HIV-1 Interacts with Human Endogenous Retrovirus K (HML-2) Envelopes Derived from Human Primary Lymphocytes

ABSTRACT

Human endogenous retroviruses (HERVs) are viruses that have colonized the germ line and spread through vertical passage. Only the more recently acquired HERVs, such as the HERV-K (HML-2) group, maintain coding open reading frames. Expression of HERV-Ks has been linked to different pathological conditions, including HIV infection, but our knowledge on which specific HERV-Ks are expressed in primary lymphocytes currently is very limited. To identify the most expressed HERV-Ks in an unbiased manner, we analyzed their expression patterns in peripheral blood lymphocytes using Pacific Biosciences (PacBio) single-molecule real-time (SMRT) sequencing. We observe that three HERV-Ks (KII, K102, and K18) constitute over 90% of the total HERV-K expression in primary human lymphocytes of five different donors. We also show experimentally that two of these HERV-K env sequences (K18 and K102) retain their ability to produce full-length and posttranslationally processed envelope proteins in cell culture. We show that HERV-K18 Env can be incorporated into HIV-1 but not simian immunodeficiency virus (SIV) particles. Moreover, HERV-K18 Env incorporation into HIV-1 virions is dependent on HIV-1 matrix. Taken together, we generated high-resolution HERV-K expression profiles specific for activated human lymphocytes. We found that one of the most abundantly expressed HERV-K envelopes not only makes a full-length protein but also specifically interacts with HIV-1. Our findings raise the possibility that these endogenous retroviral Env proteins could directly influence HIV-1 replication.

IMPORTANCE Here, we report the HERV-K expression profile of primary lymphocytes from 5 different healthy donors. We used a novel deep-sequencing technology (PacBio SMRT) that produces the long reads necessary to discriminate the complexity of HERV-K expression. We find that primary lymphocytes express up to 32 different HERV-K envelopes, and that at least two of the most expressed Env proteins retain their ability to make a protein. Importantly, one of them, the envelope glycoprotein of HERV-K18, is incorporated into HIV-1 in an HIV matrix-specific fashion. The ramifications of such interactions are discussed, as the possibility of HIV-1 target tissue broadening and immune evasion are considered.

INTRODUCTION

Mobile genetic elements constitute 45% of the human genome, with those of retroviral origin amounting to approximately 5% (1). These endogenous retroviruses (HERVs) have colonized the germ line; therefore, they are spread through vertical passage from one generation to the next. Most members of the HERV-K group have lost their ability to replicate due to epigenetic silencing or mutagenesis (2). HERV-Ks are among the most recent HERVs to invade the human genome (3, 4); therefore, they are the ones most likely to harbor functional open reading frames (ORFs). Some of the younger HERV-Ks are estimated to be less than one million years old, and some of them (e.g., K113 and K115) are insertionally polymorphic in the human populations (5).

There are at least 89 HERV-K proviruses integrated in the human genome (4). Their expression profiles are dependent on the cell type analyzed and the stimuli driving such expression (6–8). HERV-K protein expression has been detected in testicular germ cell tumor patients (9) as well as in the plasma samples of lymphoma and breast cancer patients (10). A close association between HERV-K expression and the nonadherent malignant transitions, as well as cell-to-cell fusion of melanoma (11, 12), have also been described.

HERV-K expression also has been associated with HIV-1 infection. HERV-K RNA and virus-like particles (VLPs) have been detected in the plasma of infected patients (10, 13). HIV-1-dependent reactivation of HERV-Ks has been attributed to the expression of the HIV accessory proteins Tat and, to lesser extent, Vif (14, 15). HERV-K antibodies and cytotoxic T lymphocyte (CTL) responses against HERV-K antigens are found in HIV-1-infected patients (16–18). Moreover, HERV-K reactivation can lead to the killing of HIV-infected cells by the host immune system (15). It has been suggested that harnessing these anti-HERV CTLs could lead to new approaches for the development of HIV-1 vaccines and immunotherapeutics (15).

Although retroviruses have long been known to be able to interact with each other, particularly at the gag and env levels (19), putative interactions between HIV-1 and HERV-Ks are an underappreciated area of retrovirology. Such interactions could, however, profoundly impact HIV properties such as infectivity, the range of tissues it can infect, and the cytopathic damage caused to the host cells. A recent report described that the expression of a consensus sequence HERV-K Gag had a negative influence on HIV-1 particle release and infectivity caused by the coassembly of HERV-K gag with HIV-1 gag (20).

