variation differentially regulates influencing HIV disease outcome.

Multiple genome-wide studies have identified associations between outcome of HIV infection and polymorphisms in/around the gene encoding the HIV co-receptor CCR5, but the functional basis for the strongest of these associations, rs1015164 A/G, is unknown. We found that rs1015164, located 34KB downstream of CCR5, marks variation in an ATF1 binding site that controls expression of the antisense long non-coding RNA CCR5AS. Knockdown or enhancement of CCR5AS expression resulted in a corresponding change in CCR5 expression on CD4 + T cells. CCR5AS interfered with interactions between the RNA-binding protein Raly and the CCR5 3’ untranslated region, protecting CCR5 mRNA from Raly-mediated degradation. Reduction in CCR5 expression through inhibition of CCR5AS diminished infection of CD4 + T cells with CCR5-tropic HIV in vitro. These data represent a rare determination of the functional importance of a genome-wide disease association where expression of a long non-coding RNA affects level of HIV infection and disease progression. in site; unlikely to be directly responsible for differential However, a variant located in the first intron of CCR5AS, rs2027820, may directly affect expression levels of this gene, as it is located within a binding site for the transcription factor ATF1 and it is in virtually perfect LD with rs1015164. ATF1 belongs to the cAMP-regulated enhancer-binding (CREB) family of transcription factors and is expressed in a wide variety of cell types 41 . We show that rs2027820G, which is in LD with rs1015164A, allows binding of ATF1 and enhances protein expression in a luciferase assay relative to rs2027820A, indicating that the rs2027820G variant enhances CCR5AS expression levels, whereas the A variant confers poor binding of ATF1 and relatively poor expression of CCR5AS. reporter psicheck2 vector. Hut-78 cells were transfected with either the non-targeting siCon or siRaly. Seventy-two hours after siRNA transfection, the CCR5 3’UTR luciferase constructs were transfected into siCon or siRaly transfected Hut-78 cells. Twenty four hours after transfection of the luciferase constructs, the cells were lysed and luciferase activity was measured. Renilla and firefly luciferase activity were estimated by dual luciferase assays and presented as the normalized ratio of Renilla vs Firefly luciferase activity. The data represent six replicates in each group from two independent experiments. The mean ±SE are depicted as horizontal and vertical bars for each group, respectively. Non-parametric Wilcoxon-Mann-Whitney tests were used for statistical comparisons and two tailed p value is indicated.