Learn more: PMC Disclaimer | PMC Copyright Notice
Graf1 Controls the Growth of Human Parainfluenza Virus Type 2 through Inactivation of RhoA Signaling
ABSTRACT
Rho GTPases are involved in a variety of cellular activities and are regulated by guanine nucleotide exchange factors and GTPase-activating proteins (GAPs). We found that the activation of Rho GTPases by lysophosphatidic acid promotes the growth of human parainfluenza virus type 2 (hPIV-2). Furthermore, hPIV-2 infection causes activation of RhoA, a Rho GTPase. We hypothesized that Graf1 (also known as ARHGAP26), a GAP, regulates hPIV-2 growth by controlling RhoA signaling. Immunofluorescence analysis showed that hPIV-2 infection altered Graf1 localization from a homogenous distribution within the cytoplasm to granules. Graf1 colocalized with hPIV-2 P, NP, and L proteins. Graf1 interacts with P and V proteins via their N-terminal common region, and the C-terminal Src homology 3 domain-containing region of Graf1 is important for these interactions. In HEK293 cells constitutively expressing Graf1, hPIV-2 growth was inhibited, and RhoA activation was not observed during hPIV-2 infection. In contrast, Graf1 knockdown restored hPIV-2 growth and RhoA activation. Overexpression of hPIV-2 P and V proteins enhanced hPIV-2-induced RhoA activation. These results collectively suggested that hPIV-2 P and V proteins enhanced hPIV-2 growth by binding to Graf1 and that Graf1 inhibits hPIV-2 growth through RhoA inactivation.
IMPORTANCE Robust growth of hPIV-2 requires Rho activation. hPIV-2 infection causes RhoA activation, which is suppressed by Graf1. Graf1 colocalizes with viral RNP (vRNP) in hPIV-2-infected cells. We found that Graf1 interacts with hPIV-2 P and V proteins. We also identified regions in these proteins which are important for this interaction. hPIV-2 P and V proteins enhanced the hPIV-2 growth via binding to Graf1, while Graf1 inhibited hPIV-2 growth through RhoA inactivation.
INTRODUCTION
Rho GTPases are members of the Ras superfamily of 20- to 30-kDa small GTPases. They are highly conserved in eukaryotes and act as molecular switches to regulate essential cellular functions. To date, at least 22 members of the Rho GTPases have been identified in mammalian cells (1, 2). The most well characterized members, namely, RhoA, Cdc42, and Rac1, affect a variety of cellular activities, including actin reorganization, apoptosis, intracellular trafficking, and cell polarity (1,–5). Rho GTPases regulate cellular activities by coordinating with other host proteins such as focal adhesion kinase (FAK) and Akt. It is important for viruses to establish an environment that facilitates their growth by controlling these cellular activities. Rho GTPases and their related proteins affect the life cycles of some viruses, including respiratory syncytial virus (RSV) (6, 7), Ebola virus (8), vesicular stomatitis virus (8), Epstein-Barr virus (9), influenza A virus (IAV) (10, 11), and rotavirus (12). The relationship between herpesvirus and Rho GTPases has been well investigated (13). We previously reported that Rho activation promotes syncytium formation induced by human parainfluenza virus type 2 (hPIV-2) (14). However, it remains unknown whether it also affects hPIV-2 growth.
hPIV-2 is an enveloped, single-stranded, negative-sense RNA virus which belongs to the genus Rubulavirus in the family Paramyxoviridae (15). Its genome contains six genes encoding NP, P, V, M, F, hemagglutinin-neuraminidase (HN), and L proteins. Both V and P proteins are produced from the P gene. They share an N-terminal domain but have distinct C-terminal domains due to mRNA editing (16). We previously reported the interactions of the NP, P, V, and L proteins and identified their interaction sites (17,–21). NP, P, and L proteins together with RNA genome form the ribonucleoprotein complex (RNP). Rubulavirus V proteins are found in virions, while other paramyxovirus particles generally contain little or no V protein (22), suggesting the importance of V proteins for the life cycles of rubulaviruses. Many studies have demonstrated that the V protein interacts with and counteracts several host proteins, including MDA-5 (23,–25), LGP2 (26), TRAF6 (27), STATs (28, 29), AIP1/Alix (19), and tetherin (30), most of which are important for the innate immune response. Since most of these host proteins interact with V within the C-terminal V-specific region, where seven Cys and three Trp residues are well conserved among paramyxoviruses (15), they do not interact with the P protein.
