ABSTRACT Although HIV-2 does not encode a vpu gene, the ability to antagonize bone marrow stromal antigen 2 (BST-2) is conserved in some HIV-2 isolates, where it is controlled by the Env glycoprotein. We previously reported that a single-amino-acid difference between the laboratory-adapted ROD10 and ROD14 Envs controlled the enhancement of virus release (referred to here as Vpu-like) activity. Here, we investigated how conserved the Vpu-like activity is in primary HIV-2 isolates. We found that half of the 34 tested primary HIV-2 Env isolates obtained from 7 different patients enhanced virus release. Interestingly, most HIV-2 patients harbored a mixed population of viruses containing or lacking Vpu-like activity. Vpu-like activity and Envelope functionality varied significantly among Env isolates; however, there was no direct correlation between these two functions, suggesting they evolved independently. In comparing the Env sequences from one HIV-2 patient, we found that similar to the ROD10/ROD14 Envs, a single-amino-acid change (T568I) in the ectodomain of the TM subunit was sufficient to confer Vpu-like activity to an inactive Env variant. Surprisingly, however, absence of Vpu-like activity was not correlated with absence of BST-2 interaction. Taken together, our data suggest that maintaining the ability to antagonize BST-2 is of functional relevance not only to HIV-1 but also to HIV-2 as well. Our data show that as with Vpu, binding of HIV-2 Env to BST-2 is important but not sufficient for antagonism. Finally, as observed previously, the Vpu-like activity in HIV-2 Env can be controlled by single-residue changes in the TM subunit. IMPORTANCE Lentiviruses such as HIV-1 and HIV-2 encode accessory proteins whose function is to overcome host restriction mechanisms. Vpu is a well-studied HIV-1 accessory protein that enhances virus release by antagonizing the host restriction factor BST-2. HIV-2 does not encode a vpu gene. Instead, the HIV-2 Env glycoprotein was found to antagonize BST-2 in some isolates. Here, we cloned multiple Env sequences from 7 HIV-2-infected patients and found that about half were able to antagonize BST-2. Importantly, most HIV-2 patients harbored a mixed population of viruses containing or lacking the ability to antagonize BST-2. In fact, in comparing Env sequences from one patient combined with site-directed mutagenesis, we were able to restore BST-2 antagonism to an inactive Env protein by a single-amino-acid change. Our data suggest that targeting BST-2 by HIV-2 Env is a dynamic process that can be regulated by simple changes in the Env sequence.