Human Papillomavirus 16 Infection Induces VAP-Dependent Endosomal Tubulation

ABSTRACT

Human papillomavirus (HPV) infection involves complex interactions with the endocytic transport machinery, which ultimately facilitates the entry of the incoming viral genomes into the trans-Golgi network (TGN) and their subsequent nuclear entry during mitosis. The endosomal pathway is a highly dynamic intracellular transport system, which consists of vesicular compartments and tubular extensions, although it is currently unclear whether incoming viruses specifically alter the endocytic machinery. In this study, using MICAL-L1 as a marker for tubulating endosomes, we show that incoming HPV-16 virions induce a profound alteration in global levels of endocytic tubulation. In addition, we also show a critical requirement for the endoplasmic reticulum (ER)-anchored protein VAP in this process. VAP plays an essential role in actin nucleation and endosome-to-Golgi transport. Indeed, the loss of VAP results in a dramatic decrease in the level of endosomal tubulation induced by incoming HPV-16 virions. This is also accompanied by a marked reduction in virus infectivity. In VAP knockdown cells, we see that the defect in virus trafficking occurs after capsid disassembly but prior to localization at the trans-Golgi network, with the incoming virion-transduced DNA accumulating in Vps29/TGN46-positive hybrid vesicles. Taken together, these studies demonstrate that infection with HPV-16 virions induces marked alterations of endocytic transport pathways, some of which are VAP dependent and required for the endosome-to-Golgi transport of the incoming viral L2/DNA complex.

IMPORTANCE Human papillomavirus infectious entry involves multiple interactions with the endocytic transport machinery. In this study, we show that incoming HPV-16 virions induce a dramatic increase in endocytic tubulation. This tubulation requires ER-associated VAP, which plays a critical role in ensuring the delivery of cargoes from the endocytic compartments to the trans-Golgi network. Indeed, the loss of VAP blocks HPV infectious entry at a step after capsid uncoating but prior to localization at the trans-Golgi network. These results define a critical role for ER-associated VAP in endocytic tubulation and in HPV-16 infectious entry.

INTRODUCTION

High-risk human papillomaviruses (HPVs) are major causes of human cancers, with cervical cancer being the most important (1–4). While over 200 different HPV types have been reported, HPV-16 alone is responsible for over 50% of cervical cancers, making it an important human pathogen (3). The HPV genome is a circular double-stranded DNA molecule of approximately 8,000 bp and contains early and late open reading frames (ORFs). The early ORFs encode a number of proteins whose major role is to prime infected keratinocytes for replicating the viral genome. In rare cases, where the infectious cycle fails and the expression of the viral oncoproteins E6 and E7 becomes deregulated, this can ultimately result in the development of cancer (5). The late ORFs encode the viral capsid proteins L1 and L2, which encapsidate newly synthesized viral genomes.

The viral capsid proteins L1 and L2 play an essential role in successful viral infection. The L1 protein facilitates the attachment of the virus to the extracellular matrix, mostly through binding to heparin sulfate proteoglycans (HSPGs) (6), which causes a conformational change in the viral capsid and which mediates receptor engagement and the endocytic uptake of the infectious particle (7–9), a step which also appears to require active proliferative signaling within the target cell (10, 11). The viral particles then follow a complex pathway of endocytic transport. Early in this process, the viral capsid begins to disassemble as a result of endocytic acidification, and portions of the L2 protein become exposed to the cytosolic side of the endosomes (12, 13). The cytosol-exposed regions of L2 then recruit a number of cellular proteins involved in endocytic cargo fate determination. These proteins include members of the sorting nexin family of proteins, components of the retromer, and the ESCRT complex (14–17).

During the later stages of endocytic maturation, the L2/DNA complex becomes separated from L1 through the action of cyclophilins (18), and the majority of the L1 protein is then degraded in the lysosome (19). Meanwhile, the L2/DNA complex is trafficked to the trans-Golgi network (TGN) (20), where it is believed to reside until the cell enters mitosis. During mitosis, membrane dissolution and nuclear envelope breakdown occur, which then allow the L2/DNA complex, plus a small amount of residual L1, to access the nuclear compartment and accumulate at PML oncogenic domains, where the initiation of viral gene expression is believed to occur (21–24). Central to this whole process of HPV infectious entry are endocytic trafficking and vesicular transport to the TGN.