Rho GTPases are strictly regulated by guanine nucleotide exchange factors (GEFs) and GTPase-activating proteins (GAPs). GEFs convert the GDP-bound inactive form of Rho to a GTP-bound active form, while GAPs catalyze the formation of the GDP-bound state (31). It has been speculated that there are over 80 RhoGEFs and approximately 70 RhoGAPs in the human genome (32, 33). Some GEFs and GAPs appear to have Rho GTPase specificity, as shown by p112RhoGAP and ARHGAP6 which are RhoA specific, while some GEFs and GAPs can regulate the signaling of all representative Rho GTPases, including RhoA, Cdc42, and Rac1 (32). Signaling of the Rho GTPases is determined by the kind of GEFs and GAPs present, resulting in the strict regulation of various types of Rho signaling.
In the present study, we investigated whether Rho signaling affects the hPIV-2 life cycle and identified Graf1, a GAP, as a contributor linking Rho signaling and hPIV-2 growth.
MATERIALS AND METHODS
Cells.
Vero cells were grown in Eagle's minimal essential medium (MEM) supplemented with 10% fetal calf serum (FCS). HeLa, COS, BSR T7/5 (34), and HEK293 cells and their derivatives were grown in Dulbecco's modified Eagle's MEM (DMEM) containing 10% FCS. All cells were maintained in a humidified incubator at 37°C with 5% CO2.
Antibodies and reagents.
Monoclonal antibodies (MAbs) against hPIV-2 NP protein (306-1), V/P protein (315-1), V protein (53-1), P protein (335A), L protein (70-6-2), M protein (110A), HN protein (M1-1A), and F protein (144-1A) were described previously (18, 35,–37). An MAb to myc and polyclonal Abs to myc and hemagglutinin (HA) were obtained from MBL (Nagoya, Japan). Anti-actin MAb was purchased from Wako (Osaka, Japan). Anti-RhoA MAb was obtained from Cytoskeleton, Inc. (Denver, CO). Lysophosphatidic acid (LPA) was purchased from Enzo Life Sciences (Farmingdale, NY), while Y-27632 and NSC23766 were from Wako. The inhibitors ML141 and AZA1 were obtained from Sigma-Aldrich Co. (St. Louis, MO) and EMD Millipore (Billerica, MA), respectively.
Plasmids.
The hPIV-2 V, P, NP, and L genes and their mutants cloned into the mammalian expression vector pcDL-SRα296 (17, 36), pTM-1 (38), or pCI-neo (Promega, Madison, WI) were used. An expression vector, pTB701 carrying Graf1 cDNA, was a kind gift from Yoshitaka Ono (Kobe University, Kobe, Japan). cDNAs of green fluorescent protein (GFP) and of Graf1 and its deletion mutants were fused to a myc tag at the N terminus and ligated into another expression vector, pEF4/Myc-His (Invitrogen, Carlsbad, CA). All of these constructs were confirmed by DNA sequencing.
Virus titration.
Monolayers of cells were infected with hPIV-2 (Toshiba strain) at a multiplicity of infection (MOI) of 1 or 0.01 and incubated in MEM without FCS. At the indicated hours postinfection (hpi), virus titers were measured by plaque assay using Vero cells, as described previously (30).
RhoA activation assay.
