Esterification of drugs: a tool to improve hydrophobicity and encapsulation efficiency

ii DIREITOS DE AUTOR E CONDIÇÕES DE UTILIZAÇÃO DO TRABALHO POR TERCEIROS Este é um trabalho académico que pode ser utilizado por terceiros desde que respeitadas as regras e boas práticas internacionalmente aceites, no que concerne aos direitos de autor e direitos conexos. Assim, o presente trabalho pode ser utilizado nos termos previstos na licença abaixo indicada. Caso o utilizador necessite de permissão para poder fazer um uso do trabalho em condições não previstas no licenciamento indicado, deverá contactar o autor, através do RepositóriUM da Universidade do Minho. Licença concedida aos utilizadores deste trabalho Atribuição-NãoComercial-SemDerivações CC BY-NC-ND https://creativecommons.org/licenses/by-nc-nd/4.0/ iii Agradecimentos A finalização desta tese representa a concretização de uma importante etapa do meu percurso académico, estando profundamente marcada por várias pessoas que direta ou indiretamente contribuíram para o seu sucesso. Assim, deixo-lhes aqui o meu maior e mais profundo agradecimento. Em primeiro lugar, agradeço aos meus orientadores, ao Professor Artur Cavaco Paulo e à Doutora Carla Silva. Ao Professor Artur, pela oportunidade garantida e pela orientação e ajuda em todas as fases do projeto. À Carla, que para além de ser uma mentora de excelência, nunca me deixou fraquejar, impulsionando-me sempre a ser melhor e a atingir todos os meus objetivos. Agradeçote não só pelas oportunidades, conversas e carinho, mas também pela tua profunda e sincera amizade. Agradeço à Fundação para a Ciência e Tecnologia, pelo financiamento da minha bolsa de doutoramento (SFRH/BD/121673/2016) e à Universidade do Minho e Centro de Engenharia Biológica por garantirem todas as condições para a sua execução. A todos os elementos do grupo de investigação de Bioprocessos e Bionanotecnologia, não só pela constante companhia, mas também pelo excelente espírito de cooperação e solidariedade assente durante todo o percurso. Especial gratidão à Diana, Filipa, Catarina, David e Artur, pela amizade e pelos momentos de alegria que nunca esquecerei. À Tarsila Castro, pela simpatia e dedicação na realização de todo o trabalho de modelação molecular, fundamentais nesta tese. Um especial agradecimento ao Alfredo, que com toda a sua paciência, carinho e amor, contribuiu para que qualquer adversidade fosse superada. Foste um pilar capital desde o primeiro instante. Por fim, à minha família, sem os quais não seria possível concluir esta etapa. Agradeço aos meus pais, pela sua máxima dedicação, sacrifícios infindáveis e apoio incondicional incontestável. Pelos inúmeros ensinamentos e lições. Pelo amor que sempre senti e por tanto prezarem a minha educação. Agradeço-lhes especialmente por me ensinarem a apreciar as pequenas coisas, mas não me permitirem conformar, incentivando-me à superação e a lutar pelo meu futuro. À minha irmã Windy, que mesmo distante, sempre esteve (e sempre estará) presente. Vocês sempre foram o meu principal alicerce e sempre serão a estrela Polar que me guiará. A todos o meu mais profundo e sincero agradecimento. iv STATEMENT OF INTEGRITY I hereby declare having conducted this academic work with integrity. I confirm that I have not used plagiarism or any form of undue use of information or falsification of results along the process leading to its elaboration. I further declare that I have fully acknowledged the Code of Ethical Conduct of the University of Minho. v Esterificação de drogas: uma ferramenta para aumentar a hidrofobicidade e eficiência de encapsulação Resumo A necessidade de implementação de processos mais verdes e amigos do ambiente, continua a ser uma temática de grande importância para a comunidade científica. Das diferentes estratégias disponíveis para modificar e estabilizar compostos hidrofóbicos, o uso de enzimas para sua síntese e de nanoemulsões para o seu encapsulamento, são as abordagens mais comuns. Nesta tese, foram desenvolvidas nanoemulsões compostas por oligossacarídeos lipofílicos para a encapsulação de um composto hidrofóbico, o metotrexato (MTX). Enzimas, nomeadamente lipases ou proteases, foram aplicadas como catalisadores para a biossíntese de compostos hidrofóbicos com o objetivo de desenvolver alternativas ecológicas para futuras aplicações industriais. Primeiramente, a síntese de ciclo-oligossacarídeos lipofílicos (β-ciclodextrina, γ-ciclodextrina e ciclosoforaose) foi realizada através da ligação de cadeias de palmitoílo aos grupos hidroxilo. As estruturas hidrofóbicas desenvolvidas revelaram propriedades emulsificantes quando submetidas a ultrassons. A sua estabilidade, baixo tamanho e baixa polidispersividade demonstraram o alto potencial destas emulsões como novos dispositivos para o encapsulamento e libertação do MTX. Posteriormente, foram usadas duas abordagens enzimáticas para produzir pró-drogas de MTX. Inicialmente, lipases da Thermomyces lanuginosus (TL) e da Cândida antarctica B (CALB) foram utilizadas como catalisadores para a reação do MTX com triacilgliceróis e ciclodextrinas. Um meio aquoso, sob a ação de ultrassons, revelou ser uma estratégia eficaz para a síntese de pró-drogas de MTX, com o acoplamento de compostos não tóxicos. Na segunda abordagem, foi utilizada a - quimiotripsina do pâncreas de bovino para a auto-polimerização do MTX. A protease revelou capacidade de produzir oligómeros de MTX até um máximo de 6 unidades. Foi assim estabelecida uma nova via para a produção de uma pró-droga de MTX polimérico, salientando o potencial desta protease para a catálise de substratos não naturais. Ao longo desta tese, foi também avaliado o efeito da modificação química de enzimas nas suas propriedades catalíticas. Numa primeira abordagem, as modificações foram realizadas através da PEGilação das lipases TL, CALB e de uma cutinase da fusarium solani pisi . Verificou-se que enquanto a PEGilação melhorou a atividade hidrolítica da CALB e da cutinase, a atividade da lipase TL permaneceu inalterada. O efeito dessa modificação foi também avaliado na atividade de polimerase das enzimas, na síntese de um poliéster, o poli(etileno glutarato). Das três enzimas estudadas, a lipase TL PEGuilada foi responsável pela maior produção de polímero, juntamente com um maior grau de polimerização (DP), em comparação com a forma nativa da enzima. Ambas as formas da CALB exibiram atividade de polimerase semelhante. O efeito da PEGuilação na cutinase não foi significativo, sendo detetado um pequeno aumento na conversão do polímero, em comparação com a enzima nativa, porém com menor DP. A segunda abordagem, implementada para a melhoria das propriedades catalíticas das lipases, baseou-se na ligação de pequenos ligandos hidrofóbicos, aldeídos e isotiocianatos, à lipase TL. O efeito dos ligandos na atividade hidrolítica e na estabilidade da enzima foi extensivamente estudado, tendo os dados revelado maior estabilidade a diferentes temperaturas e pHs, após a sua modificação. Além disso, a lipase modificada com 4 cadeias de dodecil (a partir do dodecil aldeído) apresentou excelente atividade catalítica, avaliada por reação com substratos de diferentes tamanhos de cadeia alifática ( p -nitrofenil acetato a p -nitrofenil palmitato). Posteriormente, foi avaliada a atividade das enzimas na transesterificação do p -nitrofenil palmitato e na esterificação do ácido oleico, usando como substratos álcoois com diferentes tamanhos de cadeia (de metanol a eicosanol). A lipase modificada com 4 cadeias de dodecil, apresentou uma atividade superior, para ambas as reações, em comparação com a lipase nativa. Este incremento foi diretamente proporcional ao tamanho do álcool utilizado como substrato. Estas estratégias demonstraram a grande potencialidade das enzimas modificadas na biossíntese de produtos industriais de alto valor acrescentado. Palavras-chave: enzimas, esterificação, metotrexato, modificação química, nanoemulsões. vi Esterification of drugs: a tool to improve hydrophobicity and encapsulation efficiency Abstract The pursuit for green and environmentally friendly processes is still a great challenge among the scientific community. From the different strategies available to modify and stabilize hydrophobic compounds, the use of enzymes for their synthesis and of nanoemulsions for their encapsulation, are the most common approaches. In this thesis, nanoemulsions composed by lipophilic oligosaccharides were developed for the encapsulation of an hydrophobic compound, methotrexate (MTX). Enzymes, namely lipases or proteases, were applied as catalysts for the biosynthesis of hydrophobic compounds aiming to develop new eco-friendly routes for future industrial applications. Firstly, the synthesis of lipophilic cyclo-oligosaccharides (β-cyclodextrin, -cyclodextrin and cyclosophoraose) was carried out through the chemical coupling of palmitoyl chains to the hydroxylic groups. The hydrophobic structures showed emulsifying properties under ultrasonic energy. Their stability, narrow size, and low polydispersity demonstrated the high potential as new nanodevices for encapsulation and delivery of MTX. Afterward, two different enzymatic approaches were used to produce MTX-prodrugs. In a first approach, lipases from Thermomyces lanuginosus (TL) and from Candida antarctica B (CALB) were applied as catalysts for the reaction of MTX with triacylglycerols and cyclodextrins. Aqueous medium, together with ultrasounds showed to be a successful route for the synthesis of MTX ester prodrugs, with the coupling of non-toxic compounds. The second approach included the use -chymotrypsin from bovine pancreas for the self-polymerization of MTX. The protease revealed the ability to produce oligomers up to 6 MTX units. A novel pathway to produce a polymeric MTX prodrug was established, unravelling the potential of this protease for the catalysis of non-natural substrates. Along this thesis, the effect of the chemical modification of enzymes on their catalytic properties was also assessed. In a first approach, the chemical modifications were performed throughout PEGylation of lipase TL, CALB and cutinase from fusarium solani pisi . The data revealed that whereas PEGylation improved the hydrolytic activity of CALB and cutinase, it did not alter the activity of lipase TL. The effect of this modification was also evaluated throughout their polymerase activity in the synthesis of a polyester, poly(ethylene glutarate). From the three enzymes, PEGylated lipase TL was responsible for the highest polymer production with the highest degree of polymerization (DP), in comparison to the native form. Both CALB forms displayed similar polymerase activity. The effect of PEGylation on cutinase was not significant being detected a small increase of polymer conversion, however with lower DP, comparing with the native enzyme. The second approach for the improvement of lipases’ catalytic properties covered the grafting of small hydrophobic aldehydes and isothiocyanates to the lipase TL. The effect of the linkers was extensively studied in terms of hydrolytic activity and stability, and the data obtained revealed higher temperature and pH stability after modification. Moreover, the lipase grafted with 4 dodecyl chains (from dodecyl aldehyde) showed outstanding catalytic activity, evaluated against a panoply of substrates differing in the aliphatic chain size ( p -nitrophenyl acetate to p -nitrophenyl palmitate). Afterward, the enzymes’ activity in the transesterification of p -nitrophenyl palmitate, and in the esterification of oleic acid was evaluated. Chainsize differentiated alcohols (methanol to eicosanol) were tested as substrates of the reactions. The modified lipase (with 4 dodecyl chains) showed superior activity for both reactions comparing to the native lipase. This increment was more directly proportional to the size of the alcohol used for the reaction. These strategies revealed high potentiality for the use of modified enzymes on the biosynthesis of industrial added-value products. Keywords: chemical modification, enzymes, esterification, methotrexate, nanoemulsions. vii Table of contents Agradecimentos ................................................................................................................... iii Resumo ............................................................................................................................... v Abstract .............................................................................................................................. vi List of Symbols and Abbreviations ........................................................................................ xiii List of Figures .................................................................................................................... xvii List of Tables .................................................................................................................... xxiv List of Schemes ................................................................................................................ xxvi Chapter I ............................................................................................................................. 1 Thesis motivation and outline ................................................................................................ 1 1.1. Motivation and objectives ........................................................................................... 2 1.2. Thesis outline ............................................................................................................ 4 Chapter II ............................................................................................................................ 7 Oil-based cyclo-oligosaccharide nanodevices for drug encapsulation ......................................... 7 Abstract ................................................................................................................................ 8 2.1. Introduction ............................................................................................................... 9 2.2. Materials and methods ............................................................................................ 10 2.2.1. Materials ............................................................................................................. 10 2.2.2. Isolation of Cyclosophoraose ................................................................................ 11 2.2.3. Synthesis of modified cyclo-oligosaccharides ........................................................ 11 2.2.4. Nuclear Magnetic Resonance spectroscopy (NMR) ............................................... 12 2.2.5. Matrix-Assisted Laser Desorption/Ionization Time-of-Flight (MALDI-TOF) ................ 12 2.2.6. Cyclo-oligosaccharide-palmitoyl conjugate nanoemulsions .................................... 12 2.2.7. Dynamic Light Scattering (DLS) ............................................................................ 12 2.2.8. Nanoparticle Tracking Analysis (NTA) ................................................................... 13 2.2.9. Differential Scanning Calorimetry (DSC) ............................................................... 13 2.2.10. Scanning Transmission Electron Microscopy (STEM) ............................................ 13 2.2.11. Molecular Dynamics Simulations.......................................................................... 13 2.2.12. Cellular viability assay .......................................................................................... 14 2.2.12.1. Cells and culture conditions ............................................................................. 14 2.2.12.2. Cell viability assay ............................................................................................ 14 2.2.13. Stability of nanoemulsions in the presence of lipase ............................................. 14 xiv DPmax: Maximum degree of polymerization DSC: Differential Scanning Calorimetry ee : enantiomeric excess ESI: Electrospray Ionization FDA: Food and Drug Administration FF: Force field FTIR: Fourier-transform infrared spectroscopy : Gamma -CD: Gamma-cyclodextrin GC-MS: Gas Chromatography – Mass Spectrometry Gly: Glycine Glu: Glutamic acid HA: Heavy-atom His: Histidine HMBC: 1H-13C Heteronuclear Multiple Bond Correlation HPLC: High-Performance Liquid Chromatography HSQC: 1H-13C Heteronuclear Single Quantum Coherence k cat: turnover number K M: Michaelis-Menten constant LCR: Lipases from Candida rugosa LE: Ligand Efficiency Leu: Leucine LGA: Lamarckian Genetic Algorithm LYP: Lysine residue with a poly(ethylene glycol) unit Lys: Lysine [M]: Monomeric repeating unit MALDI-TOF: Matrix-Assisted Laser Desorption/Ionization Time-of-Flight MD: Molecular Dynamics 𝑀𝑛: Number average molecular weight M.p.: Melting point MTX: Methotrexate MTX-CD: Methotrexate-cyclodextrin conjugate xv MW: Molecular weight 𝑀𝑤: Weight average molecular weight NHS: N - Hydroxysuccinimide NIBMA: N -(iso-butoxymethyl) acrylamide NMR: Nuclear Magnetic Resonance NPT: Isothermal-isobaric ensemble, constant number of particles (N), pressure (P) and temperature (T) NTA: Nanoparticle Tracking Analysis NVT: Canonical ensemble, constant number of particles (N), volume (V) and temperature (T) p : para (substituted) P-Cys: Pentacosadiynoyl cyclosophoraose PAL: Palmitoyl PAL-Cl: Palmitoyl chloride PBS: Phosphate Buffer Saline PDB: Protein Data Bank PDI: Polydispersity Index PEG: Poly(ethylene glycol) PLA: Poly(lactic acid) PLGA: Poly(lactide-co-glycolide) PME: Particle-mesh Ewald p -NP: p -Nitrophenol p -NPAc: p -Nitrophenyl Acetate p -NPB: p -Nitrophenyl Butyrate p -NPH: p -Nitrophenyl Hexanoate p -NPL: p -Nitrophenyl Laurate p -NPO: p -Nitrophenyl Octanoate p -NPP: p -Nitrophenyl Palmitate p -NPPh: p -Nitrophenyl Benzoate RMSD: Root Mean Square Deviation RMSF: Root Mean Square Fluctuation rt: room temperature SASA: Solvent Accessible Surface Area xvi SD: Standard Deviation SDS-PAGE: Sodium Dodecyl Sulfate Polyacrylamide Gel Electrophoresis Ser: Serine STEM: Scanning Transmission Electron Microscopy T½: Half-life time Tg: Glass transition temperature (DSC) Tm: Melting point temperature (DSC) TBMA: tert -butyl methacrylate TEMED: Tetramethylethylenediamine TGA: Thermogravimetric Analysis Thr: Threonine TL: Thermomyces lanuginosus TL1: lipase from Thermomyces lanuginosus modified with 1 unit of naphthyl isothiocyanate TL2: lipase from Thermomyces lanuginosus modified with 1 unit of naphthyl aldehyde TL3: lipase from Thermomyces lanuginosus modified with 7 units of butyraldehyde TL4: lipase from Thermomyces lanuginosus modified with 7 units of hexyl aldehyde TL5: lipase from Thermomyces lanuginosus modified with 4 units of dodecyl aldehyde TLC: Thin Layer Chromatography TNBSA: 2,4,6-Trinitrobenzene sulfonic acid Tyr: Tyrosine US: ultrasound Vmax: maximum rate WB: water bath ŋ: Yield or catalytic efficiency xvii List of Figures Figure 2.1. Synthesis of modified cyclo-oligosaccharides: A) reactional scheme of the conjugation of PAL-Cl with Cys; B) reactional scheme of the conjugation of PAL-Cl with β-CD (n=7) and γ-CD (n=8)……………………………………………………………………………………………………..16 Figure 2.2. Positive ion MALDI-TOF spectrum of: A) β-CD-PAL, B) -CD-PAL and C) Cys-PAL on DHB………………………………………………………………………………………………………………17 Figure 2.3. 1H NMR spectra of: A) Cys and Cys-PAL; B) β-CD and β-CD-PAL and C) -CD and -CDPAL……………………………………………………………………………………………………………….18 Figure 2.4. Formulations photographs after emulsification with US: I) Cys (5 mg/mL); II) Cys-PAL (5 mg/mL); III) BSA (10 mg/mL) + Cys-PAL (5 mg/mL); IV) γ-CD (5 mg/mL); V) γ-CD-PAL (5 mg/mL); VI) BSA (10 mg/mL) + γ-CD-PAL (5 mg/mL); VII) β-CD (5 mg/mL); VIII) β-CDPAL (5 mg/mL); IX) BSA (10 mg/mL) + β-CD-PAL (5 mg/mL) and X) BSA (10 mg/mL)………………………………………………………………………………………………………….19 Figure 2.5. Molecular dynamic simulations of Cys-PAL auto aggregation: I) cyclosophoraose (red) surrounded by covalently bonded palmitic chains (green), before aggregation, II) 15 units of Cys-PAL self-aggregated; and BSA-Cys-PAL aggregation: III) BSA-Cys-PAL before nanoemulsion simulation, showing 6 units of BSA (in red) and 12 units of Cys-PAL (in cyan), IV) BSA-Cys-PAL after nanoemulsion simulation (one unit of Cys-PAL surrounded by several units of BSA)……………………………………………………………………………………………………20 Figure 2.6. STEM images of nanoemulsions: A) Cys-PAL (5 mg/mL) (stained with uranyl acetate); B) γ-CD-PAL (5mg/mL); C) β-CD-PAL (5 mg/mL); D) BSA (10 mg/mL) + Cys-PAL (5 mg/mL) (stained with uranyl acetate); E) BSA (10 mg/mL) + γ-CD-PAL (5 mg/mL) (stained with uranyl acetate)…………………………………………………………………………………………..21 Figure 2.7. Stability (size and PDI) of nanoemulsions during storage at 4 °C for 60 days (values are the mean ± SD of 3 independent experiments)………………………………………………….22 Figure 2.8. BJ5ta cell line viability after 72 h of contact with different concentrations (from 0.0625 mg/mL up to 1 mg/mL) of the four different nanoemulsion formulations, compared with cells (negative control) and cells incubated with 30 % (v/v) of DMSO (death control), determined by MTS assay. Values are the mean ± SD of 2 independent experiments….…24 Figure 2.9. 1H NMR (in D2O) of MTX free (in blue) and encapsulated in -CD-PAL nanoemulsions (in red)……………………………………………………………………………………………………………26 xviii Figure 2.10. Release of MTX from the nanoemulsions over time (values are the mean ± SD of 3 independent experiments)……………………………………………………………………………….…27 Figure 2.11. STEM microphotographs of nanoemulsions with MTX encapsulated: A) Cys-PAL-MTX before MTX release; B) Cys-PAL-MTX after MTX release; C) γ-CD-MTX before MTX release; D) γ-CD-MTX after MTX release…………………………………………………………………………..28 Figure 3.1. 1H NMR spectra of A) free MTX and B) methotrexate-hexanoate conjugate (DMSOd6)…………………………………………………………………………………………………………………40 Figure 3.2. 