In this work we have made use of yeast genetic analysis to investigate the influence of the 5′HIV-TR on basal transcription. Several attempts to study HIV-1 transcription in yeast have been described, all of them focused in the transactivation capacity of Tat. Although a fusion of Tat with the DNA binding domain of Gal4 can activate the GAL1
, no Tat-dependent transactivation of the HIV-1 LTR had yet been achieved 
. Therefore, the present work is the first successful reconstruction in yeast of a transcriptional system based on HIV-1 elements.
The artificial character of the Ty1-HIV transcription unit raises the possibility of the conclusions extracted from this work being of no relevance for HIV-1 biology. Several results presented in this piece of research argue against that point of view. We show that the 5′HIV-TR not only represses transcription driven by the Ty1 promoter but can also act on a completely different promoter (GAL1
) when this is weakly active. We also show that the 5′HIV-TR induces an accumulation of RNApol II immediately downstream of the promoter, a very common situation throughout the human genome but extremely infrequent in yeast 
. It has been recently proposed that the reason for this difference is the chromatin organization of the transcription start site, which is usually covered by a positioned nucleosome in yeast and immediately upstream of a positioned nucleosome in most metazoan genes 
. We show in this work that the chromatin organization of the HIV fragment present in Ty1-HIV closely resembles the distribution of positioned nucleosomes of HIV-1 proviruses in the human genome. We also present evidence that transcription through the 5′HIV-TR in yeast is influenced by factors that have been previously shown to govern HIV-1 transcription elongation: yDSIF contributes to the repressive role of the 5′HIV-TR in basal transcription, whereas Tat and P-TEFb enhance active transcription in a 5′HIV-TR-dependent manner. Finally, the main conclusion of the genetic analysis, which is the involvement of chromatin reassembly factors in repressing HIV-1 basal transcription, has been confirmed in a human model of HIV-1 latency. Based on these considerations, we believe that the chimeric yeast-HIV system is a valid complementary tool for HIV research.
The role of the 5′HIV-TR in basal transcription has scarcely been studied, and the data available is sometimes contradictory 
, likely due to the use of transiently transfected DNA, which does not ensure a proper organization of DNA in chromatin 
. In fact, mutation in this region produced different effects on HIV-1 transcription when an integrated version was compared to a transiently transfected one 
The data shown in the present work indicates that the 5′HIV-TR contributes to maintaining low levels of basal transcription without interfering with promoter activation. The genetic analysis that we have performed shows a contribution of chromatin reassembly factors to this repressive function of the 5′HIV-TR. We have also confirmed that Spt6 and Chd1 favor a close chromatin configuration on the transcribed region of Ty1-HIV. The role of chromatin reassembly factors at this level seems to be as significant as the one played by DSIF, since the combination of the chromatin alterations produced by spt6-140 and chd1Δ fully matches those caused by spt4Δ. Deletion of the 5′HIV-TR causes a disruption of nucleosome positioning on the rest of the HIV fragment present in Ty1-HIVTARless, mimicking the patterns of spt4Δ and, to a lesser extent, of spt6-140. It is possible that this chromatin difference between the two transcription units is a consequence of the higher transcription level of Ty1-HIVTARless. Alternatively, the absence of the +1 nucleosome may destabilize the overall chromatin configuration and this would, in turn, give rise to increased transcriptional activity. The differences in the chromatin patterns of spt4Δ, spt6-140 and chd1Δ, all showing similar levels of expression, indicate that the chromatin differences are likely to be responsible for the transcription increase and not vice versa. This explanation also fits better with the accumulation of RNApolII on the 5′HIV-TR and with the results of our genetic analysis. In this scenario, productive elongation would be infrequent under non-activating conditions due to both the low number of initiation events and the positioned nucleosomes sitting on the 5′HIV-TR but, in those rare occasions when RNApol II gets through the chromatin boundary, the immediate action of chromatin reassembly factors (recruited by elongating RNApolII itself) would contribute to rebuilding the repressive chromatin configuration, avoiding a transition into an activation-prone chromatin environment ().
Several chromatin-mediated mechanisms contribute to regulating HIV-1 transcription 
: the activation of the LTR promoter is mediated by the acetylation state of its chromatin, especially by the nucleosome located upstream in the LTR (nucleosome 0) 
. An additional role of chromatin in regulating HIV-1 transcription takes place at the level of early elongation, since the positioned nucleosome covering the 5′HIV-TR (nucleosome 1) becomes remodeled in response to promoter activation, in a transcription-independent manner 
Although transcription of Ty1-HIV is more intensively repressed by chromatin reassembly factors than Ty1-HIVTARless, we do not believe that the 5′HIV-TR is specifically required for their recruitment. Spt6, FACT and Chd1 are general elongation factors, whose association to actively transcribed regions, in an RNA polymerase II-dependent manner, is well documented from yeast to metazoa 
. We favour the idea of the 5′HIV-TR being an optimal DNA sequence for nucleosome positioning. Recent genome-wide studies show the importance of 5′ sequences in specifying the location of +1 nucleosomes. In turn, these act as barriers against which other nucleosomes are packed 
. In such a DNA context the repressive action of chromatin reassembly factors would be maximal.
