No abstract

BackgroundSynonymous or silent mutations are usually thought to evolve neutrally. However, accumulating recent evidence has demonstrated that silent mutations may destabilize RNA structures or disrupt cis regulatory motifs superimposed on coding sequences. Such observations suggest the existence of stretches of codon sites that are evolutionary conserved at both DNA-RNA and protein levels. Such stretches may point to functionally important regions within protein coding sequences not necessarily reflecting functional constraints on the amino-acid sequence. The HIV-1 genome is highly compact, and often harbors overlapping functional elements at the protein, RNA, and DNA levels. This superimposition of functions leads to complex selective forces acting on all levels of the genome and proteome. Considering the constraints on HIV-1 to maintain such a highly compact genome, we hypothesized that stretches of synonymous conservation would be common within its genome.ResultsWe used a combined computational-experimental approach to detect and characterize regions exhibiting strong purifying selection against synonymous substitutions along the HIV-1 genome. Our methodology is based on advanced probabilistic evolutionary models that explicitly account for synonymous rate variation among sites and rate dependencies among adjacent sites. These models are combined with a randomization procedure to automatically identify the most statistically significant regions of conserved synonymous sites along the genome. Using this procedure we identified 21 conserved regions. Twelve of these are mapped to regions within overlapping genes, seven correlate with known functional elements, while the functions of the remaining four are yet unknown. Among these four regions, we chose the one that deviates most from synonymous rate homogeneity for in-depth computational and experimental characterization. In our assays aiming to quantify viral fitness in both early and late stages of the replication cycle, no differences were observed between the mutated and the wild type virus following the introduction of synonymous mutations.ConclusionsThe contradiction between the inferred purifying selective forces and the lack of effect of these mutations on viral replication may be explained by the fact that the phenotype was measured in single-cycle infection assays in cell culture. Such a system does not account for the complexity of HIV-1 infections in vivo, which involves multiple infection cycles and interaction with the host immune system.

Patients with chronic hepatitis B virus (HBV) infection are at an increased risk for a severe and a potentially fatal viral reactivation following anti-cancer therapy. The molecular mechanism for this induction of HBV expression is still unclear. Here, we show that treating hepatoma cell line expressing HBV with various anti-cancer cytotoxic agents results in a significant up-regulation of HBV expression. This HBV induction is at the transcriptional level and is time dependent. Interestingly, treating hepatoma cells with anti-cancer cytotoxic agents results in a robust induction of peroxisome proliferator-activated receptor-gamma coactivator-1α (PGC-1α), a metabolic and energy regulator that is normally induced in the liver under starvation conditions and that has been previously shown to strongly coactivate HBV transcription. Most importantly, HBV up-regulation following anti-cancer therapy depends on PGC-1α induction, because PGC-1α knock-down abolishes HBV induction. Finally, pretreatment of HBV-expressing cells with the antioxidant agent N-acetylcysteine attenuates the induction of both PGC-1α and HBV in response to anti-cancer treatment, suggesting that chemotherapy-associated PGC-1α induction is mediated by cellular oxidative stress that ultimately leads to HBV up-regulation. We conclude that cytotoxic anti-cancer chemotherapy has a direct and an immune system-independent effect on HBV gene expression, which is mediated by PGC-1α. Our results attribute to this metabolic regulator an unexpected role in linking anti-cancer treatment to HBV reactivation and make PGC-1α a potential target for future anti-HBV therapy, especially under conditions in which it is robustly induced, such as following anti-cancer treatment.