Mammals have evolved various mechanisms to thwart the spread of intracellular pathogens. Thus, even if the infectious agent breaches the first lines of defence, the infected cells express restriction factors that suppress its further dissemination. One such restriction factor prevents HIV-1 from leaving infected cells and is counteracted by an accessory viral protein called Vpu. On page 425 of this issue, Neil and colleagues1 characterize this factor and provide tantalizing clues about how it functionsFootnote 1.

To escape the infected cell, HIV-1 buds through the cell surface (thus becoming wrapped in an envelope derived from the host cell membrane) (Fig. 1a) and pinches off. But even if the virus reaches this stage of release, Vpu-deficient HIV-1 particles tend to remain trapped at the cell surface. What keeps them there has been a mystery.

Figure 1: Model for tetherin-mediated HIV-1 retention.
figure 1

a, HIV-1 virions assemble at the cell surface and leave infected cells by budding off from the cell membrane. Assembling virions thus become covered by a cell-membrane-derived envelope. b, Tetherin is localized on the outside of the cell surface, but is anchored in the cell membrane at both ends. Identifying tetherin as the cellular factor that prevents HIV-1 release, Neil et al.1 speculate that this protein is also taken up by the budding virion and is firmly anchored in the viral envelope. Virion- and cell-associated tetherin could then interact, preventing the release of the mature virion from the cell surface. c, The authors also propose that, through one of its membrane anchors, tetherin might connect to the cell's endocytotic machinery, which engulfs extracellular material in cell-membrane invaginations and imports it into the cell. This could lead to the reuptake of mature virions by infected cells and their subsequent degradation by the cell's digestive system.

The effect of Vpu on virus release has been perplexing, as it seems to be unrelated to the protein's other function of reducing cellular levels of CD4, the HIV-1 receptor on T cells of the immune system. Also, Vpu's effect on release does not seem to require other HIV-1 components, and its importance for virus release varies widely among cell lines.

It has emerged2,3 that Vpu counteracts a human antiviral factor that suppresses HIV-1 release by tethering mature viral particles to the cell surface after they have completely pinched off. Such retention of HIV-1 particles at the surface of infected cells can also be induced by human interferon-α, a protein that 'jump-starts' a cell's antiviral defences4. Moreover, in an earlier study5, Neil and colleagues found that Vpu counteracts the effect of interferon-α on HIV-1 release.

The authors now identify1 an interferon-α-induced human protein that fulfils all the criteria for the tethering factor antagonized by Vpu. The protein, which they aptly name tetherin, is expressed only by cells that require Vpu for HIV-1 release. Decreasing tetherin levels in cells that normally produce this protein allows the release of Vpu-deficient viruses. Furthermore, tetherin expression in cells that normally lack this protein selectively inhibits the release of Vpu-deficient HIV-1. Neil and colleagues' findings suggest that tetherin is highly potent, with only minute quantities being enough to efficiently inhibit HIV-1 lacking Vpu.

How can tetherin entrap outgoing viral particles with such efficiency? Little is known about this small protein, but one aspect is clear — both ends of tetherin are inserted in the cell membrane through its unusual pair of membrane anchors6. The central portion of the protein faces the outside of the cell6 and seems to interact with the same region of another tetherin molecule7. Thus, assuming that tetherin is incorporated into the membrane enveloping Vpu-deficient HIV-1 particles, Neil et al.1 envisage a situation in which tetherin molecules that end up in the viral envelope hold the virus back by interacting with tetherins that are associated with the cell surface (Fig. 1b). Tetherin also interacts with the cell's endocytotic — or internalization — machinery8, which might play a part in the reuptake of the trapped viruses into the cell and their degradation in intracellular compartments3 (Fig. 1c).

How does Vpu counteract the effects of tetherin? Neil et al.1 did not detect reduced tetherin levels in the presence of Vpu, although this could have been due to experimental overexpression of tetherin. A previous study9, however, found that levels of the protein now identified as tetherin are reduced by an entirely unrelated human virus, Kaposi's sarcoma-associated herpesvirus (a finding that also hints at the broad antiviral activity of this protein.) The viral protein responsible in this case, K5, is a ubiquitin ligase enzyme, which adds the molecular tag ubiquitin to proteins, marking them for degradation. K5 is structurally similar to a family of human ubiquitin ligases, at least one of which can strongly reduce the cellular levels of tetherin.

With what turns out to be remarkable foresight, the authors of this earlier study9 also tested Vpu and found that it, too, decreases the normal cellular levels of tetherin. These observations raise the possibility that Vpu uses a cellular ubiquitin ligase to dispose of tetherin, as it does for CD4. But other possibilities, such as tetherin relocalization by Vpu, rather than its degradation, are also possible.

Only HIV-1 and a handful of its cousin viruses make Vpu, which poses the question of how other related viruses deal with tetherin. For HIV-2 (the less virulent human AIDS virus), a protein that mainly facilitates viral entry substitutes for Vpu, promoting virus release10, and we will probably soon learn whether this protein also antagonizes tetherin.

Because the amino-acid sequence of tetherin differs considerably among mammals, some HIV-1-related animal viruses might find it difficult to overcome human tetherin, preventing them from becoming human viruses. Conversely, it is worth investigating whether tetherin contributes to the inability of HIV-1 to efficiently escape from most rodent cells11, which has hampered efforts to develop small-animal models of HIV-1 infection. Even in human cells, Vpu might not always be able to overcome the powerful effect of tetherin, as the release of infectious Vpu-positive HIV-1 can be inhibited with high doses of interferon-α (ref. 5). Thus, an understanding of how tetherin works, and how Vpu fends it off, could lead to strategies to limit the spread of HIV-1 and other viruses that target humans.