Ubiquitin in Retrovirus Assembly: Actor Or Bystander?

Commentary

Ubiquitin in retrovirus assembly: Actor or bystander?

Volker M. Vogt*

Department of Molecular Biology and Genetics, Cornell University, Ithaca, NY 14853

he assembly of retroviruses is decep- and this signals internalization by endocy- that has a functional relationship with Ttively simple. Only the product of the tosis. The enzymology of ubiquitin conju- ubiquitin. gag gene is required for the formation of gation is complex. The first step is activa- The core element of the late domains of a virus-like particle (1). After its synthesis tion of ubiquitin by an ATP-requiring RSV (16) (Fig. 1), murine leukemia virus in the cytoplasm Gag is targeted to the enzyme called E1, leading to conjugation (17), and Mason Pfizer monkey virus (18) plasma membrane, where Gag-Gag inter- of ubiquitin to E1 via a high-energy thio- is the sequence PPPY, which is found actions, Gag-RNA interactions, Gag- ester bond. The ubiquitin is transferred by about 150–200 amino acid residues from membrane interactions, and perhaps Gag- E1 to a second enzyme, E2. A third en- the Gag N terminus in these and other host protein interactions lead to a bulging zyme, E3, which may form a complex with retroviruses. The core element of the out of the nascent virus particle, first as a E2, then transfers the ubiquitin to amino HIV-1 late domain, PTAP (15), is located horseshoe-shaped and then as a lollipop- groups of the target protein. Although in a different part of the protein, the shaped structure that can be visualized by cells have only one E1, they may have C-terminal portion called p6 (Fig. 1). In a electron microscopy. The process of mem- dozens of different E2 and E3 enzymes, distantly related retrovirus of the same brane envelopment of the virus is called which confer specificity to the process of genus, equine infectious anemia virus budding. The last step in budding is a ubiquitination. Similarly, there are many (EIAV), the late domain also is near the C membrane fusion event that serves to different ubiquitin hydrolases that can terminus of Gag, but has the sequence pinch off the virus, releasing it into the cleave the isopeptide bonds between two YPDL (19). Remarkably, these three core medium. Late in assembly, perhaps con- ubiquitins or between ubiquitin and a sequences together with immediately ad- comitant with pinching off or just preced- target protein (11). joining sequences can supply late domain ing it, a viral protease cleaves Gag at Despite lack of overall sequence simi- function if inserted into an RSV Gag or an several sites, leading to a morphological larity, the Gag proteins of retroviruses in HIV-1 Gag missing its own late domain, change in the virus. This process is called different genera are functionally and and can do so at either the HIV-1 or the maturation and is essential for the virus to RSV location (13). The late domain func-

