Exploring coaggregation mechanisms involved in biofilm formation in drinking water through a proteomic-based approach

Journal of Applied Microbiology , 2024, 135 , lxae143 https://doi.org/10.1093/jambio/lxae143 Advance access publication date: 14 June 2024 Research Article Exploring coaggregation mechanisms involved in biofilm formation in drinking w at er through a proteomic-based approach Ana C. Afonso 1 , 2 , 3 , 4 , Manuel Simões 1 ,2 , Maria José Saa v edr a 3 , Lúcia Simões 4 , Juan M. Lema 5 , Alba Trueba-Santiso 5 ,* 1 LEPABE—Laboratory for Process Engineering, Environment, Biotechnology and Energy, Faculty of Engineering, University of Porto, , Rua Dr Roberto Frias, 4200-465 Porto, Portugal 2 ALiCE—Associate Laboratory in Chemical Engineering, Faculty of Engineering, University of Porto, Rua Dr Roberto Frias, 4200-465 Porto, Portugal 3 CITAB, Department of Veterinary Sciences, University of Trás-os-Montes e Alto Douro, 5000-801 Vila Real, Portugal 4 CEB-LABBELS, University of Minho, Campus de Gualtar, 4710-057 Braga, Portugal 5 CRETUS, Department of Chemical Engineering, University of Santiago de Compostela, Campus Vida, 15782 Santiago de Compostela, Galicia, Spain ∗Corresponding author. CRETUS, Department of Chemical Engineering, University of Santiago de Compostela, Campus Vida, 15782 Santiago de Compostela, Galicia, Spain. E-mail: [email protected] , Abstract Aim: Coaggregation, a highly specific cell–cell interaction mechanism, pla y s a pivotal role in multispecies biofilm formation. While it has been mostly studied in oral environments, its occurrence in aquatic systems is also acknowledged. Considering biofilm formation’s economic and health-related implications in engineered water systems, it is crucial to understand its mechanisms. Here, we hypothesized that traceable differences at the proteome le v el might determine coaggregation ability. Methods and Results: Two strains of Delftia acidovorans , isolated from drinking water were studied. First, in vitro motility assays indicated more swarming and t witching motilit y for the coaggregating strain (C + ) than non-coaggregating strain (C −). By transmission electronic microscopy, we confirmed the presence of flagella for both strains. By proteomics, we detected a significantly higher expression of type IV pilus twitching motility proteins in C + , in line with the motility assa y s. Moreo v er, flagellum ring proteins w ere more abundant in C + , while those in v olv ed in the formation of the flagellar hook (FlE and FilG) were only detected in C −. All the results combined suggested str uct ural and conformational differences between stains in their cell appendages. Conclusion: This study presents an alternative approach for identifying protein biomarkers to detect coaggregation abilities in uncharacterized strains. Impact Statement This study introduces a new method for identifying protein biomarkers to detect coaggregation abilities in aquatic bacterial strains. It sheds light on the complex molecular mechanisms behind this phenomenon, offering an alternative to traditional visual assays. These insights could inform inno v ativ e biotechnological processes and eco-friendly antibacterial solutions, addressing c hallenges suc h as biofilm formation and improving microbial management strategies for water quality enhancement. Ke yw or ds: aggregation; cell–cell interaction; cellular appendages; multispecies biofilm; proteomics; TEM Introduction Microbial biofilm communities are widely distributed in nature, having complex interspecies interactions that drive spatial species distribution and behavior (Booth and Rice 2020 , Shokeen et al. 2021 ). Among these interactions, coaggregation is a peculiar type, recently attracting research efforts. It consists of a high level of cell–cell interaction, characterized by a specific mechanism of recognition and adhesion of different bacterial species to each other (Rickard et al. 2003 ). Coaggregation is usually included in the biofilm formation process, facilitating structural and metabolic codependencies, and consequently contributing to the development of complex multispecies biofilm communities (Elliott et al. 2006 ). This phenomenon has mostly been studied in oral environments (Stevens et al. 2015 , Jacob and Reguera 2022 ). Nevertheless, it has also been recognized in aquatic systems (Rickard et al. 1999 ). So far, studies on coaggregation in aquatic systems have focused on the identification of coaggregating species (Rickard et al. 2002 , Simões et al. 2008 , Vornhagen et al. 2013 , Stevens et al. 2015 , Afonso et al. 2023 ), characterization of the type of molecules involved in this mechanism (Rickard et al. 2000 , Simões et al. 2008 , Afonso et al. 2023 ), and the study of environmental factors that influence coaggregation (Rickard et al. 2000 , 2004 , Ishii et al. 2005 , Min et al. 2010 ). The molecular mechanisms behind coaggregation are poorly understood in aquatic systems. Previous studies indicated the role of saccharide-protein (Rickard et al. 1999 , 2000 , Received 7 May 2024; revised 6 June 2024; accepted 13 June 2024 ©The Author(s) 2024. Published by Oxford University Press on behalf of Applied Microbiology International. This is an Open Access article distributed under the terms of the Creative Commons Attribution License ( https:// creativecommons.org/ licenses/ by/ 4.0/ ), which permits unrestricted reuse, distribution, and reproduction in any medium, provided the original work is properly cited. Downloaded from https://academic.oup.com/jambio/article/135/6/lxae143/7693735 by Irene Ruiz user on 10 July 2024 2 Afonso et al. Simões et al. 2008 , Afonso et al. 2023 ) and protein–protein interactions (Daep et al. 2008 , Afonso et al. 2023 ). In the case of oral species, a discrete number of proteins involved in coaggregation have even been identified (Kaplan et al. 2009 , Mishra et al. 2010 , Coppenhagen-Glazer et al. 2015 , Zhou et al. 2015 , Khalil et al. 2020 ), yet the same is not true for aquatic systems (Afonso et al. 2021 ). Therefore, it seems relevant to investigate the proteomic profile of coaggregating bacteria from these environments. To gain insight into the mechanisms of bacterial coaggregation in aquatic systems is vital due to its implications in biofilm formation, which can derive in severe challenges. In engineered systems such as those applied for aquaculture or drinking water (DW) potabilization and conduction, it may cause biological fouling and increase maintenance costs. Furthermore, it poses health-related challenges, as shifts in microbial communities can be unpredictable, potentially favoring pathogens or horizontal genetic transfer of antibiotic resistances (Afonso et al. 2021 ). The derived knowledge can be used to assist on the design of new biotechnological processes or new eco-friendly antibacterial agents. On the other hand, coaggregation can be desired in certain water treatment processes, such as granular biomass bioreactors. The elucidation of its mechanisms can provide crucial insights for innovative microbial management strategies, oriented to improving water quality, reducing pathogenic risks and economic costs (Afonso et al. 2021 ). Coaggregation research can also contribute to developing new eco-friendly bioremediation solutions for polluted ecosystems (Malik et al. 2003 ). The detection of coaggregation abilities, defined as the capacity for aggregation of genetically distinct cells (i.e. interspecies and interstrain), has traditionally been based on the isolation of bacteria followed by laboratory tests (visual coaggregation assays) (Afonso et al. 2021 ). This approach requires defined culturing conditions (Rickard et al. 2002 ) not applicable to most uncharacterized environmental species of interest, such as strict anaerobes or sulfur-reducing bacteria, among many others. The identification of protein biomarkers of coaggregation would be the basis for a proteomic-based alternative to the more conventional procedures. Also, it would be of interest to work with mixed communities, avoiding an extensive and sometimes challenging isolation work. At the same time, it might allow to investigate the bacteria sampled directly in their real environments. Mass spectrometry (MS)-based proteomics is a powerful tool for the individual identification and quantification of proteins present in complex proteomes, such as those extracted from a whole cell, tissue, or bacterial culture (Zhang et al. 2013 ). Proteomics also allows to compare expression levels of the identified proteins in different moments or organisms (Torres-Sangiao et al. 2022 ). Previously, proteomic investigations contributed to identify molecular mechanisms triggered upon other types of bacterial interactions between isolated bacterial species (Shokeen et al. 2021 ). In a previous study, 38 different bacterial strains were isolated from DW and screened in combinations for their coaggregation ability (Afonso et al. 2023 ). One strain, Delftia acidovorans 005P, selectively coaggregated with partner bacteria ( Citrobacter freundii 002 L, C. freundii 003 L, Pseudomonas fluorescens 008P , P . putida 011P , and Enterobacter cloacae 023 L), and facilitated biofilm development. Protein–protein interactions were demonstrated to be involved in the process (Afonso et al. 2023 ), however, the exact molecules involved remain to be determined. Here, we selected the D. acidovorans 005P strain, which has coaggregation ability, and the D. acidovorans 009P strain, which lacks this ability, to evaluate the feasibility of a proteomic approach for identifying the molecular mechanisms involved in cell–cell interactions. We define “noncoaggregating” as the absence of aggregation with genetically distinct cells (Afonso et al. 2023 ). For this purpose, we applied in vitro motility analyses and transmission electronic microscopy (TEM) characterization followed by in solution shotgun MS proteomics. This study offers a novel perspective to unravel the intricate molecular mechanism of bacterial coaggregation in strains from aquatic systems while the derived information can set the foundation for new biotechnological processes. Materials and methods Bacterial strains and culture conditions Two strains of D. acidovorans isolated from the Northern of Portugal (Afonso et al. 2023 ) were cryopreserved at −80 ◦C, in aliquots of BHI (Brain–Heart Infusion) medium (Oxoid, UK) with 15% (v/v) of glycerol. Both strains are deposited in the publicly accessible culture collection MUM (Micoteca da Universidade do Minho, Braga, Portugal) with accession codes MUM 24.11 ( D. acidovorans 005P) and MUM 24.12 ( D. acidovorans 009P). Here, the strains are now referred to as C + and C −, with the positive or negative sign indicating their ability to coaggregate. Specifically, strain 005P, which can coaggregate, is now referred to as C + , and strain 009P, which cannot coaggregate, is referred to as C −. Bacterial cells were grown in batch culture using R2A broth at room temperature (23ºC ±2), under agitation (150 rpm), until reaching the stationary growth phase. The stationary phase of growth was selected because coaggregation is growth-phase-dependent, being the maximum expressed in the stationary phase (Rickard et al. 2000 ). In vitro motility assessment Swimming, swarming, and twitching motilities were assessed as previously described (Gomes et al. 2019 , Booth and Rice 2020 ). For that, D. acidovorans strains were grown overnight and the cell density was adjusted to 5 ×10 6 cells ml −1 in fresh medium. A volume of 15 μl of bacterial suspension was then dropped in the center of agar plates prepared with tryptone at 10 g l −1 (Fisher Bioreagents, New Jersey, USA), NaCl at 2.5 g l −1 and agar (VWR Chemicals, Leuven, Belgium) at 3 g l −1 , 7 g l −1 , or 15 g l −1 , for swimming, swarming, and twitching motilities, respectively. Then, plates were incubated at 25ºC. The colonies growth was measured at 24, 48, and 72 h after incubation. Three independent assays were performed with three replicates. Transmission Electron Microscopy Bacterial cells