old Boy. JIMD Rep. 2019,45, 37–44. [PubMed] 6. Balasubramaniam, S.; Christodoulou, J.; Rahman, S. Disorders of riboflavin metabolism. J. Inherit. Metab. Dis. 2019,42, 608–619. [CrossRef] 7. Olsen, R.K.J.; Konarikova, E.; Giancaspero, T.A.; Mosegaard, S.; Boczonadi, V.; Matakovic, L.; Veauville-Merllie, A.; Terrile, C.; Schwarzmayr, T.; Haack, T.B.; et al. Riboflavin-Responsive and -Non-responsive Mutations in FAD Synthase Cause Multiple Acyl-CoA Dehydrogenase and Combined Respiratory-Chain Deficiency. Am. J. Hum. Genet. 2016,98, 1130–1145. [CrossRef] 8. Yonezawa, A.; Inui, K. Novel riboflavin transporter family RFVT/SLC52: identification, nomenclature, functional characterization and genetic diseases of RFVT/SLC52. Mol. Asp. Med. 2013 ,34, 693–701. [CrossRef] 9. Console, L.; Tolomeo, M.; Colella, M.; Barile, M.; Indiveri, C. Reconstitution in Proteoliposomes of the Recombinant Human Riboflavin Transporter 2 (SLC52A2) Overexpressed in E. coli.Int. J. Mol. Sci. 2019 , 20, 4416. [CrossRef] 10. Brizio, C.; Galluccio, M.; Wait, R.; Torchetti, E.M.; Bafunno, V.; Accardi, R.; Gianazza, E.; Indiveri, C.; Barile, M. Over-expression in Escherichia coli and characterization of two recombinant isoforms of human FAD synthetase. Biochem. Biophys. Res. Commun. 2006,344, 1008–1016. [CrossRef] 11. Liuzzi, V.C.; Giancaspero, T.A.; Gianazza, E.; Banfi, C.; Barile, M.; De Giorgi, C. Silencing of FAD synthase gene in Caenorhabditis elegans upsets protein homeostasis and impacts on complex behavioral patterns. Biochim. Biophys. Acta 2012,1820, 521–531. [CrossRef] [PubMed] 12. Wu, M.; Repetto, B.; Glerum, D.M.; Tzagoloff, A. Cloning and characterization of FAD1, the structural gene for flavin adenine dinucleotide synthetase of Saccharomyces cerevisiae. Mol. Cell Biol. 1995 ,15, 264–271. [CrossRef] [PubMed] 13. Huerta, C.; Borek, D.; Machius, M.; Grishin, N.V.; Zhang, H. Structure and mechanism of a eukaryotic FMN adenylyltransferase. J. Mol. Biol. 2009,389, 388–400. [CrossRef] [PubMed] 14. Leulliot, N.; Blondeau, K.; Keller, J.; Ulryck, N.; Quevillon-Cheruel, S.; van Tilbeurgh, H. Crystal structure of yeast FAD synthetase (Fad1) in complex with FAD. J. Mol. Biol. 2010,398, 641–646. [CrossRef] [PubMed] 15. Torchetti, E.M.; Brizio, C.; Colella, M.; Galluccio, M.; Giancaspero, T.A.; Indiveri, C.; Roberti, M.; Barile, M. Mitochondrial localization of human FAD synthetase isoform 1. Mitochondrion 2010 ,10, 263–273. [CrossRef] [PubMed] 16. Barile, M.; Giancaspero, T.A.; Brizio, C.; Panebianco, C.; Indiveri, C.; Galluccio, M.; Vergani, L.; Eberini, I.; Gianazza, E. Biosynthesis of flavin cofactors in man: implications in health and disease. Curr. Pharm. Des. 2013,19, 2649–2675. [CrossRef] [PubMed] 17. Giancaspero, T.A.; Colella, M.; Brizio, C.; Difonzo, G.; Fiorino, G.M.; Leone, P.; Brandsch, R.; Bonomi, F.; Iametti, S.; Barile, M. Remaining challenges in cellular flavin cofactor homeostasis and flavoprotein biogenesis. Front. Chem. 2015,3, 30. [CrossRef] 18. Miccolis, A.; Galluccio, M.; Nitride, C.; Giancaspero, T.A.; Ferranti, P.; Iametti, S.; Indiveri, C.; Bonomi, F.; Barile, M. Significance of redox-active cysteines in human FAD synthase isoform 2. Biochim. Biophys. Acta 2014,1844, 2086–2095. [CrossRef] Int. J. Mol. Sci. 2019,20, 6203 17 of 17 19. Giancaspero, T.A.; Galluccio, M.; Miccolis, A.; Leone, P.; Eberini, I.; Iametti, S.; Indiveri, C.; Barile, M. Human FAD synthase is a bi-functional enzyme with a FAD hydrolase activity in the molybdopterin binding domain. Biochem. Biophys. Res. Commun. 2015,465, 443–449. [CrossRef] 20. Leone, P.; Galluccio, M.; Brizio, C.; Barbiroli, A.; Iametti, S.; Indiveri, C.; Barile, M. The hidden side of the human FAD synthase 2. Int. J. Biol. Macromol. 2019,138, 986–995. [CrossRef] 21. Leone, P.; Galluccio, M.; Barbiroli, A.; Eberini, I.; Tolomeo, M.; Vrenna, F.; Gianazza, E.; Iametti, S.; Bonomi, F.; Indiveri, C.; et al. Bacterial Production, Characterization and Protein Modeling of a Novel Monofuctional Isoform of FAD Synthase in Humans: An Emergency Protein? Molecules 2018,23, 116. [CrossRef] 22. Miccolis, A.; Galluccio, M.; Giancaspero, T.A.; Indiveri, C.; Barile, M. Bacterial over-expression and purification of the 3’phosphoadenosine 5’phosphosulfate (PAPS) reductase domain of human FAD synthase: functional characterization and homology modeling. Int. J. Mol. Sci. 2012,13, 16880–16898. [CrossRef] 23. Torchetti, E.M.; Bonomi, F.; Galluccio, M.; Gianazza, E.; Giancaspero, T.A.; Iametti, S.; Indiveri, C.; Barile, M. Human FAD synthase (isoform 2): A component of the machinery that delivers FAD to apo-flavoproteins. FEBS J. 2011,278, 4434–4449. [CrossRef] 24. Huerta, C.; Grishin, N.V.; Zhang, H. The “super mutant” of yeast FMN adenylyltransferase enhances the enzyme turnover rate by attenuating product inhibition. Biochemistry 2013,52, 3615–3617. [CrossRef] 25. Weber, G. Fluorescence of riboflavin and flavin-adenine dinucleotide. Biochem. J. 1950 ,47, 114–121. [CrossRef] 26. Sebasti á n, M.; Vel á zquez-Campoy, A.; Medina, M. The RFK catalytic cycle of the pathogen Streptococcus pneumoniae shows species-specific features in prokaryotic FMN synthesis. J. Enzyme Inhib. Med. Chem. 2018 , 33, 842–849. [CrossRef] 27. Pebay-Peyroula, E.; Dahout-Gonzalez, C.; Kahn, R.; Trezeguet, V.; Lauquin, G.J.; Brandolin, G. Structure of mitochondrial ADP/ATP carrier in complex with carboxyatractyloside. Nature 2003,426, 39–44. [CrossRef] 28. Pettersen, E.F.; Goddard, T.D.; Huang, C.C.; Couch, G.S.; Greenblatt, D.M.; Meng, E.C.; Ferrin, T.E. UCSF Chimera—A visualization system for exploratory research and analysis. J. Comput. Chem. 2004 ,25, 1605–1612. [CrossRef] 29. Shapovalov, M.V.; Dunbrack, R.L., Jr. A smoothed backbone-dependent rotamer library for proteins derived from adaptive kernel density estimates and regressions. Structure 2011,19, 844–858. [CrossRef] 30. Barile, M.; Passarella, S.; Bertoldi, A.; Quagliariello, E. Flavin adenine dinucleotide synthesis in isolated rat liver mitochondria caused by imported flavin mononucleotide. Arch. Biochem Biophys 1993 ,305, 442–447. [CrossRef] 31. Barile, M.; Brizio, C.; De Virgilio, C.; Delfine, S.; Quagliariello, E.; Passarella, S. Flavin adenine dinucleotide and flavin mononucleotide metabolism in rat liver—The occurrence of FAD pyrophosphatase and FMN phosphohydrolase in isolated mitochondria. Eur. J. Biochem. 1997,249, 777–785. [CrossRef] 32. Laemmli, U.K. Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 1970,227, 680–685. [CrossRef] 33. Brizio, C.; Brandsch, R.; Bufano, D.; Pochini, L.; Indiveri, C.; Barile, M. Over-expression in Escherichia coli, functional characterization and refolding of rat dimethylglycine dehydrogenase. Protein Expr. Purif. 2004 ,37, 434–442. [CrossRef] © 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/). Int. J. Mol. Sci. 2019, 20, x; doi: FOR PEER REVIEW www.mdpi.com/journal/ijms Article Mutation of Aspartate 238 in FAD Synthase Isoform 6 Increases the Specific Activity by Weakening the FAD Binding Piero Leone 1,2, Michele Galluccio 2, Stefano Quarta 1, Ernesto Anoz-Carbonell 3, Milagros Medina 3, Cesare Indiveri 2 and Maria Barile 1,* 1 Department of Biosciences, Biotechnology and Biopharmaceutics, University of Bari, via Orabona 4 –, 70126 Bari, Italy; piero[email protected] (P.L.); [email protected] (S.Q.) 2 Department of Biology, Ecology and Earth Sciences (DiBEST), Unit of Biochemistry and Molecular Biotechnology, University of Calabria, via P. Bucci 4c, 87036 Arcavacata di Rende, Italy; [email protected] (M.G.); [email protected] (C.I.) 3 Departamento de Bioquímica y Biología Molecular y Celular, Facultad de Ciencias, Instituto de Biocomputación y Física de Sistemas Complejos (GBsC-CSIC and BIFI-IQFR Joint Units), Universidad de Zaragoza, Zaragoza 50009, Spain; [email protected] (E.A.-C.); [email protected] (M.M.) * Correspondence: mar[email protected] Supplementary Materials: Supplementary materials can be found at www.mdpi.com/xxx/s1. (a) (b) Supplementary Figure 1. (a) GTP inhibition on FAD synthesis catalyzed by D238A-hFADS6. The FAD synthesis reaction was started by the addition of purified recombinant protein D238A-hFADS6 and measured by the initial rate of fluorescence decrease (λ excitation = 450 nm, λ emission = 520 nm). FAD synthesis rate, catalyzed by purified D238A-hFADS6 (2.9 µg, 0.08 nmol), was fluorimetrically measured at 37 °C in 2 ml of 50 mM Tris/HCl pH 7.5, in the presence of 100 µM ATP, 5 mM MgCl2, 3 µM (closed circle) or 1.3 µM (open circle) FMN, and of the given GTP concentrations. Data points are fitted according to the linear equation with Grafit 3.0 software. (b) NAD+ and NADH inhibition on FAD synthesis. The FAD synthesis reaction was started by the addition of purified recombinant protein 6His-D238A-hFADS6 and measured by the initial rate of fluorescence decrease (λ excitation = 450 nm, λ emission = 520 nm). FAD synthesis rate, catalyzed by purified D238A-hFADS6 (8 µg, 0.21 nmol), was fluorimetrically measured at 37°C in 2 ml of 50 mM Tris/HCl pH 7.5, in the presence of 100 µM ATP, 5 mM MgCl2, 1.3 µM FMN, and of the given NAD+ (open circle) and NADH (closed circle) concentrations. Data points are fitted according to the exponential or linear equations for NADH and NAD+ respectively with Grafit 3.0 software. Int. J. Mol. Sci. 2019, 20, x FOR PEER REVIEW 2 of 4 (a) (b) Supplementary Figure 2. Isothermal titration calorimetric analysis of the binding of flavinic substrates to D238A hFADS6 in the absence of the second substrate. Thermogram (upper panels) and binding isotherms with integrated heat (lower panels) for the titration with (a) FMN and (b) FAD. No interaction heat is detected (a) for the titration of FMN, suggesting under the assayed conditions binding is not produced. Thermograms for the titration with FAD titration envisage binding to less than 20% of the protein molecules (suggesting a slow-binding process) under the assayed conditions, as well as KdFAD below 1 µM. Titrations were performed at 25 °C in 50 mM Hepes/NaOH, 10 mM MgCl2, pH 7.0, 5 mM -mercaptoethanol. The low stability of WT hFADS6 along the ITC assay prevented production of the corresponding thermograms. 0.0 0.5 1.0 1.5 -10 -8 -6 -4 -2 0 -0.10 -0.08 -0.06 -0.04 -0.02 0.00 0.02 0.04 010 20 30 40 50 Time (min) [FAD]T/[D238A hFADS6]T 0.0 0.5 1.0 1.5 -10 -8 -6 -4 -2 0 -0.04 -0.02 0.00 0.02 0.04 0.06 0.08 0.10 010 20 30 40 50 [FMN]T/[D238A hFADS6]T Time (min) dQ/dt (µcal/s) Q (kcal/mol of injectant) Int. J. Mol. Sci. 2019, 20, x FOR PEER REVIEW 3 of 4 (a) (b) (c) (d) Supplementary Figure 3. Dependence of flavin binding pre-steady-state kinetic parameters on the concentration of the flavin substrates. Evolution of binding parameters for the FAD biosynthesis: (a) kobs1 (closed circles) and kobs2 (open circles) and (b) their corresponding amplitudes, Amp1 (closed circles) and Amp2 (open circles), obtained when mixing WT (black) and D238A (red) hFADS6s with saturating ATP and different FMN concentrations. Evolution of binding parameters for the FAD pyrophosphorolysis: (c) kobs (closed circles) and (d) Amp (closed circles) obtained when mixing WT (black) and D238A (red) hFADS6s with saturating PPi and different FAD concentrations. Data were obtained at 25 °C in mixtures containing 100 nM of protein and 250 µM of either ATP or PPi (respectively for FAD synthesis and FAD pyrophosphorolysis) in 50 mM Hepes/NaOH, 10 mM MgCl2, pH 7.0, 5 mM -mercaptoethanol. 0.0 0.1 0.2 0.3 0.4 0.5 0.0 0.5 1.0 1.5 2.0 kobs (s -1 ) [FMN] (M) 0.0 0.1 0.2 0.3 0.4 0.5 0.00 0.05 0.10 0.15 0.20 0.25 0.30 Amplitude [FMN] (M) 0.0 0.2 0.4 0.6 0.8 1.0 0 1 2 3 4 5 6 7 kobs (s -1 ) [FAD] (M) 0.0 0.1 0.2 0.3 0.4 0.5 0.000 0.001 0.002 0.003 0.004 Amplitude [FAD] (M) Int. J. Mol. Sci. 2019, 20, x FOR PEER REVIEW 4 of 4 Supplementary Table 1. Kinetic parameters derived from stopped-flow data for the binding of FMN and FAD to WT and D238A hFADS6 in the presence of the second substrate (ATP or PPi, respectively). kon and koff the kinetic constants for complex formation and dissociation were obtained by fitting lineal data corresponding to kobs1 dependences on flavin concentration presented in Supplementary Figures 3a and 3c (n = 3, mean ± SD). Dissociation constants determined as koff/kon. Data obtained at 25 °C at 25 °C in 50 mM Hepes/NaOH, 10 mM MgCl2, pH 7.0, 5 mM -mercaptoethanol. FMN-ATP mixtures (FMN binding) hFADS6 kon (µM−1.s−1) koff (s−1) KdFMN (µM) WT 3.4 ± 0.1 0.32 ± 0.01 0.094 ± 0.004 D238A 0.55 a 0.09 a >0.18 a FAD-NaPPi mixtures (FAD binding) kon (µM−1.s−1) koff (s−1) KdFAD (µM) WT 4.5 ± 0.6 0.81 ± 0.05 0.18 ± 0.04 D238A 1.5 a 0.35 a > 0.3 a a Error in the determination of parameters for the mutant is high, at least ± 30%, due to small amplitudes in the exponential decays and/or low data reproducibility. Publication III The catalytic cycle of the antioxidant and cancer-associated human NQO1 enzyme: hydride transfer, conformational dynamics and functional cooperativity Antioxidants 2020,9, 772 6 of 22 added to all the solutions to remove trace amounts of oxygen. The stepwise reduction of the protein was achieved by light irradiation from a 250 W slide projector for different periods of time, for which the UV–visible spectrum was then recorded in a Cary 100 spectrophotometer (Agilent). Once fully reduced (no variation in the UV–visible spectra despite further irradiation), the protein was re-oxidized by breaking the anaerobic conditions and exposure to atmospheric air, and absorption spectra were recorded until complete re-oxidation. The ability of NAD + to re-oxidize the hydroquinone form of the protein (NQO1 hq ) was evaluated by using anaerobic solutions of photoreduced NQO1 (7.5 µ M), produced by following the above described procedure in specially designed cuvettes. NQO1 hq was them mixed with an NAD + solution placed in the same cuvette lateral arm, providing a final 1:1 protein:coenzyme ratio and the kinetics of the re-oxidation of the enzyme were followed in the full protein spectral range using an Agilent 8453 photodiodearray spectrophotometer (Agilent). 2.4. Stopped-Flow Pre-Steady-State Kinetic Measurements Fast HT reactions from NAD(P)H/D to NQO1 ox , as well as from NQO1 hq (generated by NQO1 ox mixed with NADH at stoichiometric ratio) to 2,6-dichlorophenol indophenol (DCPIP), were measured using a stopped-flow spectrophotometer from Applied Photophysics (SX.18MV, Applied Photophysics Ltd., Leatherhead, UK) interfaced with a photodiode array detector and under anaerobic conditions, following previously established protocols [ 68 , 69 ]. All samples were made anaerobic (in specially designed tonometers by successive evacuation and O 2 -free argon flushing) before introduction into the stopped-flow syringes. NQO1 ox (7.5 µ M) was mixed with NADH/D at concentrations ranging from 1:1 to 1:14 NQO1 ox :NADH/D ratios, while when using NADPH, the single 1:1 ratio was used. To evaluate re-oxidation, NQO1 hq (7.5 µ M) was mixed with stoichiometric amounts of DCPIP. Additionally, Dic was used as an inhibitor of both the reductive and oxidative half-reactions, adding it at 1:1 and 1:4 NQO1: Dic ratios. Reactions were studied in 20 mM HEPES-KOH, pH 7.4 with glucose/glucose oxidase (310 mM/10 units · mL −1 ), at 25 ◦ C and/or 6 ◦ C. Multiple wavelength absorption data in the flavin absorption region (400-900 nm) were collected and processed using the ProData-SX software (Applied Photophysics Ltd.). Time-dependent spectral deconvolution was performed by global analysis and numerical integration methods using Pro-Kineticist (Applied Photophysics Ltd.). Collected data were fitted to either singleor multi-step (A → B → n . . . . → Z) models allowing for estimation of the corresponding observed conversion rate constants (k obsA→B, k obsB→C , . . . ) at each NAD(P)H/D concentration, as well of the spectra of intermediate and final species [ 70 ]. A, B, n and Z are spectral species, reflecting a distribution of enzyme species at any time throughout the course of the enzyme: coenzyme interaction, including HT (or deuteride transfer, DT) or reorganization processes, and do not necessarily represent a single distinct enzyme intermediate. Since none of them represents individual species, their spectra cannot be included as fixed in the global fitting. The k obs values showing hyperbolic dependence profiles on the NAD(P)H/D concentration were fitted to a function (1) that describes binding at a single site followed by reorganization or HT/DT processes, allowing for the determination of the corresponding equilibrium constant (K d ), as well as the rate constant for the subsequent process (k) [69,70]: kobs =kA→B=kB→C=k·[NAD(P)H] [NAD(P)H]+Kd (1) Depending on the process, kmight account for the rate constant of the rate-limiting step for complex formation, k on , or for the HT/DT rate constant, k HT or k DT .K d might respectively account for the complex dissociation constant, KNADH/D d , or for a reorganization constant related to the transition between reaction intermediate species, Kreg d. Antioxidants 2020,9, 772 7 of 22 2.5. Kinetic Isotopic Effects (KIEs) For the estimation of primary kinetic isotopic effects in the HT process [ 68 ], HT or DT observed rate constants ( HT k obs or DT k obs ) from NADH/D to NQO1 ox were evaluated at different temperatures in the 5.3 − 20 ◦ C range, in samples containing equimolecular mixtures (7.5 µ M of each component) using NADH and [4R-2H]-NADD, unless otherwise indicated. Kinetic isotope effects (KIEs) on rate constants were calculated as follows: KIE =kHT kDT = HTkobs DTkobs (2) For each isotope, the fitting of the observed rates to the Arrhenius equation was calculated: k=A∗e−Ea RT (3) providing the values corresponding to Arrhenius pre-exponential (frequency) factors (A H and A D ) and activation energies (E aH and E aD ). The temperature dependence of the KIE was analyzed by combining Equations (2) and (3). Additionally, the activation enthalpies ( ∆ H ‡ ) and entropies ( ∆ S ‡ ) were calculated using the Eyring equation: ln kobs T!=ln kB h!+ (∆S‡/R)− ∆H‡ R·T!(4) where kBis the Boltzmann constant (1.3806·10−23 J·K−1) and his the Planck constant (6.626·10−34 J·s). 3. Results and Discussion 3.1. Human NQO1 Does Not Stabilize Intermediate Semiquinone States upon Photoreduction NQO1 ox exhibits the characteristic UV–visible spectra of flavoproteins, with maxima at 278, 375 and 449 nm, and shoulders at 422 and 475 nm (Figure 3). Upon photoreduction, the FAD cofactor exists in the hydroquinone state, NQO1 hq (i.e., a two-electron reduction), as denoted by the decrease in absorbance at 370 and 450 nm. Full reduction was achieved after 15 min of irradiation, with equivalent spectral features to those observed upon reduction with an excess of sodium dithionite (Figure 3). The photoreduction occurs without the appearance of any red-shifted absorbance band, indicative of the stabilization of the FAD blue-neutral semiquinone, but subtle changes in absorbance at 375, 400 and 480 nm might point to traces of the red-anionic semiquinone radical [ 71 ], as observed for other oxidases, including rat liver NQO1 [ 44 ]. Such a lack of semiquinone intermediates indicates that reduction of the semiquinone to the hydroquinone state is thermodynamically more favorable and kinetically faster than the reduction of the oxidized to the semiquinone species [ 72 ]. This observation agrees with the absence of detectable semiquinone paramagnetic signals when evaluating the redox cycle of the enzyme and with mammalian quinone oxidoreductases, which are a notable exception in that they only function by a compulsory two-electron transfer [ 73 ]. Such observations denote a less negative midpoint reduction potential of the FADH · /FADH 2 couple with respect to the FAD/FADH · one. Finally, upon mixing the photoreduced protein with atmospheric air, the initial absorbance spectrum of the oxidized protein was restored (not shown). Antioxidants 2020,9, 772 8 of 22 Antioxidants 2020, 9, x FOR PEER REVIEW 8 of 23 3. Results and Discussion 3.1. Human NQO1 Does Not Stabilize Intermediate Semiquinone States upon Photoreduction. NQO1ox exhibits the characteristic UV–visible spectra of flavoproteins, with maxima at 278, 375 and 449 nm, and shoulders at 422 and 475 nm (Figure 3). Upon photoreduction, the FAD cofactor exists in the hydroquinone state, NQO1hq (i.e., a two-electron reduction), as denoted by the decrease in absorbance at 370 and 450 nm. Full reduction was achieved after 15 min of irradiation, with equivalent spectral features to those observed upon reduction with an excess of sodium dithionite (Figure 3). The photoreduction occurs without the appearance of any red-shifted absorbance band, indicative of the stabilization of the FAD blue-neutral semiquinone, but subtle changes in absorbance at 375, 400 and 480 nm might point to traces of the red-anionic semiquinone radical [71], as observed for other oxidases, including rat liver NQO1 [44]. Such a lack of semiquinone intermediates indicates that reduction of the semiquinone to the hydroquinone state is thermodynamically more favorable and kinetically faster than the reduction of the oxidized to the semiquinone species [72]. This observation agrees with the absence of detectable semiquinone paramagnetic signals when evaluating the redox cycle of the enzyme and with mammalian quinone oxidoreductases, which are a notable exception in that they only function by a compulsory two-electron transfer [73]. Such observations denote a less negative midpoint reduction potential of the FADH·/FADH2 couple with respect to the FAD/FADH· one. Finally, upon mixing the photoreduced protein with atmospheric air, the initial absorbance spectrum of the oxidized protein was restored (not shown). Figure 3. Photoreduction of NQO1. Colored spectra correspond to different illumination time points throughout the photoreduction process. The dashed black line indicates the spectra corresponding to a sample chemically reduced by dithionite (S2O42−). 