Usage of heparan sulfate, integrins, and FAK in HPV16 infection

Abstract

Human Papillomavirus Type 16 (HPV16) is the major causative agent of cervical cancer. Studies regarding the early binding and signaling molecules that play a significant role in infection are still lacking. The current study analyses the role of heparan sulfate, integrins, and the signaling molecule FAK in HPV16 infection of human adult keratinocytes cell line (HaCaTs). Our data demonstrate that infection requires the binding of viral particles to heparan sulfate followed by activation of focal adhesion kinase through an integrin. Infections were reduced in the presence of the FAK inhibitor, TAE226. TAE226 was observed to inhibit viral entry to the early endosome a known infectious route. These findings suggest that FAK can serve as a novel target for antiviral therapy.

Introduction

Human Papillomavirus Type 16 (HPV16) is a non-enveloped virus of the Papillomaviridae family. This family of double stranded DNA viruses have high tropism for squamous epithelial tissue and have been recognized as the etiologic agent for human cancers (Bosch et al., 2002; Bosch et al., 1995; Zur Hausen, 1991). HPV16 is the genotype most often associated with cases of invasive cervical carcinoma (Bosch et al., 2002).

HPV16 infection begins with the attachment of the viral particle (virion) to the target cells. This attachment step has been suggested to be mediated by heparan sulfate and to be followed by a secondary binding event, putatively an integrin complex (Evander et al., 1997; Giroglou et al., 2001; Joyce et al., 1999; McMillan et al., 1999; Shafti-Keramat, 2003). Following attachment and binding to the putative secondary receptor at the cell surface, HPV have been shown to be internalized via several pathways including clathrin dependent, caveolin, or clathrin-caveolin independent pathways (Bousarghin et al., 2003; Day, Lowy, and Schiller, 2003; Hindmarsh and Laimins, 2007; Laniosz et al., 2009; Laniosz, Holthusen, and Meneses, 2008; Smith, Campos, and Ozbun, 2007; Spoden et al., 2008). An explanation for the various findings may be the cell type used, and the method of virions production. Clathrin-mediated endocytosis pathway has been shown using cell lines that include C-127 cells, COS-7 cells, and HaCaTs cells, alongside a battery of compounds, dominant negatives and genetic approaches (Bousarghin et al., 2003; Day, Lowy, and Schiller, 2003; Laniosz, Holthusen, and Meneses, 2008). Our previous findings show that post clathrin-mediated endocytosis, HPV16 viral particles traffic to a caveolin-1 positive vesicle and particles can be found in the endoplasmic reticulum (Laniosz, Holthusen, and Meneses, 2008). A role for dynamin in HPV16, and HPV31 infection, presumably via pinching of vesicles from the plasma membrane has been described (Abban, Bradbury, and Meneses, 2008; Smith, Campos, and Ozbun, 2007). Recently, Spoden et al. described the clathrin-, caveolin-, and dynamin-independent entry for HPV16 using HeLa cells and 293TT (Spoden et al., 2008). In this study, the authors showed the involvement of tetraspanin-enriched domains in HPV16 endocytosis.

Emerging data on the initial steps of viral infection have shown that viruses that bind to heparan sulfate and integrin complexes at the cell surface, activate cellular signaling molecules including focal adhesion kinase (FAK), PyK2, Src kinase, Rho and Rac1 GTPases, and phosphatidylinositol 3-kinase (PI3-K) (Fothergill and McMillan, 2006; Krishnan et al., 2006; Marsh and Helenius, 2006; Sharma-Walia et al., 2004). These viruses include human herpes virus 8 (HHV-8), HIV, and the JCV virus. HHV-8 has been shown to induce both FAK phosphorylation and PI3-K (Naranatt et al., 2003; Sharma-Walia et al., 2004). In 1997, Davis et al., showed that the HIV virus induces PyK2 phosphorylation (Davis, 1997a; Davis, 1997b). In addition, JCV and HHV-8 viruses have been shown to activate the Ras/MAP kinase pathway (Naranatt et al., 2003; Payne et al., 2001; Querbes, 2004). Signaling studies in HPV have shown that HPV31, and HPV16 binding and entry can result in the activation of a tyrosine kinase and PI3-Kinase signaling event (Fothergill and McMillan, 2006; Schelhaas et al., 2008; Smith, Lidke, and Ozbun, 2008). These signaling events may help cytoskeletal rearrangement and filopodia formation to promote HPV viral uptake from the extracellular matrix (ECM).

The goal of our study was to determine the attachment, secondary binding, and early signaling molecules that may be required in the infectious entry route of HPV16 virus in the human keratinocyte cell line, HaCaTs, a non-tumorigenic keratinocyte cell line derived from adult skin that have a normal differentiation capacity (Boukamp et al., 1988; Boukamp et al., 1997; Breitkreutz et al., 1998). We have observed that HPV16 pseudovirions (HPV16 PsVs) are dependent on heparan sulfate for initial attachment, preferentially infect cells in the presence of α6 integrin receptor, and induce the phosphorylation, and thus activation of FAK. Our experiments showed that FAK phosphorylation is necessary for virus entry to the early endosome. Using TAE226, a small molecular bis-anilino pyrimidine compound that inhibits FAK and PyK2 phosphorylation, we observed a significant decrease in HPV16 infection. In addition infection analysis performed in FAK knock-out cells and matched controls substantiated the biological significance of FAK activity. Our data also showed that HPV16 binding induces filopodia formation on cells, possibly enhancing virus entry as previously reported for HPV31 (Smith, Lidke, and Ozbun, 2008).

Materials and Methods

Cells

HaCaT cells originally derived from the lab of Dr. Fusenig (The German Cancer Research Center, Heidelberg, Germany) were given as a gift from Dr. Ozbun (The University of New Mexico School of Medicine, Albuquerque, NM)(Boukamp et al., 1988). The primary mouse embryonic fibroblasts FAK -/- DU3 and their matched control FAK +/+ DU17 cells described by Ilic et al, (University of California at San Francisco) were a generous gift from the Chandran Laboratory (Rosalind Franklin University of Medicine and Science (RFUMS) North Chicago, IL)(Ilic et al., 1995). Cells were cultured in Dulbecco's Modified Eagle's media (DMEM) supplemented with 10% Fetal Bovine Serum (FBS, DMEM-10), 100 IU/ml penicillin and 100 μg/ml streptomycin. KH-SV keratinocytes from a healthy donor, and BOUA-SV α6 integrin negative keratinocyte cells from a patient with Pyloric Atresia–Junctional Epidermolysis Bullosa (PA-JEB) were immortalized using simian virus 40 (laboratory of Guerrino Meneguzzi), and shared with us by Dr. Christensen (Penn State University, Hershey, PA)(Gache et al., 1998). Cells were maintained in KGM media (Lonza) containing KGM SingleQuots, bovine pituitary extract, human epidermal growth factor, insulin, hydrocortisone and gentamicin/amphotericin B selection (KH-SV) or G418 (400ug/ml) selection (BOUA-SV α6 integrin negative).

