Evidence that the endosomal sorting complex required for transport-II (ESCRT-II) is required for efficient human immunodeficiency virus-1 (HIV-1) production
- 2.1k Downloads
Egress of a number of different virus species from infected cells depends on proteins of the endosomal sorting complexes required for transport (ESCRT) pathway. HIV has also hijacked this system to bud viruses outward from the cell surface. How ESCRT-I activates ESCRT-III in this process remains unclear with conflicting published evidence for the requirement of ESCRT-II which fulfils this role in other systems. We investigated the role of ESCRT-II using knockdown mediated by siRNA and shRNA, mutants which prevent ESCRT-I/ESCRT-II interaction and a CRISPR/Cas9 EAP45 knockout cell line.
Depletion or elimination of ESCRT-II components from an HIV infected cell produces two distinct effects. The overall production of HIV-1 Gag is reduced leading to a diminished amount of intracellular virion protein. In addition depletion of ESCRT-II produces an effect similar to that seen when ESCRT-I and -III components are depleted, that of a delayed Gag p26 to p24 +p2 cleavage associated with a reduction in export of virion particles and a visible reduction in budding efficiency in virus producing cells. Mutants that interfere with ESCRT-I interacting with ESCRT-II similarly reduce virus export. The export defect is independent of the decrease in overall Gag production. Using a mutant virus which cannot use the ALIX mediated export pathway exacerbates the decrease in virus export seen when ESCRT-II is depleted. ESCRT-II knockdown does not lead to complete elimination of virus release suggesting that the late domain role of ESCRT-II is required for optimal efficiency of viral budding but that there are additional pathways that the virus can employ to facilitate this.
ESCRT-II contributes to efficient HIV virion production and export by more than one pathway; both by a transcriptional or post transcriptional mechanism and also by facilitating efficient virus export from the cell through interactions with other ESCRT components.
KeywordsHIV ESCRT Late domain Virus budding
Budding of human immunodeficiency virus-1 (HIV-1) from an infected cell is known to be mediated by the endosomal sorting complexes required for transport (ESCRT) machinery recruited by either TSG101/Vps23 (human protein and yeast homologue, respectively) of ESCRT-I and/or ALIX, both via the p6 domain of Gag, the major structural protein of the core of HIV . The amino acid motifs in p6 critical for these two pathways are PTAP and YPXnL (where X refers any amino acid and n = 1–3 residues; abbreviated to YPXL hereafter), respectively. Both pathways require the downstream ESCRT-III and VPS4 ATPase. ALIX is able to recruit CHMP4 of ESCRT-III directly, however its contribution to HIV-1 budding appears to be minimal in certain cell lines unless it is overexpressed [2, 3, 4]. For the pathway mediated by TSG101, the link between ESCRT-I and -III is unclear. In yeast, ESCRT-I, -II and -III act consecutively in the formation of multivesicular bodies and ESCRT-II is crucial for the activation of ESCRT-III .
ESCRT-II is a Y-shaped complex comprised of two EAP20/Vps25 subunits, each forming one arm, and one copy each of EAP30/Vps22 and EAP45/Vps36 which together form the third arm [6, 7, 8]. ESCRT-II binds to ubiquitinated cargos and to membranes [7, 9, 10]. Together with ESCRT-I, it deforms membranes and induces bud formation in vitro . Both copies of the ESCRT-II subunit EAP20/Vps25 are required to induce conformational changes and to activate ESCRT-III CHMP6/Vps20 for membrane scission [8, 12, 13].