3.2. The Catalytic Cycle of NQO1 To evaluate the kinetics of the reductive and oxidative half-reactions in the catalytic cycle of NQO1 (Figure 1A), we studied the spectral changes occurring in the band-I (400–500 nm range) of the flavin by using fast kinetics stopped-flow spectrophotometry. 3.2.1. Non-Equivalent Active Sites in the NQO1 Dimer throughout the Reductive Half-Reaction We first evaluated the kinetics of the reductive half-reaction by mixing NQO1ox with NADH at 6 °C under strict anaerobic conditions. A decrease in the absorption at the FAD band-I was quickly Figure 3. Photoreduction of NQO1. Colored spectra correspond to different illumination time points throughout the photoreduction process. The dashed black line indicates the spectra corresponding to a sample chemically reduced by dithionite (S2O42−). 3.2. The Catalytic Cycle of NQO1 To evaluate the kinetics of the reductive and oxidative half-reactions in the catalytic cycle of NQO1 (Figure 1A), we studied the spectral changes occurring in the band-I (400–500 nm range) of the flavin by using fast kinetics stopped-flow spectrophotometry. 3.2.1. Non-Equivalent Active Sites in the NQO1 Dimer throughout the Reductive Half-Reaction We first evaluated the kinetics of the reductive half-reaction by mixing NQO1 ox with NADH at 6 ◦ C under strict anaerobic conditions. A decrease in the absorption at the FAD band-I was quickly observed without detection of semiquinone traces (Figure 4A), in agreement with FAD reduction to the hydroquinone form. The final spectrum after the overall HT compares well with the fully photoreduced NQO1 and dithionite reduced spectra, suggesting full reduction of the cofactor was achieved (Figure 3vs. Figure 4A). This observation envisages a less negative reduction potential of the NQO1 FAD/FADH 2 pair regarding the NAD + /NADH redox pair (E’ 0 = − 320 mV) [ 74 ] under our experimental conditions. This agrees with the value of − 159 mV previously reported for the rat enzyme [44]. To provide further insight into the mechanism of the reductive half-reaction, we submitted these kinetic data to global spectral deconvolution (Figure 4B). Several mechanisms were evaluated, but only a minimal three-step non-reversible mechanism (A → B → C → D) (Figure 4B–D) reproduced the experimental data well. The lack of major spectral changes in the instrumental dead time suggested that species A corresponds with the initial mixing of NQO1 ox and NADH. The conversion of species A into B was very fast (initial ~10–40 ms of reaction) and contributed to nearly 75% of the decay of the band-I absorption (Figure 4B). Therefore, this step comprises the HT process from NADH to FAD and its reduction. The observed rate constants for this process, k obsA→B , showed hyperbolic dependence on NADH concentration (Figure 4E), allowing for the determination of a NADH dissociation constant (K dNADH ), as well as of a limiting HT rate constant (k HT1 ), without considering the occurrence of any reverse HT reaction. The limiting value for k HT1 of 284 ± 17 s −1 (at 6 ◦ C) indicates a very fast HT from NADH to NQO1 ox , which compares well with the steady-state catalytic constant (k cat ) of 180–200 s −1 (at 30–37 ◦ C) [ 31 , 33 , 48 ]. This agreement supports that the HT process is the rate-limiting step within the reductive half-reaction. Alternative scenarios, in which substrate binding and/or product release could be rate-limiting steps are discussed in Appendix A. Kinetic analysis on the A → B process also Antioxidants 2020,9, 772 9 of 22 allows us to determine a K dNADH of 16 ± 3 µ M, a value 10-times lower than the reported K MNADH (160–240 µ M, using DCPIP as substrate) [ 31 , 33 , 48 ]. This observation indicates a tighter interaction in the reactive NQO1 ox :NADH complex than in the subsequent complexes formed throughout the reaction, and justifies the lack of spectroscopic detection of charge–transfer complexes (CTCs). Antioxidants 2020, 9, x FOR PEER REVIEW 9 of 23 observed without detection of semiquinone traces (Figure 4A), in agreement with FAD reduction to the hydroquinone form. The final spectrum after the overall HT compares well with the fully photoreduced NQO1 and dithionite reduced spectra, suggesting full reduction of the cofactor was achieved (Figure 3 vs. Figure 4A). This observation envisages a less negative reduction potential of the NQO1 FAD/FADH 2 pair regarding the NAD + /NADH redox pair (E’ 0 = −320 mV) [74] under our experimental conditions. This agrees with the value of −159 mV previously reported for the rat enzyme [44]. Figure 4. Kinetics of the NQO1 flavin reduction by NADH/D. (A) Spectral evolution on a 0−60 s timescale after mixing NQO1 ox (7.5 μM) with NADH (7.5 μM) in 20 mM HEPES-KOH, pH 7.4, at 6 °C. Different colored lines correspond to the spectra at different reaction times. (B) Spectral deconvolution of intermediate species observed during the reaction when fitting to a four-state model and the corresponding calculated observed rate constants. In panels A and B, the dashed line represents the spectrum of NQO1 ox before mixing. (C and D) Decay of kinetic traces at 450 nm and 475 nm and fittings to the model. (E) Dependence of k obs values on the NADH/D concentration. The trace for the fitting to Equation (1) for k obsA → B values is shown as a black line. Error bars correspond to the SD for at least three different replicates. Spectral evolution (A) and deconvolution (B) are from a single measurement and representative from n > 3. To provide further insight into the mechanism of the reductive half-reaction, we submitted these kinetic data to global spectral deconvolution (Figure 4B). Several mechanisms were evaluated, but only a minimal three-step non-reversible mechanism (A→B→C→D) (Figure 4B-D) reproduced the experimental data well. The lack of major spectral changes in the instrumental dead time suggested that species A corresponds with the initial mixing of NQO1 ox and NADH. The conversion of species A into B was very fast (initial ~10–40 ms of reaction) and contributed to nearly 75% of the decay of the band-I absorption (Figure 4B). Therefore, this step comprises the HT process from NADH to FAD and its reduction. The observed rate constants for this process, k obsA → B , showed Figure 4. Kinetics of the NQO1 flavin reduction by NADH/D. ( A ) Spectral evolution on a 0 − 60 s timescale after mixing NQO1 ox (7.5 µ M) with NADH (7.5 µ M) in 20 mM HEPES-KOH, pH 7.4, at 6 ◦ C. Different colored lines correspond to the spectra at different reaction times. ( B ) Spectral deconvolution of intermediate species observed during the reaction when fitting to a four-state model and the corresponding calculated observed rate constants. In panels A and B, the dashed line represents the spectrum of NQO1 ox before mixing. ( C , D ) Decay of kinetic traces at 450 nm and 475 nm and fittings to the model. ( E ) Dependence of k obs values on the NADH/D concentration. The trace for the fitting to Equation (1) for k obsA→B values is shown as a black line. Error bars correspond to the SD for at least three different replicates. Spectral evolution ( A ) and deconvolution ( B ) are from a single measurement and representative from n >3. This initially observed HT step (A to B) was followed by another process (B to C) that led to the nearly full reduction of the flavin bound to NQO1. In fact, this process essentially accounted for the remaining 25% of the total absorption decrease at the flavin band-I, and, therefore, must also be related to an HT process. The k obsB→C values were considerably slower than those for k obsA→B , and roughly dependent on the coenzyme concentration (Figure 4E), providing a limiting rate for this second HT event, k HT2 , in the 10–15 s −1 range. Our kinetic analysis showed a final step to achieve full FAD reduction (C to D), but this process accounted for a very small and slow spectroscopic change (with rate constants 5000 times slower than for the A → B process), suggesting that it might not be of catalytic relevance. Antioxidants 2020,9, 772 10 of 22 NQO1 can also use NADPH as electron donor, which is a better reductant than NADH (with six-fold higher catalytic efficiency and a four-fold enhancement in k cat , [ 31 , 34 , 75 ]). Consequently, NQO1 ox mixed with NADPH at a 1:1 ratio shows a faster FAD reduction than when using NADH, part of which occurs in the instrumental dead time (Figure 5A). The process was best described as a two-step process (A → B → C) (Figure 5B), with at least 80% of flavin reduction occurring in the first step. The k obsA→B and k obsB→C values at stoichiometric concentrations were 261 ± 13 and 7.8 ± 0.3 s −1 , respectively, with k obsA→B being 3.7 times faster than when using the same NADH ratio, while k obsB→C values resulted in the same range. Hence, the overall HT process is faster when NADPH is the hydride donor, consequently preventing studies using higher coenzyme concentrations. Antioxidants 2020, 9, x FOR PEER REVIEW 10 of 23 hyperbolic dependence on NADH concentration (Figure 4E), allowing for the determination of a NADH dissociation constant (KdNADH), as well as of a limiting HT rate constant (kHT1), without considering the occurrence of any reverse HT reaction. The limiting value for kHT1 of 284 ± 17 s−1 (at 6 °C) indicates a very fast HT from NADH to NQO1ox, which compares well with the steady-state catalytic constant (kcat) of 180–200 s−1 (at 30–37 °C) [31,33,48]. This agreement supports that the HT process is the rate-limiting step within the reductive half-reaction. Alternative scenarios, in which substrate binding and/or product release could be rate-limiting steps are discussed in Appendix A. Kinetic analysis on the A→B process also allows us to determine a KdNADH of 16 ± 3 μM, a value 10-times lower than the reported KMNADH (160–240 μM, using DCPIP as substrate) [31,33,48]. This observation indicates a tighter interaction in the reactive NQO1ox:NADH complex than in the subsequent complexes formed throughout the reaction, and justifies the lack of spectroscopic detection of charge–transfer complexes (CTCs). This initially observed HT step (A to B) was followed by another process (B to C) that led to the nearly full reduction of the flavin bound to NQO1. In fact, this process essentially accounted for the remaining 25% of the total absorption decrease at the flavin band-I, and, therefore, must also be related to an HT process. The kobsB→C values were considerably slower than those for kobsA→B, and roughly dependent on the coenzyme concentration (Figure 4E), providing a limiting rate for this second HT event, kHT2, in the 10–15 s−1 range. Our kinetic analysis showed a final step to achieve full FAD reduction (C to D), but this process accounted for a very small and slow spectroscopic change (with rate constants 5000 times slower than for the A→B process), suggesting that it might not be of catalytic relevance. NQO1 can also use NADPH as electron donor, which is a better reductant than NADH (with six-fold higher catalytic efficiency and a four-fold enhancement in kcat, [31,34,75]). Consequently, NQO1ox mixed with NADPH at a 1:1 ratio shows a faster FAD reduction than when using NADH, part of which occurs in the instrumental dead time (Figure 5A). The process was best described as a two-step process (A→B→C) (Figure 5B), with at least 80% of flavin reduction occurring in the first step. The kobsA→B and kobsB→C values at stoichiometric concentrations were 261 ± 13 and 7.8 ± 0.3 s−1, respectively, with kobsA→B being 3.7 times faster than when using the same NADH ratio, while kobsB→C values resulted in the same range. Hence, the overall HT process is faster when NADPH is the hydride donor, consequently preventing studies using higher coenzyme concentrations. Figure 5. Kinetics of the NQO1 flavin reduction by NADPH. (A) Spectral evolution on a 0–0.5 s timescale after the mixing NQO1ox (7.5 μM) with NADH (7.5 μM) in 20 mM HEPES-KOH, pH 7.4, at 6 °C. Different colored lines correspond to the spectra at different reaction times. The inset shows the decay of kinetic traces at 450 nm and 475 nm, the fittings to the model, and the residuals at 450 nm to show the quality of the fitting. (B) Spectral deconvolution into the different species observed along the reaction from fittings to a two-step model and calculated rate constants. Spectral evolution (A) and deconvolution (B) are from a single measurement and representative from n > 3. Figure 5. Kinetics of the NQO1 flavin reduction by NADPH. ( A ) Spectral evolution on a 0–0.5 s timescale after the mixing NQO1 ox (7.5 µ M) with NADH (7.5 µ M) in 20 mM HEPES-KOH, pH 7.4, at 6 ◦ C. Different colored lines correspond to the spectra at different reaction times. The inset shows the decay of kinetic traces at 450 nm and 475 nm, the fittings to the model, and the residuals at 450 nm to show the quality of the fitting. ( B ) Spectral deconvolution into the different species observed along the reaction from fittings to a two-step model and calculated rate constants. Spectral evolution ( A ) and deconvolution (B) are from a single measurement and representative from n >3. Altogether, and considering that NQO1 ox is a dimer, these data allowed us to suggest that the reduction of each protomer within the dimer might occur at very different rates. We must note that, according to the change in the flavin band-I magnitudes for the A → B and B → C steps, a part of the reduction of the slower protomer might occur within the first A → B step. To test this hypothesis, we used Dic to slow down the reductive half-reaction of NQO1 by NADH [ 6 , 39 ]. Dic competes for the NAD(P)H binding site, blocking access to the nicotinamide part of the coenzyme and, as a consequence, preventing the HT from the nicotinamide to the flavin cofactor [ 46 ]. As shown in Figure 6, the presence of Dic causes a considerable slowdown of the overall HT processes by NADH. Moreover, the spectral evolution was best described by a two-step mechanism (A → B → C), with each of the two steps accounting for half of the spectral change corresponding to FAD reduction. The full observation of both individual processes was likely possible because of the considerable reduction in k obsA→B and k obsB→C caused by Dic, respectively 0.034 ± 0.003 and 0.0065 ± 0.0005 s −1 at stoichiometric concentrations, which implies a ~2000-fold decrease. Higher Dic ratios (1:1:4 protein/NADH/dicoumarol) produced further slowdowns of both steps (Table 1), while increasing NADH concentrations (50 µ M, 1:6.6:4 of protein/NADH/Dic) only slightly increased reduction rate constants, hardly preventing Dic inhibition. Such observations are easily explained by the higher affinity of three orders of magnitude of Dic vs. NADH (K dDic typically in the 1 − 20 nM range; [ 33 , 49 , 55 ] and Section 3.2.1). Therefore, these inhibition studies with Dic strongly supported that the reduction of the FAD cofactor at the two active sites of the NQO1 dimer occurs at different rates. Antioxidants 2020,9, 772 11 of 22 Antioxidants 2020, 9, x FOR PEER REVIEW 11 of 23 Altogether, and considering that NQO1ox is a dimer, these data allowed us to suggest that the reduction of each protomer within the dimer might occur at very different rates. We must note that, according to the change in the flavin band-I magnitudes for the A→B and B→C steps, a part of the reduction of the slower protomer might occur within the first A→B step. To test this hypothesis, we used Dic to slow down the reductive half-reaction of NQO1 by NADH [6,39]. Dic competes for the NAD(P)H binding site, blocking access to the nicotinamide part of the coenzyme and, as a consequence, preventing the HT from the nicotinamide to the flavin cofactor [46]. As shown in Figure 6, the presence of Dic causes a considerable slowdown of the overall HT processes by NADH. Moreover, the spectral evolution was best described by a two-step mechanism (A→B→C), with each of the two steps accounting for half of the spectral change corresponding to FAD reduction. The full observation of both individual processes was likely possible because of the considerable reduction in kobsA→B and kobsB→C caused by Dic, respectively 0.034 ± 0.003 and 0.0065 ± 0.0005 s−1 at stoichiometric concentrations, which implies a ~2000-fold decrease. Higher Dic ratios (1:1:4 protein/NADH/dicoumarol) produced further slowdowns of both steps (Table 1), while increasing NADH concentrations (50 μM, 1:6.6:4 of protein/NADH/Dic) only slightly increased reduction rate constants, hardly preventing Dic inhibition. Such observations are easily explained by the higher affinity of three orders of magnitude of Dic vs. NADH (KdDic typically in the 1−20 nM range; [33,49,55] and Section 3.2.1). Therefore, these inhibition studies with Dic strongly supported that the reduction of the FAD cofactor at the two active sites of the NQO1 dimer occurs at different rates. Figure 6. Kinetics of NQO1 flavin reduction by NADH in the presence of the Dic inhibitor. (A) Spectral evolution after the mixing in the stopped-flow equipment of NQO1ox (7.5 μM) with NADH (7.5 μM) in the presence of Dic (7.5 μM) in 20 mM HEPES-KOH, pH 7.4, at 6 °C on a 0–800 s timescale. Different colored lines correspond to the spectra at different reaction times. The inset shows the decay of kinetic traces at 450 nm and 475 nm, as well as the fitting to a three-state model. (B) Spectral deconvolution of intermediate species observed during the reaction upon fitting to a two-step model and calculated observed rate constants. In panels A and B, the dashed line represents the protein spectrum before mixing. Spectral evolution (A) and deconvolution (B) are from a single measurement and representative from n > 3. So far, our analyses for the reductive half-reaction of NQO1ox by NAD(P)H have pointed to the reverse HT reactions being practically negligible. To confirm this, we mixed photoreduced NQO1hq with stoichiometric concentrations of NAD+, in specially designed spectrophotometer cuvettes and under anaerobic conditions, and followed spectral changes over the time. Once the reaction components were mixed, the FAD cofactor became very slowly re-oxidized without the stabilization of any semiquinone or CTC intermediates (Figure 7). The overall protein re-oxidation resulted in a monophasic process with a rate constant of 0.0025 ± 0.0005 s−1. This parameter would relate to the apparent HT rate constant for the backward reaction (appkHT-1), being around 28,000 times slower than the corresponding forward process. Therefore, the equilibrium of the reductive half-reaction is fully Figure 6. Kinetics of NQO1 flavin reduction by NADH in the presence of the Dic inhibitor. ( A ) Spectral evolution after the mixing in the stopped-flow equipment of NQO1 ox (7.5 µ M) with NADH (7.5 µ M) in the presence of Dic (7.5 µ M) in 20 mM HEPES-KOH, pH 7.4, at 6 ◦ C on a 0–800 s timescale. Different colored lines correspond to the spectra at different reaction times. The inset shows the decay of kinetic traces at 450 nm and 475 nm, as well as the fitting to a three-state model. ( B ) Spectral deconvolution of intermediate species observed during the reaction upon fitting to a two-step model and calculated observed rate constants. In panels A and B, the dashed line represents the protein spectrum before mixing. Spectral evolution ( A ) and deconvolution ( B ) are from a single measurement and representative from n >3. So far, our analyses for the reductive half-reaction of NQO1 ox by NAD(P)H have pointed to the reverse HT reactions being practically negligible. To confirm this, we mixed photoreduced NQO1 hq with stoichiometric concentrations of NAD + , in specially designed spectrophotometer cuvettes and under anaerobic conditions, and followed spectral changes over the time. Once the reaction components were mixed, the FAD cofactor became very slowly re-oxidized without the stabilization of any semiquinone or CTC intermediates (Figure 7). The overall protein re-oxidation resulted in a monophasic process with a rate constant of 0.0025 ± 0.0005 s −1 . This parameter would relate to the apparent HT rate constant for the backward reaction ( app k HT-1 ), being around 28,000 times slower than the corresponding forward process. Therefore, the equilibrium of the reductive half-reaction is fully displaced towards the production of NQO1 hq , in agreement with the main physiological role of the enzyme in the detoxification of quinones by their reduction. Table 1. Summary of observed rate constants (k obs ) for the reductive and oxidative half-reactions involving NQO1. Measurements were carried out in 20 mM HEPES-KOH, pH 7.4 at 6 ◦ C and the ratios are indicated between brackets for each reactant. Evolution of the reaction was followed in the 400–1000 nm wavelength range using stopped-flow equipment with a photodiode array detector (n >3, mean ±SD). Sample in Tonometer 1 Sample in Tonometer 2 kobsA→B (s−1) kobsB→C (s−1) NQO1 (1) NADH (1) 78 ±1 8.9 ±0.9 NQO1 (1) 4R-NADD (1) 44 ±2 6.3 ±0.2 NQO1 (1) NADPH (1) 261 ±13 7.8 ±0.3 NQO1 (1) +Dic (1) NADH (1) 0.034 ±0.003 0.0065 ±0.0005 NQO1 (1) +Dic (1) NADH (6.6) 0.036 ±0.002 0.010 ±0.001 NQO1 (1) +Dic (4) NADH (1) 0.018 ±0.003 0.0015 ±0.0001 NQO1 (1) +Dic (1) NADPH (1) 0.036 ±0.006 0.0070 ±0.0008 NQO1 (1) +Dic (4) NADPH (1) 0.017 ±0.001 0.0020 ±0.0001 NQO1 (1) +NADH (1) DCPIP (1) >500 160 ±14 NQO1 (1) +NADH (1) DCPIP (1) +Dic (1) 38 ±3 6.3 ±1.2 NQO1 (1) +NADH (1) DCPIP (1) +Dic (4) 7.7 ±0.2 1.3 ±0.1 NQO1 (1) +NADH (1) Ferricyanide 219 ±12 29 ±4 Antioxidants 2020,9, 772 12 of 22 Antioxidants 2020, 9, x FOR PEER REVIEW 12 of 23 displaced towards the production of NQO1hq, in agreement with the main physiological role of the enzyme in the detoxification of quinones by their reduction. Table 1. Summary of observed rate constants (kobs) for the reductive and oxidative half-reactions involving NQO1. Measurements were carried out in 20 mM HEPES-KOH, pH 7.4 at 6 °C and the ratios are indicated between brackets for each reactant. Evolution of the reaction was followed in the 400–1000 nm wavelength range using stopped-flow equipment with a photodiode array detector (n > 3, mean ± SD). Sample in Tonometer 1 Sample in Tonometer 2 kobsA→B (s−1) kobsB→C (s−1) NQO1 (1) NADH (1) 78 ± 1 8.9 ± 0.9 NQO1 (1) 4R-NADD (1) 44 ± 2 6.3 ± 0.2 NQO1 (1) NADPH (1) 261 ± 13 7.8 ± 0.3 NQO1 (1) + Dic (1) NADH (1) 0.034 ± 0.003 0.0065 ± 0.0005 NQO1 (1) + Dic (1) NADH (6.6) 0.036 ± 0.002 0.010 ± 0.001 NQO1 (1) + Dic (4) NADH (1) 0.018 ± 0.003 0.0015 ± 0.0001 NQO1 (1) + Dic (1) NADPH (1) 0.036 ± 0.006 0.0070 ± 0.0008 NQO1 (1) + Dic (4) NADPH (1) 0.017 ± 0.001 0.0020 ± 0.0001 NQO1 (1) + NADH (1) DCPIP (1) >500 160 ± 14 NQO1 (1) + NADH (1) DCPIP (1) + Dic (1) 38 ± 3 6.3 ± 1.2 NQO1 (1) + NADH (1) DCPIP (1) + Dic (4) 7.7 ± 0.2 1.3 ± 0.1 NQO1 (1) + NADH (1) Ferricyanide 219 ± 12 29 ± 4 Figure 7. Kinetics of the reaction of NQO1hq with NAD+. (A) Spectral evolution after mixing (in an anaerobic cuvette) NQO1hq (7.5 μM)) with a 1:1 ratio of NAD+ in 20 mM HEPES-KOH, pH 7.0, at 6 °C on a 0–35 min timescale. Different colored lines correspond to the spectra at different reaction times. (B) Detail of kinetic traces at 450 nm and 475 nm and fitting to a single exponential function. Spectral evolution (A) is from a single measurement and representative from n > 3. 