Inhibition of HIV-1 assembly by coiled-coil domain containing protein 8 in human cells

Human Immunodeficiency Virus type 1 (HIV-1) major structure protein Gag is synthesized in the cytoplasm, assembles on the plasma membrane, subsequently buds and releases. HIV-1 viral particles incorporate a number of host proteins to facilitate or inhibit HIV-1 replication. Here we identify a new host protein, coiled-coil domain containing protein 8 (CCDC8), in HIV-1 particles. Incorporation of CCDC8 into virions is dependent on the interaction between CCDC8 and Gag matrix region. Exogenous overexpression of CCDC8 can strongly inhibit HIV-1 production, up to ~30 fold. CCDC8 is a membrane-associated protein. The interaction between exogenously expressed CCDC8 and Gag on the plasma membrane changes the assembly of Gag, and redirects it into intracellular sites, or causes Gag endocytosis. CCDC8, along with cytoskeleton protein obscuring-like1 (Obsl1) and E3 ligase Cul7, induces Gag polyubiquitination and degradation. Thus we identify a new host protein and a new pathway for HIV-1 Gag polyubiquitination and degradation. This pathway presents potential therapeutic strategies against HIV infection.

Scientific RepoRts | 5:14724 | DOi: 10.1038/srep14724 studies revealed that in patients with 3-M syndrome, approximately 70% have E3 ligase Cul7 mutations, 25% have cytoskeleton protein obscuring-like1 (Obsl1) mutations, and 5% have CCDC8 mutations 11 . Thus, 3-M syndrome links CCDC8, Obsl1 and Cul7, suggesting they could be in the same pathway. In this study, immunoprecipitation assay proves the interaction between Gag and CCDC8, and between CCDC8 and Obsl1 and Cul7. Furthermore, exogenously expressed CCDC8 causes Gag polyubiquitination and degradation. Therefore, we identified a new pathway for Gag polyubiquitination and degradation, although the significance of this pathway is unknown.

Results
Exogenous expression of CCDC8 strongly inhibits HIV-1 production. Using mass spectrometry (MS), we analyzed proteins found in HIV-1 VLPs viral Gag (vGag-RRE), or protease-negative viral Gag/ GagPol-RRE-P-(vGag/GagPol-P-), which produces unprocessed Gag and GagPol. Proteins present in each section of the gel were identified by MS. A dozen of cDNAs to identified proteins of unknown function were synthesized, and cloned into mammalian expression vector pTT5-SH5. HIV-1 Gag CAp24 production was measured upon the expression of above cDNA clones. During screening, only one of these proteins, a low-abundant species identified as coiled-coil domain containing protein 8 (CCDC8), can strongly inhibit HIV-1 CAp24 production. To further characterize the inhibition of HIV-1 production by CCDC8, we co-expressed constant amount of HIV-1 BH10 plasmid (2 μ g) and increasing amounts of plasmid coding for His-tagged pTT5-SH5-CCDC8 in HEK293T cells, and followed by Western blot analysis. We found a drastic reduction in the amount of virus produced upon exposure to CCDC8, compared with a slower rate of reduction in the negative control (Enhanced Green Fluorescent Protein [pTT5-SH5-EGFP] expression) (Fig. 1a,b). This is accompanied by a mild reduction in cellular Gag production, but a much greater reduction in cellular CAp24, suggesting that Gag proteolytic processing could be impaired or Gag degradation could be increased or both. At lower concentrations of transfected plasmid of CCDC8, such as 1-4 μ g, Gag expression is less affected and the extracellular CAp24 is reduced up to ~10-30 fold (Fig. 1a,b). In contrast, Gag and cell CAp24 are much less affected by expression of EGFP, or other cloned host proteins, such as milk-fat globule epidermal growth factor (EGF) factor 8 (MFG-E8), Zinc-alpha-2-glycoprotein (AZGP1), etc. (unshown data).