Determination of the number of viral particles

The signaling events induced at the plasma membrane are a result of the number of viral particles binding to the cells surface receptors. In our experiments, infection is detected by the expression of GFP at a multiplicity of infection (MOI) of 0.15, i.e., 0.15 infectious units/cell. Using real-time PCR we determined the number of genomes in our preparation to be 8×10⁷ viral particles or 12 DNA containing viral particles per cell when using an MOI of 0.15. To measure all the viral particles used, we measured the amount of L1 protein in our viral preparation by comparing L1 levels to a control amount of BSA in a Coomassie stained gel. We measured 8×10⁹ viral capsids in 0.15 MOI, or 1,200 viral particles per cell/experiment.

HPV16 pseudovirion (PsV) production and purification

HPV16 PsVs were made in 293TT as described by Buck et al (Buck et al., 2004). 293TT cells, bicistronic HPV16 L1 and L2 plasmid pShell, and the GFP cDNA plasmid 8fwb were a generous gift from Drs. Day and Schiller (NCI/NIH Bethesda, MD). In brief, pShell and 8fwb were transfected into 293TT cells, harvested, and after high salt extraction, PsV were purified on an optiprep gradient (27-39%). The viral fractions used were analyzed under negative staining electron-microscopy, and by Coomassie staining on a sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE). Viral titer (MOI) was determined by measuring the percentage of GFP positive cells, i.e., infected HaCaT cells after 48 hours, by FACS analysis. The number of genome containing viral particles used in experiments was quantified by real time PCR. Total number of viral particles (empty particles, and genome containing particles) was measured by determining the L1 protein levels on a Coomassie stained SDS-PAGE. ³[H] thymidine labeled PsV was generated by the addition of ³[H] thymidine to the cultures during viral production.

Antibodies and reagents

phospho-FAK (pY397), FAK, phospho-Src (pY418), and Src antibodies were purchased from Upstate Biotechnology (Lake Placid, N.Y); and anti-mouse integrin α6 (CD49f) functional blocking monoclonal antibody was purchased from Chemicon International. F58-10E4 anti heparan sulfate monoclonal antibody was purchased from a Seikagaku Corporation subsidiary, Associates of Cape Cod, Massachusetts. Lysophosphatidic acid (Lee et al.) was purchased from Sigma (St. Louis, MO) and PP2 Src family kinase (SFK) inhibitors were purchased from Calbiochem. Purified TAE226 was generously supplied by Novartis Pharma, AG, Switzerland. TAE226 was dissolved in DMSO (Sigma Aldrich) to a concentration of 10mmol/L, stored at -20°C. Working dilutions of TAE226 at varying concentrations of 0.1, 0.5, 1.0, 2.0, 2.5, 3.0, 4.0, 5.0, and 6.0 uM were made by resuspending stock solution in DMEM-10 before use. In experiments, control cells were incubated with an equal amount of DMSO diluted in DMEM-10.

Inhibitor studies

HaCaT cells were pre-incubated at 37°C with indicated doses of PP2, PP3 or TAE226 for 1 hour prior to the addition of virus. The inhibitors were maintained on cells throughout the experiment unless indicated.

Cell viability assay

The cell toxicity of inhibitors used in experiments was determined using the Cell Titer-Glo luminescent cell viability assay (Promega, Madison, WI). HaCaT cells in a 12-well tissue culture dish at 100,000 cells per well were incubated with increasing concentrations of inhibitor at 37°C for 48 hours. At 48 hours post-incubation, supernatants were collected, cells harvested with trypsin and added to their corresponding supernatants. Cell viability was assessed in accordance with the manufacturer's recommendations of the luminescent cell viability assay kit by Promega, Madison, WI. Luminescence is an indicator of the amount of ATP in the sample. Cytotoxicity of the inhibitor was measured by a statistically significant drop in ATP levels relative to that of untreated cells. Each concentration was tested three times, and the error bars represent standard deviations. Cell viability was also confirmed using a trypan blue exclusion assay.

Detection of FAK and Src phosphorylation

500,000 HaCaT cells plated on a 10 cm² dish at 37°C in DMEM-10 were serum starved in DMEM for 48 hours. Cells were then infected with HPV16 PsV typically at MOI of 0.15 in order to achieve infection of approximately 15% cells, unless indicated. At different time points, HPV16 infected cells were harvested. Cells were lysed in cold 0.5% Nonidet P-40 (NP-40) buffer (0.5% NP-40, 150mM NaCl, 10 mM Tris pH 7.5, 5 mM EDTA, and 1X protease inhibitor cocktail (PIC, GE Healthcare)). Lysates were clarified at 13,200 × g at 4°C for 10 minutes and protein concentration determined using a bicinchonic acid (BCA) protein kit (Pierce, Rockford, Ill). Equal amounts of Protein lysates were denatured at 95°C for 5 minutes, resolved on an 8% SDS-PAGE, and transferred to nitrocellulose membrane. Membranes were immunoblotted with primary antibodies pY397-FAK, pY418-Src and anti-actin at 1:1000 overnight at 4°C. Anti-mouse and anti-rabbit Alexa Fluor 680 nm and 800nm secondary antibodies were used at 1:20,000 for 30 minutes at room temperature. Equal protein loading was also confirmed using anti-FAK and anti-Src (total FAK or Src) antibodies at 4°C overnight. Odyssey Imaging System and software (Li-Cor, Lincoln, NE) were used to scan and analyze band intensity.

Confocal microscopy colocalization studies

HaCaT cells at 100,000 cells per well were plated on a 6 well plate containing glass coverslips, and incubated overnight in DMEM-10 at 37°C (Fisher Scientific, Piscataway, NJ. Catalog #12-545-80). After 24 hours, cells were serum-starved in DMEM for 48 hours at 37°C. Post 48 hours, HPV16 PsV at an MOI of 0.15 was added to cells and incubated for the described length of time at 37°C. Cells were fixed for 10 minutes in 4% paraformaldehyde in PBS at 4°C and washed 3x in 1X PBS. Cells on coverslips were permeabilized with blocking buffer (0.2% fish skin gelatin (Sigma), 0.2% Triton X-100 in PBS) for 5 minutes. Coverslips were washed 3x in 1X PBS and incubated with mouse anti-HPV16.V5 antibody (provided by Dr. Christensen), the early endosome marker goat anti-EEA1 (Santa Cruz, California), Alexa-Fluor 488 phalloidin (Molecular Probes (Invitrogen), Carlsbad, CA), and mouse anti-pFAK (Millipore, Upstate). Primary antibodies were used at a dilution of 1:100 and incubated on cells for 1 hour. The fluorescence labeled Alexa-Fluor secondary chicken anti-goat 488 and donkey anti-mouse 594 antibodies were used at a dilution of 1:2,000. All dilutions were made in blocking buffer. Nuclei were stained with TOPRO-3 (1:1000 dilution in blocking buffer) (Invitrogen), and coverslips were mounted using Prolong Gold Anti-Fade (Invitrogen). Fluorescence confocal microscopy was performed on an Olympus Fluoview 300 inverted confocal microscope and image analysis was done using Fluoview software (Olympus, Melville, NY) at the microscopy core of RFUMS. Z-stacked images were taken at 0.75 micron sections of the cell in an x, y, and z plane (7 sections/z stack).