3.2.2. Non-Equivalent Active Sites in the NQO1 Dimer throughout the Oxidative Half-Reaction Although NQO1 can reduce a wide variety of substrates [4], most of them are not appropriate for enzymatic studies due to their low solubility in aqueous solutions, their characteristic spectral properties or, in some cases, their fast reduction that precludes pre-steady-state kinetic characterization using stopped-flow spectroscopy [34,44]. We have here used DCPIP, a suitable and artificial electron acceptor often used in activity measurements of human NQO1 [4], to study its oxidative half-reaction. Nonetheless, mixing NQO1hq with DCPIP at equimolecular concentrations also resulted in the extremely fast re-oxidation of the protein, even at low temperatures, and with nearly 50% of the spectral changes occurring in the instrumental dead time (Figure 8A–C). The Figure 7. Kinetics of the reaction of NQO1 hq with NAD + . ( A ) Spectral evolution after mixing (in an anaerobic cuvette) NQO1 hq (7.5 µ M)) with a 1:1 ratio of NAD + in 20 mM HEPES-KOH, pH 7.0, at 6 ◦ C on a 0–35 min timescale. Different colored lines correspond to the spectra at different reaction times. ( B ) Detail of kinetic traces at 450 nm and 475 nm and fitting to a single exponential function. Spectral evolution (A) is from a single measurement and representative from n >3. 3.2.2. Non-Equivalent Active Sites in the NQO1 Dimer throughout the Oxidative Half-Reaction Although NQO1 can reduce a wide variety of substrates [ 4 ], most of them are not appropriate for enzymatic studies due to their low solubility in aqueous solutions, their characteristic spectral properties or, in some cases, their fast reduction that precludes pre-steady-state kinetic characterization using stopped-flow spectroscopy [ 34 , 44 ]. We have here used DCPIP, a suitable and artificial electron acceptor often used in activity measurements of human NQO1 [ 4 ], to study its oxidative half-reaction. Nonetheless, mixing NQO1 hq with DCPIP at equimolecular concentrations also resulted in the extremely fast re-oxidation of the protein, even at low temperatures, and with nearly 50% of the spectral changes occurring in the instrumental dead time (Figure 8A–C). The observed overall process was best fitted to a two-step mechanism (A → B → C), with the initial step accounting for most of the spectroscopic changes and exhibiting observed rate constants, k obsA→B , above the instrumental measurement limit (>500 s −1 , at stoichiometric reactant ratios). Conversely, the second step (B → C) shows minor, and probably biased, contributions both in terms of k obsB→C (160 s −1 ) and amplitude. These values for the rate constants in the oxidative half-reaction further reinforce that the reductive half-reaction is rate-limiting in NQO1 catalysis, also preventing additional analyses at higher NQO1hq:DCPIP ratios. Although Dic is usually reported as an inhibitor of the reductive half-reaction, we also evaluated its effect in the oxidative half-reaction. The re-oxidation of NQO1 hq by DCPIP is also significantly slowed down in the presence of Dic (1:1 of protein/NADH mixed with 1:1 DCPIP/Dic, Figure 8D,E), which might not be surprising because Dic shares the binding site with both the electron donor and acceptor [ 46 ]. The presence of Dic has a considerable effect on both k obsA→B and k obsB→C , which decreased to 38 ± 3 and 6.3 ± 1.2 s −1 , respectively, at a 1:1 ratio and even more at higher Dic concentrations (Table 1), while the amplitudes of the changes became similar for both processes (Figure 8D–F). These data indicate that the two active sites of NQO1 are also non-equivalent regarding the kinetics of the oxidative half-reaction. Antioxidants 2020,9, 772 13 of 22 Antioxidants 2020, 9, x FOR PEER REVIEW 13 of 23 observed overall process was best fitted to a two-step mechanism (A→B→C), with the initial step accounting for most of the spectroscopic changes and exhibiting observed rate constants, kobsA→B, above the instrumental measurement limit (>500 s−1, at stoichiometric reactant ratios). Conversely, the second step (B→C) shows minor, and probably biased, contributions both in terms of kobsB→C (160 s−1) and amplitude. These values for the rate constants in the oxidative half-reaction further reinforce that the reductive half-reaction is rate-limiting in NQO1 catalysis, also preventing additional analyses at higher NQO1hq:DCPIP ratios. Figure 8. Kinetics of NQO1hq re-oxidation by DCPIP (2,6-dichlorophenol indophenol). (A) Spectral evolution after mixing NQO1hq (7.5 μM) with DCPIP (7.5 μM) in 20 mM HEPES-KOH, pH 7.4, at 6 °C on a 0–0.1 s timescale. (B) Spectral deconvolution of intermediate species observed during the reaction when using a three-state model. (C) Kinetic traces at 450 nm, 475 nm and 600 nm. Experimental data as well as the fitting to the three-state mechanism are shown. (D) Spectral evolution after mixing NQO1hq (7.5 μM) with DCPIP (7.5 μM) in the presence of Dic (7.5 μM) in 20 mM HEPES-KOH, pH 7.4, at 6 °C on a 0–1 s timescale. (E) Spectral deconvolution of intermediate species obtained from analysis using a three-state model. (F) Kinetic traces at 450 nm, 475 nm and 600 nm. Experimental data as well as the fitting to the three-state mechanisms are shown. Dashed lines (panels A–B and D–E) correspond to the initial spectra of NQO1hq and DCPIP (bold), the bold black line is the addition of these two spectra (species at t = 0) and the different colored lines correspond to Figure 8. Kinetics of NQO1 hq re-oxidation by DCPIP (2,6-dichlorophenol indophenol). ( A ) Spectral evolution after mixing NQO1 hq (7.5 µ M) with DCPIP (7.5 µ M) in 20 mM HEPES-KOH, pH 7.4, at 6 ◦ C on a 0–0.1 s timescale. ( B ) Spectral deconvolution of intermediate species observed during the reaction when using a three-state model. ( C ) Kinetic traces at 450 nm, 475 nm and 600 nm. Experimental data as well as the fitting to the three-state mechanism are shown. ( D ) Spectral evolution after mixing NQO1 hq (7.5 µ M) with DCPIP (7.5 µ M) in the presence of Dic (7.5 µ M) in 20 mM HEPES-KOH, pH 7.4, at 6 ◦ C on a 0–1 s timescale. ( E ) Spectral deconvolution of intermediate species obtained from analysis using a three-state model. ( F ) Kinetic traces at 450 nm, 475 nm and 600 nm. Experimental data as well as the fitting to the three-state mechanisms are shown. Dashed lines (panels A , B and D , E ) correspond to the initial spectra of NQO1 hq and DCPIP (bold), the bold black line is the addition of these two spectra (species at t =0) and the different colored lines correspond to the spectra at different reaction times. Spectral evolution ( A , D ) and deconvolution ( B , E ) are from a single measurement and representative from n >3. Although the main enzymatic role of NQO1 is associated with the mandatory two-electron reduction of substrates, reactivity with artificial one-electron oxidants, such as ferricyanide, has been widely reported for the characterization of Saccharomyces cerevisiae Lot6p and rat and human NQO1 [76,77] . The rapid mixing of reduced NQO1 hq with ferricyanide (1:1 reduction equivalent ratio) coursed with the re-oxidation of the FAD cofactor (in ~ 0.4 s) without any trace of semiquinone Antioxidants 2020,9, 772 14 of 22 stabilization and with k obsA→B of 219 ± 12 s −1 and k obsB→C of 29 ± 4 s −1 , values considerably lower when compared to two-electron substrates. 3.3. Dynamics at the NQO1 Active Sites Differentially Contribute to the Two HT Events Representing the Reductive Half-Reaction We then used fast kinetics measurements to investigate primary KIEs (using [4R2 H]-NADD), as well as the temperature dependence of the rate constants and KIEs, in the context of the Arrhenius equation. The resulting parameters may provide information on the structural organization and dynamics at the active site of enzymes during HT catalysis [ 68 ]. Due to HT k obsA→B values being close to the instrumental limit upon increasing the coenzyme concentration (see black closed circles in Figure 4E), equimolecular concentrations of enzyme and coenzyme were the most suitable choice to overcome the technical limitations for the temperature-dependent studies. Nonetheless, since K dNADD may differ from K dNADH , the KIEs obtained in this way might be apparent. However, this does not seem to be the case, since the K dNADD at 6 ◦ C is quite similar to that of K dNADD (12 ± 2 µ M vs. 16 ± 3 µ M, Figure 4E), indicating that, at this temperature, the deuterated substrate hardly influences this parameter. Moreover, the comparison of the limiting k DT1 of 122 ± 5 s −1 with k HT1 (284 ± 17 s −1 ) resulted in a moderate value of 2.3 ± 0.5 for the KIE A→B at 6 ◦ C. On their side, DT k obsB→C values were considerably lower (Figure 4E, red open circles) with a limiting k DT2 ~ 7 s −1 and therefore a KIE B→C may also be close to a value of 2. KIEs usually exhibit maximal values at lower temperatures, suggesting this might be a nearly limiting value. The magnitude of primary H/D KIEs theoretically can reach a maximum value of 8, although there are considerably larger values reported for some enzymes [ 78 ]. Nonetheless, primary KIEs can also decrease towards 1 when the C-H bond is either broken less (i.e., earlier transition state) or more (i.e., later transition state) in the transition state structure (asymmetrical), or if the transition state is nonlinear [ 79 , 80 ]. This suggests that, in the NQO1 reductive half-reaction, the transition state, at least for the first HT, is either moving away from symmetrical or it is non-linear. The magnitude of the KIE is itself informative, but the size of its temperature dependence also serves as a key descriptor of the reaction coordinate [ 64 ]. In particular, we applied the environmentally coupled tunneling model by determining HT and DT observed rate constants, namely HT k obsA→B , HT k obsB→C , DT k obsA→B and DT k obsB→C , at different temperatures (Figure 9A). As indicated above, we used equimolecular concentrations of the enzyme and the coenzyme substrate (instead of saturating) to avoid entering the detection limit of the instrument upon increasing the temperature. The determined k obs values provided apparent KIE A→B and KIE B→C values of ~ 1.8 (Table 2), which were temperature independent (Figure 9B). Despite these values being apparent, they match with the limiting one obtained under saturating conditions and low temperature. HT k obsA→B and DT k obsA→B showed a weaker temperature dependence than HT k obsB→C and DT k obsB→C , and in both cases the Arrhenius plots for HT and DT hardly deviated from parallel lines, indicating very similar E a values (Figure 9A). E a values are moderate for the fast HT/DT event and 1.5 times larger for the second process, while ∆ E a (E aDT − E aHT ) is nearly 0 for both processes. In addition, the calculated isotope effect on the Arrhenius frequency factor (A H /A D ) strongly differed between both HT/DT events, with a value close to the unity for the faster one and about 10 for the slower one. As a consequence of the marginal differences between E aDT and E aHT and the small values for the apparent KIEs, the activation enthalpies and entropies for HT and DT are also small when these results are analyzed in the context of the Eyring equation (Equation (4), see Figure 9C and Table 3). Antioxidants 2020,9, 772 15 of 22 Antioxidants 2020, 9, x FOR PEER REVIEW 15 of 23 and DTkobsA→B showed a weaker temperature dependence than HTkobsB→C and DTkobsB→C, and in both cases the Arrhenius plots for HT and DT hardly deviated from parallel lines, indicating very similar Ea values (Figure 9A). Ea values are moderate for the fast HT/DT event and 1.5 times larger for the second process, while ΔEa (EaDT − EaHT) is nearly 0 for both processes. In addition, the calculated isotope effect on the Arrhenius frequency factor (AH/AD) strongly differed between both HT/DT events, with a value close to the unity for the faster one and about 10 for the slower one. As a consequence of the marginal differences between EaDT and EaHT and the small values for the apparent KIEs, the activation enthalpies and entropies for HT and DT are also small when these results are analyzed in the context of the Eyring equation (Equation (4), see Figure 9C and Table 3). Figure 9. Temperature dependence of kinetic parameters for the two hydride/deuteride transfer (HT/DT) processes from NADH to NQO1. (A) Arrhenius plots of kinetic constants. (B) Temperature dependence of the kinetic isotope effects (KIEs). (C) Eyring plots of kinetic constants. Table 2. KIEs for the HT in the reduction of NQO1 by NADH. All values correspond to data obtained with equimolecular concentrations of the reactants in the stopped-flow equipment. (n > 3, mean ± SD). Analysis was performed using Equations (2) and (3) (see Figure 9A,B). HT DT KIE ΔEa EaDT - EaHT (kcal·mol−1) AH/AD HTkobs a (s−1) EaHT (kcal·mol−1) AH (s−1) DTkobs a (s−1) EaDT (kcal·mol−1) AD (s−1) A→B 78 ± 1 6.1 ± 0.2 (5.3 ± 1.2)·106 44 ± 2 6.3 ± 0.4 (4.1 ± 1.1)·106 1.8 ± 0.1 0.2 ± 0.4 1.3 ± 0.6 B→C 8.9 ± 0.9 10.9 ± 0.5 (3.4 ± 0.9)·109 5.3 ± 0.2 9.8 ± 0.5 (2.6 ± 0.6)·108 1.8 ± 0.3 -1.1 ± 0.7 13 ± 6 a Values at 6 °C. The temperature independence of the KIE is generally interpreted in the context of full tunneling models, where the reaction barrier is attributed to the heavy atom motions that affect the probability of wave function overlap and little, or no, sampling of the distance of the reacting atoms [81]. Thus, our data show that both HT events in NQO1 are consistent with transitions under the barrier (i.e., quantum tunneling) and with asymmetrical or non-linear transition states. Moreover, the lower Ea and close to zero ΔEa values and, particularly, the close to unity AH/AD ratio for the fast HT process support for this event some contribution of dynamics, and/or donor–acceptor distance (DAD) fluctuations of the active site heavy atoms, to the tunneling. Thus, the fast HT process resembles the behavior most commonly found in native enzyme-mediated HT processes, in which Figure 9. Temperature dependence of kinetic parameters for the two hydride/deuteride transfer (HT/DT) processes from NADH to NQO1. ( A ) Arrhenius plots of kinetic constants. ( B ) Temperature dependence of the kinetic isotope effects (KIEs). (C) Eyring plots of kinetic constants. Table 2. KIEs for the HT in the reduction of NQO1 by NADH. All values correspond to data obtained with equimolecular concentrations of the reactants in the stopped-flow equipment. (n >3, mean ± SD). Analysis was performed using Equations (2) and (3) (see Figure 9A,B). HT DT KIE ∆Ea EaDT −EaHT (kcal·mol−1) AH/AD HTkobs a (s−1) EaHT (kcal·mol−1) AH (s−1) DTkobs a (s−1) EaDT (kcal·mol−1) AD (s−1) A→B 78 ±1 6.1 ±0.2 (5.3 ± 1.2) · 10 644 ±2 6.3 ±0.4 (4.1 ± 1.1) · 10 6 1.8 ± 0.1 0.2 ±0.4 1.3 ± 0.6 B→C 8.9 ±0.9 10.9 ±0.5 (3.4 ± 0.9) · 10 95.3 ±0.2 9.8 ±0.5 (2.6 ± 0.6) · 10 8 1.8 ± 0.3 −1.1 ±0.7 13 ±6 aValues at 6 ◦C. The temperature independence of the KIE is generally interpreted in the context of full tunneling models, where the reaction barrier is attributed to the heavy atom motions that affect the probability of wave function overlap and little, or no, sampling of the distance of the reacting atoms [ 81 ]. Thus, our data show that both HT events in NQO1 are consistent with transitions under the barrier (i.e., quantum tunneling) and with asymmetrical or non-linear transition states. Moreover, the lower E a and close to zero ∆ E a values and, particularly, the close to unity A H /A D ratio for the fast HT process support for this event some contribution of dynamics, and/or donor–acceptor distance (DAD) fluctuations of the active site heavy atoms, to the tunneling. Thus, the fast HT process resembles the behavior most commonly found in native enzyme-mediated HT processes, in which catalytic enhancement is achieved by promoting and optimizing vibrations in active sites that minimize DAD fluctuations [ 81 , 82 ]. Conversely, the second HT event also shows almost temperature-independent KIEs and still similar E aD and E aH values, but higher E a and A H /A D considerably greater than the unity (Figure 9, Table 2). The higher E aH values suggest larger reorganization energies as the main source of the E a , whereas the rest of the parameters, particularly the A H /A D ratio, support a larger contribution to the active site environment to promote a close approach between the hydride donor and the acceptor atom with little DAD sampling. Altogether, these data indicate that for the slower HT process the initial pre-organization complex situates the reacting atoms, N5 of FAD and the C4-H of the nicotinamide of NADH, at optimal tunneling distance, creating a stiffer active site for the competent HT when compared to the active site of the fast HT process. Antioxidants 2020,9, 772 22 of 22 67. Frago, S.; Goni, G.; Herguedas, B.; Peregrina, J.R.; Serrano, A.; Perez-Dorado, I.; Molina, R.; Gomez-Moreno, C.; Hermoso, J.A.; Martinez-Julvez, M.; et al. Tuning of the FMN binding and oxido-reduction properties by neighboring side chains in Anabaena flavodoxin. Arch. Biochem. Biophys. 2007,467, 206–217. [CrossRef] 68. Sanchez-Azqueta, A.; Catalano-Dupuy, D.L.; Lopez-Rivero, A.; Tondo, M.L.; Orellano, E.G.; Ceccarelli, E.A.; Medina, M. Dynamics of the active site architecture in plant-type ferredoxin-NADP(+) reductases catalytic complexes. Biochim. Biophys. Acta 2014,1837, 1730–1738. [CrossRef] 69. Perez-Amigot, D.; Taleb, V.; Boneta, S.; Anoz-Carbonell, E.; Sebastian, M.; Velazquez-Campoy, A.; Polo, V.; Martinez-Julvez, M.; Medina, M. Towards the competent conformation for catalysis in the ferredoxin-NADP(+) reductase from the Brucella ovis pathogen. Biochim. Biophys. Acta Bioenerg. 2019,1860, 148058. [CrossRef] 70. Tejero, J.; Peregrina, J.R.; Martinez-Julvez, M.; Gutierrez, A.; Gomez-Moreno, C.; Scrutton, N.S.; Medina, M. Catalytic mechanism of hydride transfer between NADP+/H and ferredoxin-NADP+reductase from Anabaena PCC 7119. Arch. Biochem. Biophys. 2007,459, 79–90. [CrossRef] 71. Massey, V.; Muller, F.; Feldberg, R.; Schuman, M.; Sullivan, P.A.; Howell, L.G.; Mayhew, S.G.; Matthews, R.G.; Foust, G.P. The reactivity of flavoproteins with sulfite. Possible relevance to the problem of oxygen reactivity. J. Biol. Chem. 1969,244, 3999–4006. 72. Clark, W.M. Oxidation-Reduction Potentials of Organic Systems; The Williams & Wilkins Company: Baltimore, MD, USA, 1960. 73. Dinkova-Kostova, A.T.; Talalay, P. NAD(P)H: Quinone acceptor oxidoreductase 1 (NQO1), a multifunctional antioxidant enzyme and exceptionally versatile cytoprotector. Arch. Biochem. Biophys. 2010 ,501, 116–123. [CrossRef] [PubMed] 74. Burton, K. The enthalpy change for the reduction of nicotinamide—Adenine dinucleotide. Biochem. J. 1974 , 143, 365–368. [CrossRef] [PubMed] 75. Lind, C.; Cadenas, E.; Hochstein, P.; Ernster, L. DT-diaphorase: Purification, properties, and function. Methods Enzymol. 1990,186, 287–301. [PubMed] 76. Merker, M.P.; Audi, S.H.; Bongard, R.D.; Lindemer, B.J.; Krenz, G.S. Influence of pulmonary arterial endothelial cells on quinone redox status: Effect of hyperoxia-induced NAD(P)H: Quinone oxidoreductase 1. Am. J. Physiol. Lung Cell. Mol. Physiol. 2006,290, L607–L619. [CrossRef] [PubMed] 77. Bongard, R.D.; Krenz, G.S.; Gastonguay, A.J.; Williams, C.L.; Lindemer, B.J.; Merker, M.P. Characterization of the threshold for NAD(P)H: Quinone oxidoreductase activity in intact sulforaphane-treated pulmonary arterial endothelial cells. Free Radic. Biol. Med. 2011,50, 953–962. [CrossRef] 78. Klinman, J.P. Moving Through Barriers in Science and Life. Annu. Rev. Biochem. 