Conceived and designed the experiments: JIL GEP GAS. Performed the experiments: JIL PJS SBB ML. Analyzed the data: JIL PJS GEP. Contributed reagents/materials/analysis tools: JIL GEP. Wrote the paper: JIL GEP GAS.
The authors have declared that no competing interests exist.
The neuroinvasive property of several alpha-herpesviruses underlies an uncommon infectious process that includes the establishment of life-long latent infections in sensory neurons of the peripheral nervous system. Several herpesvirus proteins are required for replication and dissemination within the nervous system, indicating that exploiting the nervous system as a niche for productive infection requires a specialized set of functions encoded by the virus. Whether initial entry into the nervous system from peripheral tissues also requires specialized viral functions is not known. Here we show that a conserved deubiquitinase domain embedded within a pseudorabies virus structural protein, pUL36, is essential for initial neural invasion, but is subsequently dispensable for transmission within and between neurons of the mammalian nervous system. These findings indicate that the deubiquitinase contributes to neurovirulence by participating in a previously unrecognized initial step in neuroinvasion.
Subsets of herpesviruses, such as herpes simplex virus and pseudorabies virus, are neuroinvasive pathogens. Upon infection, these viruses efficiently target peripheral nervous system tissue and establish a life-long infection for which there is no cure. Very few pathogens are known that invade the nervous system proficiently, and the mechanism by which herpesviruses achieve neuroinvasion is largely unknown. In this study, we demonstrate that a viral protease plays a critical and specific role allowing the virus to cross the threshold of the nervous system, but is dispensable for subsequent replication and encephalitic spread within the brain.
Neuroinvasive herpesviruses comprise a group of pathogens with individual members infecting different mammalian hosts. Four viruses of this group infect humans: herpes simplex virus type 1 (HSV-1), herpes simplex virus type 2 (HSV-2), varicella zoster virus (VZV) and simian herpes B virus (SHBV). The latter is associated with rare life-threatening zoonotic infections. Diseases associated with these viruses range from minor recurrent lesions to shingles, keratitis and encephalitis. All of these pathologies can occur as a consequence of viral dissemination into the nervous system. There are additional neuroinvasive herpesviruses that infect mammals other than humans. Among these, pseudorabies virus (PRV) has a broad host range and is commonly associated with severe encephalitic infections, making it a useful model for studies of neuroinvasion and pathogenesis [reviewed in 1]. In addition, the propensity of PRV to transmit neuron-to-neuron during encephalitic spread has led to its use as a self-amplifying tracer for mapping of the vertebrate neural circuitry [reviewed in 2]. PRV is closely related to VZV (both are members of the varicellovirus subgroup of the alpha-herpesvirus family), and shares a common structure, similar genetic composition and a related infectious cycle with all neuroinvasive herpesviruses
To date, mutant viruses displaying defects in neurotropism fall into two categories: neuron-specific replication mutants and axon transport mutants. Mutants of HSV-1 that fail to replicate in neurons include viruses lacking thymidine kinase activity or ICP34.5
Recently, an isolate of PRV mutated in a conserved deubiquitinase (DUB) domain of a viral structural protein, the pUL36 tegument protein, was reported to infect the nervous system with delayed kinetics following intranasal inoculation into mice
A conserved cysteine residue in the pUL36 tegument protein that is critical for the proteolytic activity responsible for deubiquitination
(A) Illustration of the PRV-Becker genome with region encoding the UL36 gene (which encodes the pUL36 protein) and neighboring UL37 gene expanded. The position of the codon change resulting in the C26A point mutation is indicated. Promoters are represented as black triangles. IR, internal repeat. TR, terminal repeat. (B) Purified extracellular virions encoding mRFP1-capsids and either a wild-type UL36 allele (WT; PRV-GS847) or UL36 allele encoding the amino-terminal point mutation (C26A; PRV-GS1652) were examined by Western blot analysis for the incorporation of the UL37 protein. The major capsid protein, VP5, was used as a loading control. (C) Comparison of plaque diameters resulting from infection of Vero cells with the virus encoding wild-type UL36 (WT) or mutated UL36 (C26A) allele. Each virus also encodes the mRFP1-VP26 fusion (red capsids). Error bars are standard error of the means (SEM). (D) Propagation kinetics of viruses encoding wild-type UL36 (WT), mutated UL36 (C26A), or a wild-type revertant of the C26A allele (Rev). All viruses also encode the mRFP1-VP26 fusion (red capsids). Infectious virions were detected as plaque-forming units harvested from either the cells (cells) or tissue culture supernatant (sups).
