Dynamin Is Required for Efficient Cytomegalovirus Maturation and Envelopment

Viruses are known to exploit specific cellular functions at different stages of their life cycle in order to replicate, avoid immune recognition by the host and to establish a successful infection. Cytomegalovirus (CMV)-infected cells are characterized by a prominent cytoplasmic inclusion (virus assembly compartment [vAC]) that is the site of virus maturation and envelopment. While endocytic membranes are known to be the functional components of vAC, knowledge of specific endocytic pathways implicated in CMV maturation and envelopment is lacking. We show here that dynamin, which is an integral part of host endocytic machinery, is largely dispensable for early stages of CMV infection but is required at a late stage of CMV maturation. Studies on dynamin function in CMV infection will help us understand the host-virus interaction pathways amenable to targeting by conventional small molecules, as well as by newer generation nucleotide-based therapeutics (e.g., small interfering RNA, CRISPR/CAS gRNA, etc.).

ABSTRACT

Cytomegalovirus secondary envelopment occurs in a virus-induced cytoplasmic assembly compartment (vAC) generated via a drastic reorganization of the membranes of the secretory and endocytic systems. Dynamin is a eukaryotic GTPase that is implicated in membrane remodeling and endocytic membrane fission events; however, the role of dynamin in cellular trafficking of viruses beyond virus entry is only partially understood. Mouse embryonic fibroblasts (MEF) engineered to excise all three isoforms of dynamin were infected with mouse cytomegalovirus (MCMV-K181). Immediate-early (IE1; m123) viral protein was detected in these triple dynamin knockout (TKO) cells, as well as in mock-induced parental MEF, at early times postinfection, although levels were reduced in TKO cells, indicating that virus entry was affected but not eliminated. Levels of IE1 protein and another viral early protein (m04) were normalized by 48 h postinfection; however, late protein (m55; gB) expression was reduced in infected TKO cells compared to parental MEF. Ultrastructural analysis revealed intact stages of nuclear virus maturation in both cases with equivalent numbers of nucleocapsids containing packaged viral DNA (C-capsids), indicating successful viral DNA replication, capsid assembly, and genome packaging. Most importantly, severe defects in virus envelopment were visualized in TKO cells but not in parental cells. Dynamin inhibitor (dynasore)-treated MEF showed a phenotype similar to TKO cells upon mouse cytomegalovirus infection, confirming the role of dynamin in late maturation processes. In summary, dynamin-mediated endocytic pathways are critical for the completion of cytoplasmic stages of cytomegalovirus maturation.

IMPORTANCE Viruses are known to exploit specific cellular functions at different stages of their life cycle in order to replicate, avoid immune recognition by the host and to establish a successful infection. Cytomegalovirus (CMV)-infected cells are characterized by a prominent cytoplasmic inclusion (virus assembly compartment [vAC]) that is the site of virus maturation and envelopment. While endocytic membranes are known to be the functional components of vAC, knowledge of specific endocytic pathways implicated in CMV maturation and envelopment is lacking. We show here that dynamin, which is an integral part of host endocytic machinery, is largely dispensable for early stages of CMV infection but is required at a late stage of CMV maturation. Studies on dynamin function in CMV infection will help us understand the host-virus interaction pathways amenable to targeting by conventional small molecules, as well as by newer generation nucleotide-based therapeutics (e.g., small interfering RNA, CRISPR/CAS gRNA, etc.).

INTRODUCTION

Endocytic pathways are important for cellular entry of several viruses (1–5); however, their role in postentry stages of virus replication is far from resolved. Maturing herpesvirus nucleocapsids undergo primary envelopment at the inner nuclear membrane, traverse through the nuclear envelope, uncoat at the outer nuclear membrane, and reach the cytoplasm, where secondary or final envelopment takes place (6, 7). The cytoplasmic stage of herpesvirus maturation has been particularly challenging to study because a myriad of host and viral factors contribute to this process (6, 8). The identity of the cellular membranes that contribute to final virus envelope has been a topic of several studies (9–11). A further challenge in these studies has been the possibility of relocalization of cellular markers during infection. To elaborate this point, biomarkers that associate with the endoplasmic reticulum (ER) may not associate with the ER during infection or the ER membranes may form completely different structures during infection. For human cytomegalovirus (HCMV), elegant 3-dimensional confocal studies have shown the organization of a virus assembly compartment (vAC) in the cytoplasm, the site of virus maturation (12, 13). The vAC consists of several host organelles organized in specific shape and capacity with early endosomes forming the core of the structure. A similar vAC has been described for mouse cytomegalovirus (MCMV)-infected cells (14). Moreover, endosomal processes have been implicated in cytoplasmic maturation of several herpesviruses (15–19). Endocytic motifs in herpes simplex virus (HSV) envelope glycoprotein B (gB) are required for proper recycling of gB from cell surface to trans-Golgi network during maturation and thereby determine the infectivity of maturing virus (20). Similar endocytic processes for internalization of pseudorabies virus and CMV gB have also been reported (21, 22), and there is evidence that endocytic membranes are used for envelopment of several herpesviruses, including HSV, varicella-zoster virus, and CMV (10, 13, 17, 23, 24).

