D-amino acids signal a stress-dependent run-away response in Vibrio cholerae

Abstract

To explore favourable niches while avoiding threats, many bacteria use a chemotaxis navigation system. Despite decades of studies on chemotaxis, most signals and sensory proteins are still unknown. Many bacterial species release D-amino acids to the environment; however, their function remains largely unrecognized. Here we reveal that D-arginine and D-lysine are chemotactic repellent signals for the cholera pathogen Vibrio cholerae. These D-amino acids are sensed by a single chemoreceptor MCP_DRK co-transcribed with the racemase enzyme that synthesizes them under the control of the stress-response sigma factor RpoS. Structural characterization of this chemoreceptor bound to either D-arginine or D-lysine allowed us to pinpoint the residues defining its specificity. Interestingly, the specificity for these D-amino acids appears to be restricted to those MCP_DRK orthologues transcriptionally linked to the racemase. Our results suggest that D-amino acids can shape the biodiversity and structure of complex microbial communities under adverse conditions.

Fig. 1. d-Arg and d-Lys signal a chemorepellent response in V. cholerae through the putative MCP VC1313.

a, Schematic representation of DAA racemization. The most abundant amino acids produced are shown (open circles, l-amino acids; filled circles, d-amino acids). OM, outer membrane; IM, inner membrane; CYT, cytoplasm. b, Representative images of V. cholerae wild type (wt) and ΔbsrV mutant expanding in 0.3% soft agar. Images are representative of experiments repeated at least three times. c, Motility and d-amino acid production analysis of BsrV derivatives; grey bars indicate relative motility compared to wt or ΔbsrV knockout mutant, catalytic site mutant (BsrV K95A) and the complemented strain harbouring bsrV under an aTc inducible promoter (pbsrV); green dots represent the amount of secreted d-amino acids. pØ, empty plasmid. Error bars represent mean ± s.d. of 3 biologically independent replicates. Significant differences (paired t-test) are indicated by **P < 0.01 or ***P < 0.001. d, Motility of ΔbsrV mutant in soft-agar plates that are chemically complemented with 5 mM d-amino acids are shown relative to the wild-type strain. Error bars represent mean ± s.d. of 6 biologically independent replicates examined over 2 independent experiments. Significant differences (one-way analysis of variance (ANOVA)) are indicated by ****P < 0.0001. e, Chemotactic response to 0.1 mM d-amino acids in capillary assays. Chemotaxis ratio (CR) was calculated relative to the control capillary containing no stimulus. CR > 1, attractants; CR < 1, repellents; CR = 1, no response. Black diamonds represent the mean of 3 independent biological replicates. Significant differences (unpaired t-test) are indicated by *P < 0.05 or ***P < 0.001. f, Thermal proteome profiling analysis of V. cholerae cell extracts exposed to d-Arg. Stabilized and destabilized proteins in the presence of d-Arg are highlighted in red and green, respectively. From the 45 MCPs found in V. cholerae’s genome (black), only 3 interacted with d-Arg: VC2161, VC1313 and VC1406. g, Chemotactic response of selected MCP candidates to d-Arg. Double mutants (ΔbsrV Δmcp) were constructed and the ability to respond to d-Arg was tested by capillary assays. ΔbsrV strain was used as background. LAA, l-amino acids; DAA, d-amino acids. Black diamonds represent the mean of 3 independent biological replicates. Significant differences (unpaired t-test) are indicated by ***P < 0.001. Source data

Fig. 2. d-Arg production and chemotactic response are coordinated by RpoS.