PRV is frequently used as a trans-synaptic neurotracer to map circuits within the mammalian nervous system [reviewed in 2]. In particular, neural circuits between the eye and brain have been extensively studied in rodents infected with PRV, which we used as the basis for our current studies of the C26A mutant virus
WT | C26A | Rev (C26) | 1∶1 |
PRV-152 | |
Anterograde (SC) |
|||||
Intravitreal |
1/5 (20%) | 4/15 (27%) | n/d | n/d | n/d |
Retrograde (EW) |
|||||
Intravitreal |
5/5 (100%) | 0/15 (0%) |
n/d | n/d | n/d |
Anterior Chamber |
3/3 (100%) | 0/4 (0%) | 3/3 (100%) | 3/3 (100%) |
4/4 (100%) |
Retrograde (RGC) |
|||||
SCN |
n/d | 2/2 (100%) | n/d | n/d | n/d |
Retrograde (CNS) |
|||||
Eyelid |
n/d | 1/11 (9%) | n/d | n/d | 10/11 (91%) |
site imaged.
site injected.
one animal showed fluorescence signal in the oculomotor nucleus.
1∶1 mixture of PRV-152 (Bartha; green fluorescence) and PRV-GS1652 (C26A; red fluorescence).
all 3 animals emitted red and green fluorescence in the EW indicative of PRV-152 and PRV-GS1652 co-infection.
SC superior colliculus.
EW Edinger-Westphal nucleus.
RGC retinal ganglion cells.
CNS central nervous system.
WT virus encoding wild-type pUL36.
C26A virus encoding C26A mutant isoform of pUL36.
Rev revertant of C26A virus (encodes wild-type pUL36).
n/d no data.
Because anterograde spread of the C26A mutant of PRV was unexpected, and did not account for its severe attenuation, the brains of these animals were also examined for evidence of infection via retrograde autonomic circuits. Intravitreal injection into the eye can expose the iris and ciliary body to viral inoculum, which receive autonomic innervation. The time course of infection and retrograde spread of PRV in these circuits is somewhat variable due to the gelatinous matrix of the vitreous body (compared to the aqueous humor of the anterior chamber, see below). Nevertheless, in these experiments wild-type virus was reproducibly observed to have infected the Edinger-Westphal nucleus (EW) of the midbrain, preganglionic parasympathetic neurons that project to neurons in the ciliary ganglion via the oculomotor nerve, which in turn innervates the smooth muscles in the iris and ciliary body in the eye to mediate pupillary constriction and lens accommodation, respectively. In contrast, capsid fluorescence was never detected in the EW following C26A virus infection (
To determine if the C26A virus was capable of intracellular transport in both the retrograde and anterograde directions, we examined axonal transport in cultured dorsal root sensory neurons by tracking red-fluorescence emissions from capsid particles by time-lapse microscopy. In these assays, the C26A virus behaved similarly to the wild-type virus, having only small decreases in axon transport during both stages of infection (
(A) Representative example of retrograde transport of an individual capsid in a dorsal root sensory axon after infection with the C26A virus, shown as a time-lapse montage. All frames are 1.68×10.8 µm. (B) Retrograde transport efficiency measured as frame-by-frame velocities of individual capsid particles in axons (as documented in panel A). (C) Anterograde transport efficiency measured by accumulation of newly replicated capsids in axons.
