|Home | About | Journals | Submit | Contact Us | Français|
Rickettsia conorii, an obligate intracellular tick-borne pathogen and the causative agent of Mediterranean spotted fever, binds to and invades non-phagocytic mammalian cells. Previous work identified Ku70 as a mammalian receptor involved in the invasion process and identified the rickettsial autotransporter protein, rOmpB, as a ligand; however, little is known about the role of Ku70-rOmpB interactions in the bacterial invasion process. Using an E. coli heterologous expression system, we show here that rOmpB mediates attachment to mammalian cells and entry in a Ku70-dependent process. A purified recombinant peptide corresponding to the rOmpB passenger domain interacts with Ku70 and serves as a competitive inhibitor of adherence. We observe that rOmpB-mediated infection culminates in actin recruitment at the bacterial foci, and that this entry process relies in part on actin polymerization likely imparted through protein tyrosine kinase and PI3-kinase-dependent activities and microtubule stability. Small-interfering RNA (siRNA) studies targeting components of the endocytic pathway reveal that entry by rOmpB is dependent on c-Cbl, clathrin and caveolin-2. Together, these results illustrate that rOmpB is sufficient to mediate Ku70-dependent invasion of mammalian cells and that clathrin- and caveolin-dependent endocytic events likely contribute to the internalization process.
Rickettsiae are Gram-negative obligate intracellular pathogens transmitted to humans via arthropod vectors (Hackstadt, 1996). They are divided into two groups, the typhus group (TG) and the spotted fever group (SFG), based on differences in the diseases that they cause and the presence of the outer membrane proteins, rOmpA and rOmpB (Vishwanath, 1991). Members of both groups are historically responsible for severe human diseases (Hackstadt, 1996) and are Category B and C Select Agents as defined by the National Institute of Allergy and Infectious Diseases (NIAID). R. conorii, the causative agent of Mediterranean spotted fever, is transmitted into the vasculature of the host by tick-bite inoculation (Hackstadt, 1996). Subsequent replication in endothelial cells may lead to localized dermal and epidermal necrosis called an eschar or tache noire (Walker et al., 1988). Further damage to the vascular endothelium and infiltration of perivascular mononuclear cells often causes increased fluid leakage into the interstitial space, ultimately resulting in a characteristic dermal rash and acute renal failure with peripheral edema (Walker et al., 1988; Hand et al., 1970). Damage to target endothelial cells especially in the lungs and brain may result in the most severe manifestations of disease including pulmonary edema and interstitial pneumonia. Although endothelial cells are the main target cell type for SFG rickettsia, R. conorii can attach to and invade different cell types in vitro and in vivo and spread via lymphatic vessels to lymph nodes or the bloodstream to various tissues including the lungs, spleen, liver, kidneys and heart (Walker and Gear, 1985). Initial clinical symptoms include those of a flu-like syndrome, often leading to misdiagnosis and inappropriate treatment. Although infections are controlled by broad-spectrum antibiotic therapy, untreated or misdiagnosed Mediterranean spotted fever is associated with severe morbidity and mortality (Yagupsky and Wolach, 1993).
Adherence to and subsequent invasion of target cells is critical for the establishment of a successful rickettsial infection. Electron micrographs of rickettsia-host cell interactions (Gouin et al., 1999; Teysseire et al., 1995) suggest that entry of R. conorii morphologically and mechanistically resembles a zipper-like invasion strategy, in which the invasion of non-phagocytic mammalian cells is mediated by the interactions between specific bacterial ligands and host receptors, leading to localized actin recruitment around the bacterium (reviewed in (Cossart and Sansonetti, 2004)). Previous work confirmed that host actin polymerization plays a crucial role in R. conorii entry and that actin dynamics during R. conorii entry are in part governed by the actin nucleating protein complex, Arp2/3. Various approaches used to disrupt signaling pathways that directly or indirectly activate the Arp2/3 complex revealed that R. conorii utilizes pathways involving Cdc42, PI 3-kinase, c-Src and other protein tyrosine kinase (PTK) activities to enter non-phagocytic cells (Martinez and Cossart, 2004).
A recent bioinformatics analysis of sequenced rickettsial genomes identified a family of genes termed sca (surface cell antigens) encoding putative outer membrane proteins (Blanc et al., 2005). Five genes in this family, namely sca0 (ompA), sca1, sca2, sca4, and sca5 (ompB) are highly conserved among the majority of SFG rickettsiae (Blanc et al., 2005; Roux and Raoult, 2000). Interestingly, the predicted proteins encoded by ompA, sca1, sca2, and ompB share homology with a family of proteins in gram-negative bacteria called autotransporters, many of which are known virulence factors (Jacob-Dubuisson et al., 2004; Henderson and Nataro, 2001). These proteins have modular structures composed of an N-terminal signal peptide, followed by a “passenger domain” that carries out the specific function(s) of the protein, and then a C-terminal “translocation module” that serves as a putative β-barrel rich pore for the secretion of the N-terminal passenger domain across the outer membrane (Jacob-Dubuisson et al., 2004).
Two Sca proteins, rOmpA and rOmpB, are expressed on the surface of nearly all SFG rickettsiae (Hackstadt et al., 1992) and have been shown to be important for attachment to mammalian cells in vitro (Uchiyama et al., 2006; Li and Walker, 1998) and for eliciting protective humoral immune responses in vivo (Feng et al., 2004b; Feng et al., 2004a; Diaz-Montero et al., 2001; Li and Walker, 1998). rOmpB, the more abundant of the two, is expressed as a pre-protein (168 kDa) and cleaved to release a passenger domain (120 kDa) from the β-barrel translocation domain (32 kDa), leaving the mature 120 kDa domain associated with the outer leaflet of the outer membrane (Hackstadt et al., 1992). The high degree of shared sequence identity and homology of rOmpB molecules expressed by closely related and divergent rickettsia suggest that rOmpB may also be functionally conserved (Blanc et al., 2005). The full-length rOmpB from R. japonica, a rickettsial species closely related to R. conorii, is sufficient to mediate adherence and invasion of Vero cells when expressed in non-invasive E. coli (Uchiyama et al., 2006). The R. conorii rOmpB β-peptide has been shown to interact with mammalian surface proteins (Renesto et al., 2006). Together, these studies emphasized an important role for rOmpB in rickettsial virulence.