Activated RhoA was detected using a RhoA Pull-down Activation Assay Biochem kit (Cytoskeleton, Inc.) according to the manufacturer's protocols. Briefly, cells grown to 70 to 80% confluence in six-well plates were starved for 24 h in serum-free medium. Cells were infected with hPIV-2 at an MOI of 1 for the times indicated on the figures. After cells were washed with phosphate-buffered saline (PBS), they were lysed in cell lysis buffer (50 mM Tris [pH 7.5], 10 mM MgCl2, 0.5 M NaCl, and 2% Igepal) supplemented with protease inhibitor cocktail (0.62 μg/ml leupeptin, 0.62 μg/ml pepstatin A, 0.14 mg/ml benzamidine, and 0.12 mg/ml tosyl arginine methyl ester) and centrifuged. For pulldown assays, cell lysates were reacted with rhotekin Rho binding domain (RBD) beads for 1 h at 4°C and washed with wash buffer (25 mM Tris [pH 7.5], 30 mM MgCl2, and 40 mM NaCl). Then, cells were suspended with sample buffer for sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE), followed by SDS-PAGE and immunoblot (IB) analysis using an anti-RhoA MAb.
Immunofluorescence analysis (IFA).
HeLa cells grown to 50% confluence in 12-well plates were transfected with the plasmids indicated on the figure using XtremeGENE HP (Roche, Indianapolis, IN) according to the manufacturer's procedures. At 8 h posttransfection (hpt), the cells were infected with hPIV-2 at an MOI of 0.1. After 48 h, the cells were fixed with 4% paraformaldehyde for 15 min at room temperature and rinsed three times with PBS. The cells were permeabilized with PBS containing 0.2% Triton X-100 for 15 min and washed three times with PBS. The cells were incubated for 60 min with the appropriate Abs and washed three times with PBS. The secondary antibody reaction was performed for 60 min with Alexa Fluor 488 goat anti-mouse IgG and Alexa Fluor 594 goat anti-rabbit IgG (Invitrogen). After cells were washed with PBS, they were analyzed with a fluorescence microscope.
Immunoprecipitation (IP) analysis.
COS cells grown in 12-well plates and BSR T7/5 cells grown in six-well plates were transfected with the plasmids indicated on the figures using XtremeGENE 9 (Roche) and XtremeGENE HP, respectively. At 48 hpt, cells were harvested and sonicated for 30 s three times in ice-cold lysis buffer containing 50 mM Tris-HCl (pH 7.5), 150 or 300 mM NaCl, and 0.6% NP-40 and centrifuged. IB and IP were performed as described previously (30, 39).
Establishment of HEK293 cells constitutively expressing hPIV-2 proteins or Graf1.
To obtain the HEK293 cell line which constitutively expresses hPIV-2 V (HEK293/V) or P (HEK293/P) protein, cells were transfected with pCI-neo expression plasmids containing the hPIV-2 V or P gene, using FuGENE 6 (Roche) according to the manufacturer's protocols. At 2 days posttransfection, the cells were transferred to 12-well plates and cultured for 3 weeks in DMEM containing 10% FCS and 1 mg/ml G418 (Thermo Fisher Scientific, Waltham, MA). To establish an HEK293 cell line which constitutively expresses Graf1 (HEK293/Graf1) and its control cell line (HEK293/ctrl), cells were transfected with an expression plasmid, pEF4/Myc-His, containing or not containing Graf1 cDNA using XtremeGENE 9 (Roche). The transfected cells were selected as described above using 600 μg/ml Zeocin (Invitrogen) instead of G418. Each clone was analyzed for protein expression levels, and clones that exhibited high expression levels were used in this study.
Depletion of Graf1 using shRNA.