Wound-healing model assay

HaCaT cells were plated in serum-free media at 300,000 cells per well. 24 hours later cells were at near confluence, a plus sign was manually scraped in the monolayer of cells using a sterile 10 ul pipette tip. After scrape, cells were washed 3x with 1X PBS to remove any floating cells. Inhibitors or control vehicle was added for the described length of time, areas of gap were marked at greater than four matched pairs of gap, and then gap closure was measured using the ruler tool on the Nikon inverted microscope TE-2000S with a 4X objective lens. Images were captured and measured using Metamorph imaging software. Area of gap closure was calculated by the percentage change in area at greater than four data points and calculated by the formula: [area of scratch wound at time zero (minus) area of scratch wound after 24 or 36 hours]/100.

Measuring of HPV16 PsV binding

To measure virus binding, ³[H] thymidine labeled virus was incubated with a monolayer of cells on ice at 4°C for 2 hours. Unbound virions were washed off by washing 4x with 1X PBS. Cells were harvested by scraping, centrifuged at 10,000 × g for 10 minutes and resuspended in 20 ul 1X PBS. Cells were spotted onto Whatman paper, allowed to dry for 10 minutes. Radioactivity was precipitated with 5% Trichloroacetic Acid (TCA) (Fisher Scientific) and counted in a Beckman Coulter scintillation counter.

Removal of cell surface heparan sulfate using Heparinase I

Heparinase I purchased from Sigma, St. Louis, MO, was reconstituted in digestion buffer (50mM HEPES, 50mM sodium acetate, 150mM NaCl, 9mM CaCl₂ and 0.1% BSA (pH 7.0)). Cultures were washed 1x with digestion buffer, treated with increasing units of heparinase I for 1 hour at 37°C,and then washed thoroughly with 1X PBS.

Determining integrin profile of cells

HaCaT cells, KH-SV and BOUA-SV α6 integrin negative cells were plated in 10cm² tissue culture plates at 2×10⁶ cells per plate. After 48 hours at 80% confluence, cells were detached with 5mM EDTA at pH 7.4 for 10 minutes at room temperature. Integrin expression levels were analyzed using an integrin mediated cell adhesion fluorimetric array (Chemicon International, Catalog number: ECM 535). Integrin array kit recognizes extracellular epitopes of human α and β integrins. In this kit, the wells of a 96 well plate contain immobilized monoclonal human integrin antibodies used to capture cells expressing integrins on their surface. Unbound cells were washed off and adherent cells on the plate were lysed and integrins were detected using molecular probes CyQuant GR dye. Absorbance was determined at 540-570 nm on a microtiter plate reader.

HPV16 Infection Assay

HaCaTs, KH-SV, BOUA-SV, DU17 or DU3 cells were plated at 50,000 in a 24 well plate. After 24 hours, HPV16 PsVs were added to cells, and bound for 2 hours on ice. Unbound virions were removed by washing 2 times (2x) with appropriate media. Cultures were then incubated at 37°C for 48 hours. Infection was measured as the percentage of GFP-transduced cells because the PsV encapsidated plasmid 8fwb encodes the GFP cDNA. Infections were analyzed by FACS at RFUMS flow cytometry core. FACS analysis was done using the CellQuest Pro software (BD Biosciences).

Flow cytometry analysis of infection

Cells infected with PsVs in the presence or absence of inhibitors were harvested and washed 3x in 1X PBS. Cells in PBS were processed by flow cytometry on a FACS-Calibur sorter (Becton-Dickinson, San Jose, CA) to determine infection efficiency based on GFP fluorescence.

siRNA knockdown of α6 integrin in KH-SV

KH-SV cells at 50% confluence on a 6 well dish (approximately 400,000 cells) were transfected with 1uM Accell siRNA in 750ul Accell delivery media per well. After 48 hours cells were re-transfected with 1uM Accell siRNA in Accell delivery media. Transfected cells were incubated for 24 hours and infected with HPV16 PsVs. HPV16 infection went for 48 hours. The level of α6 integrin knockdown was determined via flow cytometry analysis of cells using the α6 integrin pre-conjugated PE-Cy 5 Rat anti-human antibody CD49f (BD Biosciences). IgG2a PE-Cy 5 Rat anti-human isotype antibody was used to control for non-specific fluorescence. HPV16 infections on cells transfected with α6 siRNA or control siRNA were monitored via flow cytometry analysis.

Determining level of HPV16 infection by (GFP transgene) real-time RT PCR

Total RNA was isolated from infected or uninfected cells using Tri Reagent Solution, Ambion. HPV16 8fwb GFP expression was detected by real-time RT-PCR using GFP gene specific primers and Sybr Green real-time PCR mix from Applied Biosystems. The samples were normalized to 18s RNA gene expression.

Electron microscopy analysis of viral particles

5ul of optiprep purified HPV16 PsVs and the mock viral preparation were added to formvar-carbon-coated 400 mesh copper grids for 5 minutes. Grids were stained with 4% Phosphotungstic Acid (PTA) for 1 min and washed with distilled water for 30 seconds. We allowed grids to dry and negative stained micrographs were digitally acquired using a JEOL JEM-1230 transmission electron microscope at 80,000 magnification.

Statistical Analysis

We performed all experiments at least three times with each condition in triplicate. Error bars shown represent data means +/- SD. Data were analyzed for significance as needed using paired t-tests seeking 95% confidences.

Results

Initial binding steps of HPV16 PsVs on human keratinocyte immortalized cell line HaCaT

We wanted to determine if heparan sulfate at the cells surface was mediating initial PsV attachment. Our data showed that the binding of radioactive labeled HPV16 PsVs (i.e., DNA incorporating tritiated thymidine, ³[H]-HPV16 PsV) to HaCaTs (Fig 1 A, middle lane) is blocked in the presence of soluble heparin (Fig 1 A, last bar). Because addition of soluble heparin to PsVs may result in the coating of the PsVs which may prevent the binding of the virus to a non-heparan sulfate molecule, we removed cell surface heparan sulfate molecules using heparinase I (Fig S 1). Our data indicated that the enzymatic removal of heparan sulfate decreased the level of binding by over 70% (a statistical significant change) (Fig 1 B lanes 4-7). Decreased binding was not observed with increased heparinase I treatment, suggesting that either heparinase I treatment is less than 100% efficient (supported by the FACS data shown in Fig S 1A), or that there is heparan sulfate independent binding of our radiolabeled PsVs to the HaCaT cells. Next we determined the level of infection using PsVs previously incubated with soluble heparin, and the infection level of PsVs on heparinase I treated cells (Fig 1 C). Data demonstrated that pre-incubation of virions with soluble heparin blocked infection (Fig 1 C Lane 3, as compared to uncoated PsV Lane 2), while pre-incubating the cells with soluble heparin prior to addition of PsV (cells were washed prior to PsV addition) did not interfere with infection (Fig 1 C lane 7). These data suggested that soluble heparin was coating or masking epitopes on the viral capsid needed for binding. Infection was also reduced if cells were heparinase I treated prior to infection (Fig 1 C lanes 4-6). As with the binding assay, increasing the amount of heparinase I beyond 2 units did not decrease infection further. This suggested that either heparinase I treatment was less than 100% efficient or that a non-heparan sulfate binding event can lead to infection.

