Genomic Sequences of a Low Passage Herpes Simplex Virus 2 Clinical Isolate and its Plaque-purified Derivative Strain

Abstract

Herpes simplex virus 2 is an important human pathogen as the causative agent of genital herpes, neonatal herpes, and increased risk of HIV acquisition and transmission. Nevertheless, the only genomic sequence that has been completed is the attenuated HSV-2 HG52 laboratory strain. In this study we defined the genomic sequence of the HSV-2 SD90e low passage clinical isolate and a plaque-purified derivative, SD90-3P. We found minimal sequence differences between SD90e and SD90-3P. However, in comparisons with the HSV-2 HG52 reference genome sequence, the SD90e genome ORFs contained numerous point mutations, 13 insertions/delections (indels), and 9 short compensatory frameshifts. The indels were true sequence differences, but the compensatory frameshifts were likely sequence errors in the original HG52 sequence. Because HG52 virus is less virulent than other HSV-2 strains and may not be representative of wildtype HSV-2 strains, we propose that the HSV-2 SD90e genome serve as the new HSV-2 reference genome.

Introduction

Herpes simplex virus 2 (HSV-2), a causative agent of genital herpes, undergoes a primary infection in the genital mucosa and spreads to sacral ganglia where the virus establishes a latent infection in sensory neurons. Reactivation of the latent virus leads to recurrent genital infections and lesions, called genital herpes. HSV-2 causes life-threatening infections in neonates infected during delivery and increases the risk of HIV acquisition and transmission (Roizman et al., 2013). Therefore, HSV-2 is an important human pathogen, and additional antivirals and a vaccine are needed.

The herpes simplex viruses are large double-stranded DNA (dsDNA) viruses that replicate in the nuclei of host cells (Roizman et al., 2013). The HSV genome is a linear dsDNA molecule comprised of two covalently linked segments, the long (L) and short (S) segments, which each consist of unique sequences (U_L and U_S) bounded by inverted repeats (Hayward et al., 1975) (Figure 1). Several HSV-1 genomic sequences have been determined (Macdonald et al., 2012a, b; McGeoch et al., 1988; McGeoch et al., 1986; Szpara et al., 2010), but the complete genome of only one strain of HSV-2 has been determined. The only available HSV-2 genomic sequence, that of the HG52 laboratory strain (Timbury, 1971), was determined a number of years ago (Dolan et al., 1998) and defined the genome size as 154,476 basepairs (bps). The HG52 DNA sequence was recently updated by Andrew Davison and provided in GenBank (JN561323.1), but the original sequence still serves as the reference genome for HSV-2. Although there is no complete genome sequence of a low-passage clinical isolate of HSV-2, sequencing of specific genes of HSV isolates gives evidence of at least limited sequence diversity among HSV-2 isolates (Norberg et al., 2007). Sequence analysis of glycoprotein genes in African and European HSV-2 isolates showed limited sequence diversity and defined two possible genogroups, one consisting entirely of African isolates and another containing both European and African isolates (Norberg et al., 2007). In addition, laboratory strains of HSV-2 show polymorphisms in restriction endonuclease cleavage sites (Hayward et al., 1975), and there is evidence for differences in immunological and pathogenic properties of HSV-2 strains from the United States and South Africa (Dudek et al., 2011).

Figure 1. Diagram of the Structure of the Herpes Simplex Virus Genome.

The top row shows the long (L) component and the short (S) component of the HSV genome. The bottom row shows the unique sequences as a line and the boxes denote the repeated sequences. U_L = unique long component sequences; U_S = unique short component sequences; R_L and R_L′ = inverted repeats bounding the long component; R_S and R_S′ denote inverted repeats bounding the S component. a = terminal repeat also located at the L/S junction.

