ORIGINAL RESEARCH published: 07 April 2020 doi: 10.3389/fmicb.2020.00583 The Interaction of RecA With Both CheA and CheW Is Required for Chemotaxis Elisabet Frutos-Grilo 1 , Maria Marsal 2 , Oihane Irazoki 1 , Jordi Barbé 1† and Susana Campoy 1* † 1 Departament de Genètica i de Microbiologia, Universitat Autònoma de Barcelona, Barcelona, Spain, 2 ICFO-Institut de Ciències Fotòniques, The Barcelona Institute of Science and Technology, Barcelona, Spain Salmonella enterica is the most frequently reported cause of foodborne illness. As in other microorganisms, chemotaxis affords key physiological benefits, including Edited by: Hari S. Misra, enhanced access to growth substrates, but also plays an important role in infection Bhabha Atomic Research Centre and disease. Chemoreceptor signaling core complexes, consisting of CheA, CheW (BARC), India and methyl-accepting chemotaxis proteins (MCPs), modulate the switching of bacterial Reviewed by: flagella rotation that drives cell motility. These complexes, through the formation of Irfan Ahmad, University of Health Sciences, heterohexameric rings composed of CheA and CheW, form large clusters at the cell Pakistan poles. RecA plays a key role in polar cluster formation, impairing the assembly when the Juan Carlos Alonso, Centro Nacional de Biotecnología SOS response is activated. In this study, we determined that RecA protein interacts with (CNB), Spain both CheW and CheA. The binding of these proteins to RecA is needed for wild-type *Correspondence: polar cluster formation. In silico models showed that one RecA molecule, attached to Susana Campoy one signaling unit, fits within a CheA-CheW ring without interfering with the complex susana.campoy@uab.cat † These formation or array assembly. Activation of the SOS response is followed by an increase authors share senior authorship in RecA, which rises up the number of signaling complexes associated with this protein. This suggests the presence of allosteric inhibition in the CheA-CheW interaction and Specialty section: This article was submitted to thus of heterohexameric ring formation, impairing the array assembly. STED imaging Microbial Physiology and Metabolism, demonstrated that all core unit components (CheA, CheW, and MPCs) have the same a section of the journal subcellular location as RecA. Activation of the SOS response promotes the RecA Frontiers in Microbiology distribution along the cell instead of being at the cell poles. CheA- and CheW- RecA Received: 23 December 2019 Accepted: 17 March 2020 interactions are also crucial for chemotaxis, which is maintained when the SOS response Published: 07 April 2020 is induced and the signaling units are dispersed. Our results provide new molecular- Citation: level insights into the function of RecA in chemoreceptor clustering and chemotaxis Frutos-Grilo E, Marsal M, Irazoki O, Barbé J and Campoy S determining that the impaired chemoreceptor clustering not only inhibits swarming but (2020) The Interaction of RecA With also modulates chemotaxis in SOS-induced cells, thereby modifying bacterial motility in Both CheA and CheW Is Required the presence of DNA-damaging compounds, such as antibiotics. for Chemotaxis. Front. Microbiol. 11:583. Keywords: SOS response system, chemotaxis, RecA, CheA, chemoreceptor polar arrays, STED microscopy, doi: 10.3389/fmicb.2020.00583 swarming Frontiers in Microbiology | www.frontiersin.org 1 April 2020 | Volume 11 | Article 583 Frutos-Grilo et al. Chemotaxis Needs the RecA-CheA-CheW Interaction INTRODUCTION Lusetti and Cox, 2002; Patel et al., 2010; Keyamura et al., 2013) being the SOS response activator (Little and Mount, 1982; Chemotaxis allows bacteria to sense their environment and adjust Maslowska et al., 2019). When DNA damage occurs, RecA their flagellar rotation accordingly, resulting in their directed acquires co-protease activity and thus the ability to promote movement toward attractants and away from repellents (Falke the auto-cleavage, among others, of LexA, the SOS system et al., 1997; Bi and Lai, 2015). Chemoreceptors are methyl- repressor. The LexA auto-hydrolysis induces the expression accepting chemotaxis proteins (MCPs) that detect the presence of SOS genes (including recA), most of which are involved of chemoeffectors and modulate the activity of the CheA kinase in DNA repair (Sassanfar and Roberts, 1990). RecA is also that, via the CheY chemotaxis response regulator, initiates the essential for chemoreceptor polar array formation and standard signaling pathway controlling the flagellar motor (Sourjik and flagellar rotation switching (Mayola et al., 2014). In previous Wingreen, 2012). There are different types of MCPs, being work, we showed that RecA interacts with CheW (Irazoki et al., Tar and Tsr the most abundant and studied in Salmonella 2016a) impairing chemoreceptor clustering and consequently enterica and Escherichia coli (Blat and Eisenbach, 1995). In swarming motility during activation of the SOS response (Irazoki many Bacteria and Archaea, MCPs group together to form large et al., 2016a,b). Specifically, when the SOS response is induced, chemosensory arrays that contain from a few to thousands the intracellular locations of CheW and RecA changes from of chemoreceptor core complexes (Briegel et al., 2009, 2012, the poles to along the cell axis (Irazoki et al., 2016a). Only 2015; Greenfield et al., 2009). These ordered structures act as after repair of the DNA damage are the polar arrays restored “antennae,” amplifying chemoeffector sensing by cooperative (Irazoki et al., 2016a,b). However, whether the inhibition of networking (Li and Hazelbauer, 2014; Frank et al., 2016; Piñas array assembly is due to CheW titration by RecA, thereby et al., 2016). Besides chemotaxis, chemoreceptor clusters are altering the stoichiometric balance of these proteins, or to other essential for swarming motility (Cardozo et al., 2010; Santos et al., causes is unclear. 2014; Irazoki et al., 2016b) and are involved in other important Previous studies showed that the structures of the P5-CheA processes, including biofilm formation (He and Bauer, 2014; domain and CheW are paralogous (Vu et al., 2012; Piñas et al., Huang et al., 2019b), cell adhesion (Huang et al., 2017), host 2018). The mutual substitution of the P5-CheA domain and colonization (Erhardt, 2016; Johnson and Ottemann, 2018) and CheW within the hexagonal rings of chemoreceptors has also antibiotic resistance (Butler et al., 2010; Irazoki et al., 2017). been described (Bilwes et al., 1999; Park et al., 2006). Based In E. coli and S. enterica, the signaling core complexes on these observations, we hypothesized that RecA also interacts are formed by two heterotrimers of transmembrane MCP with CheA and is part of the chemoreceptor core complexes homodimers, each one coupled to a protomer of the CheA kinase comprising the chemoreceptor arrays. by the chemoreceptor adaptor protein CheW (Li and Hazelbauer, Thus, to better understand the association of the SOS response 2004, 2011; Koler et al., 2018). CheA is a dimeric histidine kinase and RecA with chemoreceptor cluster formation and chemotaxis, that presents five structural and functional domains associated we explored the interaction between RecA and the P5-CheA with: histidine-containing phosphotransfer (P1), CheY/CheB domain and identified the region involved in that interaction. binding (P2), dimerization (P3), ATP binding/catalysis (P4), and In addition, we determined the location within SOS-response- CheW binding (P5) (Bilwes et al., 1999). The architecture of the activated cells of the major chemoreceptor core unit-components chemoreceptor array has been previously elucidated, in which the and the impact of this intracellular distribution on chemotaxis. interaction between CheW and the P5-CheA domain was shown to be the key structural link between core signaling units in the arrays (Liu et al., 2012; Li et al., 2013; Briegel et al., 2014b; Cassidy MATERIALS AND METHODS et al., 2015; Piñas et al., 2016). Specifically, the interaction of CheW with P5-CheA links three core complexes [using (CheW- Bacterial Strains and Growth Conditions CheA2 -CheW) core linkers] and forms a hexagonal ring of Except when indicated, all strains were grown at 37◦ C in receptors, giving rise to a lattice of hexagonally packed receptor Luria-Bertani (LB) broth or on LB plates, supplemented, when trimers of dimers networked by P5-CheA/CheW rings (Briegel necessary, with ampicillin (100 µg/mL), chloramphenicol et al., 2012, 2014a; Liu et al., 2012; Cassidy et al., 2015). These (34 µg/mL), and/or kanamycin (10 µg/mL). The strains highly stable structures are located at the cell poles (Maddock and and constructions used in this work are described in Shapiro, 1993; Sourjik and Berg, 2000; Jones and Armitage, 2015; Supplementary Table S1. Koler et al., 2018). Several conditions can disrupt chemoreceptor array assembly. In silico Docking Analysis In S. enterica and E. coli, the absence of RecA (Gómez-Gómez RaptorX (Källberg et al., 