A VP26-mNeonGreen Capsid Fusion HSV-2 Mutant Reactivates from Viral Latency in the Guinea Pig Genital Model with Normal Kinetics

Fluorescent herpes simplex viruses (HSV) are invaluable tools for localizing virus in cells, permitting visualization of capsid trafficking and enhancing neuroanatomical research. Fluorescent viruses can also be used to study virus kinetics and reactivation in vivo. Such studies would be facilitated by fluorescent herpes simplex virus recombinants that exhibit wild-type kinetics of replication and reactivation and that are genetically stable. We engineered an HSV-2 strain expressing the fluorescent mNeonGreen protein as a fusion with the VP26 capsid protein. This virus has normal replication and in vivo recurrence phenotypes, providing an essential improved tool for further study of HSV-2 infection.

Efforts to optimize fluorescent viruses have focused on HSV-1 in which fluorescent proteins have been inserted into the N-terminal region of VP26. These viruses are infectious, but nonetheless have shown important differences from wild-type virus [2][3][4]6,7,[10][11][12]. Dimerization of GFP and YFP, leading to nuclear aggregates, has been a persistent challenge [2,3,7,10]. Previous work examined the utility of replacing dimeric fluorescent proteins with monomeric ones, which reduced but did not eliminate aggregates, and was still associated with replication defects that manifested as diminished plaque size [4,6,10,12]. Some of these previous constructs also reverted to a non-fluorescent virus, likely due to competition from non-fluorescent mutants that arise during replication in culture [10,11,13].
Fluorescent viruses have the potential for aiding in study of HSV latency and reactivation, potentially making it easier to identify loci of viral reactivation or of incipient recurrent lesions. Studies of viral reactivation from neuronal latency are limited because no fluorescent HSV mutant has been shown to reactivate in vivo with wild-type kinetics. A previously described HSV-2 that expresses a VP26-eGFP fusion protein is genetically stable and has been useful for in vitro studies, but does not reactivate with normal kinetics in vivo [14][15][16].
Construction of a fluorescent HSV-2 that replicates and reactivates with normal kinetics in vivo would enable use of the female guinea pig genital model, in which virus reactivates spontaneously to cause recurrent lesions. We based an HSV-2 small capsid protein (VP26) fusion on the most successful HSV-1 and HSV-2 constructs, which employ N-terminal VP26 fusions with monomeric fluorescent proteins. Small deletions in the N-terminus of HSV-1 VP26 appeared necessary for optimal fusion, so we engineered a corresponding deletion into HSV-2. We also used a recently described monomeric fluorescent protein derived from Branchiostoma lanceolatum that is three times brighter than eGFP, mNeonGreen, which has not been previously used in HSV capsid fusions [17]. Using the guinea pig as a well-established spontaneous reactivation model and a novel, monomeric fluorescing virus with an optimized design, we have tested the value of a newly generated tool and a novel, bright fluorescent protein, which has potential use in other viruses or protein fusions. Ideally, a capsid-modified strain would have a similar morphology to wild-type strains, replicate with similar kinetics to the parental strain, fluoresce brilliantly enough to visualize with available microscopy, create similar cytopathic effects in vitro, cause similar pathology in vivo, and reactivate spontaneously from latency.
Western Blot. Monolayers of Vero cells were infected with HSV-2 strain 333 or Nedel at a multiplicity of infection (MOI) of 10. Protein was extracted with Laemmli Buffer and separated using sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) using Nu-Page 4-12% Bis-Tris gels (Invitrogen, Carlsbad, CA, USA) and transferred to nitrocellulose membranes (iBlot™ Transfer Stack, nitrocellulose, mini, Thermo Fisher Scientific, Waltham, MA, USA). Membranes were incubated with a rabbit peptide antibody to the C-terminus of VP26 (95RRTYSPFVVREPSTPGTP112), a generous gift of Prashant Desai, at a 1:500 dilution for 1 h at RT, horseradish peroxidase-linked secondary anti-rabbit antibodies (GE Healthcare, Cincinnati, OH, USA) were used at a 1:2000 dilution for 1 h at RT and detected by chemiluminescence (ECL) reagent (GE Healthcare). Magic marker protein standard (Invitrogen) was loaded into the first lane.
Growth Curve and Plaque size comparison. Growth of HSV-2 in Vero cells was characterized as previously described [18]. Briefly, Vero cells were infected with HSV-2 Strain 333 or Nedel at a MOI of 0.01, and total virus was collected from cells at 0, 6, 12, and 24 h post-infection (hpi). Strain 333 or Nedel was quantified by standard plaque assay on Vero cells. The area of plaques was determined by NIS Element software (v4.1) on a Nikon Eclipse Ti-E fluorescent microscope (Nikon, Tokyo, Japan) and the 6-point oval tool.
Primary adult neuron infection. Dorsal root ganglia were removed from 6 week old Swiss Webster mice, dissociated enzymatically and mechanically, and plated on Matrigel-coated Lab-Tek II chamber slides (Thermo Scientific, Waltham, MA, USA) as previously described [14]. Neurons were maintained in Neurobasal A media supplemented with B27, penicillin/streptomycin, Glutamax, neurotrophic factors, and mitotic inhibitors (Life Technologies, Carlsbad, CA, USA). Three days post-plating, neurons were inoculated with 30 MOI Nedel. For productive infections, neurons were fixed and immunolabeled for A5 (Fe-A5 supernatant, DSHB, Iowa City, IA, USA) or IB4 (isolectin B5-rho, Vector Laboratories, Burlingame, CA, USA) 9 h post-infection. For reactivation studies, infected neurons were maintained in 300 µM acyclovir to establish experimental latency for 7 days, followed by deprivation of neurotrophic factors to induce reactivation, and neurons were fixed and immunolabeled 9 h later. Neurons were counted and expressed as percentage of total or subpopulations of neurons productively infected. Reactivations were similarly counted, but subpopulations were represented as percentage of total reactivations that occurred in A5+ or IB4+ neurons. Neurons were imaged on inverted fluorescent Olympus IX73 microscope (Olympus Corporation, Tokyo, Japan).