grown until the stationary phase of growth were washed three times in saline solution (NaCl 8.5 g l −1 ) and resuspended in ultrapure water at a concentration of 10 8 CFU ml −1 . For negative staining, samples were adsorbed to glowdischarged carbon-coated collodion film on 400-mesh copper grids. Then, the grids were washed with deionized water and stained with 1% uranyl acetate. Visualization was performed Downloaded from https://academic.oup.com/jambio/article/135/6/lxae143/7693735 by Irene Ruiz user on 10 July 2024 Proteomic analysis of DW bacteria 3 at 80 kV in a JEOL JEM 1400 microscope (Japan), and digital images were acquired using a CCD digital camera Orious 1100 W (Tokyo, Japan). The TEM was performed at the HEMS core facility at i3S, University of Porto, Portugal. Proteome extractions Cellular proteome extraction To prepare samples for proteome extraction, triplicate 5 ml aliquots of cultures were centrifuged at 3700 ×g for 10 min at 4ºC. The resulting pellets were utilized for cellular proteome extraction as described by Kennes-Veiga et al. ( 2022 ), while the supernatants were reserved for exoproteome extraction. Initially, the pellets were washed with phosphate buffered saline (PBS) (Sigma–Aldrich, Spain) and then resuspended in extraction buffer [50 mmol l −1 Tris buffer (Sigma–Aldrich, Spain), 1% SDS (Aplichem Panreac, USA), pH 7.5]. Subsequently, the samples were incubated at 90ºC for 20 min at 1000 rpm in a thermoshaker incubator (MS-100, Thermo- Shaker, LabGene Scientific, Switzerland). Cell disruption was achieved by transferring the samples to bead-beating tubes containing glass beads and subjecting them to four cycles of bead beating (3 min each) with 1 min intervals on ice using a cell disruptor (Scientific Industries, USA). After centrifugation at 3700 ×g for 20 min at 4ºC, the supernatants were collected. Protein concentration and purification were accomplished by precipitating the proteins with ice-cold acetone (ThermoFisher Scientific, USA) through at least two rounds of precipitation. Following centrifugation at 9000 ×g for 10 min at 4ºC, the supernatants were discarded, and the resulting pellets were resuspended in molecular grade water (MGW) (Sigma–Aldrich, Spain). Exoproteome extraction The supernatants derived from cell centrifugation underwent sterile filtration using a 0.22- μm Millex-GV PVDF filter (13 mm) (ThermoFisher Scientific, USA). Following this step, proteins were precipitated using the ice-cold acetone method as previously outlined and then reconstituted in MGW. Protein quantification by bicinchoninic acid assay assay Protein quantification was performed using the Pierce™ bicinchoninic acid assay Protein Assay Kit (ThermoFisher Scientific, USA), following the manufacturer’s instructions and utilizing a bovine serum albumin standard curve. Statistical analysis of the protein quantification data was conducted using SPSS version 28.0 (IBM Corp., USA). The significance of differences was assessed using the Wilcoxon statistical test and t -test, with a significance level set at P < 0.05 and calculations based on a confidence level of ≥95%. The results, presented as mean ±standard deviation, were obtained from three independent extractions and triplicate measurements. Triplicates were pooled together before further proteomic analyses. SDS-PAGE electrophoresis To verify the quality of the extracted samples, sodium dodecyl sulfate-polyacrylamide gel electrophoresis in denaturing conditions (SDS-PAGE) was conducted. Duplicate aliquots containing a maximum of 10 μg of protein per well were subjected to electrophoresis. Protein samples were prepared by mixing with NuPAGE LDS Sample Buffer and NuPAGE Reducing Agent following the manufacturer’s guidelines. A molecular weight marker, the PageRulerTM Plus Prestained Ladder ranging from 10 to 250 kDa (ThermoFisher Scientific, USA), was included for reference. Electrophoresis was carried out on a Bis-Tris NuPAGE 4%–12% gel (ThermoFisher Scientific, USA) at 200 V. Visualization of protein bands was accomplished using a standard blue Coomassie staining protocol. MS-based proteomics Proteins in each sample were quantified using a label-free method (Zhang et al. 2013 ). Initially, samples underwent trypsin digestion, reduction, alkylation, and subsequent desalting using ZipTipμC18 material (Merck, Germany). The resulting peptide samples were subjected to in-solution shotgun proteomics analysis (Zhang et al. 2013 ). Peptide samples containing 0.2 μg of protein were loaded onto a timsTOF Pro mass spectrometer (Bruker, Bremen, Germany) equipped with a nano-electrospray source (CaptiveSpray) and a tims- QTOF analyzer. Chromatographic separation was performed using a nanoELUTE chromatograph (Bruker) with an Aurora analytical column (C18, 250 ×0.075 mm, 1.6 μm, 120 ˚ A, IonOpticks). The nanoHPLC system utilized binary mobile phases: solvent A (0.1% formic acid in miliQ H2O) and solvent B (0.1% formic acid in acetonitrile). The analysis spanned 105 min, with a gradual increase in the B/A solvent ratio. Blank injections, lasting 60 min each, were interspersed between samples to ensure no carry-over. MS acquisition utilized collision-induced dissociation fragmentation and nanoESI positive ionization mode. PASEF-MSMS scan mode was employed for acquisition within the 100–1700 m/z range (Quiton-Tapia et al. 2023 ), with matches filtered to maintain a 1% false discovery rate at the peptide level. MS analyses were conducted at the MS and Proteomics Unit (Area of Infrastructures) of the University of Santiago de Compostela. Proteomic data analysis MS/MS spectra were analyzed using PEAKS Studio software (Bioinformatics Solutions, Canada) for protein identification and quantification utilizing the spectral counting method and Spec value. Given the incompletely characterized nature of the samples from a genomic perspective and the study’s focus, a custom database (Zhang et al. 2013 ) downloaded from NCBI protein database in September 2022 was utilized. Protein label-free quantification was performed using the