Our results suggest a novel model for the regulation of HIV transcription and the role of the SWI/SNF chromatin-remodeling complex (). We find that, on the HIV LTR, active chromatin remodeling is required for the generation of a chromatin conformation that is repressive to transcription; the BAF complex strictly positions nuc-1. Thus, in the absence of Tat, BAF is bound to the HIV promoter where it represses transcription by counteracting intrinsic nucleosome-DNA sequence preferences and positioning nuc-1 in a less energetically favorable position immediately downstream of the TSS. Upon activation, BAF is removed from the LTR, allowing the formation of nucleosomes according to their intrinsic histone-DNA sequence preferences. Upon expression, Tat first recruits acetyltransferases resulting in the acetylation of promoter histones as well as Tat itself 
. Intriguingly, another SWI/SNF complex, PBAF, specifically interacts with acetylated Tat and is recruited to the HIV promoter by Tat in vivo. Thus, the biochemically distinct chromatin-remodeling complexes BAF and PBAF display functional specificity on the HIV promoter, one repressive and the other participating in the transcriptional activation of the HIV promoter by Tat. This model presents a considerably more complex picture of the role of SWI/SNF proteins in the transcriptional regulation of HIV expression than had previously been anticipated.
Model for SWI/SNF regulation of HIV LTR transcription.
The involvement of the BAF complex in repressing basal HIV transcription raises a number of novel questions. Intriguingly, HIV-1 integrase interacts with the SWI/SNF protein INI-1 in vitro and in vivo 
. During HIV infection, incoming retroviral pre-integration complexes trigger the cytoplasmic export of the SWI/SNF component INI1 and of the nuclear body constituent PML 
. The HIV genome associates with these proteins before nuclear migration. In the presence of arsenic, PML is sequestered in the nucleus and the INI-1/preintegration complex interaction is disrupted 
. Under these conditions, the efficiency of HIV-mediated transduction is markedly increased. This observation could be explained in part by our observations that the BAF complex represses basal HIV transcription by positioning the repressive nuc-1.
Why would the HIV virus position a repressive nucleosome immediately downstream of its transcriptional start site? An attractive possibility is that BAF may remove or pull the thermodynamically favored nucleosome away from the promoter (DHS1) to allow for binding of host cell transcription factors whose binding sites are present within the DHS1. In particular, the NFKB and Sp1 consensus sites within the DHS1 have been shown to be critical for basal HIV LTR promoter activity 
. In this context, positioning of nuc-1 by BAF downstream of the TSS would make consensus sites within the DHS1 accessible for binding by sequence-specific transcription factors, and allow for the assembly of the initiation complex. Surprisingly, our high-resolution MNase nucleosomal mapping of the LTR demonstrated that DHS1 is not devoid of nucleosomes as previously implied. The DNA within this region displayed partial protection against MNase digestion, indicating the presence of loosely positioned nucleosomes, which disappeared upon PMA stimulation. Interestingly, a previous study by Workman and colleagues argued for the presence of a nucleosome over the HIV promoter (DHS1) and the formation of a ternary complex consisting of the transcription factors, histones, and DNA 
. It is important to note that in this study the HIV promoter/DHS1 sequence was reconstituted into mono-nucleosomes or placed within an array of positioned nucleosomes in vitro and outside the context of the adjacent nuc-0 and nuc-1 sequences 
. Our modeling data, depicting the high predicted affinity (nucelosome score) of the DHS1 DNA sequence (), are in agreement with these observations implying the presence of a nucleosome over the DHS1 region in vitro.
Our high-resolution MNase nucleosomal mapping data provide a detailed picture of the dynamic nucleosomal landscape of the HIV LTR, comparing the latent, PMA activated, and BAF-depleted de-repressed LTR states. Loss of BAF250a caused a dramatic re-positioning of the nucleosomes according to their intrinsic DNA-histone sequence preference; DNA encompassing nuc-1 became hypersensitive and susceptible to digestion by MNase while the DHS sites were rendered less accessible and protected from MNase, more closely resembling the predicted LTR nucleosomal structure (). These results support a nucleosome repositioning model upon BAF250a depletion, which contrasts with the observed eviction of nucleosomes downstream of the positioned Nuc-0, which occurs after PMA stimulation.