2012) custom-generated S. enterica et al., 2007; Mayola et al., 2014), the stoichiometric excess of RecA, CheW, Tar, and CheA protein structures were used in CheW (Cardozo et al., 2010; Irazoki et al., 2016b) and an the docking assays. In all cases, the available resolved structures increase in the concentration of the RecA protein prompted by of E. coli RecA (PDB: 2REB) (Story et al., 1992) and CheW the activation of SOS response (Irazoki et al., 2016b) inhibit (PDB: 2HO9), and Thermotoga maritima Tar and CheA (PDB: polar chemoreceptor array formation and suppress swarming 3JA6.C) (Cassidy et al., 2015) were used to validate the obtained motility. RecA is a multifunctional protein, it is the main bacterial 3D structures. In silico models were generated using the ClusPro recombinase and is also involved in DNA repair (Cox, 1999; server (Comeau et al., 2004). Frontiers in Microbiology | www.frontiersin.org 2 April 2020 | Volume 11 | Article 583 Frutos-Grilo et al. Chemotaxis Needs the RecA-CheA-CheW Interaction The interaction of CheA and RecA was assessed in a simple cheA tagged gene and their corresponding mutant derivatives protein-protein docking study. At least 30 of the highest-scoring were used (Supplementary Table S1). S. enterica 1recA1cheW models, in which RecA was the receptor and CheA the ligand, and background was used when cheW tagged gene was induced. In vice versa, were analyzed in duplicate. The protein structures and all cases, the tagged-gene overexpression was induced by the the obtained in silico models were visualized and analyzed using addition of 1mM of IPTG and cell lysates were obtained by PyMOL software (Schrödinger, 2010). sonication (Branson Digital Sonifier). As control, cells lysates For in silico studies of the signaling core unit and ring complex harboring the pUA1108 overexpression vector were obtained formation, the RaptorX generated structures were compared with following the same procedure. the structures documented in T. maritima (PDB:3JA6) (Cassidy The CoIP assays were performed using Pure Proteome et al., 2015) and modeled using PyMOL (Schrödinger, 2010), with Protein A magnetic beads (Millipore) coated, following the the RecA protein placed according to the identified residues of the manufacturer’s instructions, with either mouse anti-FLAG IgG or CheW-RecA and CheA-RecA interfaces (Tables 1, 2). anti-6xHis IgG monoclonal primary antibodies (Sigma-Aldrich). Before CoIP, the coated magnetic beads were pre-incubated Construction of RecA and CheA Tagged with the control lysates to minimize non-specific interactions. Two cell lysates containing the corresponding proteins were Proteins and Overexpressing Vectors mixed and incubated at 30◦ C for 1 h without shaking to allow Co-immunoprecipitation (CoIP) assays were performed using protein-protein interaction and kept at 4◦ C without shaking to proteins carrying -6xHis and -FLAG tags. Likewise, -CLIP and maintain specific interactions for 16 h. Afterward, treated coated -SNAP tagged proteins were used for STED microscopy. Plasmids magnetic beads were added to the lysate mixture for 1 h at harboring the corresponding tagged genes were constructed RT with gently shaking. Magnetic beats were then recovered, using the appropriate oligonucleotides (Supplementary washed three times and heated for 10 min at 90◦ C. Supernatants Table S2) and the HiFi DNA assembly cloning kit (NEB). The tag were separated by SDS-PAGE on a 15% polyacrylamide gel sequences were included at the 30 end of the genes preceded by a 3 × Gly linker (Supplementary Figure S1). All PCR products were digested, cloned into the pUA1108 overexpression vector (Mayola et al., 2014) and transformed into E. coli DH5α. The TABLE 2 | In vitro interaction of RecA mutant derivatives with wild-type CheA and recA and cheA tagged mutants were obtained using a site directed CheW. mutagenesis kit (Agilent). All constructions were confirmed RecA proteina Secondary structure Interaction Interaction by PCR and sequencing. In all cases, the expression of the region containing the with wild-type with wild-type corresponding tagged derivative was confirmed by Western mutated residue CheAb CheWc,d blotting (Supplementary Figure S2). Wild-type NA + + L10A Helix α1 + + Co-immunoprecipitation Assays L14A + + The CoIP assays were conducted as described previously (Mayola Q20A + − et al., 2014; Irazoki et al., 2016a) with a few modifications. H163A NR + + Briefly, cultures of S. enterica 1recA1cheA carrying the Q173A Helix α12 + + corresponding overexpression plasmid encoding a recA, or R176A + − F203 NR + + N213A Helix α13 + + TABLE 1 | In vitro interaction of CheA mutant derivatives with wild-type RecA. A214V − + K216A + + CheA protein CheA domain containing the Interaction with Y218A + + mutated residuea mutationb wild-type RecAc R222A Strand β11 − − Wild-type NAc + D224A − + M303A P3 + I228A − + L311A P3 + R243A + + G537A P5, subdomain 1 − V247A − + D587A P5, Subdomain 2 + K250A − − K590A P5, Subdomain 2 − F255A NR + + T591A P5, Subdomain 2 − Q257A Strand β 12 + + S628A P5, Subdomain 1 − K286A NR + + S646A P5, Subdomain 1, Strand β9 − Q300 Strand β 15 + + NA, not applicable. a The mutated residue and the substitution of each tagged NA, not applicable; NR, non-resolved secondary structure. a The mutated residue mutant derivative are indicated. b Unless otherwise indicated, the residue is located and the substitution of each tagged mutant derivative are indicated. b,c Results of in a non-resolved secondary structure region of the P5 CheA domain. c Results of co-immunoprecipitation (CoIP) assays using each RecA derivative and either CheA co-immunoprecipitation assays using each CheA derivative and wild-type RecA. (+) or CheW wild-type proteins. The maintenance (+) or abolishment (−) of CheA-RecA and (−) indicate the maintenance or abolishment of CheA-RecA complex formation, or CheW-RecA complex formation is shown. d Based on the results of previously respectively. described CoIP assays (Irazoki et al., 2016a). Frontiers in Microbiology | www.frontiersin.org 3 April 2020 | Volume 11 | Article 583 Frutos-Grilo et al. Chemotaxis Needs the RecA-CheA-CheW Interaction and analyzed by Western blotting using mouse anti-6xHis parameters (Sievers et al., 2011). The data were represented in a IgG1 (Merck) and rabbit anti-FLAG (Merck) and horseradish- R heat map obtained using Prism (GraphPad). peroxidase (HRP)-coupled anti-mouse IgG or anti-rabbit IgG antibodies (Acris). The membranes were developed using a HRP Stimulated Emission Depletion (STED) chemoluminiscent substrate (SuperSignalTM West Pico PLUS Microscopy Chemiluminescent Substrate, Thermo Scientific) following the S. enterica cells were labeled using SNAP-Cell 505-Star R manufacturer’s instructions. The membranes were imaged using and CLIP-CellTM TMR-Star permeable dyes, which specifically a ChemiDocTM XRS + system (Bio-Rad). recognize SNAP- and CLIP- tags, respectively, following the manufacturer’s instructions. All strains were cultured Construction of S. enterica Mutant and and coverslip mounted as described (Irazoki et al., 2016a), Tagged Strains supplemented when required with 0.08 µg mitomycin C/mL. S. enterica 1cheA and S. enterica 1cheA1cheW mutants and Previous STED imaging, samples were examined under an -SNAP and/or -CLIP tagged strains were constructed according AxioImager M2 microscope (Carl Zeiss Microscopy) to ensure to the λRed recombinase-based gene replacement method that at least 90% of the cells were correctly labeled. (Datsenko and Wanner, 2000; Supplementary Figure S1). Fluorescence immunolabeling was carried out as described pGEM-T vectors (Promega) containing -SNAP or -CLIP tags (Buddelmeijer et al., 2013), with a few modifications. The two- from pSNAP-tag (T7)-2 and pCLIPf vectors (NEB) followed color labeled samples were observed using a commercial gated- by the kanamycin cassette from pKD4 vector (Datsenko and STED microscope (Leica TCS SP8 STED 3X) equipped with Wanner, 2000) were constructed using HiFi DNA assembly a pulsed white light laser source and three depletion lasers. cloning kit (NEB) giving rise to pUA1135 and pUA1136, STED illumination of cells tagged with SNAP-Cell 505-Star was R respectively (Supplementary Table S1). These plasmids were performed using a 505 nm line, and depletion using a 592 nm used as the template to amplify the -SNAP or -CLIP tag followed line. For the CLIP-CellTM TMR-Star tag, the illumination line by the kanamycin cassette using the suitable oligonucleotides was 555 nm and the depletion source 660 nm. Fluorescent light (Supplementary Table S2). The PCR products were transformed was collected using high-efficiency single-molecule detectors into the corresponding S. enterica cells containing the pKOBEGA (SMD-HyD), using a HC PL APO CS2 100×/1.40 oil objective. plasmid (Chaveroche et al., 2000). When necessary, the antibiotic The selected areas were scanned at 600 Hz and the final resistance cassettes were eliminated using the pCP20 plasmid pixel size was 20 