structurally similar. All contain three COMMENTARY become infectious. common domains that become mature tion appears to be mediated by cellular In this issue of PNAS, three articles by proteins upon proteolytic processing. The proteins. PPPY (consensus PPxY) is a independent groups provide an important MA (matrix) domain contains a plasma sequence recognized by a family of pro- new clue to the mechanism by which ret- membrane targeting sequence, and in the tein interaction modules called WW do- roviruses bud from the plasma membrane mature virus particle MA lines the inner mains (20). YPDL (consensus YxxL) is (2–4). The articles implicate the cellular recognized by the AP-2 complex that is face of the membrane. The CA (capsid) protein ubiquitin in a late step in budding. involved in clathrin-mediated endocytosis domain participates in protein–protein in- Ubiquitin is a small highly conserved (21). Indeed AP-2 colocalizes with EIAV teractions in assembly, and the mature CA 76-aa protein well known for its role in budding structures by immunofluores- forms the core shell in the infectious virus. protein degradation by the proteasome cence techniques (22). The consequence The NC (nucleocapsid) domain recog- (5). In proteins targeted for degradation, of an HIV-1 late domain deletion is that nizes the packaging signal in the genomic ubiquitin becomes coupled via its C- the cells become covered with virus par- RNA, and the mature NC coats and con- terminal glycine residue to the ␧-amino ticles that remain tethered to the mem- group of lysine residues in the protein (or denses the RNA genome. In addition to brane by narrow stalks (14). This dramatic rarely to the N-terminal ␣-amino group), the three common mature proteins, Gag phenotype depends on the presence of an forming an isopeptide bond. Other ubiq- proteins have other polypeptide segments active protease (15), implying some func- uitin proteins are similarly coupled to whose functions have not been fully elu- tional connection between maturation lysine residues in this first ubiquitin, and cidated and which vary among virus gen- and the final steps in budding. thus the target protein ends up carrying a era. Deletion analyses of HIV-1 and the Enter ubiquitin and ubiquitin ligase. ramified polyubiquitin chain that acts as a avian Rous sarcoma virus (RSV) (Fig. 1), The hint that ubiquitin might have some- signal recognized by the proteasome (6). as well as a few other model systems, have thing to do with retrovirus assembly came By mechanisms that are incompletely un- shown that much of Gag is dispensible for over a decade ago from the observation derstood, the targeted protein is degraded budding, when budding is defined mini- that RSV particles contain some 100 mol- while the ubiquitin molecules are recy- mally as the efficient release of Gag in an ecules of this protein in an unconjugated cled. A less well-known role for ubiquitin enveloped and particulate form into the form, which represents an enrichment is in endocytosis of plasma membrane medium. The segments of Gag that are over the concentration of free ubiquitin in proteins (7, 8). For example, this process essential are called assembly domains (not the cytoplasm of the cell (23). Similar has been studied with yeast plasma mem- to be confused with the units that are amounts of free ubiquitin later were re- brane receptors (9) and with human separated by cleavage, like MA or CA). ported in HIV-1 and murine leukemia growth hormone receptor (10). In yeast a These essential segments of Gag include a single ubiquitin moiety becomes coupled sequence called the late domain because it to the cytoplasmic tail of the receptor in is required for efficient pinching off of See companion articles on pages 13057, 13063, and 13069. response to stimulation of the receptor, the virus bud (12–15). It is the late domain *E-mail: [email protected].