1H NMR of MTX-CD (α, β and γ)-conjugates in DMSO-d6 using lipase from Thermomyces lanuginosus in the US..………………………………………………………………….43 Figure 3.3. MALDI-TOF of MTX-CD conjugates: A) α-CD conjugated with 1 MTX unit, B) β-CD conjugated with 2 MTX units, C) γ-CD conjugated with 3 MTX units; the upper image shows the proposed MTX-CD conjugates; all conjugates were obtained after reaction using lipase from Thermomyces lanuginosus in ultrasound………………………………………………………..44 Figure 4.1. Proposed mechanism for the synthesis of a dimeric unit of methotrexate catalysed by -chymotrypsin………………………………………………………………………………………………. 55 Figure 4.2. A) Backbone RMSD of -chymotrypsin, simulated at 310 K and physiological pH (black trace), and at 323 K and pH 9.5 (red trace), from the initial 3D structure (1OXG); B) 3-D structure of bovine α-chymotrypsin (PDB ID: 1OXG), from X-ray diffraction (I), highlighting in stick representation the catalytic triad: His57, Asp102 and Ser195. Green for cartoon/carbon, blue for nitrogen, red for oxygen and with for hydrogens; (II) shows the electrostatic distribution, where blue corresponds to positive areas and red to negative ones………………………………………………………………………………………………………………57 Figure 4.3. A) Skeletal and 3-D methotrexate (MTX) structures at pH 7.0 and B) the deprotonated forms at pH 9.5. C) skeletal and 3-D structures of MTX dimer, at pH 9.5 and D) shows the trimer 3-D representation. The colour scheme use green for carbon, blue for nitrogen, red for oxygen and white for hydrogen………………………………………………………………………..58 Figure 4.4. Docking poses showing the interaction of α-chymotrypsin with the different methotrexate species, and the respective binding energies obtained using AutoDock 4.0: A) α-chymotrypsin with methotrexate; B) α-chymotrypsin with dimeric methotrexate and C) αchymotrypsin with trimeric methotrexate. Surface/cavities representation are in the right panel. Enzyme is represented in light grey, hydrogen bonds in yellow dashes and methotrexates following the scheme previous described in Figure 4.3…………………………60 xix Figure 4.5. 1H NMR (DMSO-d6) spectra of A) methotrexate and B) oligomeric methotrexate synthesized by α-chymotrypsin……………………………………………………………………………61 Figure 4.6. MALDI-TOF spectra of: A) control reaction carried without enzyme, and B) oligomeric methotrexate synthesized by -chymotrypsin, and ESI spectra of: C) monomeric methotrexate and D) oligomeric methotrexate where the most abundant specie (100 %) is observed at m/z 975.36 corresponding to the dimeric MTX………………………………………62 Figure 4.7. Differential scanning calorimetry (DSC) curves of methotrexate (MTX) and oligomeric methotrexate; the black line corresponds to the monomer; the grey line corresponds to oligomeric MTX synthesized by α-chymotrypsin………………………………………………….…..63 Figure 5.1. SDS-PAGE gel of native and PEGylated esterases stained with Coomassie brilliant blue; A) GRS Protein Marker Blue (from Grisp, Portugal), B) Lipase from Thermomyces lanuginosus ; C) PEGylated lipase from Thermomyces lanuginosus ; D) Cutinase from Fusarium solani pisi ; E) PEGylated cutinase from Fusarium solani pisi ………………………..75 Figure 5.2. Absolute activity of esterases (native and PEGylated forms), at time zero, after 8 h, and after 50 days of incubation at 40 °C in phosphate buffer (pH 7.8). The activity was measured against pNPB over 1 min and considering the same initial amount of protein; 1 U of enzyme activity was defined as the amount of enzyme required to convert the substrate ( p -nitrophenyl butyrate) into p -nitrophenol in 1 min……………………………………………………….…………….76 Figure 5.3. Conversion (%) vs enzyme loading (U/mg) after synthesis of poly(ethylene glutarate)………………………………………………………………………………………………………..78 Figure 5.4. Conversion (%) and degree of polymerization (DP) for all the tested esterases under the optimized conditions: 2 h under US followed by 5 h under vacuum, at 40 °C; 65 U/mg. The graph bar corresponds to the conversion (%) and the graph line to the average DP………..79 Figure 5.5. Lipase TL in water medium (A) and in substrates mixture medium (B). Optimized structures of diethyl glutarate and ethylene glycol in united-atom GROMOS 54a7 FF representation (C). Initial structures of lipase TL (green, 1TIB) and CALB (grey, 1TCA) (D). PEGylated structures based on lipase TL and CALB: LP1 (LYP98), LP2 (LYP98, LYP74, LYP223 and LYP237), LP3 (LYP98, LYP24, LYP46, LYP127), LP4 (LYP98, LYP24, LYP46, LYP223, LYP237), LP5 (LYP46, LYP74, LYP127, LYP223, LYP237) and PEG-CALB (all Lys replaced by LYP). Lipases are represented in cartoon, with the lid region highlighted in cyan, the catalytic triad in yellow, and LYP residues in magenta…………………………………………82 xx Figure 5.6. Backbone RMSD of lipase TL, CALB and PEGylated analogues in solvent (A-B) and in water (C-D)……………………………………………………………………………………….………..…..83 Figure 5.7. Middle conformations of lipase TL and PEGylated analogue LP1, in reactant mixture (A) and in water (B), highlighting the interior cavities and pockets surrounding the catalytic triad and lid regions……………………………………………………………………………………………..….84 Figure 5.8. Middle conformations of CALB and PEGylated CALB, in reactant mixture (A) and in water (B), highlighting the cavities and pockets surrounding the catalytic triad and lid regions. Cartoon in grey, catalytic triad in yellow, lid-like region in cyan, LYP residues in magenta and the orange spheres represent the empty space (cavity or pocket) on each structure………85 Figure 5.9. 1H (A) and 13C (B) NMR spectra of poly(ethylene glutarate) recorded in CDCl3…..…..87 Figure 5.10. Positive ion MALDI-TOF spectra of poly(ethylene glycol) (DPmax= 16 and DPavg= 5)………………………………………………………………………………………………………..88 Figure 5.11. FTIR spectra of A) diethyl glutarate; B) ethylene glycol; C) poly(ethylene glutarate) with a DPavg= 3; D) poly(ethylene glutarate) with a DPavg= 5……………………………………….…..88 Figure 5.12. Graphics of A) DSC curve of poly(ethylene glutarate) with Tg= 77.29 ± 1.21 °C and Tm = 195.7 °C, and the respective monomers and B) TGA curves of the starting materials (diethyl glutarate and ethylene glycol) and the formed poly(ethylene glutarate)………..….…89 Figure 6.1. Proposed reactional scheme for the enzyme modification: A) native enzyme with an exposed lysine residue represented; B) enzyme modified with an aldehyde in the lysine residue, leading to a secondary amine; C) enzyme modified with an isothiocyanate in the lysine residue, leading to a thiourea; linkers used for lipase modification: i) naphthaldehyde; ii) butyraldehyde; iii) hexyl aldehyde; iv) dodecyl aldehyde; v) naphthyl isothiocyanate; vi) phenethyl isothiocyanate; vii) octyl isothiocyanate…………………………………………….…..100 Figure 6.2. SDS-PAGE of native lipase from Thermomyces lanuginosus and modified lipases (TL1TL5)…………………………………………………………………………………………………………….102 Figure 6.3. A) Proposed reactional scheme representing the hydrolysis of the ester substrates and the designation of each substrate according to the number of carbons in the aliphatic chain; B) Absolute hydrolytic activity of native lipase (TL) and of modified lipases with aromatic linkers (TL1 and TL2), and C) with aliphatic linkers (TL3-TL5), measured against different substrates: p -nitrophenyl benzoate ( p -NPPh, 0C), p -nitrophenyl acetate ( p -NPAc, 2C), pnitrophenyl butyrate ( p -NPB, 4C), p -nitrophenyl hexanoate ( p -NPH, 6C), p -nitrophenyl xxi octanoate ( p -NPO, 8C), p -nitrophenyl laurate ( p -NPL, 12C) and p -nitrophenyl palmitate ( p - NPP, 16C)………………………………………………………………………………………………….…104 Figure 6.4 . Activity of the modified lipases versus hydrophobic area of the linkers; activity measured against p -nitrophenyl acetate ( p - NPAc , • ), p -nitrophenyl hexanoate ( p -NPH ,  ) and p - nitrophenyl laurate ( p -NPL ,  ). The colour of the symbols in the graphs corresponds to the colour of each modified enzyme represented bellow the graph………………………………………………………………………………………………….……….105 Figure 6.5. Catalytic turnover (η= k cat/ K M) of the native and modified enzymes calculated for the hydrolysis of p -nitrophenyl acetate ( p -NPAc, 2C), pnitrophenyl butyrate ( p -NPB, 4C) and p - nitrophenyl octanoate ( p -NPO, 8C)……………………………………………………………………..108 Figure 6.6. Middle structures characterized for lipase TL (A) and TL5 (C); Lid is highlighted in cyan, catalytic triad with the residues in green sticks and TL5 in blue sticks (C); (B) and (D) zoom in the active site, showing the pockets/cavities in yellow surface…………..………………….109 Figure 6.7. A) Backbone RMSD of lipase TL and TL5 in water; B) amino acids RMSF…………….110 Figure 6.8. Representation of the best docking poses, interactions and ΔGbinding between p - nitrophenyl butyrate with native TL (A) and with TL5 (C); p -nitrophenyl octanoate with native TL (B) and with TL5 (D). Enzymes are represented in grey cartoon, sticks to highlight the amino acids residues participating in the binding, substrate in cyan ball and sticks, and hydrogen bonds in green dashed…………………………………………………………….…………111 Figure 6.9. Distance (nm) along simulation time from p -nitrophenyl butyrate or p -nitrophenyl octanoate to the Ser146 in catalytic triad……………………………………………………………..112 Figure 6.10. Stability of native and modified enzymes: A) Relative activity (%) at time zero, and after 4 months in solution at 4 °C and 37 °C; B) Half-life time (in hours) of the enzymes at 60 °C; C) Effect of temperature in the enzymes’ activity after 10 min incubated at the respective temperatures (37 °C, 45 °C, 60 °C and 70 °C); D) Effect of pH in the enzymes’ activity after 10 min incubated at the respective pH (1.81, 4.46, 6.41, 7.80, 9.40 and 10.5). The results presented were obtained using p -nitrophenyl butyrate as substrate. E) and G) same as C) but using p -nitrophenyl octanoate or p -nitrophenyl palmitate as substrate, respectively; F) and H) same as D) but using p -nitrophenyl octanoate or p -nitrophenyl palmitate as substrate, respectively…………………………………………………………………..…………………113 Figure 6.11. Electrostatic potential surface (kb T ec-1) generated with PDB2PQR server and APBS plugin in PyMOL, at pH 10.5, for: A) TL with all Lys side chain protonated, and B) TL3 with xxii all Lys converted to TL3 linker. On the left exterior surface and on the right the cavities/pocket representation………………………………………………………………….……….115 Figure 6.12. Circular dichroism of native lipase (TL) and modified lipases (TL1 to TL5)……..….116 Figure 6.13. A) Fluorescence spectra of native (TL) and modified lipases (TL1-TL5) after excitation at 280 nm. B) Table of the maximum wavelength with the respective intensity value……..117 Figure 7.1. A) Representative scheme for the modification of (I) native lipase from Thermomyces lanuginosus with dodecyl aldehyde to produce (II) modified lipase TL with 4 dodecyl chains; B) MALDI-TOF of the modified lipase, confirming the grafting of 4 dodecyl chains (MW of native= 29620.3) [326]; C) Half-life time of both enzymes at different temperatures….….126 Figure 7.2. A) Reactional scheme for the transesterification reaction of p -nitrophenyl palmitate ( p - NPP) with differentiated size-alcohols (methanol, pentanol and decanol), to produce p - nitrophenol and an aliphatic ester; B) Transesterification activity of native (full lines) and modified lipase (dash lines), using methanol (green), pentanol (red) and decanol (blue) as alcohol substrates. K values calculated after 7 h of reaction; C) GC-MS chromatograms of the products of the transesterification of p -NPP with methanol, catalysed by native and modified lipase (after 7 h of reaction)………………………………………………………………….128 Figure 7.3. A) Reactional scheme for the esterification of oleic acid with differentiated chain-size alcohols; B) Reactional yield (%) of the esterification reaction of oleic acid with different alcohols catalysed by native and modified lipase TL (0.08 % w/v). Results obtained after 24 h of reaction, with the reactions performed at 37 ºC; C) K values ( K = [ester product]/[oleic acid]) regarding the effect of temperature (25, 37 and 50 ºC), in the esterification of oleic acid with alcohols (methanol, decanol and eicosanol), using native vs modified lipase (0.08 % w/v) after 30 h of reaction…………..…………………………………..132 Figure 7.4. A) 1H NMR spectrum of propyl oleate, synthesized by reaction between oleic acid and propanol catalysed by the modified lipase; B) Retention time, m/z obtained after GC-MS analysis and theoretical molecular weight of all biosynthesised esters; C) FTIR spectra of oleic acid (black) and of propyl oleate (grey) synthesized by the modified lipase…..……..134 Figure 8.1. Different fields of application where enzymes can be incorporated for the synthesis or modification of hydrophobic compounds (enzymes represented: A) lipase from Candida antarctica B PDB ID: 4K6G; B) cutinase from Fusarium solani PDB ID: 1CUS; C) - chymotrypsin from bovine pancreas PDB ID: 1YPH; D) lipase from Thermomyces lanuginosus PDB ID: 5AP9)………………………………………………………………………………138 xxiii Figure 8.2. Some of the enzyme properties improved by chemical modification (enzyme represented: lipase from Thermomyces lanuginosus , PDB ID: 1TIB, modified aleatorily with four dodecyl chains through lysine residues using PyMOL)……………………………………..145 Figure 8.3. Examples of reactions performed by lipase-catalysis……………………………………….147 Figure 8.4. Structure of lipase from Thermomyces lanuginosus (PDB ID: 1TIB) elucidating the lid in cyan and the catalytic triad Ser-His-Asp in magenta……………………………………………148 Figure 8.5. Examples of enzyme immobilization strategies………………………………………………156 Figure 8.6. Major advantages of both strategies: immobilization vs chemical modification of enzymes……………………………………………………………………………………………………….157 4 1.2. Thesis outline This thesis is divided into 8 chapters presented according to the temporal order of execution of the works. Chapter 1 describes the layout adopted, as well as the motivations behind the investigation performed. Chapters 2 to 7 are dedicated to the experimental work performed. Chapter 8 is a detailed review, which summarizes the importance and the novelty of the works here presented, comparing our findings with the data from literature, in the same field of investigation. The purpose of this final chapter was to demonstrate the major achievements obtained, which fulfilled some identified literature gaps. The content of each chapter is summarized below: Chapter II. Oil-based cyclo-oligosaccharide nanodevices for drug encapsulation This chapter was dedicated to the synthesis of new hydrophobic cyclo-oligosaccharides which demonstrated oily behaviour at low temperatures. This property was explored to produce new nanoemulsions. A full characterization of their physico-chemical properties was performed, as well as the capability to encapsulate MTX. A suitable release profile was observed, showing the promising properties of these novel nanoemulsions produced. This chapter is based on the following publication: Jennifer Noro, Ana Loureiro, Filipa Gonçalves, Nuno G. Azoia, Seunho Jung, Carla Silva, Artur Cavaco-Paulo, Oil-based cyclo-oligosaccharide nanodevices for drug encapsulation, Colloids and Surfaces B: Biointerfaces, 159 (2017) 259-267. Chapter III. Ultrasound-assisted biosynthesis of novel methotrexate-conjugates In this work, novel methotrexate (MTX) conjugates were biosynthesized, using two different lipases: lipase from Thermomyces lanuginosus and immobilized lipase from Candida antarctica B . The drug was covalently bonded to two different types of compounds: triacylglycerols and cyclodextrins. All reactions were performed in an ultrasonic bath, which was a key factor to obtain higher yields in the production of the conjugated compounds. This chapter is based on the following publication: Jennifer Noro, Rui L. Reis, Artur Cavaco-Paulo, Carla Silva, Ultrasound-assisted biosynthesis of novel methotrexate-conjugates, Ultrasonics Sonochemistry, 48 (2018) 51-56. 5 Chapter IV. -Chymotrypsin catalyses the synthesis of methotrexate oligomers A protease, -chymotrypsin, was explored for the self-polymerization of methotrexate by the production of new peptide bonds among substrate units. The tested enzyme was able to form oligomers up to 6 units, demonstrating capacity to polymerize non-natural substrates. This chapter is based on the following publication: Jennifer Noro, Tarsila G. Castro, Artur Cavaco-Paulo, Carla Silva, -chymotrypsin catalyses the synthesis of methotrexate oligomers, Process Biochemistry, 98 (2020) 193-201. Chapter V. Catalytic activation of esterases by PEGylation for polyester synthesis The effect of PEGylation on the catalytic properties of three enzymes was explored. Two lipases (from Candida antarctica B and Thermomyces lanuginosus ) and one cutinase (from Fusarium solani pisi ) were successfully PEGylated, and an improvement or preservation of their hydrolytic activity was observed. The main focus of the work was to explore their polymerase activity in the biosynthesis of a polyester, namely, poly(ethylene glutarate). It was observed that the PEGylation of lipase TL and of the cutinase enhanced their polymerase activity, comparing to their native forms. This chapter is based on the following publication: Jennifer Noro, Tarsila Castro, Filipa Gonçalves, Artur Ribeiro, Artur Cavaco-Paulo, Carla Silva, Catalytic activation of esterases by PEGylation for polyester synthesis, ChemCatChem, 11 (2019) 2490-2499. Chapter VI. Substrate’s hydrophobicity and enzyme’s modifiers play a major role on the activity of lipase from Thermomyces lanuginosus The chemical modification of lipase from Thermomyces lanuginosus with small hydrophobic compounds (isothiocyanates and aldehydes) was performed. The impact of the modification on the lipase’s activity and enzyme’s conformation was extensively studied. The chemical modification with aldehydes showed to improve their hydrolytic activity against differentiated size substrates and enhance their thermostability. This chapter is based on the following publication: Jennifer Noro, Tarsila G. Castro, Artur Cavaco-Paulo, Carla Silva, Substrate’s hydrophobicity and enzyme’s modifiers play a major role on the activity of lipase from Thermomyces lanuginosus , Catalysis Science & Technology, 10 (2020) 5913-5924. 6 Chapter VII. Chemical modification of lipase from Thermomyces lanuginosus enhances transesterification and esterification activity In this chapter, a practical application of the most active modified lipase (grafted with 4 dodecyl chains) was evaluated on transesterification and esterification reactions. A range of linear alcohols differing in chain size (methanol to eicosanol) were studied, for the transesterification of p - nitrophenyl palmitate and the esterification of oleic acid. It was observed that the modified lipase showed higher activity for both transesterification and esterification reactions, than the native enzyme, being this increment more pronounced for the longer alcohols studied. This chapter is based on the following publication: Jennifer Noro, Artur Cavaco-Paulo, Carla Silva, Chemical modification of lipase from Thermomyces lanuginosus enhances transesterification and esterification activity , submitted to: ACS Catalysis. Chapter VIII. Biotechnological approaches for the synthesis, modification and stabilization of hydrophobic compounds In this final chapter, a revision of the literature related with the research topics developed during the thesis were addressed. The results obtained were compared with the recent reports from literature, showing the direct impact on the investigation related with the stabilization and modification of hydrophobic compounds. The new compounds and techniques developed are beyond state-of-art, fulfilling the gaps identified in literature regarding the use of enzymes as catalysts for MTX-prodrugs and their practical application after chemical modification. This chapter is based on the following publication: Jennifer Noro, Artur Cavaco-Paulo, Carla Silva, Biotechnological approaches for the synthesis, modification and stabilization of hydrophobic compounds, submitted to: Critical Reviews in Biotechnology. 7 Chapter II Oil-based cyclo-oligosaccharide nanodevices for drug encapsulation 8 Chapter II Oil-based cyclo-oligosaccharide nanodevices for drug encapsulation Abstract New encapsulation nanodevices were synthesized by emulsification of cyclo-oligosaccharides fully substituted by hydrophobic palmitic chains. These highly hydrophobic compounds acquire oily-like behaviour at moderate temperatures ( 50 °C) and when submitted to ultrasounds (US) can undergo emulsification. The improved emulsifying properties of modified cyclo-oligosaccharides are suitable to produce small and narrow sized nanoemulsions with ability to encapsulate amphiphilic molecules. Both encapsulation and delivery of a therapeutic drug with amphiphilic character, methotrexate (MTX), were assessed. The physicochemical properties of the cyclo-oligosaccharide nanoemulsions containing MTX were investigated by nuclear magnetic resonance (NMR), scanning transmission electron microscopy (STEM) and dynamic light scattering (DLS). The results revealed that the modified cyclo-oligosaccharides are potential platforms for the encapsulation of bio compounds for cosmetic and pharmaceutical purposes. This chapter is based on the following publication: Jennifer Noro, Ana Loureiro, Filipa Gonçalves, Nuno G. Azoia, Seunho Jung, Carla Silva, Artur Cavaco-Paulo, Oil-based cyclo-oligosaccharide nanodevices for drug encapsulation, Colloids and Surfaces B: Biointerfaces, 159 (2017) 259-267. 