Extended Data Figure 9: Point mutations in the LRR domain of IpaH9.8 that affect binding of GBPs mitigate IpaH9.8 disruption of GBP function in vivo.
From: Ubiquitination and degradation of GBPs by a Shigella effector to suppress host defence
a, b, HeLa cells stably expressing GFP–hGBP1 were infected with indicated S. flexneri strains. The ΔipaH9.8 strain was complemented with wild-type IpaH9.8 or the L50A, R62A, or N83A single mutant (a), or the M3 triple mutant (a, b). Cell lysates collected at indicated time points after the infection were immunoblotted with GFP and tubulin antibodies (a). Fluorescence images were taken 2 h after infection (b). c, Assay of secretion of IpaH9.8-M3 in host cells. HeLa cells were infected with S. flexneri ΔipaH9.8 expressing β-lactamase (TEM1) alone or TEM1-fused IpaH9.8 (wild-type or M3 mutant). Infected cells were loaded with the CCF2–AM dye; the 460-nm fluorescence emission of the dye indicates translocation of TEM1 from the bacteria into HeLa cell cytoplasm. d, Ubiquitin-ligase activity of IpaH9.8-M3 in synthesizing free ubiquitin chains. Immunoblot of the substrate-free ubiquitin reaction with ubiquitin antibody. e, In vitro ubiquitination of mGBP2 by IpaH9.8 or the indicated IpaH9.8 mutant proteins. Flag–mGBP2 was added to the reaction mixtures, which were then immunoprecipitated with Flag antibody followed by immunoblotting with ubiquitin antibody. Coomassie blue-stained SDS–PAGE gels show the recombinant proteins added into the reaction. f, Recruitment of endogenous mGBP2 onto intracellular S. flexneri. BMDMs from Casp1−/−Casp11−/− or Gsdmd−/−Gbpchr3 mice were infected with S. flexneri WT, ΔipaH9.8 or ΔipaH9.8 expressing IpaH9.8-M3. Infected cells were fixed 2 h after infection (a time point when bacterial morphology change had not occurred) and stained with DAPI and mGBP2 antibody. Representative fluorescence images are shown (scale bar, 3 μm). Note that while the mGBP2 antibody may cross-react with mGBP1 (Extended Data Fig. 2d), we believe that the anti-mGBP2 immunofluorescence detected on the Shigella surface here should mainly reflect mGBP2 for three reasons. First, GFP–mGBP1 could not translocate to the bacterial surface in Shigella-infected HeLa cells (Extended Data Fig. 7b). Second, mGBP1 is not a degradation target of IpaH9.8 (Extended Data Fig. 6b, 6d and 8b–e), but appearance of the mGBP2 immunofluorescence in this assay was fully responsive to the presence of IpaH9.8 in S. flexneri. Third, as shown here, the mGBP2 immunofluorescence signal was absent in Gbpchr3 BMDMs infected with S. flexneri ΔipaH9.8. However, we cannot completely rule out the possibility that the mGBP2 fluorescence signal detected on the Shigella surface contains a small fraction of mGBP1 if mGBP1 can dimerize with mGBP2 and therefore be recruited to the bacterial surface. g, h, BMDMs from Casp1−/−Casp11−/− mice were infected with wild-type S. flexneri or the indicated ipaH9.8 deletion or complementation strain for 4 h. Representative images of cytoplasmic S. flexneri indicated by anti-Shigella LPS staining (g). Nuclei are stained with DAPI. Scale bar, 3 μm (upper) and 1.5 μm (lower). h, Quantification of cytoplasmic bacteria with aberrant morphology. Approximately 100 infected cells were examined for each experiment and data are represented as mean percentages ± s.d. from three replicates. Two-tailed unpaired Student’s t-test was performed (**P ≤ 0.0001). All data shown are representative of two (b, c, f–h) or three (a, d, e) independent experiments.