PNAS ͉ November 21, 2000 ͉ vol. 97 ͉ no. 24 ͉ 12945–12947 Downloaded by guest on September 24, 2021 conjugation could well be transient, being reversed by one of the many ubiquitin hydrolases (11). Nevertheless, the case for Gag ubiquitination would remain weak were it not for the third paper in this issue (4). One of the smallest assembly-compe- tent HIV-1 Gag constructs yet created is Fig. 1. Structure of HIV-1 and RSV Gag proteins. Vertical lines denote sites of cleavage by the viral protease that liberates the mature proteins indicated. The vertical black bar represents the late domain, only one-third the size of wild-type Gag indicated by the core sequence PTAP or PPPY for these proteins. (29). In studying the budding of this protein, Strack et al. (4) found that up to about one-half of all of the Gag molecules virus (MuLV) (24). In addition, a small not explain the observation for RSV that in the released virus-like particles were percentage of the mature HIV-1 and genetically fusing a copy of ubiquitin onto conjugated to one or more ubiquitins. MuLV Gag proteins that harbor the late the C terminus of Gag partially restored Ubiquitination was decreased by protea- domains was found to be ubiquitinated. budding in the inhibitor-treated cells (2), some inhibitors and absolutely depended The present papers by Patnaik et al. (2) a result analogous to that described pre- on the presence of a late domain, with and Schubert et al. (3) now establish a viously for ubiquitin-dependent endocyto- diverse core late domain sequences being putative causal link between budding and sis of a yeast membrane receptor to which able to supply this function. Among the ubiquitin, by showing that depletion of the ubiquitin had been fused (9). Thus appar- most efficacious sequences was one from intracellular pool of free ubiquitin inhibits ently it is the covalent attachment of ubiq- Ebola virus that contains partially over- budding. Both papers used the same ex- uitin to Gag that is important for budding, lapping PPxY and PTAP sequences. The perimental logic, treatment of cells with and not some other aspect of ubiquitin authors noted from database analysis that inhibitors that block proteasome function. metabolism. However, it is troublesome several other viruses that become mem- This prevents recycling of the ubiquitin that RSV Gag has not been found to be brane enveloped in their life cycle, such as that is covalently coupled to proteins des- ubiquitinated, either previously (23) or in rhabdoviruses, also have closely juxta- tined for degradation, locking up ubiq- the present study (2). Moreover, although posed PPxY and PTAP or PSAP se- uitin in a conjugated form. The rates of the level of ubiquitination of HIV-1 Gag quences. Indeed the PPxY sequence in budding by RSV Gag (2) and HIV-1 (3) decreased in inhibitor-treated cells, mu- rhabdoviruses has been suggested previ- were found to be reduced by about 3-fold, tating the two lysine residues in p6 that are ously to function in budding (30, 31). with the reduction depending on the pres- known to be acceptors of ubiquitin failed What model best accounts for the ob- ence of a late domain. Although this level to compromise budding (28). For two servations in these three papers? The of inhibition is not large, its consistency in reasons these results do not necessarily most obvious is that a ubiquitin ligase is two different retrovirus systems is impres- negate a critical role for ubiquitin-Gag pulled to the site of budding by interaction sive. Some drug-treated RSV-infected conjugation. First, it is known that ubiq- with the late domain, and there it conju- ͞ cells had large aggregates of almost crys- uitin conjugation to proteins is not neces- gates ubiquitin to Gag and or to other talline arrays of virus particles attached to sarily specific to particular lysine residues. cellular proteins (Fig. 2). A relationship the cell surface and attached to each other If an E3 enzyme bound to the late domain between ubiquitination and the late do- by thin stalks. The observation of viruses of Gag, and then it searched for lysine main sequence PPxY has precedents out- tethered to other viruses is reminiscent of acceptors nearby in three-dimensional side of retrovirology. The protein Nedd4 the recently described phenotype of an space, the low level of any particular Gag- is a multidomain polypeptide that includes MuLV late domain defect (25) and sug- ubiquitin conjugate might well have been a calcium-dependent lipid binding do- gests that viruses bud sequentially from missed in the published studies. Second, main, an E3 ubiquitin ligase domain, and exactly the same sites on the plasma membrane. In the HIV-1 experiments the proteasome inhibitors not only reduced budding, but reduced proteolytic maturation of Gag to a similar extent. In contrast to HIV-1, RSV proteolytic maturation was not affected by the inhibitors. These ap- parently contradictory observations may relate to the fact that the kinetics of RSV budding are not affected by the viral protease, whereas there is at least one report that the kinetics of HIV-1 budding are slower in absence of protease function (26). By what mechanism does proteasome inhibition adversely affect retrovirus budding? As discussed by Schubert et al. (27), one possibility derives from the buildup of defective Gag translation products that Fig. 2. Speculative representation of late stage in retroviral budding. The Gag molecules have formed otherwise would be rapidly degraded by a spherical particle ready to pinch off from the plasma membrane. Some ubiquitin molecules (red star), the proteasome. Even a few defective Gag both free and conjugated to Gag, are present in the nascent virus. A ubiquitin ligase (blue rectangle) is proteins could be imagined to interfere recruited to the site of budding by the Gag late domain, and it conjugates ubiquitin to some of the Gag with budding. Although this possibility molecules and perhaps to cellular proteins (green oval). Other cellular proteins (yellow triangle) may bind has not been rigorously excluded, it does to these ubiquitin conjugates, setting up the machinery to pinch off the virus bud. (Adapted from ref. 20.)