2019,88, 1–24. [CrossRef] 79. Nelson, S.D.; Trager, W.F. The use of deuterium isotope effects to probe the active site properties, mechanism of cytochrome P450-catalyzed reactions, and mechanisms of metabolically dependent toxicity. Drug Metab. Dispos. 2003,31, 1481–1498. [CrossRef] 80. Yoshimoto, F.K.; Zhou, Y.; Peng, H.M.; Stidd, D.; Yoshimoto, J.A.; Sharma, K.K.; Matthew, S.; Auchus, R.J. Minor activities and transition state properties of the human steroid hydroxylases cytochromes P450c17 and P450c21, from reactions observed with deuterium-labeled substrates. Biochemistry 2012 ,51, 7064–7077. [CrossRef] 81. Nagel, Z.D.; Klinman, J.P. Update 1 of: Tunneling and dynamics in enzymatic hydride transfer. Chem. Rev. 2010,110, PR41–PR67. [CrossRef] 82. Nagel, Z.D.; Klinman, J.P. A 21st century revisionist’s view at a turning point in enzymology. Nat. Chem. Biol. 2009,5, 543–550. [CrossRef] 83. Mesa-Torres, N.; Betancor-Fernández, I.; Oppici, E.; Cellini, B.; Salido, E.; Pey, A.L. Evolutionary Divergent Suppressor Mutations in Conformational Diseases. Genes 2018,9, 352. [CrossRef] [PubMed] 84. Wyman, J.; Gill, S.J. Binding and Linkage. Functional Chemistry of Biological Macromolecules; University Science Books: Mill Valley, CA, USA, 1990. 85. Gadda, G.; Sobrado, P. Kinetic Solvent Viscosity Effects as Probes for Studying the Mechanisms of Enzyme Action. Biochemistry 2018,57, 3445–3453. [PubMed] © 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/). Publication IV Discovery of antimicrobial compounds targeting bacterial type FAD synthases Ernesto Anoz Carbonell – Doctoral Thesis 134 Ernesto Anoz Carbonell contributions: Experimental work: protein purification, steady-state enzymatic activity measurements and antimicrobial activity assays. Data analysis. Manuscript writing, review and editing RESEARCH PAPER Discovery of antimicrobial compounds targeting bacterial type FAD synthetases Mar ıa Sebasti an a,b , Ernesto Anoz-Carbonell a,b,c , Bego~ na Gracia c,d , Pilar Cossio e,f , Jos e Antonio A ınsa b,c,d , Isa ıas Lans f and Milagros Medina a,b a Departamento de Bioqu  ımica y Biolog  ıa Molecular y Celular, Facultad de Ciencias, Universidad de Zaragoza, Zaragoza, Spain; b Institute of Biocomputation and Physics of Complex Systems (BIFI-IQFR and CBsC-CSIC Joint Units), Universidad de Zaragoza, Zaragoza, Spain; c Grupo de Gen etica de Micobacterias, Departamento de Microbiolog ıa, Medicina Preventiva y Salud P ublica. Facultad de Medicina, Universidad de Zaragoza, Zaragoza, Spain; d CIBER Enfermedades Respiratorias (CIBERES), Instituto de Salud Carlos III, Madrid, Spain; e Department of Theoretical Biophysics, Max Planck Institute of Biophysics, Frankfurt, Germany; f Biophysics of Tropical Diseases, Max Planck Tandem Group, University of Antioquia, Medell  ın, Colombia ABSTRACT The increase of bacterial strains resistant to most of the available antibiotics shows a need to explore novel antibacterial targets to discover antimicrobial drugs. Bifunctional bacterial FAD synthetases (FADSs) synthesise the flavin mononucleotide (FMN) and flavin adenine dinucleotide (FAD). These cofactors act in vital processes as part of flavoproteins, making FADS an essential enzyme. Bacterial FADSs are potential antibacterial targets because of differences to mammalian enzymes, particularly at the FAD producing site. We have optimised an activity-based high throughput screening assay targeting Corynebacterium ammoniagenes FADS (CaFADS) that identifies inhibitors of its different activities. We selected the three best high-performing inhibitors of the FMN:adenylyltransferase activity (FMNAT) and studied their inhibition mechanisms and binding properties. The specificity of the CaFADS hits was evaluated by studying also their effect on the Streptococcus pneumoniae FADS activities, envisaging differences that can be used to discover species-specific antibacterial drugs. The antimicrobial effect of these compounds was also evaluated on C. ammoniagenes,S. pneumoniae, and Mycobacterium tuberculosis cultures, finding hits with favourable antimicrobial properties. ARTICLE HISTORY Received 17 October 2017 Revised 28 November 2017 Accepted 28 November 2017 KEYWORDS Bacterial FAD Synthetase; high-throughput screening; Streptococcus pneumoniae; drug discovery Introduction An important innovation gap in the discovering of antibiotics has occurred during the last two decades 1 , with only five new classes available and 51 new antimicrobials in clinical development 2–4 .In addition, the selection of multi-drug resistant microorganisms 5 encourages to search for new antimicrobial drugs capable of inhibiting novel protein targets, such as those controlling the biosynthesis of essential biomolecules. Flavin mononucleotide (FMN) and flavin adenine dinucleotide (FAD) are the cofactors of flavoproteins. All living organisms contain a great number of such proteins and many of them are involved in essential functions 6–8 , including protein folding 9 , electron transport in the respiratory and photosynthetic chains 10 ,b-oxidation of fatty acids 11 , nucleotide synthesis or signal transduction 12 , among others. Lack, or low levels, of FMN and FAD lead to the accumulation of apoflavoproteins, unable to perform the flavin-dependent functions, resulting in the concomitant death of the cell or the organism 13,14 . Prokaryotic bifunctional FAD synthetases (FADS) synthesise both FMN and FAD, being therefore potential new antimicrobial targets 15 . Such hypothesis is sustained by several facts; (i) halting the production of FMN and FAD prevents, from the very beginning, all pathways that involve flavoproteins and flavoenzymes, (ii) in most bacteria the only pathway for FMN and FAD biosynthesis occurs with bifunctional FADS 13,14 , (iii) prokaryotic FADSs differ structurally and biochemically from the mammalian proteins that transform FMN into FAD 16–19 , so drugs that target these proteins are likely to be selective for bacteria and (iv) the availability of structures of several bacterial FADSs facilitates the design of both inhibitory drugs and activity assays 20–22 . CONTACT Milagros Medina [email protected] Departamento de Bioqu  ımica y Biolog ıa Molecular y Celular, Facultad de Ciencias, Universidad de Zaragoza, Pedro Cerbuna, 12. 50009 Zaragoza, Spain Supplemental data for this article can be accessed here. ß2017 The Author(s). Published by Informa UK Limited, trading as Taylor & Francis Group. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. JOURNAL OF ENZYME INHIBITION AND MEDICINAL CHEMISTRY 2018, VOL. 33, NO. 1, 241–254 https://doi.org/10.1080/14756366.2017.1411910 Bacterial FADSs have both ATP:riboflavin kinase (RFK, EC 2.7.1.26) and ATP:FMN:adenylyltransferase (FMNAT, EC 2.7.7.2) activities, being the latter reversible (FAD pyrophosphorylase) in some species 17,23 . FADSs synthesise FMN and FAD from riboflavin (RF, vitamin B2) through two sequential reactions: RF is first phosphorylated to FMN by the RFK activity, and then the FMNAT activity transfers an adenylyl group from ATP to FMN producing FAD. These catalytic activities are performed by two almost independent modules (Supplementary Figure SD1). The C-terminus module produces FMN from RF (named RFK module), while the N-terminal module transforms FMN into FAD (FMNAT module). The RFK module shows sequence and structural homology with the monofunctional eukaryotic RFKs, while the FMNAT module does not present neither sequence nor structural similarity with the proteins that synthesise FAD in mammals 16,21,24,25 . Because the enzymes leading to FAD production in prokaryotes and eukaryotes use different chemistry, and belong to different structural families, potential inhibitors that specifically target the FMNAT module of bacterial FADSs are an interesting option for the novel drug development 15 . In this work, we have used as a model the FADS from the nonpathogenic organism Corynebacterium ammoniagenes (CaFADS), which is the best known model to characterise members of the prokaryotic FADSs family 17,24,26–30 , in an activity-based highthroughput screening (HTS) assay to find potential inhibitors. The HTS hits were assayed to determine their specificity and potency for the RFK and FMNAT activities. We also studied the kinetic inhibition mechanism of the three most potent and selective inhibitors of the FMNAT activity (FMNAT hits), as well as their binding properties. Furthermore, considering the structural similarity among CaFADS and the FADSs from the human pathogens Streptococcus pneumoniae (included in the World Health Organisation [WHO] priority list of antibiotic resistant pathogens; SpnFADS) and Mycobacterium tuberculosis (the World’s leading infectious killer; MtFADS), we explored the potential antimicrobial effect of the FADS HTS hits in these microorganisms by determining their minimal inhibitory concentration (MIC). Some of the HTS hits demonstrated high FADS inhibitory activity in vitro, but their antimicrobial activity revealed that uptake of these compounds by bacterial cells could be suboptimal. Collectively, our results validate our approach for discovering antimicrobials targeting bacterial FADSs, and for identifying inhibitors that constitute a great starting point for future developments of novel antimicrobials. Methods Protein purification and quantification Recombinant CaFADS was overexpressed in BL21 (DE3) E. coli cells and purified as previously described 30 . Recombinant SpnFADS was overexpressed in E. coli strain Bl21 Star TM (DE3) and purified as previously described 23 . Pure samples were dialysed against 20 mM PIPES, pH 7.0 and quantified using the theoretical extinction coefficients e 279 ¼27.8 mM 1 cm 1 and 28.8mM 1 cm 1 for CaFADS and SpnFADS, respectively. The purity of each protein was tested by 15% SDS-PAGE. Chemicals The Prestwick Chemical Library V R , containing 1240 molecules approved by the Food and Drugs Administration (FDA) and European Medicines Agency (EMA), was selected for the HTS. Compounds were dissolved in 100% DMSO at 10 mM. All the HTS hits were subsequently acquired from Sigma Aldrich, Prestwick or Carbosynth and dissolved in 100% DMSO to prepare stock solutions at 50 and 10 mM. The purity of all compounds was >95%, as determined by High performance liquid chromatography (HPLC), thin layer chromatography (TLC), NMR, IR or basic titration. Activity-based high-throughput screening for CaFADS An activity-based HTS was performed on the 1240 compounds of the Prestwick Chemical Library V R . The assays consisted in recording the time dependent decrease in the fluorescence of the isoalloxazine ring, produced upon transformation of RF and FMN into FAD, as a consequence of the fluorescence quenching in this later flavin 27 . When either the RFK or the FMNAT activities were inhibited, less FAD was produced and, consequently, the fluorescence decrease registered in a specific time interval was less pronounced. Measurements were carried out using a multimode microplate reader, Synergy TM HT Biotek, with BRAND 96well plates pure Grade TM . To optimise the assay conditions, a previous study was performed using constant concentrations of RF, ATP and CaFADS (5, 50 and 0.4 mM, respectively) and variable concentrations of MgCl 2 (0.2–10mM) and DMSO (0–12.5% v/v). Optimum conditions were 2.5% DMSO, 10 mM MgCl 2 and sensitivity 70. HTS reaction mixtures contained 5 mM RF, 0.4 mMCaFADS, 10 mM MgCl 2 , in PIPES 20 mM, pH 7.0, 2.5% DMSO, and the corresponding compound of the chemical library at a final concentration of 250mM. Reactions were initiated through addition of 50 mM ATP, being the final reaction volume 100 ml. Controls, which contained the reaction mixture but not any chemical from the library, were added both to the first and last columns of the plate. Flavin fluorescence in each well was registered at 25 C, every 50 s during 15 min. Excitation and emission wavelengths were 440 and 530 nm, respectively. The slope of the resulting line, recorded between 0 and 6 min, was calculated for every compound, and also for the controls, as well as the fluorescence change per time unit (DF/Dt). The compounds that decreased the reaction rate below the average reaction rate of the controls minus the standard deviation could be preselected as potential inhibitors, but we reduced further the cutoff by selecting only those compounds inhibiting more than 50% of the controls activity as HTS hits. Identification of the activity inhibited by each of the HTS hits The decreasing of the reaction rate by the presence of the HTS hits might be consequence of the compounds inhibiting the RFK activity, the FMNAT one, or both of them; also, it could be a false positive due to the properties of pan assay interference compounds (PAINS). To clarify this point, we first checked that there were no PAINS among the HTS hits using the FAF-Drug4 web server 31 . Then, the RFK and FMNAT reactions were individually assayed in the presence of the HTS hits at 25C. Reaction mixtures contained 50 mM ATP, 5 mM RF, 0.4 mMCaFADS in 20 mM PIPES, pH 7.0, 0.8 mM MgCl 2 , when assaying the RFK activity, and 50 mM ATP, 10 mM FMN, 0.4 mMCaFADS in 20 mM PIPES, pH 7.0, 10 mM MgCl 2 when measuring the FMNAT reaction. Each HTS hit was tested again at 250mM for each of the two enzymatic reactions. Finally, reactions were stopped by boiling the samples at 100 C for 5 min, and the precipitated protein was eliminated through centrifugation. The transformation of RF into FMN or FAD was evaluated through flavins separation by HPLC, as previously described 27 . Those HTS hits decreasing the FMNAT activity by more than 95% of the controls, without significantly affecting the RFK activity (rates over 75% those of the controls) were selected as FMNAT hits for further assessment. When assaying the HTS hits against 242 M. SEBASTI AN ET AL. the FMNAT activity of SpnFADS, similar conditions were used but samples contained 3 mM sodium dithionite to maintain the flavin in its reduced state 23 . Data were processed as previously reported 27 . All the experiments were performed in triplicate. Determination of the potency of FMNAT hits To determine the IC 50 values of the FMNAT hits, the FMNAT activity was assayed at different concentrations of each inhibitor (0–100 mM range) and 25 C. Experiments were performed and analysed through HPLC as described above. Positive controls (without any hit compound) were included in every reaction set. DMSO concentration was kept at 2.5% in all samples. All the experiments were performed in triplicate. Determination of the inhibition mechanism of CaFADS by FMNAT hits 24, 27 and 31 The inhibition mechanism was further studied for the three FMNAT hits that showed the lowest IC 50 and minimal residual FMNAT activities, namely, 24, 27 and 31. Reaction mixtures containing 0–100 mM of each compound, 1–20 mM FMN and 400 mM ATP were used when analysing the inhibitory effect of the compound regarding the FMN substrate, while 5–400 mM ATP and 15mM FMN when analysing the effect of the inhibitor regarding the ATP substrate. All the experiments were carried out in 20 mM PIPES, 10mM MgCl 2 , pH 7.0, 2.5% DMSO at 25 C, being the final reaction volume 500 ml. The reactions were initiated by addition of CaFADS at a final concentration of 40 nM, followed by 1 min incubation. The flavin composition of the supernatant was analysed as previously described 27 . All the experiments were performed in triplicate. The effect of the inhibitors on K m and V max was determined by fitting the data sets to the Michaelis–Menten model. Additionally, data were globally fit to Lineweaver–Burk equations for competitive, uncompetitive, non-competitive or mixed inhibition, yielding K i , as well as K i 0when applying, for each compound (Equations (1–4), respectively). 1 V0¼ 1þI ½ Ki  Km Vmax 1 ½Sþ1 Vmax (1) 1 V0¼ 1þI ½ K'i  Vmax þKm Vmax 1 ½S(2) 1 V0¼ 1þI ½ Ki  Vmax þ 1þI ½ Ki  Km Vmax 1 ½S(3) 1 V0¼ 1þI ½ Ki  Km Vmax 1 ½Sþ 1þI ½ K'i  Vmax (4) Thermodynamic characterisation of binding of hits 24, 27 and 31 through isothermal titration calorimetry (ITC) ITC experiments were performed to characterise the protein’s affinity for the selected compounds, as also the thermodynamic parameters that drive the interaction. Experiments were carried out in an AutoITC200 (MicroCal) thermostated at 25 C. In these experiments, 400 mM of each compound were used to titrate 25mMCaFADS contained in a 200 mL cell. However, when saturation of the protein was not reached, higher concentrations of compounds were employed. The titrations were performed by stepwise injections of the titrating compound. Up to 19 injections of 2ml were added to the cell sample and mixed using a 1000 rpm stirrer syringe. The compounds and the protein were dissolved in 20 mM PIPES, pH 7.0, 10 mM MgCl 2 and degassed prior to titration. DMSO was added to the protein and ligand samples, until reach a final concentration of 3%. The association constant (K a ), the enthalpy variation (DH) and the binding stoichiometry (N), were obtained through non-lineal regression of the data to a model for one or two independent binding sites, implemented in Origin 7.0 (OriginLab) as previously described 27,29 . The entropic contribution (TDS), the Gibbs free energy (DG) and the dissociation constant (K d ) were obtained through essential thermodynamic equations. Docking of hits 24, 27 and 31 to the FMNAT module of CaFADS The AutoDock4.2 software 32–34 and the coordinates of a monomer from CaFADS (PDB 2X0K) 24 were used to obtain the interaction models with 24, 27 and 31. The space sampling was defined using a grid box of 90 points in each dimension, and placing the H57 NE atom of the FMNAT module as the grid box centre. The grid size was 0.375 Å. The search was performed using the lamarkian genetic algorithm, with a starting population of 150 individuals, using 25,000,000 energy evaluations and 27,000 generations. The 24, 27 and 31 initial structures for the docking protocol were optimised using the functional B3LYP with the basis set 6–31 G (d,p) and the gaussian09 software 35 . The structural poses with the lowest docking score were selected and analysed. Determination of the antibacterial activity of the HTS hits The antimicrobial activity of the HTS hits was tested by the colorimetric method of the resazurin microplate assay 36 according to broth microdilution method guidelines (CLSI; Clinical and Laboratory Standards Institute). Serial 2-fold dilutions of the HTS hits were performed in BHI medium, in 96-well microtiter plates, with a final volume of 100 ml per well. Subsequently, liquid cultures of C. ammoniagenes ATCC 6872 in logarithmic phase were adjusted to 10 6 CFU/ml in BHI broth, and 100 ml of this suspension were added to each well, making a final inoculum of 5 10 5 CFU/ ml. Plates were incubated 16 h at 37 C. 30ml of 0.1 mg/ml resazurin solution were then added to each well, and results were observed after 4 h of incubation at 37 C. Resazurin (blue) is an indicator of bacterial growth, since metabolic activity of bacteria reduces it to resorufin (pink). The minimum inhibitory concentration (MIC) is the lowest concentration of compound that does not change the resazurin colour from blue to pink. Similarly, the HTS hits were also assayed at 37 C against M. tuberculosis ATCC 27,294 and S. pneumoniae ATCC 49619 cells. In these experiments the initial cell concentration was also 5 10 5 CFU/ml, and plates were incubated for 10 h (S. pneumoniae), and 6 days (M. tuberculosis) before addition of resazurin. Results were observed after incubation with resazurin 4 h and two days for S. pneumoniae and M. tuberculosis, respectively. In these experiments, culture media were; Middlebrook 7H9 (Difco) supplemented with 10% ADC (0.2% dextrose, 0.5% V fraction BSA and 0.0003% bovine catalase) (BD Difco) and with 0.5% glycerol (Scharlau) for M. tuberculosis growth, and BHI supplemented with 4% FBS (Gibco) for S. pneumoniae growth. Statistics Results are expressed as the mean ±the standard deviation (SD) or as the mean ± the standard error (SE) of the regression. When indicated, one-way analysis of variance (ANOVA) was performed to determine statistical significance. JOURNAL OF ENZYME INHIBITION AND MEDICINAL CHEMISTRY 243 Results Identification of potential inhibitors of the CaFADS activities through HTS To identify potential inhibitors of the CaFADS activities we designed the activity-based HTS assay described in the methods section, which allowed to determine rates for the transformation of RF into FAD (via the FMN intermediate) in each one of the plate wells. Wells containing chemicals of the library and decreasing reaction rates relative to the controls (absence of chemicals) were preselected as containing compounds that are potential inhibitors of at least one of the CaFADS activities. Thus, among the 1240 compounds of the chemical library, 140 (13.5%) reduced the CaFADS activity levels below the mean of the positive controls minus twice its standard deviation, and of them, 37 (3.6%) reduced the positive controls average rate for FAD formation in a factor higher than 0.5 (Figure 1). Those 37 compounds were selected as the HTS hits (Supplementary Chart SD1). Effect of the HTS hits on the RFK and FMNAT activities of CaFADS We then move to identify which one of the activities (RFK or FMNAT), and in which extension, was affected by each one of these 37 HTS hits. With this aim, we assayed the effect of the HTS hits both on the CaFADS RFK and FMNAT activities. Figure 2 and Table 1 summarise the results. Comparison of Figure 2(A,B) shows that, in general and under the assay conditions, the 37 HTS hits produced a stronger deleterious effect on the FMNAT activity (all decreased the activity of the controls in more than 50%) than on the RFK one. The FMNAT module of CaFADS does not have sequence or structural homology with the mammalian protein but the RFK module belongs to the eukaryotic RFKs family. Therefore, we decided to continue the study with the HTS hits that inhibit the FMNAT activity, since they are more likely to be specific to the bacterial proteins. Thus, we choose the HTS hits that decreased the FMNAT activity below 5% of that of the controls, but maintained over 75% the RFK activity (Figure 2,Table 1). Thus, among (2)\(11), tiratricol (15), benzbromarone (17), oxantel pamoate (19), Chicago sky blue 6B (24), gossypol (27), flunixin meglumine (31) and oxaprozin (43) (Chart 1) were selected as FMNAT hits. It is Figure 2. Effect of the HTS hits on the RFK and FMNAT activities of CaFADS. Residual (A) RFK and (B) FMNAT activities when assayed in the presence of 250 lM of the 37 HTS hits. In (A), the columns below the dashed line present statistical significant inhibition by the corresponding hit (p<0.002, 67% remaining activity) related to the control CaFADS RFK activity. In (B), all hits produce statistical significant inhibition (p<0.0001, dashed line) when compared with the controls of the CaFADS FMNAT activity. Solid lines indicate 75 and 5% of the control RFK and FMNAT activities, respectively. The HTS hits displaying <5% and >75% of the control FMNAT and RFK activities, respectively, were selected for further study. Experiments carried out in 20 mM PIPES, pH 7.0, 2.5% DMSO at 25 C, with 7.5 lM RF, 350 lM ATP, 0.8 mM MgCl 2 (for the RFK activity) or 15 lM FMN, 350 lM ATP, 10 mM MgCl 2 (for the FMNAT activity) (n¼3; mean ± SD). Figure 1. Activity-based high throughput screening (HTS) for the discovery of inhibitors of the RFK and/or the FMNAT activities of CaFADS. (A) Example of the flavin fluorescence evolution over time for three of the identified hits and for control assays. Reaction mixtures were incubated at 25 C and contained 5 mM RF, 50 mMATP,0.4mMCaFADS, 10 mM MgCl 2 , in PIPES 20 mM, pH 7.0, 2.5% DMSO. The black symbols and lines correspond to kinetic traces at wells containing library compounds at 250 mM, while grey ones correspond to control wells. (B) Initial velocities (Dfluorescence/s) for the reactions in each of the wells of a HTS plate. Data from wells containing chemical library compounds are in black while controls are in grey. The solid line represents the average velocity obtained for the positive controls of the reaction and the dotted lines are the average velocity plus and minus the standard deviation. The letters and numbers indicate the position of the well in the plate (row and a column respectively) for each specific selected measurement. A bold dashed line indicates 50% the rates of controls. 