To deplete Graf1 expression, a DNA fragment encoding anti-Graf1 short hairpin RNA (shRNA) was inserted into a pHygH1dTO vector (pHygH1dTO-shGraf1) (40). pHygH1dTO-shGraf1 targets the sequence 5′-GATATCTGTGCTGAATGGGAGATAA-3′ (corresponding to nucleotides 1315 to 1339 of the Graf1 gene). HEK293/Graf1 was transfected with pHygH1dTO-shGraf1 using XtremeGENE 9. Stable transfectants were selected with 200 μg/ml hygromycin (Invitrogen). Clones with highly efficient depletion were used as Graf1 knockdown (KD) cells (HEK293/Graf1KD). A cell line transduced with pHygH1dTO (an empty vector) was generated as a control (HEK293/Graf1ctrl).
RESULTS
Rho activation promoted hPIV-2 growth.
LPA, an activator of Rho family GTPases (41), promotes syncytium formation by hPIV-2 (14). To see whether Rho activation affects hPIV-2 growth, we investigated hPIV-2 growth in the presence of a Rho activator (LPA) and inhibitors (Y-27632, ML 141, AZA1, and NSC23766). Y-27632 specifically inhibits the RhoA/Rho-associated coiled-coil-containing protein kinase (ROCK), one of the downstream RhoA effectors (42, 43). ML141 is a selective and allosteric inhibitor for Cdc42 (44), and AZA1 inhibits the activation of Cdc42 and Rac1 (45). NSC23766 inhibits the Rac1-specific GEF but has no effect on RhoA and Cdc42 activity (46, 47). HEK293 cells were starved for serum and treated with LPA, Y-27632, ML141, AZA1, or NSC23766. After a 16-h treatment, cells were infected with hPIV-2 at an MOI of 1. Plaque assays showed that the virus titer from LPA-treated cells was approximately 6-fold higher than that from untreated cells, while other reagents tested had no effect (Fig. 1A and andB).B). Activation of Rho signaling thus promoted not only syncytium formation but also the growth of hPIV-2.
hPIV-2 infection induced RhoA activation.
RhoA, Cdc42, and Rac1 are representative Rho family GTPases. As shown in Fig. 1A and andB,B, none of the Rho inhibitors tested had an effect on hPIV-2 growth regardless of the enhancement of hPIV-2 growth by LPA. Y-27632 indirectly inhibits RhoA signaling (via ROCK), leading to the possibility that ROCK-independent RhoA signaling is involved in hPIV-2 growth. Furthermore, Gower et al. (6) reported that RhoA activation occurred during infection with RSV, another paramyxovirus. Thus, we focused on RhoA. To study whether hPIV-2 infection causes RhoA activation, activated RhoA in hPIV-2-infected HEK293 cells was measured. First, we confirmed that HEK293 cells were infected with hPIV-2 by IB analysis using anti-hPIV-2 V/P Ab (Fig. 2A, top panel). As shown in Fig. 2A and andB,B, activated RhoA was clearly increased by hPIV-2 infection at 6 and 12 hpi without affecting total RhoA levels.
Graf1 colocalized with viral RNP (vRNP) complexes in hPIV-2-infected cells.
Rho signaling is controlled by a variety of GEFs and GAPs. We hypothesized that some GEFs and/or GAPs regulate hPIV-2 growth by mediating Rho activation. Our previous yeast two-hybrid experiments demonstrated that the hPIV-2 V protein interacts with Graf1 (also known as ARHGAP26), a GAP (unpublished data), and is involved with RhoA and Cdc42 but not Rac1 (48). This Rho GTPase specificity is consistent with data shown in Fig. 1 (no influence of NSC23766 and AZA1 on hPIV-2 growth) and in Fig. 2A and andBB (RhoA activation during hPIV-2 infection). Therefore, we focused on Graf1 as a candidate that may potentially mediate between hPIV-2 growth and Rho signaling.
First, we investigated the subcellular localization of Graf1 in hPIV-2-infected cells by IFA. Graf1 was fused to a myc tag at its N terminus since a suitable antibody against Graf1 was not available. In mock-infected HeLa cells, Graf1 is homogenously distributed within the cytoplasm (Fig. 3A, mock). However, hPIV-2 infection clearly altered Graf1 localization via granule formation (Fig. 3A, all rows for hPIV-2-infected cells). These granules overlapped with hPIV-2 NP, P, and L but not with M, HN, and F proteins. These results suggest that Graf1 colocalized with vRNP in hPIV-2-infected cells.