Fig 1. Heparan sulfate proteoglycans role in HPV16 binding and infection.

(A) HaCaT cells were incubated for 2 hours only with purified ³[H] thymidine labeled HPV16 reporter virions or with labeled virions previously incubated with 50ug/ml soluble heparin. Bound radioactivity was precipitated with TCA and counted by scintillation counter. First bar shows background cellular cpm in the absence of virus, second bar represents virions cpm in the absence of soluble heparin, and third bar represents bound virions in the presence of soluble heparin. (B) Binding of ³[H] thymidine labeled HPV16 to HaCaT cells pre-treated with heparinase I. Radioactivity was precipitated with TCA and counted on a scintillation counter. Counts: Bar 1 buffer background; bar 2 control untreated HaCaTs; bar 3 total bound virions; bars 4-7 bound virions cpm in heparinase I treated cells with 1U, 2U, 4U and 10U respectively. Each sample was plated with 1×10⁵ cells. Experiments were done in triplicate, and error bars represent sample means +/- SD. Significance level was set at *p<0.05. (C) Cells were infected with non-radioactive HPV16 GFP reporter PsVs for 48 hours. Samples: Bar 1 uninfected HaCaTs; bar 2, virus infection; bar 3 HaCaT infected with HPV16 GFP reporter PsVs pre-incubated with 50ug/ml soluble heparin; bars 4-6 represent infection of HaCaT cells pre-treated with increasing units of heparinase I prior to addition of HPV16 GFP reporter PsVs; bar 7 represents soluble heparin treated cells prior to addition of HPV16 GFP reporter PsVs. Each condition was performed in triplicate and error bars represent sample means +/- SD. Significance level was set at *p < 0.05.

Integrin profile of human keratinocytes

Post-initial interaction of PsVs with heparan sulfate, it is hypothesized that a conformational change occurs leading to a secondary binding event (Selinka et al., 2003). This secondary binding has been suggested to involve an integrin molecule, and a likely candidate is the α6 integrin molecule (Evander et al., 1997; McMillan et al., 1999). Thus, before addressing the role of integrins in virus infection in HaCaT cells, KH-SV and BOUA-SV α6 integrin negative cells, we determined the integrin profile of these cells. We assayed for existing integrins at the cell's surface using an integrin profiling kit (a modified Elisa), and we used FACS analysis to determine the level of α6 (α6 is not part of the profile kit). Levels of integrins were compared to background levels. We determined that HaCaT cells contained β1, α2, β2, α3, β4, β6, possibly αV (Fig 2 A1) and α6 (Fig 2 A2). KH-SV and BOUA-SV α6 integrin negative cell lines had above background levels of β1, α2, β2, α3, α5, β6, and possibly αV (Fig 1 B1, C1). KH-SV cells also had above background levels of α1 and β4. We confirmed the presence and absence of α6 integrin in these latter cell lines (Fig 2 B2 and C2, respectively).

Fig 2. Integrin Profile of HaCaTs, KH-SV α6 +/+ and BOUA-SV α6 -/- cells.

Integrin profiling of HaCaTs (A1), KH-SV α6 +/+ (B1) and BOUA-SV α6 -/- cells (C1) was performed using the Alpha/Beta Integrin mediated cell adhesion array combo kit. Baseline controls (cells alone) were used for standardizing and measuring the increase in relative fluorescence units for each cell type. Flow cytometry analysis of α6 expression on HaCaTs (A2), KH-SV α6 +/+ (B2) and BOUA-SV α6 -/- cells (C2).

Infection is lowest in α6 integrin null cells

We proceeded to compare the infection levels in the three human keratinocyte cell lines using HPV16 PsVs (Fig 3). Data showed that infection was most efficient in HaCaT cells (42%) as compared to KH-SV α6 cells (18%) and BOUA-SV α6 negative cells (6%). The correlation between α6 levels and infection suggested that α6 may be important in infection. Also, the low level of infection observed in α6-null cells suggested that infection may not be not 100% dependent on α6 integrin.

Fig 3. HPV16 infection in human keratinocyte lines.

HPV16 GFP reporter PsVs were added to HaCaTs, KH-SV α6 +/+, and BOUA-SV α6 -/- cells. Percent infection was determined via flow cytometry analysis. Infection of HaCaT cells bar 1, KH-SV α6 +/+ cells bar 2, BOUA-SV α6 -/- cells bar 3. Error bars show the standard deviation of three experiments in which 10,000 cells were analyzed for GFP expression to obtain the percent of infected cells. Significance level was set at *p < 0.05.

HPV16 PsVs induce FAK phosphorylation upon binding to HaCaTs

We wanted to determine if a signal transduction event was induced upon PsVs binding to target cells because of the proposed role of signaling events in virus entry and infection (Fothergill and McMillan, 2006; Krishnan et al., 2006; Sharma-Walia et al., 2004). The primary activated molecule upon integrin binding has been shown to be FAK; activated FAK can be differentiated from quiescent FAK by means of phosphorylation at -Tyr³⁹⁷ (David J. Sieg, 1999; Krishnan et al., 2006). In order to determine if HPV16 induced FAK phosphorylation, we analyzed the levels of phosphorylated FAK (pFAK-Tyr³⁹⁷) after virus binding to serum starved HaCaT cells (Fig 4). Cells were serum starved in order to normalize the levels of pFAK-Tyr³⁹⁷. Although the protein levels of total FAK (tFAK) were unchanged, FAK phosphorylation (i.e., pFAK-Tyr³⁹⁷) increased 3.3 fold and 5.0 fold after 10 and 30 minutes respectively post PsV binding as compared to control uninfected cells (Fig 4 A lanes 2, 3 as compared to lane 1). We detected increases in pFAK-Tyr³⁹⁷ level at an MOI of 0.15, but did not observe an increase using an MOI of 0.015 (data not shown) or 0.0015 at 5-30 minutes (Fig 4 A lanes 4-7). Lysophosphatidic Acid, a known mitogenic activator, induced FAK phosphorylation at -Tyr³⁹⁷ in HaCaTs after 10 minutes (Fig 4 A lane 8). Using an MOI of 0.15 we addressed the kinetics of FAK phosphorylation during early infection (Fig 4 B and C). As before, we observed an increase in pFAK-Tyr³⁹⁷ as early as 5 minutes and determined that there was a biphasic increase in pFAK-Tyr³⁹⁷ with peaks at 10 minutes and 30 minutes. These data showed that binding of HPV16 PsV induced an increase in pFAK-Tyr³⁹⁷ in a time dependent manner. Significance was assessed using a two-tail T-test at *p<0.05 (Fig 4 B asterisk compared to no virus control lane 1). pFAK-Tyr³⁹⁷ fold induction varied somewhat from experiment to experiment, but were always statistically significant and had similar biphasic kinetics (n=5). All pFAK-Tyr³⁹⁷ values were normalized against total FAK and actin.

Fig 4. HPV16 induces FAK-Tyr³⁹⁷ phosphorylation in HaCaT cells.

