Defects in cellular sorting and retroviral assembly induced by GGA overexpression
© Joshi et al; licensee BioMed Central Ltd. 2009
Received: 9 June 2009
Accepted: 29 September 2009
Published: 29 September 2009
Skip to main content
© Joshi et al; licensee BioMed Central Ltd. 2009
Received: 9 June 2009
Accepted: 29 September 2009
Published: 29 September 2009
We previously demonstrated that overexpression of Golgi-localized, γ-ear containing, Arf-binding (GGA) proteins inhibits retrovirus assembly and release by disrupting the function of endogenous ADP ribosylation factors (Arfs). GGA overexpression led to the formation of large, swollen vacuolar compartments, which in the case of GGA1 sequestered HIV-1 Gag.
In the current study, we extend our previous findings to characterize in depth the GGA-induced compartments and the determinants for retroviral Gag sequestration in these structures. We find that GGA-induced structures are derived from the Golgi and contain aggresome markers. GGA overexpression leads to defects in trafficking of transferrin receptor and recycling of cation-dependent mannose 6-phosphate receptor. Additionally, we find that compartments induced by GGA overexpression sequester Tsg101, poly-ubiquitin, and, in the case of GGA3, Hrs. Interestingly, brefeldin A treatment, which leads to the dissociation of endogenous GGAs from membranes, does not dissociate the GGA-induced compartments. GGA mutants that are defective in Arf binding and hence association with membranes also induce the formation of GGA-induced structures. Overexpression of ubiquitin reverses the formation of GGA-induced structures and partially rescues HIV-1 particle production. We found that in addition to HIV-1 Gag, equine infectious anemia virus Gag is also sequestered in GGA1-induced structures. The determinants in Gag responsible for sequestration map to the matrix domain, and recruitment to these structures is dependent on Gag membrane binding.
These data provide insights into the composition of structures induced by GGA overexpression and their ability to disrupt endosomal sorting and retroviral particle production.
The Gag polyprotein precursors are the key structural elements driving retroviral particle production. The N-terminal matrix (MA) domain of the Gag precursor is important for plasma membrane (PM) targeting and membrane binding. Following Gag targeting, the process of assembly proceeds via Gag multimerization mediated primarily by sequences within the capsid (CA) and nucleocapsid (NC) domains. Finally, the "late" domains in Gag mediate the terminal step in particle production - the pinching off of the virion from the infected cell membrane [1–5]. Concomitant with virus release the particle undergoes maturation, a structural reorganization of the virion that results from a highly concerted catalytic cascade mediated by the viral protease (PR) [5, 6].
While the Gag precursor proteins are the sole viral determinants required for the production of immature virus-like particles (VLPs), a number of host factors have been implicated in various steps of the virus assembly and release pathway. Retroviral late domains are known to interact with components of the endosomal sorting machinery. For example, the HIV-1 Gag precursor protein, Pr55Gag, contains in its p6 domain a Pro-Thr/Ser-Ala-Pro [P(T/S)AP] motif that binds Tsg101, a component of the endosomal sorting complex required for transport-I (ESCRT-I) [7–10], and a Tyr-Pro-Xn-Leu (YPXnL, where X is any amino acid and n = 1-3 residues) motif that interacts with the ESCRT-associated factor Alix [11–13]. The physiological role of the ESCRT machinery is to promote the biogenesis of vesicles that bud into the lumen of late endosomes to form multivesicular bodies (MVBs) . The recruitment of Tsg101 from the cytoplasm to endosomal membranes occurs via interaction between Tsg101 and hepatocyte growth factor-regulated tyrosine kinase substrate (Hrs) [15–17]. The delivery of cargo proteins to MVBs usually requires the recognition by ESCRT machinery of monoubiquitin moieties attached to the cytoplasmic domains of the cargo. In yeast, disruption of any of the components of the ESCRT complexes and associated factors (known as class E VPS proteins) results in failure of proper sorting of ubiquitinated proteins and induction of an aberrant class E compartment [18–20]. Retroviral Gag proteins are also ubiquitinated [21–23], and while ubiquitination of Gag does not appear to play an essential role in virus budding , in the absence of a functional late domain ubiquitin can serve to promote virus release .