The NCBI reference sequence strain, HSV-2 strain HG52 (Accession number NC_001798), shows variable virulence, in that plaque stocks showed LD₅₀ values ranging from greater than 10⁵ PFU to less than 10³ PFU by cranial inoculation (Taha et al., 1988). When inoculated peripherally, HSV-2 HG52 seems to be relatively avirulent. Very limited paralysis was observed when HG52 was inoculated in mice by the footpad route (Subak-Sharpe et al., 1984). In a study involving corneal infection of mice, 10⁶-10⁷ PFU of HG52 showed no mortality while similar dosages of HSV-2 strains 333 or 186 showed 100% mortality (Mitchell et al., 1990). The LD₅₀ for other HSV-2 strains is in the range of 1-6 × 10³ PFU via the vaginal route (Blakeney et al., 2005; Dudek et al., 2011). Furthermore, unlike other HSV-2 strains, HSV-2 HG52 is unable to shut off host protein synthesis. This defect is likely due to a frameshift mutation in the U_L41 gene, which encodes the viral host shutoff protein (Everett and Fenwick, 1990). Despite the knowledge of the host shutoff defect, the full genotype of HG52 responsible for the attenuated phenotype is currently not known.

In total, these considerations highlight the urgent need for a complete genome sequence of a low passage isolate of HSV-2. In this study we have generated the complete genome sequence of an early passage virulent African HSV-2 clinical isolate, SD90e, as well as a plaque-purified version of this isolate. We propose that this new clinical isolate genome could serve as a new HSV-2 reference genome.

Results

Comparison of an HSV-2 early passage genome with the HG52 reference genome

The HSV-2 SD90 virus, which was isolated at an STD clinic in South Africa (Lai et al., 2003), has been passaged only a limited number of times in cell culture and shows high virulence in a mouse genital model (Dudek et al., 2011), similar to most other clinical HSV-2 strains (Blakeney et al., 2005; Dudek et al., 2011). We determined the genomic sequence for an SD90 early passage isolate, named SD90e, using viral DNA purified from infected cell lysates by double banding on sodium iodide (NaI) density gradients. Illumina sequencing of this sample generated ∼6000-fold sequence coverage. Assembly of Illumina data using de novo assembly programs gave a number of blocks of sequence, or contigs, as observed for HSV-1 (Szpara et al., 2010). To refine the assembly and fill in large gaps, we mapped the sequence reads to the individual U_L, U_S, R_L and R_S regions containing 200 bp of overlapping sequence based on the HG52 reference sequence. This gave 5 large and 3 small contigs. The resulting contigs were used to manually reconstruct a full genome sequence by 1) mining Illumina raw reads when necessary to connect the contigs, 2) generating R_L and R_S inverted repeats bounding their respective unique sequences by inverting a copy of each sequence, and 3) assignment of numbers of repetitive sequences (a sequences) at the termini and the junction of R_L and R_S to match the HG52 reference sequence. There was one region in each of the ICP4/RS1, ICP34.5/R_L1, U_L36, and ICP0/R_L2 ORFs whose sequence was ambiguous. We therefore used the primers listed in Table 1 and the conditions described in Materials and Methods to PCR-amplify these regions from SD90e DNA, and we then sequenced the amplified DNA by conventional Sanger sequencing. The sequences in the repeated regions in the ICP4/RS1, ICP34.5/R_L1, and ICP0/R_L2 ORFs were determined using the primers in Table 1.

Table 1. Primer Sets Used in this Study.

Amplified region	Amplimer size, bp*	Forward primer (nt position in HG52)	Reverse primer (nt position in HG52)
ICP4/RS1	240	GCGGACGACGAGGAG (150559-150573)	GTAGCGCGCGTAGAAGG (150782-150798)
ICP34.5/RL1	221	CTACGCCGAGCCCAG (416-430)	CTGTCGTCCGCCTGG (652-666)
ICP0/RL2	583	AAGAGGCGCGGGTCGG (4430-4445)	TGACCGTCTTGTTCACGTAAGGC (4990-5012)
UL36	577	GGAGTCTGCGTCGGAGTGTTC (71890-71910)	GGCAGTGCCGCCTTCTC (72450-72466)

^*Size predicted by HG52 GenBank sequence.