In the context of HIV-1 budding, previous reports showed that siRNA mediated depletion of the ESCRT-II subunit EAP20 did not appear to affect virus or virus-like particle (VLP) release [14, 15] and it was concluded that the infectivity of viruses was unaffected up to 48 h after knockdown . Effects at later time points were not investigated in detail although knockdown of CHMP6 reduced infectious HIV-1 production four to five fold at 48 h; this however was ascribed to a general effect on protein trafficking. More recently, using a giant unilamellar vesicular system and full length myristylated HIV-1 Gag, it was shown that whereas ALIX can directly recruit the key ESCRT-III subunit CHMP4, ESCRT-I can only recruit CHMP4 when ESCRT-II and CHMP6 are present as intermediary factors . ESCRT-II was also believed to be dispensable for normal cytokinesis  but a more recent report shows that EAP20/Vps25 is required for the recruitment of CHMP6 for ESCRT-mediated abscission .
Following an initial observation that ESCRT-II knockdown appeared to inhibit HIV-1 production we analysed its role in viral replication using a range of assays including: knockdown with interfering RNA, dominant negative protein expression and the use of an EAP45 knockout cell line. We present evidence consistent with ESCRT-II having at least two roles in efficient HIV-1 production, one of which is enhancement of viral budding.
Knockdown of ESCRT-II with shRNA reduces the production of infectious virus
Sequences targeted by the shRNA expression plasmids used
To ensure that this was not an artefact of the VSV-pseudotype system the experiment was repeated using wild type HIV. Virus production was again analysed 96 h post-transfection and similar results were obtained (Fig. 1h, i, j). Knock down of individual ESCRT-II components thus impairs HIV-1 protein production. There is a decrease in intracellular p24 protein detected but, most markedly by 96 h, we also noted a relatively greater decline in supernatant p24 and viral infectivity compared to the fall in intracellular p24 suggestive of an additional budding defective phenotype.
Disrupting ESCRT I/ESCRT-II interaction inhibits production of infectious virus
We used two mutants (an isolated GLUE domain and the EAP45 mutant IERK159-162AAAA, designated H0m below) to seek evidence of a direct interaction occurring between ESCRT-I and ESCRT-II in HIV particle budding. As determined by Western blot, transfection of GLUE alone led to overexpression (Fig. 2b). H0m transfected and untransfected cells showed no difference in EAP45 expression levels due to interdependency of the ESCRT-II subunits (data not shown). The untagged H0m could not be distinguished from wild type EAP45 on Western blot due to the few amino acid changes, but an HA-tagged variant verified that the mutations in H0m did not prevent its expression (Fig. 2c). The transfection of either mutant did not affect cell survival (Fig. 2d).
Viral plasmids were co-expressed with either wild type EAP45 or one of the two mutant EAP45 proteins. Compared to cells transfected with the control firefly luciferase, wild type EAP45 had no effect on the production of infectious wild type virus (Fig. 2e). By contrast, overexpression of GLUE or expression of H0m, significantly and reproducibly resulted in a decline in intracellular p24 but a proportionally larger fall in extracellular p24 and virion infectivity at 96 h post-transfection (Fig. 2e). Since the GLUE domain is the cargo-binding domain of ESCRT-II and both GLUE and H0m interact less efficiently with ESCRT-I than with EAP45 , their expression in HIV-1-producing cells likely diverted cargos that are relevant to viral particle production from the ESCRT machinery. Thus a direct interaction between ESCRT-I and ESCRT-II would appear to facilitate the normal production of infectious HIV-1.
Knockdown of either TSG101 or EAP20 produces a similar phenotype
Knockdown of ESCRT-II has an additive effect on release of YPXL mutant virus
Virus production is affected in CRISPR/Cas 9 EAP45 knockout cells
A role for EAP45 solely at a transcriptional or early post transcriptional stage would not explain the disproportionate decrease in virion release from ESCRT-II knocked down cells compared to the intracellular levels of viral p24 protein (Fig. 1). In addition the rate of Gag processing appeared to be impaired in the EAP45 knockout cells, especially the terminal cleavage of p24/p2 to p24 compared to the parental cells (Fig. 5).