The experiments described were repeated using a protease-negative virus (HIV-1 BH10 P − ). Again, co-expression of CCDC8 with HIV-1 BH10 P − resulted in a very strong reduction in virus production accompanied by a milder reduction in cellular Gag (Fig. 1c). This data indicate that the inhibition of HIV-1 production by CCDC8 is independent on viral protease activity. Thus, the rapid reduction in CAp24 seen in Fig. 1a,b is probably not the primary factor responsible for the decrease in viral production, but could reflect a reduction in Gag proteolytic processing or an increase in Gag degradation. Meanwhile, we found that 8 μ g plasmid of CCDC8 or an HIV-1/CCDC8 ratio at 1:4 in the transient transfection inhibited the expression of Gag (Fig. 1a,c). Thus, in the later experiments, we used 1-4 μ g plasmid of CCDC8 or a Gag/CCDC8 ratio at 1:1 or 1:2.
Moreover, without affecting expression of Gag, CCDC8 also inhibits vGag/GagPol-RRE or vGag-RRE production, which encodes only viral Gag/GagPol or viral Gag without any regulatory proteins (Fig. 1d). This data support the notion that the inhibition of HIV-1 production by CCDC8 is independent on viral protease activity. Meanwhile, the CCDC8 inhibition efficiency for vGag/GagPol-RRE is comparable to the HIV-1 whole genome containing plasmid BH10 ( Fig. 1a-d), suggesting that HIV-1 does not encode regulatory proteins to antagonize CCDC8. It is unlike APOBEC3G, which activity is antagonized by HIV-1 regulatory protein Vif 7 .
Furthermore, CCDC8 has no effect upon the production of β -actin ( Fig. 1c) or the production of exogenous V5-tagged lysyl-tRNA synthetase (Fig. 1e). These data suggest that the effect of exogenous overexpression of CCDC8 upon extra-cellular virus production is specific.
HIV-1 inhibition by CCDC8 was also repeated in other cell lines, such as Hela cells. However, the inhibitory effect of CCDC8 overexpression on HIV-1 production was less effective in Hela cells than in HEK293T cells (unshown data).
CCDC8 also reduces viral infectivity. Viral infectivity upon the overexpression of CCDC8 was measured in Hela-TZMbl indicator cell line. Either the same volume of virus-containing culture supernatant or the same amount of CAp24 was used to infect Hela-TZMbl cells. When the same volume of viral supernatant was used, infectivity of HIV-1 upon the overexpression of CCDC8 was reduced >17 fold, compared to the negative control EGFP (Fig. 1f). This will partially reflect a reduction in production of extracellular viral particles or reduced viral infectivity per particle. Indeed, upon infection of cells with normalized amount of CAp24 (2 ng), we found that infectivity per particle was still reduced by 6-7 fold (Fig. 1g).
CCDC8 interacts with HIV-1 Gag matrix. Next, we examined the ability of endogenous CCDC8 to be packaged into HIV-1 viral particles. Endogenous CCDC8 is incorporated into both Gag VLPs and protease-negative BH10 P- (Fig. 2a). In protease-positive BH10 virus, CCDC8 is processed, probably by viral protease (Fig. 2a). We also asked whether CCDC8 is on the virion surface or within the virion particles. Isolated BH10 P-virions were exposed to subtilisin protease before lysis (Fig. 2b). Subtilisin does not penetrate the viral membrane, but processes gp120 and any contaminant proteins present on    VLPs Virus S1 P100 S100 S1 P100 S100 HIV-1 P + EGFP HIV-1 P + C8 the external side of the viral membrane. These results confirm the presence of CCDC8 within the virions (Fig. 2b). To further map the region of interaction between Gag and CCDC8, we constructed a Gag mutant lacking all but the first 8 amino acids in matrix (Fig. 2c). This mutant keeps the myristoylation signal at the N terminus of Gag and can still bind to membrane and produce extracellular Gag VLPs. In Fig. 2c, Western blots of viral lysates probed with either anti-Gag or anti-CCDC8 show that endogenous CCDC8 is no longer found in VLPs containing the matrix deletion in Gag, suggesting that CCDC8 may interact with the matrix region in Gag.