Basal transcription of HIV-1 is highly dependent on the chromatin environment of integration sites 
. Nevertheless, mutations affecting the sequences located at the 3′ border of nucleosome 1, which increase its stability, produce a general reduction in basal transcription, irrespective of the integration site 
. In contrast, the deletion of 60 nucleotides within the sequence covered by this nucleosome, which presumably destabilize it, makes basal transcription even more dependent on the integration site than the wild type 
. This data is fully compatible with our yeast results and provides support for chromatin configuration playing a role in repressing HIV-1 transcription at the level of early elongation. A similar mechanism has been reported for the human hsp70
, where the repressive chromatin configuration covering its transcribed region becomes remodeled during activation, allowing poised RNApolII to complete elongation 
The most relevant conclusion of our genetic analysis is the involvement of chromatin-reassembly factors in repressing HIV-1 basal transcription. We have confirmed that their function is not restricted to our chimeric yeast-HIV system but they are also playing a role in HIV-1 basal transcription in human cells. It has been reported that integration into regions of compacted chromatin, i.e. centromeric heterochromatin, causes HIV promoter inactivity, probably due to the inaccessibility of the basal transcriptional machinery or the inability of transcription factors to overcome the repressive chromatin state imposed 
. In the latent integrations, this repressive state can be overcome by exogenous stimulation with mitogens or cytokines. We have found that, by knocking-down Spt6 and Chd1, the HIV promoter integrated in the context of latency increases its level of expression.
When we deplete Spt6 or Chd1 from the heterogeneous population of HIV latently infected cells, only a small proportion of integrant promoters are activated. This indicates that only a subset of integrations are either negatively regulated by these factors, or able to be reactivated by solely depleting them. In some chromatin environments, transcription may be completely inhibited, such that chromatin disruption by basal transcription would not be an issue and chromatin reassembly factors would play no role. Similarly, TSA treatment and YY1 depletion are not able to reactivate all latent integrations. We had predicted that individual integrations would have a more consistent response to the depletion of reassembly factors. In fact, we observed that individual clones responded differently, both amongst themselves and to the two distinct shRNAs. Still, not all cells of a clonal population respond equally. This behavior is also observed in response to TSA and resembles position effect variegation 
Depletion of Spt6 and Chd1 may cause a deficit of chromatin reassembly in those rare events, during non-activated conditions, in which RNApolII would be able to overcome the chromatin barrier (). This chromatin change would facilitate additional rounds of transcription under non-activating conditions and, eventually, an increase in transcription factor loading and PIC assembly. If transcription were enough to produce some Tat protein, then promoter activity would be reinforced. Inhibition of HDAC activity with TSA has a similar effect, supporting the idea that discrete modifications of chromatin compactness may be sufficient to switch a repressed provirus to active.
It has been shown absence of chromatin reassembly factors in yeast provokes activation of cryptic promoters located in the body of transcribed genes. In the absence of transcription, no PIC assembles on these cryptic promoters, due to the inhibitory action of chromatin; under active transcription, chromatin reassembly factors ensure rapid nucleosome deposition after RNApol II passage, avoiding the activation of cryptic promoters 
. HIV proviruses integrated in highly transcribed genes are usually latent 
. Several mechanisms of transcriptional interference have been proposed to explain this phenomenon 
. It has been shown that read-through transcription from an upstream promoter can interfere with HIV transcription by disturbing PIC assembly 
. Recent published evidence indicates that transcriptional interference is caused by the elongating form of RNA pol II, transcribing through the latent HIV copy 
. This situation parallels that of yeast cryptic promoter and, accordingly, it is likely that chromatin reassembly factors also play an important role in maintaining HIV latency by transcriptional interference (). The integration site of one of the individual clones tested in our depletion experiments (clon 27) is a highly transcribed gene. We found a clear reactivation of the latent HIV copy of this clone when chromatin reassembly factors were depleted. In fact, when Spt6 was depleted, this clone showed the maximal reactivation level reached in the whole set of experiments.
When taken together, our results indicate that Spt6 and Chd1 participate in the mechanism that controls the equilibrium between activation and repression of HIV-1 expression when the provirus is integrated in the human genome, which depends greatly on the chromatin environment of the integration site. Disturbance of this equilibrium, by depleting chromatin reassembly factors for example, makes some of the latent integrations become activated. These factors may also play a general role in repressing basal transcription throughout the human genome. Therefore, the specificity of chromatin reassembly factors in repressing viral basal transcription should be carefully evaluated before considering them as therapeutic targets against HIV-1 latency.