The opportunistic pathogen Stenotrophomonas maltophilia utilizes a type IV secretion system for interbacterial killing

Abstract

Bacterial type IV secretion systems (T4SS) are a highly diversified but evolutionarily related family of macromolecule transporters that can secrete proteins and DNA into the extracellular medium or into target cells. It was recently shown that a subtype of T4SS harboured by the plant pathogen Xanthomonas citri transfers toxins into target cells. Here, we show that a similar T4SS from the multi-drug-resistant opportunistic pathogen Stenotrophomonas maltophilia is proficient in killing competitor bacterial species. T4SS-dependent duelling between S. maltophilia and X. citri was observed by time-lapse fluorescence microscopy. A bioinformatic search of the S. maltophilia K279a genome for proteins containing a C-terminal domain conserved in X. citri T4SS effectors (XVIPCD) identified twelve putative effectors and their cognate immunity proteins. We selected a putative S. maltophilia effector with unknown function (Smlt3024) for further characterization and confirmed that it is indeed secreted in a T4SS-dependent manner. Expression of Smlt3024 in the periplasm of E. coli or its contact-dependent delivery via T4SS into E. coli by X. citri resulted in reduced growth rates, which could be counteracted by expression of its cognate inhibitor Smlt3025 in the target cell. Furthermore, expression of the VirD4 coupling protein of X. citri can restore the function of S. maltophilia ΔvirD4, demonstrating that effectors from one species can be recognized for transfer by T4SSs from another species. Interestingly, Smlt3024 is homologous to the N-terminal domain of large Ca2+-binding RTX proteins and the crystal structure of Smlt3025 revealed a topology similar to the iron-regulated protein FrpD from Neisseria meningitidis which has been shown to interact with the RTX protein FrpC. This work expands our current knowledge about the function of bacteria-killing T4SSs and increases the panel of effectors known to be involved in T4SS-mediated interbacterial competition, which possibly contribute to the establishment of S. maltophilia in clinical and environmental settings.

Fig 1. S. maltophilia uses the X-T4SS to induce E. coli cell death in a contact-dependent manner.

(A) Schematic representation of the organization of the chromosomal virB1-11 and virD4 genes encoding the X-T4SSs of S. maltophilia K279a and X. citri 306. The amino acid identities (%) between homologues are shown. (B) Bacterial competition assay using S. maltophilia strains (wild-type, ΔvirD4 and complemented strains ΔvirD4 virD4_smlt and ΔvirD4 virD4_xac and E. coli (naturally expressing β-galactosidase). A serial dilution of E. coli (1:4) was mixed with constant amounts of S. maltophilia, spotted onto LB-agar containing IPTG and X-gal and incubated for 24 h at 30°C. Representative image of three independent experiments. (C) CFUs ratio of either wild-type or ΔvirD4 S. maltophilia (attacker) to E. coli (target) recovered after 5 h of co-culture in solid or liquid media. CFUs ratios of mixed cultures at the initial time-point was included as a control. (D) Quantification of E. coli target cell lysis using the cell-impermeable compound CPRG. The same bacterial strains described in (B) were used. Graph represents the means and standard deviation (SD) of three independent experiments performed in triplicate. The slopes in the linear part of the curves (between 50 and 100 min) is proportional to the amount of β-galactosidase released by the lysed E. coli cells. (E) Representative images of time-lapse microscopy showing wild-type S. maltophilia interacting with E. coli-RFP (upper panel) at the single cell level. Images were acquired every 10 min. Dead/lysed E. coli-RFP cells are indicated by white arrows. Interaction between S. maltophilia ΔvirD4 and E. coli-RFP strains (lower panel) did not induce target cell lysis. Timestamps in hours:minutes. Scale bar 5 μm. (F) Percentage of dead/lysed E. coli cells after cell-to-cell contact with Stenotrophomonas strains over a 100 min timeframe.

Fig 2. S. maltophilia kills Gram-negative species and duels with X. citri in a T4SS-dependent manner.

(A) Percentage of dead/lysed Klebsiella pneumoniae, Salmonella Typhi and Pseudomonas aeruginosa after cell-to-cell contact with Stenotrophomonas strains over a 100 min time frame. (B) Percentage of T4SS-deficient X. citri cells that lysed after cell-to-cell contact with S. maltophilia cells. (C) Percentage of X. citri cells that lysed after cell-to-cell contact with S. maltophilia cells (left) and % of S. maltophilia cells that lysed after cell-to-cell contact with X. citri cells (right). Cells were counted per interaction over a 300 min time frame. (D) Representative images of time-lapse microscopy showing S. maltophilia wild-type and ΔvirD4 strains interacting with the T4SS-deficient X. citri ΔvirB-GFP strain at the single cell level. Dead/lysed X. citri ΔvirB-GFP cells are indicated by white arrows. (E) S. maltophilia wild-type and ΔvirD4 strains interacting with the X. citri-GFP strain (functional T4SS) at the single cell level. Dead/lysed X. citri cells are indicated by white arrows and dead/lysed S. maltophilia cells are indicated by yellow arrows. Timestamps in hours:minutes. Scale bar 5 μm. Images were acquired every 15 min.

