Towler, James Charles
Transcriptome activity of human cytomegalovirus (strain Merlin) in fibroblasts, epithelial cells and astrocytes.
PhD thesis, University of Glasgow.
Full text available as:
Global human cytomegalovirus (HCMV) gene expression was investigated during replication in three permissive human cell types in tissue culture; human foetal foreskin fibroblasts (HFFF-2), human retinal pigmented epithelial (RPE) cells, and human astrocytoma cells (U373Mg). A custom HCMV DNA microarray based on recent re-assessments of HCMV coding potential was designed. The Merlin strain of HCMV was used for these studies because it has been reported to contain the complete set of ORFs, only one of which (UL128) is mutated. The UL128 gene locus is invariably mutated in HCMV isolates propagated in fibroblasts, and the premature termination mutation of UL128 greatly enhanced infectious yields from Merlin passaged in HFFF-2 cells. The HCMV (Merlin) microarray consists of 60-mer oligonucleotide probes derived from both 3’- and 5’-proximal regions of each recognised ORF, and 3’-proximal probes for novel ORFs that have been proposed from in silico studies. Probes were also included for several previously reported ORFs that are now considered to non-protein coding. Positive and negative-sense bacterial sequence probes were included on the array, and were used in conjuction with spiked-in cognate RNAs in the cDNA synthesis reaction as controls for normalisation. The quality of the printed HCMV microarray was assessed and its specificity validated using Cy3-labelled cDNAs prepared from total RNA extracted from mock-infected and HCMV-infected cells at 96 h PI. Hybridisation conditions were then investigated to achieve optimal specificity and sensitivity of cDNA binding to cognate probes on the array.
Prior to commencing the microarray studies, the growth characteristics of HCMV (Merlin) were compared in each of the three cell lines. One-step virus growth curves revealed differential virus replication kinetics in the three cell types. Compared to infected HFFF-2 cultures, there was a 24 h delay in exit of virus from the viral eclipse phase in RPE and U373Mg cells and little or no release of infectivity to the extracellular medium. Differential growth kinetics in the three cell types were not due to differences in the ability of HCMV to enter cells or to induce virus gene expression, nor was it due to gross differences in the temporal expression kinetics of immediate-early, early and late protein synthesis in the three cell types. However, differences in the amount of protein made were evident, with viral protein expression lowest in RPE cells. Differences in virus growth kinetics were probably due to differences in the numbers of virus particles assembled, and/or their maturation and egress. It was therefore considered valid to compare the temporal kinetics of HCMV (Merlin) transcript expression in the three cell types in order to identify genes that were regulated differently.
cDNAs were prepared from total RNA extracted from infected cells at 12, 24, 48 and 72 h PI, or at 72 h PI only from mock-infected cultures. After collection of the raw data, receiver operating characteristic (ROC) analysis was performed using the positive and negative control signals in order to determine cutoff points that discriminated between true-positive and true-negative hybridisation signals. The GeneSpring gene expression analysis software was used for statistical analysis of the microarray data. Human fibroblast cells have been extensively used in HCMV research, and so the HFFF-2 cell type was used as a reference cell type for the microarray work. Differential expression of a virus gene then relates to differences in amounts and/or expression kinetics between HFFF-2 and RPE cells, or between HFFF-2 and U373Mg cells. To identify differentially expressed virus genes, combined statistical tests were performed on the mean expression value for each gene from all data points over the time course, giving a single expression value for each gene in each cell type. The statistical tests then compared the expression values for individual HCMV genes in infected HFFF-2 cells against the corresponding expression values for individual HCMV genes in RPE or U373Mg cells.
