The HSV-1 pUL37 protein promotes cell invasion by regulating the kinesin-1 motor

Significance

HSV-1 packages the cellular kinesin motor into virions and subsequently uses it to infect cells. Upon entering a cell, the motor remains in an off state until the viral particle arrives at the centrosome at which point the kinesin motor promotes trafficking to the cell nucleus. This nuclear targeting process is a critical component of the neuroinvasive property of the virus as it affords sustained dynein-based retrograde transport in axons that is uninterrupted by the dormant kinesin motor. Here, we report that a viral structural protein coordinates the spatial and temporal regulation of kinesin to achieve this two-step trafficking mechanism.

Abstract

Neurotropic alphaherpesviruses, including herpes simplex virus type 1 (HSV-1), recruit microtubule motor proteins to invade cells. The incoming viral particle traffics to nuclei in a two-step process. First, the particle uses the dynein–dynactin motor to sustain transport to the centrosome. In neurons, this step is responsible for long-distance retrograde axonal transport and is an important component of the neuroinvasive property shared by these viruses. Second, a kinesin-dependent mechanism redirects the particle from the centrosome to the nucleus. We have reported that the kinesin motor used during the second step of invasion is assimilated into nascent virions during the previous round of infection. Here, we report that the HSV-1 pUL37 tegument protein suppresses the assimilated kinesin-1 motor during retrograde axonal transport. Region 2 (R2) of pUL37 was required for suppression and functioned independently of the autoinhibitory mechanism native to kinesin-1. Furthermore, the motor domain and proximal coiled coil of kinesin-1 were sufficient for HSV-1 assimilation, pUL37 suppression, and nuclear trafficking. pUL37 localized to the centrosome, the site of assimilated kinesin-1 activation during infection, when expressed in cells in the absence of other viral proteins; however, pUL37 did not suppress kinesin-1 in this context. These results indicate that the pUL37 tegument protein spatially and temporally regulates kinesin-1 via the amino-terminal motor region in the context of the incoming viral particle.

Alphaherpesviruses, such as herpes simplex virus type 1 (HSV-1), are noted for their potent neuroinvasive properties and are one of the few classes of viruses that productively enter the nervous system and subsequently reemerge to spread host to host (1). Infections begin with fusion of the viral envelope with a cellular membrane and the deposition of the internal virion components, the DNA-containing capsid and an assortment of tegument proteins, into the cytosol (2, 3). Most of the tegument proteins disperse in the cell, but a few remain capsid bound (4, 5). Two of the latter, pUL36 and pUL37, promote capsid microtubule-based transport to the nucleus (6–10). The pUL36 tegument protein associates with two opposing microtubule motors: dynein–dynactin and conventional kinesin-1 (11, 12). Dynein–dynactin brings the capsid to the centrosome, which in neurons is responsible for long-distance retrograde axonal transport, and kinesin-1 subsequently routes the capsid from the centrosome to the nucleus (12). The spatial and temporal coordination of these cellular motors targets nuclear delivery of the viral genetic information in multiple cell types, including epithelial cells and neurons, and is critical for the establishment of life-long latency in the host nervous system.

We have recently reported that the kinesin-1 motor is packaged into the tegument of virions, which the virus then uses during the following round of infection. For example, during the infection of an epithelial cell, progeny virions are produced preloaded with the epithelial kinesin-1 motor, and upon subsequent infection of a neuron, the epithelial kinesin-1 is used to traffic capsids to the neuronal nucleus. We refer to this repurposing of a cellular factor as a proviral component of the virion architecture as “assimilation” (12). This finding suggests that the assimilated kinesin-1 motor is regulated so that it does not interfere with the opposing dynein motor during retrograde transport and only becomes active upon arrival at the centrosome.

The pUL37 tegument protein remains capsid associated upon entering a cell by virtue of its interaction with pUL36 (13) and is essential for efficient nuclear delivery (7). Our lab previously reported that mutating a conserved surface-exposed region of pUL37 (region 2; “R2”) causes the incoming capsid to display an aberrant bidirectional transport phenotype that results in a near-zero net displacement in axons and an inability of the virus to invade the nervous system in vivo (14). This stochastic back-and-forth motion could be explained if the assimilated kinesin-1 motor was prematurely active before capsids arrived at the centrosome.

In this report, wild-type HSV-1 and HSV-1 mutated in pUL37 R2 (R2*) were propagated on hTERT-immortalized human retinal pigmented epithelial (RPE) cells knocked-out for kinesin-1 to produce virions lacking assimilated kinesin (referred to here as “kinesin-less” virions). We recapitulate prior findings that HSV-1 R2* fails to sustain dynein-based retrograde axonal transport in primary sensory neurons and further demonstrate that retrograde axonal transport was intact when HSV-1 R2* was made kinesin-less. Despite the phenotypic restoration of retrograde axonal transport, kinesin-less HSV-1 R2* was defective for subsequent centrosome-to-nucleus trafficking as was expected for particles lacking assimilated kinesin-1. These findings demonstrate that pUL37 is a regulator of the kinesin-1 microtubule motor.

