Fig 1.
The A. baumannii T6SS gene cluster: Systematic review of genes important for the T6SS activity.
(A) Schematic model of A. baumannii T6SS nanomachine and representation of the genetic organization of a vgrG locus and the T6SS cluster from A. baumannii 17978. The T6SS gene cluster of A. baumannii ATCC 17978 covers a 23 kb region that includes 18 putative genes predicted to encode for T6SS components. Genes encoding the Membrane Complex (MC), Tube Complex Tail (TTC), and the Baseplate (BP) in blue, yellow and green respectively. T6SS associated genes are represented in grey. 4 proteins encoded by genes only identified in A. baumannii T6SS gene clusters (asa, Acinetobacter type six secretion system-associated, in bold). According to our results and previous studies, some of those proteins have a membrane protein signature (green stars), are essential for the T6SS functionality (red stars), and others are not essential (black stars). (B) Western blot assays probing for Hcp secretion and RNA polymerase (RNAP) in whole-cell pellets (P) and supernatants (S), by A. baumannii ATCC 17978 wildtype (WT, parental strain) and several mutants. The RNAP was used as lysis control. The supernatants of each strain of A. baumannii are isolated, concentrated and then analyzed by denatured 12.5%-polyacrylamide gel electrophoresis (PAGE). Immunodetected proteins are indicated on the right. Molecular weight markers (in kDa) are indicated on the left. (C) Bacterial competition experiments. Survival of E. coli rifampicin-resistant after incubation with ATCC 17978 WT and several mutants. The A. baumannii predator strains indicated are on the x axis and the log-transformed surviving E. coli CFU count is on the y axis. Bars are mean values and dots are three independent replicates.
Fig 2.
A. baumannii T6SS essential for the MC biogenesis.
(A) Fluorescent cluster of the TssMsfGFP protein in A. baumannii tssM-sfGFP strain. On the left panel, the images represent the average over a time lapse acquisition of 61 frames (10s / frame). scale = 2 μm. The right panel represents the histograms of cluster distribution of the main axis of the cell with a preferential accumulation at the poles of the cell. (n > 1000 cells, two biological replicates). Comparison of the cellular distribution of TssM-sfGFP foci between the WT and tsmK mutant (right panels). (B) Statistical analysis of the amount of fluorescence foci in the different mutants. On the left, a graphical representation of the percentage of cells with zero (n = 0), one (n = 1), two (n = 2), or more than three (n = 3) foci in the different A. baumannii mutants. On the right, a graphical representation of the average number of foci per cell in the different mutants. The experiment was performed in triplicate on three different groups of cells. (C-E) Dynamic study of the fluorescence foci of the TssM-sfGFP in WT and tsmK mutant. Fluorescence microscopy in TIRF mode captured an average of 61 images every 10 seconds, revealing multiple fluorescent foci within the cells. Cell masks were generated using the Cellpose cyto2 model, and FIJI’s MicrobeJ plugin, set in "rod-shape" mode. The lower panel shows the average of images processed with a bandpass FFT filter and background subtraction that enhances the foci contrast. (D) Kymographs derived from intensity profiles measured along the cell contour, representing time (61 images every 10 seconds) on the vertical axis and cell contour length in microns on the horizontal axis. Traces of foci movement along the cell contour are evident, with fixed and stable foci yielding vertical traces, some corresponding to the whole acquisition duration (the height of the kymograph = 610 seconds). (E). Violin plots for the quantitative analysis of foci traces between the WT and tsmK mutant (3 biological replications). The difference in means was tested by t.test (R software) and yields a p value < 0.0001. (F) Biogenesis of the A. baumannii-T6SS MC and the original degradation of the tail after contraction. Time-lapse fluorescence microscopy recordings showing localization and dynamics of the mCherryClpV and TssMsfGFP fusions proteins. Individual images were taken every 40 sec. The positions of the foci are indicated by an asterisk. The scale bars are 1 μm. The lines (from top to bottom) represent the phase contrast, the mCherry channel, the sfGFP channel and the superposition of the two channels. Below, a schematic representation of the sequential biogenesis of the T6SS membrane complex.
Fig 3.
Determination of the protein-protein interaction network (PPIN) of A. baumannii T6SS membrane complex.
