We have made insertion mutants in α9 of HXB2 gp41 and observed similar phenotypes like recent JRFL mutants: insertion of alanine (653+A), but not glutamine (653+Q), severely attenuated membrane fusion. To understand the underlying mechanism, we performed the fusion inhibition assay by corresponding mutant C34 peptides. Both mutant C34 peptides added at the beginning of the coculture of the effector and target cells showed less efficient inhibition of membrane fusion, which was similar to wildtype C34 added after 30 min of coculture, indicating slow association of mutant C34 peptides with the N-terminal heptad region of gp41. Due to uninterpretable CD profiles of C34 and N36, we tested the longer peptide pairs (N46 and C42) and observed CD profiles indicative of weak α-helix formation. The melting temperatures for N46-C42 pairs of 653+A, 653+Q, and wild type were 56.8 ℃ , 59.8℃, and 96℃, respectively. Taken together, our data suggested that the phenotypic difference in membrane fusion between 653+A and 653+Q (or wild type) was not based on the stability of the six-helix bundle (6HB), but due to differences in the kinetics of 6HB formation. Further, we examined additional insertions (E, R, I, and L) at position 653, for which only I and L showed fusion recovery similar to Q, suggesting that the polar nature of glutamine was not a phenotypic determinant.