9 2.1. Introduction The administration of drugs with poor solubility and low rate of dissolution through different routes has been a major challenge for the pharmaceutical research. Several techniques have been attempted to drive drugs solubility, bioavailability, and dissolution properties, namely solubilisation [16, 17], cosolvency [18] and solid dispersions [19, 20]. The encapsulation of drugs into nanodevices has also been explored to overcome the unfavourable solubility and inappropriate interactions of drugs with other chemical compounds during delivery that are responsible by their toxicological characteristics [21]. Moreover, the nanodevices may protect the drug from degradation and increase selectivity of the drug modifying the pharmacokinetic and drug tissue distribution profile, which ensure a high safety and biocompatibility [21, 22]. Among various particulate systems, liposomes and nanoparticles have been applied for site specific delivery of drugs. However, they possess size dependent properties, stability, and scale-up problems. Lipidbased delivery systems and water-in-oil microemulsions have overcome the frequency of administration of certain drugs, e.g. MTX, by providing the control delivery to absorption sites but any of them was able to achieve the target or minimize the adverse side effects [23]. The encapsulation of drugs has been explored by using several other devices, namely micellar nanonetworks [24], magnetic microcapsules [25], PLA-PEG-PLA nanoparticles [26], dendrimeric nanodevices [27], protein nanoemulsions [2, 28] among others. Still, some of them showed an almost completely release of the drug after one day [24, 25]. Cyclo-oligosaccharides complexation has been presented as a promising strategy to tackle the mentioned formulation issues. Their ability to form inclusion complexes with a wide variety of hydrophobic compounds provide significant advantages related with changes of the physicochemical and biological properties of the guest molecules. The complexation can also provide protection of the component against light and oxidation, its solubilisation, handling improvement and stability [29]. Cyclosophoraoses (Cys) are unbranched cyclic β-1,2-D-glucans containing 17-23 glucose residues produced by Rhizobium and Agrobacterium species both intracellularly and extracellularly [30]. The targeting of non-polar chemicals like ergosterol, luteolin, vitamin D3 and naproxen has been reported as Cys complexation examples [31, 32]. A flexible backbone structure with doughnut-like ring shape is attributed to Cys and is responsible for the atypical induced-fit type complexation with hydrophobic molecules. Besides complexation systems, Cys can also be applied as a novel biosourced saccharide catalyst for chemical reactions as reported by Dindulkar et al . [33] They reported the use of microbial Cys for the synthesis of therapeutically important versatile indolyl 4 H -chromenes 10 via a one pot three-component Knoevenagel–Michael addition–cyclization reaction of salicylaldehyde, 1,3-cyclohexanedione/dimedone, and indoles in water under neutral conditions. The chemical modification of Cys has been also assessed by Kim et al. , to increase its hydrophobicity [34]. Their studies revealed the improvement of the solubility of insoluble flavonoids, comparing with the non-modified Cys [34]. They also described a new pentacosadiynoyl cyclosophoraose (P-Cys) synthesized using a biosourced cyclo-oligosaccharide with intrinsic complexation capacity [30]. Their results indicated that P-Cys is a useful potential platform for the encapsulating emulsification of bioactive molecules for cosmetic and pharmaceutical applications. The main aims of this study were firstly to chemically modify the cyclo-oligosaccharides (Cys, βcyclodextrin (CD) and γ-CD) by reaction of the hydroxyl groups with an hydrophobic compound, palmitoyl chloride (PAL-Cl), and to produce cyclo-oligosaccharide-PAL-based nanoemulsions via ultrasound-assisted methodology. This methodology has been previously exploited by Silva et al . [35] and Loureiro et al . [28]. Microspheres of bovine serum albumin (BSA) and silk fibroin were produced by applying ultrasound in a biphasic system consisting of an aqueous protein solution and an organic solvent. The protein microspheres were dispersed in an aqueous media where the protein remains at the interface covering the organic solvent. These nanodevices can improve the solubility of hydrophobic drugs encapsulated in the organic phase of the particles. The modified Cys hold typical fusion properties (melting point of around 50 °C) which it turns powder into a viscous liquid. Taking advantage of this phase-change, nanoemulsions of Cys-PAL and CysPAL/BSA were produced applying ultrasound in a biphasic system consisting of an aqueous solution (BSA/buffer) and an organic phase (Cys-PAL). The overall physicochemical and biological characterizations of developed nanoemulsions were assessed. A comparison with other cyclooligosaccharides, namely βand -CD was performed. MTX was used as a model guest molecule to test the entrapment ability of the developed nanodevices. The efficiency of entrapment and the drug release over time were evaluated by fluorescence microscopy. 2.2. Materials and methods 2.2.1. Materials All compounds were purchased from Sigma-Aldrich or TCI chemicals, and used without further purifications. Column chromatography was made using silica gel 60Å, with particle size of 70-200 µm as stationary phase, and ethyl acetate, chloroform or ethanol as mobile phase for the separation of the components. TLC plates (silica gel 60 F254) were revealed under a UV lamp. 11 Human skin fibroblasts (BJ5ta cell line) (ATCC, CRL-4001) were obtained from American Type Culture Collection (LGC Standards, UK) and all culture media and supplements were purchased from Sigma-Aldrich. T75 flasks and 96-well tissue culture polystyrene plates were obtained from TPP, Switzerland. MTS assay was acquired from Promega, USA. MTX sodium salt was purchased from TevaGuard pharmaceutical. 2.2.2. Isolation of Cyclosophoraose Cys (cyclic β-1,2 glucans) was isolated from Rhizobium leguminosarum biovar viciae VF39 following the procedures found in the literature [36, 37]. MALDI-TOF and NMR techniques were applied to confirm the structure and the average molecular weight. 2.2.3. Synthesis of modified cyclo-oligosaccharides The modified cyclo-oligosaccharides were all synthesized using a similar procedure. Briefly, to a solution of cyclo-oligosaccharide (84-125 mg, 0.027-0.106 mmol) in dry pyridine (4-5 mL) under nitrogen atmosphere, was dropped the PAL-Cl (0.5-1.0 mL, 1.65-3.30 mmol) using a water bath at 50 °C. The mixture was stirred overnight at the same temperature. The solvent was removed by co-evaporation with toluene in a rotary evaporator. The formed solid was further purified by column chromatography to obtain the pure product (0.35-0.48 g, ŋ= 60.4-74.9 %). Cys (84 mg, 0.027 mmol); PAL-Cl (0.5 mL, 1.65 mmol); Dry pyridine (5 mL); 50 °C; overnight. Column chromatography 50 % CHCl3/50 % AcOEt. Beige solid (0.35 g, 0.020 mmol, 74.9 %). Melting point= 48-50 °C (Gallenkamp apparatus). δH (CDCl3): 0.88 (t, J= 6.8 Hz), 1.28 (br s), 1.63 (m), 2.35 (t, J= 7.6 Hz), 3.0-5.5 (br s) ppm. MALDI-TOF 20045 m/z. β-CD (120 mg, 0.106 mmol); PAL-Cl (1 mL, 3.30 mmol); Dry pyridine (4 mL); 50 °C; overnight. Column chromatography CHCl3/EtOH (6:1). Beige solid (0.392 g, 0.064 mmol, 60.4 %). Melting point= 44-46 °C (Gallenkamp apparatus). δH (CDCl3): 0.88 (t, J= 6.8 Hz), 1.25 (br s), 1.62 (m), 2.34 (m), 3.49 (m), 3.97 (m), 4.10 (m), 4.51 (m), 4.95 (m) ppm. γ-CD (125 mg, 0.096 mmol); PAL-Cl (0.9 mL, 2.31 mmol); Dry pyridine (5 mL); 50 °C; overnight. Column chromatography 50 % CHCl3/50 % AcOEt. Beige solid (0.48 g, 0.068 mmol, 70.7 %). Melting point= 49-50 °C (Gallenkamp apparatus). δH (CDCl3): 0.89 (t, J= 6.8 Hz), 1.31 (br s), 1.63 (m), 2.36 (m), 3.50 (m), 3.87 (m), 4.01 (m), 4.14 (m), 4.48 (m), 4.90 (m), 5.03 (m) ppm. 12 2.2.4. Nuclear Magnetic Resonance spectroscopy (NMR) 1H and 13C NMR spectroscopy were performed using a Bruker Avance III 400 (400 MHz for 1H and 100 MHz for 13C). CDCl3, DMSO-d6 or D2O were used as deuterated solvents, using the peak solvent as internal reference. Multiplicities are indicated as: t as triplet, br s as broad singlet and m as multiplet. 2.2.5. Matrix-Assisted Laser Desorption/Ionization Time-of-Flight (MALDI-TOF) MALDI-TOF mass spectra were acquired on a Bruker Autoflex Speed instrument (Bruker Daltonics GmbH) equipped with a 337 nm nitrogen laser. The matrix solution for the measurements was prepared by dissolving a saturated solution of 2,5-dihydroxybenzoinc acid (DHB) in 100 % ethanol. Samples were spotted onto a ground steel target plate (Bruker part nº 209519) and analysed in the linear positive mode using factory-configured instrument parameters suitable for a 10-30 kDa m/z range (ion source 1: 19,5kV; ion source 2: 18,3kV). Time delay between laser pulse and ion extraction was set to 130 ns, and the laser frequency was 25 Hz. 2.2.6. Cyclo-oligosaccharide-palmitoyl conjugate nanoemulsions The nanoemulsions composed of cyclo-oligosaccharide (alone or in the presence of BSA) were produced by application of US. The mixtures containing the cyclo-oligosaccharide (alone or in presence of BSA) (4mL) in phosphate buffer saline (PBS) 10 mM at pH= 7.4 were previously submitted to a pre-heat at 60 °C for 10 min. Then, the mixture was subjected to US using a highintensity ultrasonic horn (20 KHz Sonics & Materials Vibracell CV 33, 3 mm diameter titanium microtip) positioned at the middle of the mixture. The ultrasonication was performed at 60 °C for 3 min with an amplitude of 40 %. Control samples with unmodified oligosaccharides were also performed using the same conditions. 2.2.7. Dynamic Light Scattering (DLS) The mean diameter (nm), polydispersity index (PDI) and zeta-potential (mV) of the nanoemulsions were measured in a Zetasizer Nano ZS (Malvern Instruments) at 25 °C. Prior to DLS measurements, the samples were diluted with PBS at pH 7.4 (for size) and with ultrapure water (for zeta-potential). In both cases, a dilution of 5x was made, followed by a filtration using 0.45 µm filters, (SpartanTM 30 RC, Whatman, GE Healthcare). All measurements were read in triplicate, being the results described as mean ± standard deviation. 13 2.2.8. Nanoparticle Tracking Analysis (NTA) NTA was used to access quantitatively the obtained nanoemulsions. The experiments were performed using a NanoSight NS500 instrument (Salisbury, UK). This system includes a charge coupled device camera that allows visualization and tracking Brownian motion of laser-illuminated particles in suspension. The measurements were made at room temperature and each video sequence was captured over 60 sec. The samples were 10x diluted with water and filtered (Millipore filters with pore size of 0.45 μm) and then injected into the system. 2.2.9. Differential Scanning Calorimetry (DSC) All measurements were conducted on a power-compensated DSC instrument (DSC 6000 Perkin Elmer) with a nitrogen flux of 20 mL/min, using stainless steel capsules in the temperature range of 20-100 °C (heating rate: 10 °C/min, powder sample weight: 3-7 mg). The DSC device was calibrated using indium and zinc, both of high purity. The samples were stored at selected levels of humidity (relative humidity of 45 %) and temperature (20–22 °C) for 24 h prior to the analyses and each sample was measured at least two times, in order to validate the results. 2.2.10. Scanning Transmission Electron Microscopy (STEM) STEM was performed with a NOVA Nano SEM 200 FEI microscope. The carbon coated copper grids (400 meshes, 3 mm diameter) were placed in contact with the nanoemulsions and the excess of solvent was instantly absorbed by a filter paper. Another set of samples was prepared by the same way with the addition of a final staining with uranyl acetate. For this, the grids were placed on the top of a drop of uranyl acetate (0.2 % v/w) for 5 min. The excess of the liquid was removed as described above and the grids were washed with water and dried at room temperature. 2.2.11. Molecular Dynamics Simulations The simulations were performed with the software package GROMACS, using Martini force-field [38]. The system size was chosen according to the minimum image convention taking into account a cut-off of 1.2 nm. The bonds lengths were con-strained with LINCS [39]. Non-bonded interactions were calculated using a twin-range method, with shortand long-range cut-offs of 0.9–1.2 nm, respectively. Neighbour searching was carried out up to 1.2 nm and updated every ten steps. A time step of integration of 5 fs was used. A reaction field correction for the electrostatic interactions was applied using a dielectric constant of 15. Pressure control was implemented using the 20 when BSA was present in the nanoemulsions. The auto aggregation properties of Cys-PAL can justify the bigger sizes observed by DLS (Figure 2.7). Figure 2.5. Molecular dynamic simulations of Cys-PAL auto aggregation: I) cyclosophoraose (red) surrounded by covalently bonded palmitic chains (green), before aggregation, II) 15 units of CysPAL self-aggregated; and BSA-Cys-PAL aggregation: III) BSA-Cys-PAL before nanoemulsion simulation, showing 6 units of BSA (in red) and 12 units of Cys-PAL (in cyan), IV) BSA-Cys-PAL after nanoemulsion simulation (one unit of Cys-PAL surrounded by several units of BSA). 2.3.3. Microscopic observation The morphology and structure of the developed nanoemulsions (Cys-PAL, β-CD-PAL, γ-CD-PAL, BSA-Cys-PAL, BSA-β-CD-PAL and BSA-γ-CD-PAL) were investigated by STEM to assess the form and size of emulsion droplets (Figure 2.6). All the nanoemulsions present homogeneous and regular aspect with particle size between 100-200 nm, with the exception of β-CD-PAL and BSA-βCD-PAL (data not shown). These samples present much lower particle size compared to the others and by eye visualization a low stability against flocculation and coalescence was observed. The amount of sugar units played here a limiting role by hindering the nanoemulsions formation when β-CD (7 sugar units) was used. For further experiments, we excluded these samples from the studies. The results also demonstrate that Cys-PAL, γ-CD-PAL, BSA-Cys-PAL and BSA-γ-CD-PAL were successfully converted into spherical and smooth nanoemulsions by using high intensity ultrasound. This morphology is expected to offer high encapsulation potential as well as high 21 release performance and drug protection, providing the minimum contact with the aqueous environment [35]. Figure 2.6. STEM images of nanoemulsions: A) Cys-PAL (5 mg/mL) (stained with uranyl acetate); B) γ-CD-PAL (5mg/mL); C) β-CD-PAL (5 mg/mL); D) BSA (10 mg/mL) + Cys-PAL (5 mg/mL) (stained with uranyl acetate); E) BSA (10 mg/mL) + γ-CD-PAL (5 mg/mL) (stained with uranyl acetate). 2.3.4. Nanoemulsions stability over time The relative homogeneity of the developed nanoemulsions was further investigated by DLS analysis. Figure 2.7 presents the mean size and PDI of samples at the first day of production and after 60 days of storage at 4 °C. All the formulations presented narrow sized distribution between 100-200 nm in diameter with low PDI (0.1-0.3) remaining stable in nanoemulsions state at least for 60 days of storage at 4 °C. Even though the values of size slightly changed along time, the PDI values remained highly stable, which is a good indicator for future applications. The zeta-potential is an important parameter in nanoemulsions characterization which allows to predict their stability over time. Long term stability of produced nanoemulsions has been verified, which was in accordance with the zeta-potential data obtained. The nanoparticles were highly negatively charged indicating high repulsive forces with low probability of agglomeration. Cys-PAL and γ-CD-PAL presented 22 similar zeta-potential values (≈ -40 mV) while BSA-Cys-PAL and BSA-γ-PAL showed zeta-values of around - 20 mV attributed to the complexation of the cyclo-oligosaccharides with the protein. After storage for 60 days at 4 °C the surface charge of nanodroplets remained highly negative rendering high stability to the developed emulsions. Figure 2.7. Stability (size and PDI) of nanoemulsions during storage at 4 °C for 60 days (values are the mean ± SD of 3 independent experiments). The formulations containing BSA in their composition presented the lowest size values and the highest PDI. The size observed might be explained by the presence of the protein which is believed to form a shell with hydrophobic characteristics near the oily-like cyclo-oligosaccharides and hydrophilic properties near water. The PDI observed might be related with the yield of nanoemulsions formation. While samples containing only cyclo-oligosaccharides present nanoemulsions formation yields very close to 100 %, samples containing BSA might contain protein that do not contribute to the nanodroplets formation, increasing the PDI measured by DLS. These results demonstrate that the cyclo-oligosaccharides-PAL seem to be an effective carbohydratebased emulsifier. The high emulsion stability related with the low flocculation and/or coalescence obtained suggests many possibilities of uses as molecular devices for the development of cosmetic and/or pharmaceutic products with a long shelf life. 23 The size diameter of nanodroplets was also evaluated by NTA and the results were compared with the values obtained by DLS (Table 2.1). NTA, besides giving information about nanoemulsions concentration allowed us to eliminate the influence of small amounts of reagents that did not participate in the nanodroplets formation. The mean size diameter data obtained by both techniques follow the same tendency. Nanoemulsions containing only the modified cyclooligosaccharide show values of size diameter of 150-170 nm on both techniques. The same behaviour was observed for the samples containing BSA in their composition which demonstrate higher mean size values observed by NTA. Both techniques reveal only one peak, indicating the monodisperse behaviour. NTA analysis allowed to quantify the nanodroplets concentration on each sample. Similar concentration of particles per mL were measured for both Cys and γ-CD modified cyclo-oligosaccharides. From all the samples tested, the nanoemulsions containing BSA/Cys-PAL showed the highest concentration of particles after emulsification. The lowest particle concentration was observed for samples containing BSA-γ-CD-PAL. Table 2.1. Mean size diameter (nm) and size distribution measured by DLS and NTA (values are the mean ± SD of 3 independent experiments) 2.3.5. Cytotoxicity of nanoemulsions The effect of Cys-PAL (5 mg/mL), γ-CD-PAL (5 mg/mL), BSA (10 mg/mL)-Cys-PAL (5 mg/mL) and BSA (10 mg/mL)-γ-CD-PAL (5 mg/mL) on the viability of human BJ5ta cells was evaluated by the MTS assay. Figure 2.8 shows BJ5ta cell viability in the presence of different concentrations of nanoemulsions and of DMSO at 30 % (v/v), as negative control. The results demonstrated that the nanoemulsions containing the modified cyclo-oligosaccharides do not induce cytotoxicity in cells after 72 h of contact even at the highest concentration (1 mg/mL). Nanoemulsions containing BSA induced some cytotoxicity, but only for the highest concentration tested (1 mg/mL). These results DLS NTA Nanoemulsions Z-average (nm) PDI Mean (nm) SD Conc. particles (particles/mL) Cys-PAL (5mg/mL) 144.1 ± 1.05 0.093 ± 0.01 151.5 ± 0.71 52.5 ± 3.5 6.29 ± 6.3 (E9) γ-CD-PAL (5mg/mL) 162.9 ± 1.19 0.059 ± 0.01 157 ± 14.1 65 ± 9.9 8.45 ± 2.9 (E9) BSA (10mg/mL)-Cys-PAL (5mg/mL) 137.4 ± 0.23 0.201 ± 0.01 173.5 ± 3.5 52 ± 4.2 11.3 ± 0.02 (E9) BSA (10mg/mL)-γ-CD-PAL (5mg/mL) 131.2 ± 1.97 0.240 ± 0.01 182 ± 7.1 47.5 ± 9.2 1.91 ± 1.1 (E9) 24 indicate that modified cyclo-oligosaccharides have potential as biocompatible encapsulating emulsifiers for further applications in pharmaceuticals, foods or cosmetics. Figure 2.8. BJ5ta cell line viability after 72 h of contact with different concentrations (from 0.0625 mg/mL up to 1 mg/mL) of the four different nanoemulsion formulations, compared with cells (negative control) and cells incubated with 30 % (v/v) of DMSO (death control), determined by MTS assay. Values are the mean ± SD of 2 independent experiments. 