To further examine trans-synaptic retrograde transport
Stereotactic injection route resulting in exposure of viral inoculum to SCN neurons. RGC ipsilateral projection to the SCN indicates route of viral transmission to the eye. (A) Representative image of virus fluorescence in the SCN of a coronal brain slice (region imaged is indicated by the doted box in right panel). 3 V, third ventricle. Scale bar = 40 µm. (B) Virus detected in the eye following retrograde transmission from the SCN is seen as punctate fluorescence in the RGCs of the ganglion cell layer (GCL) and in bipolar/amacrine cells of the inner nuclear layer (INL) of the retina. The bright fluorescent band near the top of the image is autofluorescence emitted from the retinal pigmented epithelium (RPE) at the back of the retina, and is not of viral origin. Scale bar = 10 µm.
The data up to this point indicated that the C26A virus was incapable of spread in a subset of neural circuits, but the basis for the selective loss of function could not strictly be attributed to an inability to transport in either retrograde or anterograde circuitry. Because the C26A defect was observed only in retrograde circuits innervating tissues in the anterior chamber of the eye (i.e. iris and ciliary body), which are less efficiently infected following intravitreal injections, animals were next injected in the anterior chamber of the eye directly. This infection route immediately exposes the iris and neighboring ciliary body to the viral inoculum and provides a more reliable infection of the EW with all neuroinvasive strains of PRV examined to date
Anterior chamber injection route resulting in exposure of viral inoculum to the iris. The route of viral encephalitic spread is indicated: autonomic oculomotor nerve innervation of the iris from the ciliary ganglion (CG), which in turn receives innervation from parasympathetic neurons of the Edinger-Westphal nucleus (EW) of the midbrain. (A) Representative coronal images of EW (shown as a dashed box in coronal illustration) following anterior chamber injection or either wild-type or C26A virus. (B) Co-infection with PRV-152 and the C26A virus. Diffused GFP fluorescence and punctate RFP capsid signals are emitted from PRV-152 and the C26A viruses, respectively. Scale bars = 10 µm.
We next considered that the C26A virus typically propagated to titers 10-fold reduced relative to wild type (
Spread of C26A virus to the EW in the co-injection experiment indicated that PRV-152 had complemented the C26A defect. For complementation to occur, both viruses would have had to replicate in a common cell prior to entering the nervous system. This provided an incentive to examine the iris and ciliary body of the co-injected animals for co-infected cells (
(A) Images of the ciliary body following co-infection with PRV-152 (diffuse GFP signal) and the C26A virus (punctate RFP signal) in the anterior chamber. Cells infected with both viruses are evident in the merged image. (B) Illustration of the peripheral tissues in the eye (iris and ciliary body) exposed to viral inoculum and imaged in these studies. (C) C26A virus fluorescence from nuclei of cells in the iris following anterior chamber injection. Scale bars = 10 µm.
To confirm that the C26A virus had a specific defect in neuroinvasion from tissues innervated by the PNS, PRV-152 and the C26A virus were independently injected into the skin and muscle of one eyelid each per animal; the smooth muscle of the eyelid receives both sympathetic and parasympathetic innervation
Example of a single animal injected with PRV-152 and the C26A virus independently injected in opposing eyelids. Retrograde transmission of PRV-152, but not the C26A virus, to the locus cerulus (LC) is indicated in the fluorescence images. Region of brain imaged is illustrated as the two dashed boxes in a coronal brain slice (right panel). 4 V, fourth ventricle. Scale bar = 10 µm.
