Identification of late assembly domains of the human endogenous retrovirus-K(HML-2)
© Chudak et al.; licensee BioMed Central Ltd. 2013
Received: 5 June 2013
Accepted: 11 November 2013
Published: 19 November 2013
Skip to main content
© Chudak et al.; licensee BioMed Central Ltd. 2013
Received: 5 June 2013
Accepted: 11 November 2013
Published: 19 November 2013
Late assembly (L)-domains are protein interaction motifs, whose dysfunction causes characteristic budding defects in enveloped viruses. Three different amino acid motifs, namely PT/SAP, PPXY and YPXnL have been shown to play a major role in the release of exogenous retroviruses. Although the L-domains of exogenous retroviruses have been studied comprehensively, little is known about these motifs in endogenous human retroviruses.
Using a molecular clone of the human endogenous retrovirus K113 that had been engineered to reverse the presumed non-synonymous postinsertional mutations in the major genes, we identified three functional L-domains of the virus, all located in the Gag p15 protein. A consensus PTAP tetrapeptide serves as the core of a main L-domain for the virus and its inactivation reduces virus release in HEK 293T cells by over 80%. Electron microscopy of cells expressing the PTAP mutant revealed predominantly late budding structures and budding chains at the plasma membrane. The fact that this motif determines subcellular colocalization with Tsg101, an ESCRT-I complex protein known to bind to the core tetrapeptide, supports its role as an L-domain. Moreover, two YPXnL motifs providing additional L-domain function were identified in the p15 protein. One is adjacent to the PTAP sequence and the other is in the p15 N-terminus. Mutations in either motif diminishes virus release and induces an L-domain phenotype while inactivation of all three L-domains results in a complete loss of particle release in HEK 293T cells. The flexibility of the virus in the use of L-domains for gaining access to the ESCRT machinery is demonstrated by overexpression of Tsg101 which rescues the release of the YPXnL mutants. Similarly, overexpression of Alix not only enhances release of the PTAP mutant by a factor of four but also the release of a triple mutant, indicating that additional cryptic YPXnL domains with a low affinity for Alix may be present. No L-domain activity is provided by the proline-rich peptides at the Gag C-terminus.
Our data demonstrate that HERV-K(HML-2) release is predominantly mediated through a consensus PTAP motif and two auxiliary YPXnL motifs in the p15 protein of the Gag precursor.
The Gag precursor plays an essential role in retroviral assembly and release [1, 2]. It contains domains that mediate an association with the lipid bilayer, Gag-Gag subunit-interactions and egress from the producer cell. Successful budding ending with membrane scission to release retroviral particles depends on short peptide motifs of this protein termed ‘late (L)-domains’. Deletions or explicit mutations in viral L-domains freeze the budding step and prevent the separation of virions from the plasma membrane [3, 4].
To date, three characteristic classes of L-domains have been defined, namely PT/SAP, YPXnL and PPXY. An L-domain extends beyond the largely preserved core amino acids. A functional PT/SAP class of L-domains can be as large as 12 amino acids. L-domains function by directly or indirectly linking the Gag precursor proteins to the cellular ESCRT machinery principally involved in the endosomal sorting of cargo proteins and the biogenesis of multivesicular bodies. This machinery consists of about 25 cellular proteins that form four major complexes termed ESCRT-0, -I, -II and -III .