(A) Western blot probing binary interaction between TagX, TssM, TssL and TsmK respectively. (B) Co-purification of the 3-, 4- and 5-protein complexes. (A-D) Detergent-solubilized extract of E. coli BL21(DE3) cells producing the indicated protein were submitted to an affinity purification step on Strep- or His- Trap. The detergent-solubilized total membrane (L), and eluate (E) were subjected to denaturing 12.5%-polyacrylamide gel electrophoresis (PAGE) and immunodetected with the appropriate antibody. The composition of the protein mix is indicated on the top of each panel, the protein that binds to the column is underlined. Immunodetected proteins are indicated on the right. Molecular weight markers (in kDa) are indicated on the left. Tags: H, 6×His; S, Strep-tag; F, FLAG; HA, Hemagglutinin; V, VSVG. (C) Negative controls. (D) Western blot probing the binary interaction between TsmK and TssK (left panel), including negative control (TssK alone, right panel). (E) Native pull-down of TssM from A. baumannii cells. Western blot probing the identified interaction partners (left panel). Mass spectrometry analysis of TssM copurified interactants (right panel).(F) Schematic interaction network determined by co-purification and topology of the A. baumannii T6SS-MC full-length proteins. Interactions between proteins are represented with arrows: TsmK is in the center of the interaction network (red arrows). Indirect interactions (native pull-down are shown by dashed double-head arrows.
Fig 4.
Structural prediction study of TsmK an Acinetobacter-specific protein related to Ketoacyl synthases.
Analysis of the TsmK predicted structure. (A) AlphaFold2 confident score (predicted LDDT per position) mapped on the model. (B) Topology of the secondary elements of the predicted model for TsmK. The diagram depicts the secondary structure organization of the A. baumannii TsmK model generated with AlphaFold2. The major and minor β-sheets, as well as the long loop, are highlighted with colored squares. Inconsistent regions are represented with pale colors. Five antiparallel β-strands, β1, β7, β6, β4, and β5 assemble into a β-sheet (referred to as the major β-sheet). The α-helices are present at the loops formed between each pair of consecutive β-strand and decorate both sides of the β-sheet. AlphaFold2 predicted a second β-sheet (referred to as minor β-sheet) not present in TrRosetta and RaptorX predictions. This β-sheet is formed with the two β-strands of the converging hairpin (β9 and β10) and the β11, β12, and β8 (C) Structural superimposition of the three structural models of A. baumannii TsmK. The major and minor β-sheets as well as the long loop are highlighted with colored squares. Structural precision ranges between 0 and 15 Å (blue to red) (D) Topology of the secondary elements of the Ketoacyl synthase domain from Acyltransferase type I polyketide synthase (PKS) (PDB 4TKT). The homologous regions between TsmK and the KS domain of the polyketide synthase are highlighted in dashed colors.
Fig 5.
Characterization of TssM location.
(A) TssM location determined by sucrose gradient. A. baumannii ATCC 17978 strains were grown, and the outer (OM) and inner (IM) membranes were separated by a continuous sucrose gradient (35%-60%). After centrifugation, 750ul samples were taken from the top to the bottom of the gradient. The samples were subjected to denaturing 12.5%- polyacrylamide gel electrophoresis (PAGE) and immunodetected with the appropriate antibody (including synthetics polyclonal TssM antibody). Immunodetected proteins are indicated on the right. Molecular weight markers (in kDa) are indicated in the left. Tags: GspCsfGFP (T2SS IM protein), Omp28mCherry (OM porin). (B) Dynamics of fluorescent clusters of the TssM protein. Cells were recorded in TIRF illumination during 10 min each 10 seconds (a). The column in panel (a) shows the corresponding average image of the 4 cells. Panel (b) shows 6 time points of these 4 cells. The arrows point to the cluster fluctuations (line 2) or their lateral displacement (line 3). The lower panels (c) are kymographs that illustrate 4 types of movements: fixed clusters (F), intermittent clusters (I), lateral displacement (L), cell without clusters (WC). The kymographs were constructed from the intensity profile of the cell contour. Scale: 2 μm (C) In A. baumannii ATCC 17978, the fusion of sfGFP to TssM C-terminal (TssM-sfGFP) and mCherry to a majority porin (Omp28-mCherry) was used to observe the location of the TssM C-terminal. The porin was used as a proxy to label the outer membrane. (D) Structured illumination microscopy (SIM) images of the TssMsfGFP regarding the Omp28mCherry label. The scale bar is 1μm. (E) Box plot representing the distances measured between the TssM foci and the mCherry label in a WT, a clpV mutant and a tssB-tssC mutant. The difference in position is equivalent to the distance between the two observed fluorophores and was measured in n = 60 bacteria (WT_ ΔclpV: t 3.1331, df = 105.39, p-value = 0.00224; WT_ ΔtssBΔtssC: t = 0.024481, df = 100.92, p-value = 0.9805; ΔclpV_ΔtssBΔtssC: t = -2.6278, df = 95.461, p-value = 0.01002).
