ABSTRACT
Herpes Simplex viruses (HSV-1 and HSV-2) encode up to 15 glycosylated and unglycosylated envelope proteins. Four of these, gB, gH, gL, and gD, are essential for entry and mediate cell-cell fusion when co-expressed in uninfected, receptor-bearing cells. However, their contributions to HSV-1 tropism and the selection of entry routes are unclear. To begin addressing this, we previously pseudotyped VSV lacking its native glycoprotein, G, with HSV-1 glycoproteins gB, gH, gL, and gD. This novel VSVΔG-BHLD pseudotype recapitulated several aspects of HSV-1 entry: it could enter murine C10 cells, required gB, gH, gL, gD, and a cellular receptor for entry, and was sensitive to neutralization by gB and gH/gL antibodies. Here, we screened six additional HSV-1-susceptible cell lines and found that only two, C10 and CHO-HVEM cells, reproducibly supported a receptor-dependent entry by VSVΔG-BHLD. We then compared VSVΔG-BHLD and HSV-1 entry routes into these two cell lines using a combination of chemical and genetic inhibitors of cellular uptake pathways. We discovered that the VSVΔG-BHLD pseudotype not only has a narrower tropism but also uses entry pathways different from those used by HSV-1. We conclude that while the four essential HSV-1 entry glycoproteins enable entry in certain contexts, they are insufficient for entry into any HSV-1-susceptible cell nor do they specify native HSV-1 entry routes. We hypothesize that the HSV-1 envelope proteins outside the essential four (so-called “non-essential”) contribute towards the tropism and the selection of native HSV-1 entry routes. Our work draws attention to the need for systematic investigation of the HSV-1 entry mechanisms and the roles of the envelope proteins that were long considered non-essential in the selection of target cells, routes of entry, and pathogenesis.
AUTHOR SUMMARY Different viruses enter cells by diverse routes, but how that choice is made is not always clear. Understanding the mechanisms behind these choices is vital for finding strategies to prevent viral infections. In enveloped viruses, viral proteins embedded in the envelope accomplish this task. While most enveloped viruses encode one or two envelope proteins, Herpes Simplex viruses (HSV) encode up to 15. Four of these are deemed essential for entry (gB, gH, gL, and gD) whereas the rest have been termed non-essential. While these four proteins are essential, their contributions to HSV-1 cellular tropism and entry pathways have not been fully elucidated. Here, we generated virions that have only the four essential HSV-1 glycoproteins on their surface. We show that the VSVΔG-BHLD pseudotype has a narrower tropism than HSV-1 and uses different entry pathways. Thus, the four essential HSV-1 entry glycoproteins alone do not define HSV-1 tropism or specify native entry routes. We hypothesize that the HSV-1 envelope proteins outside the essential four may contribute towards tropism and entry route selection. Our work emphasizes the need to investigate the roles of the so-called non-essential envelope proteins in HSV entry. This is important because HSV enters natural target cells, epithelial cells and neurons, by different, poorly defined routes. Mechanistic understanding of HSV entry is essential for understanding its pathogenesis and developing new strategies to prevent HSV entry and spread.
INTRODUCTION
All viruses must enter cells to initiate infection, and different viruses accomplish this task by different mechanisms: some penetrate cells at the plasma membrane while others hijack host endocytic pathways. Enveloped viruses – those in which the nucleocapsid is surrounded by a cell-derived lipid bilayer – penetrate the cells by merging their lipid envelopes with a cellular membrane, either the plasma membrane or the membrane of an endocytic vesicle following endocytosis (Fig 1) [reviewed in (1, 2)]. Viruses typically have preferred entry routes, the choice of which is not always clear, and some enter different target cells by different routes. Understanding what routes viruses use to enter cells and how they select them may help inform strategies to prevent viral infections.
Herpes Simplex viruses (HSV-1 and HSV-2) are enveloped alphaherpesviruses that infect much of the world’s population for life, causing a panoply of diseases ranging from painful to life-threatening (3, 4). An enduring mystery of HSV entry is how and why the virus enters different cells by different routes: direct fusion with the plasma membrane [e.g., neurons, Vero cells, Hep2 cells, reviewed in (5)] or by endocytosis and subsequent fusion with the endosomal membrane [e.g., keratinocytes, corneal epithelial cells (6, 7)].
HSV-1 encodes 15 viral envelope proteins, 12 glycosylated and three unglycosylated [reviewed in (8, 9)]. HSV-1 entry by any route requires the coordinated efforts of four glycoproteins – gB, gH, gL, and gD, which are termed essential with regard to entry (10–12) – and a cellular gD receptor [reviewed in (5, 13, 14)]. Binding of gD to one of its three cellular receptors, nectin-1, herpesvirus entry mediator (HVEM), or 3-O-sulfated heparan sulfate (3-OS-HS) (15–17), is thought to activate the conserved heterodimer gH/gL (18, 19). In turn, gH/gL presumably interacts with and activates gB (20, 21), the conserved fusogen that mediates the fusion of the viral envelope with the cellular membrane [reviewed in (13, 14)]. The remaining 11 envelope proteins encoded by HSV-1 are incorporated into virions (9) and contribute to replication by aiding in egress (22) and immune evasion (23, 24), in addition to entry [reviewed in (8)].
In addition to being required for entry, gB, gH, gL, and gD are also sufficient for cell-cell fusion of uninfected receptor-bearing cells expressing these four glycoproteins (25, 26). However, until recently, it was unknown whether these four glycoproteins were sufficient for entry. To begin addressing this question, we previously generated VSV lacking its native glycoprotein G and pseudotyped with HSV-1 gB, gH, gL, and gD through trans-complementation, which we refer to as the VSVΔG-BHLD pseudotype (27). The VSVΔG-BHLD pseudotype efficiently entered C10 cells (B78 murine melanoma cells expressing nectin-1), and its entry recapitulated several important features of HSV-1 entry into susceptible cells, namely, the requirement for gB, gH, gL, gD, and a gD receptor, as well as sensitivity to anti-gB and anti-gH/gL neutralizing antibodies (27). However, this study left unclear whether the VSVΔG-BHLD pseudotype could enter any HSV-1 susceptible cell types or specify native entry routes.
Here, we expanded our studies to six additional HSV-1 susceptible cell lines. We found that unlike HSV-1, VSVΔG-BHLD reproducibly entered only two cell lines, C10 and CHO-HVEM, by non-native routes. Therefore, although the four essential HSV-1 entry glycoproteins alone can mediate entry in certain contexts, they do not enable entry into any HSV-1-susceptible cell and do not specify native entry routes. Differences in tropism and routes of entry could not be accounted for by either cell-surface receptor levels, their nature (nectin-1 vs. HVEM), the relative amounts of gB, gH, gL, and gD, virion morphology (VSV vs. HSV-1), or some other aspect of VSV biology. Therefore, we hypothesize that the so-called non-essential HSV-1 envelope proteins, which are missing from the VSVΔG-BHLD pseudotype, are important in specifying both HSV-1 tropism and its routes of cell entry.
RESULTS
VSVΔG-BHLD pseudotypes enter a limited repertoire of HSV-1 susceptible cells
To determine the tropism of the VSVΔG-BHLD pseudotype, we selected several commonly used HSV-1-susceptible cell lines, including four primate cell lines, HeLa, HaCaT, Vero, and SH-SY5Y, and three derivatives of rodent cell lines B78H1 (mouse) and CHO-K1 (hamster) engineered to stably express human nectin-1 [B78H1-nectin-1 (C10) and CHO-nectin-1] or HVEM (CHO-HVEM) (28, 29). Parental B78H1 and CHO-K1 cell lines served as receptor-negative controls. As expected, HSV-1 entered all receptor-bearing cells but not receptor-negative B78H1 or CHO-K1 cells (Fig 2A). VSVΔG-BHLD pseudotype also readily infected C10 cells but not parent B78H1 cells (Fig 2B), consistent with previous observations (27). VSVΔG-BHLD pseudotype also infected CHO-HVEM and CHO-nectin-1 cells in a receptor-dependent manner albeit with much lower efficiency (Fig 2B). However, no measurable infection by the VSVΔG-BHLD pseudotype was observed with HeLa, HaCaT, Vero, or SH-SY5Y cells (Fig 2B) even at MOI of 10 (S1 Fig).
