Recombinant origin and interspecies transmission of a HERV-K(HML-2)-related primate retrovirus with a novel RNA transport element

Abstract

HERV-K(HML-2), the youngest clade of human endogenous retroviruses (HERVs), includes many intact or nearly intact proviruses, but no replication competent HML-2 proviruses have been identified in humans. HML-2-related proviruses are present in other primates, including rhesus macaques, but the extent and timing of HML-2 activity in macaques remains unclear. We have identified 145 HML-2-like proviruses in rhesus macaques, including a clade of young, rhesus-specific insertions. Age estimates, intact open reading frames, and insertional polymorphism of these insertions are consistent with recent or ongoing infectious activity in macaques. 106 of the proviruses form a clade characterized by an ~750 bp sequence between env and the 3' long terminal repeat (LTR), derived from an ancient recombination with a HERV-K(HML-8)-related virus. This clade is found in Old World monkeys (OWM), but not great apes, suggesting it originated after the ape/OWM split. We identified similar proviruses in white-cheeked gibbons; the gibbon insertions cluster within the OWM recombinant clade, suggesting interspecies transmission from OWM to gibbons. The LTRs of the youngest proviruses have deletions in U3, which disrupt the Rec Response Element (RcRE), required for nuclear export of unspliced viral RNA. We show that the HML-8-derived region functions as a Rec-independent constitutive transport element (CTE), indicating the ancestral Rec-RcRE export system was replaced by a CTE mechanism.

Keywords: RNA export; endogenous retrovirus; evolutionary biology; human; infectious disease; microbiology; recombination; rhesus macaque; viral evolution; viruses; zoonosis.

Figure 1.. Structural features of rhesus macaque HML-2 proviruses.

(A) Overall structure of selected rhesus proviruses compared to consensus human-specific HML-2. HERVK11 HML-8-derived region shown in purple, with light purple indicating the HERVK11-derived region of env and dark purple indicating the long terminal repeat (LTR)-derived MER11 element. Intact open reading frames (ORFs) shown in bold, defective or missing ORFs indicated by grayed out font. Proviruses were selected to show notable structural features such as intact ORFs, shared internal deletions, and distinct LTR types. Deletions or gaps are indicated by dotted lines. (B) Structural evolution of recombinant region, including parental HML-2 and HML-8 regions for comparison. 1: recombinant region in oldest proviruses, with no MER11 deletion and intact env-derived non-coding sequence; 2: younger recombinants with MER11 deletion; 3: youngest recombinants with MER11 deletion and deletion of env-derived non-coding sequence outside the polypurine tract (PPT). Bright orange = HML-2-derived env coding sequence; pale orange = HML-2 env-derived non-coding sequence (including PPT); pale purple = HML-8-derived env coding sequence; dark purple = HML-8 LTR-derived non-coding sequence (MER11 element). (C). SERV-K/MER11 LTR types, showing the accumulation of deletions in U3 in younger LTRs compared to the ancestral SERV-K/MER11 LTR. U3, R, and U5 regions of LTRs indicated by dark gray, black, and light gray coloration, respectively.

Figure 1—figure supplement 1.. Origin of SERV-K/MER11 via RT-mediated recombination between co-packaged HERV-K (HML-2) and HERVK11 (HML-8) genomic RNAs.

(A) Recombination via two RT jumps during minus strand synthesis. (B) Sequence alignment of recombination breakpoints, comparing a reconstructed ancestral recombinant SERV-K/MER11 (SKM11anc) with the consensus human HML-2 HERV-Kcon and a rhesus macaque HML-8 sequence (rheMac10 chr2:185106142–185107371). Percent nucleotide sequence identity of HML-2 and HML-8 sequences to SKM11anc is noted to the right of each line. Red box in downstream junction outlines possible microhomology between HML-2 and HML-8 sequences.

Figure 2.. Phylogeny of SERV-K/MER11 pol and env.

Maximum likelihood phylogenies for SERV-K/MER11 pol (A) and env (B) with 1000 bootstrap replicates. SERV-K/MER11 sequences are in blue. In black, sequences from non-recombinant rhesus HML-2, human HML-2 LTR5Hs, human HML-2 LTR5A included for comparison, along with MMTV pol or env as outgroups. Nodes colored according to bootstrap replicate percentages; basal node of SERV-K/MER11 clade marked with exact bootstrap value. The trees are drawn to scale, with branch lengths measured in the number of substitutions per site.

Figure 3.. Phylogeny of SERV-K/MER11 long terminal repeats (LTRs) and LTR-based age estimates.