Dynamins and dynamin-related proteins constitute a superfamily of large self-assembling GTPases (an enzyme that can bind and hydrolyze GTP) that mediate membrane fission and fusion in biological processes such as endocytosis, vesicle trafficking, cell division, and organelle division (25). They are distinct from the small, Ras-like GTPases due to their oligomerization-dependent activation, the capacity to interact directly with membrane lipids, and their low GTP binding activity (26). Dynamins work twice in the mechanism of endocytosis: early in the constriction of the invaginating vesicle and late in its scission (27). Dynamins are known to be required for clathrin-mediated endocytosis (28, 29). In mammals, classical dynamins include dynamins 1, 2, and 3. Dynamin 1 is enriched within the brain and localizes to presynaptic terminals, and dynamin 2 has a ubiquitous tissue distribution, whereas dynamin 3 is localized in the testis and the brain (25).

Earlier, we studied the process of HCMV maturation in cells where dynamin-clathrin pathways were pharmacologically inhibited (30). One of the small molecules, dynasore, used in this study specifically inhibits dynamin function (31). In the present study, we utilized the recently established conditional triple-dynamin-knockout (TKO) mouse embryonic fibroblasts (32) to study the involvement of endocytic pathways in MCMV maturation. The use of TKO cells over the drug is preferred because adverse side effects of the dynamin inhibitor dynasore cannot be entirely ruled out. The results of this study reveal that dynamin is critical for a late stage of virus maturation. Studies on dynamin function in herpesvirus infection will help us understand the host-virus interaction pathways amenable to targeting by conventional small molecules as well as by newer generation nucleotide-based therapeutics (e.g., small interfering RNA, CRISPR/Cas9 gRNA, etc.). Targeting dynamin with pharmaceutical compounds has already been shown to have prophylactic potential against several infectious agents (reviewed in reference 33).

(This article was submitted to an online preprint archive [34].)

RESULTS

MCMV replicates to low titers in dynamin-depleted fibroblasts.

Dynamin was depleted in engineered mouse embryonic fibroblasts (MEF) by treatment with 4-hydroxytamoxifen (4-HT) for 48 h (Fig. 1A), leading to the generation of TKO cells as described previously (32). Whole-cell lysates of MCMV-K181-infected (+) or mock-infected (−) MEF showed similar levels of dynamin, whereas dynamin was reduced to insignificant levels in TKO cells. Mock-induced parental MEF and TKO cells were infected with MCMV-K181 at a high (3.0) or low (0.01) multiplicities of infection (MOIs) and monitored for virus growth. Cells were harvested in the medium at 3 or 6 days postinfection (dpi) and analyzed for PFU on wild-type MEF. At 3 dpi, MCMV titers were reduced about 10-fold for both low- and high-MOI infections in TKO cells compared to parental MEF (Fig. 1B). At 6 dpi, TKO virus titers were reduced >200-fold for low-MOI infections and ≈10-fold for high-MOI infections (Fig. 1B). All of these differences were statistically significant. Taken together, these data suggest that depletion of dynamin causes severe growth impairment of MCMV.

FIG 1.

Dynamin depletion impacts the growth of MCMV in fibroblasts. Dynamin was depleted in engineered MEF by a 4-HT-inducible knockout strategy leading to the generation of TKO cells as described earlier (32). MEF were treated with 4-HT for 2 days, and then 4-HT-containing medium was replaced with fresh 4-HT-free medium and cells were incubated for an additional 2 days. (A) Parental MEF and TKO cells were infected with the MCMV-K181 strain at an MOI of 3.0 (+) or mock-infected (−), and cell lysates were harvested at 4 hpi for immunoblot probing for dynamin. β-Actin was used as a loading control. (B) Parental MEF and TKO cells were infected with MCMV-K181 strain at an MOI of 3.0 or 0.01, and cells with media were harvested at 3 or 6 dpi before plating for virus titers on wild-type MEFs. Triplicate samples were used in experiments. A two-tailed unpaired t test with Welch’s correction (unequal variance assumption) was used for statistical analysis of differences. P values of <0.05 were considered significant (*).