To rule out contributions of VSV particle morphology, cells were also infected with two VSV pseudotype controls, VSVΔG-G and VSVΔG-PIV5. VSVΔG-G is VSVΔG pseudotyped in trans with native VSV glycoprotein G and is known to enter susceptible cells by a clathrin-mediated endocytosis (CME) (30). VSVΔG-PIV5 is VSVΔG pseudotyped with entry glycoproteins HN and F from parainfluenza virus 5 (PIV5), which enters susceptible cells by fusion at the plasma membrane (31) just as PIV5 itself. Both VSVΔG-G and VSVΔG-PIV5 pseudotypes infected all 9 tested cell lines, with varying efficiency (Fig 2C, D), which suggested that limited tropism of VSVΔG-BHLD pseudotype cannot be explained by VSV particle morphology or some other aspect of VSV biology.
Narrow tropism of VSVΔG-BHLD pseudotypes cannot be fully accounted for by variations in cell surface receptor levels
Levels of HSV-1 gD receptors nectin-1 and HVEM vary across cell lines, and susceptibility to HSV-1 infection generally correlates with surface receptor levels (29). To investigate the role of relative cell surface levels of HSV-1 receptors in VSVΔG-BHLD tropism, we measured the amount of nectin-1 and HVEM on the surface of B78H1, C10, CHO-K1, CHO-nectin-1, CHO-HVEM, HeLa, HaCaT, Vero, and SH-SY5Y cells by flow cytometry (Fig 3). As expected, neither receptor was detected on the receptor-negative cell lines B78H1 and CHO-K1 (Figs 3A, B). C10 and HaCaT cells had the highest levels of nectin-1 whereas intermediate levels of nectin-1 were detected on CHO-nectin-1, HeLa, Vero and SH-SY5Y cells (Fig 3A). As expected, CHO-HVEM cells had high levels of HVEM but no detectable nectin-1 on their surface (Fig 3B). In addition to nectin-1, HaCaT, HeLa, and Vero cells also had HVEM on their surfaces (Fig 3B) albeit in low amounts, which suggested that in these cells, nectin-1 likely functions as the primary receptor. Surprisingly, while both C10 and HaCaT cells express high levels of nectin-1, the VSVΔG-BHLD pseudotype efficiently enters only C10 cells.
To correlate VSVΔG-BHLD entry with receptor levels more directly, CHO-K1 cells were transfected with a plasmid encoding either nectin-1 or HVEM. Cell surface levels of either receptor increased with increasing amounts of transfected DNA. While VSVΔG-BHLD entry increased with increasing amounts of nectin-1 DNA, entry maxed out in cells transfected with 10 ng of DNA at around 2%, similar to entry levels observed in the CHO-nectin-1 cell line (∼3%). Additionally, the level of receptor detected in cells transfected with 10 ng of plasmid DNA and the level of receptor in CHO-nectin-1 stable cells are very similar, and support similar levels of entry (∼2% vs. 3% infected cells, respectively). However, increased nectin-1 expression did not result in increased VSVΔG-BHLD entry (S2 Fig). CHO-K1 cells transfected with a plasmid expressing HVEM showed increasing surface expression of the receptor with increasing amounts of DNA but never reached expression levels of the CHO-HVEM stable cell line. While HVEM levels did increase, a similar increase in VSVΔG-BHLD entry was not observed. Together these results suggest that while VSVΔG-BHLD entry requires a gD receptor, neither its surface expression levels, nor its type can account for the narrow tropism of the VSVΔG-BHLD pseudotype.
Differences in HSV-1 and VSVΔG-BHLD tropisms cannot be attributed to differences in the relative amounts of glycoproteins gB, gH, gL, and gD
We next tested whether differences in relative amounts of gB, gH, gL, and gD on HSV-1 vs. VSVΔG-BHLD pseudotype could account for the stark differences in tropism. Purified HSV-1 and VSVΔG-BHLD particles were analyzed for the gB, gH, gL, and gD content by Western blot (Fig 4A), and relative gB:gH:gL:gD ratios were determined by densitometry. In each virus, levels of gH, gL, and gD were normalized to gB. We observed no statistically significant differences in the relative amounts of gH, gL, or gD between HSV-1 and VSVΔG-BHLD virions (Fig 4B), and their ratios were very similar. We conclude that the differences in the relative amounts of gB, gH, gL, and gD in HSV-1 or VSVΔG-BHLD virions do not account for the differences in tropism.
Entry of both VSVΔG-BHLD and HSV-1 into C10 and CHO-HVEM cells occurs by endocytosis
HSV-1 enters different cell lines either by fusion at the plasma membrane or by endocytosis [(32, 33); reviewed in (5)]. As far as the cells used in the present study, HSV-1 enters C10, CHO-nectin-1, CHO-HVEM, HeLa, and HaCaT cells by endocytosis (7, 32, 34) and Vero and SH-SY5Y cells, by fusion at the plasma membrane (7, 33). To determine how the VSVΔG-BHLD pseudotype enters C10 and CHO-HVEM cells, we first treated cells with a hypertonic solution of sucrose, a broad inhibitor of endocytic pathways (35, 36). Entry of both HSV-1 and VSVΔG-BHLD into C10 and CHO-HVEM cells was inhibited by sucrose (Fig 5A, B, E, F) and, thus, occurs by endocytosis, just as entry of HSV-1.
As a control, sucrose also prevented the endocytic uptake of Alexa Fluor 488-labeled transferrin into both C10 and CHO-HVEM cells (Fig 5I). As expected, sucrose did not inhibit entry of VSVΔG-PIV5 into either cell line because PIV5 enters by fusion at the plasma membrane (31), thus serving as a negative control (Fig 5D, H). Entry of VSV occurs by an endocytic route (37–39), and, accordingly, sucrose blocked entry of VSVΔG-G into C10 cells (Fig 5C). Surprisingly, it did not block entry into CHO-HVEM cells (Fig 5G), suggesting that the inhibitory effect of hypertonic sucrose on VSV-G-dependent entry may be cell-type specific.
VSVΔG-BHLD entry requires dynamin but not clathrin whereas HSV-1 entry requires both
Having established that the VSVΔG-BHLD pseudotype entered cells by endocytosis, we next sought to identify its entry mechanism and systematically compare it to that of HSV-1. To do this, we used a combination of chemical and genetic means of inhibiting different cellular proteins involved in various endocytic uptake pathways (Fig 1). In most cases, two to three inhibitors per pathway were selected to establish the involvement of a particular pathway with greater confidence (40) and to rule out potential unintended effects of the inhibitors.
First, we examined the role of the clathrin-mediated endocytosis (CME), one of the most well studied and highly used endocytic pathways hijacked by viruses for entry [reviewed in (2, 41)]. CME requires both clathrin, a triskelion-shaped protein that assembles cage-like structures underneath the plasma membrane to promote receptor-mediated endocytosis (42), and a GTPase dynamin that mediates scission of the endocytic vesicle (43). We chose three commonly used dynamin inhibitors: Dynasore, Dyngo-4a, and myristyltrimethylammonium bromide (MiTMAB). Dynasore and Dyngo-4a inhibit dynamin GTPase activity, which is necessary for dynamin-mediated membrane fission (35), whereas MiTMAB binds the membrane-interacting pleckstrin homology (PH) domain of dynamin and prevents dynamin association with the membrane (44). As a clathrin inhibitor, we chose Pitstop-2, which selectively blocks CME by preventing ligand association with the clathrin terminal domain (35). Another CME inhibitor, chlorpromazine, which prevents clathrin association with the plasma membrane (35), was also tested. However, at previously published inhibitory concentrations, chlorpromazine was toxic to both C10 and CHO-HVEM cells, judging by the altered cellular morphology. Inhibitory activity of all four compounds was ascertained by their ability to inhibit CME of transferrin (Figs 6E and 7I).