Fig 6.
The A. baumannii-specific TssM C-terminal GS-linker is essential for T6SS assembly and functioning.
(A) Schematic topology of A. baumannii TssM protein. This protein has a huge periplasmic domain composed of three alpha domains (circular shape) and one beta domain (scare shape) on the C-terminal. On the C-terminal, the protein has a GxxGxxxGxxG helix (the “GS-linker”, in yellow) inserted between the last α- and the β- domain of the protein. (B) The 807 aligned TssM sequences were extracted from the set of Acinetobacter genomes with a complete T6SS operon. The height of each letter represents the information content of the corresponding residue at a given position in bits. A. baumannii strain ATCC 17978 TssM (on top of the sequence logo) sequence was used as a reference position. (C) TssM C-terminal oligomerization depends on the “GS-linker” motif. On the top, a schematic representation of the TssM C- terminal constructions used (the periplasmic domain between 930 and 1228 amino acids) to determine the interactions. Soluble extracts of E. coli BL21(DE3) cells producing the indicated protein were submitted to a simple affinity purification step on Strep-Trap. The lysate (total soluble extracts, L), and eluate (E) were subjected to denaturing 12.5%-polyacrylamide gel electrophoresis (PAGE) and immunodetected with the appropriate antibody. Immunodetected proteins are indicated on the right. Molecular weight markers (in kDa) are indicated in the left. Tags: H, 6×His; S, Strep-tag. (D) Upper panel, western blot assays probing for Hcp secretion in whole-cell pellets (P) and supernatants (S), by A. baumannii ATCC 17978 wildtype with the plasmid control, the ΔtssM mutant and the complemented mutant strain. The last two show the phenotype of ΔtssM mutant owning a plasmid pVRL2- with tssM mutated (Gly→Ala substitution in position 1172, 1175 and 1179 or a deletion of the alpha helix between the 1172 and 1179 residues). The TssM expression was induced by adding 0.1% arabinose. Lower panel, bacterial competition experiments. Survival of E. coli rifampicin-resistant after incubation with ATCC 17978 wildtype with the plasmid control, the ΔtssM mutant and the complemented mutant strain. (E) The “Small Domain Interference” (SDI) validate the importance of TssM-CTD. Dot blot quantification of Hcp secretion in wildtype with the plasmid control (pVRL2-) or a plasmid overexpressing the TssM C-terminal peptide, and in the ΔtssM mutant with a plasmid control. The C-terminal peptide was induced by adding 1% arabinose. (F) Outer membrane permeability test. Scheme representing the different protein constructs produced in E. coli BL21 (left panel). Exponential-phase cultures of E. coli overproducing different protein domains were adjusted for OD600 and serial diluted onto LB media supplemented with vancomycin 10μg/mL and 0.05mM IPTG (right panel).
Fig 7.
Model for the assembly of the A. baumannii T6SS MC.
(A) Conditional probability tree of the co-occurrence of “GS-linker motif”, TsmK, and TssJ on genomes containing both a tssB and a tssM gene. Percentages inside the rounded rectangles represent the conditional probabilities after each branching. Percentages reported on the bottom are the joint probabilities of the different GS-linker -TsmK -TssJ configurations in baumannii, non-baumannii and non-Acinetobacter genomes. Rounded red and green rectangles represent the absence and the presence of one of the two genes (tssJ or tsmK) or the GS-linker motif, respectively. (B) A. baumannii T6SS offers a unique glimpse of molecular evolution showing how the absence of a major T6SS component (like TssJ, upper panel) could be compensated by small additional domain or co-opted protein (lower panel). (1) TsmK stabilizes the pillar of the membrane complex (MC), the protein TssM. (2) Helped by the GS-linker, TssM oligomerizes to assemble a mega-membrane complex with TsmK. (3) TssL, TagX and TlsA are recruited to form the Acinetobacter T6SS MC. (4) The molecular mechanism and dynamic of outer membrane anchoring and piercing by the MC is not yet known.