We speculated that modern-day HERV-Ks are still expressed in primary lymphocytes and could play a role in the HIV-1 life cycle. Rather than using an artificial HERV-K consensus sequence which resembles that of an ancestral human-specific HERV-K (HML-2) (21), we decided to profile HERV-K expression in primary lymphocytes of five different, healthy blood donors and test those that are most widely expressed. We used single-molecule real-time sequencing (PacBio SMRT) (22) instead of other deep-sequencing platforms, such as Illumina and 454, because the short reads of the latter two platforms are inadequate to discriminate between the different HERV-Ks. We found good evidence for more than 30 distinct HERV-K Env transcripts being expressed in primary human lymphocytes. Interestingly, we found that two of the most robustly expressed HERV-Ks produced full-length and posttranslationally processed proteins and that one of them, HERV-K18 Env, was incorporated into HIV-1 particles in a matrix-dependent fashion. These findings imply that some HERV-Ks interact specifically with HIV, possibly shaping the properties of the lentivirus.

MATERIALS AND METHODS

Cell culture.

HEK293T cells were maintained in Dulbecco's modified Eagle medium (DMEM) supplemented with 10% fetal calf serum (FCS; Gibco, Invitrogen) at 37°C and 5% CO₂.

Primary human blood lymphocytes (PBLs) where obtained from anonymous healthy donors through the New York Blood Bank. Peripheral blood mononuclear cells were separated from buffy coats by Ficoll-Hypaque gradient centrifugation. The CD14⁺ cells were removed from the mononuclear fraction using a MACS CD14 isolation kit (Milteny Biotec) according to the manufacturer's directions. PBLs were cultured in RPMI 1640 (Cellgro; Mediatech) supplemented with 10% FCS, 2.3 mg/ml HEPES, glutamine, and 20 U/ml interleukin-2 (IL-2; Roche; obtained through the AIDS Research and Reference Reagent Program). PBLs were activated using αCD3/CD28 antibody-magnetic beads (Dynal, Invitrogen) for 4 days prior to RNA extraction.

RNA extraction, cDNA synthesis, and PCR.

RNA extraction was performed using TRIzol (Invitrogen) according to the manufacturer's directions. Each sample was DNase treated twice using the DNA-free kit (Ambion). One μg total RNA then was used for reverse transcription (RT) using Thermoscript (Invitrogen) reverse transcriptase and conditions specified by the manufacturer's protocol. The reverse transcription primer used consisted of 3 specific segments (Fig. 1A), from 5′ to 3′: a T7 promoter sequence (underlined), a 9mer random sequence (N), and an HERV-K Env-specific sequence (in italics) (5′-TAATACGACTCACTATAGGGNNNNNNNNNTTTCCTACAACTAGCATATAAGG-3′). The Env-specific RT primer sequence represents the majority consensus of the HERV-Ks with nucleotide sequence within the selected region. To eliminate the interference of the reverse transcription primer on the subsequent PCR amplification reaction, HERV-K env RT products were purified using a PCR purification kit (Invitrogen) per the manufacturer's instructions. One-tenth of the purified reverse transcription reaction was used to amplify the 700 bp of the HERV-K env SU region used for expression profiling. The forward primer used for this reaction had an HERV-K-specific sequence 5′-TTATCCTCCTATTTGCCTRGG-3′, while the reverse primer consisted of two segments (Fig. 1A): a 16-nucleotide (nt) donor-specific barcode and a T7 promoter sequence (5′-TTTCCTACAACTAGCATATAAGG-3′). PCR conditions included a denaturation step of 3 min at 95°C, followed by 30 cycles of 94°C for 15 s, 50°C for 30 s, and 72°C for 30 s. The polymerase enzyme used for the amplification was the high-fidelity PfuUltra II polymerase (Ambion). To monitor for genomic DNA contamination, RNA samples underwent an HERV-K-specific PCR without reverse transcription.

FIG 1.