Fig 3. Putative X-T4SS effectors (X-Tfes) and immunity proteins (X-Tfis) of S. maltophilia.

(A) Schematic representation of size and domain architectures of S. maltophilia X-T4SS effectors identified via BLASTp search using XVIPCD (Xanthomonas VirD4-interacting protein conserved domain) of X. citri T4SS effectors. Gene entries are shown for both effectors and their cognate immunity proteins with sizes shown in parenthesis. AHH: putative nuclease domain; PGB: peptidoglycan-binding domain. (B) Alignment of the XVIPCDs of the identified S. maltophilia effectors using Clustal Omega [85] and the consensus sequence logo generated by WebLogo [86] showing several highly conserved amino acids that match conserved residues of the X. citri XVIPCDs [30].

Fig 4. S. maltophilia Smlt3024 induces target cell stasis and Smlt3025 is its inhibitor.

(A) Schematic representation of smlt3024 and smlt3025 genomic organization. (B) Immunoblot showing T4SS-dependent and E. coli contact-dependent secretion/translocation of FLAG-Smlt3024. Representative image of three independent experiments. (C) Densitometry of quantitative dot blot analysis signals shown in (B). Signal intensity detected for S. maltophilia mixed with E. coli were normalized by the background signal detected for S. maltophilia alone. (D) Amino acid sequence of Smlt3024 and Smlt3025 as annotated in S. maltophilia str. K279a genome (GenBank AM743169). Coloured in blue is the Smlt3024 XVIPCD with conserved amino acids in bold. Methionine (M) residues at positions 1, 13, 45, 47 and 50 of Smlt3025 are shown in red. The predicted periplasmic localization signal of Smlt3025 beginning at Met₄₅ is shaded in grey with cleavage and lipidation predicted at the underlined cysteine. (E) Serial dilution (10-fold) of E. coli strains containing pBRA and pEXT22 constructs as indicated, spotted on LB-agar plates. Growth inhibition is observed upon expression of the pelB-smlt3024 construct (periplasmic) and can be reverted by the concomitant expression of Smlt3025_45-333 but not Smlt3025_1-333 or Smlt3025_13-333. (F) Left panel: SEC-MALS analysis showing the formation of a stable complex between Smlt3024 and Smlt3025_86-333. The continuous line corresponds to the normalized differential refractive index, and the spotted lines indicate the calculated molecular mass. Right panel: SDS-PAGE showing the apparent molecular mass of proteins eluted from different SEC peaks. (G) Time-lapse imaging of single cells expressing pBRA-pelB-smlt3024 showing reduced growth rates and smaller cell-sizes (L-arabinose) compared to the non-induced (D-glucose) and empty plasmid controls. Images were acquired every 10 min. Timestamps in hours:minutes. Scale bar 5 μm.

Fig 5. Smlt3024 delivery by X. citri reduces target E. coli cells growth speed.

(A) Upper panel: Representative images of time-lapse microscopy showing X. citri Δ8Δ2609-GFP expressing pBRA-smlt3025_45-333/3024 in contact with E. coli cells carrying empty pEXT22 plasmid. Blue arrows indicate E. coli cells that are in contact with X. citri at time zero, while red arrows show E. coli cells that were not in contact. Lower panel: Schematic representation showing the growth of E. coli cells that were either in contact (blue) or not (red) with X. citri at time zero. (B) Upper panel: Representative images of time-lapse microscopy showing X. citri strain described in (A) in contact with E. coli cells expressing Smlt3025_45-333. Red and blue arrows indicate E. coli cells as in (A). Lower panel: Schematic representation as in (A). Timestamps in hours:minutes. Scale bar 2 μm. Images were recorded every 20 min. (C) Quantitative analysis of the doubling time of E. coli cells either carrying empty plasmid or expressing Smlt3025_45-333 when in contact or not in contact with X. citri Δ8Δ2609-GFP with or without the plasmid expressing the Smlt3024/Smlt3025 X-Tfe/X-Tfi pair. Strains were grown at 28°C. Boxplots represent means ± SD. ND represents cells tagged at time zero that did not divide and were not used to calculate the average doubling times.

Fig 6. Smlt3025 crystal structure presents a topology similar to the iron-regulated protein FrpD of Neisseria meningitidis.

(A) Ribbon representation of the Smlt3025_86-333 structure (PDB 6PDK). The protein has two N-terminal beta hairpins followed by an 8-stranded antiparallel central beta sheet and a C-terminal alpha helix. Some of the loops between beta strands contain additional secondary structure elements. (B) Scheme illustrating Smlt3025 topology. In both (A) and (B), beta strands, 3₁₀ helix and alpha helices are colored green, light blue and marine, respectively. (C) Sequence conservation of Smlt3025 homologues mapped onto the Smlt3025 structure. Coloring generated by Consurf [87]. (D) Structural alignment between Smlt3025_86-333 (blue) and FrpD (orange, PDB 5EDF). Molecular orientations in (A), (C) and (D) are the same.