The three classes of L-domains found in viruses interact with different sorting complex proteins to gain access to the ESCRT pathway. The PT/SAP motif interacts with Tsg101 (tumour susceptibility gene 101), an ESCRT-I complex component, which was identified in yeast two-hybrid experiments as an HIV-1 p6-interacting protein. The depletion of Tsg101 by small interfering RNA interrupts HIV-1 budding to a large extent . On the other hand, the PPXY motif functions through a direct recruitment of Nedd4 (neuronal precursor cell-expressed developmentally down-regulated-4)-like ubiquitin ligases by binding to its multiple WW domains . Ubiquitinylation is a critical step in processes involving the ESCRT pathway and viral budding [8–10]. Direct ubiquitin fusion to Gag can functionally compensate for the absence of a retroviral L-domain and the eventual recruitment of an ubiquitin ligase to the budding particle [10, 11]. The third motif, YPXnL, associates with a protein named Alix (Apoptosis-linked gene 2-interacting protein) [12, 13]. The 97 kDa adaptor protein provides a direct link between ESCRT-I and ESCRT-III complexes. It interacts with ESCRT-III by means of its N-terminal Bro1 domain and with Tsg101 via its C-terminal proline rich region. A central region of the protein mediates binding to the YPXnL motif present, for example, in the p9 protein of equine infectious anaemia virus (EIAV) [12, 13]. Therefore, all three L-domain motifs appear to enter the same core scission complex, albeit by different routes.
Several retroviruses contain more than one type of L-domain motif, and these are often closely spaced or even overlapping. For example, Mason Pfizer monkey virus (MPMV), a close relative of HERV-K(HML-2), harbors a PSAP sequence four amino acids downstream of a PPPY motif within its pp24/16 protein . Furthermore, the human T-cell leukemia virus type I (HTLV-I) contains a bipartite PPPYVEPTAP motif and HIV-1 has, in addition to its primary PTAP motif, an additional L-domain of the YPXnL type [13, 15]. Beside the L-domains, the nucleocapsid protein is also engaged in the budding process through an interaction with the Bro1 domain of Alix .
While the budding of exogenous retroviruses has been well characterized, little is known about this process for endogenous retroviruses, although a PTAP motif was recently identified in the Gag of the human endogenous retrovirus (HERV)-K(HML-2) [17–19]. HERVs are relicts of infectious exogenous retroviruses whose proviruses became integrated into the human genome millions or at least hundred thousands of years ago [20, 21]. They comprise approximately 8% of human DNA and several of these elements have been linked to oncogenesis, neurological disorders and autoimmune diseases [22–25]. Despite the fact that all the known elements carry mutations that prevent productive replication, functional proteins and mature virus-like particles of the HERV-K(HML-2) subfamily are expressed . The study of virus assembly and other aspects of these archaic retroviruses have been facilitated by the generation of consensus sequences and reconstitution of original virus sequences that allow expression of functional proteins [17, 27–29].
In this present report, we show that HERV-K(HML-2) uses two distinct types of L-domain motifs to hijack the ESCRT pathway. Mutation of a single nucleotide in the PTAP motif of the HERV-K(HML-2) Gag p15 protein leads to a characteristic L-domain phenotype, with viruses arrested at a late stage of release. Our results indicate that the PTAP motif is the core of the principal L-domain of the virus. However, at least two auxiliary YPXnL L-domains in p15 are present that efficiently provide alternative access to the ESCRT pathway, if the primary L-domain is restricted and Alix is overexpressed allowing an assembly of the ESCRT-III complex.
Tsg101 is an essential component of the ESCRT-I complex and has been shown to interact directly with the PTAP motifs of several retroviruses [6, 33, 34]. We therefore investigated the subcellular colocalization of HERV-K(HML-2) Gag and Tsg101 to provide further evidence for a conventional role of this virus’s p15 protein PTAP motif. A codon-optimized version of the oriHERV-K113 gag sequence  was cloned in-frame upstream of the Cherry fluorescent protein and a PTAP mutant generated by substituting the threonine for alanine.