CCDC8 is acting on the post-translational step of HIV-1 replication. CCDC8 causes a strong reduction in virus production ( Fig. 1a) that is associated with a much weaker or no reduction in the production of cellular Gag ( Fig. 1a-d). This suggests that a major cause of reduced virus production is posttranslational, between Gag synthesis and virion release. We next examined whether CCDC8 could prevent Gag binding to the membrane for assembly. Cell lysates expressing HIV-1 with CCDC8 or with the control plasmid EGFP were subjected to Dounce homogenization and centrifuged at low speed to remove nuclei and unbroken cells, and the subsequent supernatant (S1) was further ultra-centrifuged, resulting in supernatant (S100) and pellet (P100). Membrane-associated Gag is enriched in P100 (Fig. 2d). Despite CCDC8 overexpression, the Gag in P100 does not significantly change compared to the control (compare the middle lanes in Fig. 2d). We further separated the P100 through membrane flotation sucrose gradient. The pattern of p55 Gag in each section of sucrose gradient from top to bottom has no apparent alteration either, despite the overexpression of CCDC8 or not (Fig. 2e). The membrane-associated Gag mainly exists in the section 2 and 3, as indicated by the membrane markers Caveolin-1 and transferrin receptor (TFR) (Fig. 2e). This suggests that overexpression of CCDC8 does not prevent Gag binding to the membranes, which include internal membrane and plasma membrane. Interestingly, CCDC8 also co-exists with membrane marker Caveolin-1 and TFR in the section 2 and 3 in the sucrose gradient (Fig. 2e).
To further confirm the localization of CCDC8, we used confocal microscopy. Figure 3a clearly shows that mCherry-tagged CCDC8 (red) localizes on the plasma membrane, where Gag assembles (green) (vGag-GFP, NIH AIDS Research Reagent Program, Catalog number 11468, from Dr. Andrew Mouland). With or without HIV-1 Gag, CCDC8 is mainly present on the membrane, and only a small part is present in the cytoplasm ( Fig. 3a and unshown data).

CCDC8 causes an internal accumulation of Gag particles in the cytoplasm.
Comparing the appearance of vGag-GFP co-expressed with CCDC8 or with empty vector (Fig. 3a,b), we found that for cells expressing vGag-GFP and empty vector, vGag-GFP is present homogeneously in the cytoplasm, representing Gag monomer or low-order Gag multimers, and on the plasma membrane as bright vGag-GFP puncta, representing high-order Gag multimers or assembly particles (Fig. 3b) 12 . When Gag and CCDC8 are co-expressed, there is an accumulation of bright vGag-GFP puncta (see the cytoplasm in the cell in Fig. 3a left panel) at 24 hours post-transfection, indicating that CCDC8 probably changes Gag trafficking. There are two possible explanations for the data; one is that Gag is not transported to the plasma membrane, but directly accumulates in the intracellular sites; the other is that Gag is first transported to the membrane but is re-internalized, or redirecting to the intracellular sites. The membrane flotation sucrose gradient data, which show that CCDC8 does not prevent Gag binding to the membrane, support the latter possibility.
Internalization of Gag from the plasma membrane results from endocytosis. To examine the kinetics of Gag accumulation, we traced the vGag-GFP at various time points after transfection. The transfected HEK293T cells were fixed and observed under confocal microscopy. At 10 hours after co-transfection of CCDC8 and vGag-GFP, around 84% of the cells containing vGag-GFP showed exclusively Gag localization at the plasma membrane. Only ~16% of the cells displayed intracellular accumulation of vGag-GFP puncta (Fig. 3c,d). As time passing, more cells presented intracellular vGag-GFP puncta accumulation. At 16 h post transfection of CCDC8 and vGag-GFP, the cells with intracellular vGag-GFP puncta accumulation peaked at ~71%, and then declined slowly (Fig. 3c,d). Notably, the visible accumulation of vGag-GFP puncta at the plasma membrane significantly preceded its accumulation in intracellular sites (Fig. 3c,d), suggesting that Gag is first transported to the plasma membrane and subsequently re-internalized. In contrast, upon co-expression of vGag-GFP and empty vector, the proportion of cells with intracellular vGag-GFP puncta accumulation remains constant throughout time at roughly 10% (Fig. 3c,d).