The U_L36 gene repeats were more refractory in that Sanger sequencing of the SD90e fragment showed at least 9 copies of the 15 nucleotide repeat (caggggcggctgggg) before the sequence appeared to move into unique sequences, but the sequence was not unequivocal. Parallel sequencing of our isolate of HSV-2 HG52 DNA showed 6 copies of the U_L36 repeat, in contrast to the 11 copies in the GenBank reference sequence. To attempt to confirm these numbers of repeats, we analyzed the PCR products in an 8% polyacrylamide gel (Fig. 2). The SD90e amplimer was approximately 550 bp, consistent with the 547 bp predicted by 9 copies of the repeated sequence while the HG52 amplimer was approximately 500 bp, consistent with the 502 bp predicted by 6 copies of the repeated sequence. Therefore, we have provisionally described the U_L36 gene of SD90e as having 9 copies of this repeated sequence, but further analysis is needed to confirm this.

Figure 2. Sizing of the U_L36 PCR Amplimers by Polyacrylamide Gel Electrophoresis.

Primers described in Table 1 for U_L36 were used to amplify viral genomic sequences from bp 71,890-72466 (HG52 reference sequence) in SD90e or HG52 DNAs. The amplimers were analyzed by electophoresis in an 8% polyacrylamide gel. Lanes 1 and 4, M = DNA size markers; Lane 2, HG52 amplimer; Lane 3, SD90e amplimer.

The combination of the Illumina and Sanger sequence analyses gave a complete SD90e genome of 154,096 basepairs (GenBank accession number KF781518). We then compared our SD90e genome assembly with the HSV-2 HG52 reference genome in GenBank (NC_001798) and the recent HG52 sequence submitted by Andrew Davison (GenBank accession number JN561323.1, determined by “sequence of viral DNA and polyA RNA harvested at 10 h after infection”) using the Genome Annotation Transfer Utility (GATU) program (Pickett et al., 2012) from the Virus Pathogen Resource (http://www.viprbrc.org). The SD90e genome contained more than 100 nucleotide differences within open reading frames compared to the HG52 reference sequence, but overall, U_S was 99.3% conserved and U_L was 99.6% conserved between the two genomes.

Nine ORFs contained a combined total of thirteen indels (insertions or deletions) (Table 2). These indels were multiples of 3 nucleotides so the reading frame was not altered. The original HG52 sequence and the more recent Davison HG52 sequence were identical in these regions (Table 2). Furthermore, our own sequence of HG52 shows the same indels (results not shown). Therefore, these indels are likely to be real sequence differences between SD90e and HG52. Because several of the affected ORFs encode essential or important viral proteins, we predict that these changes could account for the reduced virulence of the HG52 virus. For example, the SD90e ICP4 (R_S1) gene encoding the major viral transactivator protein showed three indels, a 3-codon deletion, a 1-codon insertion, and a 6-codon insertion (Table 2). The R_L2 gene encoding the ICP0 immediate early protein showed a 20-codon deletion. The U_L27 ORF encoding glycoprotein B, which is essential for viral entry, showed a 3-codon deletion near the amino terminus. The U_L32 ORF encoding an essential DNA packaging protein showed a 2-codon deletion. The U_L49 ORF encoding the VP22 virion tegument protein showed 2-codon insertion. The other ORFs affected that were encode viral proteins that play important roles in vivo. The R_L1 gene encoding the ICP34.5 antagonist of the cellular interferon response contained an 8-codon deletion and a 1-codon insertion. The U_L39 ORF encoding the large subunit of ribonucleotide reductase contained a 3-codon deletion and a 2-codon insertion. The U_S8 ORF, encoding the glycoprotein E subunit of the viral F_c receptor, contained a 3-codon insertion. Finally, the U_S4 ORF encoding glycoprotein G contained a 1-codon deletion within the HSV-2 specific region. Variation within this region could affect the ability of the standard HSV-2 reference sera to detect the HSV-2 specific antigen in the C-terminus of gG-2.

Table 2. Insertions/deletions in HG52 versus SD90e Genomic Sequences.

