ABSTRACT
Gram-negative bacterial pathogens utilize virulence-associated secretion systems to inject, or translocate, effector proteins into host cells to manipulate cellular processes and promote bacterial replication. However, translocated bacterial products are sensed by nucleotide binding domain and leucine-rich repeat-containing proteins (NLRs), which trigger the formation of a multiprotein complex called the inflammasome, leading to secretion of interleukin-1 (IL-1) family cytokines, pyroptosis, and control of pathogen replication. Pathogenic Yersinia bacteria inject effector proteins termed Yops, as well as pore-forming proteins that comprise the translocon itself, into target cells. The Yersinia translocation regulatory protein YopK promotes bacterial virulence by limiting hyperinjection of the translocon proteins YopD and YopB into cells, thereby limiting cellular detection of Yersinia virulence activity. How hyperinjection of translocon proteins leads to inflammasome activation is currently unknown. We found that translocated YopB and YopD colocalized with the late endosomal/lysosomal protein LAMP1 and that the frequency of YopD and LAMP1 association correlated with the level of caspase-1 activation in individual cells. We also observed colocalization between YopD and Galectin-3, an indicator of endosomal membrane damage. Intriguingly, YopK limited the colocalization of Galectin-3 with YopD, suggesting that YopK limits the induction or sensing of endosomal membrane damage by components of the type III secretion system (T3SS) translocon. Furthermore, guanylate binding proteins (GBPs) encoded on chromosome 3 (GbpChr3), which respond to pathogen-induced damage or alteration of host membranes, were necessary for inflammasome activation in response to hyperinjected YopB/-D. Our findings indicate that lysosomal damage by Yersinia translocon proteins promotes inflammasome activation and implicate GBPs as key regulators of this process.
INTRODUCTION
Innate immune recognition of pathogens plays a vital role in host defense. Surface and endosomal Toll-like receptors (TLRs) recognize conserved structures of pathogens, leading to transcriptional induction of inflammatory gene expression downstream from NF-κB and mitogen-activated protein kinase (MAPK) signaling (1, 2). In addition to membrane-bound pattern recognition receptors (PRRs), cytosolic receptors of the nucleotide binding domain and leucine-rich repeat-containing protein (NLR) family recognize pathogen-specific activities, including translocation of bacterial products via virulence-associated secretion systems and disruption of membranes by bacterial pore-forming proteins (3–10). The activation of a subset of NLRs results in the formation of multiprotein complexes known as inflammasomes (11, 12). The assembly of the inflammasome is necessary for the activation of caspase-1 in response to diverse pathogens and endogenous signals of tissue damage (11, 13). Active caspase-1 leads to the processing of interleukin-1β (IL-1β) and IL-18, the secretion of IL-1α, IL-1β, and IL-18, and pore formation and cell death through cleavage of Gasdermin-D (GSDMD) (14, 15).
In addition to the canonical inflammasome, cytosolic LPS or infection by Gram-negative bacteria triggers a caspase-11-containing noncanonical inflammasome (16–18). This noncanonical inflammasome induces GSDMD-mediated cell death and IL-1α secretion independently of caspase-1 in response to caspase-11-mediated sensing of cytosolic LPS (14, 15, 19–21). Caspase-11-dependent cleavage of GSDMD and subsequent pore formation also triggers activation of the canonical NLRP3 inflammasome by inducing potassium efflux (22–24). Caspase-11 activation can also be triggered by endocytosis of Gram-negative bacterial outer membrane vesicles (25), the activity of type III and type IV secretion systems of Gram-negative pathogens like Legionella and Yersinia (26, 27), or disruption of bacterial cell-containing vacuolar compartments by members of the interferon (IFN)-inducible GTPases known as guanylate binding proteins (GBPs) (28–32).
GBPs encoded on chromosome 3 (GbpChr3) are necessary for maximal noncanonical inflammasome activation in response to vacuolar Gram-negative bacteria, cytosolic lipopolysaccharide (LPS), or entry of Gram-negative bacteria into the cytosol (28, 33, 34). GBPs have been described to function by lysing Salmonella-containing vacuoles and exposing Salmonella LPS to the host cell cytosol, where it is detected by caspase-11 (34). A recent study also revealed that the GBPs recruit the immunity-related GTPase (IRG) protein IRGB10 to the surface of cytosolic bacteria, leading to bacterial lysis and activation of both canonical and noncanonical inflammasomes (35). However, GBPs can also promote the activation of caspase-11 independent of direct vacuolar lysis in the case of Legionella or Chlamydia (28, 33). In support of their lysis-independent function, GBPs were demonstrated to be important for noncanonical inflammasome activation in response to transfection of LPS into the cytosol (28).
Bacterial type III and type IV secretion systems (T3SS and T4SS, respectively) inject proteins into host cells that not only modulate cellular signaling but also trigger immune responses, including inflammasome activation. Pathogenic Yersinia bacteria inject Yop proteins that modulate a number of fundamental cellular processes, which collectively promote Yersinia virulence. Notably, two Yersinia proteins, YopM and YopK, prevent inflammasome activation in Yersinia-infected cells (36–39). YopM interferes with YopE- and YopT-mediated activation of the pyrin inflammasome by binding to protein kinase C-related kinases (PRKs) and promoting PRK-dependent phosphorylation and inactivation of pyrin (37). Interestingly, Yersinia pseudotuberculosis and Yersinia pestis YopK mutants are unable to prevent inflammasome activation despite the presence of YopM, indicating that YopM and YopK block the inflammasome via distinct, nonredundant mechanisms (36, 38, 40).