2.3.6. Stability of nanoemulsions in the presence of lipase The stability of nanoemulsions in vivo is governed by several factors namely, pH, surface charge, concentration, salts and the presence of hydrolytic enzymes [41]. We have tested the stability of the developed nanoemulsions in the presence of lipase from Thermomyces lanuginosus (activity: 407 μmol/mL/min). The enzyme remained active during all the incubation process time (lipase activity measured using the release of p -nitrophenol, from p -nitrophenyl butyrate, by absorbance at 400 nm) [42]. The stability of nanoemulsions was evaluated by particle size measurement at different time points of incubation and by quantification of palmitic acid concentration after hydrolysis. This quantification was performed separating the products of hydrolysis by liquid-liquid extraction with CHCl3, followed by NMR analysis (the isolated product was firstly dissolved in DMSOd6 to detect the presence of palmitic acid; then, to the same sample was added CDCl3, to calculate the percentage of palmitic acid cleavage from the nanoemulsions). Table 2.2 presents the mean size diameter values of nanodroplets before and after enzymatic incubation for 48 h. It is noteworthy that the size of all the samples is influenced by the same amount of catalyst in solution. It can be depicted that nanoemulsions made-up of modified cyclo-oligosaccharides were not 25 disrupted by enzymatic action. This was confirmed by the vestigial amounts of palmitic acid (< 1%) detected by NMR (data not shown). Lipase catalyst was able to disrupt the nanoemulsions containing BSA as it can be observed by the results of size diameter. After incubation for 48 h the mean size and PDI increased indicating nanoemulsions destabilization and disruption. The characteristics of these nanoemulsions make them instable per se to time storage, as it can be depicted by the higher values of PDI obtained (Figure 2.7). The NMR analysis corroborate these findings. The percentages of palmitic acid cleavage were in the range of 20-55.7 % when lipase was applied (data not shown). The self-aggregation ability of Cys-PAL gave rise to more robust nanoemulsions with high stability against hydrolysis. Table 2.2. Particle size before (t= 0 h) and after 48 h in the presence of lipase (values are the mean ± SD of 3 independent experiments) toh t48h Before incubation Lipase Size (nm) PDI Size (nm) PDI Cys-PAL (5mg/mL) 146.0 ± 1.2 0.134 ± 0.02 152.2 ± 8.7 0.118 ± 0.02 BSA (10mg/mL)-Cys-PAL (5mg/mL) 106.0 ± 0.6 0.385 ± 0.06 134.4 ± 7.1 0.283 ± 0.00 -CD-PAL (5mg/mL) 172.7 ± 2.1 0.087 ± 0.02 144.5 ± 13 0.286 ± 0.11 BSA (10mg/mL)--CD-PAL (5mg/mL) 129.7 ± 3.6 0.289 ± 0.00 128.7 ± 12 0.437 ± 0.08 2.3.7. Encapsulation and release of methotrexate As previously reported by Kim et al. [30] we also studied herein the dual properties of modified cyclo-oligosaccharides, namely encapsulation and emulsifying ability using a model compound, MTX. In its native form MTX is an hydrophobic-like drug acquiring hydrophilic character when solubilized at high pH [43, 44]. Taking advantage of the phase-change phenomena (from solid to oily-like liquid) we could predict an efficient encapsulation of this drug at the moment of nanodroplets formation. The encapsulation take place by mixing 10 mg/mL of MTX, at the water phase, with the modified cyclooligosaccharides. The solution was heated at 60 °C and after ultrasonication the samples showed a milky-like emulsion with yellow colour indicating the encapsulation and/or emulsification of MTX into the cyclo-oligosaccharides [30]. The available ester groups of the cyclo-oligosaccharides are expected to interact with the hydrophilic groups of the drug, whereas the aromatic groups would 26 interact more easily with the hydrophobic palmitoyl chain introduced by synthesis with the cyclooligosaccharides. By NMR spectroscopy it was possible to evaluate the molecular interactions between MTX and the cyclo-oligosaccharide nanoemulsions (Figure 2.9). The spectra of the commercial MTX evidence the aromatic peaks at δH 8.59, 7.72 and 6.89 ppm, which undergo chemical shifts when MTX was entrapped within nanoemulsions to δH 8.54, 7.66 and 6.79 ppm. The aliphatic peaks showed up at δH 4.79, 4.32, 3.18 and 2.32-2.02 ppm in the starting MTX, while in the nanodevices they appeared at δH 4.69, 4.32, 3.13 and 2.35-2.03 ppm. Considering the small chemical shifts observed, especially of the lipophilic part of the MTX (aromatic peaks), it was possible to predict that these groups preferentially interact with the hydrophobic palmitic chains of the cyclooligosaccharides leading to high MTX encapsulation yield. The small and narrow sized nanoemulsions containing MTX evaluated by DLS corroborate this assumption (Table 2.3). As we observed by NMR, MTX present in the solution was entrapped into nanoemulsions. We may speculate that this entrapment was near 100 %, since their physicochemical properties resemble the empty nanodevices and no free MTX was detected by DLS, and thus no additional step of separation was needed. Figure 2.9. 1H NMR (in D2O) of MTX free (in blue) and encapsulated in -CD-PAL nanoemulsions (in red). Afterwards the release profile of the drug was traced by dialysis, until the amount of MTX released reached a plateau (Figure 2.10). Analysing the data obtained, we can observe a burst release in 27 the first 24 h of dialysis. It was also evident the release of almost 90 % of the drug after 7 days of dialysis, for all the cases studied. The observations suggest that the intrinsic entrapment ability of modified cyclo-oligosaccharides functioned herein as drug encapsulators and a similar behaviour has been observed for all the cyclo-oligosaccharides tested. The encapsulation and/or entrapment ability was not, in this case, size dependent, but related with the intrinsic phase-change properties of the cyclo-oligosaccharides and their ability to interact with amphiphilic-like compounds. Figure 2.10. Release of MTX from the nanoemulsions over time (values are the mean ± SD of 3 independent experiments). In order to evaluate the structural stability of nanoemulsions after drug release, the size and PDI of the nanodroplets were evaluated by DLS. These parameters were measured immediately after nanoemulsions formation, when MTX encapsulation occurs, and after MTX release by dialysis. The nanoemulsions containing MTX revealed to be monodisperse with small and narrow sizes (Table 2.3 and Figure 2.11). After MTX release, the nanoemulsions maintained their mean size and PDI confirming their stability even after drug release. Table 2.3. Size, PDI and surface charge of nanoemulsions, before and after MTX release Before MTX release After MTX release Size (nm) PDI Zeta-potential (mV) Size (nm) PDI Zeta-potential (mV) Cys-PAL 137.4 ± 4.7 0.089 ± 0.011 -41.3 ± 6.4 131.0 ± 4.6 0.081 ± 0.024 -34.2 ± 1.7 BSA-Cys-PAL 131.9 ± 6.5 0.186 ± 0.006 -21.7 ± 0.6 127.4 ± 10.3 0.205 ± 0.035 -23.0 ± 3.0 γ-CD-PAL 140.2 ± 13.0 0.094 ± 0.029 -36.3 ± 2.9 147.2 ± 3.22 0.087 ± 0.018 -27.7 ± 4.9 BSA-γ-CD-PAL 128.0 ± 8.4 0.321 ± 0.008 -20.9 ± 2.0 128.1 ± 2.31 0.269 ± 0.014 -21.1 ± 1.9 28 Figure 2.11. STEM microphotographs of nanoemulsions with MTX encapsulated: A) Cys-PAL-MTX before MTX release; B) Cys-PAL-MTX after MTX release; C) γ-CD-MTX before MTX release; D) γCD-MTX after MTX release. 2.4. Conclusions Herein we report for the first time the synthesis of hydrophobic cyclo-oligosaccharides by fully substitution of the hydroxyl groups by an hydrophobic compound, a palmitic chain. The emulsifier properties of these hydrophobic cyclo-oligosaccharides made possible the development of stable nanoemulsions, as shown by DLS, NTA and STEM evaluations. Narrow sized, monodisperse and nontoxic nanoemulsions with great potential for amphiphilic drugs encapsulation were developed. The entrapment and delivery ability of these new hydrophobic compounds was evaluated by encapsulation of MTX. The results indicated that the intrinsic properties of modified cyclooligosaccharides were able to undergo emulsification in an hydrophobic environment due to their ability to acquire oily-like behaviour when heated. On the other hand, the water-soluble part of cyclooligosaccharides showed to be flexible and interact with the hydrophilic portion of the drug. The release of the drug revealed a sustained trend for 1 week of dialysis. Therefore, a step forward the fabrication of nanodevices for amphiphilic drugs encapsulation and delivery for cosmetic and pharmacological purposes was accomplished. A B C D 29 Chapter III Ultrasound-assisted biosynthesis of novel methotrexate-conjugates 36 delay between laser pulse and ion extraction was set to 130 ns, and the laser frequency was 25 Hz. 3.2.5. Electrospray Ionization (ESI) Electro spray ionization was performed in a mass detector Thermo Finnigan LxQ (Linear Ion Trap). Mass detector susceptible of analysis in full scan mode, SIM and MS/MS with positive ionization. Mass spectra range was between 50 and 2000. Capillary voltage was used as 29 V. 3.3. Results and discussion 3.3.1. Methotrexate-acylglycerol conjugates The catalytic activity of both, immobilized CALB and free form lipase from Thermomyces lanuginosus , against different triacylglycerols (carbon chain length between C3 and C17) was investigated using a water and an ultrasonic bath. The immobilized form was chosen for comparison due to its higher thermal stability and catalytic activity comparing with the free form. Moreover, the purification steps can be simplified using this enzyme form since a simple paper filtration removes the enzyme from the reactional medium [56]. The transesterification reactions were conducted herein using a green methodology, with water as solely solvent. MTX is poorly soluble in almost all the solvents, namely DMSO which does not allow the fully solubilization of the reactants disabling the hydrolytic activity of the esterases. The addition of the triacylglycerol to the MTX solution, formed as expected, an immiscible biphasic system, being the powder enzyme suspended at the soluble phase. After reaction we observed that for longer triacylglycerols, namely triolein (C17:1) and glycerol tristearate (C17), immobilized CALB did not presented any hydrolytic activity (Table 3.1). After removal of the reactional mixture from the US, followed by a liquid-liquid extraction of the unreacted triacylglycerols, only free MTX was recovered in the aqueous phase. This result was expectable since recently, Chiplunkar et al. reported that immobilized CALB did not hydrolyse triolein in organic medium [62]. For the short triacylglycerols used, glycerol tributyrate (C3), glycerol trivalerate (C4) and glycerol trihexanoate (C5), after 30 min under ultrasonication, we observed the disappearance of the biphasic system formed initially, and the precipitation of a solid that turns into soluble when more water was added to the reactional flask. The different catalytic behaviour of immobilized CALB depending on the size of the triacylglycerols may be justified by the poor mobility of the immobilized enzyme form hindering the catalysis of longer substrates. Considering this enzyme drawback, we 37 replaced the immobilized CALB by a liquid lipase from Thermomyces lanuginosus in the free form and use it without the addition of any other solvent. After reaction, it was possible to observe the hydrolysis of the longer triacylglycerols and the formation of the MTX-acylglycerol conjugates. Our data corroborate previous findings which highlighted an high activity of this lipase for long triacylglycerols, such as triolein [63]. The reactional scheme proposed for the conjugation of MTX with triacylglycerols is presented in Scheme 3.1. We proposed that the enzyme starts the hydrolysis of the triacylglycerol, forming an intermediate in the active site, with the final release of a fatty acid chain. This reaction occurs twice, which in turns the glycerol unit with only one side chain available. Then, the MTX binds to the active site of the enzyme, by the carboxylic group, forming an ester intermediate. The glycerol can undergo a nucleophilic attack, which led to the formation of the final conjugates. 1H NMR data did not allow us to ensure which carbonyl group was involved on the reaction, thus we therefore propose the -position carbonyl group, since it is the most reactive carboxyl group of the glutamic acid portion [64]. Scheme 3.1. Reactional scheme of methotrexate-acylglycerols conjugates synthesis. Compound 1, methotrexate; Compounds 2a-2c were produced using US in the presence of immobilized CALB. Compounds 2d and 2e were produced using US and lipase from Thermomyces lanuginosus . 2a was produced using glycerol tributyrate, 2b using glycerol trivalerate, 2c using glycerol trihexanoate, 2d using glycerol tristearate and 2e using triolein. From 1H NMR data we can infer that both lipases hydrolysed two of the three fatty acid chains of the triacylglycerols. Works reported have been recognising that some lipases can own 1,3selectivities against triacylglycerols [65]. Taking into account this enzyme property and considering that only one carbon chain remained attached to the glycerol moiety, we propose that the carbon chain remains in the second position (Scheme 3.1). 38 The use of ultrasound to assist the enzymatic reactions, due to emulsifying events, allowed to obtain higher conversion rates comparing with the results using a water bath. As can be depicted in Table 3.1, high yields of crude products were obtained as solids or oils after freeze-drying. The role of ultrasound on enzymatic reactions enhancement has been described as due to several reasons. One of them are the mechanical effects of ultrasound which promote mass transfer from the bulk solution to the enzyme. The collision of substrate molecules and the enzyme promote the reaction rate increment and the miscibility of the two reactional initial phases. Another reason relies on the steady cavitation corrosion of ultrasound which might cause the enzyme molecule or cell granulation around it to be shared by microstreaming [55], improving the mass transfer inside or outside of the enzyme. The final reason is related with the enzyme structure, which in the presence of ultrasound, become more flexible, and thus, may shift into its active configuration [58]. Considering that different enzymes own different stereo-configurations, for the same ultrasonic parameters adopted, different enzyme behaviour was observed for both immobilized CALB and free liquid lipase. This distinct behaviour is determined by several factors including the bulk of ultrasonic energy, the adaptability of the enzyme, and the micro-conditions around the enzyme. Analysing the entries 1 to 6 of Table 3.1, it is possible to observe that lipase from Thermomyces lanuginosus was able to hydrolyse the short triacylglycerols in a low extent. The reaction yields calculated by 1H NMR data were low, revealing that this enzyme was only able to hydrolyse part of the carbon chains available. Only a vestigial amount of product was detected by the unfolding of the triacylglycerols peaks on 1H NMR. High conversion yields were obtained on the conjugation of methotrexate with longer acylglycerols using the free liquid enzyme form in the presence of ultrasound. In both cases only vestigial amounts of product were detected after incubation in the water bath. Herein, the use of ultrasound to assist the synthesis reactions was again crucial for an effective conjugation. 39 Table 3.1. Experimental conditions and conversion yields for methotrexate-acylglycerol conjugates using immobilized CALB and free lipase from Thermomyces lanuginosus (TL) (results with standard deviation from at least 2 independent experiments) Entry Triacylglycerol Carbon Chain Length Reaction Conditions Yield (%)** Immobilized CALB Lipase TL 1 Glycerol Tributyrate C3 WB (40°C, 30 min) --* --* 2 US (5min ON/5min OFF; 30 min) 61.4 ± 4.0 --* 3 Glycerol Trivalerate C4 WB (40°C, 30 min) --* --* 4 US (5min ON/5min OFF; 30 min) 64.1 ± 0.4 --* 5 Glycerol Hexanoate C5 WB (40°C, 30 min) 45.8 ± 1.7 --* 6 US (5min ON/5min OFF; 30 min) 58.0 ± 7.6 --* 7 Triolein C17:1 WB (40°C, 30 min) --a --* 8 US (5min ON/5min OFF; 30 min) --a 62.6 ± 4.3 9 Glycerol Tristearate C17 WB (40°C, 30 min) --a -- a 10 US (5min ON/5min OFF; 30 min) --a 63.7 ± 6.4 a no reaction occurred; *only vestigial amount of conjugate product was detected by 1H NMR; ** the yield was calculated based on the initial number of moles and the moles of the product. To confirm the conjugation, we also evaluated the melting point of the conjugates produced. The melting point of methotrexate disodium salt is reported to be between 212-216 °C. For the conjugates produced we observed a different behaviour, they do not display a melting point, starting to decompose at around 250 °C. The conjugate composed by a hexanoate carbon chain was the most stable, starting to decompose at temperatures above 280 °C. The conjugate composed by a valerate chain (C4) was fully decomposed at 250 °C. The melting behaviour was not evaluated for the conjugates with longer carbon chains since they were isolated as a mixture of oil/solids. 1H NMR spectra of conjugates 2a-e (see Scheme 3.1) showed a unique set of peaks of MTX protons, meaning that the transesterification reaction occurred in only one of the carbonyl groups (Figure 3.1). The hydrolysed chains were eliminated during the liquid-liquid extraction (longer triacylglycerols) or during the freeze-drying process, in the case of short triacylglycerols. Figure 3.1 shows the 1H NMR of methotrexate-hexanoate conjugate and of free MTX. Significant chemical shifts were observed on the protons of methotrexate, mainly on the protons of the glutamic portion of the MTX. The amine proton signals remained untouched confirming that the conjugation occurred at one of the carboxylic groups. The amide proton (g) suffered a chemical shift from δH 8.14 to 7.85 ppm. Proton h appeared at δH 4.08 instead of 3.85 ppm. The aliphatic protons of the glutamic portion unfolded in two: proton j appeared at a higher chemical shift δH 2.12 and 2.06 instead of 1.89 ppm; protons i stayed in a 40 similar place: from δH 1.79 to 1.82 ppm. Glycerol moiety was observed under the HDO peak, between δH 3.25-3.50 ppm. Figure 3.1. 1H NMR spectra of A) free MTX and B) methotrexate-hexanoate conjugate (DMSO-d6). Electrospray ionization (ESI) and MALDI-TOF techniques were also assessed to confirm the formation of the conjugates (Table 3.2). The masses obtained either by ESI or MALDI-TOF were very similar to the theoretical values calculated, confirming the formation of conjugates between MTX and the acylglycerols. Table 3.2. Values of theoretical and experimental mass obtained by ESI and MALDI-TOF Theoretical mass Mass obtained by ESI Mass obtained by MALDI-TOF MTX-butyrate 620.6 666.6 (620.6 + 2Na+) 625.4 MTX-valerate 634.2 625.5 631.4 MTX-hexanoate 648.6 659.2 671.1 (648.6 + Na+) MTX-oleate 814.9 823.2 815.7 MTX-stearate 816.9 833.6 816.5 41 3.3.2. Methotrexate-cyclodextrin conjugates Cyclodextrins (CD) are cyclic sugars used as devices for the solubilization of hydrophobic compounds by formation of inclusion-complexes in aqueous medium. Considering their non-toxic properties, the covalent bonding of molecules to these devices can also be assessed as an effective strategy to produce pro-drugs [66]. We studied the biosynthesis of methotrexate-cyclodextrin conjugates using a similar methodology described previously for the triacylglycerols. The reactions were carried out without addition of solvents, using the free liquid lipase from Thermomyces lanuginosus as reactional medium to ensure the solubilisation of all reactants. Immobilized CALB was also tested, however due to its difficulty to access and accommodate longer substrates, an extremely low yield (around ŋ= 1 %) was obtained after reaction. The effect of ultrasound on the formation of the conjugates was also evaluated. The reactions were performed using an US bath, for 2 h with duty cycles of ultrasound of 15 minutes ON followed by 5 minutes OFF. The same reaction time was used for the reactions performed in a water bath (WB). Comparing with MTX-acylglycerol conjugates, the total time of the reactions was extended considering the size of the cyclodextrin substrates. As previously reported for MTX-acylglycerol conjugates, the use of US led to high yields of MTX-CD conjugates, and a higher number of MTX units linked to the macromolecules. The amount of MTX equivalents used was also incremented but the findings revealed that it does not led to higher degrees of modification or conversion yield. In Table 3.3 are depicted the results obtained after MTX-CD conjugation. Observing the data, we found that the degree of modification of the cyclodextrins was size dependent and greatly influenced by stereochemical impediments. The larger the size of the cyclodextrin molecule, higher was the number of MTX molecules bound. The smaller cyclodextrin used, α-CD composed by 6 glucose units, was only conjugated to one molecule of MTX, while the longer, γ-CD with 8 glucose units, was conjugated to a maximum of 3 units of MTX. 42 Table 3.3. Reaction conditions and conversion yields obtained after conjugation of methotrexate with cyclodextrins using liquid lipase from Thermomyces lanuginosus (TL) (results with standard deviation from at least 2 independent experiments) Entry Cyclodextrin Reaction Conditions Lipase TL Max. degree of modification Conversion Yield 1 α-CD WB (40 °C, 2h) 1 8.9 ± 3.4 % 2 α-CD US (15 min ON/5 min OFF) x 6 cycles 1 16.8 ± 1.8 % 3 β-CD WB (40 °C, 2h) 1 14.8 ± 3.4 % 4 β-CD US (15 min ON/5 min OFF) x 6 cycles 2 19.0 ± 0.4 % 5 γ-CD WB (40 °C, 2h) 1 3.0 ± 4.2 % 6 γ-CD US (15 min ON/5 min OFF) x 6 cycles 3 22.0 ± 2.8 % By 1H NMR spectroscopy, it was possible to observe significant differences between the spectra of the starting reactants and of the final conjugates. The decrease and change in the chemical shift of the OH protons of the cyclodextrins were the most evident alteration. As starting materials, the OH peaks of the CD appeared between δH 5.75 and 4.43 ppm. When conjugated, these peaks were observed between δH 6.58 and 4.45 ppm. Differences in the glycosyl peaks of the CD, mainly in the chemicals shifts and pattern, were also detected. The pattern of the MTX in all the cyclodextrins remains very similar in all the conjugates obtained. 43 Figure 3.2. 1H NMR of MTX-CD (α, β and γ)-conjugates in DMSO-d6 using lipase from Thermomyces lanuginosus in the US. MALDI-TOF analysis allowed us to calculate the degree of modification of the cyclodextrins with methotrexate (Figure 3.3). We confirmed a direct correlation between the size of the cyclodextrin and the degree of modification. α-CD presented the lowest modification, with only one MTX unit ( m/z = 1478) (Figure 3.3A), β-CD was conjugated with 2 units of MTX ( m/z = 2128) (Figure 3.3B), and γ-CD was conjugated to 3 MTX units ( m/z = 2692) (Figure 3.3C). Glycosyl cleavage can be observed in the MALDI-TOF spectra, with values arround m/z 170. Based on the 1H NMR and MALDI-TOF data we propose in Figure 3.3 the structure of the final MTX-CD conjugates. 44 Figure 3.3. MALDI-TOF of MTX-CD conjugates: A) α-CD conjugated with 1 MTX unit, B) β-CD conjugated with 2 MTX units, C) γ-CD conjugated with 3 MTX units; the upper image shows the proposed MTX-CD conjugates; all conjugates were obtained after reaction using lipase from Thermomyces lanuginosus in ultrasound. 