Herpes Simplex Virus 1 Can Enter Dynamin 1 and 2 Double-Knockout Fibroblasts

The human pathogen herpes simplex virus 1 (HSV-1) can adapt to a variety of cellular pathways to enter cells. In general, HSV-1 is internalized by fusion of its envelope with the plasma membrane or by endocytic pathways, which reflects the high adaptation to differences in its target cells. The challenges are to distinguish whether multiple or only one of these internalization pathways leads to successful entry and, furthermore, to identify the mode of viral uptake. In this study, we focused on dynamin, which promotes endocytic vesicle fission, and explored how the presence and absence of dynamin can influence viral entry. Our results support the idea that HSV-1 entry into mouse embryonic fibroblasts depends on dynamin; however, depletion of dynamin still allows efficient viral entry, suggesting that alternative pathways present upon dynamin depletion can accomplish viral internalization.

ABSTRACT

Dynamin GTPases, best known for their role in membrane fission of endocytic vesicles, provide a target for viruses to be exploited during endocytic uptake. Recently, we found that entry of herpes simplex virus 1 (HSV-1) into skin cells depends on dynamin, although our results supported that viral internalization occurs via both direct fusion with the plasma membrane and via endocytic pathways. To further explore the role of dynamin for efficient HSV-1 entry, we utilized conditional dynamin 1 and dynamin 2 double-knockout (DKO) fibroblasts as an experimental tool. Strikingly, HSV-1 entered control and DKO fibroblasts with comparable efficiencies. For comparison, we infected DKO cells with Semliki Forest virus, which is known to adopt clathrin-mediated endocytosis as its internalization pathway, and observed efficient virus entry. These results support the notion that the DKO cells provide alternative pathways for viral uptake. Treatment of cells with the dynamin inhibitor dynasore confirmed that HSV-1 entry depended on dynamin in the control fibroblasts. As expected, dynasore did not interfere with viral entry into DKO cells. Electron microscopy of HSV-1-infected cells suggests viral entry after fusion with the plasma membrane and by endocytosis in both dynamin-expressing and dynamin-deficient cells. Infection at low temperatures where endocytosis is blocked still resulted in HSV-1 entry, although at a reduced level, which suggests that nonendocytic pathways contribute to successful entry. Overall, our results strengthen the impact of dynamin for HSV-1 entry, as only cells that adapt to the lack of dynamin allow dynamin-independent entry.

IMPORTANCE The human pathogen herpes simplex virus 1 (HSV-1) can adapt to a variety of cellular pathways to enter cells. In general, HSV-1 is internalized by fusion of its envelope with the plasma membrane or by endocytic pathways, which reflects the high adaptation to differences in its target cells. The challenges are to distinguish whether multiple or only one of these internalization pathways leads to successful entry and, furthermore, to identify the mode of viral uptake. In this study, we focused on dynamin, which promotes endocytic vesicle fission, and explored how the presence and absence of dynamin can influence viral entry. Our results support the idea that HSV-1 entry into mouse embryonic fibroblasts depends on dynamin; however, depletion of dynamin still allows efficient viral entry, suggesting that alternative pathways present upon dynamin depletion can accomplish viral internalization.

INTRODUCTION

Herpes simplex virus 1 (HSV-1) infects its human host via skin and mucocutaneous linings and establishes a lifelong latent infection after spreading to the peripheral nervous system. Viral penetration into target cells is preceded by the complex interaction of virion envelope glycoproteins with cellular receptors (1). The first step includes attachment to heparan sulfate side chains of cell surface proteoglycans (2, 3) followed by binding of the viral glycoprotein D (gD) to a receptor which finally leads to fusion with a cellular membrane (4, 5). The major gD receptors are the intercellular adhesion protein nectin-1 and herpesvirus entry mediator (HVEM), a member of the tumor necrosis factor receptor superfamily (6, 7). The impact of each receptor during infection of the human host has yet to be defined. When we addressed their role in target tissues such as epidermis and dermal fibroblasts, we found that at least in murine skin, nectin-1 acts as the primary receptor and HVEM can functionally replace it, although less efficiently in keratinocytes than in fibroblasts (8, 9).

In general, viruses utilize endocytosis as the most common entry pathway. After binding to the cellular surface, the various endocytic pathways allow viruses to ferry their capsids through the cytoplasm, where cortical actin filaments and other cytoplasmic barriers are overcome to reach the nucleus. The envelope of HSV-1 can fuse either with the plasma membrane or with vesicle membranes after virions are internalized via endocytosis (10). Depending on the cell line and on the expression of gD receptors, the mode of HSV-1 entry can vary (11–14), suggesting that HSV-1 is capable of using a variety of entry pathways that may reflect an adaptation to differences in its target cells. Recently, we focused on the internalization pathway into human keratinocytes, as these represent the target cells for HSV-1 in vivo. Strikingly, we revealed dual internalization modes comprising direct fusion at the plasma membrane and uptake by endocytic vesicles, both in primary human keratinocytes and in the human keratinocyte cell line HaCaT (15). The contribution of each pathway to successful infection, however, remains to be shown. Our previous studies with pharmacological inhibitors support the idea that endocytic uptake contributes to HSV-1 entry into human keratinocytes (15). To date, there is no evidence for clathrin- or caveolin-mediated endocytosis, but several host factors involved in various endocytic pathways have been suggested to accomplish internalization of HSV (15–18).

