Impact of sequence variation in the UL128 locus on production of human cytomegalovirus in fibroblast and epithelial cells.

The human cytomegalovirus (HCMV) virion envelope contains a complex consisting of glycoproteins gH and gL plus proteins encoded by the UL128 locus (UL128L): pUL128, pUL130, and pUL131A. UL128L is necessary for efficient infection of myeloid, epithelial, and endothelial cells but limits replication in fibroblasts. Consequently, disrupting mutations in UL128L are rapidly selected when clinical isolates are cultured in fibroblasts. In contrast, bacterial artificial chromosome (BAC)-cloned strains TB40-BAC4, FIX, and TR do not contain overt disruptions in UL128L, yet no virus reconstituted from them has been reported to acquire mutations in UL128L in vitro. We performed BAC mutagenesis and reconstitution experiments to test the hypothesis that these strains contain subtle mutations in UL128L that were acquired during passage prior to BAC cloning. Compared to strain Merlin containing wild-type UL128L, all three strains produced higher yields of cell-free virus. Moreover, TB40-BAC4 and FIX spread cell to cell more rapidly than wild-type Merlin in fibroblasts but more slowly in epithelial cells. The differential growth properties of TB40-BAC4 and FIX (but not TR) were mapped to single-nucleotide substitutions in UL128L. The substitution in TB40-BAC4 reduced the splicing efficiency of UL128, and that in FIX resulted in an amino acid substitution in UL130. Introduction of these substitutions into Merlin dramatically increased yields of cell-free virus and increased cell-to-cell spread in fibroblasts but reduced the abundance of pUL128 in the virion and the efficiency of epithelial cell infection. These substitutions appear to represent mutations in UL128L that permit virus to be propagated in fibroblasts while retaining epithelial cell tropism.

High-resolution human cytomegalovirus transcriptome.

Deep sequencing was used to bring high resolution to the human cytomegalovirus (HCMV) transcriptome at the stage when infectious virion production is under way, and major findings were confirmed by extensive experimentation using conventional techniques. The majority (65.1%) of polyadenylated viral RNA transcription is committed to producing four noncoding transcripts (RNA2.7, RNA1.2, RNA4.9, and RNA5.0) that do not substantially overlap designated protein-coding regions. Additional noncoding RNAs that are transcribed antisense to protein-coding regions map throughout the genome and account for 8.7% of transcription from these regions. RNA splicing is more common than recognized previously, which was evidenced by the identification of 229 potential donor and 132 acceptor sites, and it affects 58 protein-coding genes. The great majority (94) of 96 splice junctions most abundantly represented in the deep-sequencing data was confirmed by RT-PCR or RACE or supported by involvement in alternative splicing. Alternative splicing is frequent and particularly evident in four genes (RL8A, UL74A, UL124, and UL150A) that are transcribed by splicing from any one of many upstream exons. The analysis also resulted in the annotation of four previously unrecognized protein-coding regions (RL8A, RL9A, UL150A, and US33A), and expression of the UL150A protein was shown in the context of HCMV infection. The overall conclusion, that HCMV transcription is complex and multifaceted, has implications for the potential sophistication of virus functionality during infection. The study also illustrates the key contribution that deep sequencing can make to the genomics of nuclear DNA viruses.

Human cytomegalovirus transcriptome activity differs during replication in human fibroblast, epithelial and astrocyte cell lines.

Broad cell tropism contributes to the pathogenesis of human cytomegalovirus (HCMV), but the extent to which cell type influences HCMV gene expression is unclear. A bespoke HCMV DNA microarray was used to monitor the transcriptome activity of the low passage Merlin strain of HCMV at 12, 24, 48 and 72 h post-infection, during a single round of replication in human fetal foreskin fibroblast cells (HFFF-2s), human retinal pigmented epithelial cells (RPE-1s) and human astrocytoma cells (U373MGs). In order to correlate transcriptome activity with concurrent biological responses, viral cytopathic effect, growth kinetics and genomic loads were examined in the three cell types. The temporal expression pattern of viral genes was broadly similar in HFFF-2s and RPE-1s, but dramatically different in U373MGs. Of the 165 known HCMV protein-coding genes, 41 and 48 were differentially regulated in RPE-1s and U373MGs, respectively, compared with HFFF-2s, and 22 of these were differentially regulated in both RPE-1s and U373MGs. In RPE-1s, all differentially regulated genes were downregulated, but, in U373MGs, some were down- and others upregulated. Differentially regulated genes were identified among the immediate-early, early, early late and true-late viral gene classes. Grouping of downregulated genes according to function at landmark stages of the replication cycle led to the identification of potential bottleneck stages (genome replication, virion assembly, and virion maturation and release) that may account for cell type-dependent viral growth kinetics. The possibility that cell type-specific differences in expressed cellular factors are responsible for modulation of viral gene expression is discussed.