In Vivo
Animals. Female Hartley Guinea Pigs (15 initially total per group, 150-200 g, Charles River Breeding Laboratories, Wilmington, MA, USA) were intravaginally inoculated with 1 × 10 6 pfu HSV-2 (Strain 333 or Nedel) as previously described [19]. Animals that failed to produce any lesions or developed end-point criteria prior to day 14 were removed from analysis, resulting in 7 Strain 333-infected animals and 13 Nedel-infected animals. All animal experiments were performed under protocol number 1997-08 approved by the Center for Biologics Evaluation and Research Independent Animal Care and Use Committee (29 June 2015, CBER IACUC).
Replication In Vivo. Animals were evaluated daily for evidence of genital skin disease, urinary retention, hind limb paresis, paralysis, and death. Primary genital skin disease (Days 1-14 pi) was quantified by a lesion score system on a scale from 0 to 4 as follows: 0 for no disease, 1 for redness/swelling, 2 for one or two lesions, 3 for three to five lesions, and 4 for six or more lesions, coalescence of lesions or 3-5 lesions with neurologic symptoms [19]. Replication Ex Vivo. Dorsal Root Ganglia were harvested and cultured at 36 days post-infection from 5 guinea pigs intravaginally infected with Strain 333 or Nedel, as previously described [14]. Briefly, ganglia were digested in papain, collagenase, and dispase (Worthington Biochemical Corporation, Lakewood, NJ, USA) before mechanically triturating and plating on Matrigel-coated 8-well Lab-Tek II chamber slides (Thermo Scientific), followed by mechanical trituration with a pipette. Cultures were then fixed at 60 or 72 h post-plating for 5 min in 4% paraformaldehyde, gently rinsed in phosphate buffer saline (PBS) and immunolabelled as described below. Immunolabelled neuronal cultures were evaluated by fluorescence microscopy using a Nikon Eclipse Ti-E or a Zeiss LSM 710 upright confocal laser scanning microscope.
Cryosectioning. Animals were cardiac-perfused with PBS followed by cardiac-perfusion with 4% paraformaldehyde. Tissues were dissected, rinsed in PBS and sucrose-protected overnight. Sections were embedded in OCT (TissueTek ® , Sakura®Finetek, VWR catalog number 25608-9300, Radnor, PA, USA) and 10 µm sections were made with a cryostat (Leica, Wetzlar, Germany). Sections were permeabilized, blocked, and immunolabelled Immunolabelling and microscopy of plaques, explanted ganglia, and cryosections. Antibodies Post-permeabilization and blocking, samples were labelled with the primary antibody for 24 h at 4 • C, secondary antibody for 24 h at 4 • C, labelled with Nissl stain for 2 h at room temperature before DAPI staining. Cryosections also were treated with TrueBlack™ for 30 s in order to remove autofluorescence from lipofuscin that might otherwise potentially obscure our observations [20,21]. Nikon Eclipse Ti-E was used to take phase contrast and fluorescent images. A Zeiss LSM 710 upright confocal laser scanning microscope was used to provide more detailed fluorescent imaging. Uninfected and strain 333-infected ganglia were used to establish the maximum exposure times and gain that could be used before the emergence of autofluorescence. Single-stained explanted ganglia and cryosections were used to ensure that fluorescence was not bleeding over into other channels. Images of explanted ganglia were taken at 12, 24, and 48 h time points to establish a timeline of the emergence of fluorescence.