Previous work has identified a 70kDa protein, Ku70, from mammalian cell lines as a receptor involved in R. conorii entry (Martinez et al., 2005). Ku70 is multifaceted in its cellular roles and compartmentalization, as it is found localized to the nucleus where it complexes with Ku80 and PARP to form the DNA-PK complex for DNA non-homologous end joining (NHEJ); in the cytoplasm, where it sequesters Bax from the mitochondria as an inhibitor of the intrinsic apoptotic pathway; and at the plasma membrane, where as a possible heterodimer with Ku80, has been shown to mediate cell-cell adhesion, extracellular matrix (ECM) attachment through interaction with fibronectin, and (reviewed in (Muller et al., 2005)) matrix invasion by modulation of matrix metalloproteases during angiogenesis or tumor invasion (Monferran et al., 2004b). Although ubiquitously expressed in mammalian cells, Ku70 is localized in abundance on the surface of a subset of cell types including non-transformed endothelial cells, as well as tumor cell lines including HeLa and Vero (Muller et al., 2005). By affinity chromatography approaches, rOmpB from R. conorii was identified as the sole rickettsial ligand of Ku70 (Martinez et al., 2005). These results suggested that rOmpB-expressing SFG rickettsia may have evolved strategies to utilize Ku70 on the surface of target cells target cells to initiate signals leading to rickettsial uptake; however, the mechanisms through which rOmpB and Ku70 induces cellular invasion remained unknown.
In this study, we further characterize the roles of R. conorii rOmpB in early bacterial-host interactions. Using a heterologous E. coli expression system, we determined that expression of rOmpB is sufficient to mediate association with and invasion of non-phagocytic mammalian cells, and that this invasion process is Ku70-dependent. We find that purified recombinant rOmpB passenger domain interacts with Ku70 and additionally functions as a competitive inhibitor of bacterial attachment. Through the use of pharmacological inhibitors, we show that rOmpB-Ku70 mediated bacterial uptake relies in part on actin polymerization, microtubule stability and protein tyrosine kinase and phosphatidyl inositide 3-kinase activities. We also observe that the E3 ubiquitin ligase, c-Cbl, is involved in rOmpB-mediated uptake, and that depletion of components of the endocytic machinery, namely clathrin and caveolin-2, inhibits rOmpB-mediated invasion of HeLa cells. Our findings continue to stress the importance of rOmpB in the rickettsial entry process and provide the first insight into the signaling involved in Ku70-dependent internalization.
To study the contribution of rOmpB and Ku70 in R. conorii invasion, we adapted an Escherichia coli-based heterologous protein expression system for the study of rickettsial antigens (Uchiyama et al., 2006). The full-length R. conorii ompB gene either including or excluding the endogenous rOmpB signal sequence was cloned into the E coli isopropyl-D-β-thiogalactopyranoside (ITPG) inducible expression vector, pET-22b, resulting in plasmids pJJM104 and pYC9 respectively (Figure 1A). The plasmid, pET22-RJPOB, encodes the full-length R. japonica rOmpB allele and was used as a positive control (Uchiyama et al., 2006). The plasmids were transformed into the E. coli expression strain, BL21(DE3), and the resulting strains were induced for surface protein expression. Biochemical fractionation of the induced E. coli strains indicates that rOmpB localizes to the outer membrane of rOmpB-expressing strains, as detected by Western immunoblotting with anti-6XHis antisera and rabbit hyper-immune sera against R. conorii (αRc7). To determine whether rOmpB expressed in E. coli is surface-exposed, IPTG-induced and uninduced E. coli BL21(DE3) carrying the pET22-b, pYC9 or pET22-RJPOB were fixed and processed for immunofluorescence using the αRc7 antibody. Figure 1C shows positive rOmpB staining in IPTG-induced rOmpB-expressing strains, which is absent in uninduced or empty vector strains. Negative propidium iodide staining indicates that αRc7-positive staining is not the result of membrane permeabilization, illustrating that rOmpB is indeed surface exposed in the E. coli heterologous expression system.
We then assayed for the ability of rOmpB-expressing E. coli BL21(DE3) strains to mediate attachment to mammalian cells using immunofluorescence-based and colony-forming unit-based assays, as previously described (Martinez and Cossart, 2004; Martinez et al., 2000), relative to those containing the empty vector. E. coli BL21(DE3) (pET22-RJPOB), a strain encoding rOmpB from the closely related SFG rickettsia, R. japonica (Uchiyama et al., 2006) was utilized as a positive control in these assays. As shown in Figure 2A, expression of either full-length rOmpB from R. conorii, pYC9, or R. japonica, pET22-RJPOB, is sufficient to mediate cell association to cultured HeLa cells compared to E. coli expressing the empty vector, pET-22b. Quantification of adherence using a colony-forming unit-based assay confirmed these results (Figure 2B, see Experimental Procedures).
We sought to interrogate whether rOmpB expression in E. coli could mediate invasion of mammalian cells. HeLa cells infected for 60 min with E. coli expressing the R. conorii or R. japonica rOmpB allele were examined by scanning electron microscopy (SEM). Many of the cell-associated, rOmpB-expressing E. coli appeared entangled in microvilli (Figure 3A, left panels), which are present in great abundance on the surface of HeLa cells. In some cases, rOmpB-expressing E. coli infected HeLa cells exhibited highly suggestive membrane rearrangements implicative of bacterial internalization (Figure 3A, arrows). In contrast, E. coli expressing the empty vector were scarce and difficult to find; those on the cell surface appeared passively associated and did not exhibit any sort of interaction with the microvilli (data not shown). Examination of cells infected with rOmpB-expressing E. coli using transmission electron microscopy (TEM) showed internalized, often vacuole-enclosed bacteria (Figure 3B), demonstrating a role for rOmpB in mediating bacterial invasion. To quantitatively assess the ability of rOmpB to stimulate bacterial internalization in mammalian cells, we utilized a gentamicin protection assay (see experimental procedures). Consistent with our microscopy data, E. coli expression of full-length rOmpB protein in the absence of other putative rickettsial virulence factors was able to mediate invasion of non-phagocytic HeLa cells (Figure 3C). Similar observations and results were seen in Vero cells (data not shown). The presence of the endogenous rOmpB signal sequence in conjunction with the pET-22b vector-encoded PelB signal sequence had no apparent influence on rOmpB-mediated cell-association and invasion (data not shown); therefore E. coli BL21(DE3) (pJJM104) and E. coli BL21(DE3) (pYC9) were used interchangeably.