Recent studies have shown that contacts between the endosome and the endoplasmic reticulum (ER) play a critical role in ensuring the correct delivery of cargoes from the endosome, via the retromer, to the TGN (25). This appears to require VAP, which plays an essential role in inducing actin nucleation, endosomal budding, and tubulation. While HPV is known to utilize components of the retromer to complete infectious entry, there is currently no information on how incoming virions might affect the rates of endosomal tubulation or on whether virion transport to the TGN requires well-defined components of the ER that are involved in retromer function, such as the VAP complex. Therefore, in this study, we were interested in analyzing how infection with HPV-16 affects endocytic tubulation. To do this, we made use of a well-characterized marker of tubulating endosomes, MICAL-L1 (molecules interacting with CasL-like 1). Together with a number of other proteins, MICAL-L1 aids in membrane shaping (26), and the loss of MICAL-L1 inhibits the recycling of many cargoes back to the cell surface, including epidermal growth factor receptor (EGFR) (27). We find that infection with HPV-16 induces a robust increase in the level of MICAL-L1 expression and subsequent increases in the level and extent of endocytic tubulation. Furthermore, using a VAP knockout cell line, we show that this ER component is essential for obtaining efficient virus infection. Indeed, the loss of VAP reduces endocytic tubulation and blocks virus infection at a point prior to entry into the TGN but after capsid disassembly.

RESULTS

HPV-16 infection stimulates endosomal tubulation.

Endosomal tubulation is a critical step in the sorting of endocytic cargoes, and we were interested in examining whether HPV-16 infectious entry had any influence on the levels of endosomal tubulation taking place within the cell. To investigate this, HeLa cells were infected with 5-ethynyl-2′-deoxyuridine (EdU)-labeled HPV-16 pseudovirions (PsVs), and infected cells were then fixed at 2 h, 8 h, and 24 h postinfection along with an uninfected control. Immunofluorescence analysis was then performed to detect the pattern of MICAL-L1 expression at these different time points, and the distribution of HPV-16 PsVs within the cells was also monitored by staining for EdU-labeled DNA encapsidated in the PsVs. The results obtained are shown in Fig. 1A. In uninfected cells, MICAL-L1 appears to be expressed at a low level and is present as small dots and slightly elongated tubules in some of the cells. This is very similar to the distribution pattern of MICAL-L1 reported previously (26, 28, 29). However, following infection with HPV-16 PsVs, there is a dramatic alteration in the pattern of MICAL-L1 expression. By the 2-h time point, there is a significant increase in the number of tubulating endosomes, and this is even more apparent by 8 h postinfection. At this time, there also appears to be a generalized overall increase in the level of MICAL-L1 protein expression. In contrast, by 24 h postinfection, the number of tubulating endosomes appears to be reduced. Upon closer inspection of the tubulating endosomes (Fig. 1A, right), there is a clear indication of the colocalization of the EdU-labeled HPV-16 PsVs with MICAL-L1 in the vesicular structures, and this is verified through quantification in Fig. 1C and D.

FIG 1.

HPV-16 infection induces endosomal tubulation. (A) HeLa cells were infected with HPV-16 PsVs using 150 viral genome equivalents per cell (vge/cell), and cells were fixed at 2, 8, and 24 h postinfection. Uninfected (UI) cells were used as a control. Reporter DNA that is encapsidated in PsVs is detected by EdU labeling (red), whereas endogenous MICAL-L1 is stained with MICAL-L1 antibody (green) as an endosomal tubulation marker. Images were captured by confocal microscopy. The right-hand column shows blown-up images. HPV-16 PsVs partially colocalize with MICAL-L1 in HeLa cells, as indicated by arrows. (B) Western blot showing the expression of MICAL-L1 in response to infection with HPV-16 PsVs. Lane 1 represents uninfected cells, whereas lanes 2 to 4 represent MICAL-L1 expression at 2, 8, and 24 h postinfection. Tubulin was used as a loading control. (C) Quantification of MICAL-L1 localization with EdU in HeLa cells. At least 300 cells under each condition from three independent experiments were analyzed. (D) Colocalization was analyzed using Pearson's correlation coefficients using Fiji software on 10 individual cells. Statistical significance between uninfected cells and cells at 8 h postinfection was determined by Student's t test (****, P < 0.00001).

To examine whether infection with HPV-16 PsVs can indeed increase the levels of expression of MICAL-L1, cells were infected and then harvested, along with an uninfected control, at 2 h, 8 h, and 24 h postinfection. The levels of MICAL-L1 were then analyzed by Western blotting, and the results in Fig. 1B show a steady increase in the levels of MICAL-L1 after virus infection, with a peak being attained at 8 h postinfection, which is in agreement with the results obtained from the immunofluorescence analyses.