YopK is translocated into host cells and regulates T3SS translocation by interacting with the proteins YopB and YopD (36, 41). YopB and YopD are both integral transmembrane proteins and form the pore, or translocon complex, through which other Yops are injected into the host cell. Both YopB and YopD are required for functional translocation of the secreted effector Yops (42–44). In addition to being required for translocation of other Yops, YopB and YopD are also injected into host cells (42, 45, 46). Critically, in the absence of YopK, both YopB and YopD are hypertranslocated, resulting in the activation of both canonical and noncanonical inflammasomes (42, 45, 46). Moreover, we previously found that inflammasome activation by the yopK mutant is abrogated by point mutations of the YopB/-D-specific chaperone LcrH that eliminate hypertranslocation of YopB/-D but do not affect the translocation of other Yops (46). These findings indicate that hypertranslocation of YopB/-D, rather than another effector protein or other molecule, is responsible for inflammasome activation and imply that regulating the injection of pore-forming translocon proteins is important for limiting innate responses to the T3SS. Precisely how hypertranslocated YopB and/or YopD mediates inflammasome responses remains unclear, but translocon proteins of other pathogens, including the Salmonella translocon protein SipB and the translocon of the opportunistic pathogen Aeromonas, also trigger inflammasome activation, indicating that inflammasome activation is a conserved response to T3SS translocon proteins (47, 48).
As translocon proteins are amphipathic and membrane active (49–51), they potentially have the capacity to interact with internal membranes of cellular organelles, analogous to the influenza M2 protein, which forms a channel in the membrane of the Golgi apparatus that triggers inflammasome activation by inducing ion flux (5). Here, we found that translocated YopD accumulated at high frequency in Lamp1-positive (Lamp1+) cellular compartments and that Gbp2 showed increased recruitment to YopD-containing compartments in cells infected with Yersinia yopK mutants. We also found that YopD costained with a marker of endosomal membrane damage, Galectin-3 (Gal3), and that this staining was inhibited by YopK. These data suggest that translocon proteins may damage endosomal compartments and that YopK limits this damage. Notably, macrophages lacking the five GBP proteins encoded on chromosome 3 (GbpChr3) were deficient in noncanonical inflammasome activation in response to hyperinjection of the translocon proteins. These findings demonstrate that the GBPs contribute to noncanonical inflammasome activation by Yersinia type III translocon proteins, either by directly damaging endosomal-lysosomal compartments that are marked by bacterial translocon proteins or by responding to translocon-induced damage to lysosomal compartments.
RESULTS
Translocated YopD associates with the late endosome/lysosome marker LAMP1.Hypertranslocation of the pore proteins YopD and YopB induces inflammasome activation in response to Yersinia bacteria that lack the translocation regulator YopK (46). We previously observed that translocated YopD and YopB formed large globular structures in the cell that were distinct from the bacteria themselves (46). This finding raises the possibility that injected translocon proteins were either aggregating or interacting with cellular components. Both protein aggregates and perturbance of cellular organelles have been linked to inflammasome activation (5, 34, 52–55). Interestingly, we found that translocated YopD colocalized with LAMP1 regardless of whether bone marrow-derived macrophages (BMDMs) were infected with either a ΔyopEJ or a ΔyopEJKYersinia pseudotuberculosis strain (Fig. 1A and B). However, consistent with our previous observations (46), ΔyopEJKYersinia-infected cells contained higher overall levels of translocated YopD than ΔyopEJYersinia-infected cells, and ΔyopEJYersinia-infected cells exhibited minimal levels of caspase-1 punctum formation (Fig. S1A in the supplemental material and data not shown). As the total levels of translocated YopD varied between cells infected by the ΔyopEJ or the ΔyopEJK strain, we analyzed the percentages of LAMP1 colocalized with YopD relative to total LAMP1 in the cells, which was constant across the two infections. Interestingly, the percentage of LAMP1 that colocalized with YopD was significantly greater in cells infected with the ΔyopEJK strain than in those infected with the ΔyopEJ strain (Fig. 1C). These data suggested that YopK does not prevent the recruitment of translocated YopD to LAMP1 compartments and that the increased levels of YopD translocated into ΔyopEJKYersinia-infected cells that associate with LAMP1-positive compartments could be important for caspase-1 punctum formation. In support of this hypothesis, ΔyopEJKYersinia-infected cells containing caspase-1 puncta showed a significantly higher frequency of YopD colocalized with LAMP1 (Fig. 1D and E; Fig. S1B and C). We observed that the other translocon protein, YopB, also exhibited some colocalization with LAMP1, although the overall levels of YopB translocation, as well as of LAMP1 colocalization, were lower than for YopD (Fig. S1D to F). Consistent with our prior studies, (46), YopD translocation required the translocon itself, as we did not detect any translocated YopD in YopB-deficient cells (Fig. S1G).
Translocated YopD colocalizes with LAMP1-containing lysosomes. (A) Single z-planes of confocal microscopy images of BMDMs that were pretreated with YVAD and left uninfected or infected with either ΔyopEJK or ΔyopEJYersinia for 2 h. Cells were stained for YopD (green), Yersinia (red), and LAMP1 (blue). White outlines represent cells as determined by CellMask blue staining. (B) Percentages of translocated YopD that colocalized with LAMP1 as the fractions of total YopD in ΔyopEJK and ΔyopEJYersinia-infected cells. (C) Percentages of LAMP1 that colocalized with translocated YopD as the fractions of total LAMP1 in cells. Each data point indicates the result for a single cell. (D) Cells were pretreated with YVAD, infected with ΔyopEJKYersinia as described above, and stained with antibodies against LAMP1 (blue), YopD (green), or caspase-1 (red) to identify caspase-1 punctum-positive cells. Representative staining is shown. (E) Percentages of YopD colocalizing with LAMP1 in caspase-1 punctum-positive or -negative cells were quantified. Graphs show pooled data from three independent experiments, each of which gave similar results. Three hundred seventy-six cells (A to C) and 237 cells (D, E) were analyzed. Error bars represent SEM. Scale bar represents 10 μm. *, P < 0.05; ****, P < 0.0001.