Compare module within PEAKS Studio (version 10.6, Bioinformatics Solutions Inc., Canada), with strain C + sample serving as the control. The Spec value reported here reflects this comparison. MS data processing was conducted in Protein Group mode. The MS proteomics data have been deposited in the ProteomeXchange Consortium via the PRIDE partner repository (data will be made available upon request). Initial stringent filtering involved manually selecting the first protein from each protein group, with the others disregarded. Additionally, only proteins identified with at least two unique peptides were considered in this study (Zhao and Lin 2010 ). Graphs were generated using Google Colab, an online platform that allows for the execution of Python code in a web-based environment, using the Seaborn and Matplotlib libraries. Unipept Desktop 3.0. was employed for categorizing the identified proteins into Gene Ontology categories of Cellular compartments and Molecular functions (Gurdeep Singh et al. 2019 ). Downloaded from https://academic.oup.com/jambio/article/135/6/lxae143/7693735 by Irene Ruiz user on 10 July 2024 4 Afonso et al. Ta b l e 1. Results of the in vitro motility assays performed with Delftia acidovorans C + and C −strains. Colon y gro wth halo (mm) Delftia acidovorans C + Delftia acidovorans C- Swimming 24h 49.00 ±0.00 ∗43.00 ±0.07 ∗ 48h > 75.00 ±0.00 75.00 ±0.00 72h > 75.00 ±0.00 > 75.00 ±0.0 Swarming 24h 14.50 ±0.07 ∗8.00 ±0.14 ∗ 48h 14.50 ±0.07 ∗8.50 ±0.07 ∗ 72h 14.50 ±0.07 ∗8.50 ±0.07 ∗ Twitching 24h 13.00 ±0.14 ∗10.50 ±0.07 ∗ 48h 13.00 ±0.14 ∗11.00 ±0.14 ∗ 72h 13.00 ±0.14 ∗11.00 ±0.14 ∗ The results are presented as mean ±standard deviation of three independent assays and triplicates. Motility halos were measured after 24, 48, and 72 h. Diameter of the initial drop was 7.60 ±0.13 mm. Statistically significant differences ( P < 0.05, t -test) between strains for the same incubation time are indicated with a ( ∗) symbol. Figure 1. TEM (negative staining; 1% uranyl acetate) of Delftia acidovorans strain C + and strain C −revealed the presence of flagella. The scale bars are 0.5 μm. Results In vitro motility assays First, we assessed the motility of the two strains by spreading the bacteria in different semi-solid agar medium and measuring their growth as previously described (Gomes et al. 2019 , Booth and Rice 2020 ) (Table 1 ). Interestingly, the results of this test indicate differences between the coaggregating and the non-coaggregating strains. The strain C + presented higher values ( P < 0.05) on the three types of mobilities (swarming > twitching > swimming). However, in the case of swimming, the difference was minor and disappeared after 48 h of incubation. Concerning swarming motility, the halo growth on the colonies of C + was > 40% higher than in the case of C −during the first 24 h. Regarding twitching motility assay the growth halo was higher in C + than in C −during all the assays. The detection of these different motility capacities suggests potential differences in cell appendages. TEM analyses To confirm the presence of cellular appendages, samples of the two strains were prepared using the negative staining technique for subsequent analysis by TEM. Negative staining is a valuable and well-accepted technique in microbiology, particularly for visualizing cellular surface structures such as bacterial appendages (Mörgelin 2017 ). TEM analyses revealed the presence of flagella, observable in both strains (Fig. 1 ). However, pili-like filaments were not observed. No additional information was possible to get from this analysis. It was, therefore, of great interest to assess differences at the proteome level. For this purpose, we followed a shotgun MS-based approach, analyzing both cellular proteomes and exoproteomes of both strains. Proteome extractions and identifications The protein concentrations in the extracts from cellular proteome samples were 1467.1 ±162.1 and 2027.6 ±66.6 ( μg ml −1 ) in C + and C −, respectively. In the case of extracellular proteome, samples were 732.1 ±103.5 and 952.1 ±81.3 ( μg ml −1 ) in C + and C −, respectively. The migration pattern of duplicate sample aliquots in SDS-PAGE electrophoresis ( Fig. S1 ) indicated the extraction protocol used was effective for proteins from all sizes range. The number of proteins identified on the MS-bases analyses were 3324 and 2825 ( Tables S7 , S8 , S9 , and S10 ) for cellular proteomes and 436 and 540 for extracellular proteomes ( Tables S9 and S10 ) of C + and C −, respectively. Cellular proteome analyses The proteins identified from each strain were categorized by Gene Ontology categories of biological functions for an indepth study of the proteomes. The most common function identified for the two strains was translation ( Table S2 ) followed by protein folding, cell division, cell wall organization-, and chemotaxis in both cases. A total of 256 and 244 func- Downloaded from https://academic.oup.com/jambio/article/135/6/lxae143/7693735 by Irene Ruiz user on 10 July 2024 Proteomic analysis of DW bacteria 5 Figure 2. Cellular components associated with the peptides identified for the coaggregating D. acido v orans C + (upper bar) and the non-coaggregating D. acido v orans C −(down bar) for cellular proteome samples. “Other minor components” encompasses 44 and 25 components, respectively. Categorization into Gene Ontology classification was done by Unipept Desktop 3.0. The results correspond to the proteomic analysis of three pooled extractions. tions were identified for D. acidovorans C + and D. acidovorans C −, respectively. The proteins were also categorized by cellular location (Fig. 2 ). In general, cytoplasmic protein components were the most identified, followed by integral component of membrane and then plasma membrane proteins. It stands out that the categories “Bacterial-type flagellum filament” and “outer membrane-bounded periplasmic space” were more expressed in D. acidovorans C + than in D. acidovorans C −(Fig. 2 ). Cellular appendages Using the individual