We next wanted to determine whether the induction of tubulating endosomes was limited to HeLa cells or whether it could be observed in normal keratinocytes following HPV-16 infection. To do this, we infected normal immortalized keratinocytes (NIKS) with HPV-16 PsVs and analyzed the pattern of MICAL-L1 staining at different time points postinfection. The results in Fig. 2 also show a clear induction of endosomal tubulation in NIKS following infection with HPV-16 PsVs, although the kinetics of induction appear somewhat slower than those in HeLa cells. Similar results were also obtained with HaCaT cells (data not shown).

FIG 2.

Endosomal tubulation is not limited to HeLa cells. NIKS were infected with HPV-16 PsVs (150 vge/cell) and fixed at 2, 8, and 24 h postinfection. Uninfected (UI) cells were used as a control. The right-hand column shows blown-up images.

Endosomal tubulation does not require HPV-16 L2.

We were next interested in ascertaining whether endosomal tubulation was a result of virus endocytosis or related to a specific function of the viral L2 protein. In order to do this, HeLa cells were infected with either HPV-16 PsVs or HPV-16 virus-like particles (VLPs), which contain only the major capsid protein L1. The cells were then fixed and stained for MICAL-L1, HPV-16 L1, or EdU-labeled DNA at different times postinfection. The results in Fig. 3A show a clear induction of endosomal tubulation following infection with HPV-16 PsVs, with significant colocalization of the transduced DNA with MICAL-L1 at the 2-h and 8-h time points postinfection. Interestingly, a similar induction of endosomal tubulation was also obtained following infection with HPV-16 VLPs (Fig. 3C), although we have been unable to determine whether there is any colocalization of L1 (shown in Fig. 3B) and MICAL-L1 due to antibody constraints. These results indicate that virus-induced endosomal tubulation is not linked to a specific function of L2 but is instead related more to the endocytic uptake of incoming virus particles.

FIG 3.

HPV-16 L1-only VLPs also induce endosomal tubulation. (A) HeLa cells were infected with WT HPV-16 PsVs (150 vge/cell) and fixed at 2, 8, and 24 h postinfection. Uninfected cells were used as a control. Reporter DNA that is encapsidated in PsVs is detected by EdU labeling (red), whereas endogenous MICAL-L1 is stained with MICAL-L1 antibody (green). (B) HeLa cells were infected with HPV-16 L1-only VLPs, which are detected by using an anti-L1 antibody. (C) MICAL-L1 staining in a parallel experiment at different times after infection with HPV-16 VLPs.

HPV-16 infection is VAP dependent but MICAL-L1 independent.

Having shown that infection with HPV-16 PsVs induces endosomal tubulation, we were next interested in determining whether MICAL-L1 itself is actually required for virus infection. Therefore, MICAL-L1 expression was knocked down by using targeted small interfering RNA (siRNA), and after 48 h, the cells were infected with HPV-16 PsVs. After a further 48 h, the cells were harvested, and luciferase activity was measured. As shown in Fig. 4A, the loss of MICAL-L1 results in only a very minor decrease in infectivity, indicating that the reported roles of MICAL-L1 in membrane remodeling do not play an important part in HPV-16 PsV infectious entry. MICAL-L1 knockdown was verified by Western blotting of the cell lysates probed for MICAL-L1 and α-tubulin, as shown in Fig. 4B.

FIG 4.

(A) Loss of MICAL-L1 does not inhibit infection by HPV-16 PsVs. HeLa cells were transfected with siRNA targeting MICAL-L1 or scrambled siRNA as a control. Cells were infected with HPV-16 PsVs (50 vge/cell) carrying a luciferase reporter plasmid at 48 h posttransfection. After a further 48 h, the cells were harvested, and luciferase activity was measured by using a luminometer. The values were normalized to those of scramble siRNA-transfected cells. The data shown are the mean luciferase readings from three independent experiments, where bars indicate standard errors. (B) The efficiency of MICAL-L1 knockdown was analyzed by Western blotting of the cell lysate used for the luciferase assay. (C and D) HeLa cells were transfected with scrambled siRNA (C) or siRNA against MICAL-L1 (D) and infected with HPV-16 PsVs (150 vge/cell). Tubulation was detected with MICAL-L1 (green) and Syndapin2 (sea green).

An alternative marker for tubulating endosomes is Syndapin2. Therefore, we were interested in determining whether the loss of MICAL-L1 also resulted in a loss of Syndapin2-positive tubulating endosomes. Cells were transfected with control or MICAL-L1 siRNAs (siMICAL-L1) as described above and then infected with HPV-16 PsVs. As shown in Fig. 4C, there is again a very clear induction of tubulating endosomes at 8 h postinfection, as determined by both MICAL-1 and Syndapin2 staining, with a very clear colocalization of MICAL-L1 and Syndapin2 in these structures. In contrast, in MICAL-L1 siRNA-transfected cells (Fig. 4D), there are fewer Syndapin2-positive tubulating endosomes, but they are still apparent. This suggests that tubulation induced by HPV-16 infection can still occur in the absence of MICAL-L1.