While YopD exhibited some colocalization with other markers of other cellular organelles (GP130, TOM20, or Calnexin), the degree of colocalization with these organelles was lower and fewer cells had high levels (60% or more) of YopD colocalization with these cellular markers (Fig. S2). Overall, these data suggest that YopD translocated by the Yersinia T3SS preferentially colocalizes with LAMP1-containing compartments.
Translocated YopD associates with marker of endosomal damage Gal3.Type III translocon proteins are amphipathic and insert into host membranes, potentially causing membrane damage or perturbance. Indeed, it was recently demonstrated that a marker of endosomal damage, Galectin-3 (Gal3) is recruited to endosomal compartments of bacteria that contain either type III or type IV secretion systems (56). Interestingly, vacuoles containing intracellular Yersinia lacking YopK recruit more Gal3 than vacuoles containing YopK-sufficient Yersinia, suggesting that YopK limits T3SS-induced perturbation of the endosomal vacuole or limits cellular sensing of T3SS-induced vacuolar disruption (56). Given that we previously observed significant levels of translocated YopD within host cells apart from the Yersinia-containing vacuole (46), we examined whether intracellular compartments that contained Yersinia-free YopD also recruit Gal3. Notably, we found that translocated YopD also colocalized with Gal3 (Fig. 2A). Both the overall volume of translocated YopD colocalizing with Gal3 and the percentage of translocated YopD that colocalized with YopD were significantly higher in ΔyopEJKYersinia-infected cells than in cells infected with ΔyopEJYersinia (Fig. 2B and C). Additionally, a greater proportion of total cellular Gal3 was recruited to YopD in the ΔyopEJKYersinia-infected cells than in the ΔyopEJYersinia-infected cells (Fig. 2D), indicating that T3SS translocon components that are translocated into the target cells also damage cellular compartments that are distal to the bacterial-cell-containing vacuole. The Gal3 staining we observed was specific, as it was not detected in Gal3−/− bone marrow-derived macrophages (BMDMs) (Fig. S3A). Notably, while we observed YopD staining in both ΔyopEJK and ΔyopEJYersinia-infected cells, most of the YopD staining in ΔyopEJYersinia-infected cells was associated with the bacteria themselves (Fig. 2A to C). Translocon proteins could potentially be trafficked to endolysosomal compartments from the plasma membrane or could insert into endosomal membranes following translocation into the cytosol of target cells. In order to test whether trafficking through the endocytic pathway was required for translocated YopD/-B to induce inflammasome activation, we treated cells with chloroquine, which disrupts the pH gradient of the endolysosomal network and inhibits endosome-lysosome trafficking (57, 58). As expected, chloroquine treatment reduced the survival of intracellular Salmonella enterica serovar Typhimurium strain SL1344, which requires a low-pH compartment to replicate (Fig. S3B). However, chloroquine-treated cells exhibited equivalent levels of cytotoxicity in response to YopB/-D hypertranslocation, indicating that translocon components do not require trafficking from the plasma membrane or early endocytic compartments in order to induce inflammasome activation (Fig. 2E).
Translocated YopD associates with Galectin-3, a marker of lysosomal damage. (A) Single z-planes of confocal microscopy images of BMDMs that were primed overnight with IFN-γ and pretreated with YVAD 2 h prior to infection with either ΔyopEJK (ΔEJK) or ΔyopEJ (ΔEJ) Yersinia for 2 h. Cells were stained for Yersinia (red), YopD (green), and Galectin-3 (blue). White outlines represent cells as determined by CellMask blue staining. White scale bar represents 10 μm. (B, C) Volumes (B) and percentages (C) of translocated YopD that colocalized with Galectin-3 in ΔyopEJK and ΔyopEJYersinia-infected cells. Each dot represents a single cell. Graphs show combined data from 3 independent experiments. Five hundred seventy-six cells were analyzed. ***, P < 0.001; ****, P < 0.0001. (D) Percentages of total cellular Galectin-3 that colocalized with translocated YopD in ΔyopEJK and ΔyopEJYersinia-infected cells. (E) LDH release was measured 4 h postinfection in cells primed with LPS and treated with 50 μM chloroquine at the time of infection with ΔyopEJK Yersinia. Data are representative of three independent experiments. (B to D) Each point represents a single cell. Error bars represent SEM. Graphs show combined data from three independent experiments. Three hundred seventy-seven cells were analyzed. ****, P < 0.0001.