HSV-1 entry into both C10 and CHO-HVEM cell lines was inhibited by all four inhibitors (Figs 6A, C and 7A, E) indicating that HSV-1 enters both cell lines by CME. While entry of VSVΔG-BHLD pseudotypes into C10 and CHO-HVEM cells was sensitive to all dynamin inhibitors (Fig 6B, D), surprisingly, Pitstop-2 did not block VSVΔG-BHLD entry (Fig 7B, F), which ruled out CME as the entry mechanism. Collectively, these observations suggested that HSV-1 enters both C10 and CHO-HVEM cells by CME whereas VSVΔG-BHLD pseudotypes utilize a dynamin-dependent, clathrin-independent entry route. These results were the first indication of possible differences in entry routes of HSV-1 and VSVΔG-BHLD.
As expected, entry of VSVΔG-G into C10 cells was blocked by all three dynamin inhibitors and the clathrin inhibitor (Fig 7C and S3A Fig), confirming previous reports of VSV entry into other cell types by CME (30, 39). Like C10 cells, VSVΔG-G entry into CHO-HVEM cells was blocked by Pitstop-2 (Fig 7G). However, its entry into CHO-HVEM cells was only blocked by two of the dynamin inhibitors, Dynasore and Dyngo-4a, and not by MiTMAB (S3C Fig). Given the results that two of the dynamin inhibitors and clathrin inhibition blocked VSVΔG-G entry into CHO-HVEM cells, it is likely that the entry mechanism is CME. In contrast, VSVΔG-PIV5 entry into CHO-HVEM cells was not blocked by any of the four inhibitors (Fig 7H and S3D Fig), consistent with the previous report of entry into other cell types by fusion at the plasma membrane (31). Surprisingly, VSVΔG-PIV5 entry into C10 cells was blocked by Dyngo-4a (S3B Fig), hinting at a potential role for dynamin in PIV5 entry into C10 cells. However, given that neither the other dynamin inhibitors, Dynasore and MiTMAB, nor Pitstop-2 blocked entry (Fig 7H and S3B Fig), Dyngo-4a may be blocking VSVΔG-PIV5 entry by an off-target effect.
Cholesterol is important for entry of both HSV-1 and VSVΔG-BHLD
Given that VSVΔG-BHLD entry into both C10 and CHO-HVEM cells did not require clathrin, we next evaluated contributions of known clathrin-independent endocytic (CIE) mechanisms. Caveolin-dependent endocytosis is a major CIE route (45) that is hijacked by viruses such as SV40 or Japanese encephalitis virus (46, 47). Caveolin-1 is cellular protein that, similarly to clathrin, promotes membrane curvature and subsequent endocytosis through the formation of caveolae (48). Caveolin-dependent entry requires plasma membrane cholesterol for proper caveolin-1 association with the membrane (48, 49).
Entry of both HSV-1 and VSVΔG-BHLD into C10 and CHO-HVEM cells decreased when cholesterol was removed from cellular membranes using a cholesterol-depleting agent, methyl-β-cyclodextrin (MβCD) (Fig 8A-D). Efficiency of cholesterol depletion was confirmed by showing that association of cholera toxin subunit B (CTB) with C10 and CHO-HVEM cells, which is dependent on membrane cholesterol (50), was decreased with MβCD treatment (S4F Fig). Entry of VSVΔG-G was insensitive to cholesterol depletion (S4A, C Fig), as observed in other cell types (51, 52). By contrast, cholesterol was important for VSVΔG-PIV5 entry into C10 cells (S4B Fig) but not into CHO-HVEM cells (S4D Fig), suggesting that the role of cellular cholesterol in PIV5 entry is cell-type dependent.
Entry of neither HSV-1 nor VSVΔG-BHLD into CHO-HVEM cells was reduced by the knockdown of caveolin-1 (Fig 8E), similarly to the VSVΔG-G and VSVΔG-PIV5 (53) control viruses (S4E Fig). Successful knockdown was verified by western blot (Fig 8F). Surprisingly, no caveolin-1 was detected in C10 cells (Fig 8G). To rule out species-dependent recognition of murine vs. hamster caveolin-1, 3T12 cells (murine fibroblasts) were tested for the presence of caveolin-1. Indeed, caveolin-1 was detected in 3T12 cells (Fig 8G), confirming that C10 cells express no detectable caveolin-1. We conclude that cellular cholesterol plays an important role in the entry of both HSV-1 and VSVΔG-BHLD but that neither virus utilizes caveolin-1-mediated endocytosis for entry into C10 and CHO-HVEM cells.
Neither HSV-1 nor VSVΔG-BHLD utilize macropinocytosis for entry, but NHE1 and Rac1 are important for VSVΔG-BHLD entry in a cell-dependent manner
We next evaluated the potential involvement of macropinocytosis, another common endocytic viral entry pathway. We selected three known inhibitors of macropinocytosis: the actin filament disruptor cytochalasin D (CytoD), the Na+/H+-exchange (NHE) inhibitor 5-(N-Ethyl-N-isopropyl)amiloride (EIPA), and the Rac1 inhibitor NSC23766 (35). EIPA blocks macropinocytosis by blocking Na+/H+ exchange proteins (NHE), which decreases the intracellular pH and inhibits small GTPase function important for macropinocytosis (54). NSC23766 blocks macropinocytosis by inhibiting the activity of the small GTPase Rac1 (55). Inhibitory activity of all three compounds was confirmed by their ability to inhibit macropinocytosis of rhodamine-B-labeled 70-kDa dextran (Fig 9E).
HSV-1 entry into either C10 or CHO-HVEM cells was not appreciably inhibited by any of the macropinocytosis inhibitors (Fig 9A, C). By contrast, VSVΔG-BHLD entry was reduced by NSC23766 and, to a lesser extent, EIPA (though not statistically significant) in both C10 and CHO-HVEM cells but not by cytochalasin D (Fig 9B, D). While the inhibitory effect of the NSC23766 and EIPA would appear to implicate macropinocytosis as the route of VSVΔG-BHLD entry into C10 and CHO-HVEM cells, the lack of actin involvement argues against it. This is because assembly of filamentous actin is essential for the formation of membrane ruffles during macropinocytosis (56) and is, thus, the major hallmark of macropinocytosis (57). Without actin polymerization, cells cannot support fluid phase uptake processes such as macropinocytosis. Likewise, we did not observe any appreciable co-localization of the VSVΔG-HLD particles with 70 kDa rhodamine B labeled dextran, a fluid phase uptake marker (S5 and S6 Figs). Taken together, our results suggest that neither HSV-1 nor VSVΔG-BHLD utilize macropinocytosis for entry into C10 or CHO-HVEM cells. Efficient VSVΔG-BHLD entry, however, requires NHE and Rac1 activity in both cell lines, and we hypothesize that these proteins play a role in VSVΔG-BHLD entry independently of their role in macropinocytosis.
Although VSV enters cells by CME rather than macropinocytosis, it has been shown to require filamentous actin to achieve full engulfment of the viral particle by the plasma membrane during endocytosis (30). Indeed, VSVΔG-G entry into C10 cells was modestly reduced by cytochalasin D and EIPA but not NSC23766 (S7A Fig) whereas its entry into CHO-HVEM cells was somewhat reduced by EIPA but not cytochalasin D or NSC23766 (S7C Fig). As VSVΔG-PIV5 fuses at the plasma membrane, its entry was not blocked with any of the three inhibitors of macropinocytosis, and, in fact, was increased in the presence of EIPA in C10 cells (S7B, D Fig).
