Comprehensive Analysis of HERV Transcriptome in HIV+ Cells: Absence of HML2 Activation and General Downregulation of Individual HERV Loci

Abstract

Human endogenous retrovirus (HERV) expression is currently studied for its possible activation by HIV infection. In this context, the HERV-K(HML2) group is the most investigated: it has been proposed that HIV-1 infection can prompt HML2 transcription, and that HML2 proteins can affect HIV-1 replication, either complementing HIV or possibly influencing antiretroviral therapy. However, little information is available on the expression of other HERV groups in HIV infection. In the present study, we used a bioinformatics pipeline to investigate the transcriptional modulation of approximately 3250 well-characterized HERV loci, comparing their expression in a public RNA-seq profile, including a HIV-1-infected and a control T cell culture. In our pilot study, we found approximately 200 HERV loci belonging to 35 HERV groups that were expressed in one or both conditions, with transcripts per million (TPM) values from 1 to >500. Intriguingly, HML2 elements constituted only the 3% of expressed HERV loci, and in most cases (160) HERV expression was downregulated in the HIV-infected culture, showing from a 1- to 14-fold decrease as compared to uninfected cells. HERV transcriptome has been inferred de novo and employed to predict a total of about 950 HERV open reading frames (ORFs). These have been validated according to the coding potential and estimated abundance of the corresponding transcripts, leading to a set of 57 putative proteins potentially encoded by 23 HERV loci. Analysis showed that some individual loci have a coding potential that deserves further investigation. Among them, a HML6 provirus at locus 19q13.43 was predicted to produce a transcript showing the highest TPM among HERV-derived transcripts, being upregulated in HIV+ cells and inferred to produce Gag and Env puteins with possible biological activity.

Keywords: HERV; HERV expression; HIV; HML2; RNAseq; endogenous retroviruses.

Figure 1

Graphical overview of the bioinformatics pipeline and the identified expressed human endogenous retrovirus (HERV) loci and transcripts. Modulation is based on at least a 3-fold change in the transcripts per million (TPM) value between the two conditions.

Figure 2

Landscape of HERV loci expressed (TPM value ≥ 0.1) in one or both conditions (HIV- and/or HIV+ T cells), divided in taxonomic viral groups. The number of expressed members (left axis) as well as their percentage with respect to the group total members (right axis) are reported.

Figure 3

Overview of the individual HERV loci modulated in the presence of HIV infection. (A) Number of HERV loci for each group; (B) Genomic context of integration: for intragenic HERVs (yellow) the sense (s) or antisense (as) orientation with respect to the surrounding gene is also indicated.

Figure 4

Focus on the puteins encoded by HML6 locus 4796. Puteins and the originating transcripts have been aligned to the locus genomic position, annotated with the retroviral genes as previously identified [43]. The presence of human genes (Gencode version 32) and repetitive elements (RepBase) is also indicated. The puteins overlapping with reported exons are highlighted with a red rectangle.

Figure 5

HERV puteins alignment with the respective consensus sequences from Dfam and Vargiu et al. [9]. In the latter, conserved domains have been annotated: dark yellow Z = Zinc finger motif, gp = glycoprotein, HR1-HR2_TM = transmembrane heptad repeats 1 and 2, RT = reverse transcriptase, light yellow z and D = zinc- and DNA-binding domains, respectively. Colored residues indicate divergent amino acids as compared to the reference, which is highlighted by the yellow bar.