Because previous data showed that Gag assembly occurs on the plasma membrane 12 , these data suggest that CCDC8 changes Gag trafficking. Namely, after Gag meets CCDC8 on the plasma membrane, Gag is internalized instead of budding out. This is in accordance with the membrane flotation data that Gag still binds to the membrane upon overexpression of CCDC8 (Fig. 2e). It is possible that Gag localizes to the membrane because of a low level of CCDC8 at the beginning of transfection. To test this, we performed CCDC8 transfection 10 hours earlier than Gag transfection to allow for a CCDC8 build-up (Supplementary Figure 1a). The scenario appeared again that visible accumulation of vGag-GFP puncta at the plasma membrane significantly preceded its accumulation in intracellular sites (Supplementary Figure 1a). These data support the notion that Gag internalization is facilitated by its  interaction with membrane-associated CCDC8. To further confirm CCDC8-mediated Gag internalization, we co-expressed CCDC8 and vGag-GFP with dominant negative Dynamin-K44A, an endocytosis inhibitor. Dynamin-K44A significantly reduced the CCDC8-mediated Gag internalization (P = 0.0005, Fig. 3e, Supplementary Figure 1b), suggesting that CCDC8-mediated Gag internalization results from endocytosis. It appears that the meeting of Gag with CCDC8 on the plasma membrane is a prerequisite for Gag endocytosis. Virion endocytosis usually goes into endolysosome pathway and virions always accumulate in endolysosomes. Thus, we determined whether CCDC8 induced this phenotype. vGag-GFP was co-expressed with His-tagged CCDC8, and cells were fixed and stained with antibody against CD63, an endolysosome marker. Supplementary Figure 2a shows that the CCDC8-mediated internalized vGag-GFP does not colocalize with CD63 (3D structure, Pearson's colocalization coefficient ~0.1). This data suggest that CCDC8-mediated endocytosis of Gag does not go into endolysosome pathway.
Gag undergoes CCDC8-dependent polyubiquitination and degradation. Since the reported role of CCDC8 mutations in 3-M syndrome links it to Obsl1 and E3 ligase Cul7, we tested whether Obsl1 or Cul7 could also inhibit HIV-1 production. Unlike CCDC8, co-expression of HIV-1 with Obsl1 or Cul7 does not inhibit the production of Gag CAp24 (Fig. 4a).
Next, we checked the metabolic half-life of Gag upon overexpression of CCDC8 using a pulse-chase experiment. The culture medium was complemented with the same amount of radioisotope 35 S-labeled methionine and cysteine, and Gag in cell lysate was chased using immunoprecipitation at the indicated times (Fig. 5c,d). Normally, HIV-1 Gag will bud out and 35 S-labled Gag will decrease over time. CCDC8 apparently reduced the half-life of Gag protein (Fig. 5c,d). The half-life of Gag upon overexpression of CCDC8 is around ~20 minutes, compared to ~80 minutes in the counterpart. This indicates that CCDC8 accelerates Gag degradation.

Discussion
This study identifies a new virion-incorporated protein, CCDC8, and its associated pathway that inhibits HIV-1 assembly, and triggers HIV-1 Gag endocytosis, polyubiquitination, and degradation. CCDC8 is a membrane-associated protein, and exogenous overexpression of CCDC8 results in a strong reduction in HIV-1 virion release.
Different techniques were used in this study, including Western blot, imaging test, immunoprecipitation and pulse-chase test. All these data were in consistence and suggested the conclusion of this study. In immunofluorescence test, Gag was tagged GFP and CCDC8 was tagged mCherry. The bright Gag-GFP puncta represent high-order Gag multimers or Gag assembly complexes because of the intensity of their fluorescent signal. The loss of extracellular HIV-1 Gag production appears to be due to their internalization by endocytosis, as demonstrated in Fig. 3e using the endocytic inhibitor, Dynamin-K44A. The loss may be due to CCDC8's interaction with the matrix region in Gag, which is responsible for the incorporation of CCDC8 into the virions. Moreover, inhibition of Gag assembly by CCDC8 is dependent on its elevated expression in a dose-dependent manner, since millions of Gag proteins bind to the plasma membrane for assembly. Gag proteolytic processing to release CAp24 occurs after the Gag arrival to the plasma membrane for assembly 1,15 . This explains an observed rapid decrease in CAp24, and a slow decrease in p55 Gag in the cell lysate when CCDC8 was co-expressed with HIV-1 (Fig. 1a,b). The apparent inhibition of CAp24 production in the cell lysate by CCDC8 is more likely a second effect of assembly inhibition.