To further investigate the pUL37 R2 mechanism, HSV-1 was packaged with mutated kinesin-1 lacking its autoinhibitory activity. Surprisingly, HSV-1 carrying constitutively active kinesin-1 maintained processive retrograde movement and trafficked to neuronal nuclei on par with virus packaged with wild-type kinesin-1. Similarly, HSV-1 packaged with a truncated kinesin (aa1-561) that lacks the carboxy-terminal kinesin-light chain and cargo binding domains also supported infection. These results point to the kinesin-1 motor domain being directly targeted by the virus; however, pUL37 failed to suppress kinesin-1 on its own indicating that other components of the viral particle contribute to motor suppression. Nevertheless, pUL37 localized to centrosomes in the absence of other viral proteins, reinforcing that a pUL37-centrosome interaction could trigger the release of kinesin-1 suppression and the routing of capsids to the nucleus.

Results

The HSV-1 pUL37 R2 Effector Suppresses Assimilated Kinesin-1.

We previously reported that HSV-1 incorporates a cellular kinesin-1 motor as a structural component of its virions and that HSV-1 can be artificially produced lacking kinesin-1 (aka, “kinesin-less” virions) (12). Specifically, hTERT-immortalized retinal pigmented epithelial (RPE) cells express a single isoform of the kinesin-1 heavy chain, KIF5B, which was previously knocked out to produce cells lacking kinesin-1 (Fig. 1B) (15). Using the WT and KO cells, stocks of HSV-1 were produced that either possessed or lacked the kinesin-1 motor, respectively. HSV-1 was also produced carrying mutated forms of kinesin-1 (Fig. 1C). Packaging of mutated kinesin-1 was achieved by producing HSV-1 on knock-out cells that were transduced with a retroviral vector encoding the mutated KIF5B heavy chain fused to the mNeonGreen fluorescence protein. The modified kinesins referred to as “H, I” or “C, H, I” carried mutations in the kinesin-1 coiled-coil (“C”), hinge (“H”), and isoleucine-alanine-lysine (IAK; “I”) domains, as indicated (Fig. 1A). These mutations were designed to impair autoinhibition of the KIF5 motor as previously reported (16). The truncated kinesin referred to as “561” encoded amino acids 1-561 of the kinesin-1 heavy chain (17, 18). Five stocks each of wild-type HSV-1 and HSV-1 R2 mutant (R2*) were produced: either in the presence or absence of kinesin-1 or in the presence of one of the three mutated kinesin-1 isoforms.

Fig. 1.

Overview of virus production in the presence or absence of kinesin-1 (KIF5B) and three KIF5B mutant isoforms. (A) Schematic of kinesin-1 (KIF5) heavy chain highlighting three regions contributing to motor autoinhibition. (B) Western blot of hTERT-RPE cells expressing endogenous KIF5B (WT), knocked out for KIF5B (KO), and KO expressing a double (H, I), triple (C, H, I), or truncated (561) KIF5B mutant. (C) Diagram of HSV-1 production. Producer hTERT-RPE cells are infected with HSV-1 that is either wild type (WT) or R2 mutant (R2*). The harvested virus either contains or lacks a kinesin-1 motor based on the producer cell type. These virus stocks are used to infect normal sensory neurons or epithelial cells for further study.

We hypothesized that the aberrant motion previously reported for HSV-1 R2* could be explained if the R2 effector was a kinesin-1 suppressor, such that when mutated the assimilated kinesin-1 would compete with the dynein–dynactin motor thereby interfering with processive retrograde axonal transport (14). To test this model, retrograde axonal transport of HSV-1 was monitored by time-lapse fluorescence imaging in primary explants of dorsal root ganglia within the first hour postinfection. Recombinant viruses were used that express the mCherry fluorescent protein fused to the pUL25 capsid protein, and the dynamics of capsid transport were assessed by kymograph analysis (Fig. 2A) (10, 19). The motion of wild-type capsids was predominantly retrograde, consistent with long-distance trafficking to the soma and nucleus, whereas the R2 mutant lacked sustained retrograde transport. These results were consistent with previous reports (14, 20). However, HSV-1 R2* made kinesin-less did not display the transport defect and instead moved retrograde in axons on par with wild-type HSV-1 kinetics. This phenotype was noted in terms of the percent time of retrograde and opposing anterograde transport events per capsid (Fig. 2B), frequency of retrograde transport interruptions (Fig. 2C), and overall net displacement (Fig. 2D), with the latter approximating zero for the HSV-1 R2 mutant. These results show that in the absence of virion-incorporated kinesin-1, mutation of the pUL37 R2 effector had no impact in retrograde axonal transport, functionally demonstrating that R2 suppresses the assimilated kinesin-1 during retrograde axonal transport to the centrosome.

Fig. 2.

The HSV-1 pUL37 R2 effector promotes capsid retrograde axonal transport by suppressing assimilated kinesin-1. (A) Example retrograde axonal transport kymographs of WT HSV-1 produced from WT RPE cells (Left) and R2* HSV-1 produced from WT RPE cells (Right). Distance (d) and time (t) axis are indicated in the Right panel. (B) Frequency of capsid motion in axons: % time retrograde (light), % time anterograde (medium), and % time stopped (dark). (C) Average number of interruptions (reversals and stops) per retrograde moving capsid during a 20-s period. (D) Net displacement of capsids after 20 s. Positive values indicate movement toward neuronal soma (retrograde displacement). Error bars are SD; ****P < 0.0001.