Dynamin depletion interferes with MCMV early as well as late gene expression.

To explore the impact of dynamin on virus entry, parental MEF and TKO cells were infected with MCMV-K181 at an MOI of 3, and cell lysates were analyzed for expression of immediate-early (IE1; m123) protein. At 4 and 24 h postinfection (hpi), IE1 was detected in both MEF and TKO cells; however, the levels of IE1 were reduced in TKO cells (Fig. 2), indicating that virus entry was affected but not completely eliminated in TKO cells. Similarly, reduced levels of early protein m06 were detected in TKO cells compared to MEF at 24 hpi. Expression levels were detected to be similar in TKO and MEF for both IE1 and m06 at 48 hpi. In contrast, late viral protein m55 was expressed at lower levels in TKO cells at 48 hpi. Altogether, these data indicate that MCMV enters less efficiently in dynamin-depleted fibroblasts but establishes infection, albeit with a compromised expression of late viral proteins.

FIG 2.

Dynamin depletion reduces the entry of MCMV in fibroblasts and interferes with late protein expression. Parental MEFs and TKO cells were infected with MCMV-K181 at an MOI of 3.0 (+) or mock infected (−). Cells were harvested at 4, 24, and 48 hpi and probed for immediate-early (IE1; m123), early (m06), and late (m55) viral proteins. β-Actin served as loading control.

There is a certain possibility that inefficient virus entry in TKO cells, as evidenced by reduced IE1 expression at early times postinfection (Fig. 2), leads to an overall delayed replication cycle. To probe this further, we performed a full single-step virus growth curve analysis. Parental MEF and TKO cells were either mock infected or infected with MCMV-K181 at an MOI of 3.0, and cells and medium were harvested followed by quantification of PFU. As summarized in Fig. 3A, viral growth defect in TKO cells was evident as early as 3 dpi and continued up to 7 dpi. We performed a cell viability test of mock- and MCMV-K181-infected MEF and TKO cells up to 7 dpi in parallel to rule out the possibility that this growth defect could be due to a viability disadvantage in TKO cells (Fig. 3B). Uninfected MEF and TKO cells were >95% viable until 5 days in cell culture. At 7 days, TKO cells showed more cell death compared to MEF. During MCMV infection, both TKO and MEF cells showed significant cell death starting at 1 dpi; however, TKO cells showed more resistance to MCMV-induced cell death, especially at 3 and 5 dpi.

FIG 3.

Impact of dynamin depletion on MCMV growth and cell viability. (A) Parental MEF and TKO cells were infected with MCMV-K181 at an MOI of 3.0, and cells with media were harvested at 0 to 7 dpi followed by estimating virus titers on wild-type MEF. (B) Parental MEF and TKO cells were infected with MCMV-K181 at an MOI of 3.0 (+) or mock infected (−), and cell viability at the indicated time points was determined by trypan blue exclusion assay. Triplicate samples were used in experiments. A two-tailed unpaired t test with Welch’s correction (unequal variance assumption) was used for statistical analysis of differences. P values of <0.05 were considered significant (*).

To probe whether the observed growth defect in TKO cells (Fig. 3A) reflects a defect in virus replication or release from cells, we separated cells and supernatants at different times postinfection and evaluated the viral titers (Fig. 4). The results indicate significant reduction in both cell- and supernatant-associated virus in TKO cells at 3 dpi for both low and high MOIs. A similar trend was observed at 6 dpi for low MOIs. At high MOIs, the supernatant titers for TKO cells were significantly reduced but cell-associated titers were equivalent to MEF cells at 6 dpi, likely because only a small fraction of both MEF and TKO cells are viable at 5 and 7 dpi (Fig. 3B) and the majority of virus particles are expected to be released into the medium. In summary, the data corroborate the results from growth analysis (Fig. 1 and 3) and protein expression studies (Fig. 2) that virus growth is delayed in TKO cells; however, it also indicates that virus growth in TKO cells catches up with wild-type MEF at late time postinfection and the growth defect observed at this time is almost entirely due to a defect in virus release.

FIG 4.