The ESCRT machinery was originally discovered in yeast where it was found to be essential for the sorting of cargos into intraluminal vesicles for degradation. Studies in yeast and in vitro reconstitution experiments showed that ESCRT-0, I, -II, -III and the Vps4-Vta1 complexes are sequentially involved in this pathway (reviewed in ). ESCRT-II proteins are crucial for this; overexpression of ESCRT-II rescues yeast ESCRT-I deletion mutants but not vice versa .
Humans possess a conserved homologous pathway and the role of ESCRT in HIV-1 budding is well established  with recent evidence suggesting a scaffolding interaction between the viral Gag protein and ESCRT components at the site of virus assembly . The involvement of ESCRT-II has been the subject of debate for HIV and certain other systems. Specific cellular cargos including ferroportin  and CXCR4  require ESCRT-II for their lysosomal degradation following internalisation, while others (Kaposi sarcoma-associated herpesvirus ubiquitinated MHC-I) do not . For epidermal growth factor receptor (EGFR), one of the most widely studied ESCRT cargos, the data are conflicting [14, 19, 20].
Budding and release of HIV-1 is a topologically similar process to endosomal vesicle formation in which the membrane also evaginates from the cytoplasm. ESCRT-I, -III and the ESCRT-associated protein ALIX have been firmly implicated in the process. The p6 region of Gag contains two late domains, the PTAP and YPXL motifs that bind TSG101 and ALIX, respectively (reviewed in [32, 33]). Both of the TSG101 and ALIX pathways require the recruitment of certain ESCRT-III subunits and VPS4 ATPase [34, 35]. ALIX is able to recruit ESCRT-III directly , however in certain cell lines its contribution to HIV-1 budding is a minor one unless it is overexpressed [4, 37]. For the predominant pathway mediated by TSG101, the link between ESCRT-I and -III is still unclear. Langelier et al. concluded that HIV-1 release was ESCRT-II-independent since the depletion of the ESCRT-II subunit EAP20 by small interfering RNA did not appear to cause a significant effect on viral release and infectivity as measured up to 48 h post transfection in a pseudotyped viral system. A similar phenotype was noted using wild type virus harvested at 24 h post transfection. They did however report that, at delayed time points, knockdown of ESCRT-II had a deleterious effect on virus production . Pincetic et al. showed that the fusion of EAP20 or EAP45 to the C-terminal of HIV-1 Gag containing a PTAP motif mutation did not rescue budding of virus-like particles whereas it did rescue ASLV ; their work also highlighted the different cell membranes involved in budding used by HIV and ASLV. It is unknown however whether the fused EAP20 and EAP45 used in their studies were properly folded and fully functional. Two copies of EAP20 are required to activate ESCRT-III [8, 12, 13] and isolated EAP45 is unstable without the other ESCRT-II partners due to its extensive exposed hydrophobic regions [7, 8, 19, 20]. How ESCRT-III is activated without ESCRT-II was unexplained . Recently however a reconstituted assembly model of HIV using myristylated Gag and giant unilamellar vesicles, presented evidence that ESCRT-II was necessary for the TSG101-mediated pathway to function for viral particle assembly .
We have now used several established and novel approaches to study this controversial area including two different knockdown protocols with two different viral strains in different cell lines and also a CRISPR/Cas 9 EAP45 knockout cell line. In both HeLaM and 293T cell lines a defect was observed in viral export when ESCRT-II was knocked down. Importantly in no case was there a detectable decrease in cell viability at the late time points assayed, indicating that our findings were specific and did not reflect a generalised effect on vesicular sorting in the cell (Figs. 1a, 2d, Additional file 4: Figure S4C). EAP20 plays an important role in recruiting CHMP6 for correct cytokinesis  and the expression of ESCRT-II components is mutually dependent [19, 20 and this study]. However, the depletion of EAP45 does not noticeably affect cell proliferation suggesting the level of EAP20 is sufficient for normal cell division.