Packaging Constitutively Active Kinesin-1 Does Not Impair HSV-1.

Kinesin-1 possesses an autoinhibitory mechanism that holds the motor in an off state when not bound to cargo (21). To determine whether the R2 effector suppresses kinesin-1 by engaging this mechanism, HSV-1 was packaged with constitutively active mutants of kinesin-1: “H, I” and “C, H, I” (Fig. 1A). These viruses were then studied following infection of explanted dorsal root ganglia as before. Wild-type HSV-1 packaged with the constitutively active motors displayed normal retrograde axonal transport dynamics, indicating that HSV-1 suppressed these motor isoforms and prevented interference with dynein-based retrograde axonal transport (Fig. 2). Nevertheless, HSV-1 R2* packaged with the constitutively active motors displayed aberrant transport dynamics consistent with the R2* phenotype, demonstrating that the constitutively active kinesins were assimilated into HSV-1 (otherwise the R2* viruses would have displayed processive retrograde transport as previously noted in the absence of kinesin-1). This latter finding supports that HSV-1 suppressed the constitutively active kinesin-1 motors during retrograde axonal transport. To examine this further, HSV-1 was packaged with truncated kinesin-1 [aa1-561], which lacks the kinesin-light chain and cargo binding domains of the carboxy terminus and is also a constitutively active mutant (17, 18). As with the “H, I” and “C, H, I”, “561” was effectively used by the virus for nuclear trafficking and was suppressed by pUL37. Taken together, these results indicate that the suppression of kinesin-1 enforced by the R2 effector is mediated independently of the kinesin-1 intrinsic autoinhibitory mechanism, carboxy-terminal cargo binding domain, and carboxy-terminal kinesin-light chain binding domain.

Assimilated Kinesin-1 Does Not Exhibit Continuous Processive Motion.

The data support that R2 mutation allows for premature kinesin-1 activity that opposes dynein during the retrograde axonal transport stage of neuronal infection, resulting in a tug-of-war phenotype. However, whereas dynein normally provides sustained retrograde transport of capsids for long-distance trafficking to the centrosome, the assimilated kinesin-1 is only required for trafficking the short distance to nuclei once the capsid arrives at the centrosome (12). Therefore, whether kinesin-1 in the assimilated state can drive long-distance transport was examined by treating the HSV-1 R2 mutant with ciliobrevin to favor release of dynein from microtubules by inhibiting its ATPase activity (22–24). Ciliobrevin was effective at inhibiting retrograde axonal transport of wild-type HSV-1 and this was reversible; however, ciliobrevin had no detectable effect on the tug-of-war motion of the HSV-1 R2 mutant (Fig. 3). Therefore, relieving opposition from the dynein motor was insufficient for the assimilated kinesin-1 to power sustained capsid trafficking anterogradely in axons.

Fig. 3.

Inhibiting dynein does not promote HSV-1 anterograde axonal transport. (A) Experiment timeline. (B) Frequency of capsid motion in axons: % time retrograde (light), % time anterograde (medium), and % time stopped (dark). No significant differences in the anterograde component of motion were observed. (C) Net displacement of capsids after 10 s. Positive values indicate movement toward neuronal soma (retrograde displacement). Error bars are SD; ** P<0.01.

pUL37 Localizes to the Centrosome but Does Not Impact Vimentin Distribution in Uninfected Cells.

To further interrogate pUL37 function, hTERT-RPE cells were produced that stably express GFP fused to either WT pUL37 or the R2 mutant. A pUL37-enriched spot was consistently observed in cells and was determined to be the centrosome by immunofluorescence (Fig. 4A). pUL37 localization at the centrosome occurred independently of R2 and kinesin-1 and did not notably alter centrosome integrity based on pericentrin fluorescence levels (Fig. 4B). To examine whether pUL37 was sufficient to suppress cell endogenous kinesin-1, the vimentin intermediate filament network, which is dependent on kinesin-1, was imaged by immunofluorescence microscopy (15). In KIF5B KO cells, vimentin filaments were distributed in a distinctive arrangement surrounding the nucleus. Neither WT pUL37 nor the R2 mutant impaired the integrity of the vimentin network, indicating that the viral protein is not sufficient for kinesin-1 inhibition in this context (Fig. 4C).

Fig. 4.

pUL37 intrinsically associates with the centrosome but is insufficient for kinesin-1 suppression. (A) hTERT-RPE cells were stained with anti-GFP, anti-pericentrin (centrosome), and DAPI (nucleus) to document protein localization in cells transduced with the indicated GFP constructs. Arrows indicate centrosome locations based on pericentrin fluorescence. Values overlaid on the merged images indicate the percentage of cells with enriched GFP emissions at the centrosome. (B) Quantitation of centrosomal GFP intensities (Top) and corresponding pericentrin fluorescence intensities (Bottom). Data were extracted from samples used to produce representative images in panel (A). Error bars are SD; ***P < 0.001. (C) hTERT-RPE cells were stained with anti-GFP, anti-vimentin, and DAPI (nucleus) to document kinesin-1 activity in cells. (Scale bar, 10 µm.)