YopK limits recruitment of Gbp2 to YopD-containing compartments.Galectins mark intracellular membranes that are damaged by cellular stress or infection, which promotes the recruitment of autophagy machinery or IFN-inducible guanylate binding proteins (GBPs) to degrade damaged or pathogen-containing vacuoles (34, 59, 60). Indeed, GBPs function as sensors of endolysosomal membrane integrity (56) and play an important role during both canonical and noncanonical inflammasome activation in response to infection by Gram-negative bacteria. GBPs can contribute to inflammasome activation by binding to and disrupting pathogen-containing vacuoles or bacteria themselves, thus enabling the detection of cytoplasmic LPS (28, 34, 61–63). Whether GBPs might be recruited to vacuoles containing translocated microbial products but not pathogens themselves has not been investigated. Interestingly, we observed that a fraction of translocated YopD colocalized with Gbp2 (Fig. 3A and B) and that the volume of this translocated YopD colocalizing with Gbp2 was increased in BMDMs infected by Yersinia lacking YopK (Fig. 3B). Consistent with our prior studies, ΔyopEJYersinia-infected cells also contained YopD staining, but the vast majority of this YopD colocalized with the Yersinia cells themselves (Fig. 3A, gray arrows). Gbp2 was not detected in cells from GbpChr3−/− BMDMs, indicating the specificity of the anti-Gbp2 antibody (Fig. S4A). We previously demonstrated that Yersinia strains expressing YopD in-frame deletion mutants that are incapable of translocating substrates but can still form pores in membranes are unable to trigger inflammasome activation (46, 51). Interestingly, these strains are also deficient for recruitment of Gbp2, whereas Yersinia strains expressing YopD deletion mutants that retain translocation capacity and induce inflammasome activation recruit significant levels of Gbp2 (Fig. S4B). These data suggest that the presence of translocation-competent YopD plays an important role in the recruitment of GBPs to Yersinia-containing vacuoles and YopD-containing compartments. Nevertheless, the individual loss of Gbp2 or Gbp5 did not affect inflammasome activation in response to YopK-deficient Yersinia (Fig. S4C). The immunity-related GTPase IRGB10, which mediates the recruitment of GBPs to cytosolic bacteria to promote inflammasome activation, was also not required for inflammasome activation in response to YopK-deficient Yersinia (Fig. S4B). These data suggested that while Gbp2, and potentially other Gbps, is recruited to translocon-containing compartments, either another GBP could be responsible for inflammasome activation or redundancy among GBPs could allow additional GBPs to contribute to Yersinia translocon-induced inflammasome activation in the absence of Gbp2.
Gbp2 is recruited to translocated YopD. (A) Single z-planes of confocal microscopy images of BMDMs that were primed overnight with IFN-γ and then infected with ΔyopEJK (ΔEJK) or ΔyopEJ (ΔEJ) Yersinia expressing mCherry or left uninfected for 2 h. Cells were stained for YopD (blue) and Gbp2 (green). White outlines represent cells as determined by CellMask blue staining. White scale bar represents 10 μm. Overlap between translocated YopD and Gbp2 appears as cyan (solid white arrows); overlap between Gbp2 and Yersinia is yellow (white-outlined arrows); overlap between Yersinia and YopD is magenta (gray arrows in ΔyopEJYersinia-infected cells). (B) Quantification of volumes of translocated YopD that colocalized with Gbp2 in ΔyopEJK and ΔyopEJYersinia-infected cells. Each dot represents the result for a single cell. Error bars represent SEM. Graphs show combined data from three independent experiments. Five hundred two cells were analyzed. ***, P < 0.001.
GBPs encoded on chromosome 3 contribute to Yersinia T3SS-induced inflammasome activation.To test whether other GBPs might contribute to inflammasome activation by Yersinia translocon components, we infected BMDMs from GbpChr3−/− mice, which lack all five GBPs encoded on chromosome 3 (64) and are defective for inflammasome activation in response to cytoplasmic LPS, as well as cytosolic bacteria (28). Critically, BMDMs lacking all 5 GBPs encoded on chromosome 3 exhibited decreased cytotoxicity and secretion of both IL-1β and IL-1α in response to YopK-deficient Yersinia, whereas the secretion of an inflammasome-independent cytokine, IL-6, was unaffected (Fig. 4A to D). The requirement for GBPs in inflammasome activation was specific to Yersinia infection, as GbpChr3−/− macrophages exhibited similar levels of cytotoxicity and IL-1 cytokine secretion in response to Salmonella, which activates the canonical NLRC4-dependent inflammasome (Fig. 4A to C). Finally, caspase-1 processing was abrogated in GbpChr3−/− BMDMs following infection with yopK mutant Yersinia but not Salmonella, which demonstrates that GBPs act upstream from caspase-1 processing to promote inflammasome activation in response to the Yersinia T3SS (Fig. 4E).
GBPs encoded on chromosome 3 contribute to Yersinia T3SS-induced inflammasome activation. (A to D) B6 and GbpChr3−/− BMDMs were infected with ΔyopEJK (ΔEJK), ΔyopEJ (ΔEJ), or ΔyopB (ΔB) Yersinia or Salmonella (STm) or treated with LPS. (A) Cytotoxicity was determined by LDH release. Data are representative of three independent experiments. *, P < 0.05. (B to D) Supernatants from B6 and GbpChr3−/− BMDMs infected with indicated bacterial strains or treated with LPS were assayed for levels of secreted IL-1β, IL-1α, and IL-6. Data are representative of three independent experiments. *, P < 0.05; NS, not significant. (E) B6 and GbpChr3−/− BMDMs were left uninfected (UI) or infected with indicated bacterial strains. Cell lysates were assayed for pro- and processed caspase-1 and actin by immunoblotting; data are representative of 3 independent experiments. (F) B6 and GbpChr3−/− BMDMs were primed overnight with IFN-γ and infected with ΔyopEJK or ΔyopJKYersinia or Salmonella (STm) or treated with LPS or LPS plus ATP, and cytotoxicity determined by LDH release. Error bars represent SEM. Data are representative of three independent experiments. **, P < 0.01; NS, not significant.