In addition to serving a well-established role in retrovirus budding, host cell factors have also been reported to function in promoting Gag trafficking to the PM. We previously reported that the phospholipid phosphatidylinositol-(4,5)-bisphosphate [PI(4,5)P2] is a key cellular cofactor for HIV-1 Gag targeting to the PM  via a direct MA-PI(4,5)P2 interaction [27, 28]. The clathrin adaptor protein complexes 1, 2, and 3 (AP-1, 2, and 3) [29–31], suppressor of cytokine signaling 1 (SOCS1) , the kinesin KIF4 [33, 34], staufen 1 , and plenty of SH3s (POSH)  have all been implicated in Gag targeting to the PM. We demonstrated previously that overexpression of the Golgi-localized, γ-ear containing, Arf-binding (GGA) proteins inhibits the production of HIV-1 and equine infectious anemia virus (EIAV) particles by impairing the association of Gag with membrane . The impairment in Gag-membrane binding induced by GGA overexpression was linked to functional disruption of the endogenous ADP ribosylation factors (Arfs) .
The GGA proteins are a family of monomeric clathrin adaptors primarily localized at the trans-Golgi network (TGN), although also reported to be present in late endosomes [38–41]. Three GGA proteins (GGA1, 2, and 3) are expressed in mammals and two in yeast and there is ample evidence demonstrating that these proteins help package cellular cargo into clathrin-coated vesicles . GGAs are comprised of four distinct domains: 1) an N-terminal Vps27, Hrs, and STAM homology (VHS) domain that binds cargo proteins and cargo protein receptors [e.g., cation-dependent mannose 6-phosphate receptors (CD-MPR)] bearing Asp-X-X-Leu-Leu [DXXLL] motifs; 2) a GGA and Tom (GAT) domain that binds Arfs, ubiquitin, and Tsg101; 3) a hinge region that recruits clathrin; and 4) a C-terminal γ-adaptin ear homology (GAE) domain that binds several proteins including rabaptin 5, epsinR, and γ-synergin. The GAT domain is responsible for recruitment of GGA proteins to membrane via interaction with GTP-bound Arf .
Previously, we reported that GGA overexpression induces the formation of large swollen vacuolar compartments that sequester Arf proteins. The compartments induced by GGA1, but not those induced by GGA2 or GGA3, also sequester HIV-1 Gag . We and others have also demonstrated in previous studies that overexpression of dominant-negative or full-length components of the ESCRT complexes and associated machinery gives rise to the formation of aberrant compartments that disrupt retroviral particle production. For example, overexpression of full-length Tsg101 (TSG-F) induces a class E-type compartment that potently disrupts HIV-1 budding but has minimal effect on EIAV release [42, 43]. Overexpression of the C-terminal portion of Tsg101 (TSG-3') generates aggresome-like structures and severely inhibits the release of HIV-1, murine leukemia virus (MLV), and Rous sarcoma virus (RSV) but does not disrupt EIAV release [42–44]. A dominant-negative, ATPase-deficient mutant of Vps4 is highly disruptive to the release of a number of retroviruses including HIV-1, EIAV, RSV, and feline immunodeficiency virus (FIV) [8, 11, 25, 43, 45–47]. It is noteworthy that each of the compartments described above is morphologically distinct and imposes selective defects on specific steps in the retroviral assembly and release pathway. For example, GGA overexpression leads to a defect in Gag trafficking to the PM, whereas TSG-F, TSG-3', and dominant-negative Vps4 overexpression inhibits particle budding and release.
In the current study, we characterized the compartments induced by GGA overexpression with regard to their composition as well as their impact on retrovirus assembly and cellular endosomal sorting pathways. Our data demonstrate that GGA overexpression causes various sorting defects as measured by recycling of CD-MPR, internalization of transferrin receptor (TfR), and the subcellular localization of proteins like Tsg101, ubiquitin, and Hrs. The determinants for HIV-1 Gag sequestration in GGA1-induced compartments were mapped to the MA domain. These data provide novel insights into the defects in cellular sorting and retrovirus assembly induced by GGA overexpression.
We next sought to define the composition and subcellular origins of the GGA-induced structures. GGA-overexpressing cells were stained with antibodies specific for the endoplasmic reticulum (ER)-resident proteins calnexin and calreticulin, the trans-Golgi network (TGN) marker TGN46, the Golgi matrix marker GM130, the early endosome marker early endosomal antigen 1 (EEA1), and the late endosome marker CD63. We observed that the GGA-induced structures did not contain ER or late endosome markers (data not shown) but did display a high degree of colocalization with TGN46 (Pearson coefficient of correlation R ~ 0.7-0.8) (Figure 1B). Interestingly, only the GGA3-induced structures stained for GM130 (R ~ 0.8), a detergent-insoluble component of the Golgi matrix peripherally associated with the cis compartment (Figure 1B) . These results suggest that GGA overexpression induces the formation of aberrant compartments derived from the Golgi but the specific composition of compartments induced by overexpression of GGA1, GGA2, or GGA3 is not identical.