a, Left: localization of sfGFP-MCP_DRK expressed from its native locus and promoter at exponential and stationary phase. Representative micrographs of 3 independent replicates are shown; at least 3 images were acquired per timepoint. Scale bar, 2 μm. Right: demographic analysis of the fluorescence intensity of sfGFP-MCP_DRK relative to cell length (n > 500 cells). The colour bar represents fluorescence intensity. b, Bottom: growth-phase dependent subcellular localization of sfGFP-MCP_DRK. Filled grey circles, OD₆₀₀ of V. cholerae grown in TB; open tradewind-blue circles, percentage of cells with sfGFP-MCP_DRK polar foci. Top: western blot against sfGFP-MCP_DRK using anti-GFP-specific antibody at the indicated OD₆₀₀. Samples were normalized and loaded with equal protein amount. c, Growth curve of V. cholerae grown in MSR6 minimal medium supplemented with 0.04% w/v d-glucose and 0.4% w/v succinate (grey line) at the indicated timepoints. The corresponding percentage of cells with sfGFP-MCP_DRK foci (open tradewind-blue circles) is also indicated. Red (glucose depletion) and purple (succinate depletion) arrows indicate carbon starvation. Quantification of % of cells with polar foci in b and c was based on a single experiment where >500 cells were counted from 3 different images per timepoint. d, Left: micrographs showing the expression of sfGFP-MCP_DRK in a ΔrpoS mutant background. Representative micrographs of 3 independent replicates are shown; at least 3 images were acquired per timepoint: scale bar, 2 μm. Right: demographic analysis of the fluorescence intensity of sfGFP-MCP_DRK relative to cell length. The colour bar represents fluorescence intensity. e, Chemotaxis response of ΔrpoS mutant to d-Arg. Black diamonds represent the mean of 3 independent biological replicates. Error bars represent mean ± s.d. of 3 biologically independent replicates. Significant differences (unpaired t-test) are indicated by ***P < 0.001. f, DAA production by the different strains. Significant differences (unpaired t-test) are indicated by **P < 0.01 or ***P < 0.001. Source data

Fig. 3. Structural basis of the MCP_DRK chemoreceptor binding d-Arg and d-Lys.

a, Domain architecture of MCP_DRK protein generated by alphafold^–, showing one subunit in rainbow colours, with the N terminus in blue and C terminus in red. The full-length MCP_DRK transmembrane chemoreceptor contains a ligand binding domain (LBD), a transmembrane (TM) and a signalling domain. b, MCP_DRK-LBD monomers bound to d-Arg (purple spheres) or d-Lys (orange spheres) are coloured in rainbow colours from the N terminus (blue) to the C terminus (red). c, Superimposition of the LBD domain of MCP_DRK crystallized with d-Arg (in purple) and d-Lys (in orange) with the aspartate-bound Tar receptor from E. coli (PDB ID: 4z9i in green), the citrate bound MCP2201 chemoreceptor from C. testosterone (PDB ID: 5XUB, in blue) and the more promiscuous quinate-bound PcaY_PP chemoreceptor from P. putida (PDB ID: 6S38, in yellow). The Tar and the PcaY_PP receptor ligand binding sites are located at the dimer interface and are therefore shown as dimers. d, The binding site of MCP_DRK-LBD bound to d-Arg (in purple). e, The binding site of MCP_DRK-LBD bound to d-Lys (in orange). f, Functional analysis of the d-Arg binding pocket was performed by substituting residues Asn43, Asp47, Thr48, Thr105, W107, Glu111 and Asn176 with an alanine. Chemotactic response to d-Arg was then analysed for Δmcp_DKR cells expressing either wild-type or mutant MCP_DRK under its native promoter. ΔbsrV strain was used as background. −, not complemented Δmcp_DRK. Black diamonds represent the mean of 3 independent biological replicates. Significant differences (unpaired t-test) are indicated by *P < 0.05 or **P < 0.01. Source data

Fig. 4. Specificity and conservation of MCP_DRK.

a, Tree of MCP_DRK orthologues. The cladogram was constructed on the basis of MCP_DRK sequence homology. Orthologues selected for an extended comparison are highlighted in bold. b, Left: selection of MCP_DRK orthologue representatives and the conservation of d-Arg binding residues. Only the residues that differ from the reference protein (VC1313, that is, MCP_DRK from V. cholerae El Tor strain N16961, boxed) are labelled. Coloured circles indicate the amino acid type (N, asparagine; D, aspartic acid; T, threonine; W, tryptophan; E, glutamic acid; S, serine; A, alanine; G, glycine). Right: genetic context of the chemoreceptor in several species. Both the MCP and the racemase are highlighted. The percentage of total protein identity (%) is shown. c, Chemotactic response of MCP_DRK orthologue proteins to d-Arg. The chemotactic response to d-Arg was tested for Δmcp_DKR cells complemented with several orthologue MCPs expressed under the mcp_DKR native promoter. ΔbsrV strain was used as background. −, not complemented Δmcp_DRK; VC, V. cholerae; VV, V. vulnificus; VPA, V. parahaemolyticus; VFU, V. furnissii; VA, V. anguillarum; AF, Aliivibrio fischeri; AH, Aeromonas hydrophyla; VO, V. owensii. Black diamonds represent the mean of 4 independent biological replicates. Significant differences (unpaired t-test) are indicated by **P < 0.01. Source data

