Varthakavi, Smith, Heimann-Nichols and Rose reply:

Our previous manuscript described CAML as a host restriction factor that blocks HIV-1 release (Nat. Med. 14, 641–647, 2008). We did not assess the function of CAML as an inhibitor of Vpu− HIV-1 release in 293T cells. The 293T cells used in our laboratory did not fit the Vpu− virus 'permissive' phenotype that has been ascribed to them by Kühl et al. (Nat. Med. 16, 155–156, 2010) and others. We have since ascertained that the permissive 293T cells (a gift from W. Sundquist's laboratory (University of Utah)) show lower amounts of CAML compared to restrictive HeLa and lymphocytic cells. We also did not test CAML against HIV-2 or macaque simian immunodeficiency virus, as did Kühl et al. (Nat. Med. 16, 155–156, 2010). Rather, we compared the ability of HIV-2 Env protein and Vpu to rescue Vpu− HIV release from simian cells. In summary, we believe that the source of cells, variation across individual cell clones and other differences in the experiments in the two reports may account for the reported discrepancies.

The CAML functional data were validated in independent experiments performed in our laboratory and that of P. Spearman. Both groups continue to see an inhibitory effect of human CAML on Vpu− mutant virus release from COS-7 cells. The precise mechanism by which CAML interferes with HIV-1 release needs to be further investigated. In this regard, we concur with the findings from P. Spearman and his colleagues that CAML does not enhance tetherin function and does not contribute to tethering mutant virus particles to the cell surface (Nat. Med. 16, 238, 2010). However, our ongoing studies in HeLa cells suggest a notable effect of CAML on HIV-1 Gag trafficking and release (V.V., E.H.-N. and K.A. Brewer, unpublished data). Our new data, which involve live imaging of cells, shows that overexpression of CAML causes a major redistribution of Gag protein to internal cellular compartments. In contrast, Gag localizes predominantly to the cell surface in cells that are depleted of CAML. Consistent with this, we have recently shown by electron microscopy that HeLa cells overexpressing CAML and infected with Vpu− mutant virus show several vesicular structures in the cytoplasm that contain virus particles. (V.V., E.H.-N. and K.A. Brewer, unpublished data). On the basis of these and other emerging data from our laboratory (V.V., E.H.-N. and K.A. Brewer, unpublished data), we propose that CAML inhibits virus release by targeting a different step in the HIV-1 life cycle than tetherin. We, therefore, stand by the original data (Nat. Med. 14, 641–647, 2008) and have repeatedly found a role for CAML in the arrest of Vpu− mutant virus release from human cells.