Dynamin depletion reduces both cell-associated and cell-free virus levels. Parental MEF and TKO cells were infected with MCMV-K181 at an MOI of 3.0 or 0.01, and cells with media were harvested at 3 or 6 dpi. Media and cells were separated by low-speed centrifugation, and viral loads in the supernatant (S) and cells (C) were quantified by determining the titers on wild-type MEF. Triplicate samples were used in experiments. A two-tailed unpaired t test with Welch’s correction (unequal variance assumption) was used for statistical analysis of differences. P values of <0.05 were considered significant (*).

Lack of dynamin does not impair the localization of early and late viral proteins.

In order to understand the impact of dynamin depletion on virus protein trafficking, we examined the localization of MCMV early and late proteins in MEF and TKO cells at 48 hpi by immunofluorescence assay (IFA). MCMV immediate-early (IE1; m123) protein was expressed and localized to the nucleus in both parental MEF and TKO cells (Fig. 5, top two panels). MCMV early (m04) protein was expressed in the cytoplasm but concentrated around the nuclear periphery in both cell types (Fig. 5, middle two panels). Similarly, MCMV late (m55) protein was expressed in the cytoplasm of both cells in diffuse, as well as punctate (possibly virion associated), forms (Fig. 5, bottom two panels). Collectively, these data indicate that dynamin depletion does not affect the expression and localization of early to late viral proteins that is evident in IFA.

FIG 5.

Dynamin depletion does not affect the localization of early and late viral proteins in infected cells. Parental MEF and TKO cells were infected with MCMV-K181 at an MOI of 3.0, fixed for IFA at 48 hpi, and stained for MCMV immediate-early (IE1; m123), early (m04), and late (m55) proteins. MCMV proteins labeled as green (left column), DNA (in nuclei) detected by Hoechst (middle column), and composite images (overlay, right column) from the same field of each panel are shown. IE1 protein localized to the nuclear compartment, whereas m04 and m55 proteins localized to the cytoplasm concentrating at the periphery of nuclei in both MEF and TKO cells.

Lack of dynamin affects the formation of vAC.

vAC is known to be the site of cytoplasmic virus maturation. Since the growth data (Fig. 4) showed a defect in virus release at late times postinfection, we investigated the formation of vAC in TKO and MEF cells to analyze any defects that would translate to a defect in virus envelopment and release. Mock-infected MEF showed the presence of perinuclear Golgin-97 staining, consistent with the presence of Golgi stacks (35) (Fig. 6). Similar perinuclear Golgin-97 staining was also observed in mock-infected TKO cells. Infected MEF showed a perinuclear Goglin-97 ring formation, as observed in HCMV-infected fibroblasts and marks the vAC (12, 36). In infected TKO cells, the Golgin-97 accumulated in the perinuclear region, but none of the cells examined (>100) showed the typical ring formation. Thus, the data indicate that assembly of vAC is compromised in TKO cells.

FIG 6.

Formation of vAC is compromised in dynamin-depleted cells. Parental MEF and TKO cells were mock infected or infected with MCMV-K181 at an MOI of 3.0, fixed for IFA at 48 hpi, and stained for Golgin-97. Golgin-97 labeled as green (left column), DNA (in nuclei) detected by Hoechst (middle column), and composite images (overlay, right column) from the same field of each panel are shown.

MCMV nuclear stages are intact in dynamin-depleted fibroblasts but cytoplasmic virus maturation is significantly impaired.

Parental MEF and TKO cells were infected with MCMV-K181 at an MOI of 3.0. At 72 h postinfection, the cells were fixed for processing and imaging using a transmission electron microscope. Both cell types showed typical infected cell morphologies with a kidney-bean-shaped nucleus and the presence of nuclear and cytoplasmic inclusions (Fig. 7A and E). The nucleus of both MEF and TKO cells contained all three types of capsids (A [empty], B [scaffold-containing], and C [DNA-containing]) reported for herpesviruses (Fig. 7B and F) (8). Quantification of these capsid types revealed similar proportions in both cell types (Fig. 7I), indicating intact nucleocapsid maturation. In contrast, very few virus particles were observed in the cytoplasm of TKO cells (Fig. 7G and H); however, these particles contained genomic DNA, and tegument proteins were evident on the surfaces of these capsids (Fig. 7G inset). Virus envelopment was not evident in TKO cells, whereas several enveloping (Fig. 7C, inset) or enveloped virus particles were present in the cytoplasm of parental MEF (Fig. 7C and D). Another striking difference was the presence of intact Golgi stacks in TKO cells (Fig. 7G and H), which were fragmented to different degrees in MEF (Fig. 7C and D), indicating increased vesiculation. Examination of the cytoplasm revealed no virions or partially enveloped particles in TKO cells (Fig. 7G, H, and J) in contrast to a significant number of virions and enveloping particles in the parental MEF (Fig. 7C, D, and J). In summary, the data indicate that cytoplasmic maturation is severely compromised in TKO cells.