Previous work identified Ku70 as a receptor for R. conorii invasion in mammalian cells, and rOmpB as a Ku70 ligand (Martinez et al., 2005). To determine whether Ku70 functions in rOmpB-mediated bacterial invasion, we performed invasion assays on HeLa cells transfected with either a scrambled control siRNA or Ku70 siRNA. Reduction of endogenous Ku70 protein levels in HeLa cells did not significantly affect rOmpB-mediated cell-association (Figure 4A), in agreement with a previous study examining R. conorii adherence to Ku70 siRNA-treated mammalian cells (Martinez et al., 2005). In contrast, invasion of rOmpB-expressing E. coli strains into Ku70-depleted HeLa cells decreased to levels comparable with that of the non-invasive, vector control (Figure 4B). These results suggest that while rOmpB may interact with multiple eukaryotic plasma membrane proteins to mediate cell association, interactions via Ku70 productively mediate bacterial entry.
Numerous studies (Uchiyama et al., 2006; Martinez et al., 2005) suggest that a functional component of the rOmpB protein is contained within the passenger domain. To assess this hypothesis, we expressed and purified the R. conorii rOmpB passenger domain (aa 36–1334) fused to Glutathione-S-Transferase (GST-rOmpB36–1334) under native conditions (Figure 5A). To determine whether the purified GST-rOmpB36–1334 fusion interacts with mammalian cells, we incubated monolayers of HeLa cells with equivalent microgram quantities of either purified GST or GST-rOmpB36–1334, washed and fixed the cells and then processed them for immunofluorescence using anti-GST antisera. As shown in Figure 5B, GST-rOmpB36–1334 bound to HeLa cells in a punctate staining pattern (upper panels) similar to that previously observed for plasma membrane associated Ku70 (Martinez et al., 2005). In contrast, GST alone did not bind to HeLa cells (lower panels). Since we had demonstrated that GST-rOmpB36–1334 could bind to mammalian cells, we sought to determine whether the R. conorii passenger domain could functionally compete for host receptor binding and prevent rOmpB-mediated cell association. As shown in Figure 5C, pre-incubation of HeLa cells with 100 μg/ml of GST-rOmpB36–1334, but not GST, reduced rOmpB-mediated adherence to HeLa cells. These results suggest that the recombinant purified rOmpB passenger domain is properly folded and functional. To investigate if the purified rOmpB passenger domain could interact with Ku70, we performed protein affinity assays using detergent-solubilized HeLa cell lysates incubated with glutathione-sepharose beads coupled with either GST as a control or GST-rOmpB36–1334. Western immunoblot analysis of elutions from these pull-down assays using anti-Ku70 antisera revealed that GST-rOmpB36–1334, but not GST alone, interacts with Ku70 from HeLa cells lysates (Figure 5D), in accordance with the finding that Ku70, purified from mammalian cell extracts, interacts with rOmpB from R. conorii lysates (Martinez et al., 2005). To determine whether the interaction between rOmpB36–1334 and Ku70 precludes any mammalian protein intermediate, we expressed and purified His-tagged Ku701–609 from bacterial lysates (Figure 5E, left panel) and tested this for association with GST or GST-rOmpB36–1334 coupled to glutathione-sepharose beads. As shown in Figure 5E (right panel), the bacteria-derived Ku70 associates specifically with rOmpB, indicating that Ku70, in the absence of any other mammalian factors, can interact with rOmpB. Together, these results demonstrate that the purified rOmpB passenger domain interacts with Ku70 and also competitively inhibits rOmpB-mediated bacterial association with the cell.
Previous work had shown that R. conorii entry into epithelial cells involves actin cytoskeletal rearrangements at the site of bacterial contact. These actin rearrangements are thought to be a consequence of signaling through cellular protein tyrosine kinases and phosphoinositide 3-kinases leading to Arp2/3 activation (Martinez and Cossart, 2004). To investigate the involvement of actin in rOmpB-mediated invasion, HeLa cells infected for 60 min with E. coli BL21(DE3) expressing the R. conorii or R. japonica rOmpB alleles were stained for surface-associated E. coli and cellular actin, then examined by confocal fluorescence microscopy (Figures 6A and B). In numerous cases, rOmpB-expressing E. coli appeared to induce localized actin recruitment, indicated by the halo of actin around the bacterium (Figure 6A & B, arrows).
To further query the requirement for actin in the rOmpB-Ku70 mediated invasion process, we employed pharmacological inhibitors of actin (cytochalasin D), as well as phosphotyrosine kinases (genistein), microtubules (nocodazole) and phosphoinositide 3-kinases (wortmannin), players shown to function in activation or stimulation of actin polymerization. We performed gentamicin protection assays using rOmpB-expressing E. coli strains and HeLa cells pre-treated with either DMSO (control), or increasing concentrations of the pharmacological inhibitors, and assayed for the loss of the invasion phenotype. Neither DMSO nor the drugs at their highest concentrations appeared to affect bacterial or mammalian cell viability by colony forming unit enumeration and trypan blue staining, respectively (data not shown). As shown in Figure 6C and D, R. conorii and R. japonica rOmpB-mediated invasion of HeLa cells was inhibited in a concentration dependent manner, in cells treated with cytochalasin D and genistein, in most cases reducing invasion to levels observed of the non-invasive control (Figure 6A and B, red dashed line). Cells treated with increasing concentrations of wortmannin exhibited only an intermediate inhibitory effect on invasion, reminiscent of that seen in R. conorii infections (Martinez and Cossart, 2004).
Previous studies had implicated microtubule dynamics as playing a critical role in uptake of invasive “zippering” pathogens, such as Campylobacter jejuni (Biswas et al., 2003) and in regulating Arp2/3 mediated actin cytoskeletal changes (Campellone et al., 2008). We therefore, investigated whether inhibition of microtubule stability would affect rOmpB-mediated invasion of HeLa cells. A shown in Figure 6C and D, destabilization of the cells’ microtubule cytoskeletal infrastructure with nocodazole reduced rOmpB-Ku70 mediated invasion to levels similar to the non-invasive control (dashed line in Figure 6C and D). The comparable effects on invasion seen between E. coli expressing the R. conorii and R. japonica rOmpB allele suggests rOmpB may stimulate entry similarly among related rickettsial species. Furthermore, the mechanism for rOmpB-Ku70-mediated internalization correspondingly mirrors that previously observed in the R. conorii invasion process (Martinez and Cossart, 2004). Taken together, these results suggest that the integrity of actin and microtubule cytoskeletal structures is crucial for rOmpB-mediated invasion.