Spatial and Temporal Regulation of Assimilated Kinesin-1 Drives Nuclear Delivery of HSV-1 in Neurons and Epithelial Cells.

Although the pUL37 R2 effector promoted dynein-based retrograde axonal transport, in the absence of assimilated kinesin-1 this effector function became dispensable during this stage of infection. Despite being competent for long-distance retrograde transport, a kinesin-less HSV-1 R2 mutant was expected to accumulate at the centrosome because the motor is necessary to redirect capsids arriving at the centrosome to the nucleus (12). This prediction was initially tested in explanted dorsal root ganglia by live-cell imaging. Whereas wild-type HSV-1 abundantly localized to the nuclear surface by 3 to 4 hpi, HSV-1 R2* was largely absent from the nucleus and the neuronal soma in general, consistent with the inability of the latter to sustain retrograde trafficking in axons (Fig. 5A). In contrast, capsids of kinesin-less forms of both wild-type and R2* HSV-1 had accumulated in the neural soma but were clustered adjacent to the nucleus.

Fig. 5.

HSV-1 R2 mutant capsids lacking packaged kinesin accumulate at the centrosome in neurons and epithelial cells. (A) Live imaging of dorsal root ganglion sensory neurons infected with either WT or R2* HSV-1 encoding pUL25/mCherry (capsid) and counterstained with Hoechst (nucleus). Images were captured 4 hpi. (Scale bar, 10 µm.) (B) Immunofluorescence microscopy of hTERT-RPE cells fixed 4 hpi following infection with either WT or R2* HSV-1 and stained with anti-VP5 (capsid), DAPI (nucleus), and anti-pericentrin (centrosome). (Scale bar, 10 µm.) Magnified regions centered on centrosomes are shown in the corresponding panels below. (C) Automated analysis of capsid distribution in fixed hTERT-RPE cells. Capsids were scored as either at the nuclear rim (dark), centrosome (medium), or cytoplasmic away from the nucleus and centrosome (light). ***P < 0.0001.

Because the R2 effector and assimilated kinesin-1 are both required for infection of epithelial cells as well as neurons, an immunofluorescent study was performed to examine the localization of incoming capsids relative to the centrosome and nucleus in RPE cells (12, 14). Results were consistent with the live-cell observations from the dorsal root ganglia and confirmed that kinesin-less forms of HSV-1 specifically accumulate at the centrosome (Fig. 5B). An automated analysis pipeline was used to determine the distribution of capsids relative to the nucleus and the centrosome. A capsid that had zero distance from the nuclear rim, but a positive distance from the centrosome, was categorized as nuclear, and a capsid that had zero distance from centrosome but positive distance from the nuclear rim was categorized as centrosomal. Quantitative analysis showed that most wild-type HSV-1 capsids were at the nuclear rim with very little detected at the centrosome, whereas R2* capsids were randomly dispersed across cells (Fig. 5C). These phenotypes were unchanged when the assimilated kinesin-1 was constitutively active, indicating that the autoinhibition mechanism intrinsic to kinesin-1 was not only dispensable during retrograde transport but was also dispensable for subsequent nuclear trafficking. Nevertheless, the assimilated kinesin-1 was an effector of centrosome-to-nucleus trafficking, as kinesin-less forms of either the wild-type or R2* HSV-1 were preferentially distributed at the centrosome relative to the nucleus. These results reinforce that R2 is a suppressor of the assimilated kinesin-1 that prevents premature kinesin-1 activity that would otherwise interfere with dynein-based capsid trafficking to the centrosome (Fig. 6). The R2* HSV-1 defect is overcome when the R2* HSV-1 was made kinesin-less, but at the cost of impairing subsequent centrosome-to-nucleus trafficking. We conclude that the R2 effector in the pUL37 tegument protein choreographs kinesin-1 activity to drive capsids first to centrosomes and then to nuclei. Trafficking to centrosomes is sustained over long distances making this mechanism uniquely compatible with both epithelial and neuronal invasion and is fundamental to the retrograde axonal transport that underlies the robust neuroinvasive property of HSV-1 seen in vivo.

Fig. 6.

Model of HSV-1 retrograde transport. (Top panel) WT HSV-1 capsid with assimilated KIF5. pUL37 R2 suppresses the KIF5 activity during retrograde transport. (Second panel) Kinesin-less WT HSV-1 capsid. Removing KIF5 allows capsid to transport retrogradely. (Third panel) R2* HSV-1 capsid with assimilated KIF5. pUL37 R2* fails to suppress the KIF5 activity, which prohibits continuous retrograde movement. (Fourth panel) Kinesin-less R2* HSV-1 capsid. R2 effector function is dispensable in the absence of KIF5 during the retrograde transport step.

Discussion

Mammalian viruses of the Alphaherpesvirinae subfamily, including HSV-1, establish life-long latent infections in the ganglia of the peripheral nervous system. Before this can happen, these viruses first replicate in exposed mucosal epithelial cells and then transmit to nerve endings in the peripheral tissues. As with infection of any cell type, entry of these viruses into nerves occurs by a fusion event between the viral envelope and the cell, resulting in the deposition of the capsid and surrounding tegument proteins into the cytosol. The capsid traverses the length of the nerve by engaging in microtubule-based retrograde axonal transport mediated by the dynein–dynactin motor complex: a process that requires several hours of sustained transport to complete in vivo (25). Alphaherpesviruses engage the host transport machinery directly while in the cytosol via the capsid-bound pUL36 tegument protein, as opposed to being transported in an endosome as is the case for other neuroinvasive viruses including poliovirus and rabies virus (11, 26–29). Once the capsid reaches the neural soma, resident in the peripheral ganglia, the capsid is routed from the centrosome to the nucleus where the genome is delivered. Although several unrelated viruses are capable of neuroinvasion, alphaherpesviruses are noted for their proficiency at this process in otherwise healthy tissue.