FIG 7.

MCMV nuclear stages are intact in TKO cells, but cytoplasmic virus maturation is significantly impaired. Transmission electron micrographs of MEF (A to D) and TKO (E to H) cells infected with MCMV-K181. Cells were infected at an MOI of 3.0 and fixed for processing at 3 dpi. (A and E) A single infected cell showing the nucleus, as well as cytoplasm. (B and F) Infected cell nucleus illustrating A (black arrows), B (black arrowheads), and C (white triangular arrows) capsids. (C, D, G, and H) Cytoplasmic section illustrating DNA containing capsids (black arrowheads), partially enveloped/enveloping capsids (white arrows), and virus-like particles that are difficult to type morphologically (black arrow). The inset in panel C magnifies and illustrates a partially enveloped virion, and the inset in panel G illustrates nonenveloped DNA-containing capsids with evidence of intact tegument. Intact Golgi stacks are evident in panels G and H, whereas a fragmented Golgi stack is prominent in panels (C and D). (I) Nuclear capsids per cell were quantified in MEF (n = 6) and TKO cells (n = 7). A representative image of each type of capsid is shown under each graph. (J) MCMV particles per cell in the cytoplasm of MEF (n = 8) and TKO cells (n = 7) were quantified. A representative example of the virion, partially enveloped particle, and irregular particle is shown under each graph. A two-tailed unpaired t test with Welch’s correction (unequal variance assumption) was used for statistical analysis of differences. P values of <0.05 were considered significant (*).

Dynasore-treated cells mimic the dynamin knockout phenotype.

To rule out any unknown peculiarity in TKO cells that may be responsible for the virus maturation defects evident in these cells, we treated wild-type primary MEF with a 50 µM concentration of an established dynamin inhibitor (dynasore) and subsequently infected them with MCMV-K181 to study virus entry and growth. Dynasore is a small molecule that is well established to specifically abolish dynamin activity in cells without an impact on cell viability (31). Dynasore-treatment resulted in a decrease in IE1 gene expression at 4 h postinfection, but this expression was normalized at 48 h postinfection (Fig. 8A), similar to the results obtained for TKO cells (Fig. 2). Analysis of virus growth at low (0.05) and high (3.0) MOIs indicated significant differences between dynasore-treated and mock-treated cells (Fig. 8B). These results correlate with the results obtained for TKO cells (Fig. 1 and 3). Thus, the phenotype we observed in TKO cells is indeed due to the deficiency of dynamin function and is not an aberrant effect of dynamin depletion on a single cell type.

FIG 8.

Dynasore-treated MEF allow virus entry and gene expression but compromise virus growth. (A) MEF were treated with dynasore (50 µM) or mock treated (DMSO) and infected with MCMV-K181 at an MOI of 3.0. Cell lysates were harvested at 4 and 48 hpi for immunoblot probing for MCMV IE1 protein. β-Actin was used as a loading control. (B) Dynasore- or mock-treated cells were infected with MCMV-K181 at an MOI of 0.05 or 3.0, and cells were harvested at 3 or 6 dpi before plating for virus titers on wild-type MEF. Triplicate samples were used in experiments. A two-tailed unpaired t test with Welch’s correction (unequal variance assumption) was used for statistical analysis of differences. P values of <0.05 were considered significant.

DISCUSSION

Endosomal membranes have been implicated in herpesvirus maturation; however, the role of specific endocytic pathways in herpesvirus morphogenesis remains largely unexplored. In the current work, we show that dynamin-mediated endocytic pathways are important for MCMV maturation. We utilized recently characterized triple-dynamin-knockout cells for these studies to provide convincing evidence that these pathways are important at a late stage of the MCMV life cycle that involves virus morphogenesis, gain of infectivity, and egress of mature particles.