Recent studies have shown that many pathogens, including Listeria monocytogenes and Yersinia pseudotuberculosis, utilize the “zippering” pathway of invasion, high-jacking the clathrin-, caveolin and ubiquitin-mediated endocytosis of host receptors to gain access to the nutrient-rich intracellular environment of mammalian host cells (Bonazzi and Cossart, 2006; Hamon et al., 2006; Veiga and Cossart, 2005). Invasion of “zippering” bacteria into non-phagocytic mammalian cells is therefore a combination of signalling events leading to localized actin rearrangements at entry foci coupled with the endocytosis of host cell receptors (Veiga et al., 2007). Previous studies have also demonstrated that the E3 ubiquitin ligase c-Cbl mediated ubiquitin modification of Ku70 correlates with R. conorii entry (Martinez et al., 2005) and that rOmpB is sufficient to mediate invasion of non-phagocytic mammalian cells (Uchiyama et al., 2006). As ubiquitin modification of Ku70 appears to be an important event during rickettsial entry, we first determined the contribution of c-Cbl to the rOmpB-dependent entry process using siRNAs against endogenous human c-Cbl. As shown in Figure 7A, inhibition of endogenous c-Cbl expression in HeLa cells reduced rOmpB-mediated uptake to levels similar to those observed for cells treated with Ku70 siRNA.
A recent study has demonstrated that the InlB-mediated invasion of L. monocytogenes involves the sequential recruitment of clathrin and actin, respectively, to entry foci and that the recruitment of clathrin is required for the localized actin rearrangements involved in efficient bacterial entry (Cossart and Veiga, 2008). We next determined if major components of the endocytic pathway also contribute to rOmpB-mediated invasion. As shown in Figure 7B, transfection of siRNAs directed against caveolin-1 (Cav1), caveolin-2 (Cav2) and clathrin heavy chain (CLTC) were able to reduce endogenous expression of these proteins. Interestingly, rOmpB-mediated invasion of HeLa cells was diminished in cells transfected with siRNAs against Cav2 and CLTC, but not Cav1, suggesting that rOmpB-mediated uptake is independent of Cav1 (Figure 7B). Together, these results suggest a role for c-Cbl, clathrin and caveolin 2-mediated endocytosis in the entry of rOmpB-expressing bacteria into non-phagocytic mammalian cells.
The ability of spotted fever group (SFG) rickettsiae to bind to and invade target mammalian cells is a critical initial event during pathogenesis. A previous report identified a mammalian protein, Ku70, as a receptor involved in R. conorii invasion and had identified the rickettsial autotransporter protein, rOmpB, as a ligand (Martinez et al., 2005). These results suggested that the interactions between Ku70 and rOmpB are important to initiate signals leading to rickettsial uptake. Here we demonstrate that rOmpB from both R. conorii and the closely related SFG rickettsia, R. japonica, are sufficient to mediate invasion of non-phagocytic mammalian cells in a Ku70-dependent manner.
rOmpB is an abundant surface protein expressed by rickettsiae and exhibits homology to a family of modular proteins in Gram-negative bacteria called autotransporters (Blanc et al., 2005). We have shown that rOmpB from R. conorii, when expressed in E. coli, can mediate association to and invasion of non-phagocytic mammalian cells. A previous study has suggested that the rOmpB β-peptide domain may serve as a rickettsial adhesin (Renesto et al., 2006); our results indicate that the passenger domain also functions in cellular recognition and may also be responsible for adherence to target cells. While rOmpB from R. conorii and R. japonica presumably interact with multiple eukaryotic ligands at the plasma membrane, our findings suggest that interactions with Ku70 are integral in the invasion process. Given the high degree of shared sequence identity and homology in the passenger domains of rOmpB molecules expressed by closely related and divergent rickettsia (Blanc et al., 2005), we propose that Ku70 is likely utilized by multiple rickettsial species to gain entry into non-phagocytic mammalian cells and that disruption of this interaction in part can block invasion.
The importance of rOmpB in rickettsial pathogenesis was highlighted by the observation that anti-rOmpB antibodies protect mice from an otherwise lethal challenge of R. conorii (Feng et al., 2004b; Feng et al., 2004a) in a model of Mediterranean spotted fever. Interestingly, anti-rOmpB antisera raised against R. conorii rOmpB protein are cross protective against a closely related SFG rickettsia, R. rickettsii, and a phylogenetically divergent rickettsia, R. australis, (Feng et al., 2004b; Feng and Walker, 2003; Stenos and Walker, 2000) likely due to the high degree of sequence identity shared amongst rOmpB molecules. A previous study demonstrated that antisera against a specific region in the R. rickettsii rOmpB passenger domain (aa 451-1308) elicited protective humoral immunity in a murine model of infection. In contrast, antisera raised against part of the predicted rOmpB β-peptide (aa 1335-1704) were not protective (Diaz-Montero et al., 2001). A corresponding sub-region in the R. conorii rOmpB passenger domain (aa 545-848), is involved in the activation of effector CD8+ T-cells (Li et al., 2003). We have shown that a purified GST-rOmpB36–1334 passenger domain fusion binds to Ku70 in a biochemical pull-down assay. We have also demonstrated that GST-rOmpB36–1334 binds to mammalian cells in a punctate pattern, and competitively inhibits R. conorii and R. japonica rOmpB-mediated adherence. Interestingly, a previous report showed that the R. conorii rOmpB passenger domain, but not the β-peptide, interacts with the Ku70 N-terminal domain (aa 1-535) (Martinez et al., 2005) and that a monoclonal antibody (mAb) directed against an exposed Ku70 epitope (aa 506-541) does not block R. conorii attachment to cells, but effectively blocks internalization. We predict that a surface exposed rOmpB domain is likely involved not only in eliciting protective immune responses, but also recognizing Ku70. Identification of the domains involved in rOmpB-Ku70 interactions will be crucial to the development of novel anti-rickettsial therapies.