Overview of the p15 L-domain mutants used in the study
RAPYPQPPTRRLNP A APPSR
RAPSR QPPTRR ANPTAPPSR
RAPSR QPPTRR ANP A APPSR
VISR ET A K AEGK
VISR ET A K AEGK
RAPYPQPPTRRLNP A APPSR
VISR ET A K AEGK
RAPSR QPPTRR ANPTAPPSR
VISR ET A K AEGK
RAPSR QPPTRR ANP A APPSR
Analogous to the rescue of HERV-K(HML-2) YPXnL defects by ectopic Tsg101 expression described above, previous studies have documented that overexpression of Alix can rescue budding defects of HIV-1 PTAP L-domain mutants [12, 42]. We therefore used our HERV-K(HML-2) PTAP mutants to determine whether overexpression of Alix also increases their release. Cotransfection of FLAG-Alix plasmids indeed partially abrogated the PTAP defects in mutants with one or two wild type YPXnL motifs. Surprisingly, release of even the virus with mutations in both YPXnL motifs was rescued, reaching levels of up to 50% of the wild type (Figure 8). As well as being in line with the identified functional YPXnL motifs, these findings also potentially indicate the presence of at least one additional, unidentified cryptic YPXnL motif. Alternatively, low expression and incorporation of chromosomally encoded Gag proteins into nascent viral particles might also explain the rescue by high cytoplasmic levels of Alix. Indeed, we previously demonstrated very low levels of HERV-K(HML-2) transcripts in HEK 293T cells by sensitive RT-PCR .
The human genome encodes about 90–100 well-preserved proviruses of the HERV-K(HML-2) family . Several of these proviruses code for Gag proteins that can form retroviral particles upon expression. Using electron microscopy and other methods, we and others have demonstrated that these particles bud in a C-type manner and are in principle able to ‘pinch off’ from the cell membrane and undergo maturation [17, 27, 45]. To identify the protein motifs of HERV-K(HML-2) that govern the release of these ancient retroviruses, we generated a reconstituted HERV-K113 element in which non-synonymous postinsertional mutations were reverted (termed oriHERV-K113). The lack of such mutations renders consensus sequences or reconstituted sequences of endogenous retroviruses and retroelements suitable for general functional analyses of their proteins [17, 27–29, 46].
Guided by knowledge of the common position and consensus sequences of late domains in exogenous retroviruses, we performed an in silico screen of the oriHERV-K113 sequence and identified a potential PTAP motif and two YPXnL motifs in the p15 protein localized between the matrix and capsid subunits of Gag. One of the YPXnL domains, a YPETL sequence, is found N-terminal in the p15 protein. The second, a YPQPPTRRL sequence, is located just 10 amino acids upstream of the PTAP motif with only a single intervening amino acid. Closely spaced L-domains are very frequently observed in retroviruses and some other enveloped viruses [47, 48]. In the lentivirus HIV for example, a PTAP and a YPXnL L-domain is located in the phosphoprotein p6 at the C-terminus of the Pr55Gag. HERV-K(HML-2) encodes two proline and glutamine rich peptides at the C-terminus with unknown functions . Although a conventional L-domain motif is not evident in these peptides, sequences closely resembling a PPPY motif are present, making this a region of interest. Deletion mutants unable to express these peptides showed no impairment or defect in particle release, strongly arguing against them having an L-domain function.
In contrast to this C-terminal deletion mutant, a T254A substitution in the PTAP sequence (known to inactivate this L-domain) reduced particle release in HEK 293T cells by a factor of approximately six, and thin section EM revealed a classical late domain phenotype with an abundance of particles at the membrane unable to finalize the last steps of release and ‘pinch off’. A similar defect in particle release in a PTAP mutant of the HERV-K(HML-2) consensus sequence has also been recently reported . The function of this sequence as core of an L-domain has been further substantiated by colocalization studies of the HERV-K(HML-2) Gag with the ESCRT-I complex protein Tsg101. While full length Tsg101 is found in association with Gag at presumed budding sites, this is not the case if the PTAP mutant is expressed, indicating that the PTAP motif at position 253 is the only Tsg101 binding site for this virus’s Gag protein. This is supported by the fact that overexpression of Tsg101 fails to increase the release of the PTAP mutant. Consistent with previously published observations with HIV and other viruses, overexpression of Tsg101 resulted in a moderate decrease of wild type release that can be explained by a partial disorder of cellular endosomal sorting pathways [41, 49]. On the other hand, transfection of HEK 293T cells with a Tsg101 expression plasmid increased significantly the release of both YPXnL L-domain mutants and of the version carrying mutations in both YPXnL motifs. Inactivation of each of the presumed two YPXnL L-domains reduced particle release and induced a late phenotype, although less dramatically than did mutation of the PTAP motif. By thin section EM we observed changes in quantity but no morphologic differences in the PTAP and the YPXnL mutants.