In addition to the ability of YopK to limit the recruitment of GBPs to the vacuole of internalized Yersinia (56), our data indicate that YopK also limits the recruitment of Gbp2 to YopD-containing compartments that are distinct from the bacteria themselves (Fig. 3). However, as ΔyopEJKYersinia are internalized due to the absence of the antiphagocytic protein YopE (65), an alternative explanation for our observations is that GBPs respond to YopD from internalized bacteria. To test this possibility, we compared the pyroptosis of wild-type (WT) and GbpChr3−/− BMDMs in response to ΔyopEJK and ΔyopJK mutant bacteria. Importantly, ΔyopJK bacteria are not internalized efficiently due to YopE-mediated blockade of phagocytosis (Fig. S5) (66, 67), but they translocate high levels of YopD due to the absence of YopK (Fig. S5). The ΔyopEJK and ΔyopJK strains also induce similar levels of inflammasome activation due to the lack of YopK (Fig. 4F) (36, 46). Importantly, the absence of chromosome 3-encoded GBPs affected the inflammasome responses to infection by both ΔyopEJK and ΔyopJK strains similarly, and as expected, LPS-plus-ATP-induced NLRP3 inflammasome activation was unaffected by the loss of GbpChr3. Thus, GbpChr3-dependent inflammasome responses to the Yersinia T3SS are driven in large part by the responses to hyperinjected YopB/YopD.
Guanylate binding proteins promote noncanonical inflammasome activation by Yersinia translocon proteins.Injection of Yersinia translocon proteins induces canonical NLRP3 and noncanonical caspase-11 inflammasome activation (27, 46). The canonical inflammasome components ASC and NLRP3 are necessary for translocon-induced release of IL-1β but dispensable for translocon-induced pyroptosis, which depends largely on caspase-11 (27, 46). GBPs could thus potentially affect one or both aspects of translocon-induced inflammasome activation, as GBPs have been implicated in the activation of both canonical and noncanonical inflammasomes (31, 32, 59, 68). GbpChr3−/− BMDMs exhibit reduced levels of caspase-11-dependent pyroptosis in response to transfection of LPS or to infection by Gram-negative bacteria that enter the host cytoplasm (28). Gbp5 and -2 also promote canonical inflammasome activation by lysis of cytoplasmic bacteria and activation of an Aim2 inflammasome due to released bacterial DNA or by direct binding of Gbp5 to NLRP3 (63, 69, 70). We therefore sought to test whether GBPs were acting in one or both of these inflammasome pathways in response to hyperinjection of the Yersinia translocon. Notably, both Casp11−/− and GbpChr3−/− BMDMs showed significantly reduced frequencies of caspase-1 puncta in comparison to the numbers in B6 BMDMs after ΔyopEJK strain infection, and no significant difference in the numbers of active caspase-1 puncta was observed between Casp11−/− and GbpChr3−/− BMDMs (Fig. 5A and B). Consistent with these results, direct comparison of pyroptosis and inflammasome-dependent cytokine release revealed no significant difference in these measures of inflammasome activation between Casp11−/− and GbpChr3−/− BMDMs in response to the Yersinia T3SS (Fig. 5C to E). Altogether, these data indicate that the GBPs encoded on chromosome 3 play a key role in caspase-11 noncanonical inflammasome activation by hyperinjection of Yersinia translocon proteins and implicate translocon-induced damage to endolysosomal compartments in this response.
Guanylate binding proteins promote noncanonical inflammasome during Yersinia infection. (A) Collapsed z-stacks of confocal microscopy images of B6, Casp11−/−, and GbpChr3−/− BMDMs that were primed overnight with IFN-γ, pretreated with YVAD, and infected with ΔyopEJKYersinia for 2 h. Cells were stained for YopD (green) and caspase-1 p10 (red). White outlines represent cells as determined by CellMask blue staining. White scale bar represents 10 μm. (B) Percentages of cells that were caspase-1 puncta positive among WT, Casp1/11−/−, and GbpChr3−/− BMDMs. Each dot is representative of a single field. Data are pooled from three independent experiments. (C to E) Supernatants from B6, Casp11−/−, and GbpChr3−/− BMDMs infected with indicated bacterial strains or treated with LPS or LPS plus ATP were assayed for LDH release and levels of secreted IL-1β and IL-1α. Error bars represent SEM. Data are representative of three independent experiments. (F) Supernatants from B6, Casp11−/−, and GbpChr3−/− BMDMs infected with ΔyopJK or ΔyopEJKYersinia were assayed for LDH release. Data are representative of two independent experiments. UI; uninfected; *, P < 0.05; ***, P < 0.0005; ****, P < 0.0001; NS, not significant.
DISCUSSION
Specialized secretion systems and pore-forming toxins of bacterial pathogens play important roles in establishing infection but are also sensed by host innate immune components that trigger antibacterial responses (71). Successful bacterial pathogens have thus evolved strategies to mask these virulence factors from detection or to blunt the host response that is initiated (72–74). Indeed, a number of studies in several different systems have revealed that immune sensing of virulence factor activity promotes pathogen clearance (36, 75–79).
Pathogenic Yersinia species express a T3SS that manipulates numerous aspects of cellular physiology and immune signaling. The activity of the T3SS itself, as well as the activity of translocated effectors, triggers inflammasome activation, which promotes Yersinia clearance in vivo (36, 80). However, Yersinia also inhibits inflammasome activation through the activities of YopK and YopM (36, 38, 39). Recent studies have revealed that YopJ can cooperate with YopM in activated macrophages to limit inflammasome activation (78, 81), although YopJ blockade of NF-κB and MAPK signaling triggers the activation of caspase-8 and caspase-1 in naive macrophages (82, 83).