Numerous studies have implicated that entry of R. conorii morphologically and mechanistically resembles a zipper-like invasion strategy (Gouin et al., 1999; Teysseire et al., 1995), similar to that utilized by Listeria monocytogenes, Yersinia enterocolitica and uropathogenic E. coli (UPEC). In this mechanism, invasion is mediated by the sequential interactions of specific bacterial ligands and host receptors that lead to localized actin recruitment and “zippering” of the plasma membrane around the bacterium (Cossart and Sansonetti, 2004). Previous work had also shown that the entry of R. conorii into non-phagocytic mammalian cells requires the activation of host signaling pathways, involving the tyrosine phosphorylation of host proteins, the activation of protein tyrosine kinases (c-Src) and the activation of lipid kinases (PI 3-Kinase), ultimately leading to localized Arp 2/3 driven actin polymerization (Martinez and Cossart, 2004). Our pharmacological inhibitor studies show that rOmpB-mediated uptake closely parallels that of R. conorii, suggesting that rOmpB is sufficient to trigger these signaling pathways leading to bacterial entry. Interestingly, inhibition of microtubule dynamics by nocodazole treatment also inhibited rOmpB-mediated invasion of mammalian cells, as has been previously observed for the invasive, “zippering” pathogen, Campylobacter jejuni (Biswas et al., 2003). Exactly how microtubule and actin rearrangements could be regulated in the uptake process is unclear; however, a recent report identified a modular protein in mammalian cells termed WHAMM (WASP homolog associated with actin, membranes, and microtubules) that can bind to microtubules and promote membrane tubulation and stimulate Arp2/3 mediated actin rearrangements (Campellone et al., 2008). Disruption of microtuble integrity with nocodazole may also indirectly disrupt the regulation of Arp2/3 mediated actin rearrangements that are crucial for invasion of mammalian cells. The mechanism by which microtuble dynamics and actin rearrangements are regulated during bacterial invasion warrants further investigation.
Recent evidence has shown that zippering invasive pathogens can also highjack the host cellular endocytic machinery to gain entry into non-phagocytic cells (Veiga and Cossart, 2005). For example, the invasive pathogen, Listeria monocytogenes induces the mono-ubiquitination of receptors HGFR (c-Met) (Veiga and Cossart, 2005) and E-cadherin and triggers the recruitment of clathrin and caveolin-1 to entry foci (Veiga et al., 2007; Bonazzi and Cossart, 2006). Mechanistic similarities between the entry of L. monocytogenes and R. conorii suggested that ubiquitin modification of mammalian proteins, including the receptor Ku70, may also be important during SFG rickettsia invasion. Martinez and co-workers showed that R. conorii induces a rapid ubiquitination of Ku70 and that inhibition of the E3 ubiquitin ligase c-Cbl by siRNAs reduces R. conorii entry into HeLa cells (Martinez et al., 2005). These results suggested that c-Cbl-mediated ubiquitination of Ku70 may be a critical signaling intermediate in R. conorii entry (Martinez et al., 2005). Using siRNAs, we confirmed that rOmpB-mediated entry is dependent on c-Cbl. Whether rOmpB is sufficient to mediate Ku70 ubiquitination during the entry process is currently being investigated. We also determined that rOmpB-mediated invasion was in part dependent on clathrin and caveolin-2. In contrast to that observed for InlA-mediated uptake of L. monocytogenes (Bonazzi and Cossart, 2006), inhibition of caveolin-1 had little effect on rOmpB-mediated invasion of HeLa cells. Interestingly, the entry of non-zippering pathogens, Salmonella typhimurium and Shigella flexneri, was found to be independent of clathrin (Bonazzi and Cossart, 2006), suggesting that pathogens have evolved alternate strategies to gain access to the intracellular environment. These data support clathrin-mediated endocytosis as a general mechanism usurped by zippering pathogens to enter non-phagocytic mammalian cells and suggest that distinct utilization of caveolin-1- and caveolin-2-dependent pathways may be attributed to differences in receptor utilization. Interestingly, a recent report demonstrated that the InlB-mediated invasion of L. monocytogenes requires a sequential recruitment of clathrin and actin to entry foci. Inhibition of clathrin expression by siRNA transfection disrupted the recruitment of actin to sites of entry suggesting that clathrin recruitment is a required for the activation of actin rearrangements (Cossart and Veiga, 2008). Taken together, these results suggest that stimulation of signals at the level of the plasma membrane receptor may be coordinated to not only recruit and activate components of the host cell endocytic machinery but also to recruit localized actin to sites of bacterial attachment, ultimately leading to bacterial internalization of “zippering” pathogens.
Although ubiquitously expressed in mammalian cells, Ku70 has been found associated with the plasma membrane in high abundance on a subset of cell types, including endothelial cells, macrophages, solid tumors under hypoxic conditions and tumor cell lines, including HeLa and Vero (Muller et al., 2005). Endothelial cells represent the major target cell for SFG rickettsia in vivo and we suggest that rickettsia have evolved a mechanism to bind to and utilize the abundant protein, Ku70, at the plasma membrane to trigger internalization. A recent study demonstrated that plasma membrane associated Ku70 plays an important role in cell-cell adhesion and in the attachment of cells to some extra-cellular matrix (ECM) proteins (Monferran et al., 2004a). However, little else is known about the function and the regulation of Ku70 at the plasma membrane. In the absence of recognizable protein interacting motifs, such as SH2 and SH3 domains, present in other mammalian receptors, it is difficult to rationalize how Ku70 can transmit signals eventually leading to bacterial uptake. One hypothesis is that Ku70 may interact with other surface proteins to facilitate rickettsial entry. Through yeast two-hybrid screens (Y2H), several groups have identified Ku70 interactors, including, matrix-metalloprotease nine (MMP-9), a protein that plays a central role in wound healing, angiogenesis, arthritis and tumor metastasis (Monferran et al., 2004b) and the ADP-ribosylation factor 6 (Arf6), a small GTP-binding protein that regulates membrane traffic and the actin cytoskeleton at the plasma membrane (Schweitzer and D’Souza-Schorey, 2005). In addition, under certain conditions, Ku70 was found to associate with the epidermal growth factor receptor (EGFR) on mammalian cells (Bandyopadhyay et al., 1998), although the significance of this interaction to Ku70 function remains unclear. Whether these proteins play a role in rickettsia entry is currently being investigated. Interestingly, a recent report elucidated a critical role for non-muscle myosin IIA (NMMIIA) in the transport of a nuclear protein, nucleolin, to the plasma membrane where nucleolin can then interact with various ligands (Huang et al., 2006). NMMIIA is an actin motor protein involved in the modulation of the actinomyosin cytoskeleton and regulating cell-shape, changes in cell migration, secretion or cell division (Sellers, 2000). Since actin polymerization plays a crucial role in rOmpB-mediated internalization, one possibility is that NMMIIA aids in modulating actin contractility at the endocytic cup. Alternatively, NMMIIA may be involved in the trafficking of Ku70 to the plasma membrane, similar to that observed with nucleolin. Interactions of rOmpB with Ku70 may trigger signaling events that couple actin rearrangements and recruitment of components of the endocytic machinery to entry foci. The potential role of this myosin in rickettsial entry is intriguing and warrants further investigation.