nm. (Datsenko and Wanner, 2000). The selected cells were screened along the z axis and The S. enterica 1recA1cheA and 1recA1cheA1cheW strains the brightest plane was chosen. At least three different were constructed by transduction as previously described representative cells were obtained for each sample. The (Campoy et al., 2002), using the P22int7(HT) bacteriophage and images were deconvoluted using the Lightning GPU-based S. enterica 1recA (UA1927), as donor strain (Mayola et al., 2014). Deconvolution Leica package. Images for publication were The absence of the prophage in the transductants was determined processed and prepared using Fiji ImageJ software (National by streaking them onto green plates as described previously Institutes of Health). (Davis et al., 1980). In all cases, gene substation in all constructs was verified by Chemotaxis Capillary Assays PCR using suitable primers followed by sequencing. Chemotaxis assays were conducted as previously described (Mayola et al., 2014). Briefly, 1 µL capillary tubes (Microcaps, Residue Conservation Percentage Drummond Scientific Co.) filled with either tethering buffer or To determine the percent conservation of the involved residues, 10 mM L-aspartate dissolved in tethering buffer (Block et al., all complete genomes from Salmonella specie and one random 1983) were placed in contact with 2 mL of the corresponding genome of each genus of the Enterobacteriaceae family were cell suspension in the chemotaxis chambers formed by placing downloaded from the GenBank database. RecA, CheA, and three V-shaped bent needles (40 mm 18G needle, Nipro). After CheW S. enterica ATCC 14028 protein sequences (ACY89831.1, incubation at 30◦ C for 1 h, the exterior of capillaries was rinsed ACY88793.1, and ACY88792.1, respectively) were used as queries under a stream of sterile distelled water. Then capillary tubes were in the Basic Local Alignment Search Tool (tBLASTn) to emptied and the cell concentration was determined by plating. identify similar protein sequences, limited using the previously Chemotaxis ratios were calculated as the ratio of viable bacteria searched genomes. RecA matches were obtained for 501 inside capillary tubes with vs. without aspartate. Salmonella and 23 Enterobacteriaceae genomes (Supplementary Data Sheet S1), CheA matches for 492 Salmonella and 13 Chemoreceptor Polar Clustering Assay Enterobacteriaceae genomes (Supplementary Data Sheet S2) The chemoreceptor polar cluster arrays were visualized and CheW matches for 501 Salmonella and 13 Enterobacteriaceae as previously described (Mayola et al., 2014), with a few genomes (Supplementary Data Sheet S3). The results were modifications. Briefly, overnight cultures of the corresponded filtered based on cut-offs for the e-value (<10−20 ) and coverage tagged strains were grown at 30◦ C in tryptone broth, (>75%). Multiple sequence alignment and data analysis were supplemented, when needed, with ampicillin and/or 40 µM carried out using the Clustal Omega local server with standard IPTG. Overnight cultures were diluted 1:100 in tryptone broth Frontiers in Microbiology | www.frontiersin.org 4 April 2020 | Volume 11 | Article 583 Frutos-Grilo et al. Chemotaxis Needs the RecA-CheA-CheW Interaction supplemented with IPTG and incubated at 30◦ C until an OD600 of 0.08–0.1 was reached, Then cells were collected by centrifugation, washed once using ice-cold tethering buffer (10 mM potassium-phosphate pH 7, 67 mM NaCl, 10 mM Na-lactate, 0.1 mM EDTA, and 0.001 mM l-methionine) and resuspended in 20–100 µL of the same buffer. Then cells were stained with the permeable dyes SNAP-Cell 505-Star and CLIP- R CellTM TMR-Star, following the manufacturer’s instruction. Finally, cells were fixed with paraformaldehyde, resuspended in 1 × PBS, mounted on 35 mm poly-L-lysine-pre-coated coverslips using Mowiol-DABCO mounting medium and air-dried. FIGURE 1 | Co-immunoprecipitation assays of S. enterica RecA and CheA. The samples were examined under an Axio Imager M2 Cell-lysates prepared from S. enterica 1recA1cheA cultures overexpressing either RecA-FLAG- or CheA-6xHis-tagged proteins were incubated together microscope (Carl Zeiss Microscopy) equipped with the to allow interaction of the proteins. Co-immunoprecipitation (CoIP) was appropriate filter set [green channel: GFP (Zeiss filter set 38); red performed by adding magnetic beads coated with anti-FLAG (Beads FLAG) or channel: Rhod (Zeiss filter set 20)]. Cell fields were photographed anti-6xHis antibodies (Beads His), attached proteins were recovered and and at least 350 cells were visually inspected. All images were separated by SDS-PAGE. The CoIP controls consisted of mixtures containing acquired under identical conditions. Each experiment was only RecA-FLAG or CheA-6xHis overexpressing lysates. The presence in the recovered supernatants of each tagged-protein was assessed by Western performed at least in triplicate using independent cultures; a blotting using both anti-FLAG and anti-6xHis primary antibodies. The minimum of 1,050 cells from each studied strain were therefore presence or absence of RecA-FLAG, CheA-6xHis, or both tagged proteins in analyzed. The images presented in the figures are representative the corresponding lysate is indicated. The experiments were done at least in of the entire image set. ImageJ software (National Institutes triplicate. Black arrows show the position of CheA-6xHis and RecA-FLAG. +, added protein; -, non-added protein; MW, molecular mass marker, in kDa. of Health) was used to quantify the number of clusters and to prepare images for publication. Statistical Analysis 2REB) (Story et al., 1992) and the T. maritima CheA (PDB: The results of the chemotaxis and chemoreceptor-clustering 3JA6.C) (Cassidy et al., 2015), which includes the P3-, P4-, and assays were statistically evaluated using a one-way analysis P5-CheA domains. Balanced-coefficient docking models were of variance (ANOVA) with Prism (GraphPad), as previously considered to be the most accurate for the analysis of the described (Brennan et al., 2013; Raterman and Welch, 2013; RecA-CheA interaction (Comeau et al., 2004). Thirty of the Mayola et al., 2014). The analyses were followed by the highest-scoring models were analyzed for each combination of Bonferroni multiple comparison post hoc test. A p-value < 0.01 RecA receptor protein and CheA ligase and for the reverse was considered to indicate statistical significance. In all cases, the combination. Although the spatial arrangement was not exactly error bars in the figures indicate the standard deviation. the same in each combination, the putative interacting regions were considered to be those repeated in all of the studied models (Figure 2). RESULTS As expected, the results were similar to those obtained for the CheW-RecA interaction. In CheA, both P5 subdomains (1 RecA and CheA Interaction and 2) interacted with RecA. In some of the in silico models, Given the structural similarities of the P5-CheA domain and residues of the P3 domain were also exposed to the RecA-CheA CheW (Vu et al., 2012; Piñas et al., 2018), we conducted co- interface (Figure 2 and Table 1). With respect to RecA, the immunoprecipitation (CoIP) assays to determine whether, as putative interface with CheA was located in the NH2 -terminal with CheW, RecA is able to interact with CheA (Arifuzzaman and central domains (at α1, α12, α13, β11, and β15) (Figure 2 et al., 2006; Mayola et al., 2014). Thus, RecA-FLAG and and Table 2). CheA-6xHis tagged proteins were overexpressed in S. enterica To confirm these interaction interfaces, site-directed 1recA1cheA strains carrying the corresponding plasmids mutagenesis was used to construct the corresponding mutant (Supplementary Table S1). When both recombinant proteins derivatives for each protein. The 21 RecA and 8 CheA residues were present in the protein mixture, anti-FLAG antibody-coated were selected based on their exposure and their potential ability beads recovered both RecA-FLAG and CheA-6xHis from the to mediate RecA-CheA pair formation (Tables 1, 2). With the supernatants (Figure 1). When anti-6xHis antibody-coated beads exception of the RecA A214V mutant, in which the Ala residue were added to the mixture, RecA-FLAG proteins were also was changed to a Val, all selected residues were converted to an recovered along with CheA. These results demonstrated the Ala (Tables 1, 2), which is considered to be non-reactive amino in vitro pairing of RecA and CheA. acid (Cunningham and Wells, 1989). The corresponding recA An in silico modeling experiment was then conducted, and cheA gene mutants constructed in vitro were 6xHis-tagged aimed at identifying the putative RecA and CheA residues and the effects of the substitutions on the RecA-CheA interaction participating in the interaction of these proteins. Protein-protein were determined by CoIP assays using the corresponding interaction docking was performed with RaptorX (Källberg et al., FLAG-tagged wild-type RecA or CheA protein (Figure 3). 