Delivery determinants of an Acinetobacter baumannii type VI secretion system bifunctional peptidoglycan hydrolase

Abstract

Acinetobacter baumannii is a Gram-negative opportunistic pathogen and is a common cause of nosocomial infections. The increasing development of antibiotic resistance in this organism is a global health concern. The A. baumannii clinical isolate AB307-0294 produces a type VI secretion system (T6SS) that delivers three antibacterial effector proteins that give this strain a competitive advantage against other bacteria in polymicrobial environments. Each effector, Tse15, Tde16, and Tae17, is delivered via a non-covalent interaction with a specific T6SS VgrG protein (VgrG15, VgrG16, and VgrG17, respectively). Here we define the regions of interaction between Tae17 and its cognate delivery protein VgrG17 and identify that amino acids G1069 and W1075 in VgrG17 are essential for Tae17 delivery via the T6SS, the first time such specific delivery determinants of T6SS cargo effectors have been defined. Furthermore, we determine that the Tae17 effector is a multidomain, bifunctional, peptidoglycan-degrading enzyme that has both amidase activity, which targets the sugar-peptide bonds, and lytic transglycosylase activity, which targets the peptidoglycan sugar backbone. Moreover, we show that the Tae17 transglycosylase activity is more important than amidase activity for the killing of Escherichia coli. This study provides molecular insight into how the T6SS allows A. baumannii strains to gain dominance in polymicrobial communities and thus improve their chances of survival and transmission.IMPORTANCEWe have shown that the Acinetobacter baumannii T6SS effector Tae17 is a modular, bifunctional, peptidoglycan-degrading enzyme that has both lytic transglycosylase and amidase activities. Both activities contribute to the ability to degrade peptidoglycan, but the transglycosylase activity was more important for the killing of Escherichia coli. We have defined the specific regions of Tae17 and its cognate delivery protein VgrG17 that are necessary for the non-covalent interactions and, for the first time, identified specific amino acids essential for T6SS cargo effector delivery. This work contributes to our molecular understanding of bacterial competition strategies in polymicrobial environments and may provide a window to design new therapeutic approaches for combating infection by A. baumannii.

Keywords: Acinetobacter baumannii; amidase; lytic transglycosylase; peptidoglycan hydrolase; toxic effector; type VI secretion system.

Fig 1

The Tae17 AlphaFold2 model has a four-domain structure with two likely active sites. (A) Tae17 AlphaFold2 model colored according to confidence. (B) Tae17 sequence mapped by structural domains. Each domain is indicated in a different color and residue numbers are shown. (C) Tae17 AlphaFold2 structure mapped by domains, colors are as used in panel B. (D) Alignment of the C. jejuni lytic transglycosylase 7LAM (green) and the proposed Tae17 lytic transglycosylase domain (light blue, residues 234–446). Active site resides are indicated by pink boxes: E278 (E390 in 7LAM), H276 (R388 in 7LAM; labeled +), and K436 (K505 for 7LAM). (E) Alignment of the proposed Tae17 amidase domain (dark blue, residues 469–582) to the full-length Ralstonia pickettii T6SS effector Tae3 (orange, PDB ID: 4HZ9). Active site residues are indicated by a pink box: C23 and H81 for Tae3 and the equivalent C488 and H550 for Tae17. For panel E, 4HZ9 chains B and C are removed as these are from the immunity protein Tai3.

Fig 2

Activity of the A. baumannii AB307-0294 T6SS predicted bifunctional effector Tae17. (A) Growth curves were generated for recombinant E. coli strains harboring a gene encoding Tae17^WT, Tae17^E278A, Tae17^C488A, or Tae17^E278A,C488A (each with an added PelB leader sequence) on the vector pBAD30 (Table 2). Expression of the recombinant proteins was either induced with arabinose (solid markers and lines) or repressed with glucose (open markers and dotted lines) and optical density was measured at 0 h, 1 h, 3 h, and 5 h. (B) Viable counts were performed on the same cultures at 0 h, 1 h, 3 h, and 5 h. On both graphs, each data point represents the mean of three biological replicates, and the vertical bars show ±SD. The statistical significance of the difference in means was analyzed by ANOVA with Tukey’s multiple comparisons test. ***P < 0.001, ****P < 0.0001. (C) Degradation of FITC-labeled peptidoglycan following incubation for 2 h with purified lysozyme (control), Tae17^WT, Tae17^C488A, Tae17^E278A, or Tae17^E278A,C488A. For purification of the Tae17 proteins, see Fig. S1. (D) An image of a Tris-Tricine gel following electrophoresis of the following samples: Gram-negative dansylated LII (DAP LII), polymerized DAP LII (Pol. DAP LII), and polymerized DAP LII following incubation for 2 h with purified Tae17^WT, Tae17^C488A, Tae17^E278A, or Tae17^E278A,C488A. (E) An image of a Tris-Tricine gel following electrophoresis of Gram-positive dansylated LII (L-Lysine LII), polymerized L-Lysine LII (Pol. L-Lysine LII), and polymerized L-Lysine LII following incubation for 2 h with purified Tae17^WT, Tae17^C488A, Tae17^E278A, or Tae17^E278A,C488A.