HSV-1 and VSVΔG-BHLD differ in their requirements for Rab GTPases for entry
Endocytic entry by many viruses also requires small GTPases known as Rabs. Rab GTPases are important for the formation of specific endosomal compartments in the cell (58). Given that different viruses penetrate endocytic membranes at distinct endosomal maturation stages – for example, VSV fuses with membranes of early endosomes whereas influenza A virus fuses with membranes of late endosomes (59) – proper formation of these endosomal compartments is essential for viral entry. GTPases Rab5 and Rab7 are important for the maturation and formation of early endosomes and late endosomes/multi-vesicular bodies (MVBs), respectively (60). Overexpression of dominant negative (DN) forms of either Rab5 (Rab5DN) or Rab7 (Rab7DN) suppress early and late endosome formation, respectively (61, 62).
To identify the endosomal compartment(s) required for HSV-1 or VSVΔG-BHLD entry, C10 and CHO-HVEM cells were transfected with fluorescently tagged Rab5DN or Rab7DN expression constructs and then infected. Overexpression of either Rab5DN or Rab7DN did not significantly reduce HSV-1 entry into either cell line (Fig 10A, E) or entry of VSVΔG-BHLD into CHO-HVEM cells (Fig 10F). However, VSVΔG-BHLD entry into C10 cells was reduced in the presence of Rab5DN and, to a lesser extent, Rab7DN (Fig 10B). These results suggest that HSV-1 entry into both cell lines and VSVΔG-BHLD entry into CHO-HVEM cells are either independent of the endosomal maturation state or occur very early after internalization, prior to the formation of early or late endosomes. By contrast, VSVΔG-BHLD likely enters the cytosol of C10 cells out of early endosomes. This observation marked the first instance of a difference in VSVΔG-BHLD entry into C10 vs. CHO-HVEM cells.
As expected, entry of VSVΔG-G into both C10 and CHO-HVEM cells was reduced in the presence of Rab5DN (Fig 10C, G), in accordance with reports of VSV fusing with membranes of early endosomes (39). VSVΔG-PIV5 entry into C10 cells was insensitive to either Rab5DN or Rab7DN (Fig 10D). While VSVΔG-PIV5 entry into CHO-HVEM cells was reduced by both Rab5DN and Rab7DN in a statistically significant manner (Fig 10H), the differences between mean values were less than 10%. Therefore, Rab5DN and Rab7DN have a minimal effect on VSVΔG-PIV5 entry as expected for a virus that fuses with the plasma membrane.
Small GTPase Arf6 regulates entry of HSV-1 but not VSVΔG-BHLD
Another small GTPase, ADP-ribosylation factor 6 (Arf6), which is involved in regulating vesicular trafficking (63), regulates the endocytic entry of HIV, Coxsackievirus, and Vaccinia virus (40, 64, 65). To probe the role of Arf6 in entry, cells were treated with NAV-2729, which blocks Arf6 interaction with guanine exchange factors (GEFs) thereby preventing its activation (66). HSV-1 entry into both C10 and CHO-HVEM cells was inhibited by NAV-2729 (Fig 11A, E) whereas VSVΔG-BHLD entry into either cell line was not inhibited, which suggests that HSV-1 entry requires Arf6 activity but entry of VSVΔG-BHLD does not (Fig 11B, F).
As anticipated, VSVΔG-PIV5 entry was not inhibited by NAV-2729 in either C10 or CHO-HVEM cells (Fig 11D, H). However, VSVΔG-G entry was inhibited by NAV-2729 in both C10 and CHO-HVEM cells, suggesting that Arf6 could be involved in VSV endocytosis (Fig 11C, G). These results point to a previously unappreciated role of Arf6 in HSV-1 and VSV entry.
Entry of VSVΔG-BHLD pseudotype requires endosomal acidification in a cell-dependent manner
As HSV-1 and VSVΔG-BHLD enter cells by endocytosis, we assessed the role of endosomal acidification in entry. Previous work showed that HSV-1 entry into CHO-HVEM cells was sensitive to inhibitors of endosomal acidification ammonium chloride (NH4Cl) and monensin (32), and thus likely required endosomal acidification. Conversely, HSV-1 entry into C10 cells was insensitive to either NH4Cl or bafilomycin A1 (BFLA), another inhibitor of endosomal acidification (34).
To investigate the role of endosomal acidification in entry, we used BFLA, NH4Cl, and monensin. All three inhibitors effectively blocked endosomal acidification as evidenced by decrease in Lysotracker fluorescence in inhibitor-treated cells (Fig 12E). VSVΔG-G, which requires low pH as a trigger for membrane fusion during endocytic entry, was sensitive to all three inhibitors (S8A, C Fig). In contrast, VSVΔG-PIV5, which enters cells by fusion at the plasma membrane, was insensitive to any of the inhibitors (S8B, D Fig).
HSV-1 entry into C10 cells was inhibited by NH4Cl and monensin but not by BFLA (Fig 12A), the latter in agreement with a previous report (34). HSV-1 entry into CHO-HVEM cells was inhibited by NH4Cl and monensin, which was consistent with the previous report (32), but not by BFLA (Fig 12C). VSVΔG-BHLD entry into C10 cells was inhibited by all three endosomal acidification inhibitors (Fig 12B) whereas VSVΔG-BHLD entry into CHO-HVEM cells was inhibited only by NH4Cl (Fig 12D).
Although HSV-1 entry into C10 and CHO-HVEM cells was sensitive to NH4Cl and monensin, the lack of inhibition in the presence of BFLA, an endosomal V-ATPase inhibitor (Fig 1), suggests that the effects of NH4Cl and monensin on HSV-1 entry may be independent of their effects on endosomal acidification. VSVΔG-BHLD entry into C10 cells was sensitive to all three endosomal acidification inhibitors and likely depends on endosomal acidification. Finally, VSVΔG-BHLD entry into CHO-HVEM cells could only be inhibited by NH4Cl and is unlikely to require endosomal acidification. We hypothesize that the discrepancy in the inhibitory effects among the three inhibitors could potentially be due to the distinct mechanisms by which BFLA, NH4Cl, and monensin raise endosomal pH.
DISCUSSION
Decades of research have established glycoproteins gB, gH, gL, and gD as essential for HSV-1 entry (10–12). These four glycoproteins are also sufficient for cell-cell fusion when co-expressed in uninfected, receptor-bearing cells (25, 26). While these studies greatly increased our understanding of the HSV-1 entry and fusion mechanisms, the specific contributions of these four glycoproteins to cellular tropism and the selection of entry routes remain incompletely understood in part due to the presence of 11 other proteins in the HSV-1 envelope. To begin addressing this, we utilized a VSV-based pseudotype containing HSV-1 gB, gH, gL, and gD. Being devoid of other HSV-1 proteins, the VSVΔG-BHLD pseudotype provided a deconvoluted platform to identify contributions of the core set of four essential glycoproteins to HSV-1 cellular tropism and the selection of entry routes.
Previously, we showed that the VSVΔG-BHLD pseudotype efficiently entered C10 cells and that its entry recapitulated several important features of HSV-1 entry into susceptible cells: the requirement for gB, gH, gL, gD, and a gD receptor, and sensitivity to anti-gB and anti-gH/gL neutralizing antibodies (27). Here, we expanded this study to six additional HSV-1-susceptible cell lines and found that out of seven, only C10 and CHO-HVEM cells supported appreciable VSVΔG-BHLD entry. We then showed that the VSVΔG-BHLD pseudotype entered these two cell lines by different mechanisms as judged by the differences in sensitivity to endocytic inhibitors (Fig 13). Therefore, while gB, gH, gL, and gD are sufficient for entry in certain contexts, they do not enable entry into any HSV-1-susceptible cell type and do not specify native entry routes. We hypothesize that the so-called non-essential HSV-1 envelope proteins contribute towards the tropism and the selection of entry routes, likely, in a cell-specific manner.