3.4. Conclusions In the present study we developed new methotrexate-acylglycerols and methotrexate-cyclodextrin (α, β and -CD) conjugates via enzymatic transesterification or esterification reaction assisted by ultrasound. We verified that ultrasound played a crucial role on the final conversion yields and degree of modification. The ultrasonic system was not only advantageous over the traditional water bath (WB) in terms of reaction rates, but was also responsible for the products purity and selectivity, namely for the isolation of cyclodextrins with high amount of MTX units conjugated. The therapeutic association of non-toxic triacylglycerols and cyclodextrins, are herein presented as promising therapeutic compounds since they may prevent the development of transport resistance of the drug, which is often observed during the clinical use of methotrexate. 45 Chapter IV -Chymotrypsin catalyses the synthesis of methotrexate oligomers 52 oligomers), and a maximum number of 27,000 generations. Additionally, the binding site preference and interactions were identified through the AutoDock tools. We analysed docking results looking at the binding energy and the Ligand Efficiency (LE), which measures the binding energy per ligand heavy-atom (kcal/HA) to understand how the increase in the oligomer size contributes to the binding. 4.3. Results and discussion 4.3.1. -Chymotrypsin-catalysed synthesis of oligomeric MTX The use of proteases as polymerization tools for the synthesis of peptides is being explored by several researchers [75, 76]. Considering the chemical structure of methotrexate, composed by two carboxylic acids and two primary aromatic amines, we considered it as a promising template to evaluate the polymerase activity of -chymotrypsin (Scheme 4.1). Moreover, the production of a polymeric drug would represent a promising strategy for the pharmaceutical industry regarding its increased interest in the commercialization of these type of drugs [96]. As an amide prodrug, the oligomeric MTX is expected to have higher stability in the human body than the monomeric structure. Moreover, the amide linkages of the oligomeric units, being more prone to hydrolysis by enzymatic action, would improve drug metabolism and decrease dosage administration [97, 98]. The improved metabolism is expected to result in a less toxic drug with lower side effects than the commercial monomeric specie [97, 98]. Methotrexate di-sodium salt was used as starting material to ensure water solubility and provide hydrophilic character to the produced oligomers. Scheme 4.1. Reactional scheme for the oligomerization of methotrexate. α-Chymotrypsin was chosen for the MTX oligomerization based on its well-known ability to polymerize peptides [76], and its high selectivity for the hydrolysis of aromatic amino-acids, such 53 as tyrosine, phenylalanine, and tryptophan [99]. The reactional conditions, namely temperature, pH, and time, were optimized to achieve the maximum oligomerization degree (Table 4.1). Table 4.1. Reactional conditions tested for the -chymotrypsin-assisted synthesis of methotrexate oligomers Entry Temperature pH Time[a] Methotrexate oligomers[b,c] 1 RT 7 1 week 1 2 50 °C 7 1 week 1 3 50 °C 7.8 1 week 1 4 50 °C 9.5 48 h 3 5 50 °C 9.5 72 h 3 6 50 °C 9.5 96 h 5 7 50 °C 9.5 1 week 6 8 50 °C 9.5 2 weeks 6 9 50 °C 9.5 1 week 1[d] [a]Maximum time tested. [b]Detected by MALDI-TOF, where 1 unit corresponds to the monomer. [c]Maximum oligomeric units detected. [d]Control reaction without enzyme. Most of the proteases have an optimum working pH between 7.0-9.0. Based on this feature, on the low basicity of the aromatic amines and on the lack of reactivity of the pteridine ring, the reactions were conducted under basic medium. Similarly to the previously observed by Qin and coworkers [76] for the synthesis of –(KL)x– peptides catalysed by -chymotrypsin, we confirmed that no oligomers were produced when the reactions were carried out at neutral pH (Table 4.1, entries 1-3). For an efficient reaction synthesis, it was critical to perform the reactions at alkaline pH, using NaOH as base ( 9.5, Table 4.1, entries 4-8).The best reactional outcomes were obtained herein when performing the reactions at 50 °C, which is in accordance with the optimum temperature postulated by other authors for the same protein [100]. During reaction, samples were withdrawn, and continuously monitored by MALDI-TOF, until 2 weeks of reaction (Table 4.1, entries 4-8). The best reactional outcome was achieved after 1 week of reaction with the formation of oligomers composed by 6 monomeric units (DPmax). Longer reactional times led to the recovery of similar-sized MTX oligomers, however in lower amount. The DP average calculated (1.5) indicates that the most abundant specie in the reactional medium was the methotrexate dimer, confirmed by MALDI-TOF and ESI (Table 4.2). After one week of reaction, the enzyme converted 33 % of the monomer molecules into oligomers, while the reaction carried 54 out in the absence of enzyme did not lead to the formation of any oligomeric species (Table 4.1, entry 9 and Table 4.2). Table 4.2. Conversion rate (%), average and maximum degree of polymerization (DPavg and DPmax) analysed by MALDI-TOF and electrospray ionization (ESI), for the oligomerization of methotrexate, with and without -chymotrypsin With -chymotrypsin Control (without enzyme) Conversion rate[a] 33.4 % 0 % DPavg[b] 1.5 1 DPmax[b] 6 1 ESI 975.4 g/mol[c] 453.28 g/mol[d] [a]Calculated by the Carothers equation. [b]Calculated by MALDI-TOF. [c]Most abundant specie, corresponding to the dimer. MW dimer= 956.8, where in ESI: 975= 956 + Na+ - 4H+. [d]Monomer MW= 454 g/mol or 498 g/mol (di-sodium form), where in ESI: 453= 454 - 1H+. In order to ensure an efficient oligomerization, it was imperative to infer the maintenance of the basic pH during reaction. For this, the pH was continuously monitored and after 1 week of reaction, one observed a pH decrease from 9.5 to 8, and to 7.5, after 2 weeks. We may assume that as the reaction occurs, the NaOH was consumed by the low basic amines in the pteridine ring, which were deprotonated, leading to a pH decrease. The addition of NaOH was crucial to increase the amines reactivity and proceed with the oligomerization. It was noteworthy that the activity of α-chymotrypsin remained almost unaltered during the first week of reaction (activity of 0.61 U/mg), decreasing only 15 % after 2 weeks (activity of 0.52 U/mg). The high stability of the catalyst ensured the efficient catalysis for longer periods of incubation. 4.3.2. Synthesis of dimeric MTX – proposed mechanism Methotrexate, mainly composed by aromatic rings, is presented herein as a suitable substrate for α-chymotrypsin. Based on the data obtained, we have proposed a mechanism for the synthesis of a dimeric unit of methotrexate catalysed by α-chymotrypsin (Figure 4.1). The reaction is predicted to start by the nucleophilic attack of the OH terminal group of the serine, in the catalytic triad, to one of the carboxylic groups of MTX (step 1), leading to the formation of the tetrahedral 55 intermediate (step 2). The carboxylic group in the glutamic moiety, which suffers the nucleophilic attack as depicted in Figure 4.1, is reported as the most reactive [79]. This carboxylic group is also less sterically precluded, which is an important feature to afford oligomerization and an effective catalysis by the protease. After release of a water molecule, an MTX-enzyme complex is formed, through the formation of an ester bond between MTX and the serine residue of the catalytic triad. This reactive bond easily suffers a nucleophilic attack by one of the amines in the pteridine ring of another MTX unit (step 3). The dimeric MTX is then formed and released from the enzyme, where the catalytic triad of -chymotrypsin is restored (step 4). Figure 4.1. Proposed mechanism for the synthesis of a dimeric unit of methotrexate catalysed by -chymotrypsin. Considering the data previously described, we may predict that the reaction was not restricted to the synthesis of dimeric units, but it can go further to longer species (DPmax=6). The size of the oligomers was however dependent on the affinity of MTX oligomers to the enzyme’ active site. Regarding the data obtained, we may infer that α-chymotrypsin was able to synthesize species until 56 a maximum of 6 MTX units. However, longer species were hampered to be synthesized due to stereochemical constrains resulting from the molecular weight of MTX (498 g/mol), which is comparably higher than the natural substrates of this enzyme (amino acids, peptides). For a better understanding of these findings, molecular dynamics simulations were performed to evaluate the affinity of MTX and oligomeric MTX to the active site of the enzyme and evaluate the role of the substrate size on the enzyme activity. 4.3.3. Molecular Dynamics Simulations Bovine α-chymotrypsin (PDB ID: 1OXG) [82] was evaluated under different conditions of temperature and pH during Molecular Dynamics (MD) Simulations. Firstly, the enzyme was evaluated at 310 K and physiological pH, then at 323 K and pH 9.5; this second simulation was performed to mimic the experimental conditions used. RMSD results (Figure 4.2A), indicate that the experimental conditions established are the ideal to study the enzyme behaviour, since the enzyme demonstrates to be more stable at the higher temperature and pH. Figure 4.2B shows the α-chymotrypsin crystallographic structure (I) highlighting the active site, where the catalytic triad composed by His57, Asp102 and Ser195 takes place, and the electrostatic distribution for this enzyme (II). 57 Figure 4.2. A) Backbone RMSD of -chymotrypsin, simulated at 310 K and physiological pH (black trace), and at 323 K and pH 9.5 (red trace), from the initial 3D structure (1OXG); B) 3-D structure of bovine α-chymotrypsin (PDB ID: 1OXG), from X-ray diffraction (I), highlighting in stick representation the catalytic triad: His57, Asp102 and Ser195. Green for cartoon/carbon, blue for nitrogen, red for oxygen and with for hydrogens; (II) shows the electrostatic distribution, where blue corresponds to positive areas and red to negative ones. MTX and MTX oligomeric structures were predicted at PM6 level [93], a semi-empirical method that performs geometry optimization and charge distribution, making the structures suitable to be used for molecular docking. In the case of MTX, the neutral and negative forms (at pH 9.5) were evaluated. The structures presented in Figure 4.3 were used in Molecular Docking experiments. 58 Overall, monomer (MTX), dimer and trimer have an increasing flexibility due to the increasing number of torsions. Dimer and trimer (Figure 4.3, C-D) optimized with PM6, can undergo more "bent" conformations when submitted to docking, due to the higher number of torsions. Figure 4.3. A) Skeletal and 3-D methotrexate (MTX) structures at pH 7.0 and B) the deprotonated forms at pH 9.5. C) skeletal and 3-D structures of MTX dimer, at pH 9.5 and D) shows the trimer 3-D representation. The colour scheme use green for carbon, blue for nitrogen, red for oxygen and white for hydrogen. 59 Docking experiments were conducted in order to understand the interactions between MTX and αchymotrypsin, and the role of this enzyme in the synthesis process. We selected an area that enwrap the active site with enough space to comprise the MTX units (monomer, dimer or trimer) to dock. Figure 4.4 shows the correlation between the MTX and the enzyme, and the binding energy involved for each case studied. MTX complexed with -chymotrypsin, presented a binding energy of -6.94 kcal/mol (-29.04 kJ/mol) and a Ligand Efficiency (LE) of -0.19 kcal/HA (-0.79 kJ/HA). The monomer has hydrogen bonds with Ala56, Gly59, Ser96, Tyr94 and Thr104, and van der Waals interactions were found with His57 and Asp102. For dimeric MTX, Gly59 and Tyr94 were also involved in hydrogen bonds, as well as Lys90 and Asn95. The binding energy was -6.35 kcal/mol (-26.61 kJ/mol) in this case, and a LE of -0.1 kcal/HA (-0.42 kJ/HA). His57 was still a close contact for MTX dimer. The trimer binds to the enzyme with -5.45 kcal/mol (-22.80 kJ/mol) and LE of -0.06 kcal/HA (- 0.25 kJ/mol). Similarly, His57 and Asp102 interact with the trimer. In addition, five hydrogen bonds were found, two with Ser96 and the others with Leu97, Asn91, and Thr104. In all the three cases, electrostatics interactions were also present, between the carboxylic acids of MTX and the positive region observed in Figure 4.2B II. Farhadian and co-workers recently reported docking experiments with -chymotrypsin [101, 102]. However, the complexes were not formed in the active site, where the catalytic triad take place. Some other works addressing the docking near the catalytic site do not reveal the complex binding energy for the association of a molecule to the protein [103, 104]. Other larger molecules have been described to efficiently bind to -chymotrypsin, however these molecules are used as inhibitors, and therefore cannot be compared to our work [104]. The lack of similar docking approaches makes it difficult to compare our binding preferences and energies with other studies. Moreover, the conversion of the binding energy to association constants is also hard to obtain. In docking experiments, the protein is static, the whole long-range electrostatics is not considered, and no water molecule is present. Therefore, the use of the equation ΔG = -RT ln Ka, would lead to a rough approximation. All these interactions demonstrate the ability of MTX to bind to the protease active site, leading to the formation of oligomeric MTX. Even though being possible more interactions of longer MTX forms with the protease, the binding energy decreases discretely as well as the LE per heavy-atom. One may infer that this was due to the energy penalties resulting from the increased number of torsions and desolvation of the bulky forms of MTX (dimer and trimer), since they can undergo 60 more conformations, as previously observed during docking experiments. The results indicated a decrease of the binding energy as the oligomer chain grows, justifying that only a maximum of 6 units of MTX could be biosynthesized by α-chymotrypsin. Figure 4.4. Docking poses showing the interaction of α-chymotrypsin with the different methotrexate species, and the respective binding energies obtained using AutoDock 4.0: A) αchymotrypsin with methotrexate; B) α-chymotrypsin with dimeric methotrexate and C) αchymotrypsin with trimeric methotrexate. Surface/cavities representation are in the right panel. Enzyme is represented in light grey, hydrogen bonds in yellow dashes and methotrexates following the scheme previous described in Figure 4.3. 61 4.3.3.1. NMR Spectroscopy By 1H NMR spectroscopy, it was possible to identify differences between MTX oligomers and the starting material (monomer). In the spectra of free MTX, the signals appeared at δH 8.56, 8.13 (NH), 7.62, 7.38 (NH), 6.81, 6.58 ppm (NH2), in the aromatic region (Figure 4.5A). The spectra of oligomeric MTX (Figure 4.5B) show significant chemical shifts in the pteridine ring, δH 8.15 (b), 8.14 (NH), 7.62, 6.78 ppm. The pattern of the signals remained the same as in the monomer MTX, confirming that the amide bond (near proton g) was not hydrolysed by the protease during the oligomerization reaction. Figure 4.5. 1H NMR (DMSO-d6) spectra of A) methotrexate and B) oligomeric methotrexate synthesized by α-chymotrypsin. 4.3.3.2. MALDI-TOF and ESI MALDI-TOF spectra displays the pattern of MTX oligomers, catalysed by α-chymotrypsin (Figure 4.6B), and of the control reaction without enzyme, showing the peak of the monomer below m/z 500 (Figure 4.6A). By ESI the ion peak corresponding to the monomer was observed at m/z 453 68 evaluate their catalytic performance on the biosynthesis of poly(ethylene glutarate), from diethyl glutarate and ethylene glycol. The reactions were carried out in the absence of water using the substrates as reactional medium. Molecular dynamics simulations in organic (reactant mixture) and aqueous medium were assessed in order to study the effect of PEGylation on the lipase’s conformation and on the access of the substrates to the active site cavity. A complete characterization of the synthesized oligomers was performed by proton and carbon nuclear magnetic resonance spectroscopy (1H and 13C NMR), matrix-assisted laser desorption/ionizationtime of flight (MALDI-TOF), Fourier-transform infrared spectroscopy (FTIR), differential scanning calorimetry (DSC) and thermogravimetric analysis (TGA). 5.2. Materials and methods 5.2.1. Materials Lipase from Thermomyces lanuginosus (solution, ≥ 100,000 U/g), lipase from Candida antarctica B (0.3 U/mg), O -[2-(6-Oxocaproylamino)ethyl]- O ′-methylpolyethylene glycol (PEG, MW 5000 Da), ethylene glycol (≥ 99 %), diethyl glutarate (≥ 99 %), 2,4,6-trinitrobenzene sulfonic acid (5 % (w/v) in H2O), pnitrophenyl butyrate ( p -NPB, ≥ 98 %) and sodium cyanoborohydride (95 %) were purchased from Sigma-Aldrich. Tetrahydrofuran (HPLC grade, Fisher Chemical) was used without further purification. Ultrafiltration was performed with Ultracel 10 kDa regenerated cellulose ultrafiltration discs, 47 mm (Millipore) with ultrapure water (Milli-Q). The expression and production of cutinase from Fusarium solani pisi (EC 3.1.1.74) was performed following the procedure reported by Araújo et al . [120]. 5.2.2. Synthesis 5.2.2.1. General procedure for the PEGylation of esterases The PEGylation of the esterases was performed using the procedure reported by Mayolo-Deloisa et al . [121]. Briefly, the esterase (12 mg/mL) was solubilized in phosphate (100 mM) and NaBH3CN (20 mM) buffer (pH= 5.1) followed by the addition of the PEG-aldehyde (esterase 1:4 PEG w/w). The reactional mixture was placed at 4 °C, overnight, under stirring. The separation of the unreacted PEG and buffer was carried out by ultrafiltration using a 10 kDa regenerated cellulose membrane housed in an ultrafiltration device using ultrapure water. The PEGylated esterase was obtained as a white solid after freeze-drying for 2 days. 69 5.2.2.2. General procedure for synthesis of poly(ethylene glutarate) Ethylene glycol was added to a 50 mL round-bottom flask, followed by the addition of the diethyl glutarate (equimolar amount) forming a biphasic mixture. The esterase was added, and the suspension was placed in an ultrasonic bath (USC600TH, VWR International Ltd., USA; frequency 45 kHz and power of 120 W) programmed to not exceed the 45 °C. After sonication for 2 hours, the round-bottom flask was transferred to a rotary evaporator (Heidolph, Germany) at 40 °C, 120 rpm, to complete 7 hours of total reactional time. Tetrahydrofuran was added to the reactional mixture and the enzyme was removed by filtration. The solvent was removed in the rotary evaporator and the final solution formed a colourless oil. 1H NMR (CDCl3): δH 1.23 (t, J = 6.8 Hz, CH3), 1.91-1.96 (m, CH2), 2.37-2.42 (m, CH2), 3.78-3.81 (m, CH2), 4.10 (q, J = 7.2 Hz, CH2), 4.18-4.20 (m, CH2) and 4.28 (s, CH2) ppm. 13C NMR (CDCl3): δC 14.1 (CH3), 19.8 (CH2), 19.9 (CH2), 20.0 (CH2), 20.3 (CH2), 32.7 (CH2), 32.8 (CH2), 32.9 (CH2), 33.0 (CH2), 33.1 (CH2), 33.2 (CH2), 33.3 (CH2), 60.4 (CH2), 60.9 (CH2), 61.9 (CH2), 62.1 (CH2), 65.9 (CH2), 172.5 (C=O), 172.6 (C=O), 172.7 (C=O), 172.8 (C=O), 173.0 (C=O), 173.1 (C=O), 173.2 (C=O), 173.3 (C=O), 176.5 (C=O), 176.7 (C=O) and 176.9 (C=O) ppm. 5.2.3. Enzyme characterization 5.2.3.1. SDS-PAGE SDS-PAGE electrophoresis of lyophilized esterases was performed by solubilize them in water and the samples were loaded on polyacrylamide Gel Electrophoresis (SDS-PAGE) gel 12.5 %. The gel was stained with Coomassie brilliant blue solution to analyse size and purity. 5.2.3.2. Esterase activity The activity of all the esterases used was determined by a continuous spectrophotometric assay using p -nitrophenyl butyrate ( p -NPB) as substrate. One unit of enzyme activity was defined as the amount of enzyme which catalyses the production of 1 μmol p -nitrophenol per minute. The standard assay was performed at 37 °C in a final volume of 4 mL containing p -NPB (6 mM), the enzyme and the assay buffer (K2HPO4 buffer, pH 7.8, 50 mM). The reaction was initiated by the addition of the enzyme. The hydrolysis of p -NPB was monitored by the formation of the p - nitrophenol at 400 nm [122]. The measurements were conducted in a Synergy Mx Multi-Mode Reader from BioTek (USA). 70 5.2.3.3. Protein quantification The quantification of the protein concentration was performed by using the DC protein assay (BIORAD). 5.2.3.4. Degree of PEGylation The degree of enzyme PEGylation was indirectly evaluated by colorimetric titration. This methodology occurs by the reaction of 2,4,6-trinitrobenzene sulfonic acid (TNBSA) with the free amine residues at the surface of the enzymes. Knowing the total amount of amine residues available in each protein, this quantification allowed us to calculate the amount of PEG chains coupled to each esterase. The procedure was followed as reported by Castillo et al . [123]. 