Fig 3

The effect of C-terminal truncations on the ability of VgrG17 to deliver Tae17 and kill E. coli prey. (A) A schematic representation of wild-type and C-terminal truncated VgrG17 proteins. The majority of A. baumannii VgrG sequence (white region) is conserved and predicted to participate in the formation of the stalk formed by the trimerization of VgrG proteins at the T6SS tip. This region is not shown in its entirety. The C-terminal end of the protein (green) is unique to VgrG17 and specific regions within it are predicted to interact with Tae17 for its delivery. (B) The ability of C-terminal truncated VgrG17 proteins VgrG17_1–1,085 and VgrG17_1–1,074 to deliver Tae17 via the T6SS was investigated using an interbacterial killing assay. The A. baumannii predator strain AB307-0294Δtse15Δtde16ΔvgrG17 (strain AL4060) containing empty vector (EV), VgrG17, VgrG17_1–1,085, or VgrG17_1–1,074 was co-incubated for 3 h with E. coli prey strain DH10B harboring an empty pWH1266 vector (AL3340). To check the growth of the prey in the absence of a predator, the E. coli prey was also grown alone (Media Only). Each data point represents the mean of two technical replicates, the horizontal bars show the mean of the biological replicates, and the vertical bars show ±SEM. The dotted line signifies the limit of detection of the assay. The statistical significance of the difference in means was analyzed by ANOVA with Tukey’s multiple comparisons test. n = 7–8. ****P < 0.0001.

Fig 4

VgrG17 amino acids G1069 and W1075 are important for the delivery of the effector Tae17. (A) Schematic representation of VgrG17 with the amino acid sequence in the region of interest (1,051–1,101) shown. Amino acids targeted for dual alanine substitutions are shown in underlined pairs while amino acids targeted by single alanine substitutions (G1069, Y1070, W1075, and H1076) are shown in bold. The white region of VgrG17 indicates the portion predicted to participate in the formation of the stalk formed by the trimerization of VgrG proteins at the T6SS tip, while the region in green is predicted to be the general region available for interaction with the effector Tae17. (B) An Interbacterial killing assay using E. coli DH10B harboring empty pWH1266 vector as prey, co-cultured for 3 h with the A. baumannii predator strain AB307-0294Δtse15Δtde16ΔvgrG17 (AL4060) as predator, provided with empty pBASE (EV) or a pBASE-derivative encoding wild-type VgrG17, or VgrG17 with a single amino substitution as shown. As a prey growth control, E. coli DH10B harboring empty pWH1266 was cultured alone (Media Only) (C) Western immunoblot assessing Hcp secretion (top) or intracellular amounts of Hcp (whole-cell lysates [WCL]; bottom) in the A. baumannii wild-type AB307-0294 (WT), the tssM mutant (negative control for Hcp secretion), or AB307-0294Δtse15Δtde16ΔvgrG17 strain provided with plasmids encoding wild-type VgrG17 or one of the VgrG17 alanine mutants. The secretion of the T6SS needle protein, Hcp, is used as an indicator of a functional T6SS in A. baumannii. (D) Interbacterial killing assay using E. coli DH10B harboring a plasmid encoding immunity proteins Tdi16 and Tsi15 as prey (to neutralize T6SS effectors Tse15 and Tde16). The A. baumannii predator strain in this assay was AB307-0294ΔvgrG17 (AL4329) containing EV, or a pBASE-derivative encoding wild-type VgrG17, or VgrG17 with a single amino substitution as shown. (E) Western immunoblot assessing Hcp secretion (top) or intracellular amounts of Hcp (WCL; bottom) produced by the A. baumannii wild-type AB307-0294 (WT), the tssM mutant (negative control for Hcp secretion), or the AB307-0294ΔvgrG17 predator strain provided with plasmids encoding wild type VgrG17 or one of the VgrG17 alanine mutants. For panels B and D, each data point represents the mean of two technical replicates, the horizontal bars show the mean of the biological replicates, and the vertical bars show ±SEM. The dotted line signifies the limit of detection of the assay. The statistical significance of the difference in means was analyzed by ANOVA with Tukey’s multiple comparisons test. n = 4–12. ***P = 0.0001, ****P < 0.0001. For panels C and E, images were cropped to only include the ~10 kDa section of the gel containing the Hcp protein. Supernatant samples were concentrated 10× (10×sup).

Fig 5

AlphaFold2 model of VgrG17_1010-1085 interacting with Tae17 with predicted hydrogen bonds. While the AlphaFold2 interaction was modeled with the full-length Tae17, only the interacting region is shown here for clarity. (A) AlphaFold2 model of VgrG17_{1,010–1,085} interacting with Tae17_1–156. VgrG17 is shown in green. For Tae17, the Ig-like domain is shown in pink, the LysM domain is shown in purple, and the linker region is colored gray. Edge-to-edge contact backbone hydrogen bonds are shown in red. Residues of interest mutated during the alanine scanning mutagenesis are shown. The amidase and lytic transglycosylase domains are removed for clarity but are included in Fig. S4. (B) The same AlphaFold2 model of VgrG17_{1,010–1,085} interacting with Tae17_1–156 but rotated forward 90 degrees. The VgrG17 N- and C-termini and TTR domains are indicated.