In HeLaM cells the effect of EAP45 knockdown on viral budding was masked to some extent by a decline in Gag production as measured by intracellular p24 (Fig. 1) but this decrease can now be explained as ESCRT-II appears to have an additional earlier effect on HIV protein production (Figs. 5, 6). Both knockdown and knockout experiments document a decline in intracellular Gag production indicating that there may be a transcriptional or post transcriptional effect of ESCRT-II. The magnitude of this effect seems to be cell type dependent. In 293T cells, the total p24-associated viral products are relatively unaffected (data not shown) consistent with a previous report . However, in HelaM and HAP1 cells a more significant decrease in p24-associated viral products is seen (Figs. 1, 5). ESCRT-II has been shown to have transcriptional effects in other systems  but this has not been investigated in detail in the context of HIV. Despite the reduction in Gag production partially obscuring the later budding effect to some degree the latter is strikingly clear in the EAP45 knockout cell line (Figs. 7, 8) suggesting that ESCRT-II although not indispensable for production of infectious virus, contributes significantly to efficient particle release.
Viruses released from the PTAP mutated provirus transfected cells may be using the intact YPXL-ALIX pathway. Interestingly the fold reduction in release ratio between WT and PTAP mutant from the control cell line is similar to that from the knockout cells suggesting the proportion of viruses using YPXL-ALIX pathway is unaltered. The additive defect in infectious virus production observed after TSG101 or EAP20 are knocked down in the context of a co-transfected YPXL mutant virus, suggests that both TSG101 and EAP20 are indeed involved in the budding process (Additional file 3: Figure S3, Additional file 4: Figure S4). The time course release ratio from pulse-chase experiments corroborates this (Fig. 4). The effect seen is less dramatic but pulse chase conditions may be suboptimally sensitive to detection of alterations in viral assembly, which is very rapid [39, 40].
One of the hallmarks of the defect caused by inhibiting the ESCRT pathway is a generalised delayed rate of Gag processing which is especially marked at the final stage of cleavage of p24 from p24-p2 [2, 14, 27, 41]. We consistently see the same defect in EAP20 depleted cells (Fig. 3, Additional file 3: Figure S3) and EAP45 knock out cells (Figs. 5, 7). Similar effects in Gag processing occur in the side by side TSG101 and EAP20 knockdown conditions. The exaggerated effect of a PTAP mutant transfected into the EAP45 knock out cell line (Fig. 7d) implies that ESCRT-I and ESCRT-II may play similar or complementary roles at the early stage of virus budding. Although not cell lines for which HIV is tropic, all of these monolayer cells allow a precise assessment of a single step viral life cycle and facilitate standardisation of the infectivity assay. In addition suspension cells such as lymphocytes are less easily synchronised and cell division might be affected by ESCRT-II knockdown, which would likely impinge directly on HIV replication. The findings from gene depletion experiments (Figs. 1, 3, 4, 5, 7) together with a consistency of effect using a dominant negative interference approach (Fig. 2) strongly suggest that ESCRT-II contributes to efficient virus budding and production of infectious HIV-1 and that the effect of ESCRT-II is mediated through its interaction with ESCRT-I.
It has been suggested that the self-assembly of Gag is sufficient to deform the membrane into a bud making a role for ESCRT-II dispensable. However three dimensional analysis of the viral budding site argues that ESCRT contributes to membrane deformation during the early stage of Gag assembly and that membrane scission is not the only role of ESCRT . ESCRT-II participates in membrane deformation as shown by an in vitro reconstitution assay  and by modelling of the crystal structure of ESCRT-II-CHMP62 with membranes . Our EM data confirm that in the absence of ESCRT-II components budding is slowed or arrested and help to resolve some of the controversy surrounding HIV virus release.
The ESCRT system of proteins is involved in more than one process of viral production in HIV-1. ESCRT-II contributes to efficient transcriptional and/or post-transcriptional generation of HIV Gag. Additionally, there is clear evidence that the presence of ESCRT-II makes a significant contribution to the budding and viral export process of the virus and that this involves a co-ordinated link with ESCRT-I.