Results from this study demonstrate that HSV-1 invasion is dependent on the spatiotemporal regulation of opposing microtubule motors and identifies the pUL37 tegument protein as an effector of this regulation: specifically, the R2 conserved surface-exposed region on pUL37 (30, 31). While the importance of the R2 effector in neurons and other cells was previously recognized from findings that alphaherpesviruses mutated in R2 fail to sustain microtubule transport, the molecular mode of action of R2 was undefined (14, 32). Recent work from our lab identified the presence of an assimilated kinesin-1 motor packaged within the alphaherpesvirus tegument that is preconfigured to traffic the virus from centrosomes to nuclei after entering cells (12). These findings implied that before the assimilated kinesin-1 motor performs its function at the centrosome it transports there with the capsid, yet in a manner that does not interfere with dynein-based retrograde transport. A possible unifying explanation of these phenomena was that R2 suppresses the assimilated kinesin-1 motor prior to arrival at the centrosome.

The prediction that R2 was a suppressor of the assimilated kinesin-1 was addressed by producing virions in the absence of the microtubule motor. If the R2* transport defect was due to aberrant, premature, activity of the assimilated kinesin-1 motor then the R2* defect would be absent when virions were produced without kinesin-1. In support of this model, removing the packaged kinesin allowed the HSV-1 R2 mutant to transport retrogradely in axons indistinguishably from WT virus. In addition to advancing our understanding of pUL37, this finding provides insight into the HSV-1 invasive mechanism. Specifically, restoration of retrograde axonal transport did not rescue delivery to nuclei as kinesin-less HSV-1 R2* was defective for centrosome-to-nucleus trafficking. We infer that the two-step invasive process exhibited by HSV-1 requires suppression of kinesin-1 during dynein-based retrograde transport, followed by release of R2 suppression at the centrosome to trigger kinesin-1 activity and route capsids to the nucleus. In principle, suppression could be relieved by removal of pUL37 from capsids, which is consistent with some observations; however, we have not seen evidence of pUL37 loss (20, 33). Posttranslational modification of pUL37 is reported, including phosphorylation and ubiquitination, and may also contribute to this process (10, 34).

Two approaches were used to narrow down how pUL37 R2 exerts its suppressive activity upon kinesin-1. First, viruses were produced carrying kinesin-1 that had mutations designed to specifically prevent the motor from stably adopting its folded autoinhibited state (mutants “H, I” and “C, H, I”), or to remove the carboxy-terminal half of the protein altogether (truncated mutant aa1-561). Truncated mutants analogous to aa1-561 are commonly used as constitutively active mutants for in vitro assays (e.g., microtubule gliding assays). Unexpectedly, infectivity was intact when any of these mutant kinesins were assimilated by HSV-1. Therefore, HSV-1 can capture and package, and pUL37 can suppress, all three mutant forms of the motor. This was particularly remarkable in the case of the 561 truncation and implies that HSV-1 interactions with kinesin-1 are focused on its motor domain. In hindsight, this may not be surprising given that the virus uses kinesin-1 to serve its own specific purpose. Second, pUL37 was stably expressed in hTERT-RPE cells in the absence of other viral proteins. In this paradigm, pUL37 did not display evidence of inhibiting endogenous cellular kinesin-1. This was assessed by examining the vimentin intermediate filament network, which is dependent on kinesin-1 for maintenance; in the absence of kinesin-1 the vimentin network collapses around the nucleus (15). We hypothesize that pUL37 suppression may only function in the context of the viral particle with assimilated kinesin-1. However, pUL37 intrinsically localized to the centrosome, and future studies of pUL37-centrosome interactions will likely reveal how pUL37 suppression is relieved, which may include posttranslational modifications such as ubiquitination, phosphorylation, and proteolysis.

Based on these findings, we expected that inhibiting dynein would reverse the direction of capsid transport in axons to the anterograde direction when R2 was mutated. This did not prove to be the case. Although ciliobrevin inhibited the retrograde transport of WT capsids, it did not visibly alter the tug-of-war motion of R2* capsids. There are several possible explanations for this result, including that inhibiting dynein activity may impair kinesin activity (35) or that posttranslational modification of the assimilated kinesin-1 may occur at the centrosome to enhance motor processivity (36–40). Alternately, the virus may instead benefit from constraining kinesin-1 activity. In this model, the activation of the assimilated kinesin-1 at the centrosome would produce short runs away from the centrosome, while maintaining opposing dynein transport similar to the R2* phenotype observed in axons. Runs occurring on perinuclear microtubules would predispose capsid docking and capture at a nuclear pore; however, if a run occurred on a microtubule heading away from the nucleus, the constrained kinesin-1 activity would restrict the capsid from going astray to the cell periphery and instead returning to the centrosome by dynein where the process could repeat. This process could be further enhanced by perinuclear-enriched microtubule acetylation (41–44). Facilitated diffusion-to-capture is a possible solution to HSV-1 nuclear targeting that would not require additional levels of regulation.