Graf1 is predicted to mainly consist of four domains: the N-terminal Bin/amphiphysin/Rvs (BAR), pleckstrin homology (PH), RhoGAP (GAP), and C-terminal Src homology 3 (SH3) domains (Fig. 3B). To investigate which Graf1 domain(s) was affected by hPIV-2 infection, we generated N-terminal myc-tagged segments of Graf1 consisting of amino acids (aa) 19 to 226 (BAR), aa 268 to 368 (PH), aa 391 to 565 (GAP), and aa 705 to 759 (SH3). In uninfected cells, while the distribution of BAR, PH, and GAP was similar to that of full-length (FL) Graf1, SH3 distribution was mainly nuclear (Fig. 3A, top row, and B). hPIV-2 infection did not affect the subcellular distribution of BAR, PH, and GAP (Fig. 3C, myc-BAR, myc-PH, and myc-GAP). In contrast, SH3 partially formed granules in hPIV-2 infection, similar to FL Graf1 (Fig. 3C, myc-SH3). Furthermore, these granules colocalized with the hPIV-2 P protein. These results indicate that the SH3 domain of Graf1 is important for its colocalization with hPIV-2 proteins.
hPIV-2 P and V proteins interact with Graf1 via their common N-terminal region.
Since the distribution of Graf1 was affected by hPIV-2 infection, it is likely that Graf1 interacts with either the hPIV-2 NP, P, and/or L protein. We addressed this possibility by coimmunoprecipitation experiments. hPIV-2 NP, V, and P genes and their mutants were transfected into COS cells together with N-terminal myc-tagged Graf1 (Fig. 4A). myc-Graf1 was coimmunoprecipitated with V protein (Fig. 4B, bottom panel, lane 3), consistent with our previous yeast two-hybrid study. The P and V/P common region also coimmunoprecipitated with myc-Graf1, while the P-specific region did not (Fig. 4B, bottom panel, lanes 4 to 6), indicating that the N-terminal 164 aa of the P and V proteins are important for the interaction with Graf1. In contrast, an NP-Graf1 interaction was not observed (Fig. 4B, bottom panel, lane 7). To investigate the L protein-Graf1 interaction, BSR T7/5 cells were transfected with myc-Graf1 and the hPIV-2 L or P gene. As shown in Fig. 4C, L could not coimmunoprecipitate with myc-Graf1 (Fig. 4C, bottom panel, lane 4). Next, we examined the interaction between Graf1 and N-terminally truncated mutants of the P protein consisting of aa 48 to 395 (PΔN47) and aa 111 to 395 (PΔN110) (Fig. 4D). Coimmunoprecipitation of myc-Graf1 with PΔN47 but not with PΔN110 was observed (Fig. 4E, bottom panel, lanes 5 and 6). In contrast, deletion of the C-terminal region did not affect the interaction with Graf1 (Fig. 4F and andG,G, bottom panel, lanes 5 to 8). These results indicate the importance of the N-terminal common region of the V and P proteins for their interaction with Graf1.
We previously reported that C-terminal Cys (zinc finger) and Trp residues of the V protein are important for the interaction with several host proteins, including STATs (29), AIP1/Alix (19), TRAF6 (27), and tetherin (30). In order to investigate the involvement of these residues in the interaction with Graf1, mutant V proteins (C193/197A, C209/211/214A, C218/221A, and W178H/W182E/W192A) were tested. IP analysis revealed that all of these mutants retained the ability to interact with Graf1 (Fig. 4H, bottom panel, lanes 5 to 8).
C-terminal SH3 domain-containing region of Graf1 is important for interaction with hPIV-2 P and V proteins.