Endocytosis refers to a process of invagination of the cargo protein and cell membrane to form a pocket, internalization and membrane scission 16,17 . Clathrin-dependent endocytosis is typical. However, ubiquitin-dependent endocytosis has also been described; for example, ubiquitin-mediated endocytosis has been reported for the membrane-bound epidermal growth factor (EGF) receptor (EGFR). When stimulated with low-doses of EGF, EGFR is internalized almost exclusively through the clathrin-dependent pathway, while at higher concentrations of EGF, endocytosis of EGFR is through a clathrin-independent, ubiquitin-dependent pathway 18,19 . This also implies the possibility that CCDC8-mediated Gag polyubiquitination could be the triggering signal for Gag endocytosis.
In this study, we report that Gag is polyubiquitinated, even though mono-and di-ubiquitination of Gag p6 for promoting virus budding have been reported 14,20 . Dr. Ulrich Schubert group reported Gag polyubiquitination in Gag p6 mutants S40F 21 and deltaPTAP 14 , but those were not wild-type Gag. When this manuscript was in preparation and submission, Nityanandam reported that BCA2/Rabring7 targets Gag and induces Gag polyubiquitination 22 . In this study, Gag polyubiquitination is CCDC8-dependent (Fig. 5), and associated with the CCDC8 pathway proteins 10,11 . Due to the association of CCDC8 with the plasma membrane (Figs 2 and 3), we theorize the following scenario: multimer Gag assembly units interact with CCDC8, Obsl1, and Cul7 at the plasma membrane, causing Gag polyubiquitination. The polyubiquitination signal might trigger Gag endocytosis, and then Gag complex enters into 26S proteasome, resulting in an eventual degradation of the complex by the ubiquitin-proteasome system (Fig. 5e  model). However, we still cannot exclude another possibility that polyubiquitination of Gag occurs in the cytoplasm, since Cul7 localizes in the cytoplasm too. The relationship between CCDC8-mediated Gag polyubiquitination and endocytosis is still elusive.
It is unknown what mechanism underlies the growth-related disorder 3-M syndrome. The serum growth hormone (GH) level of 3-M patients is usually normal, and insulin-like growth factor 1(IGF-1) is normal or low 10 . An intriguing recent finding is that Cul7-Fbw8 can degrade the insulin receptor substrate 1 (IRS-1) through Skp1-Fbw8-ROC1 (SCF complex), suggesting the mechanism of 3-M syndrome is related to IRS-1 23 . As well, transcription microarray data showed that the insulin-like growth factor 2 (IGF-2) is down-regulated in 3-M patients 24 . Therefore, we speculate that 3-M syndrome is related to a disorder of insulin-like growth factor or receptor signaling. Interestingly, another 3-M related protein, Scientific RepoRts | 5:14724 | DOi: 10.1038/srep14724 insulin-like growth factor II mRNA binding protein 1 (IMP1, or IGF2BP1), was reported to inhibit HIV-1 assembly 25,26 . It may be informative to investigate more generally the potential of 3-M syndrome pathway proteins in the inhibition of HIV-1 assembly.
Endogenous expression of CCDC8 is similarly low in many human tissues. In some cell lines, such as Jurkat T lymphocyte and HEK293 cells, the endogenous expression is a little higher (GeneCard, http:// www.genecards.org/cgi-bin/carddisp.pl?gene= CCDC8). Apparently, no human cells or cell lines endogenously express sufficiently high levels of CCDC8 to accomplish HIV-1 inhibition. Using siRNA technique, we down-regulated the expression of CCDC8, and found that HIV-1 Gag production was not significantly or modestly altered in HEK293T cells. Therefore, this study is not able to address the role of endogenously expressed CCDC8.