The current studies were influenced by our earlier studies on HCMV where we utilized laboratory strains that utilize a glycoprotein-mediated fusion mechanism at plasma membrane to enter the cells instead of endocytosis (30). The HCMV entry pathways in different cell types have been studied in detail, and it is well known that laboratory strains enter the cells via a pH-independent fusion mechanism at the plasma membrane (37, 38). We used clathrin and dynamin inhibitors in the study described above to reveal a role for endocytic processes on HCMV maturation. The data from these studies indicated an impact of pharmacological inhibition of dynamin-clathrin pathways on HCMV maturation; however, virus entry and early gene expression remained intact. To be able to extend the study of virus biology in an appropriate animal model, we utilized an established dynamin-knockout mouse cell model that has been extensively characterized (32, 39, 40) and is free from any side effects that chemical inhibitors may have on cells. It also provides the ability to test MCMV instead of HCMV, which would be useful for future in vivo studies looking to characterize the effect of endocytic inhibitors in a mouse model of CMV infection. This is important because dynamin inhibitors have already shown a therapeutic potential against several infectious agents (reviewed in reference 33).

After successfully establishing a nearly complete depletion of dynamin in TKO cells, we measured its impact on MCMV growth and yield. Dynamin depletion had significant impact on virus growth at low, as well as high, MOIs. Although analysis of early viral gene expression revealed an impact on early times of infection (4 and 24 h), these differences were normalized by 48 h, indicating that the reduction in virus growth in TKO cells observed at late times postinfection is unlikely due to defects in virus entry or early gene expression. To further investigate this point, we analyzed the expression and distribution of early to late viral proteins in infected cells. The expression of early proteins (IE1, m06) was at equivalent levels at 48 hpi; however, late protein (m55) expression was reduced, indicating that the defects in CMV replication in dynamin-depleted cells relate to late steps in virus replication that include the expression of late genes. It is likely that MCMV partially utilizes an endosome-dependent mechanism for entry in addition to fusion at plasma membrane; thus, IE1 levels at early time points, which directly correlate with virus entry, are reduced. However, IE1 levels are normalized by 48 h postinfection. Considering that we are probing whole-cell lysates, a difference in infectivity would be reflected in overall lower protein levels throughout the infection, which does not seem to be the case here. It is also unlikely that trafficking-dependent proteins would show lower levels in Western blots of whole-cell lysates. Localization of viral proteins (IE1, m04, and m55) in TKO cells was not significantly different from paternal MEF. These results rule out the possibility that virus growth defects could be due to an impact on early viral gene expression or abnormal localization of major viral proteins in dynamin-depleted cells. Investigation of cellular protein (Golgin-97) localization revealed that vAC does not form properly in TKO cells; therefore, cytoplasmic virus maturation and egress are likely impaired.

The possibility of a defect at a late stage of virus maturation was analyzed by ultrastructural detailed analysis of infected cells. Nuclear stages of virus replication, including capsid assembly and DNA packaging, were intact based on the numbers and types of capsid particles present in the nuclei of TKO cells versus parental MEF. Equivalent proportions of A, B, and C capsid forms (8) were observed in both cell types (Fig. 7I). Most importantly, the presence of similar numbers of C capsids (DNA packaged capsids) in two cell types indicate that dynamin did not influence the stages of virus replication up to the point of production of DNA packaged capsids, which go on to become infectious virions. Thus, the defects would either be at the nuclear egress of packaged capsids or at the cytoplasmic stage of virus maturation and egress. Further, a block at nuclear egress was ruled out on the basis of the absence of any large buildup of assembled particles at the inner nuclear membrane in ultrastructural images (Fig. 7E and F). Moreover, a few virus particles were present in the cytoplasm of TKO cells, indicating that nuclear egress could not be completely blocked. These cytoplasmic virus particles in TKO cells were mostly unenveloped (Fig. 7G) or appeared to be morphologically abnormal/degrading (Fig. 7H, arrowhead). This is not unusual since the absence of an envelope would ultimately lead to the degradation of naked virus particles in the cytoplasm. We saw a similar phenotype of capsids degrading in the cytoplasm of pp150 mutant virus-infected cells in our earlier study (36). This degradation happened despite the relocalization of viral DNA from the nucleus to the cytoplasm (41). Since viral DNA cannot exit the nucleus independent of virus capsids (and even if it did, it would degrade rapidly due to strong cytoplasmic nucleases), the most convincing explanation is that capsids carry viral DNA to the cytoplasm but are unable to maintain their integrity in the absence of essential inner tegument proteins.