The neuroinvasive property of the alphaherpesviruses is of interest as a potential target to combat infection but also as a resource that could be retooled to deliver material, such as gene therapies, into the nervous system. Regarding the former, R2* viruses including the HSV-1 R2* used in the present study show potential as live-attenuated vaccines that elicit responses in peripheral tissues but are incapable of invading the peripheral nervous system (14, 32, 45, 46). The current study identifies the R2 effector in the pUL37 capsid-bound tegument protein as a regulator of assimilated kinesin-1 that coordinates the trafficking of incoming capsids to nuclei in epithelial and neuronal cells.

Materials and Methods

Cells.

Vero (African green monkey kidney epithelial) cells were obtained from ATCC and grown in Dulbecco’s modified Eagle’s medium (DMEM; Invitrogen, 11965-118) supplemented with 10% bovine growth serum (BGS; Rocky Mountain Biologicals, FGR-BBT). hTERT-RPE (immortalized human retinal pigmented epithelial) cells were obtained from V. Gelfand (Northwestern University) and were grown in DMEM supplemented with 10% fetal bovine growth serum (FBS; Gemini, 100106) and 1 mM sodium pyruvate (Invitrogen, 11360070). BGS and FBS were reduced to 2% when performing infections. Cell lines were tested for Mycobacterium contamination using the PlasmoTest kit (InvivoGen) and authenticated by the source. No commonly misidentified cell lines were used in the study.

hTERT-RPEΔKIF5B cells were previously described (15). Derivatives of hTERT-RPEΔKIF5B cells that expressed recombinant forms of KIF5B were produced by Moloney Murine Leukemia Virus (MMLV) transduction. Three MMLV constructs were produced. The first encoded KIF5B carrying mutations in the hinge region (referred to herein as “H”) and isoleucine-alanine-lysine (IAK; referred to herein as “I”) motif near the C terminus. The second encoded the same mutation plus a third mutation in the coiled-coil 2 (CC2; referred to herein as “C”) region of the motor. Both constructs encoded a C-terminal translational fusion to mNeonGreen (mNG). The third encoded a truncated KIF5B that express only the first 561 amino acids (1 to 561; referred to herein as “561”). The KIF5B(H, I)-mNG and KIF5B(C, H, I)-mNG constructs were recently described (16). Both constructs were subcloned into the pQCXIN MMLV backbone (Takara Bio, 631514) by the InFusion procedure (Takara Bio) using PCR primers AATTGATCCGCGGCCGCCACCATGGCAGACCTG and ATTCCGGATCCGTTACTTGTACAGCTCGTCCATGCC (underlined sequences are complementary to the KIF5B coding sequence). The truncated KIF5B was subcloned using PCR primers AATTGATCCGCGGCCGCCACCATGGCAGACCTG and ATTCCGGATCCGTTAGTCTTTTAGTAAAGATGCCATCATC. The PCR product was recombined into pQCXIN cut with NotI and PacI. The resulting MMLV clones were saved as pGS7729, encoding KIF5B(H, I)-mNG, pGS7730 encoding KIF5B(C, H, I)-mNG, and pGS7731 encoding KIF5B(561). Retrovirus particles were packaged in GP239 cells transiently expressing the vesicular stomatitis virus glycoprotein (VSV-G). At 48 h posttransfection, the medium was collected and passed through a 0.2 µm filter (EMD Millipore, SLFG05010). Supernatants containing the transducing particles were frozen at −80 °C for long-term storage. hTERT-RPEΔKIF5B cells were freshly passaged to yield 30 to 40% confluence in six-well dishes the following day, and the media were removed and replaced with 1 mL of transducing supernatant for 6 h and then replaced with fresh DMEM supplemented with 10% FBS and 1 mM sodium pyruvate. Beginning 2 d later, cells were placed under selection with 2 mg/mL G418 (EMD Millipore, 345812) for 1 wk. Expression of the recombinant KIF5B proteins in the resulting polyclonal cell populations, hTERT-RPEΔKIF5B(H, I), hTERT-RPEΔKIF5B(C, H, I), and hTERT-RPEΔKIF5B(561) was verified by western blot.

pBecker3 expressing UL37-GFP or UL37[R2]-GFP was previously described (14). Derivatives of hTERT-RPE or hTERT-RPEΔKIF5B cells that expressed GFP, UL37-GFP, or UL37[R2]-GFP were produced by Moloney Murine Leukemia Virus (MMLV) transduction. Inserts from pEGFP-N1 were prepared by restriction enzyme digest using BglII and MfeI and ligated to pQXCIN MMLV backbone cut with BamHI and EcoRI. The resulting MMLV clones were saved as pGS7560, encoding GFP, pGS7561, encoding UL37-GFP, and pGS7562, encoding UL37[R2]-GFP. Cells were transduced by methods described above and verified expression by western blot.

Viruses.