(A) Maximum likelihood phylogeny of rhesus macaque HML-2 LTRs with 100 bootstrap replicates. SERV-K/MER11 clade colored blue, and shaded according to species specificity; dark blue proviruses are found in all Old World monkeys (OWM), light blue are cercopithecine specific, and royal blue are rhesus macaque specific. One clade of non-recombinant rhesus proviruses has been collapsed for clarity. Nodes colored according to bootstrap replicate values. Δ1, Δ2, Δ1 + Δ2, and Δ3 mark estimated last common ancestor for each LTR deletion or combination of LTR deletions (see Figure 1C). (B) Proviral integration times estimated using the sequence divergence between cognate 5′ and 3′ LTRs. Proviruses with no differences between their LTRs were plotted at 250,000 years. Red and green shaded areas mark range of estimated ages for the last common ancestors of the OWM and Catarrhine primates, respectively. SKM11 = SERV K/MER11. (C). Proviral integration times in the past 5 million years. Estimated ages of human and gorilla-specific HML-2s included for reference.

Figure 3—figure supplement 1.. Allele-specific PCR screening.

(A) 3 primer allele-specific PCR for amplifying both full-length provirus, solo long terminal repeat (LTR), and pre-integration empty site alleles for specific proviral insertions. Primers specific to the 5′ and 3′ flanking sequences for 23 proviruses were designed, and used along with a primer specific to the 5′ UTR of SERV-K/MER11 proviruses to screen 14 rhesus macaque (RM) and 2 crab-eating macaque (CEM) genomic DNA samples for the presence or absence of proviral insertions. Examples of PCR screening results for 19-18145532_RM10 (B) and SERV-K1 (C). Gel electrophoresis of PCR products for each sample, with proviral amplicons outlined in blue and empty site amplicons outlined in red. No solo LTR alleles were identified for these proviruses.

Figure 4.. Structures of gibbon proviruses.

Structure of gibbon SERV-K/MER11 provirus compared to rhesus SERV-K/MER11 and human HML-2. HERVK11 HML-8-derived region shown in purple, with light purple indicating the HERVK11-derived region of env and dark purple indicating the long terminal repeat (LTR)-derived MER11 element. Intact open reading frames (ORFs) shown in bold, defective or missing ORFs indicated by grayed out font. Deletions or gaps in alignment shown by dotted lines.

Figure 5.. Phylogeny and long terminal repeat (LTR) deletions of gibbon and golden snub-nosed monkey (GSM) SERV-K/MER11 proviruses.

(A, B) Pol and LTR maximum likelihood phylogenies of gibbon and GSM SERV-K/MER11 proviruses with 100 bootstrap replicates. Rhesus SERV-K/MER11 and non-recombinant proviruses were included for comparison; large clades of rhesus proviruses collapsed for clarity. Nodes colored by bootstrap confidence level, branches colored by provirus type and species specificity. Black = non-recombinant rhesus, dark blue = OWM SERV-K/MER11, light blue = cercopithecine specific, royal blue = rhesus specific, purple = GSM, red = gibbon. (C) Schematic of deletions in U3 of rhesus, GSM and gibbon SERV-K/MER11 LTRs, compared with a human HML-2 full-length 968 bp LTR (HERV-Kcon). Dotted lines mark deletions relative to the HERV-Kcon LTR. Approximate location of 433 bp HML-2 Rec Response Element (RcRE) noted by red box.

Figure 5—figure supplement 1.. 10-73510510 provirus is the result of intra- or interchromosomal recombination.

(A) Screenshot of chromosome 10 of the northern white-cheeked gibbon genome in the UCSC Genome Browser, demonstrating synteny between chr10 and human chr12 and chr19. This is a visualization of the alignment of human chr12 and chr19 to gibbon chr10 in the hg19 Chain/Net track. Purple regions align with chr19, and green regions align with chr12. Red boxes outline regions containing two SERV-K/MER11 insertions. (B) Regions containing proviral insertions expanded to show local genomic context, with flanking regions aligning to different human chromosomes. (C) SERV-K/MER11 solo long terminal repeat (LTR) 10-5132425_gib and provirus 10-73510510_gib, with flanking sequence. (D) Mismatched target site duplications of 10-5132425_gib and 10-73510510_gib. (E) Potential recombination scenario: chr10 of the white-cheeked gibbon initially composed of two regions, one syntenic to human chr19, and one to human chr12. Homologous recombination between two proviruses in the opposite orientation to each other, one in the chr19 region and one in the chr12 region, leads to a reciprocal translocation on chr10, resulting in the current structure, with two regions syntenic to chr19 and two to chr12.

Figure 6.. SERV-K/MER11 chimeric rec-like transcript and Rec response element deletions.

(A) Comparison of human HML-2 rec transcript to SERV-K/MER11 sRec transcript. Both are doubly spliced sub-genomic transcripts. The second coding exon of sRec is encoded by the HML-8-derived region, in purple, with light purple denoting the HML-8 env-derived region, and dark purple the HML-8 long terminal repeat (LTR)-derived region. The HML-2 Rec Response Element (RcRE) in the U3 region of the LTR is also marked. (B) HML-2 Rec protein compared to chimeric SERV-K/MER11 sRec. Red region of sRec is homologous to Rec, with black bars denoting amino acid differences. Purple region is the 62 aa HML-8-derived exon, with black bars again showing amino acid differences.

