SHP-2 Mediates Cryptosporidium parvum Infectivity in Human Intestinal Epithelial Cells

The parasite, Cryptosporidium parvum, induces human gastroenteritis through infection of host epithelial cells in the small intestine. During the initial stage of infection, C. parvum is reported to engage host mechanisms at the host cell-parasite interface to form a parasitophorous vacuole. We determined that upon infection, the larger molecular weight proteins in human small intestinal epithelial host cells (FHs 74 Int) appeared to globally undergo tyrosine dephosphorylation. In parallel, expression of the cytoplasmic protein tyrosine phosphatase Src homology-2 domain-containing phosphatase 2 (SHP-2) increased in a time-dependent manner. SHP-2 co-localized with the C. parvum sporozoite and this interaction increased the rate of C. parvum infectivity through SH2-mediated SHP-2 activity. Furthermore, we show that one potential target that SHP-2 acts upon is the focal adhesion protein, paxillin, which undergoes moderate dephosphorylation following infection, with inhibition of SHP-2 rescuing paxillin phosphorylation. Importantly, treatment with an inhibitor to SHP-2 and with an inhibitor to paxillin and Src family kinases, effectively decreased the multiplicity of C. parvum infection in a dose-dependent manner. Thus, our study reveals an important role for SHP-2 in the pathogenesis of C. parvum. Furthermore, while host proteins can be recruited to participate in the development of the electron dense band at the host cell-parasite interface, our study implies for the first time that SHP-2 appears to be recruited by the C. parvum sporozoite to regulate infectivity. Taken together, these findings suggest that SHP-2 and its down-stream target paxillin could serve as targets for intervention.


Introduction
The coccidian protozoan parasite, Cryptosporidium parvum, is a major public health concern known to cause outbreaks of gastroenteritis (cryptosporidiosis) worldwide [1]. Based on morbidity, mortality, and economic burden data, the World Health Organization (WHO) has listed the recruitment of the tyrosine phosphatase, SHP-2, is essential for efficient infection and raises the possibility that phosphatases may play an important role in the C. parvum infection process.
Viable C. parvum oocysts (Iowa isolate) were originally purchased from Waterborne Inc. (New Orleans, LA), and continually propagated at the USEPA facility in Cincinnati, OH, using immunosuppressed CF-1 mice. The mice were purchased from Charles River Laboratories, Inc. The animals were allowed to acclimate for 1-3 days, and given food and water ad libitum. All procedures in this study were done in compliance with the US EPA Institutional Animal Care and Use Committee regulations. For immunosuppression, the mice were given an alternative day dosing of dexamethazone and tetracycline in the drinking water. After immunosuppression for 7 days, mice were dosed with 10 5 oocysts, per oral administration using a ball-end dosing needle. Animals were maintained in raised stainless steel mesh grid floors, and their feces were collected every 36 hours. Each fecal collection was processed by sieving and multiple flotations, as previously described [20]. Mice were euthanized using CO 2 gas, containing at least 5% O 2 . Purified oocysts were stored at 4°C in reagent-grade water containing 100 U/ml penicillin and 100 μg/ml streptomycin. All oocyst lots were used within two months of shedding.
For experiments using immunofluorescence assays, approximately 6 x 10 5 FHs 74 Int cells were seeded onto 18 mm round coverslips into each well of a 12-well Costar tissue culture plate (Corning, Inc., Corning, NY), and cultured for 24 h. For experiments using protein expression assays, cells were cultured in 150 cm 2 peel-back Techno Plastic Products (TPP) cell culture flasks for 24 h. Prior to treatment with inhibitor or oocysts, all cells were serum-and EGFdeprived overnight. Cells were stimulated with specific inhibitor, vehicle, or media (negative control) for 2 h prior to inoculation with oocysts. Prior to inoculation onto the monolayer, oocysts were pre-treated for excystation, as described above. Cells were cultured under standard conditions at 37°C and 5% CO 2 during the infection process and treatment with inhibitors.