Previously we showed that GGA overexpression modulates retrovirus assembly by inhibiting Gag targeting to the PM. GGA overexpression also induced the accumulation of large vacuolar structures, which in the case of GGA1 sequestered HIV-1 Gag . In the current study, we investigated in detail the morphology and composition of GGA-induced compartments and the impact of their formation on normal cellular physiology and retrovirus release. We observed that GGA-induced compartments stained positively for Golgi and aggresome markers, and GGA overexpression caused defects in endosomal sorting. Moreover, GGA overexpression also altered the localization of various host proteins important for retrovirus assembly. Overexpression of ubiquitin led to a dissolution of the GGA-induced structures and a partial rescue of HIV-1 release. Study of determinants for HIV-1 Gag sequestration in GGA1-induced compartments demonstrated a role for the MA domain. Thus, GGA overexpression not only inhibits retrovirus assembly and Gag localization but also alters the cellular sorting pathways important for normal cellular physiology.
GGA proteins when expressed exogenously in cells whether transiently or stably have been reported to show a localization pattern similar to that of their endogenous counterparts [48, 49]. Indeed, tagged GGA expression vectors have been widely used in the field of cell biology to substitute for the use of GGA antibodies. However, we observe defects in endosomal sorting induced by even moderate levels of GGA overexpression and the formation of vacuolar compartments in which the exogenously expressed GGA proteins are localized. This phenotype is similar to the GGA expression pattern reported previously, namely, "compaction and fizzling" of Golgi stacks and accumulation of fragmented vacuolar-like blobs that contain Golgi markers [52, 68, 69]. GGA overexpression also led to striking changes in TfR distribution and recycling of CD-MPR. Following ligand binding to the TfR, both receptor and ligand are rapidly internalized and transported to acidic compartments in which the receptor is released and recycled back to the cell surface [70, 71]. However, in cells overexpressing the GGA proteins, there was no evidence of the bright cell surface staining with fluorescently tagged transferrin that is observed in cells not overexpressing the GGAs. This loss in transferrin binding was due to sequestration of TfR in GGA-induced compartments. GGA overexpression led to the accumulation of CD-MPR in the Golgi with a concomitant depletion from the periphery. This alteration in CD-MPR recycling has also been described previously in cells expressing a dominant-negative fragment of GGA1 comprising the VHS-GAT domain but lacking the hinge and GAE domains .
GGA overexpression also led to alterations in the subcellular localization of several host proteins implicated in retrovirus assembly and release. We previously observed that GGA overexpression led to sequestration of Arfs . Moreover, GGA overexpression also caused a shift in the localization of overexpressed Tsg101 and Hrs (Figure 5). Of particular interest were the observations that GGA-induced compartments sequestered ubiquitinated cargo and providing exogenous ubiquitin not only resolved the GGA-induced structures but also rescued HIV-1 release (Figure 6). We speculate that GGA-induced structures are formed as a result of misfolded protein accumulation, which due to ubiquitin sequestration are incapable of being targeted for proteasomal degradation [73, 74]. Providing exogenous ubiquitin could facilitate destruction of misfolded proteins leading to the disappearance of GGA-induced structures. Interestingly, this phenomenon was not observed for TSG-3'-induced structures despite their ability to sequester ubiquitinated cargo. The observations that BFA did not disrupt the GGA-induced compartments and that these compartments were still induced by the Arf-binding-deficient GGA mutants (Figure 2) argue that formation of the GGA-induced structures is independent of Arf binding and GGA membrane association. These GGA-induced structures possess some properties of aggresomes [for example, they stain with the aggresome marker GFP-250 (Figure 3)] but are clearly distinct in terms of composition and inhibitory activity from the aggresome-like structures induced by TSG-3'. Most notably, TSG-3'-induced structures primarily block virus budding and do not interfere with EIAV release [42, 43], whereas GGA-induced structures impair Gag association with membrane and display inhibitory activity against HIV-1 and EIAV (; this study). The distinct phenotypes of these compartments is likely due to differences in the proteins that they sequester. Purification of these aberrant structures and proteomic characterization of their composition would potentially provide additional insights into host cell machinery required for Gag trafficking to the plasma membrane and virus budding.