In Vitro Characterization
Our goal was to develop and characterize a fluorescent HSV-2 variant that could reactivate in vivo with wild type HSV-2 kinetics and maintain its fluorescence into the reactivation phase. One of the most successful previous approaches with HSV-1 and HSV-2 was chosen, but the fluorescent protein was replaced by the three-fold brighter monomeric mNeonGreen fluorescent protein as an N-terminal fusion to UL35 (VP26) with the first 21 base pairs of UL35 deleted (corresponding to the first 7 codons), under control of the native UL35 promoter. We constructed this HSV-2 mNeonGreen VP26 fusion by homologous recombination, and designated the recombinant as "Nedel". In order to confirm that the fluorescent protein had been fused to the minor capsid protein, we performed a Western blot using an antibody to VP26 ( Figure 1). The wild-type capsid protein is 12 kDa; when fused to the 24 kDa mNeonGreen, the fusion mutant VP26 migrates consistent with the expected size of 36 kDa.    To ensure that viral infection was not limited to Vero cells, we assessed the ability of Nedel to infect cultured primary adult murine DRG neurons. Nedel productively infected a similar percentage of total cultured DRG neurons compared to HSV-2 Strain 333 (39.5% and 40.9%, respectively), with preferential productive infection observed in A5+ neurons (Figure 3a), similar to previous reports [16]. Reactivation of Nedel also occurred comparable to Strain 333; approximately 70% of total reactivations occurred in IB4+ neurons, for both Strain 333 and Nedel, as opposed to less than 1% of total reactivations detected in A5+ neurons (Figure 3b). mNeonGreen expressed by Nedel during reactivation was clearly visualized within the immunolabeled primary adult cultured neurons (Figure 3c).  To ensure that viral infection was not limited to Vero cells, we assessed the ability of Nedel to infect cultured primary adult murine DRG neurons. Nedel productively infected a similar percentage of total cultured DRG neurons compared to HSV-2 Strain 333 (39.5% and 40.9%, respectively), with preferential productive infection observed in A5+ neurons (Figure 3a), similar to previous reports [16]. Reactivation of Nedel also occurred comparable to Strain 333; approximately 70% of total reactivations occurred in IB4+ neurons, for both Strain 333 and Nedel, as opposed to less than 1% of total reactivations detected in A5+ neurons (Figure 3b). mNeonGreen expressed by Nedel during reactivation was clearly visualized within the immunolabeled primary adult cultured neurons (Figure 3c). To ensure that viral infection was not limited to Vero cells, we assessed the ability of Nedel to infect cultured primary adult murine DRG neurons. Nedel productively infected a similar percentage of total cultured DRG neurons compared to HSV-2 Strain 333 (39.5% and 40.9%, respectively), with preferential productive infection observed in A5+ neurons (Figure 3a), similar to previous reports [16]. Reactivation of Nedel also occurred comparable to Strain 333; approximately 70% of total reactivations occurred in IB4+ neurons, for both Strain 333 and Nedel, as opposed to less than 1% of total reactivations detected in A5+ neurons (Figure 3b). mNeonGreen expressed by Nedel during reactivation was clearly visualized within the immunolabeled primary adult cultured neurons (Figure 3c).  In standard growth curves, to study in vitro replication of Nedel, monolayers of Vero cells were infected at a multiplicity of 0.01 PFU/cell with Strain 333 or Nedel, and virus was quantified by plaque assay from cultures harvested at indicated times post-infection (Figure 4a). Replication of Nedel was similar to that of Strain 333 over the course of 24 h (Figure 4a, Mann-Whitney U test, p = 0.90). By transmission electron microscopy (after epoxy-resin embedding) there was no significant difference between Nedel and Strain 333 in capsid size (~100-125 nm diameter), or virion morphology (Figure 4b).
We did not observe proteinaceous aggregates or capsid aggregation, as has been reported for other fluorescent HSV [10]. In standard growth curves, to study in vitro replication of Nedel, monolayers of Vero cells were infected at a multiplicity of 0.01 PFU/cell with Strain 333 or Nedel, and virus was quantified by plaque assay from cultures harvested at indicated times post-infection (Figure 4a). Replication of Nedel was similar to that of Strain 333 over the course of 24 h (Figure 4a, Mann-Whitney U test, p = 0.90). By transmission electron microscopy (after epoxy-resin embedding) there was no significant difference between Nedel and Strain 333 in capsid size (~100-125 nm diameter), or virion morphology ( Figure  4b). We did not observe proteinaceous aggregates or capsid aggregation, as has been reported for other fluorescent HSV [10].