(A) Cells were incubated with MOI 0.0015 HPV16 PsV for 5, 10, 15 and 30 minutes (Lanes 4-7 respectively) and MOI 0.15 HPV16 PsV for 10 and 30 minutes (Lanes 2 and 3 respectively); uninfected cells are shown in lane 1. Control LPA (20 ng) induced FAK-Tyr³⁹⁷ is shown in lane 8. Equal amounts of NP-40 lysates resolved by 8% SDS- PAGE, were blotted for phosphorylated FAK-Tyr³⁹⁷ (pFAK blot), total FAK (tFAK blot), and actin. (B) Graphical representation of temporal pFAK-Tyr³⁹⁷ induction evaluated by Odyssey densitometry of western blot shown in C. Uninfected cells assigned a value of 1 for comparison lane 1. (C) HaCaT cells were infected with HPV16 (MOI 0.15) at the indicated times post-infection. Cell lysates were resolved on an SDS-PAGE and incubated with pFAK-Tyr³⁹⁷, total FAK (tFAK) and Actin antibodies. Protein levels were measured using the Odyssey densitometer. (D) Lysates from uninfected HaCaT cells (lane 1), mock-infected HaCaTs with 8fwb mock viral fraction (lane 2), or HPV16 PsVs (lane 3) were resolved on an SDS-PAGE and incubated with phosphorylated FAK-Tyr³⁹⁷ (pFAK), total FAK (tFAK) and Actin. Uninfected cells are assigned a value of 1 for comparison with values of pFAK fold increase. Protein levels were measured using the Odyssey densitometer. Significance level was set at *p< 0.05.

To confirm that the pFAK-Tyr³⁹⁷ induction observed were not due to non-virion contaminants, i.e., a non-viral protein in our optiprep purified HPV16 (L1/L2/8fwb PsVs) preparation, we performed identical experiments with a control “mock” virus production. The cells for the mock virus production fractions were transfected with only 8fwb, i.e., L1 and L2 proteins were not expressed and thus no PsVs were produced. PsVs and mock preparation were analyzed in a Coomassie blue stained SDS polyacrylamide gel and negative stain electron microscopy (EM) (Fig S 2). As is evident, there are non-viral proteins visualized in both PsV and mock PsV preparation by Coomassie staining, but there are no viral particles observed in the mock preparation by EM. We used these fractions to determine if FAK phosphorylation was specific to PsV containing fractions, and indeed FAK phosphorylation only occurred in PsV containing fractions (Fig 4 D Lane 3 as compared to Lane 2). All batches of viral preparations used in pFAK-Tyr³⁹⁷ induction were similarly evaluated for non-specificity and purity of viral preparation (n=5).

FAK inhibitor TAE226 prevents HPV16 pseudovirion induced FAK-Tyr³⁹⁷ phosphorylation

Interaction of cell surface integrins with the ECM induces recruitment of FAK from the cytoplasm leading to the autophosphorylation of FAK -Tyr³⁹⁷. In vitro, bis-anilino pyrimidine FAK inhibitor TAE226 has been shown to inhibit extracellular matrix-induced autophosphorylation of FAK-Tyr³⁹⁷ (Ta-Jen Liu, 2007). In order to confirm that HPV16 infection was indeed inducing an increase in pFAK-Tyr³⁹⁷, we analyzed pFAK-Tyr³⁹⁷ levels and cellular pFAK-Tyr³⁹⁷ expression in the presence of TAE226 (courtesy of Novartis). We first analyzed the cytotoxicity levels of TAE226 on HaCaT cells (Fig 5 A), and determined that a loss of cell viability was observed at greater than 5.0uM of TAE226 (*p< 0.05). Our experiments on HaCaT cells were thus performed with the nontoxic doses of 0.1, 1.0 and 2.0uM of TAE226. The induction of pFAK-Tyr³⁹⁷ (Fig 5 B lanes 2-5) was not observed in the presence of the three TAE226 concentrations (Fig 5 B lanes 6, 7, 8). Induction of pFAK-Tyr³⁹⁷ at 30 minutes was confirmed using 20% FBS as previously described (Fig 5 B lane 9) (Krishnan et al., 2006).

Fig 5. Phosphorylation of FAK is blocked in HaCaT using TAE226.

(A) Cytotoxicity of TAE226 was not observed on HaCaTs up to 2 uM concentration. Increasing concentrations of 0um, 0.1uM, 1.0uM, 2.0uM, 5.0uM and 10uM were tested. Cell viability was assessed after 48 hours as a function of the level of ATP. Doses of 5.0uM and 10.0uM were cytotoxic to the cells (Significance level was set at *p<0.05). (B) HaCaT cells were infected with HPV16 virus (MOI 0.15) for: 5’, 10’ 15’ and 30 minutes in the absence of TAE226 (lanes 2-5), in the presence of TAE226 at 0.1, 1.0 and 2.0 uM of (lanes 6-8); or incubated with 20% FBS for 30 minutes (lane 9). Equal amounts cell lysates were run on an 8% SDS-PAGE. Levels of FAK phosphorylation (pFAK-Tyr³⁹⁷), total FAK (tFAK), and actin blots (actin) were determined using corresponding antibodies. Values were normalized based on tFAK values which were normalized to actin control. Uninfected cells are assigned a value of 1 for comparison with values of pFAK fold increase. The percent inhibition of FAK phosphorylation (Bottom of lanes 6-8) was determined by dividing the decrease in FAK phosphorylation in inhibitor treated cells post 30 minutes infection to FAK induction post 30 minutes infection in the absence of inhibition. (C-N) Visualization of FAK phosphorylation by immunofluorescence confocal microscopy. In panels E, G, I, K and M cells were infected with HPV16 MOI 0.15 for 0, 10, 30, 60 minutes and 4 hours respectively. In panels F, H, J, L, and N cells pre-incubated with 2.0uM of TAE226 were infected for 0, 10, 30, 60 minutes and 4 hours respectively. Immunofluorescence was performed with anti-pFAK-Tyr³⁹⁷; nuclei are visualized with Topro-3 (blue). Confocal images were obtained using a 60X oil objective. Control uninfected cells stained for pFAK and nuclei are shown (C, D).

FAK has been shown to selectively localize at focal adhesion sites on the plasma membrane and upon ligand binding to an integrin receptor, FAK is phosphorylated and thus activated at Tyr³⁹⁷ (Hildebrand, Schaller, and Parsons, 1993). We employed immunofluorescence assays to observe pFAK-Tyr³⁹⁷ upon virus binding and infection. In this assay, HaCaT cells were infected with HPV16 PsVs for 10 minutes, 30 minutes, 60 minutes or 4 hours at 37°C or virus was added to cells on ice to allow virus to bind for 2 hours (0 minutes). We showed that in uninfected cells there was minimal to no detected pFAK-Tyr³⁹⁷ in HaCaT cells in the presence or absence of TAE226 (Fig 5 C, D). Binding of HPV16 PsVs did not induce a significant change in the levels of pFAK-Tyr³⁹⁷ in the presence or absence of TAE226 (Fig 5 E, F). The levels of pFAK-Tyr³⁹⁷ detected in infected cells in the absence of TAE226 increased with time and peaked at 30 minutes (Fig 5 G, I, K) and decreased to what appears to be basal levels after 4 hours (Fig 5 M). In contrast we did not observe an increase in pFAK-Tyr³⁹⁷ in cells infected in the presence of TAE226 (Fig 5 H, J, L, N). These immunofluorescence experiments showed that indeed TAE226 was able to interfere with FAK phosphorylation, a finding that parallels our levels of FAK phosphorylation in the protein western blot analysis of HPV16 infected cells pre-treated with TAE226.