Immunofluorescence analysis of entry
Infection was stopped at 2.5 h post-inoculation by removing the culture medium. Non-adherent sporozoites and oocysts were removed by washing the infected and non-infected (control) cells three times with cold 1X HBSS. Cells were then fixed onto the coverslip with 4% paraformaldehyde for 10 min, permeabilized with 0.25% Triton X-100 (Sigma, St. Louis, MO) for 10 min, and blocked using fresh 1% bovine serum albumin (BSA) in 1X HBSS. Intracellular parasites were labeled with biotin-conjugated Vicia villosa lectin (VVL) (Vector Laboratories, Burlingame, CA) as previously described [22] at a 1:333 dilution of VVL in 1% BSA. After 1 h incubation, the coverslips were washed several times, followed by a second 1 h incubation with Cy5-conjugated streptavidin. For staining of SHP-2, the slides were then washed again several times and stained with 1:500 dilution of antibodies to SHP-2, followed by more washes and a fourth 1 h incubation with Cy3-conjugated anti-mouse IgG antibodies.
An LSM 700 confocal microscope (Zeiss, Jena, Germany), was used to detect Cy3 and Cy5 emissions (492/510 nm and 650/670 nm respectively). Serial Z-stack images (0.45 microns thick) were captured and Zen 2010 software was employed to compile and generate the 3D images. The Cy5 and Cy3 images (artificially colored in red and green, respectively) were overlaid, resulting in C. parvum appearing red, host cell protein appearing green, and any co-localization appearing yellow. Stained slides were screened for the presence of infection at several magnification levels including 400X, 630X, and 1000X. For quantification of parasites, the number of intracellular C. parvum were counted per field of view at 200X.

Protein purification and Western blot
After 2-6 h of infection time, flasks were washed three times with cold 1XHBSS to remove non-adherent sporozoites and oocysts. Cells were lysed and solubilized using RIPA lysis and extraction buffer (Life Technologies, Inc., Carlsbad, CA) supplemented with Halt protease and phosphatase inhibitor cocktail (100X) (Life Technologies, Inc). Lysates were clarified by centrifugation at 13,000xg for 10 min at -9˚C, followed by quantification using BCA Protein Kit (Life Technologies, Inc.). Proteins were resolved by SDS-PAGE using 4-15% Mini-PROTEAN1 TGX™ precast gels (BioRad, Hercules, CA), and transferred onto a low fluorescence PVDF membrane. The membranes were then blocked by incubation in Odyssey Blocking Buffer (Li-Cor, Inc., Lincoln, NE) for 1 h followed by incubation with the appropriate antibody diluted in the blocking buffer overnight. Detection was done using ECL Plex Cy3 or Cy5 conjugate secondary antibodies (GE Healthcare, Noblesville, IN), followed by excitation with a Typhoon Fluorescence Imager (GE Healthcare). Generated images were quantified using Image J software. Experiments were performed on at least three separate occasions.

SH2 domain profiling assay
An SH2 domain-based receptor tyrosine kinase (RTK) profiling kit (Signosis, Inc., Santa Clara, CA) was used to detect active SHP-2. The kit contained microtiter wells coated with the specific SH2 domains of several proteins, including SHP-2 (referred to as PTPN11 in this kit). Using an ELISA format, protein lysates with and without C. parvum infection were exposed to the immobilized proteins. Bound proteins were labeled using biotin-conjugated pTyr antibodies, followed by streptavidin-conjugated HRP for detection. Absorbance was measured spectrophotometrically at 450nm.

Statistics
The densitometry analyses of the Western blots are representative of three independent experiments. One-way ANOVA was used as a computation to test various means for equality, using a significance level of 0.05. In cases where the difference between all the means was found to be significant, a post-hoc analysis was done to do a pairwise comparison of the different means. For post-hoc analysis, the Bonferroni Multiple Comparisons Procedure was used and significance was set at either p<0.02 or p<0.05, as specified.