(A) Graphic representation of the diversity in the 700 bp of the HERV-K env SU region used to probe HERV-K expression. Sequences were aligned with MAFFT (23) and graphically represented with Jalview (55). (B) Schematic of the double-tagging strategy where first each cDNA molecule is individually labeled by a sequence of 9 random nucleotides (IDs), and then it is given a donor-specific barcode during the PCR. (C) Illustration of the PacBio sequencing steps used in this report, with ligation of the bell-shaped adaptors to the PCR-amplified HERV-K env region, primer annealing, and the polymerase-dependent reiterative synthesis of the fluorescent product.

Single-molecule sequencing.

Preparation and sequencing of PCR product-derived libraries was performed according to the manufacturer's instructions, reflecting the XL-C2 sequencing enzyme and chemistry. Libraries were purified, processed for repair of the DNA ends, and then converted to SMRTbell templates by the ligation of hairpin adapters (bells) using the PacBio RS DNA template preparation kit. After ligation, the library underwent exonuclease treatment to remove unligated DNA fragments. Sequencing was performed as reported previously (22). Briefly, fluorescent phospho-linked nucleotides were incorporated onto SMRTbell circular templates by immobilized single DNA polymerase molecules (Fig. 1C). Data collection was performed in real time (multiple times for each nucleotide of template) during the rolling-circle replication and recorded in 45-min movies. The reiterating nature of the process allows for the generation of the consensus and the avoidance of systematic error beyond fluorophore-dependent error rates (22). Data were then processed with PacBio RSI CCS resequencing secondary variant analysis using circular consensus (CCS) reads to generate variant (.VCF) files for downstream analyses.

Sequence analysis.

A Python3 pipeline was written to process the “filtered CCS subreads” provided from the initial sequencing process. The sense sequence and reverse complement of the reads were initially screened for the presence of both the reverse transcription and PCR amplification primer sequences. Reads containing both primers and an internal cDNA reverse transcription identity (cDNA ID) of ≥5 bp were sorted by donor using the Levenshtein distance of the sequenced donor barcode to identify the candidate donor. Within the set of reads corresponding to each donor, duplicate reverse transcription tags were used to identify and remove duplicates caused by PCR amplification. Unique reads were subsequently matched by local BLASTn (BLAST 2.2.28+) to a database of HERV-K sequences obtained from the literature (4) and were filtered for those containing hits of >200 bp with >95% homology to the highest-scoring hit within this region. Given the marginally higher error rate of PacBio CCS reads compared to those of other sequencing technologies, a 95% identity limit was chosen to maximize the number of reads accepted into the pipeline. In reality, an average of 98.7% homology to the best match was seen for reads remaining at the end of the filtering. The identity of the best-scoring HERV-K was recorded for these reads, and the matching region of the read was extracted and analyzed for chimerism (due to in vitro recombination) against the HERV-K database using UCHIME (USEARCH 7.0.1001) (23). Reads in which chimerism was detected were excluded, allowing summation of the remainder to give the number of reads corresponding to each HERV-K for each donor. Raw counts for each HERV-K were converted to the percentage of the total reads for each donor, allowing standardization of the read counts between donor samples and downstream comparison. The defined envelope region was aligned using MAFFT (24), and the phylogenetic relationships were inferred using the neighbor-joining method (25) in MEGA5 (26). The evolutionary distances (number of base substitutions per site) were computed using the maximum composite likelihood method (27), removing ambiguous positions for each sequence pair.

Plasmids.

Envelopes of HERV-K102, K18, and KII were amplified from the oligo(dT)-mediated reverse transcription of total RNA from primary PBLs. PCR primers for the amplification of full-length env were the following: forward, 5′-GATCAAGCTTATGAACCCATCAGAGATGCAAAG-3′; reverse, 5′-GATCCCCGGGCTACACAGACACAGTAACAATCTG-3′ (HinDIII and SmaI restriction sites are underlined). Envelopes were cloned into expression plasmid pTR600 (28) using restriction sites HinDIII and SmaI and verified by sequencing. The vector contains an HERV-K RcRe (Rec response element) segment, a sequence necessary for RNA nuclear export, that is similar to HIV RRE (29).