To identify the region important for the interaction with hPIV-2 V and P proteins, five mutants of Graf1 with N-terminal myc tags were generated. These mutants consist of aa 1 to 368 (BAR+PH), aa 1 to 565 (BAR+PH+GAP), aa 268 to 565 (PH+GAP), aa 268 to 759 (PH+GAP+SH3), and aa 391 to 759 (GAP+SH3) (Fig. 5A). P protein and FL Graf1 or its mutants were coexpressed in COS cells, and the cell lysates were subjected to IP analysis. As shown in Fig. 5B, PH+GAP+SH3 and GAP+SH3 as well as FL Graf1 could immunoprecipitate with P protein (Fig. 5B, bottom panel, lanes 6 to 8). The region of aa 565 to 759 containing the C-terminal SH3 domain of Graf1 is thus necessary for the interaction with P protein. The interaction between Graf1 mutants and V protein was also investigated. To eliminate nonspecific interactions between V and the myc tag, lysis buffer containing 300 mM NaCl was used. As for the case with the Graf1-P interaction, V protein was selectively coimmunoprecipitated with SH3 domain-containing Graf1 mutants (Fig. 5C, bottom panel, lanes 7 to 9).
Graf1 negatively regulated the growth of hPIV-2.
To investigate the effects of Graf1 on hPIV-2 growth, an HEK293 cell line constitutively expressing Graf1 (HEK293/Graf1) and a control cell line (HEK293/ctrl) were generated (Fig. 6A). hPIV-2 was inoculated into HEK293/Graf1 and HEK293/ctrl cells at an MOI of 0.01. We confirmed that the virus titer from HEK293/ctrl cells was similar to that from normal HEK293 cells (data not shown). Plaque assays demonstrated that the virus titer from HEK293/Graf1 cells was approximately 5- to over 10-fold lower than that from HEK293/ctrl cells (Fig. 6B). For further confirmation of the inhibition of hPIV-2 growth by Graf1, Graf1 was knocked down in HEK293/Graf1 cells (Fig. 6C). This knockdown cell line (HEK293/Graf1KD) and its control cell line (HEK293/Graf1ctrl) were infected with hPIV-2 at an MOI of 0.01. The virus titer from HEK293/Graf1KD cells increased approximately 10-fold compared to that from HEK293/Graf1ctrl cells at 48 and 72 hpi (Fig. 6D). These results indicated that Graf1 negatively regulated hPIV-2 growth.
Graf1 inhibited hPIV-2-mediated RhoA activation.
Since Graf1 inactivates RhoA via its GAP activity, it is likely that Graf1 has a role in downregulation of RhoA activation during hPIV-2 infection. HEK293/ctrl, HEK293/Graf1, and HEK293/Graf1KD cells were infected with hPIV-2 at an MOI of 1 for the hours indicated on the figure. The activated RhoA level in these cells was measured. hPIV-2 infection caused an increase in the level of activated RhoA in HEK293/ctrl cells at 6 and 12 hpi (Fig. 7A), which is similar to that in normal HEK293 cells (Fig. 2A and andB).B). In contrast, overexpression of Graf1 inhibited hPIV-2-mediated RhoA activation at any time tested postinfection (Fig. 7B). In HEK293/Graf1KD cells, RhoA was activated by hPIV-2 infection at 6 and 12 hpi (Fig. 7C).
hPIV-2 P and V proteins promoted hPIV-2-mediated RhoA activation.
