Understanding the mechanisms that Chlamydia species use to gain entry into host cells is complicated by the ability of this obligate intracellular pathogen to enter host cells via multiple routes and by the inability to carry out classical bacterial genetics. In this study, we have circumvented these obstacles by using RNAi to carry out a large scale forward genetic screen in D. melanogaster, a surrogate host with less functional redundancy, to identify host proteins required for early steps in C. trachomatis infection. Our screen confirmed some previously known host targets and has, most importantly, identified for the first time the activation of PDGFR and Abl kinase signaling pathways as key events in the pathogenesis of C. trachomatis infections. We demonstrate that in mammalian cells, PDGFRβ functions as a receptor for C. trachomatis binding and that once bound, bacterial internalization can occur either through activation of PDGFRβ or through independent activation of Abl kinase. Activation of these kinases culminates in phosphorylation of the Rac guanine nucleotide exchange factor, Vav2, and several actin nucleators, including WAVE2 and Cortactin, that ultimately promote efficient uptake of this obligate intracellular parasite.
The initial step in Chlamydia
binding is thought to be a reversible, electrostatic interaction with heparan sulfate-like glycosaminoglycans followed by an irreversible interaction with an unknown receptor 
. We provide compelling data that at least one receptor for C. trachomatis
binding is PDGFRβ. We demonstrate that RNAi-mediated depletion of PDGFRβ or addition of a neutralizing antibody to PDGFRβ significantly decreases bacterial binding to mammalian cells. Consistent with its role as a receptor, phosphorylated PDGFRβ is recruited to the site of EB binding.
There are several possible models to explain the interactions between C. trachomatis and PDGFRβ. The bacterium may bind directly to PDGFRβ. Alternatively, the interactions may be indirect. For example, EBs could bind directly to the growth factor receptor ligand (PDGF), which would in turn facilitate binding and activation of PDGFRβ. Another possible scenario is that heparan sulfate (either on the surface of C. trachomatis or on the surface of host cells) could bind to PDGF, facilitating subsequent interaction with PDGFRβ. Finally, EBs, heparan sulfate, PDGF, and PDGFRβ could form a complex, with heparan sulfate and PDGF serving as bridging molecules. Although EGFR was not involved in C. trachomatis binding to host cells under our experimental conditions, suggesting specificity for PDGFRβ we cannot rule out a possible role for other growth factor receptor interactions.
The biological properties of PDGFRβ fit well with the known characteristics of EB binding and entry. The receptors are ubiquitously expressed in cultured cells, consistent with the known ability of C. trachomatis
to enter most cell types in vitro
. PDGFRβ is known to be highly expressed in the uterus and ovaries as well as in macrophages, tissues and cells susceptible to C. trachomatis
infections in vivo 
. PDGFRβ internalization occurs by both clathrin-dependent and clathrin-independent pathways (reviewed in 
), consistent with the reported diversity in C. trachomatis
internalization mechanisms (reviewed in 
). PDGFRβ signals to regulators of the actin cytoskeleton 
that are activated in response to C. trachomatis
infection, including Vav2 (this work), Cortactin (this work), WAVE2 (this work and 
), and Rac 
. PDGF can bind heparan sulfate proteoglycans, and this interaction can enhance PDGF-induced signaling of PDGFR 
, potentially explaining the contribution of this heparan sulfate to Chlamydia
binding. Finally, activation of and entry through a growth factor receptor pathway may serve to promote host cell survival and prevent apoptosis early during infection, which is vital for the successful growth and dissemination of an obligate intracellular parasite. Indeed, the phosphatidylinositol-3 kinase (PI3K) pathway contributes to resistance of C. trachomatis
infected cells to apoptosis 
, although the mechanism by which this pathway is activated remains to be determined. Since PDGFR can exert an antiapoptotic effect in a PI3K-dependent manner 
, we speculate that bacterial binding to PDGFRβ may activate the PI3K pathway.
The results presented here indicate that C. trachomatis binding leads to activation of PDGFRβ and Abl kinase signaling pathways, which operate in a redundant manner to ensure failsafe and efficient uptake of this obligate intracellular parasite into mammalian cells. While inhibition of either Abl kinase or PDGFR alone has a minimal effect on bacterial entry, inhibition of both kinases either by STI571 treatment or by inhibiting PDGFRβ in cells where Abl is deleted or depleted significantly decreases internalization. Furthermore, a STI571-resistant allele of Abl kinase is capable of supporting entry in the presence of drug, indicating that Abl kinase alone is sufficient for entry. We note that in S2 cells, depletion of Abl kinase alone is sufficient to decrease vacuole formation. This may reflect the absence of functional redundancy in Drosophila or may indicate an additional essential role for Abl kinase in post-entry events.
Our results indicate that during infection, C. trachomatis
-induced activation of PDGFRβ is not necessary for activation of Abl kinase even though PDGFRβ signaling has been shown to activate Abl kinase in other settings 
. How Abl kinase is activated upon EB binding remains to be determined. Abl kinase activation may occur through activation of another as yet unidentified surface receptor or it may be activated by TARP or other translocated bacterial effectors 
Our work demonstrates that C. trachomatis
infection leads to tyrosine phosphorylation and recruitment of several key molecules involved in Rac-dependent actin rearrangements that are known to be regulated by PDGFR and Abl kinase 
. These include Vav2, a Rac GEF as well as WAVE2 and Cortactin, activators of the Arp2/3 complex in the Rac pathway. Although our IF data using the 4G10 antibody ( and ) suggests that Abl kinase is the major kinase responsible for phosphorylating EB-associated proteins, our western blot data () indicates that PDGFR activity also contributes to tyrosine phosphorylation of WAVE2, Vav2, and Cortactin. Since Abl kinase and PDGFR function redundantly for C. trachomatis
entry, the indispensable role of Abl kinase in tyrosine phosphorylation of EB-associated proteins and the dispensable role of Abl in C. trachomatis
invasion can be explained by: (i) Abl kinase may phosphorylate a large fraction of EB-associated proteins, while PDGFR phosphorylates a subset of functionally redundant proteins, (ii) some of the proteins phosphorylated by Abl kinase may have multiple tyrosine residues (such as TARP), making Abl appear to be the major kinase, and/or (iii) not all Abl targets play a role in entry.