Dynamin controls several distinct endocytic pathways, of which clathrin-mediated endocytosis is the best studied (19). The large GTPase dynamin assembles into helical polymers at the necks of budding vesicles and catalyzes membrane fission and fusion (20–22). The role of dynamin during HSV-1 internalization is still controversial and seems to depend on the cell type and entry pathway (16–18, 23). We found a dynamin-dependent uptake of HSV-1 in keratinocytes and dermal fibroblasts but not in primary hippocampal neurons (9, 15). The dynamin inhibitor dynasore blocks infection of keratinocytes by interfering with the penetration of virions at the plasma membrane, which, in turn, inhibits uptake both via fusion of the viral envelope with the plasma membrane and via vesicle formation (15).

To explore the underlying mechanisms of dynamin involvement in HSV-1 entry and to address the impact of dynamin on the dual uptake mechanisms that include envelope fusion with the plasma membrane and internalization via vesicles, we used mouse embryonic fibroblasts from conditional dynamin knockout mice for infection studies (24). Mammalian genomes contain three dynamin genes. Dynamin 1 (Dnm1) is highly expressed in neurons and dynamin 2 (Dnm2) is ubiquitously expressed, whereas dynamin 3 (Dnm3) is mainly present in brain and testis (25). Dynamin 2 knockout (KO) mice die during early embryogenesis, emphasizing the major impact of this dynamin isoform at the cell and organism levels (24). As conditional dynamin 2 KO cells express dynamin 1 (26) but not detectable levels of dynamin 3 (24), we explored HSV-1 entry into conditional dynamin 1 and 2 double KO (DKO) fibroblasts. These cells exhibit a proliferation defect and multiple signaling defects (27, 28) and are impaired in clathrin-mediated endocytosis, as clathrin-coated pits still form but do not mature to free vesicles (24). Using DKO fibroblasts as an experimental tool, we obtained evidence that HSV-1 can enter fibroblasts that lack dynamin 1 and 2, although HSV-1 entry into control fibroblast cells still expressing dynamin is dynamin dependent. Strikingly, our electron microscopy (EM) studies suggest viral uptake via fusion at the plasma membrane and by vesicle formation in the dynamin-expressing control as well as in the dynamin DKO cells.

RESULTS

Entry of HSV-1 and Semliki Forest virus into cells lacking dynamin.

Based on the overexpression of a dominant negative dynamin mutant and the application of the dynamin inhibitor dynasore, we conclude that dynamin plays an essential role during HSV-1 entry into keratinocytes and dermal fibroblasts (9, 15). In this study, we utilized mouse embryonic fibroblasts that conditionally lack dynamin 2 in addition to dynamin 1 (24) to explore the impact of dynamin. The DKO cells were obtained after exposure to 4-hydroxytamoxifen (4-OHT) for 7 days, inducing the inserted Cre recombinase (24). After this time, proliferation was reduced in the DKO compared to the control cells, whereas the overall morphologies were comparable (data not shown). Immunofluorescence staining of F-actin and α-tubulin demonstrated no obvious change in the actin cytoskeleton and the microtubule network, respectively (Fig. 1a). When we visualized endophilin-A2, a dynamin-recruiting protein, we confirmed the more clustered distribution in the DKO cells relative to the control cells (Fig. 1b), which is in line with arrested clathrin-coated pits in the absence of dynamin (24). Furthermore, we confirmed the depletion of the dynamin isoforms upon 4-OHT treatment. As shown by quantitative reverse transcription-PCR (qRT-PCR), dynamin 1 and 2 transcripts were strongly reduced in DKO compared to control cells (Fig. 1c). As a control, we confirmed the downregulation of dynamin 1 and 2 expression in each experimental setting by RT-PCR (see Fig. 2c, 4b, and 5d).

FIG 1.

Characterization of dynamin DKO fibroblasts. (a) Immunostainings with DAPI (blue) as nuclear counterstain show no differences in the actin cytoskeleton (red) and the tubulin network (green) between control and DKO fibroblasts. (b) Immunostainings of endophilin-A2 (gray) demonstrate its accumulation in puncta only in DKO cells. Confocal projections and merged images show representative cells. Bars, 50 μm. (c) Quantitative PCR confirmed minor levels of dynamin 1 and 2 transcripts in DKO cells after 4-OHT induction for 7 days. The relative numbers of transcripts were determined in at least three independent experiments and are shown as means plus SDs. P values of ≤0.05 (*) are shown.

