Structure-function analyses of the bacterial zinc metalloprotease effector protein GtgA uncover key residues required for deactivating NF-κB

Abstract

The closely related type III secretion system zinc metalloprotease effector proteins GtgA, GogA, and PipA are translocated into host cells during Salmonella infection. They then cleave nuclear factor κ-light-chain-enhancer of activated B cells (NF-κB) transcription factor subunits, dampening activation of the NF-κB signaling pathway and thereby suppressing host immune responses. We demonstrate here that GtgA, GogA, and PipA cleave a subset of NF-κB subunits, including p65, RelB, and cRel but not NF-κB1 and NF-κB2, whereas the functionally similar type III secretion system effector NleC of enteropathogenic and enterohemorrhagic Escherichia coli cleaved all five NF-κB subunits. Mutational analysis of NF-κB subunits revealed that a single nonconserved residue in NF-κB1 and NF-κB2 that corresponds to the P1' residue Arg-41 in p65 prevents cleavage of these subunits by GtgA, GogA, and PipA, explaining the observed substrate specificity of these enzymes. Crystal structures of GtgA in its apo-form and in complex with the p65 N-terminal domain explained the importance of the P1' residue. Furthermore, the pattern of interactions suggested that GtgA recognizes NF-κB subunits by mimicking the shape and negative charge of the DNA phosphate backbone. Moreover, structure-based mutational analysis of GtgA uncovered amino acids that are required for the interaction of GtgA with p65, as well as those that are required for full activity of GtgA in suppressing NF-κB activation. This study therefore provides detailed and critical insight into the mechanism of substrate recognition by this family of proteins important for bacterial virulence.

Keywords: GtgA; NF-κB; Salmonella enterica; bacterial effectors; bacterial pathogenesis; metalloprotease; substrate specificity; type III secretion system (T3SS); virulence factor.

Figure 1.

Substrate specificity of GtgA, GogA, PipA, and NleC. A, immunoblot analysis of 293ET cells cotransfected with plasmids encoding the indicated GFP-tagged effector proteins and FLAG-tagged NF-κB subunits. The abundance of endogenous p65 and p50 was analyzed using anti-p65 and anti-p50 antibodies, whereas FLAG-tagged RelB, cRel, and p100 were detected using an anti-FLAG antibody. An anti-tubulin antibody was used as a loading control. Immunoblots shown are representative of three independent experiments. B, 5 μm His₆–SUMO–p65(20–188) was incubated with 0.1 μm of the indicated His₆–GST-tagged effector protein for 5 h at 37 °C. The reaction was quenched by the addition of 2× Laemmli buffer, and proteins were separated and visualized by SDS-PAGE followed by Coomassie Blue staining. The arrow indicates the larger cleavage product generated by NleC relative to GtgA, GogA, and PipA. Immunoblot analysis using an anti-GST antibody was done to confirm equal amounts of each GST-tagged effector protein. The Coomassie Blue-stained polyacrylamide gel and immunoblot are representative of three independent experiments. C, quantification of His₆–SUMO–p65(20–188) cleavage in B. Data are presented as percent cleavage relative to control sample (−) and represent the mean ± S.E. of three independent experiments, for which individual data points are indicated.

Figure 2.

P1′ site is a critical determinant of GtgA, GogA, and PipA substrate specificity. A, sequence alignment of residues in close proximity to the GtgA, GogA, PipA, and NleC cleavage site in p65 and the corresponding residues in cRel, RelB, p105/p50, and p100/p52. Uniprot numbers are Q04206, Q04864, Q01201, P19838, and Q00653. The cleavage sites targeted by each enzyme is highlighted. B–E, immunoblot analysis of 293ET cells transiently transfected with plasmids encoding the indicated GFP-tagged effector and FLAG-tagged NF-κB subunit. Immunoblotting of whole-cell lysates was done with an anti-GFP and anti-FLAG antibody. Lysates were also blotted with an anti-tubulin antibody as a loading control. Immunoblots are representative of at least three independent experiments.

Figure 3.

Crystal structures of GtgA alone and in complex with the p65 N-terminal domain. A, crystal structure of Zn²⁺-free GtgA(20–228)^E183Q in complex with p65(20–188). The α-helices and β-strands in GtgA are labeled A to H and 1 to 4, respectively. In the p65 NTD, the α-helices are labeled A′ and B′, and the β-strands are labeled 1′ to 9′. The p65 cleavage site residues Gly-40/Arg-41 are colored yellow and shown in a stick representation. p65 residues Pro-47 to Pro-59 are colored orange. The chloride ions are shown as gray spheres. B, surface and cartoon representation of GtgA in complex with the p65 NTD. GtgA (left) and p65 (right) are colored according to the electrostatic surface potential (positive blue, negative red), as calculated using Adaptive Poisson-Boltzmann Solver (APBS) in PyMOL (44). C, crystal structure of Zn²⁺-bound GtgA(20–228)^E183Q. The α-helices and β-strands in GtgA are labeled A to H and 1 to 4, respectively. The Zincin-like catalytic core is colored as in Ref. ; α-helices are colored teal, and the β1β2 β-sheet and the active-site upper rim residues are colored purple. The zinc ion and chloride ion are shown as pink and gray spheres, respectively. A close-up view of the catalytic zinc ion and active-site residues in a stick representation are shown as an inset. Zinc-coordinating residues are colored orange, Tyr-224 green, and Gln-183 yellow. D, surface representation of GtgA(20–228)^E183Q colored according to the active-site subdomains. With the exception of the right wall, which is colored in green, the N-terminal subdomain is colored in different shades of blue, and the C-terminal subdomain is shown in orange. E, solvent-accessible surface representation colored according to the electrostatic surface potential (positive blue, negative red) of the structure of GtgA(20–228)^E183Q.