Sequences of complete human cytomegalovirus genomes from infected cell cultures and clinical specimens.

We have assessed two approaches to sequencing complete human cytomegalovirus (HCMV) genomes (236 kbp) in DNA extracted from infected cell cultures (strains 3157, HAN13, HAN20 and HAN38) or clinical specimens (strains JP and 3301). The first approach involved amplifying genomes from the DNA samples as overlapping PCR products, sequencing these by the Sanger method, acquiring reads from a capillary instrument and assembling these using the Staden programs. The second approach involved generating sequence data from the DNA samples by using an Illumina Genome Analyzer (IGA), processing the filtered reads by reference-independent (de novo) assembly, utilizing the resulting sequence to direct reference-dependent assembly of the same data and finishing by limited PCR sequencing. Both approaches were successful. In particular, the investigation demonstrated the utility of IGA data for efficiently sequencing genomes from clinical samples containing as little as 3 % HCMV DNA. Analysis of the genome sequences obtained showed that each of the strains grown in cell culture was a mutant. Certain of the mutations were shared among strains from independent clinical sources, thus suggesting that they may have arisen in a common ancestor during natural infection. Moreover, one of the strains (JP) sequenced directly from a clinical specimen was mutated in two genes, one of which encodes a proposed immune-evasion function, viral interleukin-10. These observations imply that HCMV mutants exist in human infections.

Sequential mutations associated with adaptation of human cytomegalovirus to growth in cell culture.

Mutations that occurred during adaptation of human cytomegalovirus to cell culture were monitored by isolating four strains from clinical samples, passaging them in various cell types and sequencing ten complete virus genomes from the final passages. Mutational dynamics were assessed by targeted sequencing of intermediate passages and the original clinical samples. Gene RL13 and the UL128 locus (UL128L, consisting of genes UL128, UL130 and UL131A) mutated in all strains. Mutations in RL13 occurred in fibroblast, epithelial and endothelial cells, whereas those in UL128L were limited to fibroblasts and detected later than those in RL13. In addition, a region containing genes UL145, UL144, UL142, UL141 and UL140 mutated in three strains. All strains exhibited numerous mutations in other regions of the genome, with a preponderance in parts of the inverted repeats. An investigation was carried out on the kinetic growth yields of viruses derived from selected passages that were predominantly non-mutated in RL13 and UL128L (RL13+UL128L+), or that were largely mutated in RL13 (RL13-UL128L+) or both RL13 and UL128L (RL13-UL128L-). RL13-UL128L- viruses produced greater yields of infectious progeny than RL13-UL128L+ viruses, and RL13-UL128L+ viruses produced greater yields than RL13+UL128L+ viruses. These results suggest strongly that RL13 and UL128L exert at least partially independent suppressive effects on growth in fibroblasts. As all isolates proved genetically unstable in all cell types tested, caution is advised in choosing and monitoring strains for experimental studies of vulnerable functions, particularly those involved in cell tropism, immune evasion or growth temperance.

High-throughput sequence analysis of variants of human cytomegalovirus strains Towne and AD169.

The genomes of commonly used variants of human cytomegalovirus (HCMV) strains Towne and AD169 each contain a substantial mutation in which a region (U(L)/b') at the right end of the long unique region has been replaced by an inverted duplication of a region from the left end of the genome. Using high-throughput technology, we have sequenced HCMV strain Towne (ATCC VR-977) and confirmed the presence of two variants, one exhibiting the replacement in U(L)/b' and the other intact in this region. Both variants are mutated in genes RL13, UL1, UL40, UL130, US1 and US9. We have also sequenced a novel AD169 variant (varUC) that is intact in U(L)/b' except for a small deletion that affects genes UL144, UL142, UL141 and UL140. Like other AD169 variants, varUC is mutated in genes RL5A, RL13, UL36 and UL131A. A subpopulation of varUC contains an additional deletion affecting genes IRS1, US1 and US2.

Genotypic analysis of two hypervariable human cytomegalovirus genes.