FIG 2.

HSV-1 infection of dynamin DKO fibroblasts. (a) Control and DKO fibroblasts were infected with HSV-1 at a multiplicity of infection (MOI) of 2 or 20 PFU/cell for 3 h, and the number of ICP0-expressing cells was determined in at least three independent experiments. The results are shown as means + SDs. (b) Upon infection at 20 PFU/cell for 3 h, immunostainings showed comparable numbers of ICP0-expressing cells (green) in control and DKO cultures with DAPI (blue) as a nuclear counterstain. Examples of nuclear ICP0 (arrowhead) and cytoplasmic ICP0 (arrow) are indicated. Overlays of immunofluorescence analyses are shown. Bar, 50 μm. (c) As a control for the cells shown in panel b, the depletion level of dynamin 1 and 2 was demonstrated by RT-PCR.

FIG 4.

Dynasore inhibits HSV-1 infection efficiency only in control fibroblasts. (a) Control and DKO fibroblasts were pretreated with 4% DMSO or 40 μM dynasore or left untreated for 30 min at 37°C, followed by infection with HSV-1 at 20 PFU/cell for 3 h in the presence of dynasore. As a control, dynasore was washed out after the 30-min pretreatment and prior to infection. Stainings demonstrate the reduced number of ICP0-expressing cells (green) only in dynasore-treated control cells with DAPI (blue) as a nuclear counterstain. Overlays of immunofluorescence analyses are shown. Bar, 50 μm. (b) As a control, the depletion level of dynamin 1 and 2 was demonstrated by RT-PCR. (c) After pretreatment of control and DKO cells with DMSO or increasing concentrations of dynasore followed by infection for 3 h, the numbers of ICP0-expressing cells were determined in at least three independent experiments and are shown as means + SDs. P values of ≤0.05 (*) are shown.

FIG 5.

Entry into control and dynamin DKO fibroblasts is actin independent. (a) Control and DKO cells were pretreated with 0.1 μM latrunculin A (lat A) for 5 min. Stainings with phalloidin (red) demonstrate the initiation of actin depolymerization. DAPI (blue) is shown as a nuclear counterstain. (b) Pretreated cells were infected with HSV-1 at 20 PFU/cell for 2 h in the presence of latrunculin A, followed by incubation in the absence of latrunculin A for 1 h. Overlays of immunofluorescence analyses show nearly all cells with nuclear ICP0 (green) and the level of depolymerized F-actin at 3 h p.i. Bar, 20 μm. (c) The numbers of ICP0-expressing cells were determined after infection of DMSO- or latrunculin A-treated control or DKO cells at 3 h p.i. in at least three independent experiments and are shown as means + SDs. (d) As a control, the depletion level of dynamin 1 and 2 was demonstrated by RT-PCR.

To address whether the efficiencies of HSV-1 entry differ in the presence and absence of dynamin, we performed infection studies and determined successful entry by visualizing viral infected cell protein 0 (ICP0) in individual cells. ICP0 is expressed once the viral genome is released into the nucleus and localizes in nuclear foci, but it relocalizes to the cytoplasm during later infection (29, 30). Hence, the visualization of ICP0 expression allows one to determine successful virus entry into single cells. Surprisingly, immunostainings showed comparable numbers of ICP0-expressing cells in the control and dynamin DKO cells after infection at 20 PFU/cell (Fig. 2b). To better characterize minor differences in infection efficiency when dynamin was present or absent, we reduced the virus dose to 2 PFU/cell. Again, no difference in the number of infected cells was observed as early as 3 h postinfection (p.i.) (Fig. 2a). In addition, the ratio of cells with cytoplasmic and nuclear localization of ICP0 even indicated no delay in the onset of entry and early infection when dynamins were absent (Fig. 2b).

As HSV-1 entered dynamin DKO cells quite efficiently, we concluded that the virus adapted its mode of internalization to a dynamin-independent uptake pathway. To explore whether only HSV-1 can enter via these alternate entry modes, we performed infection studies with Semliki Forest virus (SFV), a well-studied enveloped RNA virus that was first shown to adopt clathrin-mediated endocytosis as its internalization pathway (31–34). When we infected control and dynamin DKO cells with SFV and visualized single cells undergoing viral replication, we found high numbers of infected cells in both the absence and presence of dynamin (Fig. 3). While nearly all control cells were infected at 9 h p.i., infection efficiency was slightly reduced in the absence of dynamin (Fig. 3b). At 16 h p.i., however, we observed a strong cytopathic effect in both the control and dynamin DKO cells, with nearly all cells infected (Fig. 3a). This suggests that SFV infection is delayed only in cells that lack dynamin.

FIG 3.