pCRV1 K-rev, pCRV1/gag-PR-pol, and pCRV1/env (21), expressing the consensus sequence of HERV-K rec, gag-pol, and env, were a gift from Paul Bieniasz (Aaron Diamond AIDS Research Center, The Rockefeller University). pNL4.3 E-Luc (30) was a kind gift from Cecilia Cheng-Mayer (Aaron Diamond AIDS Research Center, The Rockefeller University). The plasmids encoding SIVmac251 Gag-Pol pSIV3 (31) were a gift from Dan Littman (Skirball Institute of Biomolecular Medicine, NYU Medical Center). Plasmid phCMV-intron gag-pol, encoding murine leukemia virus (MLV) gag-pol, was a gift from François-Loic Cosset (LVRTG, ENS de Lyon–U412 INSERM, Lyon, France). HIV-1 vectors pCRV1 gag-pol, pCRV1 gag, and pCRV1 gagΔMA (32) were also a gift from Paul Bieniasz (Aaron Diamond AIDS Research Center, The Rockefeller University). HIV/simian immunodeficiency virus (SIV) matrix chimeras were cloned into the pNL4.3 E-Luc backbone using overlap PCR and standard molecular biology procedures.

Transfections, concentration of virus-like particles (VLPs), and Western blotting.

Transfections were carried out in a 24-well format in HEK293T cells using 0.5 μg HERV-K env-expressing plasmid (except for HERV-Kcon; given its high expression, we used 10 ng) and 0.25 μg pCRV1 K-rev and 3 μg/ml polyethyleneimine (Polysciences Inc.). Forty-eight hours after transfection, cells were lysed with 1% SDS lysis buffer (1% sodium dodecyl sulfate, 50 mM Tris-HCl, 150 mM NaCl, and 5 mM EDTA), and proteins separated on a 10% polyacrylamide gel (Invitrogen). After transfer to a polyvinylidene difluoride (PVDF) membrane (Pierce), proteins were probed with α HERV-K Env mouse monoclonal antibody (MAb) HERM 1811-5 (Austral Biologicals) and with horseradish peroxidase (HRP)-conjugated secondary antibodies (Sigma). Membranes were then developed with SuperSignal West Femto (Pierce), and detection was carried out with the ProteinSimple FluorChem E imaging system.

HERV-K env-pseudotyped VLPs were produced by cotransfecting HEK293T cells in a 24-well format using 0.5 μg viral vector, 0.25 μg HERV-K env plasmids, and 0.1 μg pCRV1 K-rev. Concentrated viral pellets were obtained 48 to 72 h after transfection by centrifuging supernatants over a 6% iodixanol (OptiPrep; Invitrogen) cushion at 20,000 relative centrifugal force units (RCFs) at 4°C for 4 to 5 h. Viral pellets were solubilized in 1% SDS lysis buffer and analyzed by Western blotting. The following antibodies were used for the analysis of the VLPs: α HERV-K Env HERM 1811-5, α-SIV p27 antibody (55-2F12) (NIH AIDS Reagent Program), α-HIV-1 p24 monoclonal antibody (183–H12-5C) (NIH AIDS Reagent Program), α-HIV-1 patient serum (which mainly recognizes HIV-1 Gag polyprotein), and α-MLV p12 monoclonal antibody (a gift from Paul Bieniasz).

Immunofluorescence.

HEK293T cells plated on coverslips coated with bovine fibronectin (EMD Millipore) were cotransfected with HERV-K Env-encoding vectors alongside an expression vector encoding a calreticulin-red fluorescent protein (RFP) fusion protein as an endoplasmic reticulum (ER) marker (Origene). Twenty-four hours after transfection, cells were fixed with 3% formaldehyde (Tousimis, Rockville, MD) in phosphate-buffered saline (PBS), permeabilized with PBS and 0.2% Triton X-100, blocked in 1% nonfat milk in PBS, and probed with α HERV-K Env HERM1811-5 (Austral Biological) as a primary antibody and Alexa-Fluor 488 goat anti-mouse as a secondary antibody (Molecular Probes, Invitrogen). Microscopy was carried out on a Zeiss LSM510 inverted confocal microscope at the Microscopy Shared Resource Facility of the Icahn School of Medicine at Mount Sinai. Images were processed with Zeiss Aim Image Examiner software.

RESULTS

HERV-K env profiling in primary human lymphocytes.