identification from PEAKS software presented in Tables S7 and S8 , we specifically searched for appendages related proteins. As shown in Fig. 3 , different proteins, related to cellular appendages like pilus and flagella were identified in both strains. In our study, “Type IV pilus twitching motility PilT protein” were only detected in the aggregating strain (Fig. 3 ), together with an overexpression of “Type IV-A pilus assembly ATPase PilB”. Regarding flagella-related proteins, “Flagellar M-ring protein FliF” was only detected in D. acidovorans C + , while “flagellar biosynthesis protein FlgE” and “Flagellar hook protein FlgE” were only detected in D. acidovorans C −(Fig. 3 ). Other cellular proteome features The differential presence of chaperone proteins in both cellular proteome samples is a noteworthy finding. In the present work, although chaperone-like proteins are expressed in the two strains, the non-coaggregating strain was the one that expressed more types ( Table S3 ), and some of the common proteins were more abundantly expressed in C −(i.e. chaperonin GroES, ATP-dependent chaperone ClpB, molecular chaperone DnaK). Methyl-accepting chemotaxis proteins (MCPs) are membrane receptors that initiate the chemotaxis signal transduction cascade to control the direction of the flagellar motor (Cooper et al. 2021 ) and were also present in the cellular proteome of both bacteria. Evident differences were not observed in the expression of most chemotaxis proteins between them, except for the absence of the CheY protein in strain C + . Another interesting finding was the greater expression of the protein called “morphology and auto-aggregation control protein” in the coaggregating strain compared to C −(Fig. 3 ). This is in line with previous tests in which autoaggregation of both strains has already been demonstrated, with a higher score for D. acidovorans C + (Afonso et al. 2023 ). Figure S2 shows the difference in the autoaggregation scores of the two strains. Downloaded from https://academic.oup.com/jambio/article/135/6/lxae143/7693735 by Irene Ruiz user on 10 July 2024 6 Afonso et al. Figure 3. Matrix bubble plot showing the different expression of selected proteins from the cellular proteomes of D. acidovorans C + and D. acidovorans C −strains related to cell appendages. The size of each circle is representative of the relative abundance of each protein on its corresponding proteome sample. The lowest relative abundance value was 1 for C −“fimbrial protein” and the highest was 12 for C + “morphology and auto-aggregation control protein.” The results correspond to the proteomic analysis of three pooled extractions. Figure 4. Cellular components associated with 211 for D. acidovorans C + (upper bar) and 271 peptides identified for D. acidovorans C −(down bar) for e x oproteome samples. The results correspond to the proteomic analysis of three pooled extractions. Exoproteome analyses Delftia acidovorans C + was previously suggested to be a producer of "public goods” (Afonso et al. 2023 ) defined as molecules that provide a collective benefit, usually through release into the extracellular environment (Smith and Schuster 2019 ). To complement our study, we specifically separated and evaluated the proteins present in their exoproteome, being this defined as those proteins located in the extracellular proximity derived from cell secretion, other mechanisms of protein export or cell lysis (Armengaud et al. 2012 ). As expected, a smaller number of proteins was obtained here in comparison with the cellular proteome samples. The number of proteins identified on this analysis was lower, with 200 proteins for D. acidovorans C + and 273 for D. acidovorans C −. The highest protein contents were quantified for D. acidovorans C − ( Table S1 ) ( P < 0.05) in both the cellular proteome and the exoproteome. Also, less categories of cellular components and fewer biological functions were represented, as expected from the sample fractionation applied before the analyses (Fig. 4 ). A total of 31 biological functions were detected for D. acidovorans C + and 42 for D. acidovorans C −( Table S4 ). The differential identifications were 23 for strain C + and 92 for C −( Tables S5 and S6 ). Figure 5 shows the relative abundance of flagella proteins and chaperone proteins identified in the exoproteome. In summary, proteins related to the flagellar hook (FlgL proteins) were more expressed in strain C-, while “flagellar biosynthesis protein FlgL” was only identified for the coaggregating strain. Discussion Based on research into coaggregation with oral bacteria, it is clear that bacterial coaggregation ability depends significantly on various surface features, such as proteins and appendages, that facilitate intercellular interactions. Surface proteins, often adhesins, mediate the initial adhesion between bacterial cells by recognizing and binding to specific receptors on neighboring cells (Chagnot et al. 2012 ). This adhesive function is pivotal for the formation of bacterial aggregates and subsequent biofilm development. Moreover, cellular appendages such as Downloaded from https://academic.oup.com/jambio/article/135/6/lxae143/7693735 by Irene Ruiz user on 10 July 2024 Proteomic analysis of DW bacteria 7 Figure 5. Heatmap showing the different expression of chaperone proteins and proteins related to flagellum in D. acido v orans C + and D. acido v orans C −e x oproteome. T he color grading is representativ e of the relativ e abundance of proteins in the sample (Spec v alue) (y ello w: higher abundance; dark blue: absence). The results correspond to the proteomic analysis of three pooled extractions. pili, fimbriae, and flagella play crucial roles in coaggregation by promoting physical interactions between bacteria). However, none of these mechanisms are currently well understood for bacteria isolated from water systems, emphasizing the need for further research to elucidate the unique factors driving coaggregation in aquatic environments. Studying cellular appendages is crucial for