SFV infection of dynamin DKO fibroblasts. (a) Control and DKO fibroblasts were infected with SFV at 1 PFU/cell. Stainings indicate cells with replicating viral RNA genomes (dsRNA) (green) at 9 and 16 h p.i. with DAPI (blue) as a nuclear counterstain. Overlays of immunofluorescence analyses are shown. Bar, 50 μm. (b) Upon infection with SFV at 1 PFU/cell for 9 h, the numbers of infected control and DKO cells were determined in two independent experiments as replicates and are shown as means + SDs.

Overall, we observed efficient viral entry in the absence of dynamin not only for HSV-1 but also for SFV, which is known to depend on clathrin, AP2, and dynamin. These findings further support the idea that the dynamin DKO cells can adopt and/or activate alternative pathways that allow the viruses to enter in a dynamin-independent manner.

Dynasore interferes with HSV-1 entry in the presence of dynamin but not in cells lacking dynamin.

As HSV-1 efficiently entered DKO cells, we addressed whether entry into the control fibroblasts still expressing dynamin is indeed mediated by dynamin. To investigate the impact of dynamin, we pretreated the control fibroblasts with dynasore, an inhibitor of dynamin GTPase activity (35). Addition of up to 40 μM dynasore impaired HSV-1 infection in a concentration-dependent manner (Fig. 4a and c). These results are comparable to our recent observations in primary dermal fibroblasts, although the level of inhibition was slightly higher in the primary cells (9). No inhibitory effect was observed after addition of dimethyl sulfoxide (DMSO), the solvent used to dilute dynasore, or when dynasore was washed out prior to infection (Fig. 4a and c). Thus, our findings support a dynamin-dependent entry into the control fibroblasts.

When we compared the effects of dynasore on infection efficiency in the control and dynamin DKO cells, we observed no decrease in the number of infected DKO cells when the DKO cells were treated with up to 40 μM dynasore (Fig. 4a and c). Higher concentrations, such as 80 μM dynasore, were not suitable, as cell attachment to the coverslips was disturbed. These results demonstrate that dynasore impaired HSV-1 infection only in the presence of dynamin. When the DKO cells adopted dynamin-independent uptake strategies, the virus could use these alternative pathways, which may only be present upon dynamin depletion.

HSV-1 enters fibroblasts in an actin-independent manner in both the presence and absence of dynamin.

Dynamin is intimately linked to actin, most obviously at endocytic sites, where both play a role in membrane remodelling during endocytosis (21). Studies with dynamin DKO cells have demonstrated that F-actin acts upstream of dynamin at clathrin-coated pits (24). HSV-1 can engage the actin cytoskeleton at multiple steps during infection. The role of actin dynamics during viral entry most likely depends on the cell type and the corresponding internalization pathway (36). As HSV-1 efficiently entered control and DKO fibroblasts, we investigated whether actin plays opposing roles in the two different uptake pathways. Hence, we pretreated control and dynamin DKO cells with latrunculin A, which blocks actin polymerization (37). Detachment of the DKO fibroblasts was avoided by seeding the cells on collagen-coated coverslips. Pretreated cells were infected for 2 h in the presence of latrunculin A. To facilitate visualization of ICP0-expressing cells, the drug was replaced with medium for 1 h to allow at least some cell recovery (Fig. 5b). Control stainings confirmed that nearly no ICP0-expressing cells were detected at 1 h p.i. (data not shown). Interestingly, both control and dynamin DKO cells were infected after latrunculin A-induced impairment of the cytoskeleton, suggesting that viral entry is independent of actin polymerization irrespective of whether HSV-1 enters via a dynamin-dependent or -independent pathway (Fig. 5a to c). We cannot exclude minor effects on the efficiency of viral entry when actin assembly is impaired; however, our results demonstrate no obvious difference between control and DKO cells.

Internalization of HSV-1 relies on two distinct internalization pathways in both control and dynamin DKO cells.

To characterize how HSV-1 is internalized in control and dynamin DKO cells, we performed EM studies. Previous analysis of primary murine fibroblasts suggests that HSV-1 can enter both by direct fusion with the plasma membrane and via endocytic vesicles (9). In this study, we compared internalization into control and DKO fibroblasts to explore whether HSV-1 bypasses endocytic uptake and is internalized by fusion with the plasma membrane when entry is dynamin independent. After infection with 400 PFU/cell for 30 min, cells were processed for EM analyses. We omitted preadsorption of virus on ice, as the DKO cells detached as soon as they were shifted from incubation on ice to 37°C. While approximately 80% of the virus particles were found to be attached to the cell surface at 30 min p.i., we observed the internalized particles either as free cytoplasmic nucleocapsids or as enveloped particles in vesicles (Fig. 6). Interestingly, there was no obvious difference in the distribution of internalized viruses between cells expressing or lacking dynamin as shown by the quantification of virus particles (Fig. 6d). Sometimes, more than one virus particle was found in enlarged vesicles in control as well as in DKO cells; however, the impact of these enlarged vesicles remains to be shown. The results suggest that HSV-1 uptake can occur both via direct fusion with the plasma membrane and via an endocytic pathway. As many of the capsids were located just underneath the plasma membrane, we assume a direct fusion process of the viral envelope. However, we cannot exclude that some of the free capsids might be released from endosomes.