Although the three mammalian GGA proteins are closely related, in our studies they have shown some notable differences: 1) depletion of GGA3, and to a lesser extent GGA2, stimulated HIV-1 release whereas GGA1 depletion did not . 2) Whereas the compartments induced by overexpression of each of the three GGA proteins stained positively for the TGN marker TGN46, only those induced by GGA3 were positive for the detergent-insoluble cis-Golgi marker GM130 (Figure 1). 3) Compartments induced by GGA1 trapped HIV-1 Gag, whereas those induced by GGA2 or GGA3 did not . It remains to be determined whether the GGA proteins interact directly with Gag or whether differential association with other host cell factors could explain the ability of GGA1 but not GGA2 or GGA3 overexpression to trap Gag. In this study, we observed that Fyn10deltaMA and the non-myristylated HIV-1 Gag mutant were not sequestered in GGA1-induced compartments, indicating a requirement for the MA domain and Gag membrane binding in Gag trapping.
In contrast to the Gag sequestration data, in virus release assays both Fyn10FullMA (Joshi and Freed, unpublished) and Fyn10deltaMA  were resistant to inhibition mediated by GGA overexpression. This observation suggests that Gag sequestration is not the sole cause of inhibition mediated by GGA1 overexpression. It is also possible that the Fyn targeting signal, which confers strong membrane binding ability, is able to rescue the inhibition. Interestingly, EIAV particle production is also inhibited by GGA overexpression  and EIAV Gag is sequestered in GGA1-induced structures (Figure 7B). Moreover, while overexpression of GGA2 and GGA3 mutants defective in Arf binding (the GGA-NA mutants) was no longer inhibitory to HIV-1 and EIAV release, the GGA1-NA mutant was still capable of inhibition and Gag sequestration . Overall, these findings indicate intriguing differences in the mechanisms by which overexpression of the three GGA proteins inhibit retroviral particle production.
In the current study, we extend our previous findings and characterize in detail the structures induced by GGA overexpression with respect to their impact on cellular sorting pathways and their ability to inhibit retroviral particle production. We observe that the compartments induced by GGA overexpression are distinct from those formed by dominant-negative components of the ESCRT pathway, both in terms of their effect on cellular functions and with respect to the step in the retrovirus assembly and release pathway that they disrupt. A better understanding of these host components will not only further our knowledge of cellular sorting processes but also help define the requirements for host factors in retrovirus assembly and release.
HeLa cells were cultured in DMEM supplemented with 5% fetal bovine serum (FBS) and 2 mM glutamine. All transfections were performed using Lipofactamine2000™ reagent (Invitrogen) as per the manufacturer's instructions.
Plasmids expressing full-length Myc-tagged GGA proteins were kindly provided by J. Bonifacino (NIH, Bethesda MD) [39, 48, 49]. The GGA3 mutant defective in ubiquitin binding (L276A)  was constructed by site-directed mutagenesis. The full-length HIV-1 proviral clone pNL4-3  and the myristylation-deficient mutant pNL4-3/1GA  have been previously reported. The ubiquitin expression vector pCW7  was provided by R. Kopito (Stanford University, CA). An HIV-1 derivative bearing at its N terminus ten amino acids from the membrane targeting sequence of Fyn (Fyn10fullMA)  and its derivative lacking the MA domain (Fyn10deltaMA)  have been described. HA-tagged Tsg101 expression constructs TSG-F and TSG-3' have been reported [42, 77]. The Hrs expression vector was constructed by amplifying the Hrs coding region from a cDNA library by PCR and inserting the amplified fragment between the BamHI and XbaI sites of pcDNA3.1. The EIAV Gag expression vector pPRE/Gag has been reported previously [43, 78]. The aggresome marker GFP-250  was kindly provided by E. S. Sztul (University of Alabama at Birmingham, Birmingham AL). The pEGFPhVPS4A (E228Q) expressing an ATPase-deficient mutant of VPS4A fused to GFP was a gift from P. Woodman (University of Manchester, UK) . The anti-GGA1 antibody was a kind gift from R. Kahn (Emory University, Atlanta, GA) ; anti-GGA2 and GGA3 antibodies were purchased from BD Biosciences. The mouse anti-GM130 antibody was from Transduction Laboratories (Lexington, KY) and sheep anti-TGN46 was from Serotec (Oxford, United Kingdom). EIAV anti-p26 antibody was kindly provided by R. Montelaro (University of Pittsburg). EGF-Texas Red and Transferrin Alexa Fluor 594 (Tf-594) conjugates were from Molecular Probes (Invitrogen). The anti-polyubiquitin antibody was from Biomol (clone Fk1). The anti-CD-MPR antibody (clone 22d4) was obtained from University of Iowa Developmental Studies Hybridoma bank and anti-TfR antibody was from Zymed. BFA was purchased from Calbiochem. Protein A beads were from Invitrogen and ubiquitin-agarose beads were obtained from Sigma.