While the above-described results demonstrate that MICAL-L1 per se is not required for efficient virus infection, another key element involved in vesicular processing and transport to the TGN is the ER-associated VAP protein. During endosomal tubulation, the ER and endosomal membranes come into close contact, and VAP ensures efficient WASH-dependent actin nucleation and cleavage of the budding endosome. Therefore, we next wanted to determine whether VAP was required for virus infection. To do this, wild-type (WT) HeLa and VAP double-knockout (DKO) HeLa cells were infected with HPV-16 PsVs, and after 48 h, luciferase activity was measured. The results in Fig. 5A demonstrate that the loss of VAP results in a dramatic decrease in the ability of HPV-16 PsVs to infect cells, suggesting that the ER-associated VAP protein, which is required to tether the tubulating endosome to the ER, plays an important role in the HPV-16 infectious entry pathway.

FIG 5.

HPV-16 infection depends on the integral ER protein VAP. (A) WT HeLa and VAP DKO HeLa cells were infected with HPV-16 PsVs or McPyV for 48 h. Relative luminescence, which indicates PsVs/McPyV infectivity, was measured and is plotted as a bar graph. The data shown here are the mean luciferase readings derived from 3 independent experiments, where the bars indicate standard errors. (B and C) WT HeLa cells (B) and VAP DKO HeLa cells (C) were infected with EdU-labeled HPV-16 PsVs (150 vge/cell) and fixed at 2, 8, and 24 h postinfection. PsVs are stained red, and MICAL-L1 is stained green. Representative images from three independent experiments are shown. Images were captured by confocal microscopy. The right-hand panel shows blown-up images. (D) Western blot showing the expression of MICAL-L1 in HeLa and VAP DKO cells in response to infection with HPV-16 PsVs (100 vge/cell). (E) The tubulation length (micrometers) in WT HeLa and VAP KDO HeLa cells was quantified by using the measurement-and-dimensioning function of the LSM 510 system. At least 100 cells under each condition from three independent experiments were analyzed. Bars indicate standard errors.

We also extended these studies to determine whether the loss of VAP would also negatively affect infection by another virus whose entry is endosomally mediated. For this, we analyzed Merkel cell polyomavirus (McPyV) PsVs carrying a luciferase reporter construct. The results in Fig. 5A also show a marked decrease in the infectious entry of McPyV when VAP expression is knocked down, indicating that the requirement of VAP for virus infectious entry is not restricted to HPV.

Having found that VAP was required for efficient virus infection, we next wanted to determine whether it was also involved in the tubulation induced following HPV-16 infection. Wild-type and VAP knockout HeLa cells were again infected with EdU-labeled HPV-16 PsVs, and tubulation was monitored at different times postinfection by performing immunofluorescence staining for MICAL-L1. As shown in Fig. 5B, there is again a dramatic increase in endosomal tubulation by 8 h after infection with HPV-16 PsVs in wild-type HeLa cells, but this is largely absent in VAP knockout cells (Fig. 5C), with a great reduction in the length of the tubulating endosomes (Fig. 5C and E). We also analyzed whether MICAL-L1 protein levels were also affected following infection in VAP knockdown cells. The results in Fig. 5D again show a marked increase in the levels of the MICAL-L1 protein in wild-type HeLa cells but only a minimal change in VAP knockout cells. Taken together, these results demonstrate that VAP plays an important role in the induction of MICAL-L1 and endosomal tubulation seen following HPV-16 infection.

Having found that VAP plays an important role in HPV-16 infectious entry, we were interested in determining whether we could detect incoming virally transduced DNA colocalizing with VAP. HeLa cells were infected with HPV-16 PsVs that contained EdU-labeled DNA, and at 8 h postinfection, the cells were fixed and stained for VAP. The results in Fig. 6A show clear evidence of virally transduced DNA colocalizing with VAP at this time point. Unlike for MICAL-L1, however, we see no evidence that infection with HPV-16 PsVs has any effect on the levels of VAP protein expression, as determined by Western blotting (Fig. 6B).

FIG 6.