5.2.3.5. Molecular Dynamics Simulations Molecular Dynamics (MD) simulations were performed on lipase from Thermomyces lanuginosus (lipase TL, PDB ID: 1TIB) [124], on lipase B from Candida antarctica (CALB, PDB ID: 1TCA) [125], and on its PEGylated forms, to understand the role of the lid and active site cavity in different environments. Lipases were modelled in the simple point charge water model, for control, and in a mixture of diethyl glutarate and ethylene glycol, respecting the same experimental proportion. In both cases, a cubic box with an approximate volume of 530 nm3 was used, with the enzyme centralized and Na+ ions to neutralize the system. One stage of energy minimization was performed using a maximum of 50,000 steps with steepest descent algorithm. Position restraints (with force constant of 1000 kJ·mol−1·nm−2) were applied to all heavy atoms at the initialization steps, the first using an NVT ensemble and the second, with NPT. The temperature was maintained constant with Vrescale algorithm [84] and the pressure, was regulated at 1 atm, with the Parrinello-Rahman barostat [85]. In the control situation (enzyme in water), a temperature of 310 K was used, but for the second situation, enzymes in the mix of substrates, a temperature of 313 K was chosen to mimic exactly the experimental procedure. The following coupling constants were considered: τT=0.10 ps and τP=2.0 ps. After that, all systems were submitted to MD simulations during 40 ns, in an NPT ensemble, without position restraints. All simulations were performed using the GROMACS 5.1.4 version [86, 126], within the GROMOS 54a7 force field (FF) [87, 88]. The Lennard-Jones interactions were truncated at 1.4 nm and we use particle-mesh Ewald (PME) [89] method for electrostatic interactions, also with a cut-off of 1.4 71 nm. The algorithm LINCS [90] was used to constrain the chemical bonds of the proteins and the algorithm SETTLE [127] in the case of water. To design and simulate the box containing the substrates diethyl glutarate and ethylene glycol as solvents, we had to optimize and parameterize these molecules. For this we run a PM6 calculation [93] with Gaussian09 software [92] followed by submission of the resulting optimized structures at ATB server (Automated Force Field Topology Builder) [128, 129]. As result, we obtained optimized structures with a GROMOS 54a7 FF parameter associated to each one. The PEGylated systems were designed replacing Lysine (Lys) residues to a new type of Lysine, named as LYP, where a PEG chain was linked. To prevent high computational costs, we connect only three PEG units to each chosen lysine side chain. The GROMOS topology necessary for this new residue, LYP, was also obtained using ATB server. In the case of lipase TL, 5 types of PEGylation took place until the limit of 5 of 7 Lys PEGylated: the analogue LP1 has LYP only in position 98, LP2 has LYP98 and the Lys PEGylated in β-sheets (positions 74, 223 and 237), LP3 has LYP98 and the Lys PEGylated in -helices (positions 24, 46 and 127), LP4 present LYP98 and 4 random Lys PEGylated (positions 24, 46, 223 and 237) and LP5 is the only case where Lys98 was not PEGylated, but 5 random Lys were replaced by LYP (positions 46, 74, 127, 223 and 237). For CALB, an analogue with all Lys PEGylated was designed. This was done to correspond to the results obtained experimentally. From MD simulations, we computed a cluster analysis from GROMACS package, with the singlelinkage method, to determine the middle structure of each enzyme, i.e., this technique adds structures that are below a RMSD cut-off, generating more or less populated clusters and, within the largest cluster, it finds a middle structure that is the most representative of the whole simulation. We also follow the changes in the lid conformation and in the region involving the catalytic triad through visualization analysis with PyMOL [130]. 5.2.4. Polymer characterization 5.2.4.1. Nuclear Magnetic Resonance spectroscopy (NMR) All NMR spectra, namely 1H NMR, 13C NMR, Distortionless Enhancement by Polarization Transfer (DEPT), 1H-13C Heteronuclear Single Quantum Coherence (HSQC) and 1H-13C Heteronuclear Multiple Bond Correlation (HMBC) were carried out on a Bruker Avance III 400 spectrometer (400 MHz for 1H and 100 MHz for 13C). Deuterated chloroform (CDCl3, Cortecnet, France) was used as 72 NMR solvent, and the peak solvent used as internal reference. Signal multiplicity was given as: s (singlet), t (triplet), q (quartet) and m (multiplet). 5.2.4.2. Matrix-Assisted Laser Desorption/Ionization Time-of-Flight (MALDI-TOF) MALDI-TOF mass spectra were acquired on a Bruker Autoflex Speed instrument (Bruker Daltonics GmbH) equipped with a 337 nm nitrogen laser. The procedure was followed as previously reported [79, 131]. 2,5-dihydroxybenzoic acid (DHB) or α-cyano-4-hydroxycinnamic acid (CHCA) were used as matrix. Samples were analysed in the linear positive or negative mode. The number average (𝑀𝑛) and weight average molecular weight (𝑀𝑤), the polydispersity index (PDI= (𝑀𝑤 / 𝑀𝑛)), the average and maximum degree of polymerization (DPavg, DPmax) (DPavg= 𝑀𝑛/repeating unit of the oligomer) were calculated based on the MALDI-TOF spectra obtained, based on the m/z values and intensity, following the equations: 1) 𝑀𝑛 =∑𝑛𝑖 𝑀𝑖 ∑𝑛𝑖 2) 𝑀𝑤 =∑𝑛𝑖 𝑀𝑖2 ∑𝑛𝑖 𝑀𝑖 Where ni is the relative abundance of each peak in the MALDI-TOF spectra and M i is the m/z corresponding to each peak. 5.2.4.3. Fourier-transform infrared spectroscopy (FTIR) Infrared spectra were recorded on a FTIR Bomem MB using NaCl cells. The samples were analysed over the range 500-4000 cm-1, with a spectral resolution of 4 cm-1. All spectra were an average of over 20 scans. 5.2.4.4. Differential Scanning Calorimetry (DSC) DSC measurements were conducted on a power-compensated DSC instrument (DSC 6000 Perkin Elmer) with a nitrogen flux of 20 mL/min, using stainless steel capsules in the temperature range of 20-250 °C (heating rate: 20 °C/min, sample weight: 2-3 mg). The DSC device was calibrated 73 using indium and zinc, both of high purity. The samples were freeze-dried, prior to the analyses and the sample was measured at least six times, to validate the results. 5.2.4.5. Thermogravimetric Analysis (TGA) TGA analysis was performed in a Perkin Elmer TGA 4000. The calibration was performed with metals, such as Nickel, Alumel and Perkalloy, based on their Curie Point Reference. The temperature range was 30-800 °C (heating rate 20 °C/min, sample weight: 12-16 mg) and the nitrogen flow rate was 20 mL/min (3 bar). 5.3. Results and discussion 5.3.1. PEGylation and catalytic properties of PEGylated esterases Most of the soluble enzymes are highly stabilized in the presence of high concentrations of polyethylene glycol. The presence of a medium with a high viscosity is expected to prevent the undesired changes in enzyme structure promoted by denaturing agents (e.g., high temperatures, strong basic conditions, extreme pH values) and, therefore, the stability of the soluble enzymes greatly increases. Based on these observations, our approach was to build such a layer around the enzyme to stabilize it, by PEGylation of the primary amine groups available. The PEGylation of lipase from Thermomyces lanuginosus (lipase TL), lipase from Candida antarctica B (CALB) and cutinase from Fusarium solani pisi (CUT) was performed as reported in the literature, using an monofunctional PEG with an aldehyde group (MW 5000 Da) [121]. The reaction occurred at acidic pH (5.1) in the presence of a reducing agent, sodium cyanoborohydride, as proposed in Scheme 5.1. Scheme 5.1. Proposed mechanism for the PEGylation of esterases; A) Lysine residue of an esterase; B) PEG-Aldehyde; C) Esterase-lysine-PEG as imine intermediate; D) Esterase-lysine-PEG. 74 The TNBSA assay allowed the quantification of the free amine residues at the surface of the esterases which were not modified by PEGylation. Indirectly, we were able to infer that, under the conditions described, the PEGylation of CALB resulted in a protein with 100 % of the exposed amines linked to PEG. After modification of lipase TL and cutinase, only 59 and 78 % of the amino groups were linked to PEG, respectively (Table 5.1). Depending on the position of the amine residues at the enzyme’s surface, stereo-chemical impediments might influence the degree of PEGylation. More exposed amines are more prone to covalently react with the PEG available chains [132]. Regarding the methodology used to PEGylate, it is likely that the amine residue of the N - terminus had been also PEGylated. This assumption is based on the use of acidic pH during the reaction, which led to the activation of the N -terminus [121]. Table 5.1. Percentage of amine residues modified by PEGylation and activity of esterases, before and after PEGylation Esterase Amine modification[a] Activity[b] (U/mgprotein) CALB --- 60 PEGylated-CALB 100 % 132 Lipase TL --- 26 PEGylated-Lipase TL 59 % 27 Cutinase --- 260 PEGylated-Cutinase 78 % 1061 [a]obtained by TNBSA assay. [b]calculated by the hydrolysis of p -nitrophenyl butyrate over 1 min and considering the same initial amount of protein; 1 U of enzyme activity was defined as the amount of enzyme required to convert the substrate ( p -nitrophenyl butyrate) into p -nitrophenol in 1 min. The PEGylation was also ascertained by SDS-PAGE electrophoresis as a complementary methodology (Figure 5.1). 75 Figure 5.1. SDS-PAGE gel of native and PEGylated esterases stained with Coomassie brilliant blue; A) GRS Protein Marker Blue (from Grisp, Portugal), B) Lipase from Thermomyces lanuginosus ; C) PEGylated lipase from Thermomyces lanuginosus ; D) Cutinase from Fusarium solani pisi ; E) PEGylated cutinase from Fusarium solani pisi . Lipase from Thermomyces lanuginosus has a typical visible band at 30 kDa and when PEGylated a new band appears at around 40 kDa (corresponding to 2 units of PEG covalently bond to the protein). Cutinase display a representative band at 22 kDa, while its PEGylated form shows an intense smear with two pronounced bands in the range of 25 and 45 kDa. The PEGylated forms present also an evident smear suggestive of an increase of the molecular weight incremented by PEGylation. As a consequence of the reduction in the number of free amino groups, the enzyme derivatives (PEGylated forms) showed a smaller electrophoretic mobility toward the cathode than the unmodified esterase. Due to the presence of stabilizers in the medium, the native and PEGylated CALB were not successfully revealed by SDS-PAGE electrophoresis. The activity of enzymes is a parameter greatly influenced by the PEGylation procedure. Thus, the hydrolytic activity of the esterases against p -nitrophenyl butyrate was evaluated, before and after PEGylation, and the results obtained reveal a different catalytic behaviour of the catalysts after the chemical modification. We have found that PEGylation greatly enhanced the CALB and CUT catalytic activities comparing to their native state. This was a surprising result since only few examples of increased activity after PEGylation can be found in the literature, as previously mentioned [112]. Comparing to their native forms, the activities of PEGylated CALB and CUT increased 2-fold and 4-fold after PEGylation, respectively. According to our data, the PEGylation of 76 lipase TL did not influence its catalytic activity which remained unaltered after chemical modification. The addition of free PEG to the native enzymes’ medium was also evaluated and the results revealed no effect of the stabilizer on their hydrolytic activity (data not shown). It was noteworthy that all PEGylated catalysts remained stable for at least six months of storage at room temperature. To evaluate their stability under processing conditions, we incubated the enzymes at 40 °C and measured the activity over time (activity measured using p -NPB as substrate). From the results obtained one might infer that PEGylation conferred stabilizing effects to the modified esterases. We observe that both native and PEGylated enzyme forms, remained active for more than 50 days under the same storage conditions, with minimal activity loss. During all the reactional process at 40 °C, no loss of activity was registered based on the results obtained after 8 h of incubation (Figure 5.2). The temperature of incubation has been described in literature as a differential factor for PEG stabilization performance. The high thermal stability achieved herein at 40 °C might be explained by the high viscosity of PEG layers surrounding the enzymes. Although PEGylation is being described as a methodology prone to inactivate some proteins [133], the results obtained clearly demonstrate that this methodology allows to improve esterase’s activity and stability. Figure 5.2. Absolute activity of esterases (native and PEGylated forms), at time zero, after 8 h, and after 50 days of incubation at 40 °C in phosphate buffer (pH 7.8). The activity was measured against pNPB over 1 min and considering the same initial amount of protein; 1 U of enzyme activity was defined as the amount of enzyme required to convert the substrate ( p -nitrophenyl butyrate) into p -nitrophenol in 1 min. 0 200 400 600 800 1000 CALB PEG-CALB TL PEG-TL CUT PEG-CUT Activity (U/mgprotein) 0h 8h 50 days 77 5.3.2. Effect of PEGylation on the polymerase activity of esterases The synthesis of poly(ethylene glutarate) catalysed by immobilized CALB was previously investigated by us [54], using diethyl glutarate and ethylene glycol diacetate as starting reagents. Herein, we replace the ethylene glycol diacetate by ethylene glycol, which would allow to obtain a greener sub-product, ethanol, instead of ethyl acetate. A greener and environmentally friendly methodology is also associated with the proposed method due to the absence of solvents on the reactional mixture, being all the reactions carried out in bulk. Moreover, mild reaction conditions of temperature (40 °C) and short reactional times were used, reducing therefore the energy costs associated to the process. Regarding the previous results published [54] we proceed with the processing optimizations (data not shown) and established the best reactional conditions for the polymerization of the proposed polyester: 2 h under US followed by 5 h under vacuum at 40 °C, following the reaction presented in Scheme 5.2. Scheme 5.2. Reactional scheme for the synthesis of poly(ethylene glutarate) catalysed by esterase; A) diethyl glutarate; B) ethylene glycol; C) poly(ethylene glutarate); D) ethanol. Taking into account the best reactional procedure established, the enzyme loading (2 to 130 U/mg) was tested to evaluate the best conditions to attain the highest synthesis conversion (Figure 5.3). Results from Figure 5.3 reveal that, when comparing to their native forms, the highest product conversion was obtained when using both PEGylated-lipase TL and PEGylated-CUT enzymes. CALB display however an opposite behaviour after PEGylation, giving rise to slightly lower conversion levels. One can also observe that all enzymes tested showed a similar trend between 2 and 130 U/mg of enzyme loading, reaching however different levels of conversion, depending on the enzyme form and source. At the maximum enzyme loading, PEGylated lipase TL revealed the highest polymerase performance comparing to its native form, reaching levels of conversion of around 90 % whereas PEGylated-CUT converts only 60 % of the monomers. The reactions were also carried out with native enzymes in the presence of free PEG in the reactional medium. The data obtained at these conditions showed lower conversions than compared with native enzymes 84 glutarate. This is evident in Figure 5.7 (A), which highlights the active site of lipase TL and LP1 in reactant mixture. Figure 5.7. Middle conformations of lipase TL and PEGylated analogue LP1, in reactant mixture (A) and in water (B), highlighting the interior cavities and pockets surrounding the catalytic triad and lid regions. The CALB middle structures demonstrate that the medium has a similar impact on the enzyme’s cavity region. In water, the PEGylation seems to not disturb the structure and the active site cavity which remains unaltered. In the reactant mixture (Figure 5.8), although PEG stabilizes CALB, it does not generate a great enlargement of the cavity, promoting a more discrete effect on this enzyme, when compared with the effect towards TL. Regarding the similar conversion yield and DP obtained experimentally for both CALB forms, a negligible effect of PEG on the CALB structure was expected. CALB owns a large acyl and small alcohol, clefts, ideal for the polymerization of the substrates tested. MD studies reveal that PEG did not disturb the active site of the enzyme, thus maintaining the size of both clefts which resulted in similar polymer conversion, independently on the enzyme form used. 85 Figure 5.8. Middle conformations of CALB and PEGylated CALB, in reactant mixture (A) and in water (B), highlighting the cavities and pockets surrounding the catalytic triad and lid regions. Cartoon in grey, catalytic triad in yellow, lid-like region in cyan, LYP residues in magenta and the orange spheres represent the empty space (cavity or pocket) on each structure. Recently, MD studies and other modelling techniques have been used to understand the mechanisms of interfacial activation in lipase TL, CALB and other lipases [141-143]. However, these studies were in general performed using water as medium, or a water-lipid/water-oil interface. We show herein an innovative approach to demonstrate the contribution of PEGylation on the final enzyme arrangement. By comparing simulations in water and in the reactant mixture, we were able to confirm that the starting organic reactants, diethyl glutarate and ethylene glycol, create the best environment to enhance the catalytic properties of lipase, i.e., an open access to the catalytic triad. 5.3.4. Poly(ethylene glutarate) characterization The results of the biosynthesis are summarized in Table 5.2, including the number average molecular weight (𝑀𝑛), the weight average molecular weight (𝑀𝑤), the polydispersity index ((𝑀𝑤)/ 𝑀𝑛)) and the average and maximum degree of polymerization, for all the reactions performed. 86 The synthesized polymers displayed 𝑀𝑛 values between 415.97 and 799.99 g/mol, and 𝑀𝑤 values between 527.03 and 1031.61 g/mol. All reactions gave rise to polymers with good polydispersity, whereas PEGylated-CALB displayed the more homogeneous (1.04) and PEGylatedTL the more heterogeneous (1.48). As previously stated, PEGylated-lipase TL showed higher DP than its native form. This PEGylated esterase, also showed the highest maximum degree of polymerization (16 units). Both forms of CALB have the same average and maximum DP. The native form of cutinase showed better performance than its PEGylated form. A possible explanation for all these results was already discussed elsewhere in this chapter. Table 5.2. Number average molecular weight (𝑀𝑛),weight average molecular weight (𝑀𝑤), polydispersity (PDI),average degree of polymerization (DPavg),maximum degree of polymerization (DPmax) (calculated by MALDI-TOF) and conversion rate, after poly(ethylene glutarate) biosynthesis (2 h under US followed by 5 h under vacuum at 40 °C with the 65 U/mg of enzyme) 5.3.4.1. NMR 1H NMR spectra of the new polymers formed have a similar pattern independently on the enzyme used and the degree of polymerization obtained. In Figure 5.9A is represented the 1H NMR spectra of poly(ethylene glutarate). The monomer, ethylene glycol, presented only one peak in the spectrum which appeared at δH 3.73 ppm. These protons, suffered a significant chemical shift when the synthesis occurred, appearing as a singlet at δH 4.30 ppm (e). The terminal ethylene glycol unit (protons a and b) appeared as two distinct peaks, one at δH 3.82 ppm (protons a) and the other at δH 4.21 ppm (protons b), both as multiplets. The glutarate moiety in the polymer do not showed significant changes comparing with the same protons in the monomer. Only the terminal part Esterase 𝑴𝒏 𝑴𝒘 𝑴𝒘 / 𝑴𝒏 DPavg DPmax Conversion rate (by 1H NMR) CALB 580.89 625.18 1.08 4 8 91.3 ± 1.3 % PEGylated-CALB 597.91 624.32 1.04 4 8 84.0 ± 5.9 % Lipase TL 415.97 590.78 1.42 3 10 29.0 ± 1.4 % PEGylated-Lipase TL 691.49 1021.40 1.48 4 16 87.8 ± 8.7 % Cutinase 799.99 1031.61 1.29 5 15 19.6 ± 16 % PEGylated-Cutinase 428.02 527.03 1.23 3 8 33.9 ± 13 % 87 (protons h and i) showed a decrease in the signal intensity depending on the degree of polymerization. Figure 5.9. 1H (A) and 13C (B) NMR spectra of poly(ethylene glutarate) recorded in CDCl3. In the 13C NMR spectra (Figure 5.9 B) the expectable peaks related with the new polymer were observed. The C=O were observed between δC 172.5 and 176.9 ppm. The terminal CH3 (i) appeared at δC 14.1 ppm and CH2 (h) at δC 60.4 ppm, while the other carbons of the glutarate (c, d, f, g) moiety were observed between δC 19.8-20.3 and 32.7-33.3 ppm. The ethylene glycol moiety (e) appeared at δC 61.9 and 62.1 ppm. The terminal a and b carbons were observed at δC 60.9 and 65.9 ppm, respectively. The pattern obtained was in accordance to the previous poly(ethylene glutarate) reported [54], and the chemical shifts in the 13C NMR are typical of polyesters [144]. 5.3.4.2. MALDI-TOF The MALDI-TOF mass spectrum of the formed poly(ethylene glutarate), recorded in linear positive mode, displays a typical isotopic distribution between 500 and 2500 m/z (Figure 5.10). From the spectra one can depict a repetition unit mass between each two peaks, corresponding to the monomeric repeating unit ([M]= 158). This confirmed the presence of the monomeric repeating unit in the polymer main chain. Similar spectra were obtained for the polyester synthesized by the other esterases studied, showing different DPs (data not shown). These results are in agreement with the formation of the proposed polymer, with DPavg= 5, as suggested by 1H NMR and 13C NMR data. The pattern obtained for the polyester is in accordance with similar polyesters described in the literature [145]. 88 Figure 5.10. Positive ion MALDI-TOF spectra of poly(ethylene glycol) (DPmax= 16 and DPavg= 5). 5.3.4.3. FTIR, DSC and TGA FTIR analysis was also conducted to evaluate the chemical changes after enzymatic synthesis. From the data obtained (Figure 5.11), as expected, one can observe that the stretch of the OH terminal group was more pronounced for low degrees of polymerization. Also, the OH appeared in the polymer at ν 3500 cm-1 while in the monomer was observed at ν 3300 cm-1. Regarding the C=O bond, it was observed at ν 1700 cm-1 for monomers and polymers. Figure 5.11. FTIR spectra of A) diethyl glutarate; B) ethylene glycol; C) poly(ethylene glutarate) with a DPavg= 3; D) poly(ethylene glutarate) with a DPavg= 5. 