The impact of the mutations regarding the degree of virus release indicates that in HEK 293T cells the PTAP motif is the dominant L-domain and each of the YPXnL motifs play a supportive role. We obtained similar results with SK-Mel13 cells, with a more pronounced inhibition (up to 95%) of release by the PTAP mutant (data not shown). However, in other cell types the PTAP motif might be of less relevance if, for example, more Alix and less Tsg101 is expressed. The relevance of different types of L-domains is known to be cell type dependent . For efficient budding retroviruses must recruit the ESCRT-III complex. It provides the mechanical means for scission of the virus from the cell membrane. This complex can be recruited via the ESCRT-I or ESCRT-II route, though Alix appears to be used in both pathways. Alix interacts with the ESCRT-III complex so that by high Alix levels or overexpression of this protein the downstream ESCRT-III complex is assembled and budding can be restored. However, the process is much more efficient if the ESCRT-I complex is available .
In HEK 293T cells, mutation of both YPXnL L-domains has a synergistic effect on particle release compared with that of individual motif mutants, resulting in a stronger inhibition than that seen with the PTAP mutation. This is in line with the significant reduction in particle release following siRNA mediated down regulation of Alix or Tsg101 expression. The close proximity of the PTAP motif to one of the YPXnL L-domains suggests that mutation in one of these domains might also inactivate the other. This, however, appears not to be the case as variants carrying mutations in both motifs (YPXnL2 + PTAP) are more impaired than the single motif mutants. Moreover, overexpression of Tsg101 is able to partially rescue the release of the variant carrying an inactivating mutation in the nearby YPXnL motif, demonstrating the presence of a functional PTAP motif. In turn, Alix overexpression can partially rescue a virus that carries inactive PTAP and N-terminal YPXnL1 motifs, providing further evidence that YPXnL motifs function as L-domains in HERV-K(HML-2) Gag. However, the triple mutant carrying inactivating mutations in both YPXnL L-domains plus the PTAP domain is still partially rescued by Alix overexpression. Therefore, there may be one or more additional cryptic sites with low affinity for Alix that allow interaction with this protein (if overexpressed) and provide access to the ESCRT machinery. Alternatively, low-level expression of Gag from proviruses encoded in the chromosomes of HEK 293T cells might, under such circumstances, be sufficient to allow release of this mutant. Although HEK 293T cells express HERV-K(HML-2) transcripts at barely detectable levels , incorporation of Gag from endogenous proviruses with functional L-domains is very likely since even HIV Pr55Gag has been shown to be integrated into nascent HERV-K(HML-2) particles . Even a minor fraction of Gag proteins with functional L-domains per particle is sufficient to confer egress from the cell  and our results with Tera-1 cells that express considerably higher levels of endogenous Gag support this hypothesis. The expression of a minority of Gag proteins with functional L-domains might be the reason why many tumor cells are able to release HERV-K(HML-2) particles. The particle production of the PTAP and triple mutants in Tera-1 was only moderately impaired, with a 5-fold reduction in the release of the triple mutant compared to complete inhibition in HEK 293T cells. However, further experiments are needed to confirm or refute the presence of cryptic Alix binding sites in the HERV-K(HML-2) Gag protein and to analyze the effects of incorporation of low fractions of Gag proteins with functional L-domains.