However, it is well accepted that proteasomal degradation of pathogen-derived proteins plays a key role in antigen presentation by MHC class I molecules 14 . Intriguingly, CCDC8 inhibits the viral Rev-RRE source native-codon Gag assembly, but not humanized or synthesized-codon Gag. Namely, CCDC8 can distinguish between "self " and "non self " HIV-1 Gag (unshown data). This mechanism is central to immune system recognition of pathogens. It is possible that the human immune system can "sense" HIV-1 attack by CCDC8-medicated Gag proteasomal degradation to present antigens to the immune system.
The CCDC8 gene is highly conserved among mammalian species, including human, chimpanzee, rhesus monkey, dog, cow, mouse, and rat (http://www.ncbi.nlm.nih.gov/gene/83987). CCDC8 was also reported to be a cofactor required for p53-mediated apoptosis following DNA damage in humans 27 . Therefore, CCDC8 is a very important factor relevant to cell growth, apoptosis, and probable anti-viral infections, although its exact functions are unknown. Further work is needed for clarifying its roles in all of these cellular processes.
In conclusion, this finding sheds light on a new pathway for inhibiting HIV-1 assembly. It presents a potential target for pharmaceutical drug development against HIV-1.

Methods
All experiments were carried out in accordance with protocols approved by

Mass Spectrometry analysis. The viral lysates were resolved by SDS-PAGE gel and stained by
Coomassie-blue. Samples were cut into small pieces and be analyzed by nanoliquid chromatography/ tandem mass spectrometry (nano-LC-MS/MS) in Institute For Research in Immunology and Cancer (IRIC), University of Montreal, Montreal, Canada.

Membrane flotation analysis.
Membrane flotation analysis is performed as previously described 28 .
Briefly, forty-eight hours posttransfection, the cells were treated with hypotonic Tris-EDTA (TE) buffer (20mM Tris-HCl ph7.4, 1 mM EDTA, 0.01% β -mercaptoethanol) supplemented with a protease inhibitor cocktail (Roche). After Dounce homogenization, the cell homogenate was then centrifuged at 1500 g for 30 minutes to remove the nuclei and unbroken cells. The supernatant (S1) was then centrifuged for 1 h at 100,000 g in SW55Ti rotor (Beckman) at 4 °C, resulting in the supernatant (S100) and the pellet (P100). The P100 was resuspended in TNE buffer (20 mM Tris-HCl ph 7.5, 100 mM NaCl, 1 mM EDTA). A total of 200 μ l of each lysate was mixed with 1ml of 85.5% sucrose (w/v) in TNE buffer, placed at the bottom of ultracentrifuge tubes, and overlaid with 2.5 ml of 65% sucrose and 1.5 ml of 10% sucrose. The samples were centrifuged at 4 °C in a SW55Ti rotor for 16 hours at 35,000 rpm. Nine equal fractions were collected from the top to the bottom. in the figures. After the indicated time posttransfection, the cells were fixed in 4% paraformaldehyde for 5 min and followed by cold methanol at − 20 °C for 15 min. The cells stained with or without antibody, and DAPI, and examined using Zeiss confocal microscope.
Metabolic labelling and pulse chase experiment. HEK293T cells were transfected with vGag-RRE, Rev, empty vector or CCDC8. Two days post-transfection, the cells were washed with PBS, and starved for 30 minutes in methionine-cysteine free medium supplemented with 2% dialyzed fetal bovine serum (Invitrogen). Then, the cells were labeled for 20 min in medium containing 200 μ Ci [ 35 S] methionine and cysteine (Express [ 35 S] Protein Labelling Mix, NEG07200, Perkin Elmer Life Sciences). The pulse was ended by adding prewarmed DMEM supplemented with 10% fetal bovine serum and 5 mM of unlabeled methionine and cysteine. After various chase time periods, the cells were pelleted on ice and washed with ice cold PBS. For immunoprecipitation, the cells were lysed, and Gag was immunoprecipitated with anti-Gag antibody. The proteins were separated by 10% SDS-PAGE gel. The gel was fixed by buffer containing 40% ethanol, 10% acetic acid, 10% glycine at room temperature for 2 hours and amplified (NAMP 100, GE life Science) for 15 minutes, dried the gel and put the film on − 80 °C freezer, then quantified by ImageJ.