Comparing the microarray data from HFFF-2 and RPE infected cells, 13 HCMV ORFs (UL4, UL16, UL45, UL148, IRS1, US11, US12, US13, US14, US15, US18, US19, US20), were found to be differentially expressed, and this was confirmed by examination of the expression kinetics of the individual genes. When the microarray data from infected HFFF and U373Mg cells were compared, 26 ORFs appeared to be differentially expressed. However, the microarray showed that late HCMV genes were expressed at unusually early times (24 h PI) in U373Mg cells; in contrast to the expression of their protein products. The data suggested that in U373Mg cells, either the HCMV transcriptome cascade was completed more rapidly, or that there was a breakdown in regulation of transcription control. Most of the 26 ORFs differentially expressed in HFFF-2 and U373Mg cells (identified by combined statistical testing) are expressed at a significantly higher level in U373Mg cells, but 7 were made in significantly reduced amounts (UL4, IRS1, US12, US14, US18, US19 and US20), and these were considered more likely candidates for differential expression, and were also differentially expressed in RPE cells. The disregulation of transit through the HCMV transcript cascade in U373Mg cells is thought to be due to the fact that p53 is mutated in this cell line. It has been reported that p53 mutants including the p53 mutation in U373Mg cells are capable of activating transcription from the HCMV MIEP and that the minimal promoter sequence is a TATA box. It may be that mutant p53 activates HCMV early and late promoters resulting in an accelerated transit through the HCMV transcription cascade in infected U373Mg cells.
Of the 13 ORFs differentially expressed in HFFF-2 and RPE cells, UL4 is reported to be under both transcriptional and translational control. Cellular and viral transcription factors are involved in both positive and negative regulation of the UL4 promoter making it a good candidate for differential regulation in different cell types, although the function provided by UL4 is unknown. The immediate-early IRS1 gene is an important viral transactivator required throughout the virus replication cycle, and also self regulates its gene expression though an internal ORF. The functions of the US12 family genes (US12, US13, US14, US15, US18, US19 and US20) are unknown, but it has been suggested that they have a role in virus particle tegumentation, envelopment, and egress from infected cells. However, the identification of multiple members of the US12 gene family as differentially regulated should be interpreted with caution, since US18, US19 and US20 produce 3’ co-terminal transcripts, and it is probable that other family members also share polyadenylation sites. Down-regulation of US12 family genes with potential roles in virus maturation and egress are consistent with the impaired release of virus to the extracellular medium from RPE and U373Mg cells. Compared to HFFF-2 cells, the down-regulation of UL16 and US11 in RPE and U373Mg cells is interesting since the genes have immune evasion functions, and these cell types are located in immune-privileged organs, i.e. the brain (astrocytes) and eye (retinal epithelia). As yet, it remains unclear why the expression of two immune evasion genes from a total of ten, are down-regulated in these cell types. The down-regulation of UL45 in RPE cells may play a role in virus dissemination in the eye. While the HCMV UL45 gene product is a component of the virus tegument and might supply an important function early in the virus replication cycle, a UL45 mutant exhibited a growth defect in fibroblasts that changed cell-cell spread characteristics. Interestingly, M45, the MCMV homologue of HCMV UL45, is reported as a determinant of endothelial cell tropism.
With respect to proposed novel ORFs identified by in silico analysis, in most cases, we found no evidence for transcript expression. Of those that gave positive hybridisation signals, most might be explained by overlapping transcripts from genes in the same region and coding in the same direction. Other novel ORFs lie within regions of the genome now considered to be non-coding, but where transcripts have previously been reported, while the remainder may represent genuine coding ORFs. The lack of signal for previously described ORFs that are now considered non-protein coding confirms their status as discounted genes.
In order to test the microarray system, the temporal expression kinetics of selected virus genes were investigated by alternative methods including; real-time PCR, and northern blots to check the identity of specific transcripts, and where antibodies were available, western immunoblotting to confirm the expression kinetics of specific proteins. The expression kinetics obtained for specific genes both differentially and non-differentially expressed by these various methods were entirely consistent with those obtained for the same genes with the HCMV microarray. It was concluded that the Merlin microarray system was a valid and reliable research tool for the investigation of HCMV gene expression.
Actions (login required)