Fig. 6

Downregulation of BORIS and epigenetic remodeling show concordant activation of SVA transcription in K562 cells. a The two SVA elements with highest expression in testis with the position of BORIS ChIP-seq peaks in K562. The absence of strong BORIS binding indicates that BORIS is unlikely to act as SVA activator in testis. Genomic coordinates are in kb. The SVA-D shown is intergenic, while the SVA-B is antisense intronic. b RT-qPCR shows the downregulation of BORIS in K562 clones with stable integration of Tet-On inducible anti-BORIS shRNA constructs (site 1 and site 2), 48 h after shRNA induction. Un-infected K562 cells and a clone with the integrated empty vector were controls. Only the experiments in the presence of doxycycline (Dox+) are shown. c Immunoblotting with anti-BORIS mAbs demonstrates fourfold–fivefold depletion of BORIS protein in shRNA clones. The quantification of relative BORIS amount (white numbers) was performed using LiCor software and alpha-tubulin as a reference. d RNA-seq analysis of differential expression of 2223 SVA elements longer than 1 Kb mapped in the human genome versus K562 infected with the empty vector (SVAs that were constitutively silent were not included). Shown are the distributions of ratios of RNA-seq difference in: BORIS KD K562 cells (paired two-tailed t test p value <0.001), K562 treated with 5-AzadCyD (5Aza) (paired two-tailed t test p value <0.001), and K562 treated with DZNep (paired two-tailed t test p value = 0.001). The RNA-seq reads enrichments for SVA elements were normalized to the total number of reads in each individual experiment. The mean and standard deviation diagrams are shown on top of each graph. The graphs demonstrate the overall increase in the shift toward higher SVA expression from BORIS KD to 5Aza and especially DZNep treatments. Vertical blue lines correspond to the unchanged