HeLaM, a derivative of HeLa cells  and 293T cell lines were obtained from ATCC. TZMbl, carrying two HIV-1 LTR-driven reporter genes, firefly luciferase and E. coli β-galactosidase [43, 44], is a HeLa cell clone that stably expresses high levels of CD4 and CCR5 and was obtained from NIH AIDS Research and Reference Reagent Program. The HAP1 EAP45 CRISPR/Cas 9 knockout and control cell lines were constructed and purchased from Haplogen GmbH, Vienna, Austria. They were grown in Iscove’s modified Dulbecco’s medium (IMDM) supplemented with 10% fetal calf serum (FCS). HeLaM, 293T and TZMbl cells were grown in Dulbecco’s modified Eagle medium (DMEM) supplemented 10% FCS. All cell lines were grown at 37°C and 5% CO2 incubator.
pSVC21∆BglII is derived from a pHXB2 infectious clone but with a BglII fragment in the env gene (7041–7621) deleted [45, 46]. The BH10 strain was a kind gift of Michael Laughrea, McGill University, Canada. The analogous BglII deletion provirus was created in BH10. YPXL and PTAP deletion mutants of pBH10∆BglII were constructed using site directed mutagenesis [27, 47]. pCMV-VSV-G encodes the VSV glycoprotein for pseudotyping. pLAI is a full-length molecular clone of HIV-1 strain LAI for the expression of wild type virus . pTER vector  was obtained from van der Wetering (Centre for Biomedical Genetics, The Netherlands) and was used for cloning shRNA into BamHI and XhoI sites to generate individual pTER-shRNA. pGEX4T1-EAP20, pGEX4T1-EAP30, pGEX4T1-EAP45  and pEGFPC2-HSV1TK were obtained for the cloning of genes of interest to pEF vector containing human elongation factor 1α according to Lee et al. . Mutagenesis PCR was used to introduce the IERK 159-162 AAAA mutation in the H0 helix of EAP45 .
The manufacturer’s protocol for CellTiter-Glo Luminescent Cell Viability Assay (Promega) was adapted for 96-well half-area plates. Briefly, the cell culture plate was equilibrated at room temperature for 30 min. CellTiter-Glo Reagent (50 µl) was added to each well and the plate was placed on an orbital shaker (750 rpm, Titramax 100, Heidolph) for 2 min to lyse cells. After a further 10-min incubation at room temperature, luminescence was read with a 1 s integration time using GloMax 96 Microplate Luminometer (Promega).
Pseudotyped and wild type virus production with shRNA or EAP45 expression
To produce pseudotyped virus, HeLaM cells of 50–80% confluency were co-transfected in a 24-well plate format with 44 ng pSVC21ΔBglII and 15.6 ng pCMV-VSV-G in an optimised ratio of 3.7:1.3 using Fugene HD (Roche). A total amount of 200 ng pBH10∆BglII and pCMV-VSV-G was co-transfected into 50–80% HAP1 cells in a 24-well plate using TurboFectin 8.0 (Origene). For wild type virus, cells were transfected with 40 ng pLAI. Transient expression of shRNA was performed by co-transfecting pTER-shRNA with the viral plasmids. For pseudotyped virus, 120 ng shRNA expression plasmid together with the above mentioned quantities of DNA plasmids for pseudotyping were transfected. For wild type virus, 50 ng shRNA expression plasmid and 40 ng LAI were co-transfected. In the EAP45 mutant experiments HeLaM were transiently co-transfected with 200 ng EAP45 mutant expression plasmids. 200 ng EAP45 mutant expression plasmids was also used for the wild type virus study. In all the above experiments, viruses were harvested at 48 or 96 h post-transfection and the supernatant was used to infect TZMbl cells to assay for infectivity. The medium was replaced 24 h prior to harvesting at 96 h to study the released viruses at late point. Supernatant was inactivated in Empigen (0.1%; Sigma-Aldrich) in PBS. Virus-producing cells were lysed in 150 μl/well of 1% Empigen in Passive Lysis Buffer (Promega). Extracellular and intracellular CA-p24 levels were quantified by ELISA (Aalto) with a slight modification .