Single-molecule real-time sequencing (PacBio SMRT [22]) offsets some of the major weaknesses of short-read next-generation sequencing by providing sequencing reads over an order of magnitude longer than those of other current technologies. Fundamentally, compared to previous short-read technologies, PacBio SMRT sequencing improves the ability to unambiguously align sequences containing repetitive regions. Applied to the sequencing of PCR products, PacBio reads maintain the entire product as an uninterrupted sequence, allowing reliable identification against reference libraries with levels of similarity equivalent to those of HERV-Ks (7). A comprehensive description of the features and advantages of this technology were published previously (33, 34).

In order to determine which of the 89 HERV-K (HML-2) proviruses identified in the most recent human genome assembly (GRCh37/hg19) (http://www.ncbi.nlm.nih.gov/assembly/2758/) are expressed in primary human lymphocytes, we identified a highly divergent region flanked by conserved sequences to allow reverse transcription and amplification using the same sets of primers. The sequence that best met these criteria spans 700 nucleotides within the surface protein (SU) of HERV-K Env. Figure 1A shows a graphical representation of the diversity of the HERV-K env region used, which exhibits a 78% consensus level.

We reverse transcribed RNA extracted from CD3/CD28-activated peripheral blood lymphocytes (PBLs) of five anonymous healthy blood donors. To account for PCR amplification bias and sequencing errors, each cDNA molecule was labeled with a unique nine-nucleotide identification tag (cDNA ID) during the reverse transcription reaction. Moreover, in order to sequence several samples simultaneously, a donor-specific barcode was added during the PCR. Thus, as shown in Fig. 1B, each sequence displayed a unique combination of cDNA ID, donor barcode, and HERV-K env sequence. This approach allows us to calculate the number of distinct HERV-K cDNAs sampled (even if similar in sequence) while avoiding resampling bias during PCR amplification (35).

A total of 3,944 high-quality reads mapping to 32 different HERV-Ks were found in the five samples. Figure 2 shows the sequencing results of the PCR-amplified 700-nucleotide HERV-K env region obtained from a single PacBio SMRT run. The sequence analysis was carried out by excluding duplicates (identified by primer IDs) and retaining only those reads with cDNA primer IDs and a recognizable donor barcode. Unique sequences were compared against the HERV-K database (4). Individual HERV-Ks were assigned when the reads reached ≥95% identity within the hit region. Finally, sequences were assessed for the presence of recombination, as this has been documented to be a source of artifacts (36, 37), and recombinants were excluded.

FIG 2.

Expression profile of 32 HERV-Ks in primary peripheral blood lymphocytes from five different donors. A 700-nt-long HERV-K env fragment was amplified from cellular RNA by RT-PCR and sequenced using the single-molecule, deep-sequencing SMRT PacBio platform. Three HERV-Ks, K102, K18, and KII (HML-2 1q22, HML-2 1q23.3, and HML-2 3q12.3, respectively), make up over 90% of the total HERV-K (HML-2) expression in primary lymphocytes.

Overall, a total of 32 different HERV-K env transcripts were expressed in one or more of the five donors (Fig. 2 and 3). The three most highly expressed env genes were those of HERV-K K102 (HML-2 1q22), HERV-K K18 (HML-2 1q23.3), and KII (HML-2 3q12.3). (GenBank accessions numbers JN675014.1, JN675013, and JN675019.1, respectively). While the overall HERV-K expression varied among donors, K102, K18, and KII were highly expressed in all five of them. Given the robust expression of these three HERV-Ks in activated lymphocytes, it is conceivable that HIV-1 encounters these HERV-Ks during the early and late steps of its viral life cycle.

FIG 3.

HERV-K (HML-2) env phylogenetic tree combined with a heat map of their expression. Loci in red denote HERV-K (HML-2) env genes that are expressed in primary lymphocytes. Different background colors denote different subgroups: green, LTR5A; purple, LTR5B; blue, LTR5-Hs. The numbers in the heat map represent the relative expression (in percentage) of each HERV-K Env, while the colors correspond to different percentage ranges.

Figure 3 shows a heat map representation of the complete HERV-K env expression profile combined with a phylogenetic analysis of the HERV-K env relationships. It illustrates a clear clustering of low-level HERV-K expression in the LTR5A subgroup, which is less evident in LTR5B and LTR5-Hs subgroups (LTR5 grouping was derived from reference 4). On the other hand, the three HERV-Ks that make up most of the total HERV-K env expression all belong to the LTR5-Hs subgroup. These observations are consistent with the fact that these proviruses represent the most recent integrations in the germ line (38, 39); therefore, they show less evidence of the inactivating processes of mutation and recombination.