This study identifies three L-domains in a partially reconstituted prototypical HERV-K(HML-2) virus. The major L-domain, providing access to the ESCRT complex via the Tsg101 protein, is a consensus PTAP motif in the C-terminus of the p15 protein. Based on mutagenesis studies and rescue experiments with Alix overexpression, this L-domain was shown to be supported by two auxiliary YPXnL motifs also located in the p15 protein of the virus. By providing essential information concerning the relevant functional domains facilitating HERV-K(HML-2) particle release, these results advance our understanding of the biology of these ancient and inherited elements preserved within our genomes.
HEK 293T, SK-Mel13 and Tera-1 cells were cultured in complete Dulbecco’s modified Eagle medium (DMEM) supplemented with 10% fetal bovine serum, L-glutamine (2 mM), penicillin (50 U/ml) and streptomycin (50 μg/ml).
For the generation of the pBSKoriHERV-K113, the proviral sequence of HERV-K113 (GenBank AY037928) was amplified from the RP11-12X BAC-library plasmid  and cloned into the small pBlueskript SK+ (pBSK) vector from Stratagene using Apa I and Not I as restriction sites. The identification of the presumed 25 non-synonymous postinsertional mutations has been described elsewhere [17, 24, 28]. These mutations and the three mutations in the 3′LTR (see Additional file 1) were introduced into pBSK-HERV-K113 by site-directed mutagenesis using the QuikChange® Multi Site-Directed Mutagenesis Kit from Agilent Technologies to produce pBSKoriHERV-K113.
The L-domain mutants were also generated by site-directed mutagenesis using the same Kit and the following primers: PTAP-: 5′-gccgcccactaggagacttaatcccgcggcaccacctagtagacagggtagtg-3′; YPXnL1: 5′-caattacaggaggtgatatctagagaaacgttaaaattag-3′ and gtatggatatctagagaaacggcaaaagcagaaggaaaaggtccag; YPXnL2: 5′-ggcagggcgccatccagacagccgcccactaggag-3′ and 5′-cagccgcccactaggagagctaatcctacggcaccac-3′. The primer 5′-caacaaactggggcattctgaattcagccatttgttcc-3′ was used for the introduction of the stop codon after NC to generate the ΔQP-mutant and the primer 5′-cctcccaccaggcggctgaacgcccctcccagcaggcagagcgag-3′ to introduce the PTAP--mutation into the codon-optimized Gag in the pGag-Cherry vector. The reverse complement oligonucleotides primers used in the mutagenesis reactions are not shown.
The pGag-Cherry construct was designed for immunofluorescence microscopy in which the synthetic partially codon-optimized gag sequence (described previously) was cloned using the Sac I and Kpn I sites of a pmCherry-N1 vector [17, 28]. The PTAP mutant pGagPTAP--Cherry was generated as described above using the QuikChange® Multi Site-Directed Mutagenesis Kit (Agilent Technologies). pcGNM2/Tsg-F and pcGNM2/Tsg-3′, used to express Tsg101 and Tsg-3′, respectively, were kind gifts of Eric Freed (University of Maryland, USA). The plasmid pCMV-FLAG-ALIX was kindly provided by Jörg Votteler (University of Erlangen, Germany). The small interfering RNAs (siRNA) used to down regulate Tsg101 has been previously described [6, 38] as well as the Alix specific siRNAs [39, 40]. As non-target siRNA the “AllStars Neg. Control siRNA” (Qiagen) was used.
HEK 293T cells (6 × 105) were seeded into 6-well plates and transfected the following day with 1.95 μg pBSKoriHERV-K113 or mutant constructs using Polyfect (Qiagen) according to the manufacturer’s instructions. In addition, 0.05 μg pGL3-Promotor Vector (Promega) was included for normalization. Samples of cell culture media taken 48 h after transfection were filtered (0.45 μm) and RT activity measured using the HS-Mg RT Activity Kit (Cavidi, Uppsala, Sweden). Luciferase activity in transfected cells lysed 48 h post transfection (using the Luciferase Cell Culture Lysis 5x Reagent from Promega) was measured using the Promega Luciferase Assay Kit.