Immunostaining of cells was performed as described , with minor modifications. Cells seeded onto Nunc Lab-Tek II chamber slides were rinsed with phosphate-buffered saline (PBS) and fixed with 3.7% formaldehyde in 100 mM sodium phosphate buffer (pH 7.2) for 20 min. Following fixation, cells were thoroughly rinsed with PBS, permeabilized using 0.1% Triton X-100/PBS for 2 min, and incubated for 10 min with 0.1 M glycine/PBS to quench the remaining aldehyde residues. Cells were then blocked with 3% BSA/PBS for 30 min, followed by incubation for 1 h with primary antibody appropriately diluted in 3% BSA/PBS. After 3 washes in PBS, cells were incubated for 30 min with secondary antibody diluted in 3% BSA/PBS. Cells were then washed and mounted using Aqua Poly/Mount (Polysciences Inc). Images were acquired with a DeltaVision RT microscope. To quantify colocalization, we calculated the Pearson correlation coefficient (R) values, which are standard measures of colocalization . The R values were calculated using the softWoRx colocalization module which generates a "colocalized" image from two channels. A scatter plot of the two intensities on a pixel-by-pixel basis was plotted and the R value calculated by dividing the covariances of each channel by the product of their standard deviations. For EM analysis, transfected cells were fixed using buffer containing 2% glutaraldehyde and 100 mM sodium cacodylate and stored at 4°C. Samples were then sectioned and analyzed by transmission EM .
24 h posttransfection, cells were washed three times with serum free medium (SFM) containing 20 mM HEPES and 1% BSA and incubated for 1 h at 37°C in the same medium. Cells were then placed on ice for 5 min followed by incubation for 30 min at 4°C in SFM containing 5 μg/ml EGF-Texas Red or 10 μg/ml Tf-594 (Molecular Probes). Cells were washed three times with PBS and incubated in cell culture medium at 37°C for 15 min for TfR and 30 min for EGF receptor. Finally, cells were washed with PBS, fixed and mounted using DAPI-containing medium for EGF receptor assays.
The protocol for radiolabeling and immunoprecipitation of cell and virus lysates has been described in detail previously . Briefly, transfected cells were starved for 30 min in labeling media lacking Met and Cys. Thereafter, cells were incubated for 2-3 h in labeling medium supplemented with FBS and [35S]Met/Cys. Culture supernatants were ultracentrifuged at 100,000 × g for 45 min, cell and virus lysates were immunoprecipated with HIV-Ig, resolved by SDS-PAGE followed by PhosphorImager analysis. Virus release was calculated as the percentage of virion-associated p24 (CA) relative to total (virion + cell-associated) Gag. Virus release efficiency = virion p24/(cell associated Pr55Gag + cell-associated p24+virion-associated p24) × 100.
ALG-2 interacting protein X
ADP ribosylation factors
bovine serum albumin
cation-dependent mannose 6-phosphate receptor
Dulbecco-modified Eagle's medium
early endosomal antigen
epidermal growth factor
equine infectious anemia virus
endosomal sorting complex required for transport
fetal bovine serum
feline immunodeficiency virus
γ-adaptin ear homology domain
GGA and Tom domain
Golgi-localized, γ-ear containing, Arf-binding
Golgi matrix marker 130
hepatocyte growth factor-regulated tyrosine kinase
murine leukemia virus
plenty of SH-3s
Rous sarcoma virus
sodium dodecyl sulfate polyacrylamide gel electrophoresis
signal transducing adaptor molecule
tumor susceptibility gene 101
suppressor of cytokine signaling 1
Vps27, Hrs and STAM homology
We thank F. Soheilian and S. Ablan for expert technical assistance and members of the Freed lab for helpful discussion and critical review of the manuscript. We thank J. Bonifacino for kindly providing GGA expression vectors, E. Sztul for the GFP-250 construct, A. Ono for the Fyn10deltaMA and Fyn10fullMA Gag chimeras, R. Kopito for the ubiquitin expression vector pCW7, R. Kahn for the GGA1 antibody, and P. Woodman for the GFP-tagged Vps4EQ expression vector. The HIV-Ig and TZM-bl cells were obtained from the NIH AIDS Research and Reference Reagent Program, EIAV anti-p26 antibody was kindly provided by R. Montelaro and MLV anti-p30 was a gift from A. Rein. This research was supported by the Intramural Research Program of the Center for Cancer Research, National Cancer Institute, NIH, and by the Intramural AIDS Targeted Antiviral Program. This project was funded in part with federal funds from the National Cancer Institute, NIH, under contract N01-CO-12400. The content of this publication does not necessarily reflect the views or policies of the Department of Health and Human Services, nor does mention of trade names, commercial products, or organizations imply endorsement by the U.S. Government.
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.