Exocytic vesicles and vesicles containing clathrin-coated pits are generated from Golgi fragmentation. Anti-dynamin antibodies have been shown to block the formation of these vesicles in a cell-free assay (42). This vesiculation is restored upon the addition of purified dynamin. Thus, it comes as no surprise that dynamin-depleted cells have more intact Golgi stacks compared to parental cells when infected by CMV. This vesiculation may contribute significantly to CMV envelopment, which is compromised in dynamin-depleted cells. The late virus maturation defect observed in the present study, along with an evident disruption of vAC formation, point toward the important role of dynamin-mediated vesicular pathways in CMV maturation (Fig. 9). For HCMV infection, it has been shown that vAC formation is dependent on late viral gene expression; thus, the observed disruption of vAC structure could also be an indirect result of dynamin depletion on late protein expression.

FIG 9.

Proposed model for the functions of dynamin in CMV-infected cells. (A) Dynamin plays a role in membrane remodeling at different stages of endosomal trafficking. Newly synthesized proteins in the ER are sorted in the TGN targeted for their final destination in the cell or secreted forms. TGN also receives input from the endocytic pathways (broken arrows) where dynamin is implicated. (B) Proposed points of critical activity of dynamin in CMV-infected cells are marked with an asterisk. ER, endoplasmic reticulum; TGN, trans-Golgi network; NC, nuclear capsid; vAC, virus assembly complex. The “gear-shaped” symbols on the right side of panel B indicate glycoproteins.

There is little doubt that endosomal systems contribute to the process of herpesvirus maturation; however, examples of specific virus proteins hijacking host endocytic machinery are lacking. A study based on mass spectroscopic analysis of protein interactions in HCMV-infected cells indicated that the tegument protein pp150 directly interacts with clathrin (43). pp150 has established roles in virus maturation (36), and it is certainly possible that this pp150-clathrin interaction is functional during virus maturation and egress. More specific interactions of herpesvirus proteins with endocytic systems are likely to be revealed by studying host factors, such as dynamin, that are important in the late stages of herpesvirus maturation.

MATERIALS AND METHODS

Cells.

Mouse embryonic fibroblasts (MEF) were cultured in Dulbecco modified Eagle medium (DMEM; Cellgro, Manassas, VA) containing 4.5 g/ml glucose, 10% fetal bovine serum (SAFC, Lenexa, KS), and 1 mM sodium pyruvate, 2 mM l-glutamine, and 100 U/ml penicillin-streptomycin (Cellgro, Manassas, VA) at 37°C with 5% CO₂. The deletion of dynamin in engineered triple-dynamin-knockout (TKO) cells is mediated by a tamoxifen-inducible-knockout strategy. Briefly, these cells express a Cre-estrogen receptor mutant knock-in transgene from the ROSA26 locus (44). Thus, Cre is only shuttled into the nucleus in response to tamoxifen exposure. The TKO cells were treated with a 10 µM stock of 4-hydroxytamoxifen (4-HT; Sigma, catalog no. H-6278) in 100% ethyl alcohol for 2 days, and then the medium was changed back to normal tamoxifen-free medium. Depletion of dynamin was evident at 3 to 4 days posttreatment in Western blots (described below) of whole-cell lysates.

Antibodies, immunofluorescence assays, and immunoblots.

The mouse anti-dynamin clone 41 from BD (catalog no. 610245) was used to probe for dynamin in the Western blotting. Mouse cytomegalovirus IE1 (m123), m04, m06, and m55 mouse antibodies (catalog no. HR-MCMV-08, HR-MCMV-01, HR-MCMV-02, and HR-MCMV-04) were purchased from Center for Proteomics, University of Rijeka, and used at a 1:1,000 dilution. Golgin-97 rabbit antibody was purchased from Cell Signaling Technology (catalog no. 13192S) and used at a 1:1,000 dilution. Fluorescent label-tagged secondary antibody DYLIGHT 594 was purchased from Thermo Scientific Pierce and used at a 1:1,000 in immunofluorescent assays (IFAs), as described below. Hoechst 33258 (Thermo Scientific Pierce) staining (1:3,000 dilution) identified the nuclei in IFA. Anti-β-actin antibody (AC-74; Sigma-Aldrich, St. Louis, MO) was used (1:1,000 dilution) as a control for sample loading in immunoblots (IB). Horseradish peroxidase-labeled anti-mouse IgG, IgM, and anti-rabbit IgG (catalog no. 31444 and 31460; Thermo Scientific, Rockford, IL) were used as the secondary antibodies at 1:3,000 dilutions for IBs. Blots were detected using ECL Western blotting detection reagents (GE Healthcare, Buckinghamshire, United Kingdom).