Most human cytomegalovirus (HCMV) genes are highly conserved in sequence among strains, but some exhibit a substantial degree of variation. Two of these genes are UL146, which encodes a CXC chemokine, and UL139, which is predicted to encode a membrane glycoprotein. The sequences of these genes were determined from a collection of 184 HCMV samples obtained from Africa, Australia, Asia, Europe, and North America. UL146 is hypervariable throughout, whereas variation in UL139 is concentrated in a sequence encoding a potentially highly glycosylated region. The UL146 sequences fell into 14 genotypes, as did all previously reported sequences. The UL139 sequences grouped into 8 genotypes, and all previously reported sequences fell into a subset of these. There were minor differences among continents in genotypic frequencies for UL146 and UL139, but no clear geographical separation, and identical nucleotide sequences were represented among communities distant from each other. The frequent detection of multiple genotypes indicated that mixed infections are common. For both genes, the degree of divergence was sufficient to preclude reliable sequence alignments between genotypes in the most variable regions, and the mode of evolution involved in generating the genotypes could not be discerned. Within genotypes, constraint appears to have been the predominant mode, and positive selection was detected marginally at best. No evidence was found for linkage disequilibrium. The emerging scenario is that the HCMV genotypes developed in early human populations (or even earlier), becoming established via founder or bottleneck effects, and have spread, recombined and mixed worldwide in more recent times.

Fundamental and accessory systems in herpesviruses.

Evolutionary studies have a large theoretical component and will not directly provide therapies for herpesvirus infections. However, they do provide a conceptual framework within which we can evaluate the origins of the various systems that contribute to viral lifestyle. An evolutionary context allows ancient systems that are fundamental to the replication of all herpesviruses to be distinguished from those that have developed relatively recently in order to tailor viruses to particular biological niches. Both categories are in principle accessible to intervention, either to prevent basic replicative capabilities or to reduce the advantages that the virus has in its interactions with the host. Phylogenetic data provide estimates of evolutionary rate for herpesviruses that are only between one and two orders of magnitude greater than those of their hosts. However, it is becoming apparent that certain genes have evolved much faster under selection pressures and by mechanisms that are not well understood. Nonetheless, the mutation rates of even the most highly conserved genes are sufficient to permit herpesviruses to escape from antiviral therapy. Greater understanding of the origins and functions of herpesvirus genes may lead to new insights into the determinants of pathogenesis and hence to new diagnostic and therapeutic targets.

Reconstruction of the complete human cytomegalovirus genome in a BAC reveals RL13 to be a potent inhibitor of replication.

Human cytomegalovirus (HCMV) in clinical material cannot replicate efficiently in vitro until it has adapted by mutation. Consequently, wild-type HCMV differ fundamentally from the passaged strains used for research. To generate a genetically intact source of HCMV, we cloned strain Merlin into a self-excising BAC. The Merlin BAC clone had mutations in the RL13 gene and UL128 locus that were acquired during limited replication in vitro prior to cloning. The complete wild-type HCMV gene complement was reconstructed by reference to the original clinical sample. Characterization of viruses generated from repaired BACs revealed that RL13 efficiently repressed HCMV replication in multiple cell types; moreover, RL13 mutants rapidly and reproducibly emerged in transfectants. Virus also acquired mutations in genes UL128, UL130, or UL131A, which inhibited virus growth specifically in fibroblast cells in wild-type form. We further report that RL13 encodes a highly glycosylated virion envelope protein and thus has the potential to modulate tropism. To overcome rapid emergence of mutations in genetically intact HCMV, we developed a system in which RL13 and UL131A were conditionally repressed during virus propagation. This technological advance now permits studies to be undertaken with a clonal, characterized HCMV strain containing the complete wild-type gene complement and promises to enhance the clinical relevance of fundamental research on HCMV.

The products of human cytomegalovirus genes UL23, UL24, UL43 and US22 are tegument components.

We have investigated the human cytomegalovirus (HCMV) US22 gene family members UL23, UL24, UL43 and US22. Specific antibodies were generated to identify pUL23 (33 kDa), pUL24 (40 kDa) and pUL43 (48 kDa), while pUS22 was identified by monoclonal antibody HWLF1. A C-terminally truncated UL43 product (pUL43t; 21 kDa) produced by a deletion mutant was also investigated. The UL24 and UL43 genes were expressed with early-late (gamma1) and true-late (gamma2) kinetics, respectively. Immunoblot and immuno-EM studies demonstrated that pUL23, pUL24, pUL43 and pUS22 were virion tegument components. Immunofluorescence and immuno-EM studies showed that pUL23, pUL24, pUL43 and pUL43t were located in cytoplasmic protein aggregates, manifesting two forms: complex juxtanuclear structures and smaller, membrane-bound aggregates resembling dense bodies. The complex-type aggregate is a putative site of particle maturation. Because pUL43t was present in protein aggregates, but under-represented in virus particles compared to pUL43, it was concluded that N-terminal sequences target pUL43 to protein aggregates and that C-terminal sequences are important for incorporation into particles. Since three other US22 family products (pUL36, pTRS1 and pIRS1) are documented tegument components, at least seven of the twelve US22 family genes encode tegument proteins, suggesting that the products of the remaining five genes might be similarly located. These findings demonstrate a common biological feature among most, if not all, US22 family proteins and implicate the family in events occurring immediately after virus penetration.