Our results underscore the importance of rOmpB and Ku70 in the rickettsia entry process. However, recent data suggest that other host cell receptors and rickettsia ligands likely also play important roles in the adherence and invasion of mammalian cells. Several studies have illustrated the importance of another related rickettsia autotransporter protein, rOmpA, in the attachment to host cells and in the generation of protective humoral immune responses (Li and Walker, 1998; Li et al., 1988; Anacker et al., 1987a; Anacker et al., 1987b). Sequence data mining revealed the presence of a gene family termed sca (surface cell antigens) within different SFG rickettsial genomes whose genes are predicted to encode either secreted proteins or outer membrane proteins (Blanc et al., 2005). Although many genes appear to be fragmented or split, several genes, including ompB, are present in nearly all SFG rickettsiae (Blanc et al., 2005; Roux and Raoult, 2000), suggesting that they may play important roles in rickettsial pathogenesis. Elucidating the functions of rOmpB in addition to conserved Sca proteins is crucial to understanding the complex host-pathogen interaction underlying successful SFG rickettsia infections in human hosts and may lead to the development of more efficacious therapies.
HeLa cells (ATCC, Manassas, VA) and Vero cells and were grown in Dubecco’s Modified Eagle’s Medium (DMEM) supplemented with 10% heat-inactivated fetal bovine serum, 1× non-essential amino acids (Lonza, Walkersville, MD) and 0.5 mM sodium pyruvate. Cells were grown at 37°C/5% CO2. Escherichia coli BL21(DE3) or TOP10 were grown in LB Miller broth at 37°C supplemented with ampicillin (100 μg/ml) where appropriate. Bacteria were diluted 1:20 from overnight cultures, grown to an OD600 = 0.6, and induced with 100 μM Isopropyl-β-D-thiogalactopyranoside (IPTG) for 3 h at 37°C or as otherwise indicated.
For immunoblot detection, the anti-Ku70 mAb (N3H10) was purchased from NeoMarkers (Fremont, CA). Anti-Clathrin mAb (CHC5.9) was purchased from Chemicon International (Temecula, CA). The anti-6xHis rabbit polyclonal was obtained from Covance (Berkeley, CA). Anti-GST rabbit sera (Z-5) was purchased from Santa Cruz Biotechnology (Santa Cruz, CA). The anti-c-Cbl and anti-Caveolin-1 rabbit polyclonals were purchased from Cell Signaling Technology (Danvers, MA). Anti-Cavelolin-2 rabbit antisera and the polyclonal goat anti-mouse IgM horseradish peroxidase (HRP) conjugate were purchased from Abcam (Cambridge, MA). The anti-actin mAb (AC-15), rabbit anti-mouse IgG HRP conjugate and goat anti-rabbit IgG HRP conjugate used for western immunoblot analysis; the pharmacological inhibitors, cytochalasin D, genestein, nocodazole and wortmannin were purchased from Sigma (Saint Louis, MO). AlexaFluor 488-conjugated goat anti-rabbit IgG and Cell Tracker™ Red CMTPX, used in immunofluorescence studies, and the pCR2.1 TOPO TA Cloning kit were purchased from Invitrogen (Carlsbad, CA). The gentamicin sulfate was purchased from MP Biomedicals (Solon, OH). Annealed small-interfering RNAs (siRNAs) against caveolin-1 sense strand 5′-CGAAAUACUGGUUUUACCGtt-3′ and anti-sense strand 5′-CGGUAAAACCAGUAUUUCGtc-3′; caveolin-2 sense strand 5′-GCACACAACGAUUAUAGUAtt-3′ and anti-sense strand 5′-UACUAUAAUCGUUGUGUGCtt-3′; c-Cbl sense strand 5′-CCUUAUAUCUUAGACCUGCtt-3′ and anti-sense strand 5′-GCAGGUCUAAGAUAUAAGGtg-3′; clathrin heavy chain sense strand 5′-GGGAAGUUACAUAUUAUUGtt-3′ and anti-sense strand 5′-CAAUAAUAUGUAACUUCCCtc-3′; Ku70 sense strand 5′-GGCUAUGUUUGAAUCUCAGtt-3′ and anti-sense strand 5′-CUGAGAUUCAAACAUAGCCtt-3′; and the Silencer negative control #1 siRNA were obtained from Ambion (Austin, TX). Complete protease inhibitor cocktail was purchased from Roche (Indianapolis, IN). Glutathione Sepharose™ 4B beads used in GST purifications were purchased from Amersham Biosciences (Piscataway, NJ). Chemically competent E. coli BL21(DE3) were attained from Stratagene (La Jolla, CA). The QIAquick® Spin Kit was obtained from Qiagen (Valencia, CA). The pET-22b vector was purchased from Novagen (Gibbstown, NJ). Restriction enzymes BamHI, NcoI, XhoI and KpnI were all obtained from New England BioLabs (Ipswich, MA).
ompB fragments were amplified by polymerase chain reaction (PCR) from a chromosomal preparation of Rickettsia conorii Malish 7 containing the ompB gene (Genbank accession no. AAL03623). All constructs were initially cloned into pCR2.1 (Invitrogen), digested, gel-purified using the QIAquick® Spin Kit (Qiagen) before being inserted into its expression vector. Plasmids used in this study are listed in Table 1. The full-length ompB was PCR amplified using the primers GAGCCCGGATCCAGCTCAAAAACCAAATTTTCT and GGCTCGAGGAAGTTTACACGGACTTTTAG and inserted into pET-22b (BamHI and XhoI sites) generating pJJM104. The plasmid pYC9, encoding the full-length R. conorii rOmpB excluding its endogenous signal peptide, was generated by PCR amplification of ompB from the chromosome using the primers ACCATGGCTATACAGCAGAATAGAAC and GGCTCGAGGAAGTTTACACGGACTTTTAG, then digested and ligated into pET-22b (NcoI and XhoI). The region of ompB encoding the passenger domain of the protein was PCR amplified using the primers ACCATGGCTATACAGCAGAATAGAAC and GGCTCGAGTAATCTGTTACCAAGTTGAGC, digested (NcoI and XhoI) and ligated into pGEX-2TKP (gift from Dr. T. Kouzarides, UK Gurdon Institute), generating pYC11. The plasmid pYC30, encoding a 10xHis-Ku701–609 was generated by excision of XRCC61–1827 from pYC29 (NdeI and XhoI) and ligation into pET-16b (NdeI and XhoI).