Virus.

MCMV strain K181 was grown in MEF cells. Virus stock was prepared in 3× autoclaved milk, sonicated three times, and stored at −80°C. During infection, the medium was removed from the wells of cell culture plates, and appropriately diluted virus stock was absorbed onto the cells in raw DMEM. Cells were incubated for 1 h with gentle shaking every 10 min, followed by three washes with phosphate-buffered saline (PBS). Fresh complete medium was added, and cells were incubated until the endpoint.

Cell viability assay.

Parental MEF and TKO cells grown on 12-well tissue culture plates were infected with MCMV-K181 at an MOI of 3.0 or mock infected at confluence. A 500-μl portion of fresh complete medium was added to the wells on days 3 day 6. At the designated time points, the medium was removed, and the cells were harvested by trypsinization. Cell viability was determined using trypan blue exclusion on a TC20 automated cell counter (Bio-Rad Laboratories, Hercules, CA) according to the manufacturer’s protocol.

Microscopy.

Samples were prepared using established protocols for IFA and confocal fluorescence microscopy. Briefly, mock-induced parental MEF or 4-HT-treated TKO cells were grown on coverslip inserts in 24-well tissue culture dishes and infected at an MOI of 3.0 at confluence. At the endpoint of the experiment, the cells were fixed in 3.7% formaldehyde for 10 min and incubated in 50 mM NH₄Cl in 1× PBS for 10 min to reduce autofluorescence. This was followed by washing in 1× PBS, incubation in 0.5% Triton X-100 for 20 min to permeabilize the cells, and finally a wash and incubation with primary and secondary antibodies at a 1:1,000 dilution in 0.1% bovine serum albumin in 1× PBS. Coverslips were retrieved from the wells, mounted on glass slides with a drop of mounting medium (Gel/Mount; Biomeda, Foster City, CA), and dried overnight before imaging. Images were acquired on an inverted Evos-FL microscope (Thermo Fisher Scientific, Waltham, MA) using ×100 objective. Samples for transmission electron microscopy were prepared by fixing the cells (MEF) at endpoint in 2.5% glutaraldehyde in 0.1 M cacodylate buffer (pH 7.2) for 2 h at room temperature. The cells were then washed with the same buffer and postfixed with buffered 1.0% osmium tetroxide at room temperature for 1 h. After several washes with 0.1 M cacodylate buffer, the cells were dehydrated with ethanol, infiltrated, and embedded in Eponate 12 resin (Ted Pella, Inc., Redding, CA). Cell culture plates were cracked with a hammer to release the resin after it had solidified, and ultrathin sections (60 to 70 nm) of monolayer cells were cut and counterstained using uranyl acetate and lead citrate. Examination of ultrathin sections was carried out on a Hitachi H-7500 transmission electron microscope operated at 75 kV, and images were captured using a BioScan (Gatan, Pleasanton, CA) charge-coupled device camera. The images were acquired and analyzed using DigitalMicrograph (Gatan) software.

Drug inhibition assay.

Confluent MEF monolayers were pretreated with dynasore (50 µM; catalog no. 324410; EMD Millipore Corp., Billerica, MA) for 1 h and infected with MCMV-K181 in medium containing the drug. Cells were washed with PBS and then incubated in the presence of the drug until the end of the experiment.

Virus titers.

Infected or mock-infected samples were harvested within the medium at the designated end points and stored at −80°C before titration. In some experiments, media and cells were separated by low-speed (<1,000 × g) centrifugation, and viral loads in the supernatant and cells were quantified by determining the titers on wild-type MEF. Titers were determined as described earlier (45) with some modifications. In brief, monolayers of MEF grown in 12-well plates, and serial dilutions of sonicated samples were absorbed onto them for 1 h, followed by three washes with PBS. A carboxymethyl cellulose (CMC; catalog no. 217274; EMD Millipore) overlay with complete DMEM (one part autoclaved CMC and three parts media) was added, and the cells were incubated for 5 days. At the endpoint, the overlay was removed, and the cells were washed twice with PBS. Infected monolayers were fixed in 100% methanol for 7 min, washed once with PBS, and stained with 1% crystal violet (catalog no. C581-25; Fisher Chemicals, Fair Lawn, NJ) for 15 min. Plates were finally washed with tap water and air dried, and plaques with clear zone were quantified.