244 M. SEBASTI AN ET AL. Table 1. RFK and FMN residual activities of CaFADS in the presence of the HTS hits. % Residual RFK activity 55–50 50–75 75 % Residual FMNAT activity 5 9, 29, 33, 37 14,47 2, 11, 15, 17, 19, 24, 27, 31, 43 5–50 35 38 22,32 1, 3, 4, 5, 6, 7, 8, 10, 12, 13, 16, 18, 25, 28, 39, 40, 44, 46 50 ––– – Values measured at 25 C, in 20 mM PIPES, pH 7.0, 2.5% DMSO and 0.8 or 10 mM MgCl 2 when assaying the RFK or the FMNAT activities, respectively. The final concentration of each HTS hit was 250 lM and saturating concentrations of all the substrates were used. The compounds highlighted in italics completely inhibited the FMNAT activity with minor effects on the RFK one, and were selected as FMNAT hits. 2(Tolfenamic acid, 13710-19-5) NN OH O F F F 11 (Niflumic acid 4394-00-7) O I I I OH OH O 15 (Tiratricol, 51-24-1) O O CH 3 Br Br OH 17 (Benzbromarone 3562-84-3) 19 (Oxantel pamoate, 68813-55-8) Na + Na + Na + S NH 2 OH O OO S N O OO N O CH 3 O N CH 3 N OH NH 2 S S O O O O O ONa + 24 (Chicago sky blue 6B, 2610-05-1) 27 (Gossypol, 303-45-7) 31 (Flunixin meglumine,42461-84-7) N O O OH 43 (Oxaprozin, 21256-18-8) OH N CH 3 OH OH OH OH NN CH 3 F F F COOH H H H H Chart 1. Chemical structures of compounds selected as FMNAT hits for CaFADS. JOURNAL OF ENZYME INHIBITION AND MEDICINAL CHEMISTRY 245 worth noticing that although some of these compounds are apparently structurally related with other of the HTS hits, slight differences in functional groups and geometries induce different enzymatic responses in the FMNAT or RFK activities. This is a fact of particular interest when developing specific inhibitors. To rate the power of these 9 FMNAT hits as inhibitors of the CaFADS FMNAT activity, their half maximal inhibitory concentrations (IC 50 ) and the remaining activity at 50 mM of each compound were determined (Figure 3,Table 2). The 9 compounds yielded IC 50 values in the micromolar range, reducing by more than half the activity of the controls. Considering both their IC 50 value and residual activity, the most potent inhibitors were 24 (IC 50 ¼0.4 ± 0.1 mM), 27 (IC 50 ¼0.5 ± 0.1 mM) and 31 (IC 50 ¼6.6 ± 0.6 mM). These three compounds produce residual activities below 25% of the controls. Compound 43 also showed high inhibitory potency, but due to its low solubility in the working buffer it was discarded. The inhibition mechanisms of 24, 27 and 31 To determine the inhibition mechanism of the hits 24, 27 and 31, we measured the FMNAT activity of CaFADS in the presence of increasing concentrations of each compound. Considering that this is a bi-substrate activity, for each compound we carried out two set of experiments at; (i) saturating ATP and different FMN concentrations, and (ii) saturating FMN and increasing ATP concentrations. Then we fit our experimental data to the Michaelis–Menten model, obtaining K m and k cat values. The analysis of the evolution of these constants on the hits concentrations, together with the corresponding Lineawever–Burk plots [representation of data as double inverses and fit to Equations (1–4)] (Figure 4,Supplementary Figures SD2 and SD3, Supplementary Table SP1) allowed identifying the inhibitory mechanisms of the 24, 27 and 31 hits, as well as the corresponding inhibition constants (K i or, K i and K i 0)(Table 3). The experiments carried out at saturating ATP (varying the FMN concentrations), revealed that the three compounds are strong non-competitive inhibitors of the CaFADS FMNAT activity regarding the FMN substrate (K i values around 0.08 mM, Table 3). Nevertheless, when using a constant and saturating FMN concentration but varying the ATP concentration, their inhibition mechanisms differ among them. 24 is a strong ATP uncompetitive inhibitor (K i ¼0.08 ± 0. 03 mM), therefore, it is able to bind the CaFADS-ATP complex and reduce the amount of enzyme that is available to react. 27 is a strong competitive inhibitor regarding the ATP substrate (K i ¼0. 06 ± 0.01 mM), while 31 is a considerably poorer mixed inhibitor. Thus, 31 is able to bind to both the free enzyme and the CaFADSATP complex, although binding constants indicate that binding to free CaFADS is preferred (K i 3.5 ± 1.0 mM versus K i 018.4 ± 4.0 mM, Table 3). Binding of 24, 27 and 31 to CaFADS The interaction of CaFADS with compounds 24, 27 and 31 was characterised using ITC. This is a very powerful technique because first the shape of the thermograms informs us about the number of binding sites related to the thermodynamic nature of binding. Then, fitting of experimental data to the equations describing binding models allow determining thermodynamic binding parameters, as well as binding stoichiometry at each of the different binding sites. Thus, analysis of our ITC titrations indicates that the three compounds bind the enzyme at, at least, one binding site (Figure 5(A),Table 4). Thus, we identified in CaFADS a unique binding site of moderate-low affinity for 31 (N1, K d ¼30.9 ± 2.8 mM) and two binding sites of high and similar affinity for 27 (N2, K d,av ¼0.7 ± 0.07 mM, this K d,av value is an average value, since the similarity between the two binding sites prevents them to be distinguished). These 27 and 31 binding sites are expected to be located at the enzyme FMNAT domain, since that is the inhibited activity. The interaction between the hit 24 and the enzyme resulted more complicated, because our ITC data indicate that compound 24 binds to the protein at three sites. Two of them show similar and strong affinity (K d,av ¼1.1 ± 0.1 mM) and therefore, we suggest that they might be located at the FMNAT module, since that is the activity mainly inhibited by compound 24. The third site for 24 binding has lower affinity (K d ¼161 ± 20 mM). Since this compound also mildly inhibits the RFK activity, we presume this third binding site may be at the RFK module. All bindings here characterised (with the only exception of the low affinity binding site for 24) are favoured by the enthalpic contribution (Figure 5(B),Table 4). This suggests a net gain of H-bonding and ion-pair interactions and indicates specific binding interactions. Regarding the entropic contribution to the binding free energy, it is small and favourable for 24 and 27, and only slightly unfavourable for 31 (Figure 5(B),Table 4). However, it drives the binding of 24 to the Figure 3. Dose–response curves for the FMNAT activity of CaFADS in the presence of representative hits. Values derived from these representations are included in Table 2. Experiments performed at 25 C in 20 mM PIPES, pH 7.0, 10 mM MgCl 2 , 2.5% DMSO, with 15 lM FMN and 350 lM ATP (n¼3, mean ± SD). Table 2. Effect of the FMNAT hits on the FMNAT activity of CaFADS. FMNAT hit Residual activity a (%) IC 50b (lM) 2 41.1 ± 5.2 8.9 ± 1.0 11 34.6 ± 4.9 9.0 ± 1.3 15 45.3 ± 4.2 40.7 ± 3.9 17 33.5 ± 10.1 12.8 ± 3.4 19 43.3 ± 3.9 20.8 ± 2.6 24 3.6 ± 0.2 0.4 ± 0.1 27 6.9 ± 0.8 0.5 ± 0.1 31 24.5 ± 1.9 6.6 ± 0.6 43 c 20.3 ± 5.2 1.0 ± 0.5 Experiments carried out at 25 C, in 20 mM PIPES, pH 7.0, 10 mM MgCl 2 at saturating FMN and ATP. All samples contained 2.5% DMSO (n¼3, mean ± SD). a Remaining activity in the presence of 50 lM of each compound. All data show statistical significance differences when compared with activity in the compound absence (p<0.0001). b Compounds assayed in the 0–100 lM concentration range. c This compound shows very low water solubility, so it was discarded to continue the study even though its good properties. 246 M. SEBASTI AN ET AL. RFK module (site 3 Figure 5(B)), revealing that it might be nonspecific and that could occur due to the compound hydrophobicity and rigidity 37 . To investigate how the FMNAT module of CaFADS accommodates these compounds, we performed a computational proteinligand docking (Figure 6 and Supplementary Figure SD4). In the highest-scoring docking mode, as well as for the best five poses (purple molecule in Figure 6(B), Supplementary Figure SD4, respectively), 24 1 (1 indicates the first molecule of 24 docked) is situated in the substrates binding pockets (Figure 6A) 24 . One of the 24 1 moieties binds through one of its sulphates to S164 and H31. 24 1 is in addition H-bonded to the N125 catalytic base 17 ,as well as to Y106, T127 and N131 at the loop that forms the upper flavin ring binding site (Figure 6(B)). Since our ITC data suggest two binding sites for compound 24 at the FMNAT module, we carried out an additional docking analysis to identify the second site, 24 2 (pink molecule in Figure 6(B)), using as receptor our highest-scoring FADS:24 1 model. Considering the binding energies (Table 4) this site is expected to be less populated, but given that 24 is an ATP uncompetitive inhibitor, the docking binding energy could improve if we consider the ATP substrate presence instead of the 24 1 molecule. Additionally, it is worth noting that in the presence of substrates or products, the 24 binding conformations might differ from the ones presented here. In the most favourable docking poses, the first molecule of 27 bound to the protein (27 1 ) is H-bonded by T127 and the N125 catalytic base at the binding site of the ATP phosphates 17 (purple molecule in Figure 6(C) and SD4), in agreement with 27 being an ATP competitive inhibitor. Because two binding sites were again predicted by ITC at the FMNAT site, we carried out a second docking using as receptor the highest-scoring FADS:27 1 model. 27 2 is stabilised by H-bonds with the N-terminal of the a6n helix, T165 and R168, at the ATP binding site entrance (pink molecule in Figure 6(C)). We observed direct interaction between the two 27 molecules, and 27 2 somehow resembling the 24 1 binding (purple molecule in Figure 6(B)). For 31, we only carried Figure 4. Hit 27 as an inhibitor of the FMNAT activity of CaFADS. Michaelis–Menten plots at different concentrations of 27 and saturation (A) of ATP and (B) of FMN. Lineaweaver–Burk representations with global fit to (C) non-competitive inhibition at saturating ATP and (D) competitive inhibition at saturating FMN. Reaction rates obtained in 20 mM PIPES, pH 7.0, 10 mM MgCl 2 ,at25 C, with 15 lM FMN and 10–450 lM ATP (FMN saturating) or with 350 lM ATP and 0.5–20 lM FMN (ATP saturating). All samples contained 2.5% DMSO (n¼3, mean ± SD). Table 3. Inhibition constants and mechanisms of the best FMNAT hits relative to the FMNAT activity of CaFADS. Saturating ATP Saturating FMN K i (lM) Inhibition mechanism K i (lM) K ’ i (lM) Inhibition mechanism 24 0.07 ± 0.01 Non-competitive 0.08 ± 0.03 –Uncompetitive 27 0.08 ± 0.01 Non-competitive 0.06 ± 0.01 –Competitive 31 0.09 ± 0.03 Non-competitive 3.5 ± 1.0 18.4 ± 4.0 Mixed Experimental data recorded at 25 C in 20 mM PIPES, pH 7.0, 10 mM MgCl 2 and 2.5% DMSO. Data obtained by globally fitting the experimental data to the corresponding Lineweawer–Burk inhibition model. JOURNAL OF ENZYME INHIBITION AND MEDICINAL CHEMISTRY 247 fmicb-10-00046 January 28, 2019 Time: 18:36 # 2 Sanz-García et al. Mycobacterial Aminoglycoside Acetyltransferases (ANTs), have been widely detected in most pathogenic bacteria as a major determinant of resistance; in these bacteria the presence of aminoglycoside-modifying enzymes correlated with patterns of AG susceptibility (Smith and Baker, 2002). The modified AG (either by acetylation, phosphorylation or nucleotidylation) fails to inhibit their bacterial target, the 30S ribosomal subunit (Smith and Baker, 2002). Most genes encoding aminoglycosidemodifying enzymes are plasmid-located (indicative of a potential acquisition by horizontal gene transfer processes) and confer the bacterial hosts with high levels of AG resistance (Davies and Wright, 1997). In mycobacteria, however, resistance to AGs resulted mainly from mutations of the ribosome components that prevent the drugs from inhibiting its function (Jugheli et al., 2009;Zhang and Yew, 2009;Shcherbakov et al., 2010). This is due to the fact that most mycobacterial species have either one (like Mycobacterium. tuberculosis) or two (like Mycobacterium fortuitum) ribosomal operons, hence making dominant those mutations acquired in their components (Magnet and Blanchard, 2005;Shi et al., 2013). The presence of aminoglycoside-modifying enzymes, mostly AAC, in mycobacterial species has been reported over the years (as detailed in the following sections), and the role of such AACs has been explored, originally for their contribution to AG resistance, and more recently for their role in other bacterial processes, which has resulted in the interest of developing inhibitors of these enzymes (Jana and Deb, 2005;Labby and Garneau-Tsodikova, 2013). In this review, we will summarize major findings on two mycobacterial AACs, the AAC(20)-I and Eis enzymes, that have resulted in a Copernican turn for AACs in mycobacteria, from being putative drug resistance mechanisms, to reach the status of novel drug targets. AACs IN NON-TUBERCULOUS MYCOBACTERIA The first detection of AACs in mycobacteria (Hull et al., 1984) was reported in a group of M. fortuitum isolates, an opportunistic fast-growing mycobacteria. Biochemical assays of crude extracts from M. fortuitum strains revealed the presence of AAC activity, strongly acetylating gentamicin and kanamycin A, along with other AGs. This substrate profile was consistent with that of AAC(3) enzymes that had been previously described in Pseudomonas and Enterobacteriaceae (Angelatou et al., 1982), although confirmation at the genetic or molecular levels were not done at that time. Surprisingly, the AG susceptibility profile of M. fortuitum could not be correlated with the activity of AACs, indicating that in this species AACs were not the major responsible for AG resistance; it was neither correlated with the presence of plasmids, hence suggesting a chromosomal location (Hull et al., 1984). In fact, the frequency of resistant mutants to kanamycin and amikacin in M. fortuitum and the related species Mycobacterium chelonae ranged between 10−4 and 10−7(Wallace et al., 1985). This relatively high frequency of mutations, along with the fact that AAC activity was detected at similar levels between susceptible and resistant strains, led the authors to suggest that ribosome alterations were the main factor responsible of AG resistance in these species (Wallace et al., 1985). In another study (Udou et al., 1986), altered transport or permeability of AGs was identified as a contributor to AG resistance in M. fortuitum, since ribosomes from a clinical isolate were inhibited by one tenth of the MIC of AGs: for example, the MIC of kanamycin for M. fortuitum was 50 µg/ml, and in cell-free systems, 5 µg/ml of kanamycin reduced the activity of ribosomes to 13% in comparison with drug-free controls; similar results were obtained when using gentamicin or paromomycin (Udou et al., 1986). The biochemical analysis of crude extracts from other nontuberculous mycobacteria such as Mycobacterium smegmatis, Mycobacterium phlei,Mycobacterium vaccae, and Mycobacterium kansasii, from both clinical and environmental origins, revealed similar characteristics to those found in M. fortuitum: crude extracts from all strains contained AAC enzymatic activity, but no correlation with AG susceptibility profile could be established (Udou et al., 1987;Ho et al., 2000). In other mycobacterial species such as Mycobacterium avium and Mycobacterium intracellulare, however, AAC activity could not be detected (Ho et al., 2000). In a recent study, using cell-free translation assays, ribosomes of Mycobacterium abscessus and M. smegmatis were inhibited by both AGs having a 20-amino group (tobramycin, dibekacin, and kanamycin B) and by those having a 20-hydroxyl group (amikacin, and kanamycin A). However, in M. abscessus, those AGs having a 20-amino group (tobramycin, dibekacin, and kanamycin B) were less efficient in killing the cells and inhibiting its ribosomes, which is consistent with the presence of a highly active AAC(20) activity in this species. These findings suggested that in M. abscessus, AACs could somewhat mitigate the bactericidal effect of its AGs substrates (Maurer et al., 2015). Interestingly, when characterizing the enzymatic activity of crude extracts, early reports detected that substrates such as amino sugars, malonyl-CoA, propionyl-CoA or butyryl-CoA inhibited the enzymatic activity of AACs from mycobacterial species. Such effect of non-acetyl CoA donors had never been described for the AAC(3) enzymes from other bacteria (Udou et al., 1987). Altogether, these findings strongly suggested for the very first time that, in mycobacteria, AACs could have important metabolic functions, and their contribution to AG resistance could be marginal (Udou et al., 1989). AAC(20), A NEW ENZYME COMES INTO SCENE In Providencia stuartii, a Gram-negative species, phylogenetically distant from mycobacteria, a novel class of AAC was identified as AAC(20), because of its acetylating activity against gentamicin and lack of enzymatic activity against kanamycin A. The gene encoding AAC(20)-Ia was cloned from P. stuartii (Rather et al., 1993) and found to be present in the chromosome of all isolates of this bacteria. In P. stuartii, the expression of the aac(20)- Ia gene was controlled by several transcriptional regulators (Macinga and Rather, 1999), suggesting that this enzyme could play an important role, beyond its contribution to drug resistance (Franklin and Clarke, 2001). In fact, AAC(20)-Ia contributes to Frontiers in Microbiology | www.frontiersin.org 2January 2019 | Volume 10 | Article 46 fmicb-10-00046 January 28, 2019 Time: 18:36 # 3 Sanz-García et al. Mycobacterial Aminoglycoside Acetyltransferases BOX 1 | Aminoglycosides: origins, structure, mode of action, resistance and clinical use. Most aminoglycoside antibiotics are produced by bacterial species of the genus Streptomyces, such as the antitubercular streptomycin that is produced by S. griseus being the first antibiotic identified from bacteria. Other genera producing aminoglycosides are Micromonospora and Bacillus. Many semi-synthetic aminoglycosides, such as amikacin, have also been produced (Mingeot-Leclercq et al., 1999;Magnet and Blanchard, 2005;Shi et al., 2013). Structurally, aminoglycosides are formed by an aminocyclitol (commonly, 2-deoxystreptamine) with additional amino sugars bound by glycosidic bonds. There are two large families of 2-deoxystreptamine aminoglycosides, those carrying substitutions at positions 4 and 5 of the 2-deoxystreptamine ring (including neomycin, paromomycin, lividomycin, ribostamycin and butirosin) and those being substituted at positions 4 and 6 of the 2-deoxystreptamine (including kanamycin, amikacin, tobramycin, dibekacin, arbekacin, gentamicin, isepamicin, sisomicin, and netilmicin). The carbon atoms in the sugar bound to position 4 of