Protein expression potential of HERV-K env sequences.

We asked next whether any of the three highly expressed HERV-K Env sequences identified above could yield a functional protein that could interact with HIV. We cloned full-length HERV-K102, KII, and K18 env genes into expression vectors (see Materials and Methods) that were then cotransfected into HEK293T cells with the HERV-K homolog of HIV-1 rev, i.e., rec, for optimal RNA nuclear export (21, 29). KII was analyzed despite the presence of stop codons in the coding sequence, as retroviruses have demonstrated translation read-through capacity (40). The Env protein levels were assessed by immunoblot analysis of transfected cell lysates using an antibody that recognizes the transmembrane portion (TM) of the glycoprotein and were compared to the functional Env of HERV-Kcon [a consensus sequence derived from a selected group of 10 HERV-K (HML-2) proviruses (21)]. We found that the env sequences of HERV-K102 and K18 could generate proteins with the predicted molecular weight, whereas, as expected, that of KII could not (Fig. 4A).

FIG 4.

(A) Western blot of HEK293T transfected with the three most expressed HERV-K envelopes in all donors. To allow simultaneous detection of the different envelopes, a smaller amount of HERV-Kcon-expressing plasmid was transfected than for the other three. Also, the HERV-K antibody HERM 1811-5 recognizes the transmembrane domain (TM) portion of the glycoprotein, which explains the different bands detected. (B) Colocalization of transfected HERV-K102, K18, and Kcon Env proteins with the cell's endoplasmic reticulum (ER), analyzed in permeabilized HEK293T cells using laser-scanning immunofluorescence microscopy. Microscopy was carried out on a Zeiss LSM510 inverted confocal microscope at a ×60 magnification.

It is also important to note that functional retroviral Env proteins have their precursor cleaved by cellular furin-like enzymes in the late Golgi apparatus (41, 42) and that they are highly glycosylated (43, 44), which results in the appearance of bands of different molecular weights (Fig. 4A).

Given that antibodies for the HERV-K Env surface region (SU) are not yet available, we assessed cellular localization of the HERV-K102 and K18 Env proteins in permeabilized cells with commercially available anti-HERV-K Env TM antibody. Again, HERV-Kcon Env served as a positive control. Confocal immunofluorescence microscopy shows that all three envelopes localized in the cytoplasm and specifically in the endoplasmic reticulum, which is consistent with their trafficking toward the cell surface (Fig. 4B).

Taken together, these data show that two highly expressed HERV-K RNA sequences retain their ability to be expressed and processed by the cell. Furthermore, despite the lack of antibodies that recognize the HERV-K Env SU region, immunofluorescence experiments show that HERV-K102 and K18 Env proteins localize in the ER, similar to those occupied by the functional HERV-Kcon Env.

HERV-K Env incorporation into HIV-1 particles.

Retroviruses are known to incorporate glycoproteins belonging to other viruses (45–48).

We asked whether the naturally expressed HERV-K Env proteins could be incorporated into virus-like particles (VLPs) of different retroviruses. We examined the two HERV-K glycoproteins that retain protein expression capability, HERV-K102 and K18 Env, and coexpressed them with HIV-1 ΔEnv and HIV-1 Gag-Pol-, SIVmac Gag-Pol-, and MLV Gag-Pol-expressing constructs. Supernatants of the transfected cells were concentrated and analyzed by Western blotting. We observed that HERV-K18 was incorporated into HIV-1 and MLV VLPs with considerable efficiency but not into SIV; in contrast, HERV-K102 Env failed to be incorporated in any of the retroviral VLPs tested (Fig. 5). We also detected HERV-K18 incorporation in HERV-Kcon Gag-Pol VLPs (21 and data not shown), but we could not detect Gag in the concentrated HERV-Kcon VLPs, likely due to the poor sensitivity of the commercially available antibody used.

FIG 5.

Analysis of HERV-K Env incorporation into retroviral particles. Concentrated preparations of viral particles obtained by cotransfection of HERV-K102 or K18 Env with HIV-1 Δenv, HIV-1 Gag-Pol, SIV Gag-Pol, and MLV Gag-Pol constructs were analyzed by Western blotting. HERV-K18 Env appears to be incorporated in both HIV-1 and MLV but not in SIV viral particles (upper), even though both HERV-K18 Env (middle) and VLPs are produced in all samples (lower). α-HIV-1 p24 antibody (183–H12-5C), α-SIV p27 antibody (55-2F12), and α-MLV p12 antibody were used to probe the gag genes of HIV, SIV, and MLV, respectively (lower).