All recombinants of herpes simplex virus type 1 (HSV-1) were derived from a self-excising bacterial artificial chromosome (BAC) infectious clone of strain F, pGS6000, that was derived from pYEbac102 (47, 48). pGS6807 (HSV-1 encoding a pUL25/mCherry capsid tag) was previously described (10, 49). pGS6298 (HSV-1 encoding pUL25/mCherry and R2 mutations in the UL37 gene) was previously described (14). A self-excising version of pGS6298 was produced for this study by two-step Red-mediated recombination in the GS1783 E.coli using pEP-CRE-in template and primers AGCTGGTTTAGTGAACCGTCAGATCCGCTAGCGGTCGCCACCATGCCCAAGAAGAAGAGGAAGGTGTC and CTATTGCTTTATTTGTAACCATTATAAGCTGCAATA AACAAGTTACTAATCGCCATCTTCCAGCAGG (underlined sequences are complementary to the pEP-CRE-in template) (47, 50). The resulting self-excising BAC clone, pGS7502, was verified to produce HSV-1 upon transfection into Vero cells that lacked the BAC vector sequence based on PCR analysis.

HSV-1 stocks were produced by electroporation (BTX ECM 630 set to 950 µF, 200 V, 0 Ω) of BAC clones into Vero cells and harvested when all cells exhibited full cytopathic effect (CPE). Each virus was serial passaged once on Vero cells (2 µL of harvested material was added to a confluent 10 cm plate of cells) to produce high-titer stocks. To produce HSV-1 stocks packaged with specific versions of KIF5B, or without KIF5B, the high-titer stocks were passaged an additional time through hTERT-RPE cells encoding the respective KIF5B constructs as follows. Three 15 cm dishes of hTERT-RPE cells were grown to 80 to 90% confluency and infected with HSV-1 at MOI of 10 for 48 h. The supernatants were collected into a 50 mL conical flask on ice and centrifuged for 20 min at 5,000 × g at 4 °C. The cleared supernatant was transferred to a 25 × 89 mm centrifugation tube (Seton, 7052), underlaid with a 20% (wt/vol) sucrose solution in PBS, and centrifuged at 13,000 rpm for 1 h in SW-32 rotor at 4 °C. The pellet was resuspended in 1 mL of cold PBS on ice, and 50 μL aliquots were stored at −80 °C.

Primary Neuronal Culture.

Dorsal root ganglia (DRG) from embryonic chicken (E8-E9) were cultured as previously described (14). Explanted ganglia were maintained in culture for 1 to 3 d on poly-DL-ornithine- and laminin-treated coverslips in 2 mL of DRG culture media [DMEM/F12 media (Invitrogen, 11039-012) supplemented with 0.08 g/mL bovine serum albumin fraction V powder (VWR, IC16006910), 0.4 mg/mL crystalline bovine pancreas insulin (Sigma-Aldrich, I6634), 0.4 μg/mL sodium selenite (VWR, S9133), 4 μg/mL chick transferrin (Cedarlane, CLF-811), and 5 ng/mL nerve growth factor (Sigma-Aldrich, N0513)].

Live-Imaging Fluorescence Microscopy.

DRG explants were infected with HSV-1 in 2 mL of DRG culture media with 5.0 × 10⁷ plaque-forming units (PFU). Time-lapse images were captured at 0.5 hpi using an inverted widefield Nikon Eclipse TE2000-U microscope fitted with 60×/1.4 NA objective and Evolve electron-multiplying charge-coupled device (Teledyne Photometrics). The microscope was housed in a 37 °C environmental box (In Vivo Scientific). Moving particles were detected by time-lapse fluorescence microscopy in the red-fluorescence channel at 10 frames per second (continuous 100 ms exposures) for 200 frames. Particle trajectories were measured by kymograph analysis using a multiline tracing tool with a width of 20 pixels and average background subtraction in the MetaMorph software package (Molecular Devices). The trajectories were analyzed to measure the type of motion (retrograde, anterograde, stop), the percent time in motion, and the total net displacement of individual particles.

For dynein inhibition experiments, DRG explants were cultured in 1 mL of DRG culture media in a μ-Slide open chamber (ibidi, 80286), and were infected by adding 5.0 × 10⁷ PFU of HSV-1 to the culture medium. Time-lapse images were captured at 0.5 hpi with a Nikon Ti inverted microscope fitted with a 100×/1.45 NA objective and coupled to a CSU-W1 confocal head (Yokogawa Electric Corporation) and Prime 95B camera (Teledyne Photometrics). Prior to addition of drug, viral particles were tracked in axons by time-lapse fluorescence microscopy in the red-fluorescence channel at 10 frames/s (continuous 100 ms exposures) for 100 frames. Then, 4 µL of 5 mM ciliobrevin D (EMD Millipore, 250401) in DMSO was added and gently mixed into the media, and the cultures were incubated for 5 min. The axons were imaged a second time as above. For washout, 800 µL of media was gently removed from the chamber and replaced with 800 µL of fresh DRG culture media. The washing procedure was repeated five times, and the cultures were incubated for 15 min. The axons were imaged a third time as above.