Viral Genome		Translated Sequence
RL1 ICP34.5 Neurovirulence Factor
HG52_Reference	1	MSRRRGPRRRGPRRRPRPGAPAVPR
HG52_Davison	1	MSRRRGPRRRGPRRRPRPGAPAVPR
SD90e	1	MSRRR-PRRRGPRRR--------PR
HG52_Reference	230	RAHEDGGPA-EEEEAAAAARGSSAA
HG52_Davison	230	RAHEDGGPA-EEEEAAAAARGSSAA
SD90e	230	RAHEDGGPAEEEEEAAAAARGSSAA
RL2 ICP0 Immediate Early Protein
HG52_Reference	599	SSSSASSSSASSSSASSSSASSSSA
HG52_Davison	599	SSSSASSSSASSSSASSSSASSSSA
SD90e	599	SS------------------SSSSA
UL27 Glycoprotein B
HG52_Reference	20	ASAAPAAPAAPRASGGVAATVAANG
HG52_Davison	20	ASAAPAAPAAPRASGGVAATVAANG
SD90e	20	ASAAPAAPRA---SGGVAATVAANG
UL32 DNA Packaging Protein
HG52_Reference	251	QSATRALIGTLASGGDDGAGAGAGGG
HG52_Davison	251	QSATRALIGTLASGGDDGAGAGAGGG
SD90e	251	QSATRALIGTLASGGDD--GAGAGGG
UL39 Ribonucleotide Reductase Subunit 1
HG52_Reference	111	VVALGGTSGPSATTSVGTQTSGEFLH
HG52_Davison	111	VVALGGTSGPSATTSVGTQTSGEFLH
SD90e	111	VVALGGTS---ATTSVGTQTSGEFLH
HG52_Reference	200	SSDEDTGS--ETLSRSSSIWAAGATD
HG52_Davison	200	SSDEDTGS--ETLSRSSSIWAAGATD
SD90e	200	SSDEDTGSGSETLSRSSSIWAAGATD
UL49 VP22 Tegument Protein
HG52_Reference	61	EAGYALYRDSSS--DDDESRDTARPR
HG52_Davison	61	EAGYALYRDSSS--DDDESRDTARPR
SD90e	61	EAGYALYRDSSSSEDNDESRDTARPR
RS1 ICP4 Immediate Early Protein
HG52_Reference	151	TPSMRADYGEENDDDDDDDDDDDRDA
HG52_Davison	151	TPSMRADYGEENDDDDDDDDDDDRDA
SD90e	151	TPSMRADYGEEN---DDDDDDDDRDA
HG52_Reference	201	SPRPPAPRRHHHHHHH-RRRRAPRRR
HG52_Davison	201	SPRPPAPRRHHHHHHH-RRRRAPRRR
SD90e	198	SPRPPAPRRHHHHHHHRRRRRAPRRR
HG52_Reference	300	PRPSPPRAEPAP------ARTPAATA
HG52_Davison	300	PRPSPPRAEPAP------ARTPAATA
SD90e	298	PRPSPPRAEPAPGAGAGSARTPAATA
US4 Glycoprotein G
HG52_Reference	421	SPLPAAAAATPGAGHTNTSSASAAKTP
HG52_Davison	421	SPLPAAAAATPGAGHTNTSSASAAKTP
SD90e	421	SPLP-AAAATPGAGHTNTSSAPAAKTP
US8 Glycoprotein E
HG52_Reference	30	VTSGEDVVLLPAPA---ERTRAHKLLW
HG52_Davison	30	VTSGEDVVLLPAPA---ERTRAHKLLW
SD90e	30	VTSGEDVVLLPAPAGPEERTRAHKLLW

In addition, there were 9 sequence differences between the SD90e and HG52 reference sequences that consisted of pairs of closely spaced single nucleotide indels that change the reading frame for a short distance before it goes back into the reference reading frame (Table 3). However, comparison of the recent Davison HG52 sequence and our own HG52 Illumina sequence data showed that these two HG52 sequences are identical to the SD90e sequence in these areas (Table 3). Therefore, it is likely that these differences relative to the original HG52 reference sequence are due to sequencing errors in the original HG52 sequence.

Table 3. Compensating Frameshifts in Open Reading Frames.