12946 ͉ www.pnas.org Vogt Downloaded by guest on September 24, 2021 several WW domains. Nedd4 is implicated of the lysines in the cytoplasmic tail were Two other groups also have implicated in down-regulation of the epithelial so- mutated, the receptor was still internal- ubiquitin in the function of late domains in dium channel, and it acts by binding to ized in a ubiquitin-dependent manner, retrovirus assembly, and their results pro- PPxY motifs in subunits of the channel implying that ubiquitination of the protein vide further clues to its mechanism of complex and then ubiquitinating them, itself cannot be involved (33). (b) What is action. Kikonyogol et al. (J. Leis, personal leading to endocytosis and degradation the origin of the free ubiquitin in virus communication) found the E3 ubiquitin (32). How ubiquitination of Gag or nearby particles? Are these molecules derived ligase Nedd4 to interact with the RSV late cellular proteins might function is un- from hydrolysis of ubiquitin conjugated to domain, and VerPlank et al. (C. Carter, known. Presumably further proteins in Gag or conjugated to a cellular protein in personal communication) found the E2- turn would bind to the conjugated ubiq- the vicinity of the budding site? Alterna- like protein Tsg101 to interact with the uitin, as inferred for the ubiquitin involved tively, are the free ubiquitin molecules in HIV-1 late domain. Both groups also have in endocytosis of yeast receptors, ulti- the virus simply a footprint left close to the evidence that these proteins can bind to mately recruiting cellular membrane re- site of budding by a locally high concen- their respective Gag partners in trans- modeling machinery. This machinery tration of proteins with ubiquitin binding fected cells. Although Tsg101 is clearly a probably shares elements with that used domains? (c) Why are the HIV-1 mini- ubiquitin ligase by overall sequence simi- for endocytosis, as suggested by the ob- Gag proteins far more heavily ubiquiti- larity, it lacks the active site Cys residue servation that the late domain of EIAV nated than wild-type HIV-1 Gag? Is this a needed for ligation. The significance of interacts with AP-2 (22). reflection of the presence of a ubiquitin To establish the function of ubiquitin in hydrolase in normal particles but not mini- the implied absence of E2 enzymatic func- retrovirus budding, several outstanding Gag particles, or is the size of the Gag tion remains to be established, and sug- questions need to be answered. (a) What protein or the structure of the particle the gests that the role of ubiquitination may be is the relevant target of ubiquitination? critical element? (d) What is the relation- complex. Does Gag ubiquitination play a role in ship between late domain function and Because diverse viruses appear to have budding, or is this modification simply proteolytic maturation of Gag, and why late domain sequences, we can anticipate acting as a reporter for the nearby pres- does proteolysis depend on late domain that the principles learned from further ence of a ubiquitin ligase, whose real function in HIV-1 but not RSV? (e) Fi- studies of retroviral late domains and function is to ubiquitinate a cellular pro- nally, what is the nature of the membrane ubiquitination will have wide relevance. It tein at the budding site? The data avail- fusion machinery implied by the late bud- may turn out, as has happened frequently able to date do not yet make a persuasive ding defect? Does the cortical cytoskele- in the history of virology, that deciphering case for the importance of modifying Gag ton, which has been implicated indirectly how a virus has coopted cellular machin- by ubiquitin addition, but no other targets in budding of HIV-1 (34), play a role in ery will lead to insights into fundamental have been found or suggested. In the the failure of virus particles to pinch off cellular processes that are otherwise dif- human growth hormone system, when all completely? ficult to approach.