C. parvum infection promotes activation of phosphatases
The role of phosphatase activity during C. parvum infection was investigated in two different ways. One mechanism was to evaluate changes in total concentrations of tyrosine phosphorylated proteins before and after C. parvum infection. Here, we collected total whole cell lysates of non-infected and infected cells 2 h and 4 h after infection. Immobilization of these proteins and immuno-blotting with antibody to phospho-tyrosine revealed that after infection, proteins of higher molecular weights of 125kD to 37kD were less phosphorylated after exposure to C. parvum ( Fig 1A). Image density analysis reveals that in this higher molecular weight range, there is a 37% decrease in tyrosine phosphorylated proteins at 2 h, and a 70% decrease at 4 h, suggesting that tyrosine dephosphorylation is a progressive part of the C. parvum infection process, and continues up to 4 h after infection. Another mechanism to evaluate phosphatase activity was to determine the effects of the phosphatase inhibitor, sodium orthovanadate on C. parvum infectivity. When comparing host cells exposed to the phosphatase inhibitor and host cells without the phosphatase inhibitor, we observed that C. parvum infection sites per field of view decreased significantly (p<0.05) for cells with the inhibitor, even at concentrations as low as 0.1 mM (Fig 1B). The higher doses of phosphatase inhibitor of 5.0 and 10.0 mM resulted in a more significant decrease of infection sites compared to vehicle control (p<0.02), by approximately 47% and 81%, respectively.
The tyrosine phosphatase SHP-2 plays a role in C. parvum infection SHP-2, a non-receptor tyrosine phosphatase, is normally auto-inhibited, with the N-terminal SH2 domains interacting with its catalytic protein tyrosine phosphatase domain. Complexing of the N-SH2 with a pTyr-containing ligand reduces interaction with the phosphatase domain, causing the enzyme to switch to its active state [14]. To measure the relative increase or decrease of proteins that bind to the SH2 domain of SHP-2, we used an ELISA-based profiling kit, with the SH2 domains of SHP-2 immobilized in a well of a microtiter plate (Fig 2A).  Increased levels of absorbance represent increased availability of specific ligand to the SH2 domain of SHP-2. After 2 h of infection, mean absorbance significantly increased (p<0.05) from 0.066 ± 0.01 and 0.07 ± 0.02 in uninfected controls to 0.18 ± 0.04 in infected cells, indicating that C. parvum infected cells have an increased level of activated proteins that bind to the SH2 domain of SHP-2. Interestingly, total protein levels of SHP-2 also increase significantly at 2 h and 4 h of infection. However, this expression level starts decreasing by 6 h post infection (Fig 2B and 2C), indicating that SHP-2 is a regulated protein during C. parvum infection and increased levels of the protein is needed during the initial 4 h of infection. This finding coupled with the increased levels of activated binding partners to SHP-2, suggests that SHP-2 may be one of the many phosphatases that play a role in the tyrosine dephosphorylation seen in Fig 1A. SHP-2 co-localizes with the sporozoite at the C. parvum infection site SHP-2 preferentially localizes in the cytoplasm; however, it has been proposed that the binding of SHP-2 to a pTyr-containing ligand could promote subcellular relocalization of SHP-2 to the cell membrane [13]. Therefore, we determined the cellular localization of SHP-2 after C. parvum infection. Following infection, bright speckles of SHP-2 appeared on the background of cytoplasmic SHP-2 staining (Fig 2D). In addition, staining of C. parvum showed that the sporozoites co-localized to the SHP-2 speckles (Fig 2E and 2F). The co-localization was observed at approximately 95% of infections sites. To further characterize SHP-2 and sporozoite colocalization, three-dimensional images were generated by Z-stacking the series of confocal images from just below the basal region of the cell through to just above the apical region of the cell. As seen in Fig 2G, SHP-2 formed a cap over the sporozoite, with finger-like projections extending down the sides of the sporozoite. The specific localization of SHP-2 around the sporozoite suggests that SHP-2 is recruited early in the infection process and may be involved in the initial formation of the parasitophorous vacuole around the parasite.