Fig. 5. Model of MCP_DRK-dependent chemotactic response to d-amino acids.

Under certain environmental stresses (that is, starvation), the RpoS response induces the expression of both the broad-spectrum racemase BsrV and the dedicated d-Arg/d-Lys chemoreceptor MCP_DRK. Among the d-amino acids (DAAs) produced by BsrV, d-Arg and d-Lys stand out as warning signals sensed by V. cholerae as well as other MCP_DRK-encoding species. Chemotactic run-away response to these d-amino acids enables these bacterial communities to move away and explore more favourable niches.

Extended Data Fig. 1. Effect of D-amino acids in motility.

a, Motility analysis of V. cholerae in soft agar in presence of L- and D-amino acids. Grey bars indicate relative motility of BsrV knockout mutant compared to wild-type (wt) both in presence and absence of 5 mM amino acids. b, Effect of D-amino acids (DAAs) in growth. Survival rate was calculated as relative maximum growth in late stationary phase using the not supplemented culture as reference. Cultures were supplemented with 5 mM DAAs. c, Spot assay for the indicated V. cholerae strains. Cells were grown in TB until stationary phase, serially diluted and spotted on LB agar plates supplemented with 5 mM DAAs. Error bars in a and b represent mean values ± SD of 6 and 3 biologically independent replicates. Significant differences (One-way ANOVA, p-values were adjusted for multiple comparison using Bonferroni-Dunn method) are indicated by **** (p < 0.0001). Source data

Extended Data Fig. 2. Comparison of chemotactic ability of V. cholerae wild-type and ΔbsrV strains.

a, Chemotactic ability of ΔbsrV mutant. Both chemotactic bias and swimming speed were analysed in presence of several known chemoeffectors to compare the behaviour of wild-type (wt) and ΔbsrV mutant cells. Chemotaxis was measured in microfluidic devices, monitoring cell motion from a reservoir filled with cells in motility buffer to a reservoir filled only with 100 µM of indicated compound in motility buffer. Blank, no attractant in second reservoir; L-Arg, L-arginine; AiBu, α-aminoisobutiric acid; GABA, γ-aminobutyric acid; Succ, succinate. Each data point represents the average over the 3 time points for a given experiment (n = 4) and the line represents the mean. Significant differences (unpaired t-test) are indicated by *** (p < 0.001) and was evaluated relative to the blank for the given strains. b, Swimming speed analysis of V. cholerae grown in presence of D-Arg. Cells were grown in TB (± 500 μM D-Arg) and transferred to fresh TB for swimming speed measurements. Each point represents the mean and error bars indicate the SD over 3 biological replicates. c, Motility analysis of non-chemotactic V. cholerae cells in presence and absence of D-Arg. Relative motility on soft-agar plates supplemented or not with 5 mM D-Arg of several strains compared to wild-type (wt). ΔbsrV and ΔcheY3 strains are included as controls. Error bars in c represent mean values ± SD of 3 biologically independent replicates. Significant differences (unpaired t-test, adjusted for multiple comparison using Bonferroni-Dunn method) are indicated by *** (p < 0.001). See source data. Source data

Extended Data Fig. 3. Thermal proteome profiling.

a, Schematic representation of Thermal Proteome Profiling (2D-TPP) experimental setting. 4 different conditions were tested: 0, 0.05, 0.4, 3.2 mM D-Arg. Cell extracts were prepared and treated with a range of D-Arg concentrations (1); the samples were heated to a range of temperatures (2); then soluble proteins fraction was extracted and digested to analyze by mass-spectrometry (3); finally the effect of ligand on thermal stability of each protein was evaluated by comparing the remaining soluble fraction at each ligand concentration with the untreated control for each temperature (4). b, Heatmaps for selected candidates targeting D-Arg in V. cholerae. The legend represents the log₂-transformed fold-change protein abundance for a given temperature and D-Arg concentration relative to the no D-Arg control.