2012) using, as reference structures, the E. coli RecA (PDB: The results are summarized in Tables 1 and 2. For CheA, only Frontiers in Microbiology | www.frontiersin.org 5 April 2020 | Volume 11 | Article 583 Frutos-Grilo et al. Chemotaxis Needs the RecA-CheA-CheW Interaction FIGURE 2 | In silico model of the RecA and CheA protein interaction. The predicted ternary structures of (A) S. enterica RecA (yellow) and (B) CheA (P3-, P4-, and P5-domains, pink) proteins are shown. The predicted interface of RecA and CheA is highlighted in purple and green, respectively. Residues selected for site-directed mutagenesis and their locations are also indicated. (C) Ribbon diagrams of one of the highest-scoring models of the RecA-CheA interaction. The two views of the interaction, obtained using Pymol software, are rotated 90◦ about the x-axis. P5 domain was associated with the RecA interaction; CheA decrease in the recombination activity of the residues associated mutations in the P3 domain did not disturb wild-type RecA with RecA-CheA pair formation (Supplementary Figure S3). binding (Table 1). Within the P5-CheA domain, five residues However, not all of the residues involved in the CheW-RecA (G537, K590, T591, S628, and S646) were found to be directly interaction were also associated with the CheA-RecA interaction. involved in the interaction with RecA. Their substitution by Ala Thus, RecA Q20A and R176A mutants, while unable to bind prevented RecA-CheA pair formation (Table 1 and Figure 3). CheW (Irazoki et al., 2016a), interacted with wild-type CheA Analyses of the 21 RecA mutants showed that only five were (Table 2 and Figure 3). Similarly, the involvement of residues unable to bind wild-type CheA (A214V, R222A, D224A, K250A, A214, D224, and I228 was limited to the RecA-CheA interaction, and I228A). With the exception of A214, all of the residues as they had no effect on RecA-CheW binding (Table 2). Only two were located on the β11 strand, shown in previous studies to be mutant derivatives, R222A and K250A, abolished the interactions associated with monomer-monomer interactions as well as RecA of CheA and CheW with RecA (Table 2). These results not only filament formation and stabilization (Skiba et al., 1999; Zaitsev revealed the residues associated with RecA-CheA pairing but also and Kowalczykowski, 1999; Chen et al., 2008). Recombinase demonstrated the ability of RecA to interact with both CheA and assays with the RecA mutants showed, in almost all cases, a clear CheW through different interfaces. In addition, when CheA and Frontiers in Microbiology | www.frontiersin.org 6 April 2020 | Volume 11 | Article 583 Frutos-Grilo et al. Chemotaxis Needs the RecA-CheA-CheW Interaction FIGURE 3 | Co-immunoprecipitation assays of S. enterica RecA and CheA mutant derivatives. Representative images of the CoIP of mutant derivatives that allow (CheA M303A or RecA R20A) or impair (CheA G537A or RecA R222A) RecA-CheA interaction. Each lane contains a mixture of S. enterica 1recA1cheA cell lysates containing the corresponding 6xHis-tagged overexpressed mutant derivative and the wild-type FLAG-tagged protein. The immunoprecipitates were obtained using anti-FLAG coated magnetic beads. All experiments were done at least in triplicate. The results obtained with all mutant derivatives (Tables 1 and 2) were the same as those shown in the figure. Black arrows indicate RecA and CheA protein bands. NA, not added; MW, molecular mass marker, in kDa. FIGURE 4 | Conservation of RecA-associated residues among Salmonella and other representative Enterobacteraceae. The percentage of conserved residues in the interaction between (A) RecA, (B) CheA, and (C) CheW from S. enterica was calculated based on the number of residues in the studied sequences that differed from the query sequence. The studied residues were compared with their homologs in Salmonella and in one representative of each available genus within the family Enterobacteriaceae. Identities closer to 0% are shown in red, and those closer to 100% in pink. Intermediate percentages are represented by other colors in the legend. All sequences were downloaded on November 14, 2019. CheW proteins were not present RecA protein was majorly not Enterobacteriaceae. These results pointed out that the association located at the cell poles (Supplementary Figure S4). of the SOS response with chemoreceptor signaling complexes In addition, we determined the residue conservation may be extended to Enterobacteriaceae and perhaps also to other percentages among Salmonella and Enterobacteriaceae for each families of bacteria. involved amino acid (Figure 4). The RecA residues involved in CheA and CheW interactions were highly conserved (100% identity; Figure 4A), except for residue I228 (96.4%). Among RecA as a Part of the Chemoreceptor the CheA residues associated with the RecA interaction, all Signaling Core Unit were conserved in Salmonella (100% identity), except G537, Our results also indicated the differential interaction of RecA which differed in S. bongori, resulting in a slightly lower with CheW and CheA (Table 2). RecA interfaces with CheA identity (99.2%). The CheA residues were also highly conserved and CheW do not overlap with the regions of CheA-CheW in Enterobacteriaceae (>75% identity), again except G537 binding, nor with those involved in MCP interaction (Cassidy (7.7%). Finally, for the involved residues of CheW, the results et al., 2015; Piñas et al., 2016; Huang et al., 2019a). These were similar, with 100% identity in Salmonella and >80% in observation suggested that RecA may be part of a signaling Enterobacteriaceae. According to these findings, the ability complex, a possibility explored by generating in silico interaction of RecA to interact with CheW and CheA may occurs not models that included the entire signaling core unit (Figure 5). only in Salmonella species besides S. enterica but also in The RaptorX-generated structures for all S. enterica proteins Frontiers in Microbiology | www.frontiersin.org 7 April 2020 | Volume 11 | Article 583 Frutos-Grilo et al. Chemotaxis Needs the RecA-CheA-CheW Interaction were compared with the structure of the T. maritima chemotaxis microscopy (STED), a super-resolution fluorescence imaging signaling complex (PDB:3JA6) (Cassidy et al., 2015) and technique that increases the axial resolution of biological samples modeled using PyMOL software (Schrödinger, 2010). The RecA up to 20–40 nm (Han and Ha, 2015). interaction was placed according to the residues determined to S. enterica cheA:SNAP tar:CLIP and 1recA cheA:SNAP be directly involved in the CheA-RecA or CheW-RecA interfaces tar:CLIP tagged strains were constructed and were tested in (Tables 1, 2). As seen in Figure 5, the RecA fits into the chemoreceptor clustering and swarming assays under non-DNA chemoreceptor signaling complex without allosteric interference. damage conditions to verify that tag addition did not alter their chemoreceptor array phenotypes. No changes in either chemoreceptor polar clusters or swarming motility (Cardozo RecA Interacts With Both CheW and et al., 2010; Partridge et al., 2019) were observed for S. enterica CheA in vivo cheA:SNAP tar:CLIP strain. Its phenotype was the same as To corroborate the results of the in silico models and study the that of S. enterica wild-type (Supplementary Figure S5). importance of the RecA-CheA interaction for chemoreceptor Also, S. enterica 1recA cheA:SNAP tar:CLIP was unable to polar cluster formation in vivo, the location of CheA and Tar swarm and the number of chemoreceptor polar clusters was proteins was analyzed using the stimulated emission depletion drastically reduced (Supplementary Figure S5) since RecA is FIGURE 5 | In silico model of the interaction of RecA-CheA-CheW proteins forming the core signaling complex. (A) The predicted ternary structures of S. enterica RecA (R, yellow), CheW (W, blue), CheA (P3-P5, pink), and Tar (T, gray) are represented in cartoon form. Model images are cross-sections through the receptor tip and CheA/CheW baseplate, viewed perpendicular (left) and parallell to (right) the cytoplasmatic membrane. The proteins were modeled using Pymol software according to the studied protein-protein interfaces responsible for RecA-CheW (Irazoki et al., 2016a), RecA-CheA (in this study), and CheA-CheW-Tar (Piñas et al., 2016) interactions. (B) Schematic representation of the above in silico model and also including the P1 and P2 domains. Frontiers in Microbiology | www.frontiersin.org 8 April 2020 | Volume 11 | Article 583 Frutos-Grilo et al. Chemotaxis Needs the RecA-CheA-CheW Interaction essential for both swarming and polar array cluster formation (Irazoki et al., 2016b). Then, as expected, the IPTG-induced (Mayola et al., 2014). expression of a wild-type recA gene within S. enterica cheA:SNAP The RecA complementation and overexpression assays were tar:CLIP promoted the redistribution of CheA and Tar along the performed using the pUA1108 vector containing wild-type recA cell (Figure 6B). However, the