VSVΔG-BHLD pseudotype has a narrower cellular tropism than HSV-1
HSV-1 can infect a wide range of receptor-bearing cell lines and types from many different species (8). By contrast, VSVΔG-BHLD had a narrower tropism, efficiently entering only 2 out of 7 cell lines used. Puzzlingly, the VSVΔG-BHLD pseudotype entered two engineered rodent cell lines but none of the four human and primate cell lines typically used in HSV-1 studies, HeLa, Vero, HaCaT, or SH-SY5Y even at MOI of 10. The lack of VSVΔG-BHLD entry did not correlate with the HSV-1 entry route into these cells, endocytosis in HeLa and HaCaT (7, 32) vs. plasma membrane in Vero and SH-SY5Y (7, 33). The reasons for the observed differences in tropism are yet unclear but do not appear to be due to differences in ratios of gB, gH, gL, and gD or VSV biology. Therefore, we hypothesize that the narrower tropism of VSVΔG-BHLD is instead due to the absence of one or more of the 11 HSV-1 envelope proteins outside of the core set of four. Indeed, some of these proteins have been shown to increase HSV-1 entry efficiency. For example, deletion of glycoprotein K (gK) significantly reduced HSV-1 entry into Vero cells (67). Another HSV-1 glycoprotein, gC, aids viral attachment by binding heparan sulfate moieties of cell surface proteoglycans (68) and promotes efficient entry into cells that HSV-1 enters by an endocytic route (69). Thus, it is plausible that HSV-1 glycoproteins outside of the core set of four could modulate HSV-1 entry efficiency by tuning the cellular tropism of the virus. In other words, the more efficiently HSV-1 enters a given cell type, the more likely that cell type is to be successfully infected.
VSVΔG-BHLD pseudotype is internalized differently from HSV-1
If HSV-1 gB, gH, gL, and gD were sufficient to specify the native routes of HSV-1 entry, then we would have expected the VSVΔG-BHLD pseudotype to utilize the same entry routes into C10 and CHO-HVEM as HSV-1. Indeed, entry of both viruses occurred by endocytosis, required dynamin and cellular cholesterol, and was independent of caveolin-1 or actin polymerization. However, further investigation showed that the entry requirements of VSVΔG-BHLD ultimately differed from those of HSV-1 in several regards (Fig 13).
Unlike HSV-1 entry, VSVΔG-BHLD entry was insensitive to the clathrin inhibitor Pitstop-2, which implicated CIE, rather than CME, as the entry route. Macropinocytosis is a CIE used by a number of viruses. But while VSVΔG-BHLD entry into both C10 and CHO-HVEM cells was sensitive to two inhibitors of macropinocytosis, EIPA and NSC23766, it was insensitive to the inhibitor of actin polymerization, cytochalasin D. Given the central role of actin polymerization in macropinocytosis (57), these data suggest that macropinocytosis is not the primary entry mechanism for VSVΔG-BHLD pseudotype. Indeed, VSVΔG-BHLD particles did not colocalize with a fluid phase uptake marker 70 kDa rhodamine B labeled dextran to an appreciable extent.
If the VSVΔG-BHLD pseudotype does not enter cells by macropinocytosis, why is its entry sensitive to EIPA and NSC23766? One possibility is that their respective targets, Na+/H+ exchangers and Rac1, could contribute to VSVΔG-BHLD entry by regulating, for example, the function of other cellular GTPases (54) or other downstream targets of Rac1 [reviewed in (70)]. An alternative possibility is that EIPA inhibits VSVΔG-BHLD entry due to its previously described pleiotropic effect manifested as gross reorganization of the endosomal network and changes in Na+ and H+ gradients in the cell (71, 72). Similarly, inhibition of Rac1 by NSC23766 could have a pleotropic effect because Rac1, in addition to its roles in regulating the actin cytoskeleton, is also involved in several other cellular processes (73).
VSVΔG-BHLD pseudotype and HSV-1 differ in late-stage entry requirements
Many viruses that enter by endocytosis, for example, influenza A and VSV, rely on Rab-GTPase-dependent endosomal maturation and endosomal acidification (59). VSVΔG-BHLD entry into C10 cells required Rab5, a marker of the early endosomes, and endosomal acidification. Conversely, VSVΔG-BHLD entry into CHO-HVEM cells required neither Rab5, nor Rab7, nor endosomal acidification. Entry of HSV-1 into both C10 and CHO-HVEM cells appeared independent of Rab5, Rab7, or endosomal acidification, similar to VSVΔG-BHLD entry into CHO-HVEM cells. While both Rab5 and Rab7 appeared disposable for HSV-1 entry into either cell line, Arf6, a small GTPase involved in endosomal trafficking including CME and CIE (63), may play a role. How Arf6 promotes HSV-1 entry is yet unclear, considering its numerous downstream effectors, including lipid modifying enzymes, proteins involved in endosome trafficking, and GTPase activating proteins (GAPs) and guanine exchange factors (GEFs) for other GTPases [reviewed in (63)]. Interestingly, Arf6 did not seem important for VSVΔG-BHLD entry.
Surprisingly, we observed that HSV-1 entry into both C10 and CHO-HVEM cells was insensitive to BFLA yet inhibited by NH4Cl and monensin. While BFLA specifically blocks endosomal acidification by inhibiting endosomal V-ATPase proton pumps, NH4Cl and monensin can also affect other cellular processes, e.g., vacuolization or organelle swelling. Therefore, we hypothesize that HSV-1 entry into either the C10 or the CHO-HVEM cells does not require endosomal acidification. Sensitivity of HSV-1 entry to NH4Cl and monensin could, instead, be due to their ability to interfere with other cellular processes. Both compounds alter ion content of the endosomes and cause vacuolization (74), so a change in endosomal ion concentration may reduce the ability of HSV-1 to fuse with the endosomal membrane. Indeed, binding of HSV-1 to the cell surface releases intracellular Ca2+ stores (75) and increases intracellular levels of Cl- ions (76), both of which appear important for subsequent entry.
One notable difference between VSVΔG-BHLD and HSV-1 entry into C10 cells was the apparent reliance on endosomal acidification of the former but not of the latter virus. Previous work suggested that HSV-1 entry into C10 cells does not require endosomal acidification (34). Indeed, we found that HSV-1 entry into C10 cells was insensitive to inhibition by bafilomycin A1 (BFLA). However, VSVΔG-BHLD entry into C10 cells was blocked by BFLA treatment.
Putting these results together, we hypothesize that HSV-1, which does not require Rab5/7 or endosomal acidification of entry, fuses with the endosomal membrane prior to maturation of the newly formed vesicle into an early endosome [pH ∼6.2 (77)] (Fig 13). This latter scenario is consistent with the rapid nature of HSV-1 entry into both C10 and B78A10 cells (B78 murine melanoma cells expressing HVEM) cells (t1/2 = 8-10 minutes) (34). Alternatively, endosomal maturation status does not influence HSV-1 fusion with the membrane of the endocytic vesicle. VSVΔG-BHLD entry into CHO-HVEM cells, likewise, does not require Rab5/7 or endosomal acidification. Therefore, it is likely that VSVΔG-BHLD fusion with the endosomal membrane occurs prior to delivery of the endocytic vesicle to an early endosome. In contrast, VSVΔG-BHLD entry into C10 cells requires both Rab5 and endosomal acidification. In that case, we hypothesize that VSVΔG-BHLD fuses with membranes of early endosomes. This would suggest that in C10 cells, other HSV-1 envelope proteins enable fusion prior to endosomal acidification.