YopK is a translocated effector that regulates the translocation of other Yersinia Yops (42, 84). YopK deficiency results in significantly attenuated infection in vivo, which was initially surprising given that YopK-deficient Yersinia translocate an increased amount of virulence factors into target cells (36, 76, 85, 86). However, attenuation of yopK mutant Yersinia depends on caspase-1/-11, suggesting that Yersinia regulates translocation to prevent inflammasome sensing of T3SS activity (36, 46). Indeed, we previously found that hypertranslocation of YopD and YopB, which occurs in cells infected by Yersinia yopK mutants, was responsible for triggering inflammasome activation in response to Yersinia infection (46). However, the link between hypertranslocation of YopB and YopD and inflammasome activation has been unclear.
Our studies now demonstrate that a significant fraction of translocated YopD preferentially localizes to LAMP1-positive compartments. In addition, translocated YopD also associated with Galectin-3 (Gal3), a marker of membrane damage, suggesting that YopD-containing compartments are damaged, either by the insertion of YopD itself or as a cellular response to YopD insertion. Chloroquine treatment did not inhibit inflammasome activation in response to Yersinia translocon proteins, suggesting that translocated YopD inserts into LAMP1-containing compartments rather than being inserted into the plasma membrane or early endosomes prior to trafficking to these compartments. Intriguingly, the proportions of YopD colocalization with LAMP1 were positively correlated with the recruitment of caspase-1 to large puncta, which denote assembly of active inflammasomes (87), suggesting that the recruitment of YopD to LAMP1-positive compartments correlates with inflammasome activation.
Our studies also revealed significant colocalization between YopD and guanylate binding protein 2 (Gbp2). While we did not observe a specific requirement for Gbp2 itself in inflammasome activation in response to Yersinia infection, macrophages from GbpChr3−/− mice, which lack Gbp2 as well as four additional GBPs encoded on chromosome 3 (64), showed significantly reduced levels of caspase-1 activation, cytotoxicity, and secretion of IL-1 family cytokines. Thus, one or several GBPs encoded on chromosome 3 likely play an important role in inflammasome activation in response to hypertranslocation of Yersinia translocon components, although the precise role of these GBPs and how they facilitate inflammasome activation remain to be determined.
YopK-deficient Yersinia induces both a canonical NLRP3 inflammasome and a noncanonical caspase-11 inflammasome (27). Interestingly, the noncanonical inflammasome appears to be the primary pathway activated by the Yersinia translocon. Indeed, neither NLRP3 nor ASC is required for pyroptosis in response to YopK-deficient Yersinia, although they are both necessary for the cleavage and release of IL-1β (36). Our findings that the levels of inflammasome activation are similar in Casp11−/− and GbpChr3−/− cells suggest a model whereby trafficking of YopD and YopB to late endosomal/lysosomal membranes results in damage to this compartment, leading to activation of the GBP/caspase-11 pathway and subsequent GSDMD induction of plasma membrane pores, potassium efflux, and NLRP3 inflammasome activation (Fig. 6).
Model for the role of GBPs and caspase-11 in YopB/-D hypertranslocation-induced inflammasome activation. YopK regulates injection of translocon proteins YopB (not shown) and YopD (green ovals) into target cells. Translocated YopD (and YopB, not shown) localize to LAMP1+ compartments following T3SS-mediated injection. YopD-containing compartments also colocalize with Galectin-3 (Gal-3), and YopK limits Gal-3 recruitment to translocated YopD-containing compartments. Either recruitment of GBPs or recruitment of translocon proteins promotes damage to these compartments (dashed lines), leading to caspase-11 activation. Caspase-11 activation promotes Gasdermin-D (GSDMD) cleavage, leading to potassium efflux and canonical NLRP3 inflammasome activation. GBPs could therefore contribute to noncanonical inflammasome activation either upstream or downstream from endosomal compartment damage (indicated by question marks).
While eliminating all five of the GBPs encoded on chromosome 3 significantly reduces inflammasome activation in response to Yersinia infection (Fig. 4), it is currently unclear which specific GBP or GBPs are necessary for inflammasome activation and at which step of translocon-induced inflammasome activation they participate. We cannot currently exclude the possibility that GBPs are working in combination or that multiple GBPs encoded on chromosome 3 play different roles in inflammasome activation by YopK-deficient Yersinia.
The release of a bacterial product from Yersinia-containing lysosomes could potentially be responsible for activating the noncanonical inflammasome. LPS from Yersinia pestis grown at 37°C does not activate caspase-11 (19), but whether Y. pseudotuberculosis LPS can activate the noncanonical inflammasome is not known. LPS from Y. pseudotuberculosis grown at 37°C is also a poor stimulator of Toll-like receptor 4 (TLR4), although it is slightly more stimulatory than LPS of Y. pestis (88). Indeed, a portion of Y. pseudotuberculosis LPS remains hexa-acylated at 37°C, leaving open the possibility that YopB and YopD could facilitate cytosolic delivery of Yersinia LPS, either by directly damaging lysosomes or by promoting the recruitment of GBPs that damage lysosomes, resulting in increased access of LPS to the cytoplasmic compartment. Further studies are needed to determine whether Yersinia LPS or another bacterial product is necessary for noncanonical inflammasome activation by Yersinia. Overall, our studies provide new insight into inflammasome activation by T3SS-containing pathogens and implicate GBPs as novel regulators of inflammasome activation by T3SS translocon proteins.
MATERIALS AND METHODS
Bacterial strains and growth conditions.The Yersinia pseudotuberculosis strains used here were all on a strain IP2666 background and have been previously described (46). Infection of cultured cells was performed as previously described (36, 46). Briefly, Yersinia pseudotuberculosis cultures were grown overnight with aeration in 2× yeast extract-tryptone (YT) broth at 26°C. The bacteria were diluted into fresh 2× YT broth containing 20 mM sodium oxalate and 20 mM MgCl2. Bacteria were grown with aeration for 1 h at 26°C followed by 2 h at 37°C prior to infection. Wild-type Salmonella enterica serovar Typhimurium strain SL1344 was grown overnight with aeration in LB broth at 37°C. The bacteria were diluted into fresh LB broth containing 300 mM NaCl. Bacteria were grown standing to induce Salmonella pathogenicity island 1 (SPI-1) expression.