The human cytomegalovirus genome revisited: comparison with the chimpanzee cytomegalovirus genome.

The gene complement of wild-type human cytomegalovirus (HCMV) is incompletely understood, on account of the size and complexity of the viral genome and because laboratory strains have undergone deletions and rearrangements during adaptation to growth in culture. We have determined the sequence (241 087 bp) of chimpanzee cytomegalovirus (CCMV) and have compared it with published HCMV sequences from the laboratory strains AD169 and Toledo, with the aim of clarifying the gene content of wild-type HCMV. The HCMV and CCMV genomes are moderately diverged and essentially collinear. On the basis of conservation of potential protein-coding regions and other sequence features, we have discounted 51 previously proposed HCMV ORFs, modified the interpretations for 24 (including assignments of multiple exons) and proposed ten novel genes. Several errors were detected in the published HCMV sequences. We presently recognize 165 genes in CCMV and 145 in AD169; this compares with an estimate of 189 unique genes for AD169 made in 1990. Our best estimate for the complement of wild-type HCMV is 164 to 167 genes.

Homology between the human cytomegalovirus RL11 gene family and human adenovirus E3 genes.

A significant proportion of the human cytomegalovirus (HCMV) genome comprises 12 multigene families that probably arose by gene duplication. One, the RL11 family, contains 12 members, most of which are predicted to encode membrane glycoproteins. Comparisons of sequences near the left end of the genome in several HCMV strains revealed two adjacent open reading frames that potentially encode related proteins: RL6, which is hypervariable, and RL5A, which has not been recognized previously. These genes potentially encode a domain that is the hallmark of proteins encoded by the RL11 family, and thus constitute two new members. A homologous domain is also present in a subset of human adenovirus E3 membrane glycoproteins. Evolution of genes specifying the shared domain in cytomegaloviruses and adenoviruses is characterized by extensive divergence, gene duplication and selective sequence loss. These features prompt speculation about the roles of these genes in the two virus families.

Two novel spliced genes in human cytomegalovirus.

Two novel spliced genes (UL131A and UL128) flanking UL130 were predicted from sequence comparisons between human cytomegalovirus (HCMV) and its closest known relative, chimpanzee cytomegalovirus (CCMV), and the splicing patterns were confirmed by mRNA mapping experiments. Both genes were transcribed with late kinetics and shared a polyadenylation site. Comparisons with wild-type HCMV in infected human tissues showed that three of five isolates passaged in cell culture contained disruptions of UL128, one was frameshifted in UL131A and one exhibited a deletion affecting UL131A and UL130. CCMV and the Colburn strain of simian cytomegalovirus, which have been passaged in cell culture, also exhibit disruptions of UL128. These observations indicate that expression of either one of UL128 and UL131A is deleterious to growth of primate cytomegaloviruses in cell culture. Although the functions of these genes are unknown, sequence comparisons suggest that UL128 encodes a beta-chemokine.

Genetic content of wild-type human cytomegalovirus.

The genetic content of wild-type human cytomegalovirus was investigated by sequencing the 235 645 bp genome of a low passage strain (Merlin). Substantial regions of the genome (genes RL1-UL11, UL105-UL112 and UL120-UL150) were also sequenced in several other strains, including two that had not been passaged in cell culture. Comparative analyses, which employed the published genome sequence of a high passage strain (AD169), indicated that Merlin accurately reflects the wild-type complement of 165 genes, containing no obvious mutations other than a single nucleotide substitution that truncates gene UL128. A sizeable subset of genes exhibits unusually high variation between strains, and comprises many, but not all, of those that encode proteins known or predicted to be secreted or membrane-associated. In contrast to unpassaged strains, all of the passaged strains analysed have visibly disabling mutations in one or both of two groups of genes that may influence cell tropism. One comprises UL128, UL130 and UL131A, which putatively encode secreted proteins, and the other contains RL5A, RL13 and UL9, which are members of the RL11 glycoprotein gene family. The case in support of a lack of protein-coding potential in the region between UL105 and UL111A was also strengthened.