Figure 4.

GtgA and p65 N-terminal domain interaction. A, close-up view of the interactions between the GtgA active-site cleft and the p65 NTD. GtgA is colored purple, and p65 is colored green with the exception of the p65 cleavage site residues (Gly-40/Arg-41), which are colored yellow. B, 2F_o − F_c electron density map contoured at 1σ of the GtgA zinc-coordinating residues (His-182, His-186, and Asp-193) and Tyr-224, as well as p65 residues Glu-39 and Gly-40. C, wall-eye stereo view of the p65 P1′ residue Arg-41 (colored yellow) inserted into the GtgA S1′ pocket. A and C, black dashed lines represent hydrogen bonds.

Figure 5.

Mutational analysis of GtgA–p65 interacting residues. A and B, LUMIER-binding assay. His₆–GST-tagged GtgA variants were incubated with the post-nuclear supernatant of 293ET cells expressing p65 fused via its C terminus to the Renilla luciferase. Following elution, immunoblot analysis was performed using an anti-GST antibody (A), and Renilla luciferase activity was measured to calculate the relative fold binding (B). Immunoblot is representative of three independent experiments. Data are presented as the fold change in Renilla luciferase activity relative to His₆–GST–GtgA^E183A and represents the mean ± S.E. of three independent experiments, for which individual data points are indicated. Statistical significance was computed between GtgA^E183A and each GtgA^E183A variant (*, p < 0.05; **, p < 0.01, ordinary one-way ANOVA with post hoc Dunnett's multiple comparisons test). C, 5 μm His₆–SUMO–p65(20–291) was incubated with 0.1 μm of the indicated His₆–GST–GtgA variant for 5 h at 37 °C. The reaction was then quenched by the addition of 2× Laemmli buffer, and proteins were separated and visualized by SDS-PAGE followed by Coomassie Blue staining. Immunoblot analysis using an anti-GST antibody was done to confirm equal amounts of each GST-tagged effector protein. The Coomassie Blue-stained polyacrylamide gel and immunoblots are representative of three independent experiments. D, quantification of His₆–SUMO–p65(20–291) cleavage in C. Data are presented as percentage cleavage relative to control sample and represent the mean ± S.E. of three independent experiments, for which individual data points are indicated. Statistical significances were computed between WT and each GtgA variant (*, p < 0.05; **, p < 0.01, ordinary one-way ANOVA with post hoc Dunnett's multiple comparisons test).

Figure 6.

NF-κB inhibition by GtgA variants. A, immunoblot analysis of 293ET cells co-transfected with plasmids encoding an NF-κB–dependent firefly luciferase, a constitutively expressed Renilla luciferase, and GFP or the indicated GFP–GtgA variant. 293ET cells were stimulated with 20 ng/ml TNFα for 8 h. B, confocal microscopy images of HeLa cells transiently transfected with plasmids encoding GFP or the indicated GFP–GtgA variant. Scale bar, 10 μm. C, luciferase activity was measured in cell lysates from the experiment shown in A. Data are presented as the fold change in NF-κB reporter activity between unstimulated and TNFα-stimulated 293ET cells and represent the mean ± S.E. of three independent experiments, for which individual data points are indicated. Statistical significances were calculated between WT and each GtgA variant (*, p < 0.05; **, p < 0.01, ordinary one-way ANOVA with post hoc Dunnett's multiple comparisons test).

Figure 7.

Structural comparison of GtgA and NleC. Topological and cartoon representations of GtgA (A) and NleC (B) (PDB 4Q3J). The Zincin-like catalytic core is colored as in Fig. 3C; α-helices are colored teal, and the β1β2 β-sheet and the active-site upper rim residues are colored purple. Variable features are colored in orange, blue, red, and green, and the catalytic zinc ions are shown as pink spheres. Dashed lines represent either disordered residues or regions outside the crystallized constructs. In the topological diagrams, zinc-coordinating residues are shown as white circles, and the active-site zinc is shown as a pink circle. GtgA residue Tyr-224 is shown as a blue circle, and GtgA S1′ pocket residues Asp-155, Glu-163, and Ser-179 are shown as yellow circles.

Figure 8.

DNA mimicry by GtgA. A, surface representation of GtgA colored according to its electrostatic surface potential (positive blue, negative red), in complex with the NTD of p65. B, cartoon representation of the p65 RHR in complex with DNA (PDB 2RAM) (11). In both panels, the p65 cleavage site residues (Gly-40/Arg-41) are represented as yellow sticks.