Utility of the VSV-based pseudotyping approach for studying HSV-1 entry
The utility of the VSV-based pseudotyping platform has been demonstrated on many occasions throughout the years. It has been used to study the entry mechanisms of several BSL-4 pathogens (e.g. Ebola, Marburg, and Junín viruses) that otherwise are limited to being studied in appropriate containment facilities (78). Importantly, the entry mechanisms of these pseudotypes have been shown to be similar to their respective native viruses (79), indicating that the pseudotypes typically serve as useful proxies for studying the entry of the respective native viruses. Unlike other pseudotypes and their respective native viruses, VSVΔG-BHLD pseudotype does not mimic HSV-1 entry. We hypothesize that this is due to the lack of other envelope proteins that may influence cellular tropism and entry pathways. However, other possibilities should also be considered.
While neither particle morphology nor the ratio of gB, gH, gL, and gD appear to account for the differences in HSV-1 vs. VSVΔG-BHLD entry, the differences in the lipid composition may do so. VSV and HSV-1 acquire their envelopes from different sources, the plasma membrane (PM) for VSV (78) vs. trans-Golgi network (TGN) or endosome-derived vesicles for HSV-1 (80). Although lipidomics studies suggests that the PM, TGN, and endosomes have very similar lipid composition (81), small differences in envelope lipid composition may contribute to differences in entry. While interesting, this line of inquiry is beyond the scope of the present study.
Collectively, our results implicate the so-called non-essential HSV-1 envelope proteins in both the cellular tropism and the selection of viral entry routes. Previous studies have hinted at the ability of some of these proteins to influence entry pathways. For example, gK is thought to promote entry into neurons by fusion with the plasma membrane because deletion of the N terminus of gK switches entry to endocytosis (82). gC promotes efficient endocytic entry into CHO-HVEM and primary human keratinocytes (69). Other HSV-1 envelope proteins likely play yet unappreciated roles in specifying tropism and influencing the selection of entry pathways, potentially, in a cell-specific manner. Future studies of their contributions will not only explain why HSV-1 encodes such a large number of envelope proteins but may aid in the design of better strategies for blocking entry, leading to improved treatment options for HSV-1 diseases.
METHODS
Cells
HEK293T (gift from J.M. Coffin, Tufts University), Vero (ATCC® CCL-81™), HeLa (ATCC® CCL-2™), and HaCaT cells (gift from J. Garlick, Tufts University) were grown in Dulbecco’s modified Eagle medium (DMEM; Lonza) containing high glucose, and sodium pyruvate, supplemented with L-glutamine (Caisson Labs), 10% heat inactivated fetal bovine serum (HI-FBS; Life Technologies) and 1X penicillin/streptomycin (pen/strep) solution (Corning). B78H1 cells (a gift from R.J. Eisenberg and G.H. Cohen, University of Pennsylvania) were grown in DMEM containing high glucose, sodium pyruvate, and L-glutamine supplemented with 5% FBS and pen/strep solution (1X). C10 cells (a gift from R.J. Eisenberg and G.H. Cohen, University of Pennsylvania), a clonal B78H1-derivative stably expressing human nectin-1, were grown in DMEM containing high glucose, sodium pyruvate, and L-glutamine supplemented with 5% FBS and pen/strep solution (1X) and maintained under selection for nectin-1 expression with 250 μg/ml of G418 (Selleck Chemical) as done previously (27). CHO-K1 cells were grown in Ham’s F12 medium containing 10% FBS and pen/strep solution (1X). CHO-HVEM cells, a derivative of CHO-K1 cells that stably express human nectin-1, were grown in Ham’s F12 medium containing 10% FBS and penicillin-streptomycin solution (1X), 250 ug/ml G418 and 150 ug/ml of puromycin (AG Scientific). CHO-K1 and CHO-HVEM cells were a gift from Anthony Nicola (Washington State University). SH-SY5Y cells were maintained in EMEM (Sigma catalogue number) supplemented with 15% HI-FBS and 1X-pen/strep. SH-SY5Y cells were a kind gift from Stephen Moss (Tufts University).
Plasmids
Plasmids pPEP98, pPEP99, pPEP100, and pPEP101 carry the full-length HSV-1 (strain KOS) genes for gB, gD, gH, and gL, respectively in a pCAGGS vector background. These were kindly gifted by P.G. Spear (Northwestern University). pCMV-VSV-G, which contains the full-length gene for the VSV glycoprotein, G, was a gift from J. White (University of Virginia). Rab GTPase dominant negative constructs [mCherry-Rab5DN(S34N) and dsRed-Rab7DN] were purchased from Addgene (61, 62). For consistency, the dsRed in dsRed-Rab7DN was replaced with mCherry by amplifying mCherry with the following primers: 5’-AGCGCTACCGGTCGCCACCATGGTGAGCAAGGGCGAG-3’ (forward) and 5’-AATTCGAAGCTTGAGCTCGAGATCTGAGCTTGTACAGCTCGTCCATGCC-3’ (reverse). mCherry was then cloned in frame with Rab7DN using AgeI and HindIII cut sites that were engineered into the forward and reverse primers, respectively. As our HSV-1 reporter strain uses tdTomato, eGFP-RabDN constructs were engineered. The same primers were used to amplify eGFP from pEGFP-N2. The same cloning procedure was used to replace mCherry and dsRed with eGFP in the Rab5DN and Rab7DN constructs, respectively. Isolated clones were sequenced to verify mCherry and eGFP were in frame with the DN Rab genes.
Antibodies
Nectin-1 antibody [clone CK41 (83)] conjugated to phycoerythrin (PE) was purchased from BD Biosciences. PE-isotype antibody was also purchased from BD Biosciences. HVEM antibody (R140) was a gift from Gary Cohen (University of Pennsylvania). Caveolin-1 antibody (clone 4H312) was purchased from Santa Cruz Biotechnology. β-actin antibody conjugated to horse radish peroxidase (HRP) was purchased from Santa Cruz Biotechnology (sc-47778 HRP).
Chemical inhibitors
Monensin, methyl-β-cyclodextrin, cytochalasin D, Pitstop-2, and EIPA were purchased from Sigma. Dynasore and MiTMAB were purchased from Calbiochem. Bafilomycin A1 was purchased from ApexBio. Ammonium chloride was purchased from Fisher Scientific. Dyngo-4a was purchased from Abcam. NSC23766 was purchased from Santa Cruz Biotechnology.
Viruses
Pseudotyped viral particles (VSVΔG-BHLD) were generated as described previously (27). Briefly, HEK293T cells (5.5 × 106 cells/10 cm dish) were transfected with 2.5 μg each pPEP98, pPEP99, pPEP100, and pPEP101 using polyethyleneimine (PEI at 1 mg/ml) at a 3:1 weight ratio of PEI to DNA. Twenty-four hours post transfection, cells were infected at an MOI = 3 with VSVΔG-G (VSVΔG pseudotyped with VSV G protein). Forty-eight hours post infection, supernatants were collected, cleared of cell debris (two spins at 1500xg for 10 minutes each), and stored at −80 °C. VSVΔG-BHLD titers were determined on C10 cells.
HSV-1 (GS3217, F strain) was kindly provided by G. Smith (Northwestern University). HSV-1 was propagated on Vero cells and titers were determined by plaque assay on Vero cells as previously described (HSV textbook). VSVΔG-G helper virus was generated by M. Whitt (U. Tennessee) and kindly provided by J. White (U. of Virginia). New stocks were generated similarly to the VSVΔG-BHLD pseudotypes, replacing the HSV-1 glycoproteins with pCMV-VSV-G (10 μg per 10 cm dish). VSVΔG-G titers were determined on C10 cells. VSVΔG-PIV5 was generated and kindly provided by S.P.J. Whelan (HMS). As VSVΔG-PIV5 contains the PIV5 HN and F proteins in the VSV genome, no complementation in trans was necessary. VSVΔG-PIV5 was grown on HEK293T cells and titers were determined on C10 cells.