understanding coaggregation, especially their motility abilities. In fact, flagella-like filaments have been previously described as critical factors that mediate auto and coaggregation in bacteria from oral environments (Mishra et al. 2010 , Enersen et al. 2013 ). Ishii et al. ( 2005 ) revealed that coaggregation between two strains isolated from sludge, Pelotomaculum thermopropionicum strain SI and Methanothermobacter thermautotrophicus strain H, occurred via flagella-like filaments of the SI. The authors concluded that these filaments initially allowed SI cells to approach H and, subsequently, these filaments coloided around H. In a similar manner, Mishra et al. ( 2010 ) demonstrated that coaggregation between Actinomyces oris and oral Streptococci was mediated by the type 2 fimbria shaft FimA. The application of proteomics is essential in unraveling the molecular basis of these interactions, shedding light on the complex protein networks involved. This integrated approach offers insights into microbial community dynamics and potential biotechnological approaches. An easy way to observe bacterial motility and confirm the presence of appendages such as flagella or pili is by using semi-solid media, which allow for the visualization of characteristic growth patterns as bacteria move through the medium (Marathe et al. 2014 , Gomes et al. 2019 ). Differences in twitching and swarming motility capacities were detected between the coaggregating and the non-coaggregating strain, suggesting differences in the involved cellular appendages (pilus and flagellum). As described in previous literature, twitching motility is characterized by the retraction and extension of the pilus, resulting in a kind of “walking” of bacteria on surfaces (Merz et al. 2000 ). Regarding swimming and swarming motility, these two types of movement are driven by the flagellum (McCarter and Morabe 2019 ). TEM analyses confirmed the presence of flagella in both strains; however, pili were not observed. This may be due to the fact that TEM, as a technique, involves dehydration and heavy metal staining of specimens, which can obscure delicate structures (Ivanchenko et al. 2021 ). For a more in-depth study, we performed a proteome profiling of both strains. Proteomics allowed us to detect Type IV pilus (T4P) twitching motility PilT proteins only in the aggregating strain. These findings are in line with the results obtained on the motility assays and TEM and reinforce the theory of structural differences in pilus-like filaments between both strains. In detail, the T4P extension involves protein PilB while its retraction is governed by PilT ATPases, both cytoplasmic proteins that in vitro form oblong hexamers around a central pore (Adams et al. 2019 ). PilB and PilT interface with the T4P machinery via the PilC platform protein. Briefly, when PilB is activated, PilC rotates clockwise, leading to pilus extension. On the other hand, when the PilT is engaged, its pore rotates counterclockwise, leading to retraction of the pilus (Adams et al. 2019 ). Adams et al. ( 2019 ) observed that, despite the presence of two potential retraction ATPases, PilT and PIlU, deletions of pilT lead to a total loss of pilus function in V. choleare : suggesting that PilU functions exclusively in a dependent manner on PilT. Our results, differently, point to the secretion of T4P at D. acidovorans C −being related to PilQ protein, given its significantly higher expression in this strain. Regarding flagella-related proteins, “Flagellar M-ring protein FliF” was only detected in D. acidovorans C + , while “flagellar biosynthesis protein FlgE” and “Flagellar hook protein FlgE” were only detected in D. acidovorans C −. FliF Downloaded from https://academic.oup.com/jambio/article/135/6/lxae143/7693735 by Irene Ruiz user on 10 July 2024 8 Afonso et al. protein forms two ring-shaped structures, the so-called MS rings, which come together early in flagella morphogenesis (Kubori et al. 1992 ), around elements of the type III secretion system (T3SS) (Bergeron 2016 ). FLiF and FlgE are unique proteins that constitute obligatory elements of the flagellum (MS ring and hook, respectively) and without which flagellar assembly is not possible. The non-detection of the FlgE protein by D. acidovorans C + points to differences in the assembly pathway of the flagellar basal body. Bonifield et al. ( 2000 ) had already shown that the presence or absence of detectable FlgE depended to the stage of assembly and the fact that these differences in the structure of the flagella may influence its fitness, both in terms of motility and adhesion. These differences on their detection levels might also be indicative of different sizes of flagellar motor structures. FlgL are located between the hook and the filament, forming the junction between both and providing a structural base where flagellin is inserted for the initiation of filament elongation (Hong et al. 2018 ). The FlgL junction protein is indispensable for the formation of the flagella because FlgL-deficient bacteria lack functional flagella and are immotile (Hong et al. 2018 ). The non-detection of these proteins can be explained by their structures being too small and their derived peptides suggest being below the detection limit of the shotgun approach applied in this study. We do not disregard other proteins as being involved in the aggregation of strain C + . In fact, greater expression of the protein called “morphology and auto-aggregation control protein” in the coaggregating strain compared to C −was noted. This observation is in line with a previous study, where a higher autoaggregation score for D. acidovorans C + was determined (Afonso et al. 2023 ). It remains to be elucidated if autoaggregation underlying mechanisms can also influence coaggregation in Delftia strains. Other interesting findings concern proteins that have already been described as capable of leading to changes in flagella. One of them was the absence of the CheY MCP in strain C + . This absence has already been