FIG 6.

Uptake of HSV-1 into control and dynamin DKO fibroblasts. Cells were incubated with HSV-1 at 400 PFU/cell at 37°C and prepared for EM at 30 min p.i. A particle on the cell surface (a), an enveloped particle in a vesicle (b), and a free cytoplasmic capsid (c) are shown. Bar, 200 nm. (d) In two independent experiments 174 particles in total for control cells and 205 particles for DKO cells were evaluated at 30 min p.i. The percentages of attached particles, cytoplasmic capsids, and enveloped particles in vesicles are given as means + SDs.

We then investigated whether endocytosis plays a more dominant role for dynamin-dependent than for dynamin-independent viral entry. Since endocytosis, including vesicular trafficking, is known to be blocked at low temperature, we addressed whether low temperature still allows HSV-1 entry into mouse embryonic fibroblasts and, if so, whether there is a difference between control and dynamin DKO cells. Recent studies report that HSV-1 can still infect human keratinocytes at a low temperature, such as 7°C (38). When we analyzed infection efficiency at low temperature in control and DKO cells, we used a modified infection protocol that allowed us to study internalization at various temperatures followed by virus inactivation and incubation at 37°C to enable ICP0 expression as readout for infected cells. Accordingly, cells that were infected for 2 h at 15°C or 7°C were treated with a low-pH buffer to inactivate extracellular and attached virus particles (39). This step was followed by incubation at 37°C for another 2 h to allow particles that had already penetrated at low temperature to pursue successful entry. We confirmed complete virus inactivation through the low-pH treatment by incubating control fibroblasts on ice for 1 h to allow attachment, followed by low-pH treatment and incubation at 37°C for 4 h (Fig. 7a). When the infection efficiencies of control and dynamin DKO cells were compared at various temperatures, we found a slightly reduced number of infected cells in both control and DKO cells at 15°C compared to that at 37°C (Fig. 7b). Furthermore, the localization of nuclear and cytoplasmic ICP0 indicates delayed early infection at 15°C, which seems more prominent in control than in DKO cells (Fig. 7b). ICP0-expressing cells were also found at 7°C in both control and DKO cells, although at a strongly reduced number (Fig. 7b and c). However, the number of infected cells increased at later times, suggesting that viral entry is strongly delayed at 7°C (Fig. 7c). Interestingly, infection efficiency might be slightly higher in DKO cells than in control cells at both 15°C and 7°C. To support that viral internalization takes place at 7°C, we performed EM studies. As the low temperature slows down membrane fusion processes, we observed not only free capsids in the cytoplasm but also several examples of viral envelopes that fuse with the plasma membrane (Fig. 7d). To confirm that endocytosis is impaired at 7°C, we investigated the efficiency of uptake of cholera toxin B, a glycosphingolipid-binding ligand that is internalized by different endocytic pathways such as clathrin-coated pits, caveolae, and a clathrin-, caveolin-, and dynamin-independent pathway (40, 41). When cholera toxin B was added to control cells, we observed an efficient block at both 7°C and on ice as visualized by the retention of cholera toxin B close to the plasma membrane, while the ligand showed perinuclear localization at 37°C (Fig. 7e).

FIG 7.

HSV-1 infects control and dynamin DKO fibroblasts at low temperatures. (a) Nearly all control cells which were infected with HSV-1 at 20 PFU/cell for 1 h on ice followed by 4 h at 37°C expressed cytoplasmic ICP0 (green). Acid treatment after 1 h of incubation on ice and prior to the 4 h of incubation at 37°C resulted in a block of infection. Acid treatment after 2 h at 37°C followed by further incubation for 2 h at 37°C led to ICP0 expression in nearly all cells. DAPI (blue) is shown as a nuclear counterstain. Overlays of immunofluorescence analyses are shown. Bar, 50 μm. (b) The numbers of ICP0-expressing control and DKO cells were determined after infection for 4 h at various temperatures in at least three independent experiments. (c) Control cells were infected with HSV-1 at 20 PFU/cell for 2 h at 7°C, followed by acid treatment and incubation at 37°C for 2 h (4 h p.i.) and 5 h (7 h p.i.), respectively. At 4 h p.i. ICP0 (green) was mainly expressed in the nucleus, while cytoplasmic ICP0 (green) was visible at 7 h p.i. DAPI (blue) is shown as a nuclear counterstain. Bar, 50 μm. (d) Control cells were incubated with HSV-1 at 400 PFU/cell at 7°C and prepared for EM at 2 h p.i. Fusion of the viral envelope with the plasma membrane is shown, and the arrowhead indicates the free cytoplasmic capsid. Bars, 250 nm. (e) Cholera toxin B (CT-B) was added to control cells and incubated at 7°C and 37°C, respectively, to visualize the block of endocytic uptake and transport at low temperature. As a control, cells were incubated on ice after treatment with cholera toxin B. Confocal projections and merged images show representative cells. Bar, 10 μm.