An important question remaining to be resolved is how BAF is recruited to the LTR to position nuc-1. SWI/SNF complexes have been shown to function gene-specifically, recruited by sequence-specific transcription factors to regulatory regions of target genes. The HIV LTR contains many binding sites for multiple sequence-specific host transcription factors, including SP1, NFkb, YY-1, NFAT, LBP1, etc. One possible mechanism is the recruitment of BAF by a repressive sequence-specific transcription factor bound to nuc-1. A number of transcriptional repressors contain binding sites within the region occupied by nuc-1, including LSF-1 and YY-1 
. Indeed, the transcriptional repressor YY-1 was bound to the HIV promoter under basal conditions and was displaced in response to Tat expression 
. Thus, YY-1 is a candidate transcription factor, which may recruit BAF to the HIV LTR to position the repressive nuc-1. In support of this possibility, pleiohomeotic (PHO), the Drosophila homologue of YY-1, has been shown to directly recruit the SWI/SNF complex to target genes 
. Conversely, the BAF complex, which itself may be recruited by another sequence-specific transcription factor, may allow for YY-1 binding, leading to de-acetylation of histones at the LTR, and the repression of HIV transcription under basal conditions.
Our results indicate that BAF-remodeling activity is necessary to position nuc-1 downstream of the transcriptional start site. The fact that nuc-1 becomes remodeled upon ATP depletion is consistent with this model and, within the limitations of this experiment, suggests that nuc-1 positioning downstream of the TSS is an active process driven in part by ATP and BAF activity. Complementing the ChIPs, we used FAIRE to assess changes in DNA accessibility within nuc-0, DHS1, and nuc-1 regions of the HIV LTR in response to specific depletion of BAF or PBAF. The mapping of accessible DNA within the LTR by FAIRE negatively correlated with histone H3 and H2B occupancy. Our data indicate that the specific depletion of BAF causes nucleosome repositioning in accordance with intrinsic histone-DNA sequence preferences; in the absence of BAF, the DHS1 region, whose sequence displays a higher propensity for nucleosome formation, displays higher histone density and lower accessibility, while the opposite profile is observed for the DNA sequence encompassing nuc-1. These observations suggest that the energetic cost of positioning the repressive nuc-1 downstream of the TSS is provided and driven by ATP hydrolysis. BAF positions nuc-1 by either pushing the nucleosome from its optimal sequence over DHS1 or by pulling it onto sub-optimal sequences encompassing nuc-1. Thus, the BAF complex counteracts and overrules the DNA sequence effects and intrinsically favored nucleosome position over the DHS1 region of the HIV LTR.
The dramatic increase in DNA accessibility over nuc-1 concomitant with de-repression of LTR activity observed upon depletion of the BAF complex begs the following question: why is recruitment of PBAF by Tat necessary to drive transcription at the LTR? In such a context, the recruitment by Tat of an alternative complex, PBAF, may be necessary for remodeling of the nucleosomes, formed in the absence of BAF according to their preferred DNA sequences, downstream of the TSS, leading to efficient transcription elongation (). Indeed, BRG1 was recently shown to be recruited to and facilitate RNA Pol II to overcome nucleosomal barriers during transcription elongation in vivo 
The balance between activating and repressive cofactors at the LTR is believed to determine the level of basal transcription from the LTR in the immediate early, Tat-independent phase of HIV transcription. The local availability of positive and negatively acting cofactors at the LTR therefore determines the likelihood of transcriptional silencing. Our data demonstrated a critical role for BAF in positioning nuc-1 and maintaining HIV LTR silencing. We found that depletion of BAF led to a de-repression of latent HIV in the J-Lat system reflecting HIV latency. Despite the effectiveness of modern HAART regimens, latent HIV-infected cells persist in patients, providing the main impediment to cure from HIV infection. Recently, the need for the development of new strategies to treat HIV-infected patients has been discussed 
, highlighting the necessity to deplete the latent HIV-infected reservoirs. As the catalytic subunit of the BAF complex, the enzyme BRG1 may present an attractive candidate for drug targeting to purge the latent HIV-infected reservoir in the treatment of HIV.