s also been implicated in HIV-1 cell-to-cell transmission, but this notion is not supported by other studies. We have previously investigated the sites of HIV-1 entry into cells using two complementary strategies. First, kinetic measurements of virus-cell fusion revealed that HIV-1 acquires resistance to a membrane-impermeant peptide fusion inhibitor much earlier than to low temperature, which blocked all fusion events. This result suggests that HIV-1 escapes from inhibitory peptides, such as enfuvirtide, by entering an endocytic pathway and fusing with endosomes at a 
later time. Second, single virus fusion has been imaged using particles co-labeled with a GFP-based viral content marker and a lipophilic dye incorporated into the viral membrane. The fact that the viral content release was not associated with loss of a lipid marker demonstrated that full fusion occurred in small intracellular compartments that were not connected to the plasma membrane. In stark contrast, particles that exchanged lipids with the plasma membrane, as evidenced by quick disappearance of a viral EMA401 membrane marker, rarely released their content. We found that HIV-1 exhibited a strong preference for endocytic entry, irrespective of the coreceptor tropism and of the choice of retroviral core used to produce pseudoviruses. Our data also support endocytic HIV-1 entry into permissive adherent cells, T cell lines and primary CD4+ T cells. In addition, we have Single HIV Fusion Imaged by Viral Content Release found that, upon blocking endocytosis, HIV-1 fusion and infection were inhibited, while lipid mixing at the cell surface was exaggerated. It thus appears that HIV-1 fusion with the plasma membrane is arrested at a hemifusion stage upstream of fusion pore formation. Our single virus imaging experiments have primarily employed the Murine Leukemia Virus 7949100 Gag-GFP-labeled particles pseudotyped with the HIV-1 Env. The small nucleocapsid-GFP fragment produced upon Gag-GFP cleavage by the viral protease served as the viral content marker and was released through a fusion pore. In addition, we and others labeled particles with the HIV-1 Gag-iGFP construct originally made by Benjamin Chen’s lab. Here, the “internal”GFP flanked by the protease cleavage sites is inserted between the MA and CA domains of Gag. The release of free GFP generated upon cleavage of Gag-iGFP enables the detection of 18000030 viral fusion. A recent study reported that iGFP released upon HIV-1 maturation could be trapped not only within the viral membrane, but also within the mature capsids. Two lines of evidence supported this notion: iGFP co-migrated with the purified capsids, and the majority of intact HIV-1 cores detected in the cytoplasm was GFP-positive. Intact post-fusion cores were identified based upon colocalization with the cytoplasmic bodies formed by TRIM5a, a restriction factor that recognizes the intact HIV-1 capsid. These results suggest that HIV-1 capsids can carry iGFP into the cytoplasm and that perhaps the loss of iGFP in live cell imaging experiments may report capsid uncoating, as opposed to the formation of a small fusion pore. In the extreme case that all iGFP molecules are trapped by the capsid, full fusion at the cell surface could be misinterpreted as hemifusion defined as lipid mixing without the viral content release. Although a fraction of iGFP within a mature virus should reside outside the capsid and thus be lost through a fusion pore, we nonetheless sought to test whether the full