Overall, our infection studies at 7°C indicate that HSV-1 can be internalized into control and DKO cells, implying that a block of endocytic membrane traffic inhibits neither dynamin-dependent nor the alternative pathways in the absence of dynamin but might more efficiently delay entry in the presence of dynamin.

DISCUSSION

In this study, we explored the impact of dynamin on HSV-1 entry by utilizing dynamin DKO cells. As dynamin catalyzes membrane fission and can act as a master regulator of mechanistically diverse endocytic pathways, we speculated that the investigation of its involvement in viral uptake helps to understand how HSV-1 accomplishes penetration into its host cells. To address the effect of the lack of dynamin on viral entry, we performed infection studies with DKO mouse embryonic fibroblasts in which the two isoforms, dynamin 1 and 2, are depleted upon exposure to 4-OHT (24). Surprisingly, we observed comparable efficiencies of HSV-1 entry into control and DKO cells. One might argue that the potential expression of dynamin 3 in DKO cells and the putative overlapping role of the three dynamin isoforms contribute to this effect (42). However, dynamin 1, 2, and 3 triple-KO (TKO) cells display the same endocytic defects as DKO cells and show no additional malfunctions or inhibition of fluid-phase endocytosis (42). Furthermore, DKO as well as TKO cells are still viable, although they cease proliferation (42). These findings support the idea that a very residual activity of dynamin 3 cannot trigger dynamin-dependent cellular processes in DKO cells but suggest that at least the limited cell viability is due to the activation of compensatory pathways.

In addition to HSV-1, we addressed whether DKO cells still allow entry of small viruses such as SFV, known to enter via clathrin-mediated endocytosis, as this pathway is shown to be impaired in the dynamin-deficient cells (24). Strikingly, SFV entered DKO cells efficiently, although infection was delayed compared to that of control cells. Together with the successful entry of HSV-1, the results led us to hypothesize that the absence of dynamin leads to the activation of alternative endocytic pathways which might contribute to cell viability and could serve as viral entry portals. This assumption is in line with our observation that HSV-1 entry relies on dynamin in control fibroblasts but is dynamin independent in DKO cells, as shown by inhibitor studies with dynasore. The opposing effect of dynasore in control and DKO cells argues against the described off-target effects of dynasore, such as inhibition of fluid-phase endocytosis and membrane ruffling (42), but rather points to the impact of dynamin for the uptake of HSV-1. When fibroblasts lack dynamin, HSV-1 can enter via a dynamin-independent pathway that is most likely adopted by the cells to survive.

The alternative pathways in DKO cells might be activated upon the induced lack of dynamin, or they might be already present in control cells but less prominent and not preferred for viral entry. Our ultrastructural analyses of HSV-1-infected cells revealed free capsids in the cytoplasm and enveloped particles in vesicles, strikingly, in both the presence and absence of dynamin. These results support viral entry via fusion with the plasma membrane as well as endocytic uptake. Although the presence of free capsids in the cytoplasm could also result from the release after fusion with the vesicle membrane, the localization close to the plasma membrane is highly suggestive of direct fusion. Both direct fusion with the plasma membrane and endocytosis were already observed in primary murine dermal fibroblasts, where dynamin is required for viral entry (9). However, it is still open how efficiently each internalization mode contributes to successful entry. We cannot exclude that the virus-containing vesicles observed in DKO cells are not processed and thus fail to support productive infection. It is generally assumed that fusion with the plasma membrane is dynamin independent, while uptake via vesicles depends on dynamin. Thus, one might speculate that the contribution of each uptake pathway to successful entry might differ in the absence and presence of dynamin. However, we observed no difference in the time course of ICP0 expression as readout for successful entry, implying that the efficiency of preferred endocytic uptake in control cells would be comparable to the efficiency of fusion with the plasma membrane in dynamin DKO cells. To explore the efficacy of viral entry in the absence of endocytosis, we performed infection studies at low temperature, such as 15°C and below, which is known to arrest endocytosis and vesicular transport (43). At 15°C we still observed efficient entry of HSV-1 at 4 h p.i., although entry was delayed, as indicated by the localization of ICP0 in the nucleus and about 20% fewer infected cells than at 37°C (Fig. 7b and c). As viral internalization took place even at 7°C (Fig. 7d), we conclude that HSV-1 can in principle be internalized in the absence of endocytic pathway, although at a reduced level. Viral entry at low temperatures occurred in both the presence and absence of dynamin, supporting the idea that the nonendocytic pathway could accomplish successful entry in control as well as in DKO cells. Strikingly, we observed minor differences in the numbers of infected control and DKO cells, suggesting that the nonendocytic pathway in control cells may be less efficient than in DKO cells. Future studies are needed to define the nonendocytic pathway(s) and how it is achieved relative to the presence or absence of dynamin. In addition to fusion with the plasma membrane, recent studies suggest that the uptake of Sindbis virus occurs via induction of a protein pore in the plasma membrane (44). This conclusion is based on the observation that Sindbis virus infection can take place at 5°C and 15°C; however, virus-mediated cell-cell fusion was arrested, implying that membrane fusion could be prevented at low temperatures (44). Membrane fusion at low temperatures most likely depends on the fluidity of membrane areas and might be still possible at lipid-rich domains. Taken together, our infection studies at low temperatures support nonendocytic internalization pathways of HSV-1 into murine fibroblasts. However, we are only at the beginning of understanding the precise mechanisms underlying the diverse modes of HSV-1 uptake. The challenge of future experiments is to distinguish under which conditions HSV-1 accomplishes cellular internalization that leads to productive infection.

