In the present paper, we present a PCR-based approach that allows the rapid and accurate discrimination of unrelated HSV-1 strains. The method generally used to discriminate HSV-1 strains is RFLP analysis (7
). This method requires virus culture, is time-consuming, and is highly labor-intensive. Furthermore, culture requires viable virus, which is not always obtainable from certain types of clinical samples (e.g., cerebrospinal and intraocular fluids). More recently, a system that uses PCR amplification and subsequent RFLP analysis has been developed to facilitate discrimination of HSV-1 strains, eliminating the necessity of virus culture. This method, however, is not significantly less time-consuming or labor-intensive than conventional strain differentiation (31
Conventional RFLP analysis with restriction endonucleases that recognize 6 bp (6-bp REs) is insufficient for differentiation of HSV-1 strains of a predominant genotype. The use of 4-bp REs and RFLP analyses of reiterated sequences greatly improved the differentiation rate (28
). As in our study, the RFLP analysis of reiterated sequences was based on various numbers of repeats. Use of both techniques generated similar results, verifying the applicability of either method in molecular epidemiological studies (27
). Similar hypervariable regions have been used successfully to discriminate strains of other herpesviruses like Epstein-Barr virus and human cytomegalovirus (23
To be applicable in a PCR-based assay for discrimination of different HSV-1 strains, these regions should show a considerable degree of variability and should remain stable during a relatively short time of replication. We tested the suitability of several HSV-1 hypervariable regions for discrimination of unrelated HSV-1 strains.
Due to their G+C-rich sequences, standard PCR protocols failed to reproducibly amplify the regions tested. The high G+C content increases the formation of secondary structures, preventing consistent amplification of the repeats. We tested a number of PCR conditions in order to obtain consistent DNA amplification. Addition of DMSO as a cosolvent to the reaction mixture has previously been shown to facilitate DNA amplification of G+C-rich sequences (12
). Introduction of the exonuclease activity of the Pfu
DNA polymerase enzyme in the PCR mixture prevents “skipping” of the repeats, which could result in the formation of products smaller than the actual size of the template repeat (2
). Another modification was the introduction of 7-deaza-2′-dGTP. This analogue of dGTP is equally well incorporated into DNA but exerts a lesser binding strength to dCTP than normal dGTP (6
). The use of Pfu
polymerase, 50% 7-deaza-2′-dGTP as a replacement for 100% dGTP, and 5% DMSO resulted in the most consistent amplification of the large alleles. The specificities of the amplicons were confirmed by hybridization with Re-specific probes after Southern blotting.
Analysis of subclones of HSV-1 F showed that the stability of the ReI and ReIII sequences was too low to be useful for discrimination of HSV-1 strains. In contrast, ReIV and ReVII were shown to be stable during this procedure. Thus, regions US1 (ReIV), US12 (ReIV), and US10-US11 (ReVII) were chosen for use in the discrimination of unrelated corneal HSV-1 isolates.
In agreement with previous studies, the variability in the US10-US11 region was found to be relatively low (27
). We detected only three different alleles among 37 unrelated clinical HSV-1 isolates, which is not surprising since ReVII is located within a protein-coding region, making it a target for selective pressure. More drastic changes in the length of US10-US11 could influence the translation or function of the proteins encoded by genes US10 and US11. In contrast, the ReIV-containing sequences are located in the introns of genes US1 and US12. We found 14 and 15 different alleles for regions US1 and US12, respectively, in the 37 corneal HSV-1 isolates analyzed. Comparison of the alleles from the three regions for all 37 corneal HSV-1 isolates revealed 34 unique combinations. The isolates with identical combinations were obtained at different time points, indicating that this was most likely not due to contamination during virus isolation or culture procedures.
Sequential corneal isolates from five individuals with recurrent herpetic corneal infections were analyzed. For each individual, sequential samples showed identical DNA patterns, while the patterns for samples from different patients were different. These results indicate that the recurrent infections were most likely caused by the same virus. A comparative sequence database search revealed several point mutations between different HSV-1 strains, in addition to various numbers of repeats. More detailed analysis, like sequencing of the amplicons, might provide more conclusive evidence for this assumption. This also demonstrates that these hypervariable regions remain stable during reactivation and replication of latent HSV-1 in the corneas of these individuals.
Additionally, we have also analyzed clinical samples in which no viable virus can usually be detected (4
). Re sequence-specific PCR analyses were performed with DNA isolated from affected corneal buttons and rims obtained from patients with herpetic stromal keratitis during therapeutic keratoplasty. The PCR approach proved to be sensitive enough for amplification of the low levels of viral DNA present in these samples (unpublished data). The major advantage of the approach presented is that it provides the opportunity to discriminate HSV-1 strains without virus culture or RFLP analysis, making it convenient for rapid diagnostic testing. Although not suitable for classification of HSV-1 strains, it provides a powerful tool that can be used to address questions regarding reactivation and the modes of transmission of HSV-1. For example, it could be used to assess the risk of HSV-1 transmission through cornea transplantation and other manifestations of recurrent HSV-1 infections.