Gene	Nucleotide sequence		Translated Sequence
UL6
HG52_Reference	1177	GGGTCGCGC-CCCCAG	GSRPQ
HG52_Davison	1177	GGGTCG-GCGGCCCAG	GSAAQ
HG52_Knipe1177	GGGTCG-GCGGCCCAG	GSAAQ
SD90e	1177	GGGTCG-GCGGCCCAG	GSAAQ
UL8
HG52_Reference	337	AC-GGGCCGGTGGGCC	TGRWA
HG52_Davison	337	ACGGGGCCGGT-GGCC	TGPVA
HG52_Knipe337	ACGGGGCCGGT-GGCC	TGPVA
SD90e	337	ACGGGGCCGGT-GGCC	TGPVA
UL9
HG52_Reference	847	CG-CCGCCCGGG	RRRG
HG52_Davison	847	CGCCCGCC-GGG	RPPG
HG52_Knipe847	CGCCCGCC-GGG	RPPG
SD90e	847	CGCCCGCC-GGG	RPPG
UL16
HG52_Reference	541	ACGGACACGGCCCCCGAATCC-GGG	TDTAPESG
HG52_Davison	541	ACGGA-ACCGCCCCCGAATCCGGGG	TEPPPNPG
HG52_Knipe541	ACGGA-ACCGCCCCCGAATCCGGGG	TEPPPNPG
SD90e	541	ACGGA-ACCGCCCCCGAATCCGGGG	TEPPPNPG
UL19
HG52_Reference	2624	GTCAA-GCGCATG	VKRM
HG52_Davison	2624	GTCAACGCC-ATG	VNAM
HG52_Knipe2624	GTCAACGCC-ATG	VNAM
SD90e	2624	GTCAACGCC-ATG	VNAM
UL38
HG52_Reference	907	GGGGGCCCCTGGGG-GGCG	GGPWGA
HG52_Davison	907	GGGGG-CCCTGGGGCGGCG	GGPGAA
HG52_Knipe907	GGGGG-CCCTGGGGCGGCG	GGPGAA
SD90e	907	GGGGG-CCCTGGGGCGGCG	GGPGAA
HG52_Reference	1369	TGGCG-CGCGGCGGGG	WRAAG
HG52_Davison	1369	TGGCGCCCCGGC-GGG	WRPGG
HG52_Knipe1369	TGGCGCCCCGCC-GGG	WRPAG
SD90e	1369	TGGCGCCCCGCC-GGG	WRPAG
UL46
HG52_Reference	1085	ACCTTAC-GGCG	TLRR
HG52_Davison	1085	ACC-TACGGGCG	TYGR
HG52_Knipe1085	ACC-TACGGGCG	TYGR
SD90e	1085	ACC-TACGGGCG	TYGR
US1
HG52_Reference	472	CCGG-CCAGTCTTCCGGGGATCGCGCACGCGCAC--CGGCG	PASLPG-IAHAHR
HG52_Davison	472	CCGGCCCAGTCTTCCGGGGATCGCGCAGCCGCACCGCGGCG	PAQSSGDRAAAPR
HG52_Knipe	472	CCGGCCCAGTCTTCCGGGGATCGCGCAGCCGCACCGCGGCG	PAQSSGDRAAAPR
SD90e	472	CCGGCCCAGTCTTCCGGGGATCGCGCAGCCGCACCGCGGCG	PAQSSGDRAAAPR

Comparison of an HSV-2 low passage viral genome with a plaque-purified derivative

We also determined the sequence of the genome of the SD90-3P triply plaque-purified virus (Dudek et al., 2011) using the same Illumina sequencing methods as described for SD90e. Comparison of the SD90e and SD90-3P U_L and U_S sequences showed indels in homopolymeric sequences in the intergenic sequences and only two codon changes, U_L16 V126I and U_L31 R22L in the U_L and U_S ORFs of SD90e. Therefore, the two genomes were nearly identical, showing that the process of plaque-purification of the SD90e virus resulted in minimal changes to the viral genome.