the 2-deoxystreptamine ring are named with primed numbers (’), and those in the sugar bound to positions 5 or 6 of the 2-deoxystreptamine ring are named with double-primed numbers (”). Other aminoglycosides contain aminocyclitols distinct to 2-deoxystreptamine (this is the case of streptomycin or apramycin) or are formed by fused amino sugar rings (i.e., spectinomycin) (Magnet and Blanchard, 2005;Shi et al., 2013). The following figure shows the structure of kanamycin A, an example of 4,6 di-substituted 2-deoxystreptamine aminoglycoside antibiotic. The bacterial target of aminoglycosides is the 30S small ribosomal subunit, and their global effect is the interference with protein synthesis. In bacterial cells, translation is initiated when the 30S ribosomal subunit binds the Shine-Dalgarno sequence (because of sequence complementarity between the Shine-Dalgarno sequence and the 16S rRNA molecule of the 30S ribosomal subunit), which is normally present in the 50untranslated region of mRNAs. Next, initiation factors and fMet-tRNA will join the complex that finally will recruit the 50S ribosomal subunit in order to start translation. Aminoglycoside binding to the 30S subunit does not affect the binding of mRNA and the large 50S subunit, so translation can proceed. However, aminoglycosides differ in their binding site at the 30S subunit, hence affecting the protein production at different levels. Whereas spectinomycin blocks translocation (hence being bacteriostatic), streptomycin and most 2-deoxystreptamine aminoglycosides lock the ribosome in a conformation that is prone to introducing erroneous aminoacyl-tRNAs. The accumulation of aberrant proteins in the bacteria results in cell death (Davies and Wright, 1997;Magnet and Blanchard, 2005;Shell et al., 2015). Resistance to aminoglycosides may be due to several mechanisms (Davies and Wright, 1997;Magnet and Blanchard, 2005). Reduced uptake (which can be a consequence of alterations in the composition of bacterial membrane or to metabolic conditions like anaerobiosis) or the action of efflux pumps can lead to limited intracellular concentration of aminoglycosides, hence causing resistance. Mutations in 16S ribosomal RNA or certain ribosomal proteins such as S12 (encoded by rpsL gene) lead to aminoglycoside resistance through target modification; this is also achieved after the action of methyltransferases, which introduce methyl groups in guanine or adenine nucleotides of 16S ribosomal RNA. The presence of aminoglycoside-modifying enzymes is, however, the most prevalent mechanism of aminoglycoside resistance; there are three types of aminoglycoside-modifying enzymes: aminoglycoside N-acetyltransferases (AAC), O-phosphotransferases (APH), and O-nucleotidyltransferases (ANT). Aminoglycoside-modifying enzymes are named by using the aforementioned abbreviations, followed by a number in brackets indicating the site of modification in the aminoglycoside molecule (as explained above), a Roman numeral related with substrate profile, and a lower-case letter for differentiating isoenzymes, i.e., AAC(20)-Ib. Clinically, due to their otoand nephrotoxicity and the rising prevalence of resistance, aminoglycosides are commonly reserved as a second line of treatment of serious infections. Due to their low absorption when given orally, aminoglycosides need to be administered through injections. Streptomycin was a first-line drug in the treatment of tuberculosis, and in our days, kanamycin and amikacin are listed as second line drugs against this disease. Spectinomycin is used against Neisseria gonorrhoeae infections (Suay-Garcia and Perez-Gracia, 2018). Pseudomonas aeruginosa infections in cystic fibrosis patients, septicemia, endocarditis and several other infections caused by non-tuberculous mycobacteria, Gram-positive or Gram-negative bacteria can be treated efficiently by using aminoglycosides, either alone or in combinations with other antibacterials such as the beta-lactam antibiotics (Mingeot-Leclercq et al., 1999;Magnet and Blanchard, 2005;Bassetti et al., 2018). O-acetylation of peptidoglycan affecting cell morphology and the expression of autolysins, and can use acetylated peptidoglycan precursors as donors of acetyl groups, another indication about the role of this enzyme beyond resistance (Payie et al., 1995, 1996; Clarke et al., 1996;Payie and Clarke, 1997). AAC(20) in Mycobacteria Given the high AG acetylating activity against gentamicin, tobramycin, netilmicin and its derivatives 20-N-ethyland 60-N-ethyl-netilmicin, kanamycin A and kanamycin B found in M. fortuitum (Hull et al., 1984;Ainsa et al., 1996), we launched a molecular approach aimed at characterizing the determinant of AG resistance in this species. A genomic library of M. fortuitum was transformed in M. smegmatis, and characterization of AG resistant clones allowed the identification of a M. fortuitum gene showing sequence similarity to aac(20)-Ia of P. stuartii. The enzyme of M. fortuitum was named AAC(20)-Ib (Ainsa et al., 1996) and was capable of acetylating gentamicin, but not kanamycin A, hence indicating that another AAC enzyme should be present in M. fortuitum. Similarly to P. stuartii, the aac(20)-Ib gene was found in all strains of M. fortuitum regardless of the phenotype of AG resistance, suggesting other roles for the AAC(20)-Ib in this species. Further studies (done by database searching or by southern blot analysis using the probe of aac(20)-Ib gene) demonstrated the presence of aac(20)- Igenes in other mycobacterial species, including M. tuberculosis (the major pathogenic species in this genus), Mycobacterium leprae, and M. smegmatis, indicating thus that the presence Frontiers in Microbiology | www.frontiersin.org 3January 2019 | Volume 10 | Article 46 fmicb-10-00046 January 28, 2019 Time: 18:36 # 4 Sanz-García et al. Mycobacterial Aminoglycoside Acetyltransferases of aac(20)-I genes in mycobacteria could be universal (Ainsa et al., 1997). Interestingly, the expression of the aac(20)-Id gene in M. smegmatis was driven from two promoters, and the strongest one produced a leaderless transcript having a GTG translation start codon at its 50end (Mick et al., 2008). Leaderless transcripts are those in which the transcription start site coincides with the translation start codon; although representing a rather unusual feature in the model organism E. coli, they are quite common in mycobacteria, where 25% of all transcripts are predicted to be leaderless (Shell et al., 2015); in fact, eis gene of M. tuberculosis (see the section “The Eis Protein Becomes a Novel Aminoglycoside Acetyltransferase,” below) is transcribed also as a leaderless mRNA (Zaunbrecher et al., 2009). In leaderless mRNAs, there is no Shine-Dalgarno sequence, and 70S ribosomes bind directly to the 50-end of the mRNA in order to initiate translation (Shell et al., 2015). New Roles for AAC(20) in Mycobacteria Two major evidences supported that in mycobacteria, AAC(20)- I enzymes could have additional roles other than acetylation of AGs containing 20-amino group: first, the capability of amino sugars and diverse acyl-CoA molecules to inhibit the acetylating activity of mycobacterial crude extracts (Udou et al., 1989); and second, the implication of AAC(20)-Ia (an enzyme of the same class) in acetylation of cell wall substrates in P. stuartii. In a series of laboratory mutants of P. stuartii expressing aac(20)-Ia gene at different levels, it was found that the extent of peptidoglycan acetylation correlated with the activity of AAC(20)-Ia, hence suggesting a partial contribution of this enzyme (along with other enzymes) to peptidoglycan acetylation. Many bacterial pathogens acetylate their peptidoglycan as a way to resist the action of muramidase enzymes. Under in vitro conditions, AAC(20)-Ia was able to acetylate tobramycin having acetylated peptidoglycan as donor of acetyl groups (Payie et al., 1995, 1996;Payie and Clarke, 1997). We investigated the hypothesis of mycobacterial AAC(20)-I enzymes having also a role in cell wall metabolism, and a gene knock-out mutant of M. smegmatis deleted in aac(20)-Id gene was constructed. This mutant (named EP-10) defective in AAC(20)- I activity was more susceptible to gentamicin, tobramycin, dibekacin, and netilmicin than the parental wild-type strain, and crude extracts of M. smegmatis EP-10 failed to acetylate 20-amino group containing AGs, whereas wild-type strains of M. smegmatis readily acetylated such AGs (Ainsa et al., 1997;Maurer et al., 2015). M. smegmatis EP-10 was also twofold more susceptible to lysozyme, a feature that has been associated with the extent of peptidoglycan acetylation (Ainsa et al., 1997). Hence, we concluded that in M. smegmatis, AAC(20)-Id enzyme could also contribute to acetylation of peptidoglycan, since a less extensively acetylated peptidoglycan in the knock-out strain would be more susceptible to lysozyme degradation affecting cell viability. Biochemical Analysis of Mycobacterial AAC(20)-I Enzymes The presence of an AG (20)-N-acetyltransferase gene in the genome of M. tuberculosis was intriguing. In this species, AAC activity had never been reported (Mitsuhashi et al., 1977) and the expression of the aac(20)-Ic gene [annotated as Rv1258c gene in the M. tuberculosis H37Rv genome (Cole et al., 1998)] in the surrogate host M. smegmatis could not be associated with any change in the levels of susceptibility to AGs (Ainsa et al., 1997). We constructed a knock-out mutant of M. tuberculosis H37Rv deleted in the aac(20)-Ic gene and observed that the mutant (named M. tuberculosis B1) was twofold more susceptible than the original wild-type strain to AGs containing a 20-amino group such as gentamicin, tobramycin and dibekacin, and fourfold more susceptible to 60-N-ethyl-netilmicin. This indicated that the aac(20)-Ic gene was being expressed in M. tuberculosis although at a very low level, and that the AAC(20)-Ic enzyme in M. tuberculosis would acetylate all these four AGs in the wild type strain; hence, acetylated AGs would bind less efficiently to the ribosome, and the AAC(20)-Ic enzyme would contribute to basal AG resistance in this species. In order to find the physiological role of this enzyme in M. tuberculosis, recombinant E. coli-produced AAC(20)-Ic enzyme from M. tuberculosis was studied and found to efficiently acetylate AG antibiotics containing an amino group at the 20 position (as expected; for example, Kmfor kanamycin B was 1.4 µM), and surprisingly, it was also capable of acetylating (although to a much lesser extent) other AGs such as kanamycin A (Km320 µM) and amikacin (Km968 µM) that have a hydroxyl group in the 20position, hence suggesting this enzyme to be capable of both Nand O-acetylation (Hegde et al., 2001;Draker et al., 2003). This residual activity of AAC(20)- Ic against kanamycin A and amikacin does not affect their bactericidal activity, and these two antibiotics are used as second line drugs against drug resistant M. tuberculosis (World Health Organization [WHO], 2010). The activity of M. tuberculosis AAC(20)-Ic is dependent on metal ions, being inhibited by Cu2+ and Au3+(Li et al., 2015). Another study consisted in binding covalently the AGs kanamycin A, tobramycin, neamine and neomycin B to an agarose matrix in order to quantify the extent of AG acetylation by AAC enzymes and their subsequent ability to bind an artificial probe mimicking the A-site of the ribosome (Barrett et al., 2008). In such experimental model system, the AGs acetylated by M. tuberculosis AAC(20)-Ic enzyme (only tobramycin, neamine and neomycin B) maintained their binding affinity with the probe mimicking A-site of the ribosome at detectable levels, maybe because the percent of AG acetylation by AAC(20)-Ic was low. In contrast, these AGs were very efficiently acetylated by E. coli AAC(3) (at a different amino group) and had readily lost their affinity for binding this artificial probe. These experiments suggested that subtle differences in the structure of modified AGs (i.e., acetylation at the amino group in 20or 3 position) are sufficient to drastically affect their capability of binding to the A-site probe, and this would be expected to correlate with their activity as ribosome inhibitors. In consequence, in mycobacteria, AAC(20)-I enzymes would not play a major role in resistance to these drugs (Barrett et al., 2008). High resolution crystal structures of AAC(20)-Ic complexed with AGs demonstrated that this enzyme is a member of the GNC5 acetyltransferase superfamily and suggested a role in the synthesis of mycothiol, Frontiers in Microbiology | www.frontiersin.org 4January 2019 | Volume 10 | Article 46 fmicb-10-00046 January 28, 2019 Time: 18:36 # 5 Sanz-García et al. Mycobacterial Aminoglycoside Acetyltransferases a metabolite that has a key role in regulating redox potential in mycobacteria (Vetting et al., 2002, 2003). In agreement with the preferential N-acetylating activity over the O-acetylating activity that was found in the M. tuberculosis enzyme (Hegde et al., 2001;Draker et al., 2003), M. abscessus clinical isolates were found to be more susceptible to AGs containing a hydroxyl group at the 20position (such as amikacin and kanamycin A) than to AGs with a 20-amino group (such as tobramycin, dibekacin and kanamycin B), as the latter group would be substrates of M. abscessus AAC(20)-I enzyme (Maurer et al., 2015). In fact, crude extracts of M. abscessus efficiently acetylated kanamycin B, whereas kanamycin A was not acetylated at detectable levels. In this species, deletion of the aac(20)-I gene resulted in increased susceptibility to kanamycin B, tobramycin, dibekacin and gentamicin C (all of them containing a 20-amino group) (Rominski et al., 2017). These two reports demonstrate that in M. abscessus, the presence of an AAC(20)-I enzyme contributes to decreased innate susceptibility to AGs containing a 20-amino group (Luthra et al., 2018). Becoming a Drug Target: Developing Inhibitors of AAC(20)-I The interest in developing inhibitors against AAC(20)-I enzymes came from a study in the non-tuberculous species M. abscessus (Maurer et al., 2014). It was found that AGs such as amikacin, gentamicin or tobramycin, which are normally bactericidal against E. coli, do not have such activity against M. abscessus or M. smegmatis. However, disruption of the chromosomally encoded aac(20)-I gene in these species restored the bactericidal activity of these AGs (Maurer et al., 2014). Given that AGs are used as second-line drugs in treatment of multidrug resistant (MDR) tuberculosis infections, and also in the treatment of other infections caused by non-tuberculous mycobacteria, the possibility of developing compounds that could enhance bactericidal activity of current antimycobacterial treatments became an interesting approach. To date, the only putative inhibitor of M. tuberculosis AAC(20)-Ic is andrographolide, a natural product that was identified in methanolic extracts of a plant that were capable of inhibiting growth of M. tuberculosis strains (Prabu et al., 2015). In silico analysis predicted that this compound could potentially bind with a high affinity the AAC(20)-Ic enzyme, as well as isocitrate lyase (a metabolic enzyme of the glyoxylate shunt, involved in persistence and virulence of M. tuberculosis) and other M. tuberculosis proteins (Prabu et al., 2015). However, the specificity of this binding and the ability to really inhibit such putative target proteins were not tested. Given that the gene encoding AAC(20)-Ic is not essential in M. tuberculosis, a direct link between AAC(20)-Ic inhibition and bacterial growth inhibition could be discarded. Hence, the ability of plant extracts containing andrographolide to inhibit growth of M. tuberculosis could be due to the presence of additional compounds in the extract, or to multiple effects on M. tuberculosis cells. Other in silico analysis revealed that AAC(20)-Ic enzyme from M. tuberculosis could interact with ten other proteins (including a protein of a putative RND-like efflux pump), suggesting that inhibition of AAC(20)-Ic could also impact many other metabolic processes, hence conferring this enzyme with a relevant role in drug discovery of antituberculosis agents (Joshi et al., 2013). THE Eis PROTEIN BECOMES A NOVEL AMINOGLYCOSIDE ACETYLTRANSFERASE Investigation of M. tuberculosis virulence factors lead to the identification of a protein, that was required for infecting and survival in human macrophages; this protein was named Eis for enhanced intracellular survival (Box 2) (Wei et al., 2000). Bioinformatic analysis revealed that Eis protein of M. tuberculosis was an acetyltransferase of the GCN-5 family (Samuel et al., 2007). Later on, the analysis of kanamycin resistant M. tuberculosis laboratory and clinical strains revealed mutations in the −10 and −35 regions of the eis gene promoter, which resulted in increased levels of eis mRNA and Eis protein. These mutations were related to low-level resistance to kanamycin (MIC 25 µg/ml), but not to amikacin (MIC <4µg/ml), whereas 16S rRNA mutations confer higher levels of resistance to kanamycin (MIC >80 µg/ml) and frequently cross-resistance to amikacin. These eis mutants also displayed increased levels of AAC activity, hence demonstrating that Eis was a novel class of AAC, highly divergent from all other previously known AACs. Eis is capable of acetylating kanamycin more efficiently than amikacin, and streptomycin was not found to be a substrate of Eis (Zaunbrecher et al., 2009). From then on, detection of mutations in eis promoter has become a relevant assay in clinical microbiology laboratories for determining susceptibility to kanamycin (Georghiou et al., 2012); the diagnostic and clinical implications of these tests are beyond the scope of this review. Formation of stable hexamers by Eis is required for its AAC activity (Ganaie et al., 2011;Anand and Sharma, 2018), which can acetylate multiple amine groups of different AG antibiotics, including netilmicin, sisomicin, neamine, ribostamycin, paromomycin, neomycin B, kanamycin, amikacin, tobramycin and hygromycin, resulting in mono-, di-, tri, and tetraacetylated products (Chen et al., 2011;Houghton et al., 2013a), being able to use not only acetyl-CoA but also other acyl-CoA derivatives (Chen et al., 2012a). The Eis protein works by a randomsequential bisubstrate mechanism of acetylation (Tsodikov et al., 2014). Interestingly, Eis is also able to acetylate capreomycin, a polypeptide second-line antituberculosis drug commonly used in the treatment of MDR tuberculosis infections (Reeves et al., 2013). Several metal ions such as Au3+, Cd2+and Zn2+ inhibited Eis activity in vitro (Li et al., 2015). Other mycobacterial species such as M. smegmatis and M. abscessus have ortholog (and even paralog) eis genes, although the Eis proteins have distinct biochemical features and impact on AG susceptibility in comparison with Eis of M. tuberculosis (Chen et al., 2012b;Rominski et al., 2017). For example, M. abscessus has two eis genes, and the deletion of one (but not the other one) resulted in altered AG susceptibility (Rominski et al., 2017; Luthra et al., 2018). Consistently with these findings, mutational Frontiers in Microbiology | www.frontiersin.org 5January 2019 | Volume 10 | Article 46 fmicb-10-00046 January 28, 2019 Time: 18:36 # 6 Sanz-García et al. Mycobacterial Aminoglycoside Acetyltransferases BOX 2 | Discovery and characterization of enhanced intracellular survival (Eis) protein of M. tuberculosis. A genome wide investigation of the ability of M. tuberculosis for infecting macrophages resulted in the identification of a coding sequence that, once cloned in the non-pathogenic M. smegmatis species, conferred capacity for infecting the human macrophage-like cell line U937. This gene was named eis (Rv2416c in the M. tuberculosis genome; Cole et al., 1998) for enhanced intracellular survival (Wei et al., 2000), and was detected only in pathogenic species of mycobacteria. The promoter of the eis gene is similar to consensus E. coli sigma-70 dependent promoters (Roberts et al., 2004) and it is recognized by M. tuberculosis SigA sigma factor (Wu et al., 2009). The Eis protein was found to be mostly hydrophilic, but having a hydrophobic N-terminal end, so that it could be found in the cytosolic but also in other cell fractions such as the membrane, cell wall or among the secreted proteins of M. tuberculosis (Dahl et al., 2001). In fact, antibodies against Eis could detect this protein in the sera of tuberculosis patients (Dahl et al., 2001) and in the cytoplasm of infected macrophages (Samuel et al., 2007). Also, it was found that Eis protein, directly added to cultures of human monocytes, modulated the secretion of pro-inflammatory cytokines in a similar way to that found in M. tuberculosis infected cells (Samuel et al., 2007), hence suggesting a role of Eis as an effector protein. Further studies demonstrated that Eis inhibits the extra-cellular signal-regulated kinase 1/2 (ERK1/2) and JAK pathways, and in consequence it inhibited the production of TNF-alpha and IL-4, and stimulated the production of IFN-gamma and IL-10 (Lella and Sharma, 2007). The effect of Eis in increasing production of IL-10 was found to be related to Eis-mediated acetylation of histone H3, which binds the promoter of the human IL-10 gene (Duan et al., 2016). Hence, by disturbing cross-regulation of T-cells and impairing TH1 and TH2 response, Eis could mediate M. tuberculosis pathogenicity (Lella and Sharma, 2007). In fact, an isolate of the Beijing family (which are more transmissible and virulent than other M. tuberculosis genetic