increased expression mediated and its derivatives unable to interact with CheA (RecA A214V) by IPTG of recA mutants unable to bind CheA, CheW or both or CheW (RecA R222A) or both proteins (RecA R176A) under proteins did not alter the CheA and Tar location, that remained the control of the Ptac IPTG-inducible promoter. The plasmids at the cell poles (Figure 6B). Together, the results indicate that were transformed into S. enterica cheA:SNAP tar:CLIP and 1recA the interaction of RecA with both CheA and CheW is needed for cheA:SNAP tar:CLIP strains (Supplementary Table S1), and the chemoreceptor polar array formation and that both interactions intracellular location of CheA and Tar was then determined. occur in vivo. The SNAP and CLIP tags are self-labeling enzymes derived from the human DNA repair protein O6 -alkylguanine-DNA alkyltransferase. Appropiate permeable dyes directly attach to the Location of CheA and Tar Proteins Within target protein with high reactivity and labeling specificity. The SOS Response-Induced Cells SNAP-tag binds O6 -benzylguanine derivatives (Keppler et al., In bacteria grown in liquid medium under non-DNA damaging 2003), and the CLIP-tag O2 -benzylcytosine derivatives (Gautier conditions, the polar cluster array proteins CheW, CheA, and et al., 2008). Due to these differences, the two tags, with their Tar are located mainly at the cell poles (Sourjik and Berg, 2000; permeable dyes suitable for STED imaging (SNAP-Cell 505- R Greenfield et al., 2009; Koler et al., 2018). Several studies have Star and CLIP-CellTM TMR-Star, respectively), can be employed shown that RecA also localizes at the poles (Lusetti and Cox, simultaneously to specifically label target proteins in living cells 2002; Lesterlin et al., 2014; Irazoki et al., 2016b). Our previous (Gautier et al., 2008). work demonstrated that during SOS response activation, RecA The absence of RecA impairs chemoreceptor array formation and CheW are no longer located at the cell poles but in (Mayola et al., 2014). For complementation assays, the basal small foci distributed along the cell (Irazoki et al., 2016a). expression of the wild-type recA gene cloned in the pUA118 To further understand the association of RecA with CheA vector was enough to restore chemoreceptor array formation and and the signaling core units, the location of CheA and Tar cell CheA and Tar were located again in cell poles (Figure 6A; proteins in SOS-response-induced cells was studied in S. enterica Mayola et al., 2014). Nevertheless, in the presence of a non- cheA:SNAP tar:CLIP and cheA:SNAP recA:CLIP tagged strains CheA-interacting RecA, no chemoreceptor polar clusters were by STED imaging. formed (Figure 6A) and CheA and Tar were distributed along Under non-DNA-damaging condition, RecA, CheA, and the cell. The same phenotype was observed using a RecA mutant Tar were, as expected, located at the poles of S. enterica unable to interact with CheW (Figure 6A) or with both CheA and cells (Figure 7 and Supplementary Figure S6). However, the CheW (Figure 6A). addition of a sublethal concentration of SOS-inducer resulted On the other hand, the RecA overexpression inhibited in cell filamentation and the redistribution of CheA and Tar polar cluster assembly in a wild-type genetic background (Figure 7) to follow that of RecA (Supplementary Figure S6) FIGURE 6 | Representative STED images of the locations of CheA and Tar within S. enterica. Cells containing the pUA1108 expression vector, either empty or carrying wild-type recA, a non-CheA interacting recA mutant derivative (A214V), a non-CheW interacting (R222A) or a recA mutant derivative that interacts with neither CheA nor CheW (R176A) are shown. The corresponding plasmids were included in the genetic backgrounds of (A) S. enterica 1recA cheA:SNAP tar:CLIP and (B) S. enterica cheA:SNAP tar:CLIP. CheA and Tar proteins were labeled with the permeable dyes CLIP-CellTM TMR-Star and SNAP-Cell R 505-Star, respectively. For all images, overlapped channel results are presented. The maximum intensity projection images of the obtained z-stacks are shown. All experiments were done at least in triplicate. Frontiers in Microbiology | www.frontiersin.org 9 April 2020 | Volume 11 | Article 583 Frutos-Grilo et al. Chemotaxis Needs the RecA-CheA-CheW Interaction FIGURE 7 | Representative STED images of the subcellular locations of Tar and CheA in S. enterica cheA:SNAP tar:CLIP cells in the absence or presence of SOS inducer. The Tar and CheA proteins were labeled with the permeable dyes CLIP-CellTM TMR-Star (channel 1, in green) and SNAP-Cell R 505-Star (channel 2, in red), respectively. When appropriate, mitomycin C was added at a final concentration of 0.08 µg/mL. For all images, each channel is shown both individually and overlapped. The maximum intensity projection images of the z-stacks are shown. All experiments were done at least in triplicate. and CheW (Irazoki et al., 2016a). Thus, under non-DNA- overexpressed in bacterial cells by the addition of IPTG. Under damaging conditions and during activation of the SOS response, these conditions, i.e., in the presence of mitomycin C or recA the intracellular distributions of CheA, CheW, Tar and overexpression, the absence of chemoreceptor polar clusters RecA were the same. did not lead to an inhibition of the chemotaxis response. Interestingly, in the presence of mitomycin C, chemoreceptor polar array formation in the 1recA strain are lower than SOS Response-Induced Cells Present that observed in either SOS-induced or RecA-overexpressing Normal Chemotaxis Response wild-type cells (Figure 8). Further, as previously published To further demonstrate that RecA is a component of the signaling (Mayola et al., 2014), 1recA cells were unable to respond core unit and that the structure of this unit is preserved during to a chemoeffector, an ability that was restored only by the SOS response activation, chemoreceptor polar clustering and addition of wild-type recA. Neither polar cluster array formation chemotaxis assays were performed by exposing cells of different nor chemotaxis (Figures 6, 8) were restored in a 1recA genetic backgrounds to a sublethal concentration of mitomycin C strain complemented with recA mutant derivatives unable to and then monitoring the chemotaxis response. interact with CheA, CheW or both. According to these results, As shown in Figure 8, the presence of mitomycin C the formation of active signaling core units and therefore did not affect S. enterica wild-type strain chemotaxis. The chemoreceptor polar arrays requires the binding of RecA to same results were obtained when the wild-type recA was both CheA and CheW. Frontiers in Microbiology | www.frontiersin.org 10 April 2020 | Volume 11 | Article 583 Frutos-Grilo et al. Chemotaxis Needs the RecA-CheA-CheW Interaction of the other CheW binding partners identified so far (CheA, CheW, and MCPs) (Irazoki et al., 2016a). Moreover, all residues involved in RecA-CheA and RecA-CheW interfaces were highly conserved not only in S. enterica but also in other members of Enterobacteriaceae (Figure 4). In vivo assays showed that chemoreceptor polar clustering requires the interaction of RecA with both CheA and CheW. Indeed, RecA was distributed along the cell when polar clusters are not built (Supplementary Figure S4). Further, RecA mutants unable to bind CheA, CheW, or both proteins neither restored wild-type chemoreceptor polar cluster assembly in cells with a 1recA genetic background (Figure 6A) nor abolished polar cluster formation in wild-type cells overexpressing RecA (Irazoki et al., 2016b; Figura 6B). According to these results, the FIGURE 8 | Chemotaxis ability and chemoreceptor polar clustering of interaction of RecA not only with CheW but also with CheA is S. enterica strains. Chemotaxis assay results and the percentage of essential for chemoreceptor array formation. chemoreceptor polar clusters in S. enterica cheA:SNAP tar:CLIP and 1recA Previous studies showed that the stoichiometry of cheA:SNAP tar:CLIP strains. As needed, the cells were grown in the absence (w/o MitC) or presence (w/ MitC) of mitomycin C (0.08 µg/mL). When chemoreceptor core unit components within the cell is crucial for indicated, they were transformed with either empty pUA1108 or the plasmid polar array assembly. For example, the absence or overexpression carrying wild-type recA or a recA mutant derivative [non-CheA-interacting of CheW abolishes chemoreceptor cluster formation (Avram (A214V), non-CheW-interacting (R222A) or interacting with neither CheA nor Sanders et al., 1989; Cardozo et al., 2010). The same phenotype CheW (R176A)]. The chemotaxis ratios were calculated as the ratio of viable occurs in a knock out recA mutant (Mayola et al., 2014) or when bacteria inside capillary tubes with vs. without aspartate. The results are the mean of three independent experiments. Error bars represent the standard the RecA concentration is increased, whether by SOS response deviation. activation or by its overexpression (Gómez-Gómez