Fig. 5: Transcriptional silencing of the HIV proviral genome by concomitant activation of p53 signalling and KLF2, a key transcription factor associated with the quiescence/resting state, after HIV infection.

a, Uniform manifold approximation and projection (UMAP) of scRNA-seq data derived from a healthy donor with 2 biological replicates in addition to technical replicates shows HIV+ 96-hpi cells clustering between uninfected proliferating cells and HIV+ quiescent cells. b, Distribution of +HIV 96-hpi cells expressing MYC, KLF2, p53 pathway genes, or both p53 pathway genes and KLF2. c, Level of expression of HIV provirus in cells expressing p53 pathway genes, MYC, KLF2 or p53 pathway genes and KLF2 in combination. d, Proviral transcriptional shutdown is associated with dual activation of p53 and KLF2 signalling in ex vivo HIV-infected CD4+ T cells. In each violin plot, the central box, white dot and whiskers indicate the interquartile range (25th–75th percentiles), median and 1.5× IQR. Two-tailed Mann–Whitney U-test was used for the comparisons of KLF2+p53+ vs KLF2+53−, MYC+ and KLF2−p53+, yielding U statistics of 293.0, 667.5 and 4,541.5, respectively. As the primary hypothesis tested was that the KLF2+p53+ group differed from all three comparator groups, no multiple-comparison correction was required (intersection–union framework). e, UMAP of single-cell RNA-seq data from 6 HIV+ donor CD4+ T cells. Colours show the separation of resting and ex vivo activated cells. f, Side-by-side UMAPs show the distribution of resting and ex vivo activated cells. g, Clustering study showing the presence of different resting and activated states. h, Expression pattern of key quiescence and activation markers on UMAPs identifies 3 clusters which largely consist of activated cells. The color bar indicates log1p-transformed gene expression values. i, Dotplot identifies clusters representing activated, resting and intermediate states. j, Dotplot showing the strong induction of MYC target genes in the 2 most activated and inflamed clusters. k, Trajectory analysis indicates the separation of the cells into untreated (largely KLF2+ resting cells before ex vivo reactivation) and ex vivo reactivated groups, with the majority of the reactivated group in nodes 4, 9 and 10. l, The two MYC+ clusters in j constitute over 80% of cells in nodes 4, 9 and 10 which correspond to the endpoint of the trajectory, highlighting the strong association of MYC with the reactivated state. Node colors correspond to the cluster identities defined in g, and the pie chart proportions within each node represent the relative contributions of each cluster. m,n, Expression pattern of KLF2 and MYC indicate their association with the resting and activated state, respectively. Colour bars to the right of each plot indicate the averaged expression level per node. o–r, Expression patterns of multiple markers of the activated state map to nodes 4, 9 and 10. s,t, Calculations of the AUC of ROC plots indicate that the expression patterns of CD40LG and KLF2 are excellent predictors of the activated and quiescent states, respectively.