BACKGROUND: Despite intensive investigation the mechanism by which HIV-1 reaches the host cell nucleus is unknown. TNPO3, a karyopherin mediating nuclear entry of SR-proteins, was shown to be required for HIV-1 infectivity. Some investigators have reported that TNPO3 promotes HIV-1 nuclear import, as would be expected for a karyopherin. Yet, an equal number of investigators have failed to obtain evidence that supports this model. Here, a series of experiments were performed to better elucidate the mechanism by which TNPO3 promotes HIV-1 infectivity. RESULTS: To examine the role of TNPO3 in HIV-1 replication, the 2-LTR circles that are commonly used as a marker for HIV-1 nuclear entry were cloned after infection of TNPO3 knockdown cells. Potential explanation for the discrepancy in the literature concerning the effect of TNPO3 was provided by sequencing hundreds of these clones: a significant fraction resulted from autointegration into sites near the LTRs and therefore were not bona fide 2-LTR circles. In response to this finding, new techniques were developed to monitor HIV-1 cDNA, including qPCR reactions that distinguish 2-LTR circles from autointegrants, as well as massive parallel sequencing of HIV-1 cDNA. With these assays, TNPO3 knockdown was found to reduce the levels of 2-LTR circles. This finding was puzzling, though, since previous work has shown that the HIV-1 determinant for TNPO3-dependence is capsid (CA), an HIV-1 protein that forms a mega-dalton protein lattice in the cytoplasm. TNPO3 imports cellular splicing factors via their SR-domain. Attention was therefore directed towards CPSF6, an SR-protein that binds HIV-1 CA and inhibits HIV-1 nuclear import when the C-terminal SR-domain is deleted. The effect of 27 HIV-1 capsid mutants on sensitivity to TNPO3 knockdown was then found to correlate strongly with sensitivity to inhibition by a C-terminal deletion mutant of CPSF6 (R2 = 0.883, p < 0.0001). TNPO3 knockdown was then shown to cause CPSF6 to accumulate in the cytoplasm. Mislocalization of CPSF6 to the cytoplasm, whether by TNPO3 knockdown, deletion of the CPSF6 nuclear localization signal, or by fusion of CPSF6 to a nuclear export signal, resulted in inhibition of HIV-1 replication. Additionally, targeting CPSF6 to the nucleus by fusion to a heterologous nuclear localization signal rescued HIV-1 from the inhibitory effects of TNPO3 knockdown. Finally, mislocalization of CPSF6 to the cytoplasm was associated with abnormal stabilization of the HIV-1 CA core. CONCLUSION: TNPO3 promotes HIV-1 infectivity indirectly, by shifting the CA-binding protein CPSF6 to the nucleus, thus preventing the excessive HIV-1 CA stability that would otherwise result from cytoplasmic accumulation of CPSF6.