89 The glass transition temperature (Tg) of the polymer was determined by DSC analysis. Poly(ethylene glutarate) showed a value of Tg= 77.29 ± 1.21 °C with an energy of ΔCp = 0.228 ± 0.095 J/g°C-1 (Figure 5.12A). A melting point (Tm) was observed at Tm = 195.7 °C with an associated enthalpy of ΔHm= 19.967 J/g. The pattern here obtained is very similar to the DSC analysis of polyesters reported in literature [144]. The thermal properties of the synthesized polyester were investigated by thermogravimetric analysis scanned between 30-800 °C (Figure 5.12B). Both monomers present a one-step decomposition, losing all weight at around 180 °C. The polyester presented two distinct stages of weight loss: one at around 220 °C, corresponding to 10 % of weight loss; and another at 400 °C, corresponding to 50 % of weight loss. The total material decomposition was observed near 500 °C. The thermal behaviour observed, typical for this type of polymers [146], confirms the polyester biosynthesis. Figure 5.12. Graphics of A) DSC curve of poly(ethylene glutarate) with Tg= 77.29 ± 1.21 °C and Tm = 195.7 °C, and the respective monomers and B) TGA curves of the starting materials (diethyl glutarate and ethylene glycol) and the formed poly(ethylene glutarate). 5.4. Conclusions In this work we PEGylated three esterases (lipase from Candida antarctica B , lipase from Thermomyces lanuginosus and cutinase from Fusarium solani pisi ) and compared their catalytic performance for the biosynthesis of poly(ethylene glutarate). All the enzymes were successfully PEGylated, and their hydrolytic activity, with exception of lipase TL, was improved comparing to their native form. Regarding their polymerase activity, we observed a similar performance for native and PEGylated CALB, explained mainly by their large acyl cleft. On the other hand, lipase TL 90 presented an improved performance when PEGylated. Molecular dynamics simulations, performed on lipase TL and CALB, support that the PEGylation of the lysine at the enzyme’s lid had a positive effect on the substrate accessibility to the active site. The simulations conducted in the reactant mixture medium confirm the stabilization of the enzyme by PEG in a more organic environment. The entropic stabilization by PEG conjugation also caused the motion restriction of some surface amino acid side chains, resulting in a more stable active site. The PEGylation of esterases demonstrated to be an easy, not expensive, and timeless methodology to enhance enzyme’s performance on the biosynthesis of polyesters, envisaging a diverse range of applications. 91 Chapter VI Substrate’s hydrophobicity and enzyme’s modifiers play a major role on the activity of lipase from Thermomyces lanuginosus 92 Chapter VI Substrate’s hydrophobicity and enzyme’s modifiers play a major role on the activity of lipase from Thermomyces lanuginosus Abstract Lipase from Thermomyces lanuginosus (TL) displays high affinity for long-chain substrates, such as triolein and other long-chain triacylglycerols. Aiming to broaden the substrate chain-length specificity, different aldehydes (naphthaldehyde, butyraldehyde, hexyl aldehyde and dodecyl aldehyde) and naphthyl isothiocyanate were grafted onto lipase TL, through lysine coupling. The catalytic activity of the modified lipases was investigated by reaction with substrates differing in the aliphatic chain size ( p -nitrophenyl benzoate, p -nitrophenyl acetate, p -nitrophenyl butyrate, p - nitrophenyl hexanoate, p -nitrophenyl octanoate, p -nitrophenyl laurate and p -nitrophenyl palmitate). The enzymes modified with aldehydes displayed higher activity than the enzymes modified with the isothiocyanate. The most notable results were achieved for lipase TL, grafted with 4 units of a dodecyl chain (TL5), which displayed the highest activity against all the tested substrates, being 10-fold more active than the native enzyme for smaller substrates (acetate and butyrate chains), and 2-fold for longer (laurate and palmitate chains). The kinetic parameters evaluated (Vmax, K M and k cat/ K M) also confirmed the significant catalytic performance of TL5 comparing to the native form. The increased activity, revealed by the modified lipases, was directly proportional to the size and hydrophobicity of the linkers’ aliphatic chain. Small conformational changes, either on the enzymes’ lid as on the cavity of the active site were suggested by Molecular Dynamics simulations, Circular Dichroism and Fluorescence spectroscopy. Moreover, the grafting with aldehydes or with the isothiocyanate, conferred higher thermostability to the lipase. The chemical surface modification developed efficiently improved the activity of lipase TL, broadening the substrates chain-length specificity, incrementing thereafter the substrate possibilities for industrial reactions. This chapter is based on the following publication: Jennifer Noro, Tarsila G. Castro, Artur Cavaco-Paulo, Carla Silva, Substrate’s hydrophobicity and enzyme’s modifiers play a major role on the activity of lipase from Thermomyces lanuginosus , Catalysis Science & Technology, 10 (2020) 5913-5924. 93 6.1. Introduction Lipase from Thermomyces lanuginosus (EC 3.1.1.3) was firstly isolated from compost medium containing, among other components, long-chain esters like soybean oil, corn steep liquor and starch [147]. The natural function of lipase from Thermomyces lanuginosus is described in literature as the degradation of the long chain triacylglycerols present in the compost medium [147149]. Nowadays, this lipase is produced using a recombinant strain of Aspergillus oryzae, a filamentous fungus often used for enzyme production [150], and the isolated enzyme is responsible for the lipolytic activity of Lipolase®, used for many applications like the production of flavours, biodiesel, and fine chemicals [151]. It is also commonly used for the hydrolysis of substrates such as triolein [142] and other long-chain triacylglycerols [151, 152]. This enzyme owns 1,3-stereospecificity, meaning that it specifically hydrolyses the ester group of triacylglycerols in the position 1 and 3. The hydrolysis of the position 2 occurs by migration of the acyl group to the position 1, and further hydrolysis by the enzyme. The surface modification of enzymes is a strategy often applied to improve their global properties, mainly the thermostability. The immobilization of enzymes onto solid supports, is the strategy mostly studied, with diverse associated advantages, such as easy recovery, increase of resistance and robustness, among others [153]. Lipase from Thermomyces lanuginosus is mainly used in its immobilized form, as reported by several authors, due to the higher catalytic performance acquired in the immobilized form, comparing to the free form. Cipolatti and co-workers [154], studied the adsorption of lipase TL onto PEGylated polyurethane particles, and confirmed the improvement in the production of ethyl esters, comparing to the free native enzyme. Vasconcellos et al . [155], reported the immobilization of lipase TL on nanozeolites. They observed an increase on the production yield of biodiesel using this novel nanozeolite-enzyme complex. Despite the promising results reported so far, the use of solid supports for enzyme immobilization often reduces its catalytic activity resulting from the reduced mobility imposed by the immobilization. A strategy to overcome this limitation and, at the same time, increment their catalytic activity, is the surface modification of the enzyme with small molecules like imidazolium or alkylated ammonium salts. However limited number of works have been reported in literature so far about this topic [156]. 100 hydrophobicity. The linkers were chosen based on their high reactivity and ability to covalently bond to the primary amines of the exposed lysine residues of the enzyme (total of 7 residues observed by molecular dynamics). The proposed scheme for the reactions between the enzyme and the tested linkers is represented in Figure 6.1. The degree of modification was assessed by the TNBSA assay and by MALDI-TOF spectrometry (Table 6.1). After product isolation and evaluation (MALDI-TOF and TNBSA) the data revealed that all aldehyde compounds tested were successfully linked to the enzyme, with modification of at least one lysine available (Table 6.1). Figure 6.1. Proposed reactional scheme for the enzyme modification: A) native enzyme with an exposed lysine residue represented; B) enzyme modified with an aldehyde in the lysine residue, leading to a secondary amine; C) enzyme modified with an isothiocyanate in the lysine residue, leading to a thiourea; linkers used for lipase modification: i) naphthaldehyde; ii) butyraldehyde; iii) hexyl aldehyde; iv) dodecyl aldehyde; v) naphthyl isothiocyanate; vi) phenethyl isothiocyanate; vii) octyl isothiocyanate. The aliphatic isothiocyanates (Figure 6.1, vi and vii) displayed the lowest reactivity, confirmed by the isolation of non-modified lipase. One unit of a more reactive aromatic isothiocyanate (Figure 6.1, v) was grafted onto lipase enzyme’s surface. We can assume, as it is postulated, that isothiocyanate compounds, when linked to the amine group of the lysine residue, give rise to 101 thioureas, whereas aldehydes give rise to secondary amines [81, 167]. Lysine residues are the most nucleophilic amines present in enzymes [168]. The chosen class of compounds (aldehydes and isothiocyanates) are described to be preferentially linked to this residue instead to other reactive amino acids [168]. Different reactivities between linkers might explain the differentiated modification degrees obtained experimentally. Table 6.1. Modification degree and number of modified residues evaluated by MALDI-TOF and TNBSA assay, being TL the native enzyme, and the modified enzymes: TL1 modified with naphthyl isothiocyanate, TL2 modified with naphthyl aldehyde, TL3 modified with butyraldehyde, TL4 modified with hexyl aldehyde and TL5 modified with dodecyl aldehyde MALDI-TOF (values in Da) Nº of residues modified by MALDI-TOF % of modification (by TNBSA assay) Nº of residues modified by TNBSA TL 29620 - - - TL1 29771 1 8 1 TL2 29826 1 15 1 TL3 30283 7 93 7 TL4 30392 7 97 7 TL5 30270 4 49 3-4 MALDI-TOF results revealed, in comparison with the non-modified lipase (29620 Da), an increase of the molecular weight of the enzyme after modification with all the compounds used. In all cases, the results of MALDI-TOF and TNBSA were in accordance, revealing a direct relation between the degree of modification with the number of residues modified (Table 6.1). The lipases modified with a naphthyl moiety (TL1 and TL2) revealed only one modified residue, probably due to their poor solubility and/or low reactivity of the starting material. TL3 and TL4 (with butyl and hexyl chains, respectively) were modified in all the 7 available lysines, while TL5 (with dodecyl chains) was only modified in 4 of the 7 available lysines. The length of the chains may induce a steric effect, which might explain the different modification degree obtained for the tested 102 linkers. Lysines located at positions surrounded by larger amino acids, are less prone of being modified by larger linkers. One may predict that, in the case of aldehydes, the degree of enzyme modification was inversely proportional to the size of the linker chain. The modified lipases were also analysed by SDS-PAGE to evaluate size and purity (Figure 6.2). All enzymes were obtained pure, displaying only one visible band at around 30 kDa. The small molecular weight of the linkers was not detectable by SDS-PAGE and thus no significant differences between modified lipases were observed. Figure 6.2. SDS-PAGE of native lipase from Thermomyces lanuginosus and modified lipases (TL1TL5). 6.3.2. Hydrolytic activity 6.3.2.1. Absolute activity The hydrolytic activity of a fixed amount of lipase (1 mg) was evaluated against seven substrates, differing in the size of the aliphatic chain (from 0 to 16 carbons) (Figure 6.3A). The results are expressed in U/mg, after incubation at 37 °C for 1 min (Figure 6.3B and 6.3C). From the results depicted in Figure 6.3, it is possible to perceive that the modification of lipase with the different linkers induced the improvement of the hydrolytic activity against substrates with differentiated chain-length, which are not normally hydrolysed by the native enzyme form. One may observe that for all the enzymes tested, the highest hydrolytic activity was achieved against the 103 medium-length chain substrates, p -nitrophenyl hexanoate ( p -NPH, 6C) and p -nitrophenyl octanoate ( p -NPO, 8C). Activity values higher than 100 U/mg were reached for all the lipases. An exception was observed for the enzyme TL5. This enzyme, modified with four dodecyl chains, displayed higher activity against all the substrates tested, comparing with the native lipase. For the small chain-length substrates, p -nitrophenyl acetate ( p -NPAc, 2C) and p -nitrophenyl butyrate ( p - NPB, 4C), the hydrolytic activity obtained was at least 3-fold higher than obtained with the native enzyme. For these two substrates, all other modified enzymes showed activity values similar to the native enzyme. It is also worth mentioning that this modified lipase (TL5) also displayed activity (13 U/mg) against a synthetic substrate ( p -nitrophenyl benzoate, 0C). For higher chain-length substrates ( p -NPL and p -NPP), from all the tested enzymes, TL5 displayed also the highest hydrolytic activity (2-fold). For p -NPL, TL2, TL3 and TL4 revealed higher activity than native lipase (2-fold of activity increase). These three enzymes displayed activity values of ≥100 U/mg, for substrates containing 6, 8 and 12 carbons in the aliphatic chain, while the native enzyme only reached this activity for p -NPH (6C) and p -NPO (8C). Lipase from Thermomyces lanuginosus is an enzyme composed by a lid over its active site, requiring interfacial activation. Moreover, the active site is surrounded by hydrophobic residues, which hinders its activity against small hydrophilic substrates. These features make lipase TL more able to hydrolyse longer substrates [151]. This lipase has a lysine residue near the lid (residue number 98 of the enzymes’ sequence), which is a plausible position for modification with the tested linkers. Given the linkers’ hydrophobicity, possible Van der Waals interactions with the lid and/or the hydrophobic surface surrounding the active site can occur, leading to different activity performances depending on the linker used. In this way, molecular dynamics tools were applied herein to infer the hydrophobicity of the linkers and correlate this parameter with the hydrolytic activity displayed by the modified lipases. 104 Figure 6.3. A) Proposed reactional scheme representing the hydrolysis of the ester substrates and the designation of each substrate according to the number of carbons in the aliphatic chain; B) Absolute hydrolytic activity of native lipase (TL) and of modified lipases with aromatic linkers (TL1 and TL2), and C) with aliphatic linkers (TL3-TL5), measured against different substrates: p - nitrophenyl benzoate ( p -NPPh, 0C), p -nitrophenyl acetate ( p -NPAc, 2C), pnitrophenyl butyrate ( p - NPB, 4C), p -nitrophenyl hexanoate ( p -NPH, 6C), p -nitrophenyl octanoate ( p -NPO, 8C), p -nitrophenyl laurate ( p -NPL, 12C) and p -nitrophenyl palmitate ( p -NPP, 16C). Figure 6.4 depicts the hydrophobic area of the linker versus the hydrolytic activity of the modified lipases against three different substrates, representative of short, medium, and long chain-length: p -nitrophenyl acetate ( p -NPAc), p -nitrophenyl hexanoate ( p -NPH) and p -nitrophenyl laurate ( p -NPL). From the data obtained it can be perceived that the hydrolytic activity of the enzyme modified with a specific linker was directly related with the percentage of its hydrophobic area. The highest overall activity was obtained for the lipase TL5, modified with the highly hydrophobic linker, 4 dodecyl chains (hydrophobic area higher than 70 %). We may assume that this linker can interact more favourably with the lid and/or with the hydrophobic residues surrounding the active site, eliminating the need for interfacial activation. These findings might explain the excellent catalytic performance of TL5 for all substrates, especially for the short-chain substrates. Together with an easier access 105 of the small substrates, a slight increase of the active site size may also occur, so that longer substrates can be better accommodate and further hydrolysed. All other linkers tested are less hydrophobic than the dodecyl chain differing in their hydrophobic area in about 10 % (between 55-65 %). TL3 (butyl chains) and TL4 (hexyl chains) were completely modified, which indicates that the lysine near the lid was alkylated. However, these linkers are less hydrophobic than the dodecyl chain (TL5) which can consequentially lead to a weaker interaction with the hydrophobic environment around the active site. We may infer that these modified enzymes still required interfacial activation, given that vestigial activities were achieved when used against the shorter chain substrates. The same assumption might be valid for TL1 and TL2 (both with a naphthyl moiety), in which only one lysine residue was modified. Comparing with native lipase the most pronounced activity differences regarding TL2, TL3 and TL4, were observed for p -NPL. For this substrate, the three enzymes showed activity values 2-fold higher than native enzyme. It was predicted that the modification of these lipases induced a slight change in the size cavity of the active site, which might explain the highest activity measured. Nevertheless, the effect of the cavity size was still less pronounced than the observed for TL5 when tested against longer substrates ( p -NPP, 16C). Figure 6.4. Activity of the modified lipases versus hydrophobic area of the linkers; activity measured against p -nitrophenyl acetate ( p -NPAc , • ), p -nitrophenyl hexanoate ( p -NPH ,  ) and p - nitrophenyl laurate ( p -NPL ,  ). The colour of the symbols in the graphs corresponds to the colour of each modified enzyme represented bellow the graph. 106 TL1 did not displayed superior activity than the native enzyme in neither of the tested substrates. This modified lipase was only mono substituted with an isothiocyanate group, in a lysine placed away from the active site enabling any destabilization of the lid nor of the active site cavity. It is important to state that the presence of the free aldehyde or isothiocyanate in the medium did not induced any alteration on the enzymes’ activity, confirming that the performance differences observed are induced by the surface modifications undertaken. In chapter 5, we PEGylated lipase TL, through modification of the surface lysines with a monofunctional aldehyde-PEG (5000 Da). The modified lipase displayed a higher polymerase activity for the synthesis of a polyester, poly(ethylene glutarate), comparing to the native form. However, the hydrolytic activity of the modified lipase against p -nitrophenyl butyrate, remained similar to the native enzyme [81]. The surrounding PEG allowed an easier access of the monomers to the active site, nevertheless, the substrate used for hydrolytic activity evaluation revealed similar accommodation in the active site for both native and PEGylated form. Contrarily to PEG, which is an hydrophilic macromolecule, the linkers herein studied are small hydrophobic compounds which, depending on their size, may influence the enzyme conformation and therefore its hydrolytic activity. As postulated by some authors, a proportional increment of the enzymes’ activity with the degree of modification would be expected [157]. However, for the lipase studied, this behaviour was not observed, since higher levels of modification did not correspond to the best performance results (comparing TL5 with TL3 and TL4). One may assume that the activity was not only dependent on the level of modification but was extremely influenced by the linker type, chain size, hydrophobicity, and positioning at the enzyme’s spatial. 