Extended Data Fig. 4. Chemotactic ability of V. cholerae vc1313 mutant towards Arg.

a, Comparison between D-Arg and D-Lys chemotaxis responses in V. cholerae. b, Chemotaxis response to L- and D-Arg increasing concentrations for V. cholerae ΔbsrV and ΔbsrV Δvc1313 double mutant. Significant differences (unpaired t-test) are indicated by *** (p < 0.001), n = 3. c, Chemotactic bias of V. cholerae ΔbsrV and ΔbsrV Δvc1313, measured in microfluidics devices, towards a reservoir containing either 500 µM D-Arg in motility medium or only motility medium (blank control), 1, 2 and 3 h after the preparation of the device. d, Difference between the chemotactic bias (that is, the ratio v_ch/v₀ of the chemotactic drift velocity v_ch over the cell swimming speed v₀, see Methods) towards D-Arg and the corresponding control of the experiments of (b) as a function of the time since the preparation of the microfluidic device. e, Population averaged swimming speed measured in the experiments of (b, c). Each data point of in c, d and e represents the mean and error bars ± SEM of n = 6 biologically independent replicates. Significant differences (unpaired t-test) are indicated by ns (not-significant), * (p < 0.05) ** (p < 0.01) or *** (p < 0.001). See source data. Source data

Extended Data Fig. 5. MCP_DRK genetic context and promoter transcriptional activity.

a, Genetic context of vc1313. b, Relative expression of vc1313-vc1312 and adjacent genes across 307 available V. cholerae RNAseq experiments. c, Transcriptional activity of the putative promoter regions of the genes forming the operon (vc1313, vc1312 and vc1311) in wild-type (wt) and ΔbsrV mutant backgrounds. β-galactosidase assay was performed in cultures carrying pCB192N-derivatives grown for 8 hours at 37 °C in LB. β-galactosidase activity is measured in Miller units. d, Transcriptional activity of the Pvc1313 putative promoter region at different timepoints. β-galactosidase assay was performed in cultures of wild-type (wt) and ΔbsrV mutant backgrounds carrying pCB192N-Pvc1313, grown at 37 °C in LB for the indicated time. β-galactosidase activity is measured in Miller units. Error bars in c and d represent mean values ± SD of 3 biologically independent samples.

Extended Data Fig. 6. MCP_DRK forms part of the chemotaxis F6 system.

a, Chemotaxis response to D-Arg of chemotaxis system mutants. ΔI/F9 and ΔIII/F7 showed similar response to D-Arg as control strain. No response was observed for ΔII/F6 system mutant. ΔbsrV strain was used as background. Black diamond represents the mean of 3 independent biological replicates. Significant differences (unpaired t-test) are indicated by *** (p < 0.001), n = 3. b, Co-localization of sfGFP-MCP_DRK and CheY3-dTtomato (chemotaxis system II/F6) fluorescence reporters during stationary phase. c, Localization of sfGFP-MCP_DRK in the three chemotaxis systems knockout backgrounds. Representative micrographs of all experiments are shown. Representative micrographs of 3 independent replicates are shown in b and c, at least 3 images were acquired per condition: scale bar, 2 μm. Source data

Extended Data Fig. 7. Tree of MCP_DRK orthologues.

The cladogram was constructed based on MCP_DRK sequence homology. Representative species and strains were downloaded from Biocyc database. We included the non-related E. coli Tar chemoreceptor to help root the tree (light gray). In the first column, the presence of a Bsr ortholog is represented as full (presence) and empty (absence) circles. The presence of Bsr ortholog clustered together with the racemase is also represented in a similar way in the second column. Conservancy of the LBD residues implied in D-Arg binding is also shown (residue alignment). The bar chart represents the complete protein identity percentage (%). Bacteria were clustered in 3 different groups (cluster I, II, III) depending on the presence/absence of Bsr and MCP_DRK orthologs.