To explain how the C26A virus spread retrogradely in one animal, we again considered the possibility that injection into the eyelid might have caused local tissue damage that allowed direct exposure of the inoculum to innervating axon projections. We expect that a similar event likely occurred in the one intravitreal injected animal in which C26A virus spread specifically to the oculomotor nucleus (see above). To determine if injection can, in some instances, allow for direct exposure of the inoculum to axons embedded within the peripheral tissue, the beta subunit of cholera toxin (CTB) conjugated to Alexa Fluor 594 was injected into the eyelid. CTB is a retrograde tracer, but unlike neuroinvasive herpesviruses which spread trans-synaptically, CTB labeling is restricted to the neuron cell body from which an axon projects. Furthermore, CTB must come into direct contact with the axon to label the distal neuronal cell body. Of 6 animals tested, only one showed CTB retrograde label into the brain. The absence of CNS labeling in the remaining five animals indicated that injection into the eyelid infrequently allowed for direct exposure to nerve endings. This finding was consistent with the conclusion that the C26A virus is specifically incapable of invading the mammalian nervous system from tissues innervated by the PNS, except in infrequent cases when nerve endings become directly exposed to the inoculum.
Since its discovery four years ago
Our initial experiments produced results consistent with a selective loss of retrograde trafficking in the nervous system; however, retrograde transport of the C26A virus was unmasked by either infecting neurons directly in the retina or brain, or presumably by mechanical damage at the site of injection. An important insight was fortuitously made when the C26A mutant was co-infected with a virus encoding a wild-type pUL36 protein (PRV-Bartha), resulting in the rescue of the C26A defect and retrograde spread to the brain. This trans-complementation could only occur if the C26A virus remained competent to infect cells prior to neuroinvasion, which was subsequently confirmed by examining the eye for infection. Therefore, the C26A mutation resulted in a virus that replicated in non-neuronal cells both in culture and in the iris, and also replicated and spread in cultured neurons and neural circuits
In contrast to the results presented here, a previous study that examined the role of the PRV DUB following intranasal infection reported only a general delay in neuroinvasion
Identification of the molecular determinants that promote the neuroinvasive behavior of this class of herpesviruses not only provides insight into the underlying mechanisms of viral dissemination within host tissues, but also into the relationship between tissue tropism and pathogenesis. Yet, the majority of herpesviruses are not neuroinvasive pathogens, which argues that the pUL36 DUB may perform a more fundamental role during infection. Infection of cultured cells with DUB-mutant herpesviruses has revealed delays in viral assembly, propagation kinetics and cell-cell spread
The pUL36 tegument protein is greater than 300 kDa in neuroinvasive herpesviruses such as HSV and PRV, with the amino-terminal DUB amounting to less than 1/10th of the protein's mass. In addition to pUL36 being a critical component of the herpesvirus structure that links the capsid to the tegument and envelope layers of the virion
Two recombinant viruses were made from a derivative of the PRV (Becker strain) infectious clone, pBecker3, which was previously made to encode an mRFP1-VP26 (red-fluorescent capsid) fusion allele (pGS847)
PRV152 is a derivative of the PRV-Bartha vaccine strain that encodes a GFP-expression cassette driven by the CMV immediate early promoter inserted in the gG locus, and was described previously
Stocks of PRV-GS847 (WT), PRV-GS1652 (C26A), PRV-GS1768 (revertant) and PRV152 (Bartha-GFP) were harvested from infected PK15 cells grown in DMEM supplemented with 2% bovine growth supplement (Invitrogen). High titer virus stocks were prepared for a subset of animal studies as described previously
Quantitation of viral propagation kinetics by single-step growth curve analysis was conducted as previously described, with the exception that Vero cells were substituted for PK15 cells
Imaging of entry and egress axonal transport events was performed using primary cultured neurons from embryonic chicken (E8–E10) dorsal root ganglia, as previously described
Male Sprague-Dawley and Long Evans rats (Charles River Breeding Laboratories) were maintained under a light/dark cycle of 12 h light/12 h dark with food and water available
Under isoflurane inhalation anesthetic (2.5–5%), animals received a unilateral intravitreal injection of 2 µl of either PRV-GS847 or PRV-GS1652 (1–14.5×108 plaque-forming units/ml) or a unilateral anterior chamber injection of the same quantity of PRV-GS847, PRV-GS1652, PRV152 or a 1∶1 mixture of PRV-GS1652:PRV152, over a 1-min interval using a Nanoject II nanoinjector fitted with a glass micropipette (Drummond Scientific Co, Broomall, PA). A fresh stock of virus was thawed for each injection. Animals were maintained in a biosafety level 2 facility for up to 11 days post-inoculation.