Infectivity was based on the Tat-dependent upregulation of LTR-driven firefly luciferase expression upon HIV-1 infection of TZMbl cells. Cells (~105) were seeded in 24-well plates in 0.5 ml/well of complete DMEM. Equal volumes of medium from virus-producing cells were added to TZMbl cells. DEAE-dextran (50 µg/ml; AppliChem) was added to facilitate infection. TZMbl cells were lysed at 48 h post-infection in 150 µl/well of Cell Culture Lysis Reagent (CCLR; Promega). Firefly luciferase expression was quantified using the Luciferase Assay System (Promega). Cell lysates (5 µl) were transferred to a white 96-well half-area plate. Luciferase Assay Reagent (25 µl) was added and the luminescence was measured by GloMax 96 Microplate Luminometer (Promega).
Around 1 × 105 either HAP1 control or EAP45 KO cells were seeded in each well of an 8 well multi-well glass slide (Millipore). After 24 h, the cells were infected with HIV-VSV-G containing 200 ng supernatant p24. After 24 h infection, the media was removed and washed once with PBS before fixing with formaldehyde in PBS (3.7%) for 10 min at room temperature. This was followed by washing with PBS and permeabilising with 70% ethanol at 4°C for 1 h. The probe set composing 37 oligonucleotides labelled with Quasar® 570 Dye at the 5′ end was purchased from Biosearch Technologies specifically targeting HIV Gag coding region. The hybridisation process was followed according to manufacturer’s instructions. The vectashield was applied before the coverslip was mounted and visualised under Olympus Ix81 fluorescence microscope. The TRITC channel was used for genomic RNA visualisation. The images were taken by pre-setting at 300 ms exposure time. For fluorescence quantitation, the genomic RNA positive cells were identified and the area of each individual cell was defined and quantified using ImageJ (NIH).
The Stealth siRNA (Life technologies) knockdown was performed using TransIT-TKO transfection reagent (Mirus) based on the manufacturer’s instructions. Briefly, one well of 50% confluent 293T cells in a 24 well plate was transfected with 2 µl, 20 µM Stealth siRNA to EAP20, TSG101 or control using 2.5 µl TransIT-TKO. Total amount of 40 ng pBH10∆BglII and pCMV-VSV-G was mixed with 160 ng pBluescript empty vector and followed by co-transfection with second dose of 2 µl, 20 µM siRNA 24 h post first dose transfection. The supernatant was collected 72 h post first dose transfection and clarified by low centrifugation before ELISA for p24 quantitation or Optiprep cushion ultracentrifugation for virion purification for Western analysis. Total cells were lysed with 150 µl 1× CCLR before being subjected to ELISA and Western blot analysis using mouse monoclonal antibodies to TSG101 (Abcam) and to HIV-1 p55/p24 (ARP313, NIBSC) and rabbit polyclonal antibodies to GAPDH (Abcam) and EAP20 (P Luzio, Cambridge Institute for Medical Research, UK). Quantification of Western blots was performed using ImageJ (NIH). For knockdown conditions for the pulse-chase experiments one well of fifty percent confluent 293T cells in a 6 well plate was transfected with 8 µl, 20 µM Stealth siRNA to EAP20, TSG101 or control using 10 µl TransIT-TKO. A total of 2 µg pBH10∆BglII and 135 ng pCMV-VSV-G was co-transfected with second dose of Stealth siRNA as above. After 48 h post first transfection, the cells were labelled with 35S protein labeling mix (Perkin-Elmer) as described in the pulse-chase method section.