et al., 2007; Irazoki et al., 2016b). In earlier work, we proposed that RecA prompts the titration of CheW, thus preventing chemoreceptor DISCUSSION assembly and, in turn, polar cluster array formation during activation of the SOS response (Irazoki et al., 2016a). The same Our results provide strong evidence supporting the interaction sequence of events may describe the CheA-RecA interaction. of RecA with P5-CheA domain (Figure 1), which is structurally Nevertheless, the results described herein clearly determine similar to CheW (Vu et al., 2012; Table 1). While both P5-CheA that RecA protein present different binding interfaces with subdomains 1 and 2 participate in the RecA-CheA interaction CheA and CheW, that do not overlap with those associated (Table 1), the involved CheA-residues do not overlap with those with CheA-CheW interaction or with their binding to MCPs. of the CheA-CheW interaction (Cassidy et al., 2015). For RecA, Together with the fact that, in the absence of RecA there is no the CheA binding interface is located at its NH2 -terminus, chemotaxis response (Mayola et al., 2014), suggested the direct between residues 214 and 250 (Figure 2 and Table 2). This region interaction of RecA with the chemoreceptor core unit. is mainly associated with monomer-monomer interaction as well STED imaging of the tagged strains indicated that RecA, as RecA filament formation and stabilization (Skiba et al., 1999; CheA, Tar (Figure 6 and Supplementary Figure S6) and CheW Zaitsev and Kowalczykowski, 1999; Chen et al., 2008). Moreover, (Irazoki et al., 2016a), the main components of the signaling almost all of the residues involved in RecA-CheA pair formation core unit, follow the same intracellular distribution when RecA had a very low recombinase activity (Supplementary Figure S3). concentration is increased following the activation of the SOS In silico docking established that the RecA interaction could response or recA overexpression, moving from the cell poles be fitted to the chemoreceptor signaling complex without any to along the cell axis (Figure 6 and Supplementary Figure S6; allosterical interference (Figure 5), as the P5-CheA subdomain Irazoki et al., 2016a). Furthermore, the polar cluster arrays of 1 was still able to interact with CheW subdomain 2 (also known S. enterica 1recA cells were not restored by the presence of a as interface 1) (Natale et al., 2013). Despite the similarity of recA mutant unable to bind CheA, CheW, or both, such that CheW and P5-CheA, only RecA Arg222 and Lys250 residues, chemotaxis was inhibited (Figure 8). Only the addition of wild- located at the β11 strand, were associated with both RecA-CheA type RecA reestablished chemotaxis (Figure 8). Taken together, and RecA-CheW pair formation (Table 2; Irazoki et al., 2016a). these findings demonstrate that RecA and its ability to bind CheA The rest of the identified RecA residues (Ala214, located in the and CheW are crucial for the functionality of the signaling core α13 helix, and Asp224, Ile228, and Val247, all of them in the complex and for chemotaxis. β11 strand) are only associated with CheA binding, and their The direct association of RecA with the chemoreceptor core mutation did not affect the RecA-CheW interaction (Bilwes et al., unit also explains the impaired formation of chemoreceptor 1999; Cassidy et al., 2015). RecA binding to CheW is mediated not polar clusters in the presence of increased RecA (Figure 9). only by the β11 strand but also by the α1 and α12 helices of RecA Chemoreceptor arrays consist of hexagonal lattices of MCPs (Table 2; Irazoki et al., 2016a). As in the RecA-CheA interaction, stabilized by interconnected heterohexameric rings of CheA RecA-CheW pairing does not interfere with the binding of any and CheW (Cassidy et al., 2015). The rings are formed by the Frontiers in Microbiology | www.frontiersin.org 11 April 2020 | Volume 11 | Article 583 Frutos-Grilo et al. Chemotaxis Needs the RecA-CheA-CheW Interaction FIGURE 9 | Proposed model of chemoreceptor assembly inhibition during activation of the SOS response. (A) Under non-DNA damaging conditions, hexagonally assembled chemoreceptor polar clusters are formed by three signaling core units that are stabilized by P5-CheA-CheW rings. Only one RecA molecule fits within the inner aspect of the heterohexameric ring. Under this condition, the cells are able to swarm, polar cluster arrays are formed and the cell exhibits chemotaxis. (B) Activation of the SOS response is followed by a high increase in the RecA concentration. This induces an increase in the number of RecA-associated signaling complexes. In that context, the heterohexameric ring formation is allosterically disturbed by the presence of more than one RecA molecule. As a result, chemoreceptor polar clusters are unable to form and chemoreceptor signaling units remain distributed along the cell. These cells are unable to swarm but retain a chemotaxis response. alternated interactions between P5-CheA and CheW of three even in the absence of chemoreceptors arrays, only in the unit core complexes that give rise to interface 2, composed of presence of wild-type RecA, not a RecA mutant unable to the P5-CheA subdomain 2 and CheW subdomain 1. This key bind CheA, CheW, or both (Figure 8). Our results indicates, link between signaling core complexes in the array is needed in agreement with previously reported data, that although the for cluster assembly (Natale et al., 2013; Cassidy et al., 2015). absence of polar clustering clearly impairs swarming ability As shown in Figure 9, there is enough space within a ring (Cardozo et al., 2010; Frank et al., 2016; Irazoki et al., 2016b), to fit one RecA molecule interacting with one of the three the chemotaxis pathway remains functional (Maki et al., 2000; internal face of (CheW-CheA2 -CheW) heterohexameric ring, Briegel et al., 2014b; Frank et al., 2016; Piñas et al., 2016). These without altering its hexagonally structure. When SOS response results are in Further, it is also known that chemotaxis is not activation increases the intracellular RecA concentration, the affected when E. coli cells are treated with cephalexin (Maki protein becomes associated with a greater number of signaling et al., 2000), a β-lactam antibiotic that induces the SOS response core units. However, the build-up of heterohexameric rings (Bano et al., 2014). is prevented by the high levels of RecA, which impair the The chemotaxis of bacterial cells is enhanced by the formation of CheW-P5-CheA interface 2 and consequently, the networking of chemoreceptors (Frank et al., 2016). Nevertheless, array assembly is inhibited. Frank et al. (2016) showed that the CheA kinase levels and It is important to note that the chemotaxis response is not flagellar motor switching are similar in cells with and without associated with the presence of polar chemoreceptor clusters, polar clusters. The same authors found that, under conditions unlike swarming. Thus, in the S. enterica wild-type strain, of high CheA kinase activity, cells with dispersed receptor while either SOS response activation by mitomycin C or the complexes are more sensitive to chemoeffectors than cells with overexpression of wild-type recA impaired chemoreceptor array polar clusters (Frank et al., 2016). However, cells can remodel formation and therefore swarming motility (Supplementary their chemotaxis signaling pathway to enhance swarming Figure S7; Irazoki et al., 2016b), there was no effect on chemotaxis motility (Partridge et al., 2019). The over production of RecA (Figure 8). The signaling core units were completely functional, due SOS response activation by the presence of DNA-damaging Frontiers in Microbiology | www.frontiersin.org 12 April 2020 | Volume 11 | Article 583 Frutos-Grilo et al. Chemotaxis Needs the RecA-CheA-CheW Interaction compounds or antibiotics, increase the number of signaling contributed conception and design of the study, contributed to core units associated to RecA, and consequently modulate manuscript revision, read and approved the submitted version. the architecture of chemoreceptor arrays, by impairing the heterohexameric CheA-CheW ring formation (Figure 9). Thus, the distribution of chemoreceptor signaling units modulates FUNDING not only swarming motility in cells growing on a surface, to prevent exposure to higher concentrations of SOS-inducer The authors acknowledge financial support from the Spanish compounds (Irazoki et al., 2016b), but also the chemotaxis Ministry of Economy and Competitiveness through its R&D performance of the SOS-induced cells. The high degree of funding program (BIO2016-77011-R) and the “Severo Ochoa” conservation of residues associated with the CheA-RecA-CheW program for Centres of Excellence in R&D (SEV-2015-0522), interaction in Enterobacteriaceae (Figure 4) suggested that from Fundació Privada Cellex, Fundación Mig-Puig and from the relationship between the SOS response, chemotaxis and Generalitat de Catalunya through the CERCA program. The swarming to