1. Swanstrom, R. & Wills, J. W. (1997) in Retrovi- 13. Parent, L. J., Bennett, R. P., Craven, R. C., Nelle, 24. Ott, D. E., Coren, L. V., Copeland, T. D., Kane, COMMENTARY ruses, eds. Coffin, J. M., Hughes, S. M. & Varmus, T. D., Krishna, N. K., Bowzard, J. B., Wilson, C. B., B. P., Johnson, D. G., Sowder, C., Yoshinkaka, Y., H. E. (Cold Spring Harbor Lab. Press, Plainview, Puffer, B. A., Montelaro, R. C. & Wills, J. W. Oroszlan, S., Arthur, L. O. & Henderson, L. W. NY), pp. 263–334. (1995) J. Virol. 69, 5455–5460. (1998) J. Virol. 72, 2962–2968. 2. Patnaik, A., Chau, V. & Wills, J. W. (2000) Proc. 14. Go¨ttlinger, H. G., Dorfman, T., Sodroski, J. G. & 25. Yuan, B., Campbell, S., Bacharach, E., Rein, A. & Natl. Acad. Sci. USA 97, 13069–13074. Haseltine, W. A. (1991) Proc. Natl. Acad. Sci. USA Goff, S. P. (2000) J. Virol. 74, 7250–7260. 3. Schubert, U., Ott, D. E., Chertova, E. N., Welker, 88, 3195–3199. 26. Kaplan, A. H., Manchester, M. & Swanstrom, R. R., Tessmer, U., Princiotta, M., Bennink, J. R., 15. Huang, M., Orenstein, J. M., Martin, M. A. & (1994) J. Virol. 68, 6782–6786. Kra¨usslich, H.-G. & Yewdell, J. W. (2000) Proc. Freed, E. O. (1995) J. Virol. 69, 6810–6818. 27. Schubert, U., Anton, L. C., Gibbs, J., Norbury, C., Natl. Acad. Sci. USA 97, 13057–13062. 16. Xiang, Y., Cameron, C. E., Wills, J. W. & Leis, J. Yewdell, J. W. & Bennink, J. R. (2000) Nature 4. Strack, B., Calistri, A., Accola, M. A., Palu`, G. & (1996) J. Virol. 70, 5695–5700. (London) 404, 770–774. Go¨ttlinger, H. G. (2000) Proc. Natl. Acad. Sci. 17. Yuan, B., Li, X. & Goff, S. P. (1999) EMBO J. 18, 28. Ott, D. E., Coren, L. V., Chertova, E. N., Gagli- USA 97, 13063–13068. ardi, T. D. & Schubert, U. (2000) Virology, in press. 4700–4710. 5. Hershko, A. & Ciechanover, A. (1998) Annu. Rev. 29. Accola, M. A., Strack, B. & Go¨ttlinger, H. G. 18. Yasuda, J. & Hunter, E. (1998) J. Virol. 72, Biochem. 67, 425–479. (2000) J. Virol. 74, 5395–5402. 4095–4103. 6. Thrower, J. S., Hoffman, L., Rechsteiner, M. & 30. Craven, R. C., Harty, R. N., Paragas, J., Palese, P. 19. Puffer, B., Parent, L. J., Wills, J. W. & Montelaro, Pickart, C. M. (2000) EMBO J. 19, 94–102. & Wills, J. W. (1999) J. Virol. 73, 3359–3365. R. C. (1997) J. Virol. 71, 6541–6546. 7. Rotin, D., Staub, O. & Haguenauer, T. R. (2000) 31. Harty, R. N., Paragas, J., Sudol, M. & Palese, P. J. Membr. Biol. 176, 1–17. 20. Garnier, L., Wills, J. W., Verderame, M. F. & (1999) J. Virol. 73, 2921–2929. 8. Hicke, L. (1999) Trends Cell Biol. 9, 107–112. Sudol, M. (1996) Nature (London) 381, 32. Staub, O., Gautschi, I., Ishikawa, T., Breitschopf, 9. Terrell, J., Shih, S., Dunn, R. & Hicke, L. (1998) 744–745. K., Ciechanover, A., Schild, L. & Rotin, D. (1997) Mol. Cell 1, 193–202. 21. Ohno, H., Stewart, J., Fournier, M.-C., Bosshart, EMBO J. 16, 6325–6336. 10. Strous, G. J., Van Kerkhof, P., Govers, R., Ciec- H., Rhee, I., Miyatake, S., Saito, T., Gallusser, A., 33. Govers, R., Ten Broeke, T., Van Kerkhof, P., Schwartz, hanover, A. & Schwartz, A. L. (1996) EMBO J. 15, Kirchhausen, T. & Bonifacino, J. S. (1995) Science A. L. & Strous, G. J. (1999) EMBO J. 18, 28–36. 3806–3812. 269, 1872–1875. 34. Sasaki, H., Nakamura, M., Ohno, T., Matsuda, Y., 11. Wilkinson, K. D. (1997) FASEB J. 11, 1245–1256. 22. Puffer, B., Watkins, S. C. & Montelaro, R. C. Yuda, Y. & Nonomura, Y. (1995) Proc. Natl. 12. Wills, J. W., Cameron, C. E., Wilson, C. B., Xiang, (1998) J. Virol. 72, 10218–10221. Acad. Sci. USA 92, 2026–2030. Y., Bennett, R. P. & Leis, J. (1994) J. Virol. 68, 23. Putterman, D., Pepinsky, R. B. & Vogt, V. M. 35. Jayakar, H. R., Murti, K. G. & Whitt, M. A. (2000) 6605–6618. (1990) Virology 176, 633–637. J. Virol. 74, 9818–9827.

Vogt PNAS ͉ November 21, 2000 ͉ vol. 97 ͉ no. 24 ͉ 12947 Downloaded by guest on September 24, 2021