100 ml of induced E. coli BL21(DE3) cultures were pelleted and resuspended in 10 ml 20 mM Tris (pH 8) containing 1× protease inhibitor (Roche). Cells were passed twice through the French Pressure cell (1,500 psi) and cleared by centrifugation at 10,000 g for 5 min. Inner membrane proteins were extracted by incubation with sarkosyl (final concentration of 0.5%) at room temperature for 15 min. Outer membranes were pelleted by ultracentrifugation (SW40 Ti, 32000 rpm, 30 min, 4°C) and resuspended in 2× sample buffer. Whole cell lysates, soluble/sarkosyl-solubilized and outer membrane fractions were resolved by SDS-PAGE and analyzed by immunoblotting with anti-6xHis rabbit sera (Covance) and goat anti-rabbit HRP-conjugate (Sigma).
Cell association and invasion assays were performed as described (Martinez et al., 2000). Briefly, HeLa cells were seeded into 24-well plates at 1.2×105 cells per well 24 h prior to infection. Prior to infection, the cells were washed 3× with serum-free DMEM and the medium was replaced with 950 μl of pre-warmed serum-free DMEM. Triplicate wells were then infected with 50 μl of induced bacteria (OD600=1.0) resuspended in phosphate-buffered saline (PBS). Bacterial contact with host cells was initiated by centrifugation of the plates at 200 × g for 5 minutes. Plates were incubated at 37°C/5% CO2 for 20 or 60 min to assess for cell association and invasion, respectively. In cell association assays, cells were washed 5× with PBS or serum-free DMEM, then lysed with 1 ml 0.1% Triton X-100 in ddH2O and plated onto LB-agar plates for colony forming units. Adherence frequencies were calculated as the percent of cell-associated bacteria recovered after the PBS washes out of the total bacteria in each well after the 20 min incubation. To determine bacterial invasion, cells were washed 3× with PBS or serum-free DMEM following the 1 h incubation period, then incubated for an additional 2 h with 100 μg/ml of gentamicin sulfate (MP Biomedicals) in complete DMEM. Thereafter, cells were washed 3× with PBS, then lysed with 0.1% Triton X-100 in ddH2O and plated onto LB-agar plates. Invasion frequencies were determined as the number of bacteria surviving the gentamicin challenge out of the total bacterial input following the initial 1 h incubation. In inhibitory assays, cell association and invasion were subsequently normalized as a percent relative to the infections of mock-treated cells. Infections were done in triplicate and results are representative of at least three experiments. P-values were determined by a two-tailed Student’s t-test on replicates from a single experiment.
For scanning electron microscopy (SEM), HeLa cells grown on 12 mm glass coverslips in 24-well plates were infected with 10 μl of induced bacteria (OD600=1.0) for 1 hour, washed, then fixed for 20 min with 4% paraformaldehyde (PFA) in PBS, followed by fixation in 2.5% glutaraldehyde in PBS for an additional 72 h. Samples were next serially dehydrated with increasing concentrations of ethanol, critical point dried with hexamethyldisilazane, then sputter coated with 80% platinum/20% palladium to 8 nm by Cressington Sputter Coater 208HR. The samples were visualized using Fei NovaNano SEM200 at a distance of 5 mm. For transmission electron microscopy, HeLa cells grown in 6-well plates (3×105 cells per well) were infected with 2 ml of induced bacteria (OD600=2.0) for 2 hours, washed, and then bathed in fixative (2% glutaraldehyde, 4% PFA, 0.1 M sodium cacodylate buffer). Cells were carefully scraped from the plate, pelleted at 500×g for 5 min, allowed to fix overnight at 4°C, then post-fixed with 1% OsO4 in 0.1 M sodium cacodylate buffer for 60 min. The samples were stained in 1% uranyl acetate in maleate buffer for 60 min, and then serially dehydrated with increasing concentrations of ethanol. Cells were embedded in spurr resin for 48 h at 60°C, thin sectioned (90 nm) using a Reichert-Jung Ultracut device and then post-stained in uranyl acetate and lead citrate. The samples were imaged on the FEI Tecnai F30 with a Gatan CCD digital micrograph.
SiRNA transfections were done as previously described (Martinez et al., 2005). Briefly, HeLa cells plated in 6-well plates at 1.2×105 cells per well were transfected with 10 nM of the indicated siRNA (Ambion) using Oligofectamine transfection reagent (Invitrogen), as recommended by the manufacturer. 48 h post-transfection, cells were split into 24-well plates at 1.2×105 cells per well. Cells were tested for cell association, invasion and protein levels 72 h post-transfection. Protein levels were assessed by immunoblot analysis of cell-lysates; anti-Ku70 mAb (N3H10, 200 ng/ml), anti-c-Cbl rabbit sera (1:1000), anti-β actin mAb (AC-15, 1:20000), anti-Clathrin (CHC5.9, 1:100), anti-Caveolin-1 (1:1000), anti-Caveolin-2 (1 μg/ml).
HeLa cells plated in 24-well plates at 1.2×105 cells per well were washed and starved in serum-free DMEM for one hour. 30 min prior to infection, the HeLa cells were incubated in 950 ml of serum-free DMEM containing the cytochalasin D (Sigma), genistein (Sigma), nocodazole (Sigma) or wortmannin (Sigma) at the indicated concentrations, or DMEM with 0.5% (v/v) dimethyl sulfoxide (DMSO) as a mock-treated control. We assessed cell viability by performing trypan blue staining on drug-treated cells. Bacterial viability was assessed by comparing colony-forming units of bacteria following 1 h incubation in DMEM, DMSO or the pharmacological inhibitor at the highest utilized concentration.