modulate bacterial motility in the presence of funders had no role in the design of the study, in data collection antibiotics and other injurious or potentially lethal compounds and analysis, in the decision to publish, or in the preparation may be extended to other species of this bacterial group. of the manuscript. DATA AVAILABILITY STATEMENT ACKNOWLEDGMENTS All datasets generated for this study are included in the We thank Joan Ruiz, Susana Escribano, Marc Gaona, and article/Supplementary Material. Miquel Sánchez for their technical support during some of the experimental procedures, and Pablo Loza and Jordi Andilla AUTHOR CONTRIBUTIONS for their help and guidance on some of the super-resolution microscopy experiments. EF-G, OI, and SC performed the in silico analyses and site- directed mutagenesis. EF-G and MM performed the fluorescent immunolabeling and STED imaging. EF-G and SC designed, SUPPLEMENTARY MATERIAL performed, and analyzed the rest of experiments. EF-G, SC, and JB coordinated the research, discussed the findings, and The Supplementary Material for this article can be found interpreted the results. SC wrote the first draft of the manuscript. online at: https://www.frontiersin.org/articles/10.3389/fmicb. EF-G, MM, and OI wrote sections of the manuscript. All authors 2020.00583/full#supplementary-material REFERENCES Briegel, A., Wong, M. L., Hodges, H. L., Oikonomou, C. M., Piasta, K. N., Harris, M. J., et al. (2014b). New insights into bacterial chemoreceptor array structure Arifuzzaman, M., Maeda, M., Itoh, A., Nishikata, K., Takita, C., Saito, R., et al. and assembly from electron cryotomography. Biochemistry 53, 1575–1585. doi: (2006). Large-scale identification of protein-protein interaction of Escherichia 10.1021/bi5000614 coli K-12. Genome Res. 16, 686–691. doi: 10.1101/gr.4527806.8 Briegel, A., Li, X., Bilwes, M. A., Hughes, K. T., Jensen, G. J., and Crane, B. R. (2012). Avram Sanders, D., Mendez, B., and Koshland, D. E. (1989). Role of the CheW Bacterial chemoreceptor arrays are hexagonally packed trimers of receptor protein in bacterial chemotaxis: overexpression is equivalent to absence. dimers networked by rings of kinase and coupling proteins. Proc. Natl. Acad. J. Bacteriol. 171, 6271–6278. doi: 10.1128/jb.171.11.6271-6278.1989 Sci. U.S.A. 109, 3766–3771. doi: 10.1073/pnas.1115719109 Bano, S., Vankemmelbeke, M., Penfold, C. N., and James, R. (2014). Detection of Briegel, A., Ortega, D. R., Huang, A. N., Oikonomou, C. M., Gunsalus, R. P., and induced synthesis of colicin E9 using ColE9p::gfpmut2 based reporter system. Jensen, G. J. (2015). Structural conservation of chemotaxis machinery across World J. Microbiol. Biotechnol. 30, 2091–2099. doi: 10.1007/s11274-014-1635-y Archaea and Bacteria. Environ. Microbiol. Rep. 7, 414–419. doi: 10.1111/1758- Bi, S., and Lai, L. (2015). Bacterial chemoreceptors and chemoeffectors. Cell. Mol. 2229.12265 Life Sci. 72, 691–708. doi: 10.1007/s00018-014-1770-5 Briegel, A., Ortega, D. R., Tocheva, E. I., Wuichet, K., Li, Z., Chen, S., et al. (2009). Bilwes, A. M., Alex, L. A., Crane, B. R., and Simon, M. I. (1999). Structure of Universal architecture of bacterial chemoreceptor arrays. Proc. Natl. Acad. Sci. CheA, a signal-transducing histidine kinase. Cell 96, 131–141. doi: 10.1016/ U.S.A. 106, 17181–17186. doi: 10.1073/pnas.0905181106 s0092-8674(00)80966-6 Buddelmeijer, N., Aarsman, M., and den Blaauwen, T. (2013). Immunolabeling of Blat, Y., and Eisenbach, M. (1995). Tar-dependent and -independent pattern proteins in situ in Escherichia coli K12 strains. Bio-Protocol 3:e852. formation by Salmonella typhimurium. J. Bacteriol. 177, 1683–1691. doi: 10. Butler, M. T., Wang, Q., and Harshey, R. M. (2010). Cell density and mobility 1128/jb.177.7.1683-1691.1995 protect swarming bacteria against antibiotics. Proc. Natl. Acad. Sci. U.S.A. 107, Block, S. M., Segall, J. E., and Berg, H. C. (1983). Adaptation kinetics in bacterial 3776–3781. doi: 10.1073/pnas.0910934107 chemotaxis. J. Bacteriol. 154, 312–323. doi: 10.1128/jb.154.1.312-323.1983 Campoy, S., Jara, M., Busquets, N., de Rozas, A. M. P., Padiola, I., Barbé, Brennan, C. A., DeLoney-Marino, C. R., and Mandel, M. J. (2013). Chemoreceptor J., et al. (2002). Intracellular cyclic AMP concentration is decreased in VfcA mediates amino acid chemotaxis in Vibrio fischeri. Appl. Environ. Salmonella typhimurium fur mutants. Microbiology 148, 1039–1048. doi: 10. Microbiol. 79, 1889–1896. doi: 10.1128/aem.03794-12 1099/00221287-148-4-1039 Briegel, A., Ladinsky, M. S., Oikonomou, C., Jones, C. W., Harris, M. J., Fowler, Cardozo, M. J., Massazza, D. A., Parkinson, J. S., and Studdert, C. A. (2010). D. J., et al. (2014a). Structure of bacterial cytoplasmic chemoreceptor arrays Disruption of chemoreceptor signalling arrays by high levels of CheW, the and implications for chemotactic signaling. Elife 2014, 1–16. doi: 10.7554/eLife. receptor-kinase coupling protein. Mol. Microbiol. 75, 1171–1181. doi: 10.1111/ 02151 j.1365-2958.2009.07032.x Frontiers in Microbiology | www.frontiersin.org 13 April 2020 | Volume 11 | Article 583 Frutos-Grilo et al. Chemotaxis Needs the RecA-CheA-CheW Interaction Cassidy, C. K., Himes, B. A., Alvarez, F. J., Ma, J., Zhao, G., Perilla, J. R., et al. (2015). Johnson, K. S., and Ottemann, K. M. (2018). Colonization, localization, and CryoEM and computer simulations reveal a novel kinase conformational switch inflammation: the roles of H. pylori chemotaxis in vivo. Curr. Opin. Microbiol. in bacterial chemotaxis signaling. Elife 4:e08419. doi: 10.7554/eLife.08419 41, 51–57. doi: 10.1016/j.mib.2017.11.019 Chaveroche, M. K., Ghigo, J. M., and d’Enfert, C. (2000). A rapid method for Jones, C. W., and Armitage, J. P. (2015). Positioning of bacterial chemoreceptors. efficient gene replacement in the filamentous fungus Aspergillus nidulans. Trends Microbiol. 23, 247–256. doi: 10.1016/j.tim.2015.03.004 Nucleic Acids Res. 28:E97. Källberg, M., Wang, H., Wang, S., Peng, J., Wang, Z., Lu, H., et al. (2012). Template- Chen, Z., Yang, H., and Pavletich, N. P. (2008). Mechanism of homologous based protein structure modeling using the RaptorX web server. Nat. Protoc. 7, recombination from the RecA-ssDNA/dsDNA structures. Nature 453, 489–484. 1511–1522. doi: 10.1038/nprot.2012.085 doi: 10.1038/nature06971 Keppler, A., Gendreizig, S., Gronemeyer, T., Pick, H., Vogel, H., and Johnsson, K. Comeau, S. R., Gatchell, D. W., Vajda, S., and Camacho, C. J. (2004). ClusPro: a (2003). A general method for the covalent labeling of fusion proteins with small fully automated algorithm for protein-protein docking. Nucleic Acids Res. 32, molecules in vivo. Nat. Biotechnol. 21, 86–89. doi: 10.1038/nbt765 96–99. doi: 10.1093/nar/gkh354 Keyamura, K., Sakaguchi, C., Kubota, Y., Niki, H., and Hishida, T. (2013). RecA Cox, M. M. (1999). Recombinational DNA repair in bacteria and the RecA protein. protein recruits structural maintenance of chromosomes (SMC)-like RecN Prog. Nucleic Acid Res. Mol. Biol. 63, 311–366. doi: 10.1016/s0079-6603(08) protein to DNA double-strand breaks. J. Biol. Chem. 288, 29229–29237. doi: 60726-6 10.1074/jbc.M113.485474 Cunningham, B., and Wells, J. (1989). High-resolution epitope mapping of hGH- Koler, M., Peretz, E., Aditya, C., Shimizu, T. S., and Vaknin, A. (2018). Long- receptor interactions by alanine-scanning mutagenesis. Science 244, 1081–1085. term positioning and polar preference of chemoreceptor clusters in E. coli. Nat. doi: 10.1126/science.2471267 Commun. 91, 1–10. doi: 10.1038/s41467-018-06835-5 Datsenko, K. A., and Wanner, B. L. (2000). One-step inactivation of chromosomal Lesterlin, C., Ball, G., Schermelleh, L., and Sherratt, D. J. (2014). RecA bundles genes in Escherichia coli K-12 using PCR products. Proc. Natl. Acad. Sci. U.S.A. mediate homology pairing between distant sisters during DNA break repair. 97, 6640–6645. doi: 10.1073/pnas.120163297 Nature 506, 249–253. doi: 10.1038/nature12868 Davis, R., Botstein, D., and Roth, J. (1980). Advance Bacterial Genetics: Manual for Li, M., and Hazelbauer, G. L. (2004). Cellular stoichiometry of the components of Genetic Engineering. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory. the chemotaxis signaling complex. J. Bacteriol. 