2.4 × 106 HEK 293T cells grown in 100 mm dishes were transfected with the different plasmids (25 μg each) using calcium phosphate. At four days post transfection, samples of culture media were harvested, clarified at 3345 × g for 8 min and filtered (0.45 μm) to remove cell debris. The supernatants were then centrifuged at 175,000 × g for 3 h at 4°C through a 20% sucrose cushion in a Beckman SW32Ti rotor. Virus pellets were resuspended in 60 μl 0.05 M Hepes buffer, pH 7.2 for Western blot analysis. To prepare cell lysates, transfected cells were resuspended in cell lysis buffer (1% Triton-X 100, 20 mM Tris pH 7.7, 150 mM NaCl) containing complete protease inhibitor cocktail (Roche Diagnostic). For the Tsg101 and Alix depletion assays, 4 μg of pBSKoriHERV-K113 were cotransfected with a mix of 6 μl of two specific siRNAs (20 μM) or the same amount (12 μl, 20 μM) of control siRNA using Attractene (Qiagen) according to the transfection protocol. The transfections were performed in 6-well plates in triplicates. For Western blot analyses the supernatants and the cell lysates of the triplicates were combined.
Viral lysates or pelleted virus particles were mixed with sample buffer and boiled for 10 min before being subjected to sodium dodecyl sulphate-polyacrylamide gel electrophoresis (SDS-PAGE). The proteins were transferred onto a PVDF membrane (Roth) and after blocking in blocking buffer (phosphate-buffered saline-PBS, 5% skim milk powder, 0.1% Tween), the membranes were probed with anti-CA rat sera as described previously  and a secondary horseradish peroxidase-conjugated goat-anti-rat antibody (Sigma-Aldrich). For the detection of Tsg101 the mouse monoclonal antibody 4A10 (Genetex) was used and for the detection of Alix a goat polyclonal IgG (Q-19; Santa Cruz Biotech) was applied as primary antibody with species specific secondary antibodies in both cases. The β-actin was detected with the AC-74 monoclonal antibody (Sigma-Aldrich). Proteins were visualized using Super Signal West Femto Maximum Sensitivity Substrate (Thermo Scientific) and a Kodak Medical X-ray film.
HEK 293T cells were cultured in chamber slides (Nunc) and transfected using Polyfect (Qiagen). Twenty-four hours post transfection, cells were fixed at room temperature with 2% paraformaldehyde in PBS for 30 min. Cells were then rinsed three times with PBS and permeabilized with 0.5% Triton-X100. After blocking with 1% skim milk powder in PBS for 30 min, cells were incubated with primary mouse anti-HA antibody diluted 1:200 in blocking buffer for 1 h at 37°C. Cells were then washed three times with PBS and incubated with a secondary Alexa-488 conjugated anti-mouse-IgG antibody (Invitrogen) for 1 h. After repeated washings, cell nuclei were stained using DAPI (4′,6-diamidino-2-phenylindole) at 0.2 ng/ml. Finally, the cells were mounted in Mowiol and examined using a Zeiss LSM 780 confocal laser scanning microscope.
Transfected HEK 293T cells were fixed using 2.5% glutaraldehyde in 0.05 M Hepes (pH 7.2) for 1 h at room temperature. The methods for embedding and for the preparation of thin section transmission EM images have been described elsewhere .
We thank Lars Möller, Gudrun Holland, and Dr. Kazimierz Madela for their excellent technical assistance, Oliver Hohn for helpful discussions and Steve Norley for critical reading and assistance with the preparation of the manuscript. We are indebted to Jörg Votteler (University of Erlangen, Germany) and Eric Freed (University of Maryland, USA) for providing Alix and Tsg101 expression plasmids.
This article is published under license to BioMed Central Ltd. This is an open access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.