lineages (Hanekom et al., 2011) was found to contain elevated levels of Eis protein, mediated by increased expression of SigA (Wu et al., 2009). Other key factors in host immune response to tuberculosis are also mediated by Eis, which increased production of ROS and consequently modulated processes such as autophagy, inflammation, and cell death (Shin et al., 2010). These processes are started by Eis-dependent acetylation of dual-specificity phosphatase-16 (DUSP16)-mitogen-activated protein kinase phosphatase-7 (MKP-7), which dephosphorylates the JNK protein leading to its inactivation (Kim et al., 2012;Yoon et al., 2013). Other studies have revealed the activity of Eis for acetylating arylalkylamines such as histamine, octopamine, or tyramine, suggesting novel roles for this protein in M. tuberculosis pathogenicity (Pan et al., 2018). changes in the amino acid residues lining the substrate binding site of M. tuberculosis Eis altered its substrate specificity (Jennings et al., 2013). It is important to note that in M. tuberculosis, transcription of the eis gene is activated by the regulator WhiB7 (Reeves et al., 2013). Mutations in the promoter of whiB7 gene that led to increase in the mRNA of this gene resulted in increased expression of eis gene, along with other genes such as rv1258c (encoding the Tap efflux pump; Ainsa et al., 1998), hence resulting in cross resistance to several drugs including kanamycin (mediated by Eis protein) or streptomycin (mediated by Tap efflux pump). Similarly, in M. abscessus, WhiB7 controlled the expression of one of the two eis genes in this species and also that of the erm(41) gene, which encodes a ribosomal methyltransferase that by altering target structure is associated with resistance to macrolide antibiotics (Pryjma et al., 2017;Luthra et al., 2018). Subinhibitory concentrations of clarithromycin induced the whiB7 gene and consequently decreased Eis-mediated susceptibility to AGs, such as amikacin that is currently used in the treatment of M. abscessus infections (Pryjma et al., 2017). Aminoglycosides and Beyond. . . The unusual properties of Eis acetyltransferase include its capability for acetylating peptides and proteins (Kim et al., 2012; Houghton et al., 2013b;Yoon et al., 2013), in contrast with other AACs. The M. tuberculosis nucleoid-associated protein HU (encoded by Rv2986c gene) can be acetylated by Eis on multiple lysine residues, hence decreasing its ability to interact with DNA, and altering its DNA compactation activity (Ghosh et al., 2016;Green et al., 2018). Overexpression of Eis led to a hyperacetylation of HU protein, and consequently, to a decompactation of the genome (Ghosh et al., 2016). The reverse effect (condensation of relaxed DNA) could be reached through the deacetylation of HU protein, which is mediated by a Sir2 family protein from M. tuberculosis encoded by the Rv1155c gene (Anand et al., 2017;Green et al., 2018). Controlling the architecture of DNA is a key process in any bacteria, and so, the HU protein is essential for M. tuberculosis. In fact, inhibitors of HU have been discovered (Bhowmick et al., 2014), which could act in synergy with potential inhibitors of Eis, as described in the next section. Finding Inhibitors of Eis Protein The crystal structure of M. tuberculosis Eis protein was determined by several groups (Chen et al., 2011;Kim et al., 2012), which has been useful for determining docking properties of potential inhibitory compounds; also, its comparison with the crystal structure of M. smegmatis Eis protein revealed several distinct structural features that may account for the biochemical and substrate differences between the two proteins (Kim et al., 2014). A first screening of potential inhibitors of Eis from M. tuberculosis resulted in the identification of 25 molecules (including the antiseptic chlorhexidine) that inhibited Eis with IC50 values in the low micromolar range. In addition, this inhibition was specific to the M. tuberculosis Eis protein, since these molecules could not inhibit significantly AACs from AAC(20), AAC(3), and AAC(60) families (Green et al., 2012), nor Eis protein from Bacillus anthracis (Green et al., 2015a). Later studies revealed the presence of eis-like genes in many pathogenic and non-pathogenic bacteria (remarkably, many mycobacterial species have two or even three paralogs of the eis gene), and chlorhexidine was capable of inhibiting (to different levels) all Eis proteins that were capable of acetylating AGs (Green et al., 2015b). A second screening of a larger collection of small-molecule compounds resulted in the identification of several families of compounds capable of inhibiting Eis activity. These contained diverse chemical scaffolds (Garzan et al., 2016a,b, 2017;Willby et al., 2016;Ngo et al., 2018). Besides, these inhibitors were highly selective for Eis, and did not inhibited AACs from other families (Garzan et al., 2016b). More importantly, as these inhibitors bound in the AG pocket of the Eis protein, they were able to reverse kanamycin resistance of a M. tuberculosis isolate (Garzan et al., 2016b;Willby et al., 2016;Ngo et al., 2018). Frontiers in Microbiology | www.frontiersin.org 6January 2019 | Volume 10 | Article 46 fmicb-10-00046 January 28, 2019 Time: 18:36 # 7 Sanz-García et al. Mycobacterial Aminoglycoside Acetyltransferases Some of these inhibitors, such as those based on a pyrrolo[1,5a]pyrazine scaffold, also lacked any toxicity on mammalian cell lines (Garzan et al., 2017). OTHER AMINOGLYCOSIDE-MODIFYING ENZYMES IN MYCOBACTERIA Genome-wide analysis of M. tuberculosis genome identified only one other potential AAC (encoded by the Rv1347c gene), although such enzymatic activity could not be detected on the recombinant protein (Draker et al., 2003). Later studies related the product of the Rv1347c gene with a role in the synthesis of mycobactin, the mycobacterial siderophore (Card et al., 2005). Leaving apart AACs, only a few reports of other classes of aminoglycoside-modifying enzymes have been done in mycobacteria. An APH enzyme of the APH(300) family, conferring resistance to the AG streptomycin only, has been characterized in M. fortuitum (Ramon-Garcia et al., 2006) and M. abscessus (Dal Molin et al., 2018;Luthra et al., 2018); the latter species encodes up to 11 additional putative APH enzymes (Ripoll et al., 2009). Furthermore, a putative APH of the APH(30) class, encoded by the Rv3168 gene of M. tuberculosis, was identified and expressed as a recombinant enzyme in E. coli, being related to kanamycin phosphotransferase activity (Ahn and Kim, 2013). BACK TO THE START POINT: Eis IN M. fortuitum We started this review referring to previous work that has shown that crude extracts of M. fortuitum harbored AAC activity having gentamicin, tobramycin, netilmicin and its derivatives 20-Nethyland 60-N-ethyl-netilmicin, kanamycin A and kanamycin B as substrates. So far, the only AG acetyltransferase identified in this species has been AAC(20)-Ib (Ainsa et al., 1996), which cannot explain the acetyltransferase activity against kanamycin A and 20-N-ethyl-netilmicin detected in M. fortuitum crude extracts. Later studies characterized Eis AAC in diverse mycobacterial species, but no report were been done on a putative Eis protein in M. fortuitum. In view of the universal presence of Eis proteins in mycobacteria, and its activity as AAC, we hypothesized that M. fortuitum could also have a putative Eis protein that would be responsible for the acetyltransferase activity against kanamycin A and 20-N-ethyl-netilmicin detected in crude extracts of this species, as it was reported earlier (Hull et al., 1984; Ainsa et al., 1996). Thus, we first ascertained the existence of an eis gene in the genome of M. fortuitum (Ho et al., 2012), which presented a 94% of identity with respect to the one reported in M. tuberculosis; the sequence of both Eis proteins from M. tuberculosis and M. fortuitum also presented high levels of identity (Figure 1). We designed two oligonucleotides for amplifying specifically a DNA fragment from M. fortuitum genome containing eis gene. This DNA fragment was subsequently cloned in pMV261 vector (Stover et al., 1991), which expresses genes constitutively under the control of the hsp60 gene promoter, resulting in plasmid pFS2. Given that this plasmid still contains the Tn903-derived aminoglycoside-30-phosphotransferase (aph) gene (present in the original pMV261 cloning vector) conferring kanamycin resistance as selection marker, we anticipated that M. smegmatis strains harboring pMV261 vector or its derivatives would be intrinsically resistant to kanamycin A; this would prevent from determining whether this antibiotic is a substrate of the M. fortuitum Eis protein. Additionally, the selection marker could have cross-effect with other AGs and thus interfering with the resistance phenotype conferred by eis gene. To circumvent these problems, we generated a derivative of pFS2 through the disruption of the aph gene with the ampicillin resistance cassette (bla gene) from pGEMR -T easy (Promega). The resistance to 20-N-ethylnetilmicin conferred by eis gene, as demonstrated for the parental plasmid pFS2 (see Table 1) was used as resistance marker for transformant selection and plasmid maintenance. This process resulted in plasmid pEAC which contains the M. fortuitum eis gene and no other determinant of AG resistance. The three plasmids [original empty vector pMV261 as a control, and the two plasmids (pFS2 and pEAC) containing M. fortuitum eis gene] were introduced in M. smegmatis mc2155 in order to over-express the ortholog eis gene from M. fortuitum and to elucidate its hypothetical implication in AG susceptibility. The antibiotic susceptibility assay was made, based on a double dilution protocol with the addition of resazurin dye (Palomino et al., 2002). We observed that plasmids pFS2 and pEAC produced detectable changes in AG susceptibility of M. smegmatis mc2155, which can be attributed to the expression of the plasmid-borne M. fortuitum eis gene. The major shift in the MICs was detected for 20-N-ethylnetilmicin (Table 1), since the MIC increased from 3.12 to 6.25 µg/ml in the control strains to 50–100 µg/ml in the strains containing M. fortuitum Eis, accounting for a 8to 32-fold increase; similar changes were observed for 60-N-ethylnetilmicin (8-fold increase in the MIC). A moderate decrease in the susceptibility to kanamycin A (fivefold), hygromycin (twofold to fourfold) and gentamicin (twofold to fourfold) was also observed. Finally, slight changes (twofold) in the MICs were detected for kanamycin B and capreomycin; this finding is consistent with previous reports on the activity of M. tuberculosis Eis against capreomycin (Reeves et al., 2013). The levels of susceptibility to the AGs amikacin, streptomycin and spectinomycin, and to other non-AG compounds tested (isoniazid, rifampicin, ethambutol, ciprofloxacin, tetracycline, chloramphenicol) were not altered significantly by the presence of the plasmid-encoded eis gene (Table 1 and data not shown), suggesting that either these antimicrobials are not substrates of the Eis enzyme or the corresponding acetylations (if any) might just not affect antibacterial activity. CONCLUDING REMARKS The presence and activity of AAC(20)-I and Eis AACs in mycobacteria have clearly demonstrated that their primary role is little related with susceptibility to AGs. Frontiers in Microbiology | www.frontiersin.org 7January 2019 | Volume 10 | Article 46 fmicb-10-00046 January 28, 2019 Time: 18:36 # 8 Sanz-García et al. Mycobacterial Aminoglycoside Acetyltransferases FIGURE 1 | Comparison of the amino acid sequences of Eis proteins from M. fortuitum and M. tuberculosis. Symbols under the sequence alignment: Asterisks (∗) indicate positions with identical amino acid in both proteins. Colons (:) indicate positions which have amino acids with strongly similar properties. Periods (.) indicate positions which have amino acids with weakly similar properties. TABLE 1 | MICs (µg/ml) of antibiotics of different structural families in the M. smegmatis mc2155 strains that overexpress eis gene from M. fortuitum. Strain M. smegmatis mc2155 Plasmid None pMV261 pFS2 pEAC Marker gene on plasmid None aph aph bla M. fortuitum eis gene − − + + Gentamicin 0.78-1.56 0.78 3.12 3.12 20-N-ethyl netilmicin 6.25 3.12–6.25 50–100 50 60-N-ethyl netilmicin 3.12 3.12 3.12–6.25 25 Kanamycin A 1.56 >100 >100 7.8 Kanamycin B 6.25 >100 >100 12.5 Hygromycin 15.6 15.6 31.2–62.5 62.5 Amikacin 0.39 0.39 0.39 0.39 Capreomycin 1.95 1.95 3.9 3.9 Streptomycin 0.25 0.25 0.25 0.25 Spectinomycin 62.5 31.2–62.5 62.5 62.5 The values separated by a dash indicate that growth was detected in both concentrations. The range of antibiotic concentrations spanned from 0.78 µg/ml to 100 µg/ml, and from 0.25 µg/ml to 125 µg/ml; a twofold difference in the MIC was not considered as significant. aph: aminoglycoside 30-phosphotransferase from Tn903; bla: beta-lactamase. In the case of AAC(20)-I enzymes, its presence in phylogenetically distant genera as Providencia and Mycobacterium remains to be an evolutionary mystery. In contrast with this restricted distribution of AAC(20)-I enzymes among bacteria, Eis enzymes seem to be more widely distributed, being present even in non-pathogenic and environmental species, which suggest a general function of Eis-like enzymes in bacterial metabolism, and virtually excludes any potential selection of eis genes due to the use of AGs, or its horizontal transfer from other species. It is clear that in mycobacteria ribosomal modifications constitute the major mechanism of AG resistance, given that only one or two copies of ribosomal RNA operons are present in these species, hence making likely the acquisition of mutations conferring high levels of AG resistance. Then, AAC(20)-I enzymes only contribute modestly to innate low level susceptibility to AGs, and despite other roles have been suggested in the literature for mycobacterial AAC(20)-I enzymes, their relevance as potential drug targets is still modest, especially in comparison with Eis acetyltransferase. The contribution of Eis acetyltransferase to virulence of M. tuberculosis, and the finding that Eis is related with resistance to kanamycin (a second line drug for the treatment of tuberculosis) in clinical isolates has greatly attracted the attention and promoted the interest in developing Eis inhibitors. In some way, Eis inhibitors would fall into the class of anti-virulence and antiresistance mechanisms compounds, which is a trending topic Frontiers in Microbiology | www.frontiersin.org 8January 2019 | Volume 10 | Article 46 fmicb-10-00046 January 28, 2019 Time: 18:36 # 9 Sanz-García et al. Mycobacterial Aminoglycoside Acetyltransferases in the age of antimicrobial resistance. Globally, antimicrobial resistance is a major public health threat, and multi and extensively drug resistant (MDR, XDR) tuberculosis is a case of particular concern, so progress and major advances that can be expected from the coming years will be greatly welcomed. AUTHOR CONTRIBUTIONS JA, LR, and CM contributed to the conception and design of the study. FS-G, EA-C, AL, LR, and EP-H carried out the experimental work. FS-G, EA-C, and JA wrote the manuscript. All authors contributed to manuscript revision, read and approved the submitted version. FUNDING FS-G and EA-C are recipients of FPU fellowships from the Spanish Ministry of Economy and Competitivity. JA acknowledges funding from the European Comission [More Medicines for Tuberculosis (MM4TB) grant 260872] and the Spanish Ministry of Economy and Competitivity (grants SAF-2013-48971-C2-2-R and SAF2017-84839C2-1-R). ACKNOWLEDGMENTS Dessi Marinova was acknowledged for critical reading of the manuscript. REFERENCES Ahn, J. W., and Kim, K. J. (2013). Rv3168 phosphotransferase activity mediates kanamycin resistance in Mycobacterium tuberculosis.J. Microbiol. Biotechnol. 23, 1529–1535. doi: 10.4014/jmb.1306.06048 Ainsa, J. A., Blokpoel, M. C., Otal, I., Young, D. B., De Smet, K. A., and Martin, C. (1998). Molecular cloning and characterization of Tap, a putative multidrug efflux pump present in Mycobacterium fortuitum and Mycobacterium tuberculosis.J. Bacteriol. 180, 5836–5843. Ainsa, J. A., Martin, C., Gicquel, B., and Gomez-Lus, R. (1996). Characterization of the chromosomal aminoglycoside 20-N-acetyltransferase gene from Mycobacterium fortuitum.Antimicrob. Agents Chemother. 40, 2350–2355. doi: 10.1128/AAC.40.10.2350 Ainsa, J. A., Perez, E., Pelicic, V., Berthet, F. X., Gicquel, B., and Martin, C. (1997). Aminoglycoside 20-N-acetyltransferase genes are universally present in mycobacteria: characterization of the aac(20)-Ic gene from Mycobacterium tuberculosis and the aac(20)-Id gene from Mycobacterium smegmatis.Mol. Microbiol. 24, 431–441. doi: 10.1046/j.1365-2958.1997.3471717.x Anand, C., Garg, R., Ghosh, S., and Nagaraja, V. (2017). A Sir2 family protein Rv1151c deacetylates HU to alter its DNA binding mode in Mycobacterium tuberculosis.Biochem. Biophys. Res. Commun. 493, 1204–1209. doi: 10.1016/j. bbrc.2017.09.087 Anand, S., and Sharma, C. (2018). Glycine-rich loop encompassing active site at interface of hexameric M. tuberculosis Eis protein contributes to its structural stability and activity. Int. J. Biol. Macromol. 109, 124–135. doi: 10.1016/j. ijbiomac.2017.12.058 Angelatou, F., Litsas, S. B., and Kontomichalou, P. (1982). Purification and properties of two gentamicin-modifying enzymes, coded by a single plasmid pPK237 originating from Pseudomonas aeruginosa.J. Antibiot. 35, 235–244. doi: 10.7164/antibiotics.35.235 Barrett, O. J., Pushechnikov, A., Wu, M., and Disney, M. D. (2008). Studying aminoglycoside modification by the acetyltransferase class of resistance-causing enzymes via microarray. Carbohydr. Res. 343, 2924–2931. doi: 10.1016/j.carres. 2008.08.018 Bassetti, M., Vena, A., Croxatto, A., Righi, E., and Guery, B. (2018). How to manage Pseudomonas aeruginosa infections. Drugs Context 7:212527. doi: 10.7573/dic. 212527 Bhowmick, T., Ghosh, S., Dixit, K., Ganesan, V., Ramagopal, U. A., Dey, D., et al. (2014). Targeting Mycobacterium tuberculosis nucleoid-associated protein HU with structure-based inhibitors. Nat. Commun. 5:4124. doi: 10.1038/ ncomms5124 Card, G. L., Peterson, N. A., Smith, C. A., Rupp, B., Schick, B. M., and Baker, E. N. (2005). The crystal structure of Rv1347c, a putative antibiotic resistance protein from Mycobacterium tuberculosis, reveals a GCN5-related fold and suggests an alternative function in siderophore biosynthesis. J. Biol. Chem. 280, 13978–13986. doi: 10.1074/jbc.M413904200 Chen, W., Biswas, T., Porter, V. R., Tsodikov, O. V., and Garneau-Tsodikova, S. (2011). Unusual regioversatility of acetyltransferase Eis, a cause of drug resistance in XDR-TB. Proc. Natl. Acad. Sci. U.S.A. 108, 9804–9808. doi: 10. 1073/pnas.1105379108 Chen, W., Green, K. D., and Garneau-Tsodikova, S. (2012a). Cosubstrate tolerance of the aminoglycoside resistance enzyme Eis from Mycobacterium tuberculosis.Antimicrob. Agents Chemother. 56, 5831–5838. doi: 10.1128/AAC. 00932-12 Chen, W., Green, K. D., Tsodikov, O. V., and Garneau-Tsodikova, S. (2012b). Aminoglycoside multiacetylating activity of the enhanced intracellular survival protein from Mycobacterium smegmatis and its inhibition. Biochemistry 51, 4959–4967. doi: 10.1021/bi3004473 Clarke, A. J., Francis, D., and Keenleyside, W. J. (1996). The prevalence of gentamicin 20-N-acetyltransferase in the Proteeae and its role in the O-acetylation of peptidoglycan. FEMS Microbiol. Lett. 145, 201–207. Cole, S. T., Brosch, R., Parkhill, J., Garnier, T., Churcher, C., Harris, D., et al. (1998). Deciphering the biology of Mycobacterium tuberculosis from the complete genome sequence. Nature 393, 537–544. doi: 10.1038/31159 Dahl, J. L., Wei, J., Moulder, J. W., Laal, S., and Friedman, R. L. (2001). Subcellular localization of the Iitracellular survival-enhancing Eis protein of Mycobacterium tuberculosis.Infect. Immun. 69, 4295–4302. doi: 10.1128/IAI. 69.7.4295-4302.2001 Dal Molin, M., Gut, M., Rominski, A., Haldimann, K., Becker, K., and Sander, P. (2018). Molecular mechanisms of intrinsic streptomycin resistance in Mycobacterium abscessus.Antimicrob. Agents Chemother. 62:e001427-17. doi: 10.1128/AAC.01427-17 Davies, J., and Wright, G. D. (1997). Bacterial resistance to aminoglycoside antibiotics. Trends Microbiol. 5, 234–240. doi: 10.1016/S0966-842X(97)01033-0 Draker, K. A., Boehr, D. D., Elowe, N. H., Noga, T. J., and Wright, G. D. (2003). Functional annotation of putative aminoglycoside antibiotic modifying proteins in Mycobacterium tuberculosis H37Rv. J. Antibiot. 56, 135–142. doi: 10.7164/antibiotics.56.135 Duan, L., Yi, M., Chen, J., Li, S., and Chen, W. (2016). Mycobacterium tuberculosis EIS gene inhibits macrophage autophagy through up-regulation of IL-10 by increasing the acetylation of histone H3. Biochem. Biophys. Res. Commun. 473, 1229–1234. doi: 10.1016/j.bbrc.2016.04.045 Franklin, K., and Clarke, A. J. (2001). Overexpression and characterization of the chromosomal aminoglycoside 20-N-acetyltransferase of Providencia stuartii. Antimicrob. Agents Chemother. 