SHP-2 activity is required for C. parvum infectivity
Although we had determined that SHP-2 expression and level of binding partners increased after infection to C. parvum, we did not know whether activity of SHP-2 is necessary for C. parvum infection. We examined the need for SHP-2 activity by using a SHP-2 specific chemical inhibitor, SHP-2 PTPase Inhibitor, to block the activity of SHP-2 in the host cell. Inhibitor doses of 300 and 700 μM, showed a significant reduction in the number of infection sites observed per field of view using immunofluorescence microscopy (p<0.05) by 53% and 67%, respectively (Fig 3A). In addition, while the total number of infection sites decreased in a dosedependent manner, the few infection sites which formed revealed that SHP-2 and the sporozoite remained co-localized even at the highest concentration of inhibitor tested (Fig 3B). The localization of SHP-2 around the parasite and the reduction of infection after exposure to a specific inhibitor of SHP-2, denotes that SHP-2 localization and phosphatase activity play key and necessary roles in the infectivity of C. parvum.
Paxillin is moderately dephosphorylated during C. parvum infection Paxillin, as an adaptor protein and a substrate of SHP-2, contains at least four tyrosine phosphorylation sites that can bind with SH2 domains [23]. Phosphorylation and de-phosphorylation of paxillin at these sites is thought to regulate cell adhesion, among many other cell activities. For example, the changes in the phosphorylation status of paxillin has shown to increase cell migration by decreasing cell adhesion [24,25]. Although there have been many proteins associated with the phosphorylation for paxillin, there are very few proteins identified SHP-2 Mediates C. parvum Infection that dephosphorylate paxillin. While paxillin is dephosphorylated by SHP-2 [16,26], the role of paxillin in C. parvum infection is unknown. Therefore, we determined the cellular localization and phosphorylation status of paxillin with/without C. parvum infection. After C. parvum infection, immunofluorescence staining for paxillin demonstrated that it remained localized in the cytoplasm (Fig 4A). Therefore, in contrast to that observed for SHP-2, paxillin did not localize with the sporozoite infection site. However, we did see a moderate, but statistically significant (p<0.05), dephosphorylation of paxillin, at the Tyr118 residue, as early as 2 h postinfection (Fig 4B and 4C). This moderate dephosphorylation continues even up to 6 h postinfection. On the other hand, total paxillin expression remains relatively unchanged up to 6 h. The lack of co-localization suggests that paxillin may not directly interact with C. parvum, but these observations suggest that moderate dephosphorylation of paxillin may play a role in the continued infection process of the parasite.

Possible downstream targets of SHP-2 during C. parvum infection
Since we noticed only a moderate decrease in paxillin phosphorylation, we hypothesized that SHP-2 may be acting on other downstream targets as well. Previous studies have shown that c-Src kinase, a substrate of SHP-2, is necessary for C. parvum infection [5], and thus we used the inhibitor, Saracatinib, which strongly inhibits Src kinase, and which has been shown to have inhibitory effects on paxillin [21]. Cells were treated with Saracatinib prior to infection, and infection patterns were monitored after 2 h (Fig 5A and 5B). Not only did low doses of the inhibitor significantly decrease infections per field of view (p<0.05), even a relatively low dose of 1.0 μM, reduced C. parvum infection sites by 87% (p<0.02). In addition, immunofluorescence staining confirmed the decrease in C. parvum infection (Fig 5B). Together, these observations imply that members of the Src family, possibly other than c-Src, may play a role in C. parvum infectivity, and that paxillin may also be a factor in infection.