observed by other authors, in the case of changes in the flagella basal body (Sarkar et al. 2010 ). This agrees with the differences detected in this study in the expression of basal body proteins. Another difference was in the expression of specific chaperone-like proteins that might be related to the different flagella characteristics of both strains. The role of specific chaperones in flagellar assembly was previously described in Salmonella enterica in Minamino et al. ( 2022 ) and Inoue et al. ( 2021 ). Alternatively, lower expressions of chaperons might lead to higher aggregation or coaggregation capacities as chaperones are known to block aggregation (Santra et al. 2017 ). In short, our findings suggest that differences in the cellular appendages are related to the coaggregation abilities of DW bacteria. Moreover, the combination of the results obtained in this study point that the coaggregation ability of D. acidovorans C + may rely on some factors that predispose this strain to coaggregation, including a set of proteins core for the structure of cellular appendages, chaperones, pilus, and aggregation control. Previously, Lima et al. ( 2017 ) had already pointed that a complex interaction of more than one type of protein mediate the coaggregation of the oral bacterium F. nucleatum , which is in agreement to our results. The present study offered an alternative approach to the identification of protein biomarkers useful to detect coaggregation abilities in uncharacterized aquatic bacteria, especially relevant as an alternative for bacteria not culturable by conventional techniques. It is of great interest to follow a similar approach with strains from different taxonomies and physiology, e.g. Gram positives. A more in-depth proteomic study appears like a good methodology for further elucidating all steps involved in DW bacterial coaggregation, possibly including shaving proteomics and complementary imaging techniques. Moreover, including genetic knockouts or a broader survey of isolates could enhance the robustness of our findings by providing a more comprehensive understanding of the molecular mechanisms underlying coaggregation abilities. Additionally, such approaches could offer deeper insights into the specific proteins or genetic factors that play pivotal roles in mediating coaggregation. The information obtained here can set the basis for indepth knowledge on both multispecies biofilms and the lessexplored category of non-surface-attached biofilms, which exist in the form of microbial aggregates. This research effort is crucial for the development of new antibiofilm sustainable approaches and to develop new biotechnological procedures to enhance granulation in aquaculture systems or in wastewater treatment processes dependent on granular biomass, such as for instance anaerobic ammonium oxidation bacteria (or anammox). A c kno wledg ements A warmly thank to Sabela Balboa for very useful scientific discussions during the development of this study. The authors also thank to the Mass Spectrometry and Proteomics Unit (Area of Infrastructures) of the University of Santiago de Compostela for the nLC-MS proteomic analyses. Supplementary data Supplementary data is available at JAMBIO Journal online. Conflict of interest : All the authors declare that they have no conflict of interest. Funding This study was funded by the Galician Competitive Research Group (GRC)_ ED431C-2021/37 (to J.M.L.R. and A.T.-S.), a Juan de la Cierva-Formación postdoctoral grant FJC2019- 041664-I (to A.T.-S.). Also by the i3S Scientific Platform Histology and Electron Microscopy (HEMS), member of the PPBI (PPBI-POCI-01–0145-FEDER-022122); LEPABE, UIDB/00 511/2020 (DOI: 10.54499/UIDB/00511/2020) and UIDP/00 511/2020 (DOI: 10.54499/UIDP/00511/2020), and ALiCE, LA/P/0045/2020 (DOI: 10.54499/LA/P/0045/2020), funded by national funds through FCT/MCTES (PIDDAC); CEB, UIDB/04469/2020 (DOI: 10.54499/UIDB/04469/2020) and by LABBELS—Associate Laboratory in Biotechnology, Bioengineering and Microelectromechanical Systems, LA/P/0029/2020; CITAB, UIDB/04033/2020 (DOI: 10.54499/UIDB/04033/2020); and the FCT PhD grant attributed to A.C.A (DOI: 10.54499/2020.04773.BD). Author contributions Ana C. Afonso (Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Software, Visualization, Writing –original draft, Writing –review & editing), Manuel Downloaded from https://academic.oup.com/jambio/article/135/6/lxae143/7693735 by Irene Ruiz user on 10 July 2024 Proteomic analysis of DW bacteria 9 Simões (Conceptualization, Data curation, Funding acquisition, Investigation, Project administration, Resources, Supervision, Validation, Writing –review & editing), Maria José Saavedra (Conceptualization, Investigation, Project administration, Resources, Supervision), Lúcia Simões (Conceptualization, Data curation, Investigation, Project administration, Resources, Supervision, Writing –review & editing), Juan M. Lema (Conceptualization, Funding acquisition, Investigation, Project administration, Resources, Supervision, Validation, Writing –review & editing), and Alba Trueba-Santiso (Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Project administration, Software, Supervision, Validation, Writing –original draft, Writing –review & editing). Data availability The mass spectrometry proteomics data were deposited in the ProteomeXchange Consortium via the PRIDE partner repository and data will be made available upon request. References Adams DW , Pereira JM, Stoudmann C et al . The type IV pilus protein PilU functions as a PilT-dependent retraction ATPase. PLoS Genet 2019; 10 :e0143898. https:// doi.org/ 10.1371/ journal.pgen.1008393 Afonso AC , Gomes IB, Saavedra MJ et al . Bacterial coaggregation in aquatic systems. Water Res 2021; 196 :43–1354. https:// doi.org/ 10.1 016/j.watres.2021.117037 Afonso AC , Gomes IB, Saavedra MJ et al . Drinking-water isolated Delftia acidovorans selectively coaggregates with partner bacteria and facilitates multispecies biofilm development. Sci Total Environ 2023; 875 :162646. https:// doi.org/ 