HPV-16 PsVs exhibit some colocalization with VAP. (A) WT HeLa cells were infected with EdU-labeled HPV-16 PsVs (150 vge/cell) and fixed at 8 h postinfection. Uninfected cells were used as a control. Red indicates the virus, whereas green indicates VAPB. Images were captured by confocal microscopy. The images shown here are representative of data from 3 independent repeats. (B) Western blot analysis of VAPB expression in uninfected cells (UI) and at different times after infection with HPV-16 PsVs, with α-actinin being used as the loading control.

VAP loss causes a block in infection prior to entry into the TGN.

Having shown that the loss of VAP inhibited virus infection and reduced HPV-16 PsV-induced endosomal tubulation, we were next interested in ascertaining where the block in virus infectious entry occurred. We first analyzed whether this occurred prior to capsid uncoating and made use of the 33L1-7 antibody, which recognizes an epitope on the viral L1 protein that is exposed only once the virus particle begins to disassemble. Cells were infected with Alexa Fluor 488 (AF488)-labeled PsVs, and after 7 h, the cells were fixed and analyzed by immunofluorescence. As shown in Fig. 7, there is no difference in the levels of 33L1-7 staining between wild-type HeLa cells (Fig. 7A) and VAP knockout HeLa cells (Fig. 7B). These results indicate that the block in infection upon the loss of VAP occurs after capsid disassembly.

FIG 7.

In the absence of VAP, HPV-16 infection is not blocked before capsid uncoating. WT HeLa cells (A) and VAP DKO HeLa cells (B) were infected with AF488-labeled HPV-16 PsVs (150 vge/cell) and fixed at 7 h postinfection. Uninfected cells were used as a control. Green indicates the total capsid, whereas red (33L1-7) indicates L1 in a disassembled capsid. Capsid disassembly is similar in WT HeLa and VAP DKO HeLa cells, which indicates that HPV-16 infection is not blocked before uncoating. Images were captured by confocal microscopy. The images shown here are representative of data from 3 independent repeats.

Since VAP has been shown to play an important role in endosome-to-TGN trafficking, we next analyzed whether the viral L2 protein could be detected in the TGN in VAP knockout cells. The cells were infected with HPV-16 PsVs, and after 18 h, the cells were fixed and analyzed for the presence of HPV-16 L2 and the Golgi marker Giantin. The results in Fig. 8A show colocalization between HPV-16 L2 and the TGN in wild-type cells but virtually no colocalization in VAP knockout cells, as can be seen in Fig. 8B and C. These results demonstrate that the loss of VAP results in a defect in the infectious entry pathway of HPV-16 PsVs at a step most likely involving the initiation of endosomal tubulation and prior to virus entry into the TGN.

FIG 8.

Loss of VAP causes a block of virus entry into the Golgi complex. (A and B) WT HeLa cells (A) and VAP DKO HeLa cells (B) were infected with HPV-16 PsVs (150 vge/cell) and fixed at 18 h postinfection. Uninfected cells were used as a control. Cells were stained with Giantin (green) as a Golgi marker and with mouse anti-FLAG to detect L2. In WT HeLa cells, the virus localizes at the Golgi complex, whereas in VAP DKO cells, most of the virus particles are outside the Golgi complex. This can be seen more clearly in the blown-up images at the right. (C) Quantification of L2 and Giantin colocalization. At least 150 infected cells from three independent experiments were analyzed, and the number of virus particles per cell localizing with the Golgi complex was counted. Bars represent standard errors.

Previous studies have shown that VAP knockdown results in the formation of hybrid vesicles, which stain positive for certain Golgi markers such as TGN46 and retromer components such as Vps29 (25). We were therefore interested in analyzing whether HPV-16 PsV-transduced DNA accumulates in such structures when VAP expression is ablated. In order to do this, control cells and cells lacking VAP were infected with HPV-16 PsVs containing EdU-labeled DNA. At 8 h postinfection, the cells were fixed and stained for Vps29 and TGN46, and the results obtained are shown in Fig. 9. In control cells, the transduced DNA is occasionally found in close proximity to Vps29, which is in agreement with data from previous studies showing that the retromer is required for HPV-16 infectious entry (17, 30), but the DNA is seen only very rarely in structures that stain positive for both TGN46 and Vps29. In contrast, in VAP knockout cells, there is a clear alteration in the subcellular distribution of Vps29, which is in agreement with data from previous studies (25), and a concomitant increase in the amount of HPV-16-transduced DNA that is found in vesicles staining positive for Vps29 alone and also in hybrid vesicles staining positive for both Vps29 and TGN46.

FIG 9.

