Fig. 1

HIV 5′-LTR amplification and sequencing strategy. The 5′-LTR region (structure of the reference HXB2 sequence detailed in the figure, including canonical CpG and transcription factor binding sites) was amplified by nested PCR from total CD4⁺ T cell DNA treated with sodium bisulfite. The first-round PCR generated an 856-bp product including a fragment of gag, in order to selectively amplify the 5′-LTR and not the 3′-LTR. Nested PCR was performed with primers including Illumina adaptors for next generation sequencing. A 349-bp amplicon including nine of the canonical HXB2 CpG positions in the U3-R region was generated and deep sequenced. Assuming that each read corresponded to a single amplicon, we quantified, normalized, and aligned all variants obtained for each sample (the number of reads and proportion in the sample is indicated next to each variant). Each canonical CpG position is represented with a circle. Positions resistant to bisulfite transformation were assumed to be methylated (black) and positions susceptible to bisulfite transformation were assumed to be non-methylated (white). Positions with mutations causing loss of the CpG site were also identified (gray). We then generated a summary representation of all the variants observed in each sample, using a pie chart per each methylation-susceptible position (colors represent the proportion of methylated, non-methylated and mutated variants at that specific site in the sample as above), numbered according to order of appearance in HXB2. We also included additional CpG, CHG, and CHH methylation-susceptible sites, not observed in the HXB2 sequence (marked with *). Finally, we defined a CpG Methylation Index metric per sample, averaging the proportion of methylated variants in each of the nine 5′-LTR canonical CpG sites analyzed