It is interesting that, as previously seen for HERV-K113 Env (43), in the cell lysates we observe several HERV-K Env TM bands (e.g., Fig. 5, cell lysates), while only one of them is visible in the concentrated VLPs (Fig. 5, Env TM VLPs). This indicates that only properly processed and/or correctly glycosylated HERV envelopes have the ability to be incorporated into different retroviral VLPs.

HIV-1 matrix-specific incorporation of HERV-K18 envelope.

The HIV-1 matrix protein (MA) is essential for the incorporation of HIV-1 Env proteins into viral particles (49). Therefore, we asked whether the incorporation of HERV-K18 Env into HIV-1 particles was dependent on MA. To this end, we cotransfected HERV-K18 env with HIV constructs encoding HIV-1 Δenv, gag-pol, gag only, or gag missing the region spanning residues 10 to 100 (GagΔMA) (32). As a control, we used HERV-Kcon Env. All constructs yielded VLPs as detected by the presence of pelletable HIV-1 Gag in the culture supernatants of transfected HEK293T cells (Fig. 6A). Moreover, all constructs except HIV GagΔMA efficiently incorporated HERV-K18 Env and HERV-Kcon Env, as indicated by its presence in the concentrated VLPs. These results show that incorporation of HERV-K18 Env and HERV-Kcon Env into HIV-1 is MA dependent and that the determinants of HERV-K Env incorporation appear to be similar to those of HIV-1 Env, as both require specific Gag-Env interactions (50).

FIG 6.

(A) HERV-K Env incorporation into HIV VLPs is MA dependent. Both HERV-K18 and Kcon envelopes fail to be incorporated into Gag ΔMA VLPs, even though particles are detectable in the concentrated supernatants of the transfected cells (detected with an anti-HIV patient serum), and the HERV-K Env proteins are expressed in the cell lysates. (B) Graphic representation of the HIV-1-SIV matrix chimeras used in panel C. (C) The first 26 residues of the N-terminal portion of HIV MA are important for HERV-K18 Env incorporation into HIV VLPs. VLPs containing HIV-1 SIV matrix chimeras were produced in the presence of HERV-K18 Env. The Western blots of the concentrated supernatants show the levels of HERV-K18 Env incorporation with the different matrix chimeras (top), the levels of VLPs produced (middle), and the HERV-K18 Env expression in the corresponding cell lysates (bottom).

Given that the SIVmac construct used in the previous experiment failed to mediate HERV-K18 Env incorporation, we decided to test a panel of HIV-SIV MA chimeras in order to define the region within HIV-1 MA that is required for K18 Env incorporation. The MA chimeras were constructed in such a manner as to systematically survey the whole matrix region. For comparison purposes, these chimeras were cloned in the same HIV-1 Δenv plasmid used in the earlier experiment (Fig. 6B).

Our results show that replacing the first 26 residues of HIV-1 MA with those of SIVmac MA impairs the ability of the virus to incorporate HERV-K18 Env (Fig. 6C), suggesting that this region contains the important elements required for HERV-K18 Env incorporation into HIV-1 virions. These data are consistent with the notion that mutations in the N terminus of HIV-1 MA are sufficient to prevent HIV-1 Env incorporation into HIV-1 VLPs (51, 52).

Taken together, these results show that one of the HERV-Ks expressed to the highest levels in primary human lymphocytes, HERV-K18, is specifically and efficiently incorporated into HIV-1 virions in a matrix-dependent fashion.

DISCUSSION

The interactions of HIV-1 with its cellular environment can influence its functional properties (e.g., infectivity, immune recognition, and cell tropism). The fact that in T lymphocytes, the cellular environment contains a number of preexisting endogenous retroviral components is often overlooked yet is potentially of great relevance. Indeed, in the human genome, elements that are attributable to past retroviral germ line integrations account for nearly 5% of the genome. There is abundant evidence that the expression of many endogenous retroviruses is upregulated upon HIV-1 infection (15, 53, 54). Thus, HIV-1 infection in human cells is equivalent to a coinfection by several retroviruses. Here, we analyze the expression profiles of HERV-Ks in primary human lymphocytes from five different donors, begin to characterize the functional properties of the HERV-Ks expressed at high levels in human donors, and define the potential for HIV-HERV-K molecular interactions.