The HPV-16 pseudogenome localizes with Vps29 alone and Vps29-TGN46 hybrid vesicles in VAP DKO HeLa cells. (A) WT HeLa cells and VAP DKO HeLa cells were infected with HPV-16 PsVs (150 vge/cell) and fixed at 8 h postinfection. Uninfected cells were used as a control. Cells were stained for EdU (red), Vps29 (green), and TGN46 (sea green). Yellow arrows show areas of colocalization between EdU-labeled DNA and Vps29, while white arrows show the colocalization of Vps29, EdU-labeled DNA, and TGN46. (B) Quantification of L2 and Vps29/Vps29-TGN46 colocalization. At least 100 infected cells from three independent experiments were analyzed. Bars represent standard errors.

This phenotype of hybrid vesicle formation was also observed for cells in which sorting nexin 2 (SNX2) expression was knocked down (25). We therefore proceeded to investigate whether the loss of SNX2 would also perturb infection by HPV-16 PsVs. Either wild-type or VAP knockdown cells were transfected with control siRNA or siRNA targeting SNX2 and then infected with HPV-16 PsVs containing a luciferase reporter construct. After 48 h, the cells were harvested, and luciferase activity was measured. The results in Fig. 10A show a modest decrease in infectivity in the absence of SNX2, and this was more pronounced when the assays were performed with VAP knockdown cells. Taken together, these results demonstrate that HPV-16 infectious entry is VAP and SNX2 dependent, with the loss of VAP resulting in the accumulation of incoming virions in an aberrant population of hybrid vesicular structures.

FIG 10.

Effect of SNX2 knockdown on infection by HPV16 PsVs. (A) WT HeLa and VAP DKO HeLa cells were transfected with siRNA against SNX2 or scrambled siRNA as a control and infected with HPV-16 PsVs carrying a luciferase reporter plasmid at 48 h posttransfection. After a further 48 h, the cells were harvested, and luciferase activity was measured. The values were normalized to those for scramble siRNA-transfected cells. The data shown are the average luciferase readings from three independent experiments, where bars indicate standard errors. (B) The efficiency of SNX2 knockdown was analyzed by Western blotting for SNX2 using the cell lysate used for the luciferase assay. α-Actinin was used as a loading control.

DISCUSSION

The early endosome acts as a sorting station where endocytosed proteins are initially localized into tubular extensions. These cargo-containing tubular extensions subsequently become detached from the endosome and either recycle the cargoes back to the plasma membrane or deliver them to the Golgi complex (31, 32). For HPV-16, successful infectious entry requires its endocytic transport to the TGN and its subsequent nuclear entry during mitosis and nuclear envelope dissolution. In this study, we were interested in the events that control endosomal tubulation and how this might be affected or used by HPV-16 during infectious entry. We show that infection by HPV-16 PsVs strongly promotes the process of endosomal tubulation, as determined by the recruitment of MICAL-L1 to endocytic vesicular structures. However, this is all dependent upon the ER-associated VAP protein, without which the incoming L2/DNA complex fails to reach the TGN.

A frequently used marker of endosomal tubulation is MICAL-L1, which has often been used to identify tubulation associated with the formation of vesicular structures involved in retromer-mediated recycling. While the precise function of MICAL-L1 is still unknown, it has been shown to play a role in the recycling of certain cargoes, such as EGFR (27), and it is believed to help in the process of membrane shaping during endocytic tubulation (26). Using MICAL-L1 as a marker for endocytic tubulation, we find a striking increase in the presence of MICAL-L1 in vesicular tubules following infection of HeLa cells with HPV-16 PsVs. This increase in tubulation is very quick and can be easily observed by as early as 2 h postinfection. This peaks at around 8 h postinfection, and by 24 h postinfection, the levels of endocytic tubulation are greatly reduced. Similar results were also obtained with NIKS and HaCaT cells, although the kinetics of tubulation induction were somewhat slower. Using EdU-labeled HPV-16 PsVs, we can also readily detect the presence of HPV-16 PsVs within these MICAL-L1-positive vesicles, indicating that the incoming virus is in fact being trafficked through these structures. Interestingly, this tubulation can also be induced following infection with HPV-16 VLPs, indicating that this is not linked to a specific function of the L2 protein. These assays were done routinely using 150 viral genome equivalents per cell (vge/cell) in order to be able to readily visualize incoming virus; however, titration down to 30 vge/cell also induced clear endosomal tubulation (data not shown), indicating that this activity is not due to simply exposing cells to very high levels of viral particles.