186, 3687–3694. doi: 10.1128/jb. Erhardt, M. (2016). Strategies to block bacterial pathogenesis by interference with 186.12.3687-3694.2004 motility and chemotaxis. Science 398, 185–205. doi: 10.1007/82.2016.493 Li, M., and Hazelbauer, G. L. (2011). Core unit of chemotaxis signaling Falke, J. J., Bass, R. B., Butler, S. L., Chervitz, S. A., and Danielson, M. A. (1997). The complexes. Proc. Natl. Acad. Sci. U.S.A. 108, 9390–9395. doi: 10.1073/pnas.1104 two-component signaling pathway of bacterial chemotaxis: a molecular view of 824108 signal transduction by receptors, kinases, and adaptation enzymes. Annu. Rev. Li, M., and Hazelbauer, G. L. (2014). Selective allosteric coupling in core Cell Dev. Biol. 13, 457–512. doi: 10.1146/annurev.cellbio.13.1.457 chemotaxis signaling complexes. Proc. Natl. Acad. Sci. U.S.A. 111, 15940–15945. Frank, V., Piñas, G. E., Cohen, H., Parkinson, J. S., and Vaknin, A. (2016). doi: 10.1073/pnas.1415184111 Networked chemoreceptors benefit bacterial chemotaxis performance. MBio Li, X., Fleetwood, A. D., Bayas, C., Bilwes, A. M., Ortega, D. R., Falke, J. J., et al. 7:e1824-16. doi: 10.1128/mBio.01824-16 (2013). The 3.2 Å resolution structure of a receptor: CheA:CheW signaling Gautier, A., Juillerat, A., Heinis, C., Corrêa, I. R., Kindermann, M., Beaufils, F., complex defines overlapping binding sites and key residue interactions within et al. (2008). An engineered protein tag for multiprotein labeling in living cells. bacterial chemosensory arrays. Biochemistry 52, 3852–3865. doi: 10.1021/ Chem. Biol. 15, 128–136. doi: 10.1016/j.chembiol.2008.01.007 bi400383e Gómez-Gómez, J. M., Manfredi, C., Alonso, J. C., and Blázquez, J. (2007). A novel Little, J. W., and Mount, D. W. (1982). The SOS regulatory system of Escherichia role for RecA under non-stress: promotion of swarming motility in Escherichia coli. Cell 29, 11–22. doi: 10.1016/0092-8674(82)90085-x coli K-12. BMC Biol. 5:14. doi: 10.1186/1741-7007-5-14 Liu, J., Hu, B., Morado, D. R., Jani, S., Manson, M. D., and Margolin, W. (2012). Greenfield, D., McEvoy, A. L., Shroff, H., Crooks, G. E., Wingreen, N. S., Betzig, Molecular architecture of chemoreceptor arrays revealed by cryoelectron E., et al. (2009). Self-Organization of the Escherichia coli chemotaxis network tomography of Escherichia coli minicells. Proc. Natl. Acad. Sci. U.S.A. 109, imaged with super-resolution light microscopy. PLoS Biol. 7:e1000137. doi: E1481–E1488. doi: 10.1073/pnas.1200781109 10.1371/journal.pbio.1000137 Lusetti, S. L., and Cox, M. M. (2002). The bacterial RecA protein and the Han, K. Y., and Ha, T. (2015). Dual-color three-dimensional STED microscopy recombinational DNA repair of stalled replication forks. Annu. Rev. Biochem. with a single high-repetition-rate laser. Opt. Lett. 40, 2653–2656. 71, 71–100. doi: 10.1146/annurev.biochem.71.083101.133940 He, K., and Bauer, C. E. (2014). Chemosensory signaling systems that control Maddock, J. R., and Shapiro, L. (1993). Polar location of the chemoreceptor bacterial survival. Trends Microbiol. 22, 389–398. doi: 10.1016/j.tim.2014.04.004 complex in the Escherichia coli cell. Science 259, 1717–1723. doi: 10.1126/ Huang, L., Wang, L., Lin, X., Su, Y., Qin, Y., Kong, W., et al. (2017). mcp, aer, science.8456299 cheB, and cheV contribute to the regulation of Vibrio alginolyticus (ND-01) Maki, N., Gestwicki, J. E., Lake, E. M., Laura, L., Adler, J., and Kiessling, L. L. (2000). adhesion under gradients of environmental factors. Microbiologyopen 6:e00517. Motility and chemotaxis of filamentous cells of Escherichia coli. J. Bacteriol. 182, doi: 10.1002/mbo3.517 4337–4342. doi: 10.1128/jb.182.15.4337-4342.2000 Huang, Z., Pan, X., Xu, N., and Guo, M. (2019a). Bacterial chemotaxis coupling Maslowska, K. H., Makiela-Dzbenska, K., and Fijalkowska, I. J. (2019). The SOS protein: Structure, function and diversity. Microbiol. Res. 219, 40–48. doi: 10. system: a complex and tightly regulated response to DNA damage. Environ. 1016/j.micres.2018.11.001 Mol. Mutagen. 60, 368–384. doi: 10.1002/em.22267 Huang, Z., Wang, Y. H., Zhu, H. Z., Andrianova, E. P., Jiang, C. Y., Li, D., et al. Mayola, A., Irazoki, O., Martínez, I. A., Petrov, D., Menolascina, F., Stocker, (2019b). Cross talk between chemosensory pathways that modulate chemotaxis R., et al. (2014). RecA protein plays a role in the chemotactic response and and biofilm formation. MBio 10:e2876-18. chemoreceptor clustering of Salmonella enterica. PLoS ONE 9:e105578. doi: Irazoki, O., Aranda, J., Zimmermann, T., Campoy, S., Barbe, J., and Barbé, 10.1371/journal.pone.0105578 J. (2016a). Molecular interaction and cellular location of RecA and CheW Natale, A. M., Duplantis, J. L., Piasta, K. N., and Falke, J. J. (2013). Structure, proteins in Salmonella enterica during SOS response and their implication in function and on-off switching of a core unit contact between CheA kinase swarming. Front. Microbiol. 7:1560. doi: 10.3389/fmicb.2016.01560 and CheW adaptor protein in the bacterial chemosensory array: a disulfide Irazoki, O., Mayola, A., Campoy, S., Barbé, J., and Barbe, J. (2016b). SOS mapping and mutagenesis study. Biochemistry 52, 7753–7765. doi: 10.1021/bi40 system induction inhibits the assembly of chemoreceptor signaling clusters 1159k in Salmonella enterica. PLoS ONE 11:e0146685. doi: 10.1371/journal.pone. Park, S. Y., Borbat, P. P., Gonzalez-Bonet, G., Bhatnagar, J., Pollard, A. M., Freed, 0146685 J. H., et al. (2006). Reconstruction of the chemotaxis receptor-kinase assembly. Irazoki, O., Campoy, S., and Barbé, J. (2017). The transient multidrug resistance Nat. Struct. Mol. Biol. 13, 400–407. doi: 10.1038/nsmb1085 phenotype of Salmonella enterica swarming cells is abolished by sub-inhibitory Partridge, J. D., Nhu, N. T. Q., Dufour, Y. S., and Harshey, R. M. (2019). Escherichia concentrations of antimicrobial compounds. Front. Microbiol. 8:1360. doi: 10. coli remodels the chemotaxis pathway for swarming. MBio 10, e316–e319. doi: 3389/fmicb.2017.01360 10.1128/mBio.00316-19 Frontiers in Microbiology | www.frontiersin.org 14 April 2020 | Volume 11 | Article 583 Frutos-Grilo et al. Chemotaxis Needs the RecA-CheA-CheW Interaction Patel, M., Jiang, Q., Woodgate, R., Cox, M. M., and Goodman, M. F. (2010). Sourjik, V., and Berg, H. C. (2000). Localization of components of the A new model for SOS-induced mutagenesis: how RecA protein activates chemotaxis machinery of Escherichia coli using fluorescent protein DNA polymerase V. Crit. Rev. Biochem. Mol. Biol. 45, 171–184. doi: 10.3109/ fusions. Mol. Microbiol. 37, 740–751. doi: 10.1046/j.1365-2958.2000. 10409238.2010.480968 02044.x Piñas, G. E., DeSantis, M. D., and Parkinson, J. S. (2018). Noncritical signaling role Sourjik, V., and Wingreen, N. S. (2012). Responding to chemical gradients: of a kinase-receptor interaction surface in the Escherichia coli chemosensory bacterial chemotaxis. Curr. Opin. Cell Biol. 24, 262–268. doi: 10.1016/j.ceb.2011. core complex. J. Mol. Biol. 430, 1051–1064. doi: 10.1016/J.JMB.2018.02.004 11.008 Piñas, G. E., Frank, V., Vaknin, A., and Parkinson, J. S. (2016). The source of Story, R. M., Weber, I. T., and Steitz, T. A. (1992). The structure of the E. coli RecA high signal cooperativity in bacterial chemosensory arrays. Proc. Natl. Acad. Sci. protein monomer and polymer. Nature 355, 318–325. doi: 10.1038/355318a0 U.S.A. 113, 3335–3340. doi: 10.1073/pnas.1600216113 Vu, A., Wang, X., Zhou, H., and Dahlquist, F. W. (2012). The receptor-CheW Raterman, E. L., and Welch, R. A. (2013). Chemoreceptors of Escherichia coli binding interface in bacterial chemotaxis. J. Mol. Biol. 415, 759–767. doi: 10. CFT073 play redundant roles in chemotaxis toward urine. PLoS ONE 8:e54133. 1016/j.jmb.2011.11.043 doi: 10.1371/journal.pone.0054133 Zaitsev, E. N., and Kowalczykowski, S. C. (1999). Enhanced monomer- Santos, T. M., Lin, T. Y. Y., Rajendran, M., Anderson, S. M., and Weibel, D. B. monomer interactions can suppress the recombination deficiency of the (2014). Polar localization of Escherichia coli chemoreceptors requires an intact recA142 allele. Mol. Microbiol. 34, 1–9. doi: 10.1046/j.1365-2958.1999. Tol-Pal complex. Mol. Microbiol. 92, 985–1004. doi: 10.1111/mmi.12609 01552.x Sassanfar, M., and Roberts, J. W. (1990). Nature of the SOS-inducing signal in Escherichia coli. The involvement of DNA replication. J. Mol. Biol. 212, 79–96. Conflict of Interest: The authors declare that the research was conducted in the doi: 10.1016/0022-2836(90)90306-7 absence of any commercial or financial relationships that could be construed as a Schrödinger, L. (2010). The PyMOL Molecular Graphics System, Version 1.3r1. potential conflict of interest. Available online at: https://www.scirp.org/(S(vtj3fa45qm1ean45vvffcz55))/ reference/ReferencesPapers.aspx?ReferenceID=1571978. Copyright © 2020 Frutos-Grilo, Marsal, Irazoki, Barbé and Campoy. This is an Sievers, F., Wilm, A., Dineen, D., Gibson, T. J., Karplus, K., Li, W., et al. (2011). open-access article distributed under the terms of the Creative Commons Attribution Fast, scalable generation of high-quality protein multiple sequence alignments License (CC BY). The use, distribution or reproduction in other forums is permitted, using Clustal Omega. Mol. Syst. Biol. 7:539. doi: 10.1038/msb.2011.75 provided the original author(s) and the copyright owner(s) are credited and that the Skiba, M. C., Logan, K. M., and Knight, K. L. (1999). Intersubunit proximity of original publication in this journal is cited, in accordance with accepted academic residues in the RecA protein as shown by engineered disulfide cross-links. practice. No use, distribution or reproduction is permitted which does not comply Biochemistry 38, 11933–11941. doi: 10.1021/BI991118Z with these terms. Frontiers in Microbiology | www.frontiersin.org 15 April 2020 | Volume 11 | Article 583