Inhibition of FAK activation but not Src family kinase inhibition interferes with viral infection

One of the most well studied signal transduction events post-FAK activation involves recruitment and activation of Src family kinase members (SFK), primarily pp60^c-src kinase (pSrc) (Parsons, 2003; Schaller et al., 1994). To determine if Src was induced by HPV16 PsV binding, we assessed the level of phosphorylated Src (pSrc-Tyr⁴¹⁸) upon virus binding. Our data showed that at PsV MOI of 0.15 there is an increase in Src-Tyr⁴¹⁸ phosphorylation upon virus binding that can be prevented in the presence of SFK inhibitor PP2 (data not shown). Having determined that HPV16 PsV induced FAK-Tyr³⁹⁷ and Src-Tyr⁴¹⁸ phosphorylation, we wanted to determine the biological significance of FAK and Src phosphorylation. FAK or SFK inhibitor (TAE226 or PP2, respectively) pre-treated HaCaT cells were infected with HPV16 PsV containing the green fluorescent protein (GFP) reporter plasmid and harvested after 48 hours. As shown in Fig 6 A, there was a dose-dependent decrease in infection in the presence of TAE226 with a significant 25% decrease in infection at 0.1uM compared to untreated cells at 0uM (*p< 0.02). Increasing TAE226 concentrations to 0.5uM and 1.0uM showed no significant increase in percent inhibition (p=0.28 and 0.4 respectively) when compared to cells treated with 0.1uM of TAE226. However, at 2.0uM we saw a 53% inhibition of infection compared to untreated cells (*p< 0.003) and a significant increase in percent inhibition compared to cells treated at 0.1uM (** p < 0.015). Although increasing doses of TAE226, 4uM and 6uM, gave a continued decrease in infection, these higher inhibitor concentration levels were 10% and 23% more cytotoxic to the cells compared to 2uM TAE226 treated cells (data not shown). Similar infection experiments in the presence of 10 uM of the SFK inhibitor PP2 had no effect on the percent infection with PP2 (Fig. 6 B, bar 4) as compared to untreated cells (Fig. 6 B, bar 2). Although we observed a decrease in infection with 20uM of PP2 (Fig. 6 B, bar 5), it was not statistically significant and 20uM PP2 was cytotoxic to cells in a cell viability assay (data not shown). An inactive PP2 analog, PP3, had no effect in infection at 10uM (Fig 6 B, bar 6). In these experiments, HaCaT cells treated with 2uM of TAE226 again showed a significant 57% decrease in infection, (Fig. 6 B, bar 3, *p<0.05) compared to untreated cells (Fig. 6 B, bar 2). These data demonstrated that TAE226 reduced the level of infection and that PP2 did not, suggesting that FAK and not SFK members were important for HPV16 PsV infection.

Fig 6. FAK phosphorylation influences HPV16 infection.

(A) TAE226 inhibits HPV16 PsV infection in a dose dependent matter. HaCaT cells were infected with HPV16 PsV carrying the GFP cDNA in the presence of 0uM, 0.1uM, 0.5uM, 1.0uM and 2.0uM TAE226. Infection was measured by FACS analysis for GFP positive cells. Error bar represents the standard deviations of the mean from three separate experiment, *p<0.02, **p<0.015 (Student's t test). (B) HPV16 PsV infection level was decreased with TAE226, but not by the Src Family kinase inhibitor PP2. The inactive PP2 analogue PP3 was used as control. Cells were infected with HPV16 GFP reporter PsVs, and percent infection was analyzed by flow cytometry. Error bar represents the standard deviations of the mean from three separate experiments (in triplicate). Significance was determined by Student's t test and p-value was set at *p< 0.05. (C) FAK expressing (DU17) and FAK knockout DU3 MEFs cells infected in the presence or absence of the FAK kinase inhibitor TAE226 (2uM). Cells were infected with HPV16 GFP reporter PsVs and percent infection was measured by flow cytometry. Error bar represents the standard deviations of the mean from three separate experiments (in triplicate). Significance was determined by Student's t test and p-value was set at *p<0.05. (D) FAK expressing MEFs (DU17) (uninfected control in lane 1) were infected with HPV16 PsV (lane 2), incubated with 20% FBS (lane 3), or infected with HPV16 in the presence of an anti-α6 funtional blocking integrin antibody (lane 4). No FAK expression was detected in FAK knockout cells (lanes 5-8). Lanes 1-8, NP-40 lysates were run on an 8% SDS-PAGE and levels of phosphorylated FAK-Tyr³⁹⁷ (Top bands), total FAK (middle band) and actin blots (lower band). Loading values were normalized based on total FAK expression, which was normalized to actin values. Uninfected cells FAK phosphorylation levels are assigned a value of 1 (lane 1) for comparison.

Infection in FAK knockout cells is decreased

To further address the role of FAK in HPV16 infection and for a better understanding of a FAK mediated infection, we addressed the efficiency of infection in FAK null cells. Because human FAK negative cells are currently not available we employed the use of FAK-/- DU3 MEFs and matched control FAK +/+ DU17 MEFs. We have previously shown that infection of MEFs is efficient using HPV16 PsVs (Dabydeen and Meneses, 2009). HPV16 infection in DU3 cells was significantly less compared to infection of DU17 cells (Fig 6 C first and second bars, *p<0.05) but the lower level of infection in DU3 cells suggests that there may be a FAK-related molecule that may partially compensate for the role of FAK. Infection of DU3 MEFs treated with TAE226 resulted in a significant decrease in infection (Fig 6 C last bar) *p<0.05), as compared to TAE226 treated DU17 or un-treated DU3 cells. Because PyK2, a non-receptor tyrosine kinase, has been shown to be expressed and phosphorylated in DU3 cells, and a second target of TAE226 (Cheshenko et al., 2005; Krishnan et al., 2006; Sieg et al., 1998) we are currently investigating if PyK2 can compensate for the role of FAK during PsV infection.

Blocking α6 integrin function prevents FAK-Tyr³⁹⁷ phosphorylation

Our data suggested that α6 integrin was important for infection and that FAK phosphorylation (and thus activation) upon PsV binding was induced and necessary for infection. To determine if there was a link between α6 integrin function, PsV binding and FAK activation, we performed an experiment in the presence of an α6 integrin functional blocking antibody in MEFs. FAK phosphorylation was observed at 10 minutes post infection, Fig 6 D (lane 2), and with the positive control 20% FBS, Fig 6 D (lane 3) as compared to uninfected cells in lane 1. Addition of an α6 integrin functional blocking antibody prevented the FAK phosphorylation in the presence of PsV (Fig 6 D lane 4). This demonstrates that the α6 integrin plays a role in mediating FAK phosphorylation and that blocking α6 function in the presence of PsV blocks FAK phosphorylation. FAK knockout MEFs are shown as control for FAK staining (Fig 6 D, lanes 5-8). We have also tested the ability of blocking FAK phosphorylation with functional antibodies against integrin receptors α3 and α2; we did not see a decrease in FAK phosphorylation (data not shown).