Capsid distributions in neuronal soma were imaged by confocal microscopy (see above) at 3 to 4 hpi. For these experiments, DRG explants were infected with 1.0 × 10⁸ PFU/mL of HSV-1 and imaged between 3 and 4 hpi using the confocal imaging setup described above. Nuclei were stained with 5 µL of 10 mg/mL Hoechst (Sigma-Aldrich, SA14530) 30 min before imaging, and a Z plane was chosen that transected the nucleus equatorially.

Intracellular Capsid Localization.

hTERT-RPE cells were plated on coverslips, incubated for 1 d to obtain 80% confluency, and infected at MOI 100 with 100 µM cycloheximide to prevent viral gene expression and de novo capsid assembly. Cells were fixed at 4 hpi with ice-cold methanol for 7 min at −20 °C, washed three times with ice-cold PBS, and permeabilized and blocked with PBS supplemented with 10% FBS and 0.25% saponin for 40 min at room temperature. Primary antibody was diluted in antibody solution (PBS supplemented with 10% FBS and 0.025% saponin) and gently rocked overnight at 4 °C. Primary antibodies used were rabbit anti-pericentrin (1:200) (Bethyl Laboratories, IHC-00264) and mouse anti-HSV-1 VP5 (1:400, Virusys, HA018). Primary antibodies were removed, and the coverslips were washed three times with antibody solution. After the final wash, secondary antibodies were added, and the coverslips were gently rocked for 1 h at room temperature. Secondary antibodies used were anti-mouse Alexa-488 and anti-rabbit Alexa-568 (both diluted 1:400 in antibody solution). Coverslips were washed three times with antibody solution and once in deionized H₂O and mounted on slides using ProLong Gold Antifade Mountant with DAPI (Fisher Scientific, P36931) and sealed with nail polish. Images were captured (100 ms exposures) using the confocal microscope described above. Capsids were scored for localization at either the centrosome or nucleus using an automated analysis pipeline as previously described (12).

Immunofluorescence and Quantification of Relative Protein Intensities.

hTERT-RPE cells were plated on coverslips and incubated for 1 d to obtain 80% confluency. Cells were fixed and permeabilized using the method described above, and the following primary antibodies were used: mouse anti-GFP (1:200) (Santa Cruz Biotechnology, SC-9996), rabbit anti-pericentrin (1:200) (Bethyl Laboratories, IHC-00264), and chicken anti-vimentin (1:1,500) (BioLegend, 919101). Secondary antibodies used were anti-mouse Alexa-488, anti-rabbit Alexa-568, and anti-chicken Alexa-568 (all diluted 1:400 in antibody solution). Coverslips were washed three times with antibody solution and once in deionized H₂O and mounted on slides using ProLong Gold Antifade Mountant with DAPI (Fisher Scientific, P36931) and sealed with nail polish. Images were captured with 100 ms exposures using the confocal microscope described above. Intensity measurements of anti-pericentrin and anti-GFP emissions at the centrosome (based on pericentrin label) were obtained using NIS Elements software. A linear region of interest was drawn through the center of centrosome based on anti-pericentrin fluorescence, and a line scan of pixel intensities was plotted. Centrosomal fluorescence was quantitated by measuring the corresponding area under the line scan. The percentage of cells with pUL37-GFP at the centrosome was calculated based on GFP integrated emission intensities relative to the mean background GFP intensity in cells. Cells with >ninefold enrichment of GFP at the centrosome were scored as positive.

Western Blot.

Whole-cell lysates were produced from confluent hTERT-RPE cells in 10 cm dishes that were washed with 10 mL ice-cold PBS and lysed in 400 µL of 2× final sample buffer (60 mM Tris-HCl, pH 6.8, 2% SDS, 10% glycerol, and 0.01% bromophenol blue) supplemented with 5% β-mercaptoethanol. The lysed cells were scraped into suspension and passed through three needles (18.5, 21.5, and 27.5 gauges) seven times each, and the lysates were heated at 95 °C for 5 min. The lysates were separated on 4 to 20% precast Mini Protean TGX gels (Bio-Rad) and transferred to PVDF Hybond-P membranes (GE Healthcare). Membranes were blocked with 5% nonfat milk in PBS for 1 h at room temperature and then incubated with primary antibody at 4 °C overnight. The primary antibodies are anti-KIF5B (Abcam, ab167429) diluted at 1:1,000 and anti-tubulin (Abcam, ab7291) diluted at 1:10,000. The primary antibodies were removed, and the membranes were washed three times with 1% nonfat milk in PBS and then incubated with secondary antibody for 1 h at room temperature. Donkey anti-rabbit 680 (Licor, 926-68073) and donkey anti-mouse 680 (Licor, 926-68072) were used in 1:10,000 dilutions. All antibodies were diluted in 1% nonfat milk in PBS. Blots were imaged with a LI-COR Odyssey FC imaging system at 800 nm for 2 min.

Statistical Analysis.

Statistical significance was determined by one-way ANOVA with a post hoc Tukey test for comparison of multiple experimental groups. All independent experiments are provided as biological replicates.

Supplementary Material

Appendix 01 (PDF)

Movie S1.

Retrograde axonal transport of HSV-1 in primary dorsal root ganglia sensory neurons. Representative time-lapse fluorescent microscopy recordings that contributed to the data sets in Figure 2 are presented. Synopses of the models from Figure 6 are included at right.

Data, Materials, and Software Availability

All study data are included in the article and/or supporting information.