Lastly, the role of SHP-2 in the dephosphorylation of paxillin was investigated. Because SHP-2 has been shown to cause dephosphorylation of paxillin, either directly or indirectly [13,14], it was hypothesized that exposure of host cells to SHP-2 inhibitor may cause a reversal of the phosphorylation status of paxillin. After exposure of host cells to various levels of SHP-2 inhibitor, we infected the host cells with C. parvum to see at which doses if any, a change was seen in phosphorylation status (Fig 6A and 6B). After infection, paxillin was moderately, but significantly dephosphorylated for the infected vehicle control, as expected, and the 150 μM concentration of the inhibitor (p<0.5). However in the presence of higher doses of SHP-2 PTPase Inhibitor (300 and 750 μM), paxillin phosphorylation levels increased moderately, close to the state observed in the uninfected controls, confirming that SHP-2 activity may have a moderate effect on the phosphorylation status of paxillin, following C. parvum infection. Since the change to phosphorylation in paxillin is moderate, SHP-2 may be impacting other proteins more significantly. These results combined with the data regarding decreased infection at the higher levels of SHP-2 inhibitor, suggest that even though paxillin does not co-localize with the infection site, paxillin is moderately dephosphorylated during infection, and that SHP-2 may play a role, whether direct or indirect, in the dephosphorylation of paxillin.

Discussion
In this study, we identify a novel role for the non-receptor protein tyrosine phosphatase, SHP-2, in the infection of C. parvum, a parasitic human pathogen (summarized in Fig 7). Upon infection, SHP-2 co-localizes with the C. parvum sporozoite, thereby increasing the rate of C. parvum infectivity through SH2-mediated SHP-2 activity. Concurrently, SHP-2 dephosphorylates paxillin to facilitate C. parvum infection. While paxillin does not directly interact with the sporozoite, it may play a role for infectivity through an unknown mechanism. This study lays the groundwork for further investigation of the mechanisms that regulate C. parvum infection.
A key aspect of this discovery is that SHP-2 localizes at the C. parvum infection site and that inhibition of this protein affects the ability of this parasite to infect the host cell. C. parvum, as an apicomplexan parasite, is an intracellular pathogen. The uniqueness of this parasite relies on its ability to remain extracytoplasmic by the formation of an electron dense-membrane  [27]. The formation of this vacuole requires the recruitment of several host cell proteins, including actin [28], however this is the first study that shows the role of a phosphatase in the initial infection process of C. parvum.
The molecular mechanisms of how SHP-2 initially interacts with C. parvum and what proteins might be involved in the recruitment of SHP-2 to the infection site has yet to be determined and our current efforts are focused on understanding the early developments of this interaction. It is known that microbial pathogens often inject virulence factors into the host cell upon invasion to start the process of infection. For example, the bacterial pathogen, H. pylori, releases the virulence factor, Cag A, which binds to and activates SHP-2 [29]. C. parvum attachment results in the release of several dense granules and the migration of proteins known as micronemes during the initial invasion process. Micronemal proteins are important in host cell localization and ligand-receptor attachment in apicomplexan parasites [30]. Even though just a few micronemal proteins have been identified (GP900, TRAP-C1/2, CpSCRP, and ABD), the roles of these proteins on infection of host cells are not fully understood [31,32]. It is possible that SHP-2 activation and recruitment is stimulated by one of the many C. parvum micronemal proteins.
Interestingly, SHP-2 has been shown to be a key host cell factor for virulence by other pathogens, namely, Helicobacter pylori, Kaposi's sarcoma-associated herpesvirus, enteropathogenic Escherichia coli, and Streptococcus pneumoniae [33][34][35][36][37]. SHP-2 is recruited and activated in these infection systems, and for some pathogens, expression levels of SHP-2 are upregulated to promote infection. However, the purpose for recruitment and activation of SHP-2 is pathogendependent. For example, H. pylori recruits SHP-2 during the early infection process, via a virulence protein named CagA [29,38]. On the other hand, for other pathogens, SHP-2 plays a role in pathobiology [39], and a role for suppressing the host immune response during  [36]. In this study, we show expression levels of SHP-2 increase within 2 and 4 hours of infection, and decline by 6 hours post-infection, inferring that SHP-2 upregulation is needed during the early time-points of C. parvum infection. This is corroborated by the evidence that SHP-2 is localized at the apical end and sides of the infection site, indicating that SHP-2 is needed in the initial infection process of the parasite.