To our knowledge, Vav2 has not been previously implicated in C. trachomatis
infection. We have previously observed co-localization of Cortactin with inclusions 
. While this manuscript was in preparation, Hybiske et al
reported that depletion of Cortactin results in a modest decrease in entry 
, and Carrabeo et al
demonstrated that WAVE2 and Arp3 colocalize with EBs and are required for entry 
. We now link these signaling cascades involving Rac and Arp2/3 activators to upstream events that include EB binding to and activation of PDGFR as well as activation of Abl kinase. We conclude that activation of Abl kinase and PDGFR are necessary to ensure efficient recruitment and activation of downstream signaling molecules, including WAVE2, Vav2, and Cortactin, which mediate actin polymerization and entry.
The exact relationship between Abl activation, phosphorylation of the putative type III secreted effector TARP, entry, and vacuole formation is likely to be complex. Though our findings demonstrate that Abl kinase phosphorylates TARP, inhibition of Abl kinase by several different methods did not prevent entry. One explanation for this result is, as suggested by others 
, that TARP phosphorylation is not required for entry; instead, it could be important for a post entry function, such as inclusion trafficking and/or fusion. This observation could explain why the significantly diminished phosphorylation of EBs in Abl/Arg−/−
cells did not affect entry, especially given that TARP encodes many tyrosine residues that could serve as putative phosphorylation sites. Alternatively, TARP phosphorylation was not completely abolished in the Abl/Arg−/−
cells, suggesting that other tyrosine kinases, such as Src family kinases, may target TARP. This residual TARP phosphorylation could be sufficient to mediate bacterial entry. Interestingly, Src family kinases can be activated by PDGFR signaling and regulate cytoskeletal dynamics (reviewed in 
), thus the PDGFR-Src pathway could function redundantly with Abl kinase to promote entry.
We speculate that Abl-dependent phosphorylation of TARP may provide docking sites for recruitment of additional SH2 containing signaling molecules that may increase the efficiency of entry. This could include setting up a positive feedback pathway similar to what has been recently reported for Tir phosphorylation during pedestal formation by Enteropathogenic Escherichia coli 
. Initial phosphorylation of TARP would lead to recruitment of additional kinase molecules as well as new TARP molecules, culminating in the recruitment of actin and actin polymerizing factors 
. Alternatively, by virtue of its four polyproline motifs (PxxxP) 
, TARP could bind to the SH3 domains of both Vav2 and Abl kinase. This interaction would serve to bring Abl kinase in contact with Vav2 and ensure efficient Rac activation, a mechanism that has been proposed for Vav activation by the Murine gamma-herpesvirus 68 latency protein M2 
. It is also possible that key role of Abl kinase during entry is to phosphorylate and/or recruit other actin polymerization mediators (i.e. WAVE2, Vav2, and Cortactin).
The modulation of signaling pathways involving Abl kinase and bacterial internalization is an emerging theme among important human pathogens, including Shigella flexneri 
and Group B coxsackievirus 
. In contrast to these pathogens where Abl kinase is the only kinase required for entry, our findings demonstrate that entry can occur either via an Abl kinase-dependent pathway or through activation of PDGFRβ. Our results further suggest that these two pathways function in parallel and are thus functionally redundant. Since STI571 can inhibit both Abl and PDGFR kinases, this finding may explain why bacterial entry is diminished with this drug but unaffected in Abl/Arg−/−
cells. Although we cannot rule out the possibility that other STI571-inhibitable kinases play a role in C. trachomatis
entry, such as c-Fms and Lck kinase 
, these targets are not, to the best of our knowledge, expressed in HeLa or NIH 3T3 cells, 
). Furthermore, the fact that we can recapitulate the entry defect observed with STI571 by depleting Abl kinase and inhibiting PDGFR in the same cell using alternative means (ie. Treatment of Abl/Arg−/−
with AG1295, a specific PDGFR inhibitor) provides compelling evidence that these proteins are likely the main kinases affected by STI571 in the context of C. trachomatis
Our results are consistent with the following model for C. trachomatis entry into host cells (). C. trachomatis L2 binds to and activates PDGFRβ, possibly via heparan sulfate and/or PDGF. Abl kinase is recruited to and activated at the site of EB binding in a PDGFR-independent manner and phosphorylates TARP. Activation of PDGFR and Abl kinase leads to recruitment and activation of downstream targets, including Vav2, WAVE2, and Cortactin. Rac and Arp 2/3 are recruited to the site of entry. Actin polymerization is stimulated through the WAVE2/Arp2/3 pathway, Cortactin/Arp2/3 pathway, and/or directly by TARP. Since binding and entry of C. trachomatis was not completely abolished under our experimental system, this bacterium most likely utilizes other receptors and pathways that converge with these molecules to ensure efficient Rac and Arp2/3 activation for Chlamydia uptake. In addition, Abl-dependent TARP phosphorylation may contribute to critical events after entry that are required for this obligate intracellular parasite to survive and replicate in the host cell.
Model for C. trachomatis binding and internalization.
In summary, we have applied a genome-wide RNAi-based forward genetic screen to discover that C. trachomatis hijacks PDGFRβ and Abl kinase to modulate the host cytoskeleton in order to efficiently to enter host cells.