Characterization of transcription of genomic regions harboring HERV-W elements

Abstract

Human endogenous retroviruses (HERV) comprise 8% of the human genome and can be classified into at least 31 families. A typical HERV provirus consists of internal gag, pol and env genes, flanked by two long terminal repeats (LTRs). HERV are by nature repetitive and have with few notable exceptions lost their protein-coding capacity. Therefore, HERV are consistently not included in large scale expression studies and hence little is known of their expression, regulation and potential functional significance. Aberrant expression of HERV-W has been associated with human diseases, such as multiple sclerosis and schizophrenia. It has been reported that HERV-W elements, including ERVWE1 (the so far only known HERV-W gene functionally adopted by the human host), can be transactivated in a range of human non-placental cell-lines during influenza virus infections.

Applying a recently developed technique for obtaining high resolution melting temperature analyses, transcripts containing HERV-W gag sequences were found to be expressed in non-random patterns with extensive variations in the expression between tissues, brain regions and individuals. Furthermore, the patterns of such transcripts varied more between individuals in brain regions than other tissues (I). To determine the effect of genomic context, viral structure and orientation on the transcription of HERV-W, PCR directed at specific HERV-W loci were employed on panel of normal human tissues. HERV-W elements in intronic regions were found to be expressed at higher levels than elements in intergenic regions. With regard to intronic elements, proviruses were expressed at higher levels than pseudoelements or solo LTRs. Relative to their corresponding genes, intronic elements integrated on the sense strand appeared to be transcribed at higher levels than those integrated on the anti-sense strand. Furthermore, the expression of transcripts containing intronic proviral elements appeared to be independent from that of their corresponding genes (II). When addressing mechanisms underlying transactivation of HERV-W following virus infection we detected up-regulation of spliced ERVWE1 transcripts and those encoding the transcription factor glial cells missing 1 (GCM1) which acts as an enhancer element upstream of ERVWE1. Knock-down of GCM1 by siRNA, followed by infection suppressed the transactivation of ERVWE1. In addition, ChIP assays detected decreased H3K9 trimethylation and histone methytransferase SETDB1 levels along with influenza viral proteins associated with ERVWE1 and other HERV-W loci in infected CCF-STTG1 cells (III). Finally, by analyzing publicly available RNA sequencing datasets generated from three different regions of human brains of multiple individuals, a consistent expression (0.1-0.2% of mappable reads) of HERV families was observed across three regions of brains. Spearman correlations between tissues revealed highly correlated expression levels across 475 consensus sequences. By mapping sequences aligned to the consensus sequences of HERV-W and HERV-H families to individual loci on chromosome 7, more than 60 loci from each family were identified, some of which are being transcribed. Elevated expressions of overall HERV, as well as of HERV-W family were observed in samples from both schizophrenia and bipolar disorder patients (IV).

In conclusion, our studies show that 1) HERV-W gag transcripts appear to exhibit a highly diversified expression pattern across both tissues and individuals; 2) Both LTR directed and leaky transcription of HERV-W elements contribute to their tissue-specific expression pattern; 3) Chromatin modifications potentially mediate the effect of influenza A virus infection on HERV-W expression; 4) An independent method of RNA sequencing verifies expression of HERV in human brain regions.