TAE226 prevents closure of gap in a wound healing model

Infection of HPV has been suggested to be supported by a wounded epithelium (Roberts et al., 2007). The process of wound healing requires cell migration (Mattila and Lappalainen, 2008), and FAK kinase activity has been demonstrated by Sieg and colleagues to play an important role in cell migration (Sieg, 1999). Thus, we wanted to determine if inhibiting FAK phosphorylation (i.e. activation) will inhibit migration of HaCaT cells, and if this migration loss was responsible for our observed loss of infection. In the next set of experiments, we assessed the process of migration in an in-vitro “wound healing” model in the presence and absence of the FAK inhibitor TAE226. In this wounding assay, HaCaT cells plated to confluence were wounded (scraped) in the form of a plus sign with a 10 ul pipette tip, as has been previously described (Slack-Davis et al., 2007). Inhibitors were added to the cells, and the percentage of wound closure was monitored at 24 and 36 hours. The closure of the gap in control DMSO treated cells was compared to TAE226 treated cells (Fig 7). Wound (scrape) area in control (DMSO treated cells) cells was 75% closed after 24 hours and more than 90% after 36 hours (Fig 7 panels A-C, and J cells alone bars). Addition of TAE226 at a non-toxic dose of 2uM was able to block wound healing by greater than 95% for up to 36 hours (Fig 7 D-F and J-TAE226 bars).

Fig 7. TAE226 inhibits gap closure in HaCaT wound closure model.

(A-I) HaCaT monolayers grown to near confluence were incubated in the presence or absence of 2.0uM TAE226 +/- HPV16 PsV at the time of wounding. DMSO was used as a vehicle control. Images were obtained by light microscopy with a 4X objective lens. (J) Percentage gap closure represented as a bar graph (black bars represent gap closure at 24 hours, grey bars represent gap closure at 36 hours). Error bars represent the standard deviations of the percent gap closure measured within a marked area of at least 4 data points. Gap closure was calculated by the percentage change in area at greater than four data points and calculated by the formula: (area of scratch wound at time zero) – (area of scratch wound after 24 or 36 hours)/100. Significance was determined by Student's t test and p-value set at p<0.05, * significance at 24 hours, **significance at 36 hours.

Because HPV31 was shown to induce the formation of filopodia upon binding to target cells and filopodia are necessary to aid in closing the wound via formation of adheren junctions between epithelial cells (Mattila and Lappalainen, 2008; Smith, Lidke, and Ozbun, 2008), we tested if the addition of HPV16 PsV to TAE226 pre-treated cells would induce cell migration that would counteract the loss of migration observed in the presence of TAE226. Our data showed that in this model of wound healing, TAE226 prevented migration that would lead to gap closure and addition of HPV16 PsV did not induce gap closure in the presence of TAE226. Addition of HPV16 PsV had no effect on the size of the gap in the presence of TAE226 (Fig 7 G-I, and J-TAE226 with virus bar) or in the absence of TAE226 (data not shown).

HPV16 PsV induce filopodia formation

Because wound healing requires filopodia formation and HPV16 PsV did not affect wound closure we wanted to assess if HPV16 indeed induced filopodia formation as HPV31. As compared to cells alone or to cells in the presence of TAE226 (Fig 8 A and B respectively), the induction of filopodia is highly increased with the addition of HPV16 PsV at 10 minutes (Fig 8 C, white arrow). There is a subsequent decrease in the induction of filopodia in the presence of virus at 30 minutes (Fig 8 E, white arrow). Although we observed PsVs at the plasma membrane in the presence of TAE226, we did not detect any filopodia formation in the presence of TAE226 (Fig 8 D and F, yellow arrow shows virus and actin overlap). These data demonstrate that inhibiting FAK phosphorylation can prevent HPV16 PsV induced filopodia formation. We noticed that in the absence of filopodia formation (i.e. in TAE226 treated cells), PsV co-localized with actin was observed at the plasma membrane (compare Fig 8 C to D and E to F, yellow arrow shows overlap of actin and PsV). These results (Fig 8 D and F) provide evidence that TAE226 does not interfere with virus binding to the plasma membrane suggesting that endocytosis may be inhibited.

Fig 8. TAE226 inhibits filopodia formation but not virus binding.

(A and B) Mock infected HaCaT cells in the absence and presence of 2.0uM TAE226 respectively. (C and E) HaCaT cells infected with virus for 10 and 30 minutes in the absence of inhibitor, respectively. (D and F) HaCaT cells pre-incubated with 2.0uM TAE226 infected for 10 and 30 minutes respectively. Actin is viewed using phalloidin-488 (green); HPV PsVs are stained with anti-HPV16.V5 antibody; actin overlapping with HPV PsVs is observed (yellow). Filopodia extensions (white arrows) and virus colocalizing with the phalloidin stained actin (yellow arrows) are shown. Nuclei are visualized with TOPRO-3 (blue).

TAE226 prevents HPV16 endocytosis into early endosomes

Because our data showed that TAE226 was not interfering with PsV binding to HaCaT cells, they suggested that TAE226 was interfering with endocytosis, and thus infection. We analyzed endocytosis of PsV in confluent and sub-confluent monolayers (Fig 9 and S3). Following HPV16 PsVs attachment to target cells, PsVs have been shown to be endocytosed into early endosomes. TAE226 did not prevent the binding of viral particles to HaCaTs (Fig 8 D and F, red arrow), but did prevent the overlap of PsVs with the early endosome marker EEA1 (green arrow) in both sub-confluent and confluent HaCaT monolayers (Fig 9 and S3 respectively). HPV16 PsV staining overlaps with EEA1 at 10, and 30 minutes, and is observed for the first 4 hours (Fig 9 E, G, I, K, and in Fig S3). In contrast, we did not observe any overlap of HPV16 PsV staining with EEA1 in TAE226 treated cultures (Fig 9 F, H, J, L, and in Fig S3). These data strongly support the hypothesis that in addition to preventing filopodia formation, TAE226 is blocking the initial step of HPV16 PsV endocytosis into the early endosomes.

Fig 9. TAE226 prevents HPV16 trafficking to the early endosome.

Control uninfected HaCaT cells without (A) or with TAE226 (B). HPV16 binding to HaCaT cells (C) is not affected by TAE226 (D). PsVs internalization in the absence (E, G, I, K) or presence (F, H, J, L) of TAE226 at the corresponding time. Cells were fixed, permeabilized and stained for the early endosome marker EEA1 (green), and HPV16 PsV with anti-HPV16.V5 antibody. Nuclei are visualized with TOPRO-3 (blue). HPV16 PsVs (red arrows) did not colocalize with EEA1 (green arrows) at 10, 30, 60 minutes or 4 hours in TAE226 treated cells. Confocal images were obtained using a 60X oil objective.