Figure 6—figure supplement 1.. Identification of rec-like transcript from rhesus macaque RNAseq data.

(A) Diagram of SERV-K/MER11 provirus, full-length genomic RNA and env and sRec sub-genomic mRNAs. (B) Screen shot of the Integrative Genomics Viewer showing rhesus macaque induced pluripotent stem cell (iPSC) RNAseq reads aligned to the SERV-K1 genome. Red highlighted regions mark the locations of two large deletions shared by multiple proviruses. Dotted lines mark the splice acceptors and donors for the two introns that are spliced out to make sRec. Multiple reads cross both splice junctions. (C) Examples of reads crossing the second splice junction between the two coding exons of sRec. Exon 1 in black, exon 2 in red, with intronic portions of splice site bolded in black.

Figure 7.. Constitutive transport element (CTE) activity of HERVK11 HML-8-derived region in SERV-K/MER11.

(A) Schematics of dual color lentiviral reporter and transport elements tested. Base construct is an NL4-3 HIV provirus modified to express eGFP from an unspliced transcript and mCherry from a fully spliced transcript. Transport element of interest replaces RRE. SD = splice donor, SA = splice acceptor. HML-8-derived region was tested for unspliced RNA transport function by transfection with and without sRec; dual HERV-Kcon Rec Response Element (RcRE) with and without Rec was used as a control for RcRE-like activity, and MPMV CTE as a control for CTE-like activity. (B) Fluorescent imaging of transfected cells. mCherry = transport element-independent signal, eGFP = transport element-dependent signal. Scale bar = 400 μm. (C) Quantification of RNA transport activity using flow cytometry. eGFP and mCherry mean fluorescent intensity of transfected cell populations were measured, and the ratio of eGFP/mCherry plotted for each construct as a measure of RNA transport activity. The mean and standard deviation of three replicates are plotted for each condition. See Figure 7—figure supplement 1 for gating strategy and flow dot plots for each condition.

Figure 7—figure supplement 1.. Constitutive transport element (CTE) activity of MER11 element with SERV-K/MER11 U3R and without HML-8 env region.

(A) Schematics of dual color lentiviral reporter and transport elements tested as in Figure 7. Three SERV-K/MER11 constructs were tested for unspliced RNA transport function by transfection with and without Rec or sRec. Dual HML-2 Rec Response Element (RcRE) and MPMV CTE as controls. (B) Fluorescent imaging of transfected cells. mCherry = transport element-independent signal, eGFP = transport element-dependent signal. Scale bar = 400 μm. (C) Gating scheme for quantification of RNA transport activity using flow cytometry. Cells were first gated on a forward scatter vs side scatter plot, followed by gating for single cells using forward scatter height vs forward scatter area. Lastly, untransfected cells were used to gate out the eGFP mCherry double negative population; the eGFP+, mCherry+, and eGFP+mCherry+ populations were combined for further analysis. (D) eGFP vs mCherry plots for each experimental condition, showing the gate used to define the combined single and double positive populations. (E) RNA transport activity of each element. eGFP and mCherry mean fluorescent intensity of transfected cell populations were measured, and the ratio of eGFP/mCherry plotted for each construct as a measure of RNA transport activity. The mean and standard deviation of three replicates are plotted for each condition.

Figure 8.. MER11 constitutive transport element (CTE) can functionally replace the MPMV CTE in the context of a single round vesicular stomatitis virus glycoprotein G (VSV-G) pseudotyped viral infection.

(A) MPMV proviral constructs expressing eGFP in place of Env, with either the wild-type MPMV CTE (wtCTE), no CTE (ΔCTE), or the MER11 CTE replacing wtCTE (MER11), were transfected into HEK293T cells with and without VSV-G. Viral supernatants were harvested 72 hr after transfection and used to infect HEK293T target cells. Three infections were performed, each in triplicate. Infectivity was assayed after 72 hr via fluorescent imaging (B) with one representative image shown per condition, and flow cytometry (C) using % GFP expressing cells as a measure of infectivity. For images, the scale bar = 400 μm. For flow cytometry, each data point represents one experimental replicate, with different shapes corresponding to different infection rounds.

Figure 8—figure supplement 1.. Fluorescent imaging of producer cells for infection assay.

MPMV proviral constructs expressing eGFP in place of Env, with either the wild-type MPMV constitutive transport element (wtCTE), no CTE (ΔCTE), or the MER11 CTE replacing wtCTE (MER11), were transfected into HEK293T cells with and without vesicular stomatitis virus glycoprotein G (VSV-G). Cells were imaged 72 hr after infection and viral supernatant was harvested for infections. Scale bar = 400 μm.