45, 2238–2244. doi: 10.1128/AAC.45.8.22382244.2001 Ganaie, A. A., Lella, R. K., Solanki, R., and Sharma, C. (2011). Thermostable hexameric form of Eis (Rv2416c) protein of M. tuberculosis plays an important role for enhanced intracellular survival within macrophages. PLoS One 6:e27590. doi: 10.1371/journal.pone.0027590 Garzan, A., Willby, M. J., Green, K. D., Gajadeera, C. S., Hou, C., Tsodikov, O. V., et al. (2016a). Sulfonamide-based inhibitors of aminoglycoside acetyltransferase Eis abolish resistance to kanamycin in Mycobacterium tuberculosis.J. Med. Chem. 59, 10619–10628. doi: 10.1021/acs.jmedchem.6b01161 Garzan, A., Willby, M. J., Green, K. D., Tsodikov, O. V., Posey, J. E., and GarneauTsodikova, S. (2016b). Discovery and optimization of two Eis inhibitor families Frontiers in Microbiology | www.frontiersin.org 9January 2019 | Volume 10 | Article 46 fmicb-10-00046 January 28, 2019 Time: 18:36 # 10 Sanz-García et al. Mycobacterial Aminoglycoside Acetyltransferases as kanamycin adjuvants against drug-resistant M. tuberculosis.ACS Med. Chem. Lett. 7, 1219–1221. doi: 10.1021/acsmedchemlett.6b00261 Garzan, A., Willby, M. J., Ngo, H. X., Gajadeera, C. S., Green, K. D., Holbrook, S. Y., et al. (2017). Combating enhanced intracellular survival (Eis)-mediated kanamycin resistance of Mycobacterium tuberculosis by novel pyrrolo[1,5a]pyrazine-based Eis inhibitors. ACS Infect. Dis. 3, 302–309. doi: 10.1021/ acsinfecdis.6b00193 Georghiou, S. B., Magana, M., Garfein, R. S., Catanzaro, D. G., Catanzaro, A., and Rodwell, T. C. (2012). Evaluation of genetic mutations associated with Mycobacterium tuberculosis resistance to amikacin, kanamycin and capreomycin: a systematic review. PLoS One 7:e33275. doi: 10.1371/journal. pone.0033275 Ghosh, S., Padmanabhan, B., Anand, C., and Nagaraja, V. (2016). Lysine acetylation of the Mycobacterium tuberculosis HU protein modulates its DNA binding and genome organization. Mol. Microbiol. 100, 577–588. doi: 10.1111/mmi. 13339 Green, K. D., Biswas, T., Chang, C., Wu, R., Chen, W., Janes, B. K., et al. (2015a). Biochemical and structural analysis of an Eis family aminoglycoside acetyltransferase from bacillus anthracis. Biochemistry 54, 3197–3206. doi: 10. 1021/acs.biochem.5b00244 Green, K. D., Pricer, R. E., Stewart, M. N., and Garneau-Tsodikova, S. (2015b). Comparative study of Eis-like enzymes from pathogenic and nonpathogenic bacteria. ACS Infect. Dis. 1, 272–283. doi: 10.1021/acsinfecdis.5b00036 Green, K. D., Biswas, T., Pang, A. H., Willby, M. J., Reed, M. S., Stuchlik, O., et al. (2018). Acetylation by Eis and deacetylation by Rv1151c of Mycobacterium tuberculosis HupB: biochemical and structural insight. Biochemistry 57,781–790. doi: 10.1021/acs.biochem.7b01089 Green, K. D., Chen, W., and Garneau-Tsodikova, S. (2012). Identification and characterization of inhibitors of the aminoglycoside resistance acetyltransferase Eis from Mycobacterium tuberculosis.ChemMedChem 7, 73–77. doi: 10.1002/ cmdc.201100332 Hanekom, M., Gey van Pittius, N. C., McEvoy, C., Victor, T. C., Van Helden, P. D., and Warren, R. M. (2011). Mycobacterium tuberculosis Beijing genotype: a template for success. Tuberculosis 91, 510–523. doi: 10.1016/j.tube.2011. 07.005 Hegde, S. S., Javid-Majd, F., and Blanchard, J. S. (2001). Overexpression and mechanistic analysis of chromosomally encoded aminoglycoside 20-Nacetyltransferase (AAC(20)-Ic) from Mycobacterium tuberculosis.J. Biol. Chem. 276, 45876–45881. doi: 10.1074/jbc.M108810200 Ho, I. I., Chan, C. Y., and Cheng, A. F. (2000). Aminoglycoside resistance in Mycobacterium kansasii,Mycobacterium avium-M. intracellulare, and Mycobacterium fortuitum: are aminoglycoside-modifying enzymes responsible? Antimicrob. Agents Chemother. 44, 39–42. doi: 10.1128/AAC.44.1.39-42. 2000 Ho, Y. S., Adroub, S. A., Aleisa, F., Mahmood, H., Othoum, G., Rashid, F., et al. (2012). Complete genome sequence of Mycobacterium fortuitum subsp. fortuitum type strain DSM46621. J. Bacteriol. 194, 6337–6338. doi: 10.1128/JB. 01461-12 Houghton, J. L., Biswas, T., Chen, W., Tsodikov, O. V., and Garneau-Tsodikova, S. (2013a). Chemical and structural insights into the regioversatility of the aminoglycoside acetyltransferase Eis. Chembiochem 14, 2127–2135. doi: 10. 1002/cbic.201300359 Houghton, J. L., Green, K. D., Pricer, R. E., Mayhoub, A. S., and GarneauTsodikova, S. (2013b). Unexpected N-acetylation of capreomycin by mycobacterial Eis enzymes. J. Antimicrob. Chemother. 68, 800–805. doi: 10.1093/jac/dks497 Hull, S. I., Wallace, R. J. Jr., Bobey, D. G., Price, K. E., Goodhines, R. A., Swenson, J. M., et al. (1984). Presence of aminoglycoside acetyltransferase and plasmids in Mycobacterium fortuitum. Lack of correlation with intrinsic aminoglycoside resistance. Am. Rev. Respir. Dis. 129, 614–618. Jana, S., and Deb, J. K. (2005). Molecular targets for design of novel inhibitors to circumvent aminoglycoside resistance. Curr. Drug Targets 6, 353–361. doi: 10.2174/1389450053765860 Jennings, B. C., Labby, K. J., Green, K. D., and Garneau-Tsodikova, S. (2013). Redesign of substrate specificity and identification of the aminoglycoside binding residues of Eis from Mycobacterium tuberculosis.Biochemistry 52, 5125–5132. doi: 10.1021/bi4002985 Joshi, R. S., Jamdhade, M. D., Sonawane, M. S., and Giri, A. P. (2013). Resistome analysis of Mycobacterium tuberculosis: identification of aminoglycoside 20Nacetyltransferase (AAC) as co-target for drug desigining. Bioinformation 9, 174–181. doi: 10.6026/97320630009174 Jugheli, L., Bzekalava, N., de Rijk, P., Fissette, K., Portaels, F., and Rigouts, L. (2009). High level of cross-resistance between kanamycin, amikacin, and capreomycin among Mycobacterium tuberculosis isolates from Georgia and a close relation with mutations in the rrs gene. Antimicrob. Agents Chemother. 53, 5064–5068. doi: 10.1128/AAC.00851-09 Kim, K. H., An, D. R., Song, J., Yoon, J. Y., Kim, H. S., Yoon, H. J., et al. (2012). Mycobacterium tuberculosis Eis protein initiates suppression of host immune responses by acetylation of DUSP16/MKP-7. Proc. Natl. Acad. Sci. U.S.A. 109, 7729–7734. doi: 10.1073/pnas.1120251109 Kim, K. H., An, D. R., Yoon, H. J., Yang, J. K., and Suh, S. W. (2014). Structure of Mycobacterium smegmatis Eis in complex with paromomycin. Acta Crystallogr. F Struct. Biol. Commun. 70(Pt 9), 1173–1179. doi: 10.1107/S2053230X140 17385 Labby, K. J., and Garneau-Tsodikova, S. (2013). Strategies to overcome the action of aminoglycoside-modifying enzymes for treating resistant bacterial infections. Future Med. Chem. 5, 1285–1309. doi: 10.4155/fmc.13.80 Lella, R. K., and Sharma, C. (2007). Eis (enhanced intracellular survival) protein of Mycobacterium tuberculosis disturbs the cross regulation of T-cells. J. Biol. Chem. 282, 18671–18675. doi: 10.1074/jbc.C600280200 Li, Y., Green, K. D., Johnson, B. R., and Garneau-Tsodikova, S. (2015). Inhibition of aminoglycoside acetyltransferase resistance enzymes by metal salts. Antimicrob. Agents Chemother. 59, 4148–4156. doi: 10.1128/AAC.00885-15 Luthra, S., Rominski, A., and Sander, P. (2018). The role of antibiotic-targetmodifying and antibiotic-modifying enzymes in Mycobacterium abscessus drug resistance. Front. Microbiol. 9:2179. doi: 10.3389/fmicb.2018.02179 Macinga, D. R., and Rather, P. N. (1999). The chromosomal 20-N-acetyltransferase of Providencia stuartii: physiological functions and genetic regulation. Front. Biosci. 4, D132–D140. doi: 10.2741/Macinga Magnet, S., and Blanchard, J. S. (2005). Molecular insights into aminoglycoside action and resistance. Chem. Rev. 105, 477–498. doi: 10.1021/cr0301088 Maurer, F. P., Bruderer, V. L., Castelberg, C., Ritter, C., Scherbakov, D., Bloemberg, G. V., et al. (2015). Aminoglycoside-modifying enzymes determine the innate susceptibility to aminoglycoside antibiotics in rapidly growing mycobacteria. J. Antimicrob. Chemother. 70, 1412–1419. doi: 10.1093/jac/dku550 Maurer, F. P., Bruderer, V. L., Ritter, C., Castelberg, C., Bloemberg, G. V., and Bottger, E. C. (2014). Lack of antimicrobial bactericidal activity in Mycobacterium abscessus.Antimicrob. Agents Chemother. 58, 3828–3836. doi: 10.1128/AAC.02448-14 Mick, V., Rebollo, M. J., Lucia, A., Garcia, M. J., Martin, C., and Ainsa, J. A. (2008). Transcriptional analysis of and resistance level conferred by the aminoglycoside acetyltransferase gene aac(20)-Id from Mycobacterium smegmatis.J. Antimicrob. Chemother. 61, 39–45. doi: 10.1093/jac/dkm440 Mingeot-Leclercq, M. P., Glupczynski, Y., and Tulkens, P. M. (1999). Aminoglycosides: activity and resistance. Antimicrob. Agents. Chemother. 43, 727–737. doi: 10.1128/AAC.43.4.727 Mitsuhashi, S., Tanaka, T., Kawabe, H., and Umezawa, H. (1977). Biochemical mechanism of kanamycin resistance in Mycobacterium tuberculosis.Microbiol. Immunol. 21, 325–327. doi: 10.1111/j.1348-0421.1977.tb00294.x Ngo, H. X., Green, K. D., Gajadeera, C. S., Willby, M. J., Holbrook, S. Y. L., Hou, C., et al. (2018). Potent 1,2,4-triazino[5,6 b]indole-3-thioether inhibitors of the kanamycin resistance enzyme Eis from Mycobacterium tuberculosis.ACS Infect. Dis. 4, 1030–1040. doi: 10.1021/acsinfecdis.8b00074 Palomino, J. C., Martin, A., Camacho, M., Guerra, H., Swings, J., and Portaels, F. (2002). Resazurin microtiter assay plate: simple and inexpensive method for detection of drug resistance in Mycobacterium tuberculosis.Antimicrob. Agents Chemother. 46, 2720–2722. doi: 10.1128/AAC.46.8.2720-2722.2002 Pan, Q., Zhao, F. L., and Ye, B. C. (2018). Eis, a novel family of arylalkylamine N-acetyltransferase (EC 2.3.1.87). Sci. Rep. 8:2435. doi: 10.1038/s41598-01820802-6 Payie, K. G., and Clarke, A. J. (1997). Characterization of gentamicin 20-Nacetyltransferase from Providencia stuartii: its use of peptidoglycan metabolites for acetylation of both aminoglycosides and peptidoglycan. J. Bacteriol. 179, 4106–4114. doi: 10.1128/jb.179.13.4106-4114.1997 Frontiers in Microbiology | www.frontiersin.org 10 January 2019 | Volume 10 | Article 46 fmicb-10-00046 January 28, 2019 Time: 18:36 # 11 Sanz-García et al. Mycobacterial Aminoglycoside Acetyltransferases Payie, K. G., Rather, P. N., and Clarke, A. J. (1995). Contribution of gentamicin 20-N-acetyltransferase to the O acetylation of peptidoglycan in Providencia stuartii.J. Bacteriol. 177, 4303–4310. doi: 10.1128/jb.177.15.4303-4310.1995 Payie, K. G., Strating, H., and Clarke, A. J. (1996). The role of O-acetylation in the metabolism of peptidoglycan in Providencia stuartii.Microb. Drug Resist. 2, 135–140. doi: 10.1089/mdr.1996.2.135 Prabu, A., Hassan, S., Prabuseenivasan, Shainaba, A. S., Hanna, L. E., and Kumar, V. (2015). Andrographolide: a potent antituberculosis compound that targets Aminoglycoside 20-N-acetyltransferase in Mycobacterium tuberculosis.J. Mol. Graph. Model. 61, 133–140. doi: 10.1016/j.jmgm.2015.07.001 Pryjma, M., Burian, J., Kuchinski, K., and Thompson, C. J. (2017). Antagonism between front-line antibiotics clarithromycin and amikacin in the treatment of Mycobacterium abscessus infections is mediated by the whiB7 gene. Antimicrob. Agents Chemother. 61:e01353-17. doi: 10.1128/AAC.01353-17 Ramon-Garcia, S., Otal, I., Martin, C., Gomez-Lus, R., and Ainsa, J. A. (2006). Novel streptomycin resistance gene from Mycobacterium fortuitum. Antimicrob. Agents Chemother. 50, 3920–3922. doi: 10.1128/AAC.00223-06 Rather, P. N., Orosz, E., Shaw, K. J., Hare, R., and Miller, G. (1993). Characterization and transcriptional regulation of the 20-N-acetyltransferase gene from Providencia stuartii.J. Bacteriol. 175, 6492–6498. doi: 10.1128/jb.175. 20.6492-6498.1993 Reeves, A. Z., Campbell, P. J., Sultana, R., Malik, S., Murray, M., Plikaytis, B. B., et al. (2013). Aminoglycoside cross-resistance in Mycobacterium tuberculosis due to mutations in the 50untranslated region of whiB7. Antimicrob. Agents Chemother. 57, 1857–1865. doi: 10.1128/AAC.02191-12 Ripoll, F., Pasek, S., Schenowitz, C., Dossat, C., Barbe, V., Rottman, M., et al. (2009). Non mycobacterial virulence genes in the genome of the emerging pathogen Mycobacterium abscessus.PLoS One 4:e5660. doi: 10.1371/journal. pone.0005660 Roberts, E. A., Clark, A., McBeth, S., and Friedman, R. L. (2004). Molecular characterization of the Eis promoter of Mycobacterium tuberculosis.J. Bacteriol. 186, 5410–5417. doi: 10.1128/JB.186.16.5410-5417.2004 Rominski, A., Selchow, P., Becker, K., Brulle, J. K., Dal Molin, M., and Sander, P. (2017). Elucidation of Mycobacterium abscessus aminoglycoside and capreomycin resistance by targeted deletion of three putative resistance genes. J. Antimicrob. Chemother. 72, 2191–2200. doi: 10.1093/jac/dkx125 Samuel, L. P., Song, C. H., Wei, J., Roberts, E. A., Dahl, J. L., Barry, C. E., et al. (2007). Expression, production and release of the Eis protein by Mycobacterium tuberculosis during infection of macrophages and its effect on cytokine secretion. Microbiology 153(Pt 2), 529–540. doi: 10.1099/mic.0.2006/002642-0 Shcherbakov, D., Akbergenov, R., Matt, T., Sander, P., Andersson, D. I., and Bottger, E. C. (2010). Directed mutagenesis of Mycobacterium smegmatis 16S rRNA to reconstruct the in vivo evolution of aminoglycoside resistance in Mycobacterium tuberculosis.Mol. Microbiol. 77, 830–840. doi: 10.1111/j.13652958.2010.07218.x Shell, S. S., Wang, J., Lapierre, P., Mir, M., Chase, M. R., Pyle, M. M., et al. (2015). Leaderless transcripts and small proteins are common features of the mycobacterial translational landscape. PLoS Genet. 11:e1005641. doi: 10.1371/ journal.pgen.1005641 Shi, K., Caldwell, S. J., Fong, D. H., and Berghuis, A. M. (2013). Prospects for circumventing aminoglycoside kinase mediated antibiotic resistance. Front. Cell. Infect. Microbiol. 3:22. doi: 10.3389/fcimb.2013.00022 Shin, D. M., Jeon, B. Y., Lee, H. M., Jin, H. S., Yuk, J. M., Song, C. H., et al. (2010). Mycobacterium tuberculosis Eis regulates autophagy, inflammation, and cell death through redox-dependent signaling. PLoS Pathog. 6:e1001230. doi: 10.1371/journal.ppat.1001230 Smith, C. A., and Baker, E. N. (2002). Aminoglycoside antibiotic resistance by enzymatic deactivation. Curr. Drug Targets Infect. Disord. 2, 143–160. doi: 10. 2174/1568005023342533 Stover, C. K., de la Cruz, V. F., Fuerst, T. R., Burlein, J. E., Benson, L. A., Bennett, L. T., et al. (1991). New use of BCG for recombinant vaccines. Nature 351, 456–460. doi: 10.1038/351456a0 Suay-Garcia, B., and Perez-Gracia, M. T. (2018). Future prospects for Neisseria gonorrhoeae treatment. Antibiotics 7:E49. doi: 10.3390/antibiotics7020049 Tsodikov, O. V., Green, K. D., and Garneau-Tsodikova, S. (2014). A random sequential mechanism of aminoglycoside acetylation by Mycobacterium tuberculosis Eis protein. PLoS One 9:e92370. doi: 10.1371/journal.pone.0092370 Udou, T., Mizuguchi, Y., and Wallace, R. J. Jr. (1987). Patterns and distribution of aminoglycoside-acetylating enzymes in rapidly growing mycobacteria. Am. Rev. Respir. Dis. 136, 338–343. doi: 10.1164/ajrccm/136.2.338 Udou, T., Mizuguchi, Y., and Wallace, R. J. Jr. (1989). Does aminoglycosideacetyltransferase in rapidly growing mycobacteria have a metabolic function in addition to aminoglycoside inactivation? FEMS Microbiol. Lett. 48, 227–230. doi: 10.1111/j.1574-6968.1989.tb03304.x Udou, T., Mizuguchi, Y., and Yamada, T. (1986). Biochemical mechanisms of antibiotic resistance in a clinical isolate of Mycobacterium fortuitum. Presence of beta-lactamase and aminoglycoside-acetyltransferase and possible participation of altered drug transport on the resistance mechanism. Am. Rev. Respir. Dis. 133, 653–657. doi: 10.1164/arrd.1986.133.4.653 Vetting, M., Roderick, S. L., Hegde, S., Magnet, S., and Blanchard, J. S. (2003). What can structure tell us about in vivo function? The case of aminoglycosideresistance genes. Biochem. Soc. Trans. 31(Pt 3), 520–522. doi: 10.1042/ bst0310520 Vetting, M. W., Hegde, S. S., Javid-Majd, F., Blanchard, J. S., and Roderick, S. L. (2002). Aminoglycoside 20-N-acetyltransferase from Mycobacterium tuberculosis in complex with coenzyme A and aminoglycoside substrates. Nat. Struct. Biol. 9, 653–658. doi: 10.1038/nsb830 Waglechner, N., and Wright, G. D. (2017). Antibiotic resistance: it’s bad, but why isn’t it worse? BMC Biol. 15:84. doi: 10.1186/s12915-017-0423-1 Wallace, R. J. Jr., Hull, S. I., Bobey, D. G., Price, K. E., Swenson, J. M., Steele, L. C., et al. (1985). Mutational resistance as the mechanism of acquired drug resistance to aminoglycosides and antibacterial agents in Mycobacterium fortuitum and Mycobacterium chelonae. Evidence is based on plasmid analysis, mutational frequencies, and aminoglycoside-modifying enzyme assays. Am. Rev. Respir. Dis. 132, 409–416. doi: 10.1164/arrd.1985.132. 2.409 Wei, J., Dahl, J. L., Moulder, J. W., Roberts, E. A., O’Gaora, P., Young, D. B., et al. (2000). Identification of a Mycobacterium tuberculosis gene that enhances mycobacterial survival in macrophages. J. Bacteriol. 182, 377–384. doi: 10.1128/ JB.182.2.377-384.2000 Willby, M. J., Green, K. D., Gajadeera, C. S., Hou, C., Tsodikov, O. V., Posey, J. E., et al. (2016). Potent inhibitors of acetyltransferase Eis overcome kanamycin resistance in Mycobacterium tuberculosis.ACS Chem. Biol. 11, 1639–1646. doi: 10.1021/acschembio.6b00110 World Health Organization [WHO] (2010). Treatment of Tuberculosis Guidelines. Geneva: World Health Organization. doi: 10.1099/mic.0.024638-0 Wu, S., Barnes, P. F., Samten, B., Pang, X., Rodrigue, S., Ghanny, S., et al. (2009). Activation of the Eis gene in a W-Beijing strain of Mycobacterium tuberculosis correlates with increased SigA levels and enhanced intracellular growth. Microbiology 155(Pt 4), 1272–1281. doi: 10.1099/mic.0. 024638-0 Yoon, H. J., Kim, K. H., Yang, J. K., Suh, S. W., Kim, H., and Jang, S. (2013). A docking study of enhanced intracellular survival protein from Mycobacterium tuberculosis with human DUSP16/MKP-7. J. Synchrotron Radiat. 20(Pt 6), 929–932. doi: 10.1107/S0909049513021341 Zaunbrecher, M. A., Sikes, R. D. Jr., Metchock, B., Shinnick, T. M., and Posey, J. E. (2009). Overexpression of the chromosomally encoded aminoglycoside acetyltransferase Eis confers kanamycin resistance in Mycobacterium tuberculosis.Proc. Natl. Acad. Sci. U.S.A. 106, 20004–20009. doi: 10.1073/pnas.0907925106 Zhang, Y., and Yew, W. W. (2009). Mechanisms of drug resistance in Mycobacterium tuberculosis.Int. J. Tuberc. Lung Dis. 13, 1320–1330. Conflict of Interest Statement: The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. Copyright © 2019 Sanz-García, Anoz-Carbonell, Pérez-Herrán, Martín, Lucía, Rodrigues and Aínsa. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms. Frontiers in Microbiology | www.frontiersin.org 11 January 2019 | Volume 10 | Article 46 Ernesto Anoz Carbonell – Doctoral Thesis 260 Faig, M., Bianchet, M.A., Winski, S., Hargreaves, R., Moody, C.J., Hudnott, A.R., Ross, D., and Amzel, L.M. (2001). Structure-based development of anticancer drugs: Complexes of NAD(P)H:quinone oxidoreductase 1 with chemotherapeutic quinones. Structure 9, 659–667. Fedorovich, D., Protchenko, O., and Lesuisse, E. (1999). Iron uptake by the yeast Pichia guilliermondii. Flavinogenesis and reductive iron assimilation are co-regulated processes. Biometals an Int. J. Role Met. Ions Biol. Biochem. Med. 12, 295–300. Fischer, M., and Bacher, A. (2006). Biosynthesis of vitamin B2 in plants. Physiol. Plant. 126, 304– 318. Fisher, J., Spencer, R., and Walsh, C. (1976). Enzyme-catalyzed redox reactions with the flavin analogues 5-deazariboflavin, 5-deazariboflavin 5’-phosphate, and 5-deazariboflavin 5’- diphosphate, 5’→5’-adenosine ester. Biochemistry 15, 1054–1064. Forneris, F., Binda, C., Vanoni, M.A., Mattevi, A., and Battaglioli, E. (2005). Histone demethylation catalysed by LSD1 is a flavin-dependent oxidative process. FEBS Lett. 579, 2203–2207. Foster, C.E., Bianchet, M.A., Talalay, P., Zhao, Q., and Amzel, L.M. (1999). Crystal structure of human quinone reductase type 2, a metalloflavoprotein. Biochemistry 38, 9881–9886. Fraaije, M.W., and Mattevi, A. (2000). Flavoenzymes: Diverse catalysts with recurrent features. Trends Biochem. Sci. 25, 126–132. Frago, S., Martínez-Júlvez, M., Serrano, A., and Medina, M. (2008). Structural analysis of FAD synthetase from Corynebacterium ammoniagenes. BMC Microbiol. 8, 160. Frago, S., Velázquez-Campoy, A., and Medina, M. (2009). The puzzle of ligand binding to Corynebacterium ammoniagenes FAD synthetase. J. Biol. Chem. 284, 6610–6619. Frago, S., Lans, I., Navarro, J.A., Hervás, M., Edmondson, D.E., De la Rosa, M.A., Gómez-Moreno, C., Mayhew, S.G., and Medina, M. (2010). Dual role of FMN in flavodoxin function: electron transfer cofactor and modulation of the protein-protein interaction surface. Biochim Biophys Acta 1797, 262–271. Di Francesco, A., Di Germanio, C., Panda, A.C., Huynh, P., Peaden, R., Navas-Enamorado, I., Bastian, P., Lehrmann, E., Diaz-Ruiz, A., Ross, D., et al. (2016). Novel RNA-binding activity of NQO1 promotes SERPINA1 mRNA translation. Free Radic. Biol. Med. 99, 225–233. Fuchs, K.R., Shekels, L.L., and Bernlohr, D.A. (1992). Analysis of the ACP1 gene product: classification as an FMN phosphatase. Biochem Biophys Res Commun 189, 1598–1605. Gaedigk, A., Tyndale, R.F., Jurima-Romet, M., Sellers, E.M., Grant, D.M., and Leeder, J.S. (1998). NAD(P)H:quinone oxidoreductase: polymorphisms and allele frequencies in Caucasian, Chinese and Canadian Native Indian and Inuit populations. Pharmacogenetics 8, 305–313. Galluccio, M., Brizio, C., Torchetti, E.M., Ferranti, P., Gianazza, E., Indiveri, C., and Barile, M. (2007). Over-expression in Escherichia coli, purification and characterization of isoform 2 of human FAD synthetase. Protein Expr. Purif. 52, 175–181. García-Angulo, V.A. (2017). Overlapping riboflavin supply pathways in bacteria. Crit. Rev. Microbiol. 43, 196–209. Garcia, J.I., Medina, M., Sancho, J., Alonso, P.J., Gomez-Moreno, C., Mayoral, J.A., and Martinez, J.I. (2002). Theoretical analysis of the electron spin density distribution of the flavin semiquinone isoalloxazine ring within model protein environments. J. Phys. Chem. A 106, 4729–4735. [Document text truncated for crawler view.]