10.1016/ j.scitotenv.2023.1 62646 Armengaud J , Christie-Oleza JA, Clair G et al . Exoproteomics: exploring the world around biological systems. Expert Rev Proteomics 2012; 9 :561–75. https:// doi.org/ 10.1586/ epr.12.52 Bergeron JR . Structural modeling of the flagellum MS ring protein FliF reveals similarities to the type III secretion system and sporulation complex. PeerJ 2016; 4 :e1718. https:// doi.org/ 10.7717/ PEERJ.1718 /SUPP-4 Bonifield HR , Yamaguchi S, Hughes KT. The flagellar hook protein, FlgE, of Salmonella enterica Serovar typhimurium is posttranscriptionally regulated in response to the stage of Flagellar assembly. J Bacteriol 2000; 182 :4044. https:// doi.org/ 10.1128/ JB.182.14.4044 -4050.2000 Booth SC , Rice SA. Influence of interspecies interactions on the spatial organization of dual species bacterial communities. Biofilm 2020; 2 :100035. https:// doi.org/ 10.1016/ j.bioflm.2020.100035 Chagnot C , Listrat A, Astruc T et al . Bacterial adhesion to animal tissues: protein determinants for recognition of extracellular matrix components. Cell Microbiol 2012; 14 :1687–96. https:// doi.org/ 10.1 111/cmi.12002 Cooper KG , Chong A, Kari L et al . Regulatory protein HilD stimulates Salmonella Typhimurium invasiveness by promoting smooth swimming via the methyl-accepting chemotaxis protein McpC. Nat Commun 2021: 121 12 :1–12. https:// doi.org/ 10.1038/ s41467-020 - 20558- 6 Coppenhagen-Glazer S , Sol A, Abed J et al . Fap2 of Fusobacterium nucleatum is a galactose-inhibitable adhesin involved in coaggregation, cell adhesion, and preterm birth. Infect Immun 2015; 83 :1104–13. https:// doi.org/ 10.1128/ IAI.02838-14 Daep CA , Lamont RJ, Demuth DR. Interaction of Porphyromonas gingivalis with oral streptococci requires a motif that resembles the eukaryotic nuclear receptor box protein-protein interaction domain. Infect Immun 2008; 76 :3273–80. https:// doi.org/ 10.1128/ IAI.0036 6-08 Elliott DR , Wilson M, Buckley CMF et al . Aggregative behavior of bacteria isolated from canine dental plaque. Appl Environ Microb 2006; 72 :5211–7. https:// doi.org/ 10.1128/ AEM.01060-05 Enersen M , Nakano K, Amano A. Porphyromonas gingivalis fimbriae. J Oral Microbiol 2013; 5 . https:// doi.org/ 10.3402/ jom.v5i0.20265 Gomes I , Querido M, Teixeira J et al . Prolonged exposure of Stenotrophomonas maltophilia biofilms to trace levels of clofibric acid alters antimicrobial tolerance and virulence. Chemosphere 2019; 235 :327–35. https:// doi.org/ 10.1016/ j.chemosphere.2019.06. 184 Gurdeep Singh R , Tanca A, Palomba et al . Unipept 4.0: functional analysis of metaproteome data. J Proteome Res 2019; 18 :606–15. https:// doi.org/ 10.1021/ acs.jproteome.8b00716 Hong HJ , Kim TH, Song WS et al . Crystal structure of FlgL and its implications for flagellar assembly. Sci Rep 2018; 8 :14307. https:// doi.org/ 10.1038/ s41598- 018- 32460- 9 Inoue Y , Kinoshita M, Kida M et al . The FlhA linker mediates flagellar protein export switching during flagellar assembly. Commun Biol 2021; 4 :1–12. https:// doi.org/ 10.1038/ s42003- 021- 02177- z Ishii S , Kosaka T, Hori K et al . Coaggregation facilitates interspecies hydrogen transfer between pelotomaculum thermopropionicum and methanothermobacter thermautotrophicus . Appl Environ Microb 2005; 71 :7838–45. https:// doi.org/ 10.1128/ AEM.71.1 2.7838-7845.2005 Ivanchenko MV , Indzhykulian AA, Corey DP. Electron microscopy techniques for investigating structure and composition of hair-cell stereociliary bundles. Front Cell Dev Biol 2021; 9 :744248. https: // doi.org/ 10.3389/ fcell.2021.744248 Jacob KM , Reguera G. Competitive advantage of oral streptococci for colonization of the middle ear mucosa. In: Biofilm . Vol. 4 , 2022, 100067 . https:// doi.org/ 10.1016/ j.bioflm.2022.100067 Kaplan CW , Lux R, Haake SK et al . The Fusobacterium nucleatum outer membrane protein RadD is an arginine-inhibitable adhesin required for inter-species adherence and the structured architecture of multispecies biofilm. Mol Microbiol 2009; 71 :35–47. https: // doi.org/ 10.1111/ j.1365-2958.2008.06503.x Kennes-Veiga DM , Trueba-Santiso A, Gallardo-Garay V et al . Sulfamethoxazole enhances specific enzymatic activities under aerobic heterotrophic conditions: a metaproteomic approach. Environ Sci Technol 2022; 56 :13152–9. https:// doi.org/ 10.1021/ acs.est.2c05001 Khalil HS , Øgaard J, Leo JC. Coaggregation properties of trimeric autotransporter adhesins. Microbiologyopen 2020; 9 . https://doi.org/ 10.1002/mbo3.1109 Kubori T , Shimamoto N, Yamaguchi S et al . Morphological pathway of flagellar assembly in Salmonella typhimurium . J Mol Biol 1992; 226 :433–46. https:// doi.org/ 10.1016/ 0022-2836(92)909 58-M Lima BP , Shi W, Lux R. Identification and characterization of a novel Fusobacterium nucleatum adhesin involved in physical interaction and biofilm formation with Streptococcus gordonii . Microbiologyopen 2017; 6 . https:// doi.org/ 10.1002/ mbo3.444 Malik A , Sakamoto M, Hanazaki S et al . Coaggregation among nonflocculating bacteria isolated from activated sludge. Appl Environ Microb 2003; 69 :6056–63. https:// doi.org/ 10.1128/ AEM.69.10.60 56-6063.2003 Marathe R , Meel C, Schmidt N et al . Bacterial twitching motility is coordinated by a two-dimensional tug-of-war with directional memory. Nat Commun 2014; 5 :3759. https:// doi.org/ 10.1038/ ncomms4759 McCarter LL , Morabe ML. Swimming and Swarming Motility. In: Schmidt TM (ed.), Encyclopedia of Microbiology , 4th edn, Academic Press, 2019, 380–8, https:// doi.org/ 10.1016/ B978- 0- 12- 8096 33- 8.90742- 8 . Merz A , So M, Sheetz M. Pilus retraction powers bacterial twitching motility. Nature 2000; 407 :98–102. https:// doi.org/ 10.1038/ 350241 05 Min KR , Zimmer MN, Rickard AH. Physicochemical parameters influencing coaggregation between the freshwater bacteria Sphingomonas natatoria and micrococcus luteus . Biofouling 2010; 26 :931–40. https:// doi.org/ 10.1080/ 08927014.2010.531128 Downloaded from https://academic.oup.com/jambio/article/135/6/lxae143/7693735 by Irene Ruiz user on 10 July 2024