MICAL-L1 has been reported to be required for the formation of the elongated tubular endosomal network (ETEN) that contains transferrin, which is taken up by clathrin-dependent endocytosis (33). MICAL-L1 is also present in recycling endosomal tubules (26, 28, 34–36) and also at late endosome ETENs (33). The presence of MICAL-L1 in a diverse population of endosomal tubules suggests that it could have roles to play in both cargo sorting and recycling. However, the recruitment of MICAL-L1 itself would only seem to be relevant as a marker of tubulating endosomes in the context of infection with HPV-16 PsVs, since siRNA ablation of MICAL-L1 had no major impact on the efficiency of HPV-16 infectious entry. This suggests that whatever the roles that MICAL-L1 might play in endocytic transport pathways, they are not actually important for virus infection. Indeed, using Syndapin2 as an alternative marker for tubulating endosomes, we see clear evidence of the induction of tubulating endosomes following HPV-16 infection even in the absence of MICAL-L1. Therefore, at this stage, it is not possible to determine with certainty whether tubulation per se has an important role to play in infectious entry. However, this contrasts quite markedly with the requirement for VAP, which was recently shown to have a critical function at the contact points between membranes of the ER and the endosomal transport vesicle, where it aids in actin nucleation on the budding endosome and retromer-driven transport to the TGN. The retromer regulates the retrograde transport of cargo from endosomes to the Golgi complex (37, 38), and previous studies have shown that the depletion of the retromer leads to a block in the HPV entry pathway prior to the TGN (17, 30). We now show that the lack of the integral ER protein VAP inhibits infection by HPV-16 PsVs and also inhibits HPV-16-induced endosomal tubulation. The loss of VAP also blocks the increase in MICAL-L1 protein levels that is seen following HPV-16 infection. Interestingly, VAP also appears to be important for the infectious entry process of McPyV, indicating that VAP can be important for other viruses that gain infectious entry through endocytic transport pathways.

With the help of the 33L1-7 antibody as a marker of capsid disassembly, we found that infection is not blocked prior to capsid disassembly in VAP-depleted cells. However, we found clear evidence of a dramatic decrease in the ability of the L2/DNA complex to traffic to the TGN in VAP-depleted cells. Instead, we found evidence of HPV-16 PsV-transduced DNA accumulating in previously reported hybrid structures (25), which stain positive for both the retromer marker Vps29 and the Golgi marker TGN46. Interestingly, similar structures are formed when SNX2 expression is ablated (25), and consistent with this also having a role in HPV-16 infectious entry, we observed a decrease in HPV-16 transduction of a luciferase reporter when SNX2 was ablated. Taken together, these results indicate that VAP plays an essential role in HPV-16 infectious entry and trafficking of the viral cargo through an ER-dependent process.

Finally, it is intriguing to speculate whether the increase in both tubule length and numbers in the tubular network, as depicted by MICAL-L1 staining, actually reflects an inhibition of vesicular cleavage following HPV-16 infection. Indeed, the depletion of the WASH complex, which is required for endosomal fission, results in the formation of exaggerated tubules, indicating an impairment of endosomal fission (39, 40). Therefore, it is possible that infection with HPV-16 PsVs promotes endosomal tubulation through reducing the rate of endosomal fission. Further studies will be required to determine how incoming virions affect endocytic tubulation and what advantages this may offer for the virus.

Taken together, these results indicate that HPV-16 infection profoundly affects the overall levels and extent of endosomal tubulation. Furthermore, the completion of infectious entry and access to the TGN are critically dependent upon the ER-associated VAP complex, which ensures ER-mediated cleavage and the subsequent transport of virion-containing vesicles to the TGN.

MATERIALS AND METHODS

Cell culture.

HEK293TT human embryonic kidney cells, HaCaT human skin keratinocytes, NIKS, HeLa cervical cancer cells, and VAP DKO HeLa cells (kindly provided by Pietro de Camilli) (25) were grown in Dulbecco's modified Eagle's medium (DMEM) containing 10% fetal calf serum (FBS), penicillin-streptomycin (100 U/ml), and glutamine (300 μg/ml).

Plasmids and antibodies.

Plasmid p16shell.L2-3xFLAG-thrombin-HA expresses HPV-16 L1 and L2 in a codon-optimized bicistronic manner, where L2 is FLAG and hemagglutinin (HA) tagged (41). Plasmid pGL3 luci, which carries the firefly luciferase gene, was purchased from Promega. The pXULL VLP is generated by deleting the L2 sequence from wild-type pXULL and expresses HPV16 L1 only. McPyV capsid protein expression plasmids and phGluc, which carries a Gaussia luciferase reporter gene, were kind gifts from Chris Buck.