Entry experiments
3 × 105 B78H1, C10, CHO-K1, CHO-HVEM, HeLa, HaCaT, Vero, or SH-SY5Y cells were seeded in 35 mm dishes. Cells were infected with viruses at a MOI=1. Viruses were incubated with cells at 37 °C for one hour. After one hour, viruses that had not entered were inactivated with a low pH wash (40 mM Na-citrate, 10 mM KCl, 135 mM NaCl, pH 3.0). Complete growth media was added back to cells and infections were allowed to progress for six hours prior to analysis by flow cytometry. Entry experiments in the presence of inhibitors were performed similarly except that prior to infection, C10 and CHO-HVEM cells were pretreated with the indicated inhibitors for one hour prior to infection. All inhibitors, with the exception of sucrose and methyl-β-cyclodextrin, were present during the infection and the six hours post infection prior to analysis by flow cytometry by measuring tdTomato expression (HSV-1 GS3217) or EGFP expression (VSV pseudotypes) to allow sufficient time for viral entry and expression of the fluorescent reporters.
Viral entry experiments in the presence of fluorescently labeled Rab GTPase dominant negative constructs were performed and analyzed as follows: cells infected with HSV-1 (tdTomato) were transfected with pEGFP-N2 as an empty vector control or eGFP-tagged RabDN constructs whereas cells infected with VSV pseudotypes (eGFP) were transfected with pmCherry-C1 empty vector control or mCherry-tagged RabDN constructs. The data represented are the percentage of infected cells that were successfully transfected with either the empty vector control or the RabDN constructs.
Prior to flow cytometry analysis, cells were trypsinized, resuspended in media and pelleted at 450xg for five minutes. Cells were washed with 1X PBS containing 1 mM EDTA (to prevent clumping). Cells were pelleted again at 450xg for five minutes. Cells were then resuspended in 1X PBS with 1 mM EDTA and transferred to FACS tubes. Flow cytometry was performed on a BD LSR II or FACSCalibur instrument. tdTomato expression (HSV-1 GS3217) or EGFP expression (VSV pseudotypes) were measured as a proxy for viral entry. Data analysis was done using FlowJo software (v. 8.8.7).
Analyses of nectin-1 expression and viral infectivity by flow cytometry
Nectin-1 was detected on the surface of cells by staining them with nectin-1 antibody conjugated to PE. Briefly, 106 cells were plated into 10 cm dishes. The next day, cells were lifted from the dishes with 1X-PBS containing 5 mM EDTA. Cells were then pelleted, resuspended in 300 μl FACS buffer (1X-PBS, 2% FBS, 1 mM EDTA), and divided evenly between microfuge tubes for mock, PE-isotype, or PE-anti-nectin-1 treatment. Cells were incubated with 1 μg of antibody for 30 minutes on ice with agitation every 10 minutes. After 30 minutes, cells were pelleted and washed three times with FACS buffer, re-suspended, then immediately analyzed by flow cytometry (FACSCalibur).
Virus purification and densitometry analysis
HSV-1 and VSVΔG-BHLD virions were purified and subjected to immunoblot for gB, gH, gL, and gD. Briefly, five T-175 flasks of Vero cells were infected with HSV-1 (MOI 0.01). HSV-1 was crudely purified as previously described (84). HSV-1 particles were then purified over a continuous 15-50% sucrose gradient (85). The purified band of HSV-1 was collected by puncture and aspiration. VSVΔG-BHLD virions were generated as previously mentioned (see Viruses section of Materials and Methods). VSVΔG-BHLD particles were then pelleted at 20,000 RPM. VSVΔG-BHLD virions were then resuspended and purified over a continuous 15-35% Optiprep gradient [protocol adapted from (86, 87)] and collected by puncture and aspiration. HSV-1 and VSVΔG-BHLD virions were pelleted at 20,0000 RPM. Western blots for gB, gH, gL, and gD were done using the rabbit polyclonal R68 antibody (gB), the rabbit polyclonal R137 antibody (gH), the mouse monoclonal antibody L1 (gL), and the mouse monoclonal antibody R7 (gD). Secondary antibodies from LI-COR were used in order to perform densitometry analysis using the Image Studio Lite software (IRDye® 680RD goat anti-rabbit and IRDye® 800CW goat anti-mouse). Raw densitometry values for gH, gL, and gD blots were normalized to their respective raw densitometry values for gB and reported as fold-differences to gB.
Confocal Microscopy
105 cells (C10 and CHO-HVEM) were seeded onto 12 mm glass coverslips (Chemglass) in 24 well plates. Prior to labeling cells with specific markers of different endocytosis pathways, cells were pretreated with inhibitors for one hour at 37 °C. Post pre-treatment, cells were chilled to 4 °C for 10-15 minutes and were subsequently incubated with specific endocytic markers: Transferrin-Alexa Fluor 488 (50 μg/ml, Thermo Fisher Scientific), 70-kD dextran-rhodamine B (1 mg/ml, Thermo Fisher Scientific), or Lysotracker (1 μM, Thermo Fisher Scientific). Cells were incubated with endocytic markers for 10 minutes at 4 C. After the 10-minute incubation, C10 cells were shifted to 37 °C for 10 minutes and CHO-HVEM cells were shifted to 37 °C for 30 minutes (Transferrin-Alexa Fluor 488 and Lysotracker) (88) or 40 minutes (70 kDa dextran) (89). After incubation at 37 °C, cells were washed 3 times with 1X-PBS and fixed in 4% paraformaldehyde for 15 minutes at room temperature. Cells were washed three times with 1X-PBS and incubated with 2.5 μg/ml of DAPI (ThermoFisher Scientific) diluted in 1X-PBS for 15 minutes at room temperature. Cells were washed again three times with 1X-PBS and mounted onto Prolong Gold Antifade (Life Technologies) on glass slides (Thermo Fisher Scientific). Coverslips were sealed with clear nail polish and analyzed by confocal microscopy using a Leica SPE microscope. Images were analyzed in Fiji (90). For 70-kDa dextran and VSVΔG-BHLD co-localization experiments, C10 and CHO-HVEM cells were incubated with 1 mg/ml 70 kDa dextran and VSVΔG-BHLD (MOI = 1) for 1 hour at 4 C to allow for virion attachment. Cells were then shifted to 37 C for 20 minutes. After 20 minutes, cells were prepared for confocal microscopy by fixing with 4% paraformaldehyde, permeabilized and blocked with 1X PBS containing 5% normal goat serum and 0.3% Triton X-100. Cells were then incubated with anti-gB antibody (R68) overnight at 4 C. The next day, cells were incubated with a secondary antibody labeled with FITC for 1 hour at room temperature. Slides were then prepared as described above.
siRNA-mediated knockdown
Mouse caveolin-1 siRNA, and a control siRNA were purchased from Santa Cruz Biotechnologies. 50 pmol of siRNA (0.625 μg) (cav-1 or scramble [scr]) were diluted into 100 μl of Optimem. In another tube, 3.125 μl of PEI (1 mg/ml) was diluted into 100 μl of Optimem. The diluted PEI was mixed with the diluted siRNA to a final ratio of 5:1 (w:w) of PEI to siRNA and incubated at room temperature for 30 minutes. The complex was added dropwise to CHO-HVEM cells plated in 35 mm dishes (3 × 105 cells/dish). Cells were incubated at 37 °C for 48 hours before infection and subsequent flow cytometry analysis.
SUPPORTING INFORMATION
S1 Fig. Infecting cells at with VSVΔG-BHLD at a higher MOI does not increase entry to an appreciable extent. Receptor null (B78H1 and CHO-K1) and receptor bearing cells (C10, CHO-HVEM, HeLa, Vero, HaCaT, and SH-SY5Y) were infected at MOI =1 (red) or MOI = 10 (purple). Entry efficiency was assessed by flow cytometry at 6 hours post infection.