By probing a region in the SU domain of env, we observed the expression of a total of 32 HERV-Ks, of which 56% (18/32) are found in at least three of the five donors. We consistently detected that the same three HERV-Ks, HERV-K102 (HML-2 1q22), HERV-K18 (HML-2 1q23.3), and HERV-KII (HML-2 3q12.3), make up over 90% of the total HERV-K (HML-2) expression (Fig. 2).

Interestingly, all three envelopes belong to type 1 HERV-K proviruses. The HERV-K (HML-2) group is divided into type 1 and type 2, depending on the presence or absence of a 292-nt deletion at the 5′ end of the env sequence, resulting in Env proteins of different sizes. RNA transcripts coding for full-length envelope proteins have been detected for both types of genomes (6, 55), and both type 1 and type 2 proviral genomes have remained capable of retrotransposition after hominids split from chimpanzees (56).

Our results in primary lymphocytes differ from recently published observations on HERV-K (8, 57), in terms of the total number of HERV-Ks being expressed and the frequencies at which each one is detected. We identify almost four thousand HERV-K sequences in lymphocytes from five healthy donors, whereas a previous report on HERV-K profiling found 918 sequences from 11 different samples (8). The broader array of HERV-Ks expressed we observe in the present study likely is due to a combination of the higher sensitivity of PacBio SMRT sequencing and the different cell types analyzed. The expression frequencies, on the other hand, could simply reflect the different cell types analyzed.

In agreement with our study, each of the recent reports on HERV-K showed that one or two HERV-Ks are often overrepresented compared to the others: e.g., HML-2 7q22.2 (ERVK-14) in melanoma-related transformed cells (8) and HML-2 22q11.21 (HERV-K101) in human embryonic and induced pluripotent stem cells (57). Indeed, although the HERV-Ks that we report as mostly expressed in lymphocytes were also seen in the previous studies, they were not the most prevalent ones. Expanding what was suggested by Fuchs et al. for pluripotent stem cells (57), a more extensive and systematic characterization of HERV-K expression could establish sets of HERV-Ks as expression markers for different tissues and cell types.

In order to test the potential influence of these HERV-K env genes on the HIV-1 life cycle, we tested their coding potential. We found that two out of the three env genes retain their ability to make full-length protein, and that they are localized in the endoplasmic reticulum within the cell (Fig. 4A and B). We also experimentally demonstrate that HERV-K18 Env is incorporated into HIV-1 virions but not in SIVmac (Fig. 5).

It should be noted that humans and nonhuman primates (e.g., rhesus macaques) differ substantially with respect to their HERV-K repertoires. Almost 90% of the HERV-Ks mapped to date have colonized our ancestors' genome after their divergence from macaques 25 million years ago (4). Therefore, our observation that SIVmac, which conceivably has adapted to its macaque host, fails to incorporate the Env of HERV-K18 is consistent with the fact that HERV-K18 infected the germ line of the human-gorilla common ancestor between 7 and 14 million years ago (4).

Finally, we established that incorporation of HERV-K18 Env is dependent on the N terminus region of HIV-1 matrix (Fig. 6A and C). Of note, this region contains one of the two residues, Leu12, known to play a crucial role in the incorporation of full-length HIV-1 envelope (51).

The implications of HERV-K18 Env incorporation are potentially important for the course of HIV-1 infection. Indeed, HIV-1 virions incorporating a second envelope, such as that belonging to HERV-K18, could have a broader cell type tropism by allowing the infection of CD4-negative cells or by anchoring the virion on a target cell with low CD4 expression.

Furthermore, HERV-K Env on the virion surface could provide the ability to influence cytokine production and induce immunosuppression, as recently described (58); this characteristic also has been reported for other retroviruses, including HIV-1 gp41 (59, 60).

Taken together, we provide evidence that HERV-Ks are expressed in primary lymphocytes and that they have the ability to directly interact with HIV-1. Future studies are needed to determine the extent of their influence on the HIV-1 life cycle and whether their expression can be harnessed to hinder HIV-1 replication.