Pulse-chase analysis was performed by adapting the method of L’Hernault et al. . Briefly, 293T cells were transfected with 2 µg pBH10∆BglII and 135 ng pCMV-VSV-G in 6-well plates. After 24 h, cells were washed and starved in 0.5 ml of DMEM lacking methionine and cysteine (Met-Cys-free DMEM; Sigma-Aldrich) supplemented with 4 mM glutamine and 10% FCS. Cells were then pulse-labelled for 30 min with 0.5 ml of Met-Cys-free DMEM supplemented with 110 μCi of EasyTag Express 35S protein labeling mix (11 mCi/ml; Perkin-Elmer). The medium was removed and supplemented with fresh complete DMEM with 2 mM Met and 2 mM Cys when time point zero was set. Cells were chased at 90 and 180 min post labelling. At each time point the virus-containing supernatants were harvested and clarified by low speed centrifugation. Cells were lysed with RIPA buffer (50 mM Tris–HCl [pH 7.5], 100 mM NaCl, 1% sodium deoxycholate, 0.1% SDS, 1% Triton X-100, protease inhibitor cocktail [Roche]). The capsid and Gag proteins were immunoprecipitated using 3.25 µg anti-HIV-1 p55/p24 mouse monoclonal antibody (ARP313, NIBSC) at 4°C. Protein G-Sepharose beads (Sigma-Aldrich) were then added to the mixtures for a further 1 h rotation before being spun and washed in RIPA buffer. Finally, the beads were resuspended in 50 μl 2× Laemmli buffer. Samples were boiled for 5 min before being loaded onto a 15% SDS-polyacrylamide gel. Gels were fixed and dried prior to autoradiography. Quantification was performed using ImageJ (NIH).
Electron microscopy of budding virus
The HAP1 control cells and KO cells were transfected as described above. 48 h post transfection, the supernatant was removed and the monolayer of cells was washed with 9% NaCl followed by fixation in 2% glutaraldehyde and 2% formaldehyde in 0.05 M cacodylate buffer overnight. The cells were scraped and collected by centrifugation before being embedded and thin-sectioned. The thin-sectioned samples were mounted on an EM grid and visualised using a Tecnai G2 electron microscope. Images were taken at 3,500× magnification.
BM participated in the design of the study, drafting of the manuscript and carried out molecular and cell biological assays including interfering RNA, cell knockout and EM studies. NI participated in the design of the study and carried out interfering RNA and dominant negative studies and contributed to drafting the manuscript. LP contributed to the viral assays and helped draft the manuscript. TA participated in the design and coordination of the studies and the drafting of the manuscript. AL conceived of the study, participated in its design and coordination and helped write the manuscript. All authors read and approved the final manuscript.
We thank Roger Williams (MRC Laboratory of Molecular Biology, Cambridge, UK) for many helpful discussions. We thank Jeremy Skepper in the Multi-imaging Centre for help with the electron microscopy. We thank Paul Luzio (Cambridge Institute for Medical Research, Cambridge, UK) for the kind gift of EAP20 antibody, Herald Stenmark (University of Oslo, Norway) for providing the antibodies against EAP30 and EAP45, Laurence Tiley (University of Cambridge, UK) for pCMV-VSV-G, Michael Laughrea for providing the BH10 infectious clone (McGill University, Canada) and Marc van de Wetering (Centre for Biomedical Genetics, The Netherlands) for the plasmid vector for shRNA expression. The work was supported by the Cambridge Biomedical Research Centre.
Compliance with ethical guidelines
Competing interests The authors declare that they have no competing interests.
- 27.Dussupt V, Javid MP, Abou-Jaoude G, Jadwin JA, de La Cruz J, Nagashima K et al (2009) The nucleocapsid region of HIV-1 Gag cooperates with the PTAP and LYPXnL late domains to recruit the cellular machinery necessary for viral budding. PLoS Pathog 5(3):e1000339PubMedCentralCrossRefPubMedGoogle Scholar
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.