A point mutation in a herpesvirus polymerase determines neuropathogenicity

Abstract

Infection with equid herpesvirus type 1 (EHV-1) leads to respiratory disease, abortion, and neurologic disorders in horses. Molecular epidemiology studies have demonstrated that a single nucleotide polymorphism resulting in an amino acid variation of the EHV-1 DNA polymerase (N752/D752) is significantly associated with the neuropathogenic potential of naturally occurring strains. To test the hypothesis that this single amino acid exchange by itself influences neuropathogenicity, we generated recombinant viruses with differing polymerase sequences. Here we show that the N752 mutant virus caused no neurologic signs in the natural host, while the D752 virus was able to cause inflammation of the central nervous system and ataxia. Neurologic disease induced by the D752 virus was concomitant with significantly increased levels of viremia (p = 0.01), but the magnitude of virus shedding from the nasal mucosa was similar between the N752 and D752 viruses. Both viruses replicated with similar kinetics in fibroblasts and epithelial cells, but exhibited differences in leukocyte tropism. Last, we observed a significant increase (p < 0.001) in sensitivity of the N752 mutant to aphidicolin, a drug targeting the viral polymerase. Our results demonstrate that a single amino acid variation in a herpesvirus enzyme can influence neuropathogenic potential without having a major effect on virus shedding from infected animals, which is important for horizontal spread in a population. This observation is very interesting from an evolutionary standpoint and is consistent with data indicating that the N752 DNA pol genotype is predominant in the EHV-1 population, suggesting that decreased viral pathogenicity in the natural host might not be at the expense of less efficient inter-individual transmission.

Figure 1. Viral Mutagenesis

(A) Genomic restriction fragment patterns of the wild-type Ab4 strain (lane 1), the N752 mutant (lane 2), and the revertant D752 (lane 3) were analyzed to verify that no gross rearrangements had occurred during mutagenesis (marker, 1-kb ladder [Invitrogen]). (B) Expression of the glycoprotein restored in place of BAC sequences (gp2, 250 kDa, indicated by arrow) in all constructs was confirmed by Western blot analysis, with EHV-1 strain RacL11 (with and without gp2 [29]) and the parental Ab4 BAC as controls (marker, Precision Plus [Bio-Rad]). Mutant virus constructs were also sequenced in the region of interest at each passage level, including after reisolation from horses.

Figure 2. Clinical Data from the Equine Experimental Infection Studies

Studies were conducted in a smaller pilot study (ponies) and subsequently a larger study (horses) with the revertant pol genotype virus (D752, •) and mutant virus (N752, ○). Data were collected before and after treatments. *p < 0.05. (A and B) Median cumulative clinical scores, which indicate overall symptom severity. (C and D) Median rectal temperatures (°C), with a line drawn at the cutoff temperature for fever (38.5 °C). (E and F) Median serum virus neutralizing antibody titers. Data from the first experiment on ponies (four animals/group) are shown in the left graphs (A, C, and E), those from the second experiment on horses (seven animals/group) in the right graphs (B, D, F).

Figure 3. Real-Time Quantitative PCR Data from the Equine Experimental Infection Studies

Results from studies conducted on ponies (A and C) and horses (B and D) with the revertant pol genotype virus (D752, •) and mutant virus (N752, ○). *p < 0.05. (A and B) Geometric means of normalized lymphocyte-associated viremia measured by qPCR, with EHV-1 DNA copies normalized per million cellular genomic DNA copies. (C and D) Virus titers in nasal excretions measured by qPCR. Shown are geometric mean normalized viral genome copies per milliliter of nasal swab solution. Standard deviations (error bars) are plotted, but are small compared to the y-axis scale and thus not visible for some data points. Data from the first experiment (ponies, four animals/group) are shown in the left graphs (A and C), those from the second experiment (horses, seven animals/group) in the right graphs (B and D).

Figure 4. Viral Isolation from PBMC/RK13 Coculture Experiments

Virus was isolated during the equine experimental infection studies with the revertant pol genotype virus (D752, •) and mutant virus (N752, ○). *p < 0.05. The number of culture-positive animals are shown, out of four total ponies per group (A) and seven total horses per group (B).

Figure 5. Postmortem Examination of Horses Infected with N752 or D752 Viruses

(A) Normalized viral genome load in tissues collected from horses examined postmortem: N752–263 of the mutant group, with only two tissues positive for viral genome copies (□) and D752–4 of the revertant group (▪), with multiple positive tissues throughout the CNS and lymphatic system. RTG, right trigeminal ganglion; RLN, retropharyngeal lymph node; BLN, bronchial lymph node; SMLN submandibular lymph node. (B) Representative photomicrograph of lymphocytes present in the CSF of horse D752–4, with red cytoplasmic granules typical of cytotoxic T cells or natural killer cells. (C) Representative hematoxylin and eosin-stained section of caudal thoracic spinal cord tissue from horse N752–263 (bar 700 μm) showing no abnormalities. (D and F) Representative caudal thoracic cord sections from horse D752–4 (bars indicate, respectively, 1 mm, 40 μm, and 40 μm) showing lymphocytic cuffing in the meninges (D), dilated myelin sheaths with swollen axons, surrounded by reactive mononuclear cells (E), and typical lymphocytic perivascular cuffing (F).

Figure 6. Equine PBMC Infection In Vitro

(A) Each point represents the percentage of virus-infected cells that were positive for the respective equine cellular marker in an individual horse, as measured by FACS at 48 h postinfection. Cells were infected with either the neuropathogenic D752 revertant (•) or the non-neuropathogenic N752 mutant (○) virus; dark lines show median percentages for the revertant D752 group, and grey lines median percentages for the N752 mutant group. (B) Paired differences between the two viruses (percentage of infected cells expressing the respective cellular marker: N752 minus D752 for each individual horse) are plotted. The efficiency of PBMC infection ranged from 10% to 30%.

Figure 7. Differential Sensitivity of the Pol D/N752 Virus and Protein Variants to Aphidicolin and Structural Modeling of the Pol Region of Interest

(A) Cells were infected with the Pol D752 revertant (•) or N752 mutant (○) virus, treated with aphidicolin, incubated 3 d, then lysed; final virus yield was titrated on new cells. (B) DNA was also extracted after the lysing step and qPCR performed to quantify normalized viral genome copies. (C) The DNA polymerase activity of Pol D752 and Pol N752 proteins, in the absence and in the presence of pORF18 (Pol accessory subunit), was analyzed by measuring the incorporation of [³H]dTTP into a poly(dA)-oligo(dT) template. (▪) Pol D752; (□) Pol N752; (•) Pol D752 + pORF18; (○) Pol N752 + pORF18. (D) The effect of aphidicolin on polymerase activity of Pol D752 (•) and of Pol N752 (○) was assayed by measuring the incorporation of [³H]dTTP into a poly(dA)-oligo(dT) template in the presence of pORF18. Graphs show the average of three experiments with standard deviations (error bars). Asterisk * indicates p < 0.05. (E) Ribbon diagram of EHV-1 Pol N752 is based on HSV-1 Pol crystal structure [23]. (F) Space-filling diagram highlights the region between HSV-1 Pol secondary structure elements P3 and PB on the outside surface of the palm domain. (G) Prediction of structural changes caused by the residue variation in Pol N752 as opposed to Pol D752.