6.3.2.2. Kinetic parameters The kinetic parameters of the lipase modified with the different linkers were evaluated against different substrates: p -nitrophenyl acetate, p -nitrophenyl butyrate and p -nitrophenyl octanoate (Table 6.2). The substrates were chosen based on the preliminary activity results obtained for the modified lipase, where an increase of the enzymes’ activity over the grown of the substrate chain length was observed. Given the data obtained from Table 6.2, and as expected, the best kinetic parameters were obtained for p -NPO. For this substrate, the highest Vmax was achieved with TL1, however, with a low K M and catalytic turnover, revealing that this enzyme displayed the lowest catalytic performance for p -NPO, comparing to all other enzymes tested. The K M values of TL2, TL3, TL4 107 and TL5 are similar to the values of the native enzyme, being the catalytic turnover similar or lower after modification. For the shorter substrates, TL5 displayed the best catalytic performance. The lowest K M value (23 mM) was obtained for p -NPB, being 5-fold lower than the K M value obtained for the native enzyme (110 mM). Moreover, the highest catalytic turnover (Figure 6.5) was observed for this modified enzyme against p -NPB (η= 23364 M-1 s-1), which was 9-fold higher than the value obtained for the native enzyme (η= 2552 M-1 s-1). All other modified enzymes showed a catalytic turnover similar to the native lipase. For p -NPAc, the K M value of TL5 (128 mM) was half of the K M value of the native enzyme (245 mM). This modified enzyme revealed a catalytic turnover 6-fold higher than TL, whereas no significant alterations of the kinetic values were verified for the other modified enzymes. As previously stated, lipase from Thermomyces lanuginosus is usually used for hydrolytic purposes in its immobilized form. The immobilization strategy increases its specific activity comparing to its free form, however some kinetic parameters are negatively affected [169, 170]. The strategy here applied does not restrict the enzymes’ conformation and, in some cases, even improves its kinetic parameters (TL5). Table 6.2. Kinetic parameters of native and modified lipases (Vmax (mol/mg/min), K M (mM) and η= k cat/ K M (M-1 s-1)) calculated for the hydrolysis of p -nitrophenyl acetate ( p -NPAc), p -nitrophenyl butyrate ( p -NPB) and p -nitrophenyl octanoate ( p -NPO) under the conditions: substrate concentration varied between 1 and 350 mM, enzyme content (1 mg), performed at 37 °C for 1 min p -nitrophenyl acetate ( p -NPAc) p -nitrophenyl butyrate ( p -NPB) p -nitrophenyl octanoate ( p -NPO) Enzyme Vmax K M (mM) η (M-1 s-1) Vmax K M (mM) η (M-1 s-1) Vmax K M (mM) η (M-1 s-1) TL 752 245 1517 566 110 2552 1426 56 12482 TL1 610 214 1412 626 129 2402 1671 125 6612 TL2 680 223 1516 335 63 2656 1022 45 11340 TL3 780 299 1316 554 126 2218 701 42 8439 TL4 503 253 1005 708 229 1563 610 34 9018 TL5 2535 128 9828 1061 23 23364 1391 53 12979 108 Figure 6.5. Catalytic turnover (η= k cat/ K M) of the native and modified enzymes calculated for the hydrolysis of p -nitrophenyl acetate ( p -NPAc, 2C), pnitrophenyl butyrate ( p -NPB, 4C) and p - nitrophenyl octanoate ( p -NPO, 8C). The kinetic profile of lipase TL is described to be improved as the chain-length of the substrates grows [171], as we also confirmed. Moreover, the modified enzymes obtained, revealed better kinetic performance than the native enzyme, mainly for the shorter substrates. The kinetic parameters of the modified lipases are, as mentioned previously, greatly influenced by the type of linker attached and its positioning at the enzymes’ surface. The activity data also indicated that the lack of hydrophobicity of the small substrates seems to be counterbalanced by the hydrophobic character of the linker, as the highest activity differences were observed for smaller and less hydrophobic substrates. For longer and hydrophobic substrates, the impact of the enzyme modification (TL5) seems to be irrelevant since the access to the enzymes’ active site was always ensured (Figure 6.5). In order to predict the enzyme performance and better understand the conformational changes induced by the modifications, molecular dynamics simulations were performed on TL and TL5. 6.3.3. Molecular Dynamics Simulations MD simulations were carried out using a similar strategy applied on a previous work using the same enzyme [81]. During simulations, no modifications were carried out for native TL (PDB ID: 1TIB), whereas for TL5, the enzyme was modified at Lys98 with a dodecyl chain (Figure 6.6A and 0 5000 10000 15000 20000 25000 2C 4C 8C kcat/KM(M-1s-1) Length of the substrate chain (Nº of carbons) TL TL1 TL2 TL3 TL4 TL5 109 C, respectively). This specific lysine is in one of the lid arms and any modification may cause important alterations of the lid arrangement or of the size/opening of the active site, which is crucial for the enzymes’ activity. Figure 6.6 highlights the lid, the catalytic triad, and the linker in the case of TL5 (Figure 6.6C). In these middle structures we observe that the helical lid undergoes some unfolding on its structure in both enzymes, nevertheless leading to a larger cavity for TL5 (yellow filling), in which the amino acids side chain positioning also contributes to a well-defined cavity. This conformational change was followed through RMSD and RMSF analysis (Figure 6.7), where a RMSF peak is seen around the residue 98 in TL5. Globally, the two lipases are stable in aqueous medium. Figure 6.6. Middle structures characterized for lipase TL (A) and TL5 (C); Lid is highlighted in cyan, catalytic triad with the residues in green sticks and TL5 in blue sticks (C); (B) and (D) zoom in the active site, showing the pockets/cavities in yellow surface. 116 6.3.5. Circular dichroism Circular dichroism (CD) was undertaken to evaluate the effect of the modifications performed, on the conformational structure of the enzymes (Figure 6.12). The CD spectra of the native enzyme showed the expected behaviour, described in literature [175]. Lipase TL is structurally composed by a central eight-stranded, mostly by parallel beta-sheets and five interconnecting alpha helices [151]. All modified enzymes (TL1 to TL4) revealed a profile similar to the native TL, with no significant modifications in the number of α-helices and β-sheets. Given the experimental results obtained for these enzymes, these results were expectable. TL5 revealed the most discrepant profile, displaying the lowest intensity. The decrease in the intensity can be related to a slight unfolding of the protein’s structure. Considering the experimental activity results, one may infer that the access of the active site, was facilitated by the modification with the dodecyl chain at the enzymes’ lid, which enhanced its activity for short and long substrates. This behaviour was also observed by molecular dynamics simulation that confirmed the active site enlargement and a consequent easy access of the substrates. Figure 6.12. Circular dichroism of native lipase (TL) and modified lipases (TL1 to TL5). 6.3.6. Fluorescence analysis The intrinsic fluorescence of the lipases, related to the presence of fluorophore residues, was considered, as a strategy to observe conformational fluctuations [176] induced by the compounds grafted to the lysine residues (Figure 6.13). Comparatively to native lipase, all the modified lipases -20 -10 0 10 20 30 185 205 225 245 CD [mdeg] nm TL TL1 TL2 TL3 TL4 TL5 117 showed a decrease of the fluorescence intensity, which is usually associated to an unfolding of the protein. However, the fluorescence intensity can be also affected by other factors, such as morphology, protein size and protein aggregation [177]. By circular dichroism it was only possible to observe a slight unfolding of TL5, indicating that the decrease of fluorescence that was observed for all other modified lipases (comparing to the native form), might be related to other factors. Figure 6.13. A) Fluorescence spectra of native (TL) and modified lipases (TL1-TL5) after excitation at 280 nm. B) Table of the maximum wavelength with the respective intensity value. A hypochromic blue shift (from 320 to 316 nm) was observed for TL5, while TL3 and TL4 showed their maximum intensity at the same wavelength as the native enzyme, or at a higher wavelength (TL1 and TL2). The blue shift is frequently associated to the exposure of the amino acids to a more hydrophobic microenvironment, in which the fluorophore residues are more internalized [178]. The tryptophan residue at the lid of lipase TL is one of the amino-acids responsible for the enzymes’ fluorescence [138]. Changes in the microenvironment of this residue may influence the fluorescence, which given the differences obtained in terms of activity, might indicate that the lid of the enzyme was modified. Since TL5 is modified with the most hydrophobic linker, its fluorescence decreases greatly, induced by a more hydrophobic microenvironment. 118 6.4. Conclusions In this work, chemical modifications of lipase from Thermomyces lanuginosus with different linkers (isothiocyanates and aldehydes) were for the first time explored to improve activity, stability, and affinity to differentiated substrates. The results showed that aldehyde linkers were more prone to modify the lysines of lipase comparing with the isothiocyanates. One concluded that the size and hydrophobic character of the linkers influenced greatly the enzyme activity. The longer and hydrophobic is the linker, the stronger is its interaction with the hydrophobic amino acid residues near the active site, conducting to a destabilization of the lid and to an enlargement of the active site’s cavity. This is expected to induce an improvement of the lipase performance. The implementation of the easy methodology developed, broadening the substrates chain-length specificity, increments thereafter the range of substrate possibilities that can be hydrolysed by this lipase, paving the way to the establishment of new industrial applications. 119 Chapter VII Chemical modification of lipase from Thermomyces lanuginosus enhances transesterification and esterification activity 120 Chapter VII Chemical modification of lipase from Thermomyces lanuginosus enhances transesterification and esterification activity Abstract Lipase from Thermomyces lanuginosus is one of the most explored enzymes for the esterification of several added-value industrial compounds. The modified form (grafted with 4 dodecyl chains, TL5) revealed previously higher hydrolytic activity against size-differentiated substrates, compared with the native enzyme. In this work, we evaluated the transesterification and esterification activity of native and modified lipase, using p -nitrophenyl palmitate and oleic acid as model compounds, respectively. Linear size-differentiated alcohols (from 1 to 20 carbons in the aliphatic chain) were used to explore for the first time the effect of the chain length in both transesterification and esterification reactions. The chemically modified lipase showed greater catalytic performance, than the native enzyme, being this increase directly proportional to the size of the alcohols chain used as substrates. The enormous potential and remarkable versatility of this chemically modified lipase was here demonstrated, where diverse types of esters, differing in their potential applications, were efficiently synthesized. The produced esters were fully characterized by 1H NMR, GC-MS, and FTIR. This chapter is based on the following publication: Jennifer Noro, Artur Cavaco-Paulo, Carla Silva, Chemical modification of lipase from Thermomyces lanuginosus enhances transesterification and esterification activity , submitted to: ACS Catalysis. 121 7.1. Introduction Lipase from Thermomyces lanuginosus (TL) was the first recombinant lipase produced [179], and nowadays, it remains one of the most explored enzymes for the synthesis of compounds with industrial value. Besides their natural triacylglycerols hydrolysis function, it also demonstrates esterification and transesterification activity [151]. Ester compounds are present in many daily products, being the major components of flavours, fragrances, polymers, fats, among others [180]. The use of lipases for their synthesis is considered a green and environmentally friendly practice, regarding the high reactional yields and the mild reaction conditions associated [181]. Among the different lipases, lipase TL is one of the most explored for transesterification and esterification reactions. Ashrafuzzaman and co-workers observed that immobilized lipase TL demonstrated the highest regioselectivity in the acylation of sucrose esters comparing to other lipases [182]. The transesterification reaction was carried out using divinyl esters. A similar approach was undertaken by Chávez-Flores et al. [183], in the regioselective transesterification of vinyl laurate with a probiotic sugar, lactulose, using the same form of enzyme. Corrêa and coworkers [184] studied the esterification of the flavours, geraniol and citronellol, promoted by immobilized lipase TL. The reactions were carried out by coupling oleic, lauric, and stearic acid, being the produced esters isolated in good yields (>60 %) [184]. The esterification of oleic acid with isoamyl alcohol was performed by Lage et al. [185]. The authors used lipase TL immobilized onto polymethacrylate particles as reaction catalyst. The ester, isoamyl oleate, which can be used as a biolubricant, was successfully synthesized with high conversion (85 %). Another major field of application of this lipase is in the production of biodiesel. The extensive investigation reported in this area stems from the world demand to find green and renewable sources of fuels [186]. Biodiesel can be manufactured through enzymatic catalysis, by the transesterification or esterification of oils/fats with small-length alcohols, such as methanol or ethanol. Countless sources of oils/fats can be used for this purpose. Sunflower, coconut, soybean, palm, and cotton seed oils or even wastes from the food industry, with varied composition, are some of the potential sources [187-189]. As described previously, most reports regarding transesterification and esterification reactions describe the use of lipase TL in its immobilized form. The immobilization of enzymes presents several advantages, in comparison to other methodologies, including reusability, thermostability, and others [190]. However, immobilization techniques, besides being a more expensive strategy, 122 it reduces the enzyme mobility and consequently, its catalytic performance [191]. Many reports form literature have been describing the chemical modification of lipases as an efficient methodology to overcome the limitations associated to the immobilization methodologies [157, 158, 192]. However, a lack of practical examples regarding its implementation on the synthesis of industrial added-value products can be found. In chapter 5, we have reported that the PEGylation of lipase TL improved its polymerase activity comparing with the native enzyme. Higher degree of polymerization and conversion yield were obtained in the biosynthesis of a polyester, poly(ethylene glutarate) [81]. Previously, in chapter 6, we performed the chemical modification of lipase TL by grafting small hydrophobic aldehydes and isothiocyanates to the exposed lysine residues at the enzymes’ surface. Besides increasing their thermostability, the modification of the enzyme with aliphatic aldehydes showed to improve their activity in the hydrolysis of differentiated chain-length substrates [192]. Lipases displaying both esterification and transesterification selectivities for short and long alcohols are difficult to found, especially to obtain high reactional yields. In this work, the transesterification and esterification activity of the native and modified (grafted with four dodecyl chains, TL5) lipase TL were explored by investigation of their activity towards differentiated chain-length alcohols. p -Nitrophenyl palmitate and oleic acid were used as model lipids for the transesterification and esterification reactions, respectively. A broad range of alcohols (from methanol to eicosanol, in a total of 11 alcohols) were used as substrates for the evaluation of the activity of both enzyme forms (native vs modified). 7.2. Materials and methods 7.2.1. Materials Lipase from Thermomyces lanuginosus , butanol, hexanol, oleyl alcohol, p -nitrophenol, molecular sieves 4 Å pellets (1.6 mm diameter), were purchased from Merck. p -nitrophenyl palmitate was obtained from Santa Cruz Biotechnology. Oleic acid, methanol, ethanol, propanol, pentanol, heptanol, decanol, dodecanol and eicosanol were purchased from TCI Chemicals. n -Heptane was acquired from Fischer Chemicals (HPLC grade) and dried over molecular sieves prior to usage. 123 7.2.2. Chemical modification of lipase TL The chemical modification of the lipase was performed as previously described in chapter 6 [192]. The modified lipase TL (TL5) was isolated with four grafted dodecyl chains. Native lipase TL was used after ultrafiltration for the removed of any additives. Both enzymes were used in their lyophilized form. 7.2.3. Half-life time of the enzymes (T1/2) The half-life time of the lipases (native and modified) was evaluated at different temperatures (25, 37 and 50 ºC). Prior to the evaluation of the transesterification/esterification activity, the T1/2 of the enzymes was accessed to ensure that no significant loss of activity occurred at the different temperatures tested. For this, the lipases (1 mg/mL), dissolved in phosphate buffer (pH 7.8, 50 mM), were placed in a water bath, under different temperatures (25, 37 and 50 °C). At different time intervals, the hydrolytic activity was measured against p -nitrophenyl hexanoate. The assay was executed as previously described [192]. Afterwards, the T1/2 was calculated as reported [165]. The measurements were performed in a Synergy Mx Multi-Mode Reader from BioTek (USA) in a 96 well plates. One unit of enzyme activity was defined as the amount of enzyme which catalyses the production of 1 μmol p -nitrophenol from the initial substrate per minute. 7.2.4. Transesterification activity The transesterification activity of both enzymes (native and grafted with four dodecyl units) was measured using different alcohols, to evaluate the effect of the chain length on the final activity of the enzymes. The reactions were performed following the procedure reported by Teng and Xu [193]. Briefly, the enzymes (10 mg), were added to a flask containing 10 mL of a 10 mM solution of p -nitrophenyl palmitate in dry n -heptane. Then, 60 L of 1 M of the respective alcohol was added to the flask and placed in a water bath at 37 °C, under stirring (150 rpm). Aliquots of 30 L were withdrawn at different time intervals and quenched with 1 mL of NaOH (0.1 M). Then, 200 L were placed in a 96-well plate and the p -nitrophenol released was read at 400 nm. The enzymes activity was then calculated by plotting the p -NP released over time. 124 7.2.5. Esterification activity In a flask containing the enzyme (native or modified, 0.08, 0.17 or 0.25 % w/v) was added the oleic acid (300 L) and the respective alcohol (1 equivalent) in 3 mL of n -heptane. The suspension was placed at the desired temperature (25, 37 or 50 ºC) under stirring. At different time intervals, 100 L of the solution was taken, and the volume made up until 5 mL with a solution of ethanol/acetone 1:1. This solution was then titrated with NaOH 20 mM, using phenolphthalein 0.5 % w/v as indicator. K values were then calculated for both reactions, regarding the following equation: 𝐾 = [𝐸𝑠𝑡𝑒𝑟 𝑝𝑟𝑜𝑑𝑢𝑐𝑡] [𝑆𝑡𝑎𝑟𝑡𝑖𝑛𝑔 𝑚𝑎𝑡𝑒𝑟𝑖𝑎𝑙] 7.2.6. Products characterization 7.2.6.1. Nuclear Magnetic Resonance spectroscopy (NMR) After both transesterification and esterification reactions, the n -heptane and the volatile alcohols were completely removed in the rotary evaporator (Heidolph, Germany). Chloroform was added, and the solution washed with 5 % NaHCO3 solution (3x) followed by water (2x). The organic layer was dried over MgSO4, filtered and the solvent removed in the rotary evaporator to afford the pure product as a white solid (palmitate esters) and colourless oil (oleate esters). 1H NMR (400 MHz) was then performed dissolving the products in deuterated chloroform (CDCl3), and the samples analysed in a Bruker Avance III. 7.2.6.2. Gas Chromatography – Mass Spectrometry (GC-MS) GC was performed using a Bruker SCION 436 system with a split/splitless injector coupled to a mass spectrometer (MS). Injections were carried at 250 °C in the split mode 1:10 using a Rxi-5Sil MS (Restek) column (30 m × 0.25 mm, and 0.25 μm film thickness), with a column-head pressure of 7.3 psi using helium as carrier gas. The oven temperature started at 150 °C and was held for 3 min, and the temperature increased until 280 °C at a rate of 7 °C/min. A full scan mode (50−600 m/z) was applied for the identification of the target compound. The mass spectrometer (MS) was operated in electron ionization (EI) mode at 70 eV with total ion chromatogram detection mode for quantitative determination and S / N ratio of 5. 125 7.2.6.3. Fourier-transform infrared spectroscopy (FTIR) Infrared spectra were recorded on a FTIR Platinum-ATR Bruker Alpha II. The samples were analysed over the range 400-4000 cm-1, with a spectral resolution of 4 cm-1. All spectra were an average of over 24 scans. 7.3. Results and discussion 7.3.1. Chemical modification of lipase TL Lipases are ubiquitous enzymes with outstanding catalytic properties. Besides their natural function for the hydrolysis of triacylglycerols into glycerol and fatty acids, they can perform the opposite reaction, the biosynthesis of esters. Esters are found in countless type of compounds, and are of great importance in the food industry, cosmetics, pharmaceutics, etc [184, 194]. In chapter 6, we reported the modification of lipase TL with different aliphatic/aromatic aldehydes and isothiocyanates and tested there hydrolytic activity against 7 differentiated chain-length p - nitrophenyl substrates [192]. It was observed that the lipase modified with four dodecyl chains (TL5) (Figure 7.1A II) showed an improved hydrolytic activity for all tested substrates (up to 2-fold), comparing to the native enzyme. The degree of modification was accessed through MALDI-TOF analysis (Figure 7.1B) and the TNBSA assay. Given the promising results obtained, herein we aim to explore the catalytic activity of this modified enzyme for the production of industrial added-value esters. Prior to the evaluation of the transesterification and esterification activity of both enzymes, their half-life time at different temperatures was assessed (25, 37 and 50 ºC) (Figure 7.1C). The temperatures were chosen according to the optimum temperature range of activity of the studied lipases and considering the typical array of temperatures reported for the evaluation of esterification reactions. The data revealed that at 25 and 37 °C, both enzymes showed remarkable stability until 2 months of incubation. At 50 ºC both enzymes displayed high stability, with half-life times near 200 h. In this way, both lipases showed to be suitable catalysts for longer reactional times.