Under isoflurane inhalation anesthetic (2.5–5%), animals were positioned in a Kopf stereotaxic apparatus (David Kopf Instruments, Tujunga, CA). A craniotomy was performed, the dura was exposed and a zero depth reading was made prior to the dura mater being excised. A glass micropipette attached to a Nanoject II nanoinjector (Drummond Scientific Co, Broomall, PA), angled laterally 10° to the vertical to avoid the midline third ventricle, was lowered into the hypothalamus to a region slightly dorsal to the suprachiasmatic nucleus (SCN) using coordinates determined empirically in test subjects. The micropipette was lowered 7.7–7.8 mm from the dura mater, thus avoiding involvement of the retinal axons in the optic chiasm, and approximately 200 nl (the Nanoject II was set to the 207 nl volume preset) of a 1×108 pfu/ml stock of PRV-GS1652 were ejected into the hypothalamic neuropil. The micropipette was left in place for about 1 minute before slowing retracting, and the craniotomy was packed with gelfoam and the incision was sutured.
Under isoflurane inhalation anesthetic (2.5–5%), the skin of the upper eyelid of one eye was injected with 2 µl of PRV-GS1652 (1–4.25×108 pfu/ml) and the upper eyelid of the other eye was injected with 2 µl of PRV152 (1×108 pfu/ml). In a separate set of experiments, animals received a unilateral upper eyelid injection of 2 µl of cholera toxin B subunit conjugated to Alexa Fluor 594 (Molecular Probes, Eugene, OR) using a Nanoject II nanoinjector fitted with a glass micropipette (Drummond Scientific Co, Broomall, PA).
After post-injection intervals ranging from 2–11 days, animals were deeply anesthetized with sodium pentobarbital (80 mg/kg, i.p.), and perfused transcardially with 0.9% saline followed by freshly prepared fixative consisting of 4% paraformaldehyde in phosphate buffer (0.1 M), pH 7.3. Brains were removed, stored in the same fixative containing 30% sucrose at 4°C overnight, and sectioned at 40 µm in the coronal plane on a sliding microtome equipped with a freezing stage (Physitemp Instruments Inc., Clifton, NJ). Sections were collected in phosphate buffer, mounted on subbed slides, blotted to remove excess buffer, and coverslipped with Vectashield mounting medium (Vector Laboratories, Burlingame, CA). Coverslips were sealed with fingernail polish to prevent dehydration, and slides were stored in the dark at 4°C. EGFP and mRFP1 fluorescence is stable under these conditions for several months with minimal quenching. Slides were examined using a Leica (Nussloch, Germany) DMRA light microscope equipped with epifluorescence and fitted with a microstepping servomotor in the z–axis. Images were captured using a Hamamatsu (Hamamatsu City, Japan) C4742-95 CCD digital camera under epifluorescence using either EGFP optics (412020 High Q narrow band EGFP filter; Chroma, Brattleboro, VT) or mRFP optics (#41004 HQ Texas Red filter, Chroma) and deconvolved using Openlab fluorescence deconvolution software (Improvision, Boston, MA) running on an Apple Macintosh G4 platform. Digital images were pseudo-colored, and images were prepared using Adobe Photoshop version 6.0.1. Images were enhanced for brightness and contrast.
We acknowledge the expert technical assistance of Anne Simpson, Connie King, and Allison Evans. We thank Lynn Enquist for his generous gift of VP5 antibody.