α6 integrin surface expression is necessary for HPV16 infection of KH-SV cells

Having shown that an α6 integrin functional blocking antibody prevented FAK phosphorylation, and that blocking FAK phosphorylation interfered with HPV16 infection, we addressed infection in the presence of α6 integrin targeted siRNA. In our control cells (Fig 10 C) the mean fluorescence intensity for α6 protein staining (measured by FACS) was 22,469, whereas the mean fluorescence intensity for surface expressed α6 integrin was decreased to 3,537 in α6 integrin siRNA transfected cells (Fig 10 E). Infection of α6 siRNA transfected cells (Fig 10 F) was reduced by over 90% as compared to infection of control cells (Fig 10 D). Control IgG, and uninfected cells are shown for comparison (Fig 10 A, and B respectively). Next, infection was measured by real time RNA-RT-PCR using GFP mRNA as target. Graphical representation of infection analysis is shown with the addition of siRNA only cells to ensure no background signal due to siRNA transfection. These data clearly show that a drop in α6 integrin surface expression results in a significant loss of infection.

Fig 10. siRNA Knockdown of α6 integrin expression decrease HPV16 infection in KH-SV cells.

(A) IgG isotype staining of control uninfected KH-SV cells; (B) Flow cytometry analysis of control uninfected KH-SV cells; (C-F) KH-SV cells were double transfected with α6 siRNA or control siRNA, after 72 hours cells were infected with HPV16 PsVs and level of infection measured by flow cytometry. Panels: (C) α6 integrin expression level in control cells; (D) level of HPV16 infection in control cells; (E) α6 integrin expression level in α6 siRNA expressing cells; (F) level of HPV16 infection in α6 siRNA expressing cells; (G) real time RT-PCR analysis for the expression of GFP mRNA (HPV16 PsV encapsidate GFP cDNA plasmid). Graph shows real-time GFP RT-PCR of infected and uninfected cells after 48 hours. HPV16 GFP expression in the absence of siRNA treatment was considered as 100%. No signal was detected in α6 siRNA transfected cells in the absence of HPV16 PsV infection (Last bar). Each condition was performed in triplicates.

Discussion

The findings described in this manuscript provide evidence that HPV16 initially attaches to HaCaT cells via a heparan sulfate proteoglycan (HSPG) and subsequently mediates a signal transduction event through the α6 integrin and FAK.

Our data have shown that HPV16 reporter virions binding to an immortalized keratinocyte cell line can be blocked by soluble heparin and decreased by heparinase I treatment of cells, thus supporting the need for HSPGs for viral attachment. For instance, Girouglou et al., in looking at the infectious entry pathway of HPV33 pseudovirions and HPV16 binding have shown that initial virion binding to COS-7 cells involves the HSPG receptor (Selinka, Giroglou, and Sapp, 2002) and Joyce and colleagues have shown that the L1 conserved major capsid protein of HPV11 binds to HSPGs (Joyce et al., 1999). In our experiments, soluble heparan blocked binding by over 90% and removal of heparan sulfate by heparinase I decreased binding by 75-80%. This decrease in binding corresponded to a 75-80% drop in infection. Possible reasons why we cannot attain a 100% infection block includes: the possible inefficiency of heparinase I; the inability of soluble heparin to mask all binding sites on every infectious particle; and maybe there is a low level of heparan sulfate independent infection in our culture system. Our data supports the animal studies using a cervicovaginal challenge mouse in-vivo model that demonstrated that heparinase III blocked HPV16 infection and binding (Johnson et al., 2009). It has been well documented that neutralization of HPV infection can be achieved after viruses bind to the cell surface (Day et al., 2007). This suggests that a secondary event is needed for internalization. In this manuscript our data showed that this possible secondary event is the binding of the viral particle to an integrin, and in HaCaTs, α6 may be the most critical. Using siRNA targeting α6 mRNA we were able to decrease surface expression of α6 integrin receptor, including complete loss of detection in about 30% of integrin positive KH-SV cells. This decrease was able to reduce infection by more than 90%. The remaining infection can be attributed to the inability of 100% removal of surface expressed α6 integrin receptor or to a non-α6 dependent infection event.

The role of the α6 integrin receptor in HPV16 infection led us to demonstrate for the first time that the FAK molecule is phosphorylated upon HPV16 binding to a keratinocyte cell line. The data demonstrated FAK phosphorylation as early as 5 minutes, suggesting that phosphorylation of FAK may be one of the initial signaling molecules activated upon ligand binding to HaCaT cells. The FAK phosphorylation was detected at Tyr-³⁹⁷, a site of autophosphorylation. We did observe an increase in Src phosphorylation at Tyr-⁴¹⁸, a phosphorylation event often observed after FAK activation. Although FAK and Src were both phosphorylated and thus activated, only FAK was demonstrated to be biologically important. HPV16 infection was significantly decreased in the presence of TAE226 treated cells, whereas infection levels did not decrease with the addition of a Src family kinase (SFK) inhibitor, PP2. We also showed that in the genetically altered MEFs, HPV16 induction of FAK was blocked in the presence of an α6 integrin functional antibody, suggesting that indeed FAK activation by HPV16 was dependent on α6 integrin. Although our data suggested that TAE226 prevented wound closure and filopodia formation to the same extent as PP2, we saw that inhibiting cell migration and filopodia formation is not enough to prevent infection because PP2 did not reduce infection. This suggested that TAE226 was interfering with endocytosis, and our viral entry immunofluorescence studies confirmed that entry to the early endosomes was inhibited by TAE226.

Our data is consistent with studies on the infectious pathway of HPV infection on other cell lines that had identified the HSPGs and α6β4 as attachment factors and putative receptors respectively (Giroglou et al., 2001; Joyce et al., 1999; McMillan et al., 1999; Rommel et al., 2005; Shafti-Keramat, 2003). The internalization of viruses that utilize HSPG and integrin receptor binding suggested a role for FAK and Src signaling in entry (Cheshenko et al., 2005; Krishnan et al., 2006). The role of FAK has been implicated in Herpes Simplex Virus (HSV) entry (Cheshenko et al., 2005) and human herpesvirus 8 (HHV8) both enveloped DNA viruses, and although FAK phosphorylation has been observed in the Adenovirus (a non-enveloped virus) infection of human corneal fibroblasts, its role has yet to be defined (Natarajan et al., 2002). Our data suggest that α6 integrin is mediating a signaling event that involves FAK phosphorylation and that interfering with FAK function or α6 expression can reduce or prevent infection.

Based on the findings presented in this manuscript, we hypothesize that post HSPG binding, HPV viral particles interact with α6 integrin leading to FAK phosphorylation. FAK phosphorylation is then responsible for activating downstream molecules necessary for endocytosis and intracellular trafficking. We are exploring the possibility that PyK2 may in part contribute to HPV16 infection. There exists the possibility that HPV16 entry may be partly through non-clathrin/caveolae mechanisms, e.g., tetraspanins (Spoden et al., 2008). Although the tetraspanin link has been shown in non-keratinocyte cell line thus far, the role of integrins would still be applicable because tetraspanins form complexes with integrin receptors (Berditchevski and Odintsova, 1999). We are addressing if there is a link between tetraspanins, integrins, and FAK phosphorylation during HPV 16 PsV infection.