In Vivo Characterization
The ability of Nedel to infect adult female guinea pigs via the genital tract was also assessed.

In Vivo Characterization
The ability of Nedel to infect adult female guinea pigs via the genital tract was also assessed.

Ex Vivo Characterization
Viral reactivation was also assessed ex vivo from infected animals euthanized on day 36 post-infection. Sacral dorsal root ganglia (DRG) were dissected, enzymatically digested and plated on Matrigel-coated chamber slides. In this model, the axotomy alone stimulates ex vivo reactivation [24,25]. Slides were then observed for fluorescence. Fluorescent neurons were first observed in cultures at 60 h post-plating, with maximum fluorescence observed at 72 h post-plating. Cultures were fixed and stained to label neurons and HSV-2 antigen (Figure 6a,b). The neurons from Nedel-infected animals that exhibited green fluorescence also stained positively for non-capsid HSV-2 antigen, further supporting the idea that Nedel is capable of establishing latency in these neurons and of induced reactivation ex vivo after axotomy. Higher fluorescence intensity was observed in small neurons with altered morphology, suggestive of cytopathic effect from a recent reactivation. Double instead of triple staining of cultures (Pan-Neuronal and DAPI) allowed the analysis of more neurons per field of view (Figure 6c), also showing co-localization of green fluorescence and explanted neurons from Nedel-infected animals. Using confocal microscopy, we also observed similar co-localization of green fluorescence (representing the Nedel capsid protein) and red fluorescence (representing HSV glycoproteins) within neurons (Figure 6d).

Ex Vivo Characterization
Viral reactivation was also assessed ex vivo from infected animals euthanized on day 36 postinfection. Sacral dorsal root ganglia (DRG) were dissected, enzymatically digested and plated on Matrigel-coated chamber slides. In this model, the axotomy alone stimulates ex vivo reactivation [24,25]. Slides were then observed for fluorescence. Fluorescent neurons were first observed in cultures at ~60 h post-plating, with maximum fluorescence observed at 72 h post-plating. Cultures were fixed and stained to label neurons and HSV-2 antigen (Figure 6a,b). The neurons from Nedelinfected animals that exhibited green fluorescence also stained positively for non-capsid HSV-2 antigen, further supporting the idea that Nedel is capable of establishing latency in these neurons and of induced reactivation ex vivo after axotomy. Higher fluorescence intensity was observed in small neurons with altered morphology, suggestive of cytopathic effect from a recent reactivation. Double instead of triple staining of cultures (Pan-Neuronal and DAPI) allowed the analysis of more neurons per field of view (Figure 6c), also showing co-localization of green fluorescence and explanted neurons from Nedel-infected animals. Using confocal microscopy, we also observed similar co-localization of green fluorescence (representing the Nedel capsid protein) and red fluorescence (representing HSV glycoproteins) within neurons (Figure 6d).

In Vivo Cryosections
To study the ability of the mutant virus Nedel to spontaneously reactivate in vivo from neurons, animals were cardiac-perfused and cryosections of the sacral DRG were made. Neurons were labelled with Nissl and sections with green fluorescence were confirmed to be HSV-2 positive by immunofluorescent labelling (Figure 7a-e). Because the control Strain 333 sections (e.g., as in Figure 7a) were selected at random for staining (while Nedel sections were selected by fluorescence), only one neuron (on a single slide out of twenty stained) was identified that stained positively with the HSV-2 pAb. The Nedel in vivo-infected neurons exhibited co-localization of intense fluorescence (green, upper left quadrant) and HSV-2 pAb (red, upper right quadrant) within the nuclei and cytoplasm of neurons (merged, bottom right quadrant). This pattern was consistently observed in sections from Nedel-infected animals that exhibited fluorescence from within neurons (Figure 7c-e). Thus, fluorescence in Nedel-infected ganglia indicated viral reactivation in these neurons.

