word count: 168 15 Text word count: 5468 16 17 18 2 Abstract 19Retroviral restriction factors are important regulators of viral infection, targeting vulnerable 20 steps of the virus lifecycle; steps that are also targeted by antiviral drugs. It has become clear 21 that the route of cellular infection can alter the sensitivity of HIV-1 to these agents. Using 22 39 3 evolutionary battle between virus and host and show how the virus has co-opted a host protein 40 to protect it from destruction by an antiviral mechanism. These two key findings suggest 41 potential novel treatment strategies that may reduce the viral reservoir and help our natural 42 defences take back control from the virus. 43 44 58 uncoating, proteasomal degradation of the viral capsid, and ultimately inhibition of reverse 59 transcription; however human TRIM5α is inactive against HIV-1 (4). TRIM5α can be saturated 60 4 by large doses of virus e.g. during viral dissemination via virological synapses, and can also be 61 circumvented by alterations in the viral capsid gene that prevent TRIM5α binding (5, 6). The 62 compatibility between the viral capsid sequence and the host species TRIM5 locus is a major 63 determinant of successful infection. In fact, converting a single amino acid to the equivalent in 64 rhesus macaque TRIM5α is sufficient to enable human TRIM5α to restrict HIV-1 (7). The same 65 region of capsid that determines the sensitivity of HIV-1 to TRIM5α is also the binding site for 66 the host peptidyl prolyl isomerase enzyme cyclophilin A (CypA) (8). Binding of CypA to an 67 incoming HIV-1 capsid prevents detection of reverse transcription products by macrophages, 68 and the subsequent production of IFN, but conversely binding of CypA aids restriction by TRIM5 69 of some monkey species (9). This interaction site has been the focus of an evolutionary arms 70 race between the TRIM5 locus and the simian relatives of HIV-1. In several monkey species 71 (e.g. Rhesus macaques and Owl monkeys) multiple retrotransposition events resulted in CypA 72 being inserted into the TRIM5 locus, producing a novel antiviral protein, TRIMCyp (10, 11). 73 Although TRIMCyp does not contain the usual PRYSPRY domain for viral binding, it uses the 74 CypA domain to bind the capsid of sensitive viruses. The species-specific activity of TRIM5 and 75TRIMCyp suggest that they act as cross-species barriers to retrovirus infection (12, 13), and that 76 in consequence, the TRIM5 locus has great potential to act as a target for gene therapy. To that 77 end, it has been shown that overexpression of macaque TRIM5 variants or TRIMCyp in human 78 cells results in protection from HIV-1 infection (14, 15). However, these approaches fail to take 79 account of the role of endogenous TRIM5α as both a target and signaller of the IFN system (16), 80 nor any potentially adverse consequences of overexpression of this potent antiviral element. 81 respective isotype controls (light grey). Stem cells were fixed in 2% paraformaldehyde, followed