Furthermore, we show that SHP-2 may affect the phosphorylation status of paxillin, a focal adhesion protein, during the infection process. We observe moderate, but significant dephosphorylation of paxillin as early as 2 hours post-infection. Paxillin has focal adhesion targeting regions known as LIM domains, actin binding domains, and phosphorylation sites specific for kinases and phosphatases, which allows it to act as a scaffolding protein that binds multiple proteins. In this study, we observed that the moderate dephosphorylation of paxillin at the Tyr 118 residue was rescued in the presence of increased SHP-2 inhibitor. Nevertheless, paxillin may not serve as a direct ligand of SHP-2. Previous studies have shown that Gab1, a pleckstrinhomology domain-containing docking protein, becomes tyrosine phosphorylated upon various intracellular signals and acts as a scaffolding protein by binding to paxillin and SHP-2 and causing the dephosphorylation of paxillin [26]. This theory is consistent with another study that found that the SH2 domains of SHP2 do not bind directly to phospho-peptides derived from paxillin [40]. The ability of SHP-2 to utilize scaffolding proteins may also provide a mechanism for relocalization in the cell.
Interactive partners of SHP-2 during C. parvum infection in vivo have yet to be determined. Nevertheless, based on this study and other studies relating to C. parvum pathogenesis, some inferences can be made. Integrins and Src kinases have been shown to interact with SHP-2 [41,42], and previous work in C. parvum infectivity has shown that integrin α2 and c-Src are important players in infection [5,9]. Consistent with these studies, work in our lab has shown that integrin expression is upregulated by almost 100% upon infection and Src family proteins are dephosphorylated in FHs 74 Int cells infected by C. parvum (data not shown). However, as with paxillin, we do not see any co-localization with integrin, suggesting that these proteins may function in other mechanisms of C. parvum pathogenesis, such as survival or cell adhesion. For example, upregulation of integrin might be needed for other purposes such as involvement in the multi-molecular complexes that are needed for focal adhesions linking the extracellular matrix (ECM) with the actin cytoskeleton. Indeed one study found that the forcedependent strengthening of integrin-cytoskeleton linkages requires down-regulation of focal complex dynamics by SHP-2 [17]. Another study found that the binding of integrins to the ECM induces tyrosine phosphorylation of SHPS-1 (SHP substrate) and its subsequent association with SHP-2 [41]. The reinforcement of the host cell to the ECM may provide a unique strategy of a pathogen to prevent epithelial cell turnover. Evidence of this phenomena can be seen in the bacterial pathogen Shigella, which uses its virulence factor, OspE, to prevent intestinal epithelial cell detachment by targeting integrin linked kinases to affect reinforcement of cell adhesions. Furthermore, upon OspE expression, host cells had decreased paxillin and FAK phosphorylation levels, which prevented intestinal epithelial cell detachment by targeting integrin-linked kinases to reinforce cell adhesions, and reduce epithelial cell turnover rate [43,44]. A similar mechanism may be occurring with C. parvum infection, as it is advantageous for the parasite to progress through its stages of infection before epithelial turnover.
Inhibition of SHP-2 is a promising approach to prevent C. parvum-induced pathobiology, as there are several potent and highly selective inhibitors to SHP-2 [45]. However, since SHP-2 is a widely-expressed cytoplasmic enzyme and involved in multiple cellular pathways, a targeted approach of inhibition to only infected cells would be a more pragmatic approach. Therefore, a better understanding of the initial events in recruitment and activation of SHP-2 is crucial to designing such an approach.
Our studies reveal a critical role for SHP-2 in the pathogenesis of C. parvum. We demonstrate for the first time that SHP-2 is recruited to the infection site, and necessary for infection. We further show that SHP-2 may play a role in the dephosphorylation of paxillin. More detailed investigations are required to improve the understanding of the regulation and mechanisms of SHP-2 activation, and whether the activation is through host cell proteins or through C. parvum virulence factors. Although there is still much to be learned about the mechanisms of pathogenesis of C. parvum, we have now taken the first steps to show that SHP-2 plays a pivotal role in C. parvum infection.