In Vivo Cryosections
To study the ability of the mutant virus Nedel to spontaneously reactivate in vivo from neurons, animals were cardiac-perfused and cryosections of the sacral DRG were made. Neurons were labelled with Nissl and sections with green fluorescence were confirmed to be HSV-2 positive by immunofluorescent labelling (Figure 7a-e). Because the control Strain 333 sections (e.g., as in Figure  7a) were selected at random for staining (while Nedel sections were selected by fluorescence), only one neuron (on a single slide out of twenty stained) was identified that stained positively with the HSV-2 pAb. The Nedel in vivo-infected neurons exhibited co-localization of intense fluorescence (green, upper left quadrant) and HSV-2 pAb (red, upper right quadrant) within the nuclei and cytoplasm of neurons (merged, bottom right quadrant). This pattern was consistently observed in sections from Nedel-infected animals that exhibited fluorescence from within neurons (Figure 7c-e). Thus, fluorescence in Nedel-infected ganglia indicated viral reactivation in these neurons.

Discussion
We describe the construction and evaluation of a fluorescent HSV-2 variant. Unlike previously described fluorescent HSV, this virus is genetically stable, has normal morphology without aggregation, replicates in vitro and in vivo with normal kinetics, and has a wild-type spontaneous recurrence phenotype in vivo.
We attribute the successful construction of this virus to properties of the mNeonGreen fluorescent protein, which is much brighter than other fluorescent proteins and has no known tendency to aggregate. Although we did not formally analyze the number of incorporated VP26 molecules in Nedel, this likely allows for more normal VP26 interactions within the rest of the capsid and prevents aggregation of capsid proteins before assembly and of the capsids themselves before envelopment. Monomeric NeonGreen emits a yellow-green fluorescence (λem,max = 517 nm and λex,max = 506 nm), which may limit its use in highly autofluorescent tissues like the skin.
While Nedel exhibited wild-type kinetics of acute and recurrent infection of guinea pigs, we noted reduced neurovirulence during the acute infection when an inoculum of 10 6 pfu was used. We did not construct a rescuant for Nedel since the recombinant virus otherwise behaved as wild type HSV-2 Strain 333, so we do not know if this reduced neurovirulence is due to the VP26 mutation, or another unintentionally introduced mutation. The reduced neurovirulence observed with Nedel is in the range observed with other HSV-2 strains, and did not impact establishment of latency or viral reactivation.
We examined the capacity of Nedel to reactivate in four different ways. In vitro, Nedel reactivated preferentially in primary adult cultured sensory neurons at a rate similar to wild-type virus. In vivo spontaneous reactivation evaluated the ability of reactivations to give rise to external recurrent lesions. Ex vivo reactivation after axotomy assessed the ability of the virus to reactivate from within individual neurons after a strong stimulus, and typically led to altered morphology of neurons harboring reactivated virus. Identification of green fluorescence in cryosections indicated the ability of the virus to spontaneously reactivate from these neurons in vivo. The ability to scan unstained sections for green fluorescence greatly simplified the identification of these loci of viral reactivation. We did not identify morphological changes in the neurons harboring reactivating virus in vivo, which is consistent with a previous report [24].
Nedel is the first described HSV-2 fluorescent strain with a normal reactivation phenotype. We intend to use this virus to study early reactivation events in vivo in more detail than has previously been possible. Inclusion of the mNeonGreen protein may also enhance the utility of fluorescent fusions of other herpesviruses and allow observation of infrequent or minimally produced proteins.
Author Contributions: Philip Krause, Shuang Tang, and Julianna Pieknik conceived and designed the experiments; Julianna Pieknik and Andrea S. Bertke performed the ganglia explantation jointly; Andrea S. Bertke performed the murine neuron infection assay, and Julianna Pieknik performed the virus construction, Vero cell in vitro assays, dissections, cryosectioning, immunohistochemical staining, analyzed the data and wrote the paper.