S2 Fig. Increased amounts of surface nectin-1 or HVEM do not increase VSVΔG-BHLD entry. CHO-K1 (receptor-null) cells were transfected with increasing amounts of plasmids encoding nectin-1 (pBG38) (A) or HVEM (pSC386) (B). Surface expression was analyzed by flow cytometry 24 hours post transfection. C) Cells transfected with nectin-1 or HVEM were infected with VSVΔG-BHLD at MOI = 1. Entry efficiency was assessed at 6 hours post infection by flow cytometry. In each panel, receptor-bearing stable cell line data (CHO-nectin-1 and CHO-HVEM) were inserted as points of comparison.
S3 Fig. VSVΔG-G and VSVΔG-PIV5 differ in their dependence on dynamin for entry. C10 (A and B) and CHO-HVEM (C and D) cells were pretreated with dynamin inhibitors Dynasore (80 μM), Dyngo-4a (25 μM), or MiTMAB (5 μM) and infected with VSVΔG-G or VSVΔG-PIV5 at a MOI of 1. Infectivity was quantitated by flow cytometry at 6 hours post infection. CHO-HVEM cells treated with Dyngo-4a or MiTMAB used the same DMSO control as indicated by the same bar graph appearing twice each in panels C and D. Significance was calculated using a two-tailed Student’s T-test with Welch’s correction (p < 0.05 = *; p < 0.01 = **; p < 0.001 = ***).
S4 Fig. VSVΔG-G entry does not require cholesterol whereas VSVΔG-PIV5 entry requires cholesterol in a cell-type-dependent manner. C10 (A and B) and CHO-HVEM (C and D) cells were pretreated with a cholesterol-removal drug methyl-β-cyclodextran, MβCD (5 mM) and infected with VSVΔG-G or VSVΔG-PIV5 at a MOI of 1. Infectivity was quantitated by flow cytometry at 6 hours post infection. (E) CHO-HVEM cells were transfected with a caveolin-1 siRNA (cav-1) or a scrambled control siRNA (scr) (both 50 pm) and infected with VSVΔG-G or VSVΔG-PIV5 at a MOI of 1. Infectivity was quantitated by flow cytometry at 6 hours post infection. Significance was calculated using a two-tailed Student’s T-test with Welch’s correction (p < 0.05 = *; p < 0.01 = **; p < 0.001 = ***). F) C10 and CHO-HVEM cells were treated with either a solvent control (H2O/EtOH) or methyl-β-cyclodextrin (MβCD), then incubated with cholera toxin subunit B labelled with Alexa Fluor 488. Confocal microscopy was performed on the solvent control and methyl-β-cyclodextrin treated cells. Cells were fixed, counterstained with DAPI, and imaged by confocal microscopy. Scale bar = 25 μm.
S5 Fig. VSVΔG-BHLD does not co-localize with the fluid-phase marker 70 kDa dextran in C10 cells. C10 cells were incubated with 1 mg/ml of rhodamine-B labelled 70 kDa dextran and VSVΔG-BHLD (MOI = 1) for one hour at 4°C. Cells were then shifted to 37 °C for 20 minutes. Cells were fixed, counterstained with DAPI, and imaged by confocal microscopy. gB was detected by immunofluorescence using the rabbit pAb R68 and anti-rabbit IgG conjugated to FITC. Green = gB (marker for VSVΔG-BHLD particles); Red = 70 kDa dextran. Scale bar = 25 μm.
S6 Fig. VSVΔG-BHLD, in large, does not co-localized with the fluid-phase marker, 70 kDa dextran, in CHO-HVEM cells. CHO-HVEM cells were incubated with 1 mg/ml of rhodamine-B labelled 70 kDa dextran and VSVΔG-BHLD (MOI = 1) for one hour at 4°C. Cells were then shifted to 37°C for 20 minutes. Cells were fixed, counterstained with DAPI, and imaged by confocal microscopy. gB was detected by immunofluorescence using the rabbit pAb R68 and anti-rabbit IgG conjugated to FITC. Green = gB (marker for VSVΔG-BHLD particles); Red = 70 kDa dextran. Scale bar = 25 μm.
S7 Fig. VSVΔG-G and VSVΔG-PIV5 entry does not require macropinocytosis. C10 (A and B) and CHO-HVEM (C and D) cells were pretreated with macropinocytosis inhibitors cytochalasin D (2 μM), EIPA (25 μM), or NSC23766 (200 μM) and infected with VSVΔG-G or VSVΔG-PIV5 at a MOI of 1. Infectivity was quantitated by flow cytometry at 6 hours post infection. Significance was calculated using a two-tailed Student’s T-test with Welch’s correction (p < 0.05 = *; p < 0.01 = **; p < 0.001 = ***).
S8 Fig. VSVΔG-G but not VSVΔG-PIV5 entry requires endosomal acidification. C10 (A and B) and CHO-HVEM (C and D) cells were pretreated with inhibitors of endosomal acidification BFLA (100 nM), NH4Cl (50 mM), or monensin (15 μM) and infected with VSVΔG-G or VSVΔG-PIV5 at MOI = 1. Infectivity was quantitated by flow cytometry at 6 hours post infection. Significance was calculated using a two-tailed Student’s T-test with Welch’s correction (p < 0.05 = *; p < 0.01 = **; p < 0.001 = ***).
ACKNOWLEDGEMENTS
We thank Stephen Kwok and Allen Parmelee at the Tufts Laser Cytometry Core facility for their help with FACS experiments; Roselyn Eisenberg and Gary Cohen (University of Pennsylvania) for the gifts of antibodies and cell lines; Sean Whelan (Washington University) for the gift of VSVΔG-PIV5 pseudotype; Anthony Nicola (Washington State University) for the gift of CHO-HVEM cells; Richard Longnecker (Northwestern University) for the gift of CHO-nectin-1 cells; Michael Forgac (Tufts University) for the gift of Lysotracker reagent; and Michael Whitt (University of Tennessee) for the gift of the VSVΔG-GFP pseudotyping platform. FACS experiments were performed at the Tufts Laser Cytometry Core facility.
REFERENCES
- 1.↵
- 2.↵
- 3.↵
- 4.↵
- 5.↵
- 6.↵
- 7.↵
- 8.↵
- 9.↵
- 10.↵
- 11.
- 12.↵
- 13.↵
- 14.↵
- 15.↵
- 16.
- 17.↵
- 18.↵
- 19.↵
- 20.↵
- 21.↵
- 22.↵
- 23.↵
- 24.↵
- 25.↵
- 26.↵
- 27.↵
- 28.↵
- 29.↵
- 30.↵
- 31.↵
- 32.↵
- 33.↵
- 34.↵
- 35.↵
- 36.↵
- 37.↵
- 38.
- 39.↵
- 40.↵
- 41.↵
- 42.↵
- 43.↵
- 44.↵
- 45.↵
- 46.↵
- 47.↵
- 48.↵
- 49.↵
- 50.↵
- 51.↵
- 52.↵
- 53.↵
- 54.↵
- 55.↵
- 56.↵
- 57.↵
- 58.↵
- 59.↵
- 60.↵
- 61.↵
- 62.↵
- 63.↵
- 64.↵
- 65.↵
- 66.↵
- 67.↵
- 68.↵
- 69.↵
- 70.↵
- 71.↵
- 72.↵
- 73.↵
- 74.↵
- 75.↵
- 76.↵
- 77.↵
- 78.↵
- 79.↵
- 80.↵
- 81.↵
- 82.↵
- 83.↵
- 84.↵
- 85.↵
- 86.↵
- 87.↵
- 88.↵
- 89.↵
- 90.↵