We conclude that although dynamin appears to play a major role during HSV-1 entry, the virus is able to circumvent the absence of this essential cellular component by adapting to a pathway that dynamin-deficient fibroblasts still offer.

MATERIALS AND METHODS

Cells and viruses.

The dynamin conditional DKO fibroblasts were derived from Dnm1^LoxP/LoxP/Dnm2^LoxP/LoxP mice and express a tamoxifen-inducible Cre-estrogen receptor mutant knock-in transgene from the ROSA26 locus (24). DKO cells were maintained in Dulbecco's modified Eagle's medium (DMEM)-high glucose-GlutaMAX (Life Technologies) containing 10% fetal calf serum (FCS), penicillin (100 IU/ml), and streptomycin (100 μg/ml). Ablation of dynamins 1 and 2 was induced after incubation with 5 μM 4-OHT (Sigma) for 2 days followed by incubation with 0.5 μM 4-OHT for 5 days. Noninduced DKO cells served as control cells.

Infection studies were performed with purified preparations of HSV-1 wild-type strain 17 as described previously (45). For the infection studies at various temperatures, cells were infected at 20 PFU/cell and incubated for 2 h at 37°C, 15°C, or 7°C. To inactivate bound virus, the cells were treated with ice-cold citrate buffer (0.1 M citric acid, 0.1 M sodium citrate [pH 3]) for 3 min and then washed with phosphate-buffered saline (PBS) and incubated in prewarmed medium at 37°C for the desired times.

Infection studies were also performed with Semliki Forest virus strain SFV4 (46), which was grown on HeLa cells. Cells were infected with SFV4 at 1 PFU/cell. Virus inoculum was added to cells at 37°C, defining time point zero, and replaced with medium at 1 h p.i.

RNA preparation and qRT-PCR.

RNA was isolated from cells by use of Nucleozol reagents (Macherey-Nagel). cDNA was synthesized using SuperScript II reverse transcriptase (Life Technologies). qPCRs were performed using the SYBR GreenER qPCR SuperMix Universal (Life Technologies) on the DNAengine-Opticon 2 System (Bio-Rad). The dynamin-specific primer pairs were for dynamin 1 (forward, 5′-TCTGATGCCCTCAAGATCGC-3′, and reverse, 5′-TTGTTCTCTAGCACGTCCCG-3′) and for dynamin 2 (forward, 5′-CAGAGCGCCGGCAAAAGTTC-3′, and reverse, 5′-GGCCTCCGGGTGACAATTC-3′). For normalization, GAPDH (glyceraldehyde-3-phosphate dehydrogenase) primers (forward, 5′-TGTCCGTCGTGGATCTGAC-3′, and reverse, 5′-CCTGCTTCACCACCTTCTTG-3′) were used. Efficiency for each primer pair was determined and the relative expression levels were calculated using the threshold cycle (ΔΔC_T) method. In addition, PCR was performed with Taq DNA polymerase (Life Technologies) and the following primer pairs: (i) to detect dynamin 1 transcripts (forward, 5′-TCTGATGCCCTCAAGATCGC-3′, and reverse, 5′-TTGTTCTCTAGCACGTCCCG-3′), which generate a 75-bp fragment only in control cells spanning exons 4 and 5, (ii) dynamin 2 primers (forward, 5′-CGAGAGTCTTAGTCGGGGGA-3′, and reverse, 5′-GATGCCTTTGTTGGTGCCTG-3′) generate a 388-bp fragment in control cells spanning exons 1 to 3 or a 315-bp fragment in DKO cells with a deleted exon 2, and (iii) as a control, GAPDH primers (forward, 5′-TGATGACATCAAGAAGGTGGTGAAG-3′, and reverse, 5′-TCCTTGGAGGCCATGTGGGCCAT-3′), which generate a 240-bp fragment.