In this study, the following antibodies were used: mouse anti-MICAL-L1 (Novus Biologicals), rabbit anti-PACSIN2/Syndapin2 (Abcam), rabbit anti-Giantin (Abcam), rabbit anti-TGN46 (Abcam), mouse anti-VAPB (Abcam), mouse anti-Vps29 (D-1; Santa Cruz), mouse anti-SNX2 (F-8; Santa Cruz), mouse anti-HPV-16 L1 (Santa Cruz), mouse anti-FLAG (Sigma-Aldrich), mouse anti-tubulin (Abcam), rabbit anti-actinin (Santa Cruz), and mouse anti-33L1-7 (kindly provided by Martin Sapp). AF488 donkey anti-mouse/anti-rabbit, AF647 donkey anti-rabbit, and rhodamine goat anti-mouse (Invitrogen) antibodies were used as secondary antibodies.

Cell transfections.

HeLa and VAP DKO HeLa cells were seeded into 6-well plates at a density of 7 × 10⁴ cells per well. After attachment, the cells were transfected with siRNA against MICAL-L1, siRNA against SNX2, or scrambled siRNA by using Lipofectamine RNAiMax (Invitrogen). To knock down MICAL-L1 and SNX2, ON-TARGETplus SMARTpooL (Dharmacon) siRNAs were used, whereas siSTABLE nontargeting siRNA (Dharmacon) was used as a control (scramble).

Pseudovirion production and labeling.

HPV-16 and McPyV PsVs with a packaged luciferase reporter gene (pGL3 luci and phGluc, respectively) were generated in HEK293TT cells as described previously (42). HPV-16 L1 VLPs were generated in the same manner as for PsVs. PsVs were run on SDS-PAGE gels to determine the purity and concentration of the samples. Bovine serum albumin (BSA) was used at different concentrations as a standard. For EdU labeling, growth medium was supplemented with 25 μM EdU at 12 h posttransfection during PsV production. Labeling of PsVs with a fluorophore (AF488) was performed according to the manufacturer's protocol (Molecular Probes). The packaged pGL3 DNA was quantitated by real-time PCR, using a standard curve of reporter plasmid DNA. McPyV PsVs and VLPs were used in amounts equivalent to those of HPV-16 PsVs.

Infectivity assays.

To check infectivity in cells where MICAL-L1 or SNX2 were knocked down by siRNA treatment, and VAP DKO HeLa cells, infection was performed for 48 h with a concentration of approximately 50 vge/cell, as described previously (15).

Firefly luciferase activity was monitored after 48 h as a measure of infection by using a luciferase assay system kit (Promega). Luciferase readings obtained for scrambled siRNA-treated cells and WT HeLa cells were normalized to 100%, which was then used to calculate the percent change in luciferase activity in siMICAL-L1- and siSNX2-treated cells and VAP DKO HeLa cells. Total cell protein extracts were quantified to ensure the use of equal protein inputs in the luciferase measurements. Lysates from the luminometry assays were used to measure the efficiency of MICAL-L1 or SNX2 knockdown by Western blotting.

HPV-16 PsV trafficking assay.

HeLa, HaCaT, or NIKS cells were seeded into a 6-well plate on sterile glass coverslips at a density of 2.5 × 10⁵ cells per well. Cells were infected with PsVs with EdU-labeled reporter DNA or AF488-labeled PsVs at 150 vge/cell and agitated at 4°C for 1 h to allow virus attachment to the cells. Cells were washed with phosphate-buffered saline (PBS), supplemented with growth medium, and incubated at 37°C. At different time points (2 h, 8 h, and 24 h postinfection), cells were fixed with 3.7% paraformaldehyde for 15 min at room temperature. HeLa cells were permeabilized in 0.1% Triton (diluted in PBS) for 10 min. HaCaT and NIKS cells were permeabilized in 0.2% saponin with 3% BSA for 30 min at room temperature. Cells were then washed with PBS and incubated with 0.1 M glycine to reduce background staining. Incubation with primary antibodies was performed for 1.5 h at 37°C. After washing, cells were incubated with secondary antibody for 1 h at 37°C.

For the detection of EdU-labeled encapsidated DNA, HeLa cells were permeabilized in 0.5% Triton (in PBS) for 10 min. After incubation with primary and secondary antibodies, cells were washed in 3% BSA (in PBS) and incubated for 30 min at room temperature with the Click-iT reaction mixture (Clik-iT EdU imaging kit; Molecular Probes). Cells were washed again with 3% BSA (in PBS) and water before being mounted onto glass slides. Staining was visualized by using a Zeiss Axiovert 100 M microscope, and images were analyzed by using an LSM image browser that supports the LSM 510 confocal unit.