Cell culture conditions.Bone marrow (BM) cells were grown for 7 to 8 days in Dulbecco modified Eagle medium (DMEM) containing HEPES, sodium pyruvate, 10% fetal bovine serum (FBS), and 30% L929 supernatant in a humidified incubator. Differentiated bone marrow-derived macrophages (BMDMs) were replated overnight into 24-, 48-, or 96-well dishes. In the case of gamma interferon (IFN-γ) treatment, BMDMs were given 100 units/ml overnight. GbpChr3−/− mice have been described previously (64). Casp11−/− mice were initially made by Junying Yuan (Harvard University) and kindly provided to us by Tiffany Horng (Harvard University) (89).
Cell death assay.Cytotoxicity was assayed using a lactate dehydrogenase (LDH) release assay kit (Clontech). In brief, BMDMs were infected at a multiplicity of infection (MOI) of 10 for 4 h. In some cases, cells were primed for 3 to 4 h with 500 ng/ml of LPS. An amount of 2.5 mM ATP was added to primed cells as a control for NLRP3 inflammasome activation. After 1 h, gentamicin was added. Four hours postinfection, the plate was spun down and assayed for LDH release.
Cytokine production.Cells were primed for 3 to 4 h with 500 ng/ml of LPS prior to infection with the bacterial strains indicated in the text and figure legends. Supernatants and recombinant cytokine standards (R&D) were added to MaxiZorp enzyme-linked immunosorbent assay (ELISA) plates that had been precoated with anti-IL-1α, anti-IL-1β, or anti-IL-6 capture antibodies (eBioscience). Cytokines were detected with corresponding biotinylated detection antibodies (eBioscience) and streptavidin-horseradish peroxidase (HRP). Plates were developed with citrate buffer and o-phenylenediamine dihydrochloride (OPD; Sigma-Aldrich) and read at 490 nm.
Western blotting.Western blots were performed as described previously (36). In brief, cell lysates were run on a 4%-to-12% gradient gel and transferred to a polyvinylidene difluoride (PVDF) membrane. Membranes were blotted with the primary antibodies rat anti-mouse caspase-1 p20 antibody (a gift of Vishva Dixit, Genentech) and mouse anti-actin antibody (Sigma). The secondary antibodies were goat anti-rat-HRP (Jackson ImmunoResearch) and horse anti-mouse-HRP (Cell Signaling Technology) antibodies. The blots were developed with Pierce enhanced chemiluminescence (ECL) Western blotting substrate (Fisher Scientific).
Immunofluorescence staining.BMDMs were plated at a density of 1.5 × 105 to 2 × 105 per well on glass coverslips in 24-well plates overnight. For Gbp2 staining, BMDMs were also treated with 100 units/ml IFN-γ overnight. Two hours prior to infection, cells were treated with Z-YVAD-FMK (Z-Tyr-Val-Ala-Asp[OMe]-fluoromethylketone; SM Biochemicals LLC) at a final concentration of 25 μM in order to prevent pyroptosis. BMDMs were then infected at an MOI of 5 for 2 h. mCherry-expressing Yersinia cells were used in experiments that utilized anti-Gbp2 antibody. Cells were fixed in 4% paraformaldehyde prepared from 16% paraformaldehyde (Electron Microscopy Sciences) for 20 min at room temperature and then washed twice in phosphate-buffered saline (PBS). Cells were permeabilized for 10 min at room temperature in 0.2% Triton X-100 (Sigma) or 0.2% saponin if using purified LAMP1 antibody and washed once in PBS. In instances where saponin was used for permeabilization, subsequent steps contained 0.2% saponin. Cells were blocked for 1 h at room temperature or overnight at 4°C in 10% bovine serum albumin (BSA)–PBS. Coverslips were then inverted onto primary antibody cocktails on parafilm and incubated for 1 h at 37°C. Mouse monoclonal antibodies against YopD and rabbit polyclonal antibody against total Yersinia have been described previously (38, 90, 91). Rabbit anti-caspase-1 p10 antibody was from Santa Cruz. Rat anti-LAMP1 antibody (Developmental Studies Hybridoma Bank at the University of Iowa) or purified rat anti-mouse CD107a (LAMP1) antibody (BD Pharmigen) was used to detect LAMP1. Rat monoclonal anti-human/mouse Galectin-3 antibody clone eBioM3/38 (eBioscience) was used to detect Galectin-3. Staining with rabbit anti-Gbp2 antibody (61) was performed overnight at 4°C prior to the other staining procedures. Coverslips were washed and stained with appropriate combinations of goat anti-mouse Alexa Fluor 488, goat anti-rabbit Alexa Fluor 568, goat anti-rabbit Alexa Fluor 488, goat anti-mouse Alexa Fluor 647 (ThermoFisher Scientific), and goat anti-rat Alexa Fluor 647 (Jackson ImmunoResearch) antibodies. Coverslips were washed 4 times. Coverslips were then stained with either a cocktail of Hoechst (ThermoFisher Scientific) and phalloidin Alexa Fluor 647 (ThermoFisher Scientific) for 1 h at 37°C or HCS CellMask blue stain (ThermoFisher Scientific) for 1 h at room temperature. Coverslips were washed 4 times and mounted on glass slides with Fluoromount-G (Southern Biotech). Cells were imaged with a Leica SP5 inverted confocal microscope and analyzed using Volocity 6.3.