Immunocytochemistry and antibodies.

Infected control and DKO cells were fixed with 2% formaldehyde for 10 min and permeabilized with 0.5% NP-40 for 10 min. To detect HSV-1-infected cells, samples were stained for 1 h with mouse anti-ICP0 (11060; 1:60) (47). SFV-infected cells were visualized by staining for 1 h with mouse anti-double-stranded RNA (dsRNA) 3G1 (1:100) (48), followed by incubation with the corresponding secondary antibody and DAPI (4′,6-diamidino-2-phenylindole) as nuclear counterstain, all at room temperature. In addition, fixed cells were stained with mouse anti-endophilin-A2 (1:50; Santa Cruz) or mouse anti-α-tubulin (1:200; Millipore), visualized by the corresponding secondary antibodies, and counterstained with DAPI. Staining of F-actin was performed with TRITC (tetramethylrhodamin isothiocyanate)-conjugated phalloidin (Sigma). Microscopy was performed using a Zeiss Axiophot or a Leica DM IRB/E microscope linked to a Leica TCS-SP/5 confocal unit. Images were assembled using Photoshop (Elements 2018).

Inhibitor studies.

The dynamin inhibitor dynasore (Tocris) was dissolved in dimethyl sulfoxide (DMSO) as stock solution, aliquoted, and stored at −20°C (49). Cells were treated with increasing drug concentrations diluted in DMSO for 30 min at 37°C before HSV-1 infection at 20 PFU/cell. Dynasore was present throughout infection.

Latrunculin A (Biomol), an inhibitor of actin polymerization, was dissolved in DMSO as stock solution, aliquoted, and stored at −20°C (37). Cells were seeded on coverslips coated with collagen G (Roche; 0.02 mg/ml) to avoid detachment upon drug treatment and were pretreated with 0.1 μM latrunculin A for 5 min at 37°C followed by infection with HSV-1 at 20 PFU/cell in the presence of the drug for 2 h. To allow cell recovery for visualization of ICP0, latrunculin A was replaced by medium for 1 h at 37°C.

Alexa Fluor 555-conjugated cholera toxin B (Molecular Probes) served as a control for endocytic uptake at low temperature. After serum starvation for 30 min at 37°C, cells were washed with PBS containing 0.1 mM CaCl₂ and 1 mM MgCl₂ (PBS-C/M), and cholera toxin B (1 μg/ml), diluted in PBS-C/M, was added. The cells were incubated for 30 min on ice, washed with PBS-C/M, and then incubated in serum-free medium for 45 min at 37°C or 7°C or on ice. To stop cholera toxin B uptake, cells were placed on ice followed by three washing steps with PBS-C/M (50).

Transmission electron microscopy.

Control and DKO cells were incubated with HSV-1 at 400 PFU/cell for 30 min at 37°C. At 30 min p.i. cells were fixed in 2% glutaraldehyde and 4% paraformaldehyde in PBS, pelleted, and stored overnight, all at room temperature. Cells were then washed 6 times for 10 min in cacodylate buffer and fixed for 1 h with 1% osmium tetroxide in cacodylate buffer. After washing 6 times for 10 min in water, cells were dehydrated in a graded series of ethanol and finally infiltrated with Spurr embedding resin. Samples were transferred to Spurr in BEEM capsules and the resin was polymerized for 7 h at 70°C. Ultrathin sections were cut with a Leica UC7 microtome, stained with uranyl acetate and lead citrate, and analyzed in a JEOL JEM1400 Plus operated at 120 kV. Images were recorded using a Matataki 2Kx2K camera.

Control cells were seeded on ACLAR films (Plano) and incubated with HSV-1 at 400 PFU/cell and 7°C. At 2 h p.i. cells were fixed in 2% glutaraldehyde, 2.5% sucrose, 3 mM CaCl₂, and 100 mM HEPES for 30 min at room temperature and 30 min at 4°C. After 3 washings in cacodylate buffer and fixation with 1% osmium tetroxide, 1.25% sucrose, and 1% potassium ferricyanide, cells were washed 3 times in cacodylate buffer and dehydrated in a graded series of ethanol, followed by Epon embedding. Ultrathin sections were cut with a Leica UC7 microtome, stained with uranyl acetate, and analyzed in a Zeiss EM109 operated at 80 kV. Images were taken using a TRS 2K slow-scan charge-coupled-device (CCD) camera.

Statistics.

For the statistical analyses, Student’s t tests were performed to calculate P values using the unpaired two-tailed method. Differences were considered to be statistically significant at P values of ≤0.05.