Expression and purification of rOmpB36–1334 under native conditions was done utilizing the pGEX prokaryotic expression system. Protein expression was done in E. coli TOP10 strains (Invitrogen) carrying the pGEX derivatives. Overnight cultures were diluted 1:20 into 1 L of fresh medium and grown at 37°C to mid-exponential phase. Cultures were cooled to 30°C and induced using 10 μM IPTG at 30°C for 4–6 h. Bacteria were harvested by centrifugation, resuspended in PBS containing protease inhibitor, and lysed by passing through the French pressure cell 2× (1500 psi). Lysates were cleared by ultracentrifugation (SW40Ti, 32000 rpm, 1 h, 4°C) and purified by gravity flow over glutathione sepharose beads (Amersham Biosciences), and eluted using 20 mM glutathione in PBS. Purified protein was dialyzed into PBS with 10% glycerol, then snap-frozen in liquid nitrogen and stored at −80°C. Expression of 10xHis-Ku701– 609 was done in E. coli BL21(DE3). Following induction using 1 mM IPTG at 37°C for 4 hours, bacteria were harvested by centrifugation and lysed in lysis buffer (20 mM NaH2PO4, 0.5 M NaCl, 20 mM imidazole, pH 7.4) using the French pressure cell as described above. Lysates were cleared by ultracentrifugation (SW40Ti, 32000 rpm, 1 h, 4°C) and loaded onto a 5 ml HisTrap-FF column (GE Healthcare) using an ÄKTA FPLC with UPS-900 UV absorbance monitor and Frac920 fraction collector (GE Healthcare). The column was washed with 18 column volumes (CV) of lysis buffer, 5 CV 125 mM imidazole lysis buffer and then eluted over 6 CV using a linear gradient of 125–500 mM imidazole-containing lysis buffer. Fractions containing 10xHis-Ku701–609 were pooled and dialyzed into 10% glycerol-PBS, then snap-frozen in liquid nitrogen and stored at −80°C.
Glutathione Sepharose™ beads (Amersham Biosciences) coupled to glutathione-S-transferase (GST) or GST-OmpB36–1334 were added to either 1% NP-40 lysis buffer (1% NP-40, 20 mM Tris pH 8.0, 150 mM NaCl, 10% glycerol, 1× complete protease inhibitor), cleared HeLa cell lysates from a confluent 10 cm dish extracted with 1% NP-40 lysis buffer or 5 μg of purified 10xHis-Ku701–609 in 300 μl 1% NP-40 lysis buffer. Bead-protein mixtures were rocked at 4°C for 2 h, pelleted and washed 3× with NP-40 lysis buffer. The supernatants were then aspirated and the beads were boiled in sample buffer for SDS-PAGE analysis.
To assess for surface expression of rOmpB on E. coli BL21(DE3), 200 μl of bacterial culture diluted in PBS (OD600=0.05) were applied to L-polylysine-coated coverslips in a 24-well plate and centrifuged at 200×g for 10 min to induce contact. Bacteria were then fixed with 4% paraformadehyde in PBS for 20 min, washed to remove unassociated bacteria, and allowed to air dry. Bacteria were next rehydrated in PBS for 5 min, then blocked in 2% BSA-PBS for 1 h. E. coli were stained for rOmpB by incubation with anti-Rc7 rabbit hyperimmune sera against heat-killed rickettsia (1:200) in 2% BSA-PBS for 1 h, washed 3× with PBS, followed by a 1 h incubation in the dark with AlexaFluor 488-conjugated goat anti-rabbit IgG (1:1000), and propidium iodide (10 μM) to detect membrane-compromised bacteria. Coverslips were washed 3× with PBS, and then mounted with mowiol onto slides. Images were captured on an Olympus AX-70 fluorescence microscope coupled to a charge-coupled device (CCD) camera at 1000× magnification. Images were assembled in Adobe Photoshop.
In cell association assays, following infection in 24-well plates (described under cell association assay), cells were washed and fixed with 4% PFA at room temperature for 20 min. To assess for total association of E. coli, cells were permeabilized with 0.1% Trition X-100 in PBS and subsequently stained for total E. coli with rabbit anti-E. coli antisera (1:1000) in 2% BSA-PBS for 1h, followed by AlexaFluor 488-conjugated goat anti-rabbit IgG, Texas-Red phalloidon (1:200) and DAPI (1:10,000) in 2% BSA-PBS. Images were digitally captured on a Nikon Eclipse TE2000-u microscope coupled to a CCD camera using 200× magnification and processed using Adobe Photoshop.
To assess for association of purified rOmpB with cells, HeLa cells plated at 90% confluency (1×105 cells/well) in 24-well plates were incubated for 20 min at 37°C/5% CO2 with 100 μg/ml of GST or GST-rOmpB36–1334 in serum-free DMEM. Cells were then washed 5× with PBS, fixed for 20 min at room temperature with 4% PFA in PBS and processed for immunofluorescence as follows: cell-associated GST or GST-OmpB36–1334 were stained using an anti-GST rabbit sera (1:500, Santa Cruz Biotechnology) in 2% BSA-PBS for 1h, followed with AlexaFluor 488-conjugated anti-rabbit IgG and DAPI. Images were digitally captured on a Nikon Eclipse TE2000-u microscope coupled to a CCD camera using 200× magnification and processed using Adobe Photoshop.
In actin recruitment assays, HeLa cells grown on glass coverslips in 24-well plates were infected for 1 h at 37°C/5% CO2 with 50 μl of induced E. coli (OD600=1.0). Cells were washed 5× with PBS, fixed for 20 min with 4% PFA in PBS and stained for extracellular E. coli using a rabbit anti-E. coli antisera followed by AlexaFluor 488-conjugated goat anti-rabbit IgG. Next, cells were permeabilized with 0.1% Triton X-100 and stained for cellular actin using Texas-Red phalloidon. Glass coverslips were mounted onto slides using an anti-fade reagent, mowiol. Images were captured by confocal fluorescence microscopy on an Olympus DSU spinning disk confocal microscope and back-thinned EM-CCD camera at 1000× magnification and 0.1 μm step size. Z-stack slices were analyzed in ImageJ software and single slices assembled in Adobe Photoshop.
We thank Dr. T. Kouzarides (UK Gurdon Institute, Cambridge, UK) for the pGEX-2TKP vector and Alice Cheng for the pET-16b vector; Yimei Chen and Dr. Qiti Guo for their assistance in the transmission and scanning electron microscopy, respectively; Dr. Sean P. Riley for his critical reading of this manuscript. We also thank Patricia Crocquet-Valdes (UTMB, Galveston, TX) for providing the RC7 antibody. Y.G.Y.C. additionally thanks Dr. Olaf Scheenwind, Justin Kern, Alice Cheng and Lydia Bright for their invaluable guidance and suggestions. This work was sponsored by the NIH/NIAID Regional Center of Excellence for Biodefense and Emerging Infectious Diseases Research (RCE) Program. J.J.M. acknowledges membership within and support from the Region V ‘Great Lakes’ RCE (NIH award 1-U54-AI-057153). Y.G.Y.C. acknowledges support from the Genetics and Regulation Training Grant (5 T32 GM007197). M.M.C. acknowledges support from the Molecular Cell Biology Training Grant (T32 GM007183).