As shown in Fig. 4, hPIV-2 P and V proteins interacted with Graf1. Graf1 blocked RhoA activation caused by hPIV-2 (Fig. 7). If P and V proteins affected this function of Graf1 by their interaction with Graf1, overexpression of P or V proteins might alter the amount of activated RhoA during hPIV-2 infection. To confirm this, we established HEK293 cells constitutively expressing the P (HEK293/P) or V (HEK293/V) protein. The amounts of total RhoA and activated RhoA in these cells were not affected by overexpression of the P or V protein (Fig. 8A and andB).B). We confirmed the hPIV-2 infection in these cells using IB analysis (Fig. 8C). At 1 hpi, the amount of activated RhoA in hPIV-2-infected HEK293/P and HEK293/V cells increased while it did not in normal HEK293 cells at this time postinfection (Fig. 8D, ,E,E, and andF,F, lanes 1 and 2 of top panels). At 6 and 12 hpi, the activated RhoA level in HEK293/P and HEK293/V cells was similar to that in normal HEK293 cells (Fig. 8D, ,E,E, and andF,F, lanes 3 to 6 of top panels). These results suggest that hPIV-2 P and V proteins promoted hPIV-2-mediated RhoA activation.
DISCUSSION
Our previous study demonstrated that hPIV-2-induced syncytium formation was enhanced by the activation of Rho signaling by LPA while it was suppressed by the blocking of RhoA activation by Y-27632 (14). In the present study, we report that LPA promoted hPIV-2 growth although treatment with Y-27632 did not affect growth (Fig. 1). These data suggest that the extent of hPIV-2-induced syncytium formation in the presence of Rho activator or inhibitor does not necessarily correspond to that of viral growth. On the contrary, RhoA activation did not affect the growth of RSV despite promotion of syncytium formation induced by RSV. This suggests a difference in the involvement of RhoA activation by LPA among different Paramyxoviridae family members.
Although we attempted to see a direct interaction of hPIV-2 proteins with RhoA, these interactions could not be observed by IP (data not shown). In this study, we found instead that Graf1 affects hPIV-2 growth by controlling RhoA signaling. The interaction of the C-terminal region of P with the NP protein is required for granule formation (17). NP protein interacts with L and V proteins (18, 21). Localization of Graf1 clearly overlapped with these granules in hPIV-2-infected cells (Fig. 3A). Thus, we anticipated that Graf1 would interact with the NP, P, and/or L protein(s). Although the interaction of Graf1 with the NP or L protein could not be observed (Fig. 4B and andC),C), P protein interacts with Graf1 via its region of aa 47 to 110 (Fig. 4E). P protein binds to NP protein through the regions of aa 1 to 47 and aa 357 to 395, whereas the region of aa 278 to 353 of P protein is required for the binding with L protein (17, 18). Thus, the interaction site of P protein with Graf1 seems to be different from that with the NP and L proteins. It is noteworthy that Graf1 interacted with not only the P but also V protein of hPIV-2 (Fig. 4B). Based on these results, the N-terminal common region of the P and V proteins seems to be important for the binding with Graf1. However, as shown in Fig. 4B (lanes 3 to 5 of the bottom panel), increasing the length of the protein improved binding. Thus, folding of the P or V protein might also be a determinant for the interaction with Graf1. Although it is unclear whether the subcellular localization of Graf1 overlaps with that of the V protein as shown in Fig. 3A, it is likely that Graf1 colocalizes with the V protein. The hPIV-2 V protein interacts with several host proteins via its C-terminal Trp residues in the V-specific region (19, 27,–30). Graf1 could interact with V mutants with substitutions at these Trp residues (Fig. 4H), indicating that interaction sites with Graf1 are different from those with other host proteins that have been identified to interact with V protein.
Graf1 contains BAR, PH, GAP, and SH3 domains, all of which are well conserved among RhoGEFs and RhoGAPs. BAR and PH domains function to regulate endocytosis by binding and inducing membrane curvature (49,–51). A GAP domain inactivates small GTPases by enhancing GTP hydrolysis, and the SH3 domain modulates this GAP activity. The SH3 domain of Graf1 is reported to bind several other host proteins, including FAK (48, 52), PKNβ (53), and dynamin (50). All of these host factors possess proline-rich motifs that are important for binding with the SH3 domain. In this study, we found that the region of aa 565 to 759 containing the SH3 domain of Graf1 was important for the interaction with the hPIV-2 P and V proteins (Fig. 3C and and5B5B and andC).C). These proteins also contain a proline-rich region in their N-terminal common regions corresponding to aa 65 to 103. Our data that Graf1 was coimmunoprecipitated with PΔN47 but not with PΔN110 (Fig. 4E) indicate the importance of this proline-rich region in the interaction with Graf1.