Discussion

The herpes simplex virus DNA genomes have been difficult to sequence and assemble because of the high G+C content of the DNA and the highly repetitive nature of the genomes. Several HSV-1 genomic sequences have been determined, but the complete genome of only one laboratory strain of HSV-2 has been determined. Therefore, there has been an urgent need for the complete genomic sequence of a low passage clinical isolate of HSV-2, both to confirm the original sequence and to define the genome of a fresh clinical isolate. In this study we determined the genome sequence of a limited passage HSV-2 strain, SD90e, originally isolated in South Africa, and a plaque-purified derivative of this isolate. We found that the U_L and U_S ORFs of the early passage virus, SD90e, and the plaque-purified virus differed by just two nucleotides, which was consistent with our previous results showing that the pathogenicity, as measured in paralysis dose₅₀ values, were similar for SD90e and SD90-3P (Dudek et al., 2011). The similarity of the genomes showed that plaque purification on Vero cells did not exert a strong selection pressure on the virus. Although it is still advisable to restrict the number of viral passages to limit genetic variation that might be due to selective pressures in cell culture, these results demonstrate that it is possible to clonally purify an HSV-2 isolate with minimal genetic changes.

It is important to determine the genome sequence of a low passage clinical isolate of a virus because extensive passaging of the virus in cell culture can select for mutations that allow the virus to replicate more efficiently in vitro, yet attenuate virulence in vivo. For example, laboratory strains of HCMV contain a large genomic deletion relative to clinical isolates (Chatellard et al., 1998). We did not find any large deletions in the SD90e genome relative to the HG-52 genome, and these results and other sequencing that we have done have not shown any major genomic differences between lab strains and clinical isolates (unpublished results). HG52 does contain numerous point mutations and thirteen indels compared to SD90e. Several of the indels are in ORFs encoding essential viral replication proteins and nonessential proteins that are important in HSV pathogenesis in vivo. Some or many of these changes could be the basis for the attenuated phenotype observed with HG52 in animal studies.

The HG52 sequence as originally reported contains a number of sequence differences from SD90e and from the HG52 sequences recently reported in Genbank by Andrew Davison (JN501323.1) and our own studies of HG52. The differences could be due to the heterogeneity in the original HG52 stocks as described above, but it seems more likely that the differences are the result of sequencing errors due to the inherent difficulty in sequencing the G+C rich HSV-2 DNA using conventional sequencing techniques. Although a corrected sequence for HG52 has been posted in Genbank by Davison, the original HG52 sequence remains as the reference genome sequence.

HSV-2 Genomic Sequencing

The approaches used in this study may facilitate further advances in HSV-2 genomic sequencing. Double banding of viral DNA in NaI gradients provided a pure substrate for Illumina sequencing. Illumina sequencing and assembly analyses yielded a nearly complete genomic sequence with 4 areas of low coverage and ambiguous sequence. PCR amplification using special conditions for high GC DNA and Sanger sequencing provided or confirmed the sequence of three of these regions, the ICP4/R_S1, ICP34.5/R_L1, and ICP0/R_L2 ORFs. The repetitive region in the U_L36 ORF was tentatively assigned based on Sanger sequencing, but further analysis is needed to confirm the number of repeats in this ORF. These regions are often gaps in HSV-2 sequence assemblies so these primers may be generally useful in closing gaps that arise frequently in assembled HSV-2 genomic sequences. In total, these analyses allowed the generation of a complete genomic sequence. These and further advances in mining of sequences to fill gaps should allow the generation of the large number of HSV-2 genome sequences needed for phylogenetic analysis.

Reference strain for HSV-2

HG52 shows significant attenuation relative to other HSV-2 strains as described above and contains several indels relative to SD90e. HG52 also contains a frameshift mutation in the U_L41 ORF, which has been corrected to the WT sequence in the reference sequence in GenBank (NC_001798). For all of these reasons, HG52 does not seem to be representative of other HSV-2 laboratory strains and clinical isolates. We propose that a more representative strain, such as SD90e, should be designated as the reference genome for HSV-2 for future studies. We have used this virus for genital infection challenge studies (Dudek et al., 2011), and the virus is available for other researchers to use as a prototype of a low-passage HSV-2 virus in vaccine or pathogenesis studies.