All studies whose results are shown in the main figures were performed three times independently. Fig. S1 in the supplemental material shows the analysis of data from individual microscopy experiments. The studies whose results are shown in Fig. S2 and S4 were performed two times independently.
Volocity image analysis.Analysis of microscopy images using Volocity 6.3 was performed on three-dimensional (3-D) reconstructions of the z-stacks. Single cells were identified with the find object function based on CellMask blue staining. The images were then analyzed for staining of Yersinia cells, YopD, LAMP1, Gbp2, caspase-1, and Galectin-3 using the find objects function to generate populations for each stain. Thresholds for positive and negative signals were based on no primary antibody, uninfected, and Gbp2 knockout control samples. In the case of Galectin-3, the threshold was set to differentiate all Galectin-3 from distinct Galectin-3 foci. To determine the amount of translocated YopD, the pixels in the regions defined by YopD and/or bacterial staining were separated by whether pixels in that specific region were positive or negative for each stain. Regions determined to be positive for both YopD and Yersinia staining were subtracted from the total YopD staining using the subtract function. The remaining pixels were assigned to the population “translocated YopD.” To determine the amount of colocalization between YopD and Gbp2, Galectin-3, or LAMP1 or the colocalization between Gbp2 and caspase-1, the regions defined by translocated YopD, Gbp2, LAMP1, Galectin-3, or caspase-1 punctum staining were separated into whether pixels in that specific region were positive or negative for each stain of interest. Regions that contained both stains of interest were assigned to the new population using the intersect function. The different populations were then compartmentalized into the cells. A measurement table was made for each of the compartmentalized populations. Using the analysis feature, the volume of each population per cell was calculated. The percentage of translocated or colocalized YopD was determined by dividing the volume of translocated or colocalized YopD in a cell by the volume of total or translocated YopD in that cell.
Yop translocation assay.HeLa cells were plated in a clear-bottom, black, 96-well tissue culture plate (Greiner Bio One) overnight. Cells were infected at an MOI of 5 with Yersinia strains expressing beta-lactamase fused to YopM, YopD, or glutathione S-transferase (GST) for 2 h, and gentamicin was then added. CCF4-AM was loaded into cells using the LiveBLAzer-FRET B/G loading kit (Life Technologies) for 1.5 to 2 h. The fluorescence was read using a BioTek Synergy plate reader. The response ratio was then calculated using the following formula: (blue/green ratio)/(average negative blue/green ratio).
Statistical analysis.Statistical analysis was performed using Prism 6 (GraphPad Software, Inc.). Unpaired Student's t tests were used for the analysis, with Welch's correction in cases of unequal standard deviations. Error bars in all figures represent the standard errors of the means (SEM). ELISA data and LDH data are representative of three individual experiments in which similar results were observed. Data points in panels quantifying immunofluorescence image data represent individual cells or visual fields as indicated.
ACKNOWLEDGMENTS
We thank members of the Brodsky, Shin, and Third Floor Hill laboratories for helpful scientific discussion and Sunny Shin for critical reading of the manuscript. Casp11−/− bone marrow was provided by Junying Yuan (Harvard University). We thank Maya Ivanov for generating the YopD monoclonal antibody. We thank Gordon Ruthel and the Penn Vet Imaging Core (supported by NIH grant number S10RR027128, the School of Veterinary Medicine, the University of Pennsylvania, and the Commonwealth of Pennsylvania) for technical assistance with microscopy.
This work was supported by NIH grant number R01AI103062 and a BWF PATH award (I.E.B.), Microbial Pathogenesis and Genomics Training grant number T32 AI060516 (E.E.Z.), NIH grant numbers R01AI099222 (J.B.B.) and R01AI103197, and another BWF PATH award (J.C.).
FOOTNOTES
- Received 14 September 2016.
- Returned for modification 12 October 2016.
- Accepted 31 July 2017.
- Accepted manuscript posted online 7 August 2017.
Supplemental material for this article may be found at https://doi.org/10.1128/IAI.00778-16 .
- Copyright © 2017 American Society for Microbiology.
REFERENCES
- 1.↵
- 2.↵
- 3.↵
- 4.↵
- 5.↵
- 6.↵
- 7.↵
- 8.↵
- 9.↵
- 10.↵
- 11.↵
- 12.↵
- 13.↵
- 14.↵
- 15.↵
- 16.↵
- 17.↵
- 18.↵
- 19.↵
- 20.↵
- 21.↵
- 22.↵
- 23.↵
- 24.↵
- 25.↵
- 26.↵
- 27.↵
- 28.↵
- 29.↵
- 30.↵
- 31.↵
- 32.↵
- 33.↵
- 34.↵
- 35.↵
- 36.↵
- 37.↵
- 38.↵
- 39.↵
- 40.↵
- 41.↵
- 42.↵
- 43.↵
- 44.↵
- 45.↵
- 46.↵
- 47.↵
- 48.↵
- 49.↵
- 50.↵
- 51.↵
- 52.↵
- 53.↵
- 54.↵
- 55.↵
- 56.↵
- 57.↵
- 58.↵
- 59.↵
- 60.↵
- 61.↵
- 62.↵
- 63.↵
- 64.↵
- 65.↵
- 66.↵
- 67.↵
- 68.↵
- 69.↵
- 70.↵
- 71.↵
- 72.↵
- 73.↵
- 74.↵
- 75.↵
- 76.↵
- 77.↵
- 78.↵
- 79.↵
- 80.↵
- 81.↵
- 82.↵
- 83.↵
- 84.↵
- 85.↵
- 86.↵
- 87.↵
- 88.↵
- 89.↵
- 90.↵
- 91.↵
