Reminiscent of a tetherinmediated inhibition of retroviral release. Consistent with the possibility that tetherin causes poor PERV release in NHP cells, the amount of tetherin mRNA in the NHP cell lines was 10–100 times higher than in the 293T cells. However, the relevance of this finding to the in vivo context is unclear, as the control of tetherin expression, which is class type I interferon-inducible, is different in vivo from that in the cell lines used in this study. Though the in vivo tetherin expression in NHP has yet to be investigated, species difference compared to humans may be insignificant. The block at the cell entry step is critical and, in the case of macaques and baboons, we identified a genetic defect in a PERVA receptor gene, PAR-1. Firstly, expression of a functional human PERV-A receptor rendered NHP cell lines more amenable for viral entry and propagation, indicating that entry of the virus is a critical step in the low permissivity of NHP cells. Our study on cloned PAR-1 showed that rhesus macaque, cynomolgus monkey, and baboon PAR-1 genes encode serine at a.a. 109, and are unable to support PERV-A14/220 infection. PERV poorly infects rhesus macaque cells probably via rhPAR-2, whose expression is lower than that of PAR-1. This PERV-A14/220 infection route is likely to be minor in humans, as huPAR-1 expression is more robust than that of huPAR-2 in most tissues in vivo. These NHP species therefore lack the major infection route potentially used by PERV in humans. This fact must be taken into account in pig-to-NHP transplantation, where cynomolgus macaque and baboon are currently the most often used NHP species, and have been preferentially employed in PERV transmission studies in vivo. The negative PERV transmission results in such experiments should be interpreted with caution. Intriguingly, both AGMPAR-1 and AGMPAR-2 are able to support PERV-A infection. When the same amount of huPAR-1 and AGMPAR-1 was expressed on the cell surface of PERV-A-resistant QT6 cells, the efficiency of PERV-A14/220 mediated-EGFP transduction was similar, suggesting a comparable affinity of the receptors for the virus. However, AGM cells are poorly infected by PERV-A14/220. No evidence of an Fv1/TRIM5a-like restriction activity was found, consistently with previous data in the literature. No inhibitors secreted from AGM cells weredetected in the supernatant of these cells. Tunicamycin treatment to inhibit NBKM120 glycosylation in the target cells could, however, rescue PERV-A infectivity. This effect was cell-specific rather than receptor-specific, and it correlated with cell differences in PAR glycosylation, i.e. a heavier glycosylation of the receptors in NHP cells than in 293T and QT6 cells. A possible explanation for the tunicamycin-mediated increase in PERV-A infectivity is that heavy N-glycosylation of the receptor could prevent PERV binding, and that tunicamycin treatment could relieve this block. This was, however, not supported by our Env binding assay, since tunicamycin treatment did not increase Env binding to AGM cells.