The definition of the genomic sequence of a low passage clinical HSV-2 isolate will provide the foundation for comparative sequencing of HSV-2 isolates from different geographic origins, different diseases and different anatomic sites such as genital lesions versus meningitis. This will allow a determination of whether genetic variability of HSV-2 is linked to specific herpetic diseases or increased risk of HIV-1 infection. Definition of the genetic variability of viral epitopes will also be important in assessing the selective pressure of the host immune responses on viral evolution.

Materials and Methods

Viruses

The HSV-2 SD90 virus was isolated from a genital ulcer swab sample from an HIV-1–negative male attending a sexually transmitted disease clinic in Carletonville, South Africa, in 1995 (Lai et al., 2003). The virus was passaged once in South Africa, frozen at -70°C and shipped to the Centers for Disease Control and Prevention. A frozen sample was provided to the Knipe laboratory by Ron Ballard (Centers for Disease Control and Prevention). The study of these viruses received a human subjects exemption from the Harvard Medical School IRB. The primary isolate was passaged 3 times on Vero cells at Harvard Medical School to generate the SD90e or early passage virus. SD90-3P is a clonal virus derived from SD90 through 3 successive plaque purifications on Vero cells (Dudek et al., 2011). HSV-2 strain HG52 was provided by Bernard Roizman, University of Chicago.

Preparation of Viral DNA

Vero cells were infected with the HSV-2 viruses at a multiplicity of 1 PFU/cell and incubated at 34°C for 24 hours. The infected cells were collected, and the HSV-2 DNA was purified from infected cell lysates by double banding on sodium iodide gradients essentially as described previously (Walboomers and ter Schegget, 1976), except that proteinase K was used instead of pronase. In sequence runs, preparations made in this way gave 75-90% of the sequence runs aligning with HSV-2 reference genome, which was higher than other methods of purification that were tested (results not shown).

Methods for sequencing

For conventional Sanger sequencing, HSV-2 sequences were PCR-amplified from 300 ng of sodium iodide gradient-purified viral DNA. PCR reactions contained 10% DMSO, 0.5 M Betaine (Sigma, B0300), 2.5mM of each dNTP (NEB, N0446) except for dGTP which was added as a 1:1 mix of dGTP:7-deaza-dGTP (NEB, N0445), 300 nM of each forward and reverse primer, and 1.25 units of Ex Taq polymerase (TaKaRa, RR001A) for each 50 μl reaction. PCR cycling conditions were set as suggested by the polymerase manufacturer, noting that the elongation step was always carried out at 68°C. PCR products were run on agarose gels to confirm the size and purity of the amplicon. The DNA was then purified with a QIAquick PCR purification kit (Qiagen, 28106) and sent to GeneWiz, Inc. (Cambridge, MA) for Sanger sequencing using their GC-rich sequencing protocol. Library construction and sequencing by 454 and Illumina platforms were performed at the Broad Institute as described previously (Grad et al., 2012; Lennon et al., 2010).

Assembly of Genomic Sequences

For the consensus assembly, Illumina fragment pair data were first processed using ALLPATHS-LG (version R44182) to find overlaps between fragment pairs and fill gaps where no overlap is present. This created a set of sequencing fragments that consist of the complete sequence between two ends of a paired read set. These unpaired filled fragments were then analyzed using Roche's runMapping (version vMapAsmResearch-10/14/2011) program with default parameters and a reference genome, HSV-2 HG52, which consisted of unique segments and R_L and R_S segments of the HG52 genome flanked by a small amount of additional, repetitive, sequence on each end. The runMapping tool produced a five-contig consensus built from the placements of the SD90e filled fragment reads to the aforementioned HG52 reference genome.

Research Highlights.

• First genomic sequence of a low passage isolate (SD90e) of herpes simplex virus 2 was determined.

• SD90e shows several indels in essential viral genes as compared with the HSV-2 HG52 reference DNA.

• Plaque purification of the SD90e virus resulted in minimal ORF sequence changes.

• HSV-2 SD90e is proposed to serve as a new HSV-2 reference genome.