We propose a model based on the results in this study (Fig. 9). Activated RhoA signaling appears to be important for hPIV-2. In fact, hPIV-2 infection promoted RhoA activation. Since Graf1 inactivates RhoA via its GAP activity, hPIV-2-induced RhoA activation is suppressed by Graf1. This would explain the poor growth of hPIV-2 in HEK293 cells constitutively expressing Graf1. hPIV-2 may counteract Graf1 via its interaction with P and V proteins. As shown in Fig. 8, the increase of hPIV-2-induced RhoA activation was observed in HEK293 cells constitutively expressing V or P protein but not in normal HEK293 cells at 1 hpi. P protein exists as a component of the vRNP complex in virions. The V protein of the genus Rubulavirus, including that of hPIV-2, is a structural component of the nucleocapsid in virions (54). Thus, there are both P and V proteins in the hPIV-2 virion, which might result in the counteraction of Graf1 at such an early stage of infection.
However, which steps are affected in hPIV-2 life cycle by RhoA and Graf1 remains an open question. Lundmark et al. (49) demonstrated that Graf1 regulates the endocytic pathway in which clathrin-independent carriers (CLICs) and glycosylphosphatidylinositol (GPI)-enriched endocytic compartments (GEECs) participate. Indeed, the infection of adeno-associated virus 2 is reported to require this CLIC/GEEC pathway (55). Moreover, it was reported that simian virus 40 (SV40) stimulates Graf1, resulting in the inactivation of RhoA signaling (56). This leads to successful SV40 entry, which is thought to occur via clathrin-independent endocytosis. However, the entry of hPIV-2 occurs through the fusion of the viral envelope with the cell plasma membrane rather than through endocytosis (15). Graf1 also reportedly functions as a regulator of myoblast cell fusion (57). Furthermore, RhoA activation promoted syncytium formation of hPIV-2-infected cells (14). These studies indicate the involvement of Graf1 in hPIV-2-induced syncytium formation. However, overexpression of Graf1 in Vero cells did not affect the efficiency of cell-cell fusion mediated by hPIV-2 (data not shown). It is difficult to identify the downstream target of RhoA signaling involved in hPIV-2 infection since RhoA forms complex signaling networks and has many functions. Rotavirus infection is reported to increase the level of GTP-bound active RhoA and to alter the actin and tubulin cytoskeleton (12). Jiang et al. (58) reported that NS1, an accessory protein of IAV, induces host cell cycle arrest in the G0/G1 phase and promotes viral replication by inhibiting RhoA expression and activation. The V protein of parainfluenza virus type 5 (PIV-5) was found to be involved in the G2-M arrest, and this cell cycle alteration might be caused by the interaction between PIV-5 V and the damage-specific DNA-binding protein 1 (DDB1) (59). Since V proteins of other paramyxoviruses, including hPIV-2, mumps virus, and measles virus, interact with DDB1, it is likely that the hPIV-2 V protein plays a role in cell cycle alteration. Exploring the involvement of Graf1 and RhoA activation in the life cycles of other paramyxoviruses might lead to a better understanding of the relationship between Graf1 and RhoA during virus infection. Further investigations will be needed to determine whether hPIV-2-mediated RhoA activation affects these alterations.
Nonetheless, the present study has demonstrated that RhoA signaling closely correlates with hPIV-2 life cycles and has identified Graf1 as an important contributor linking them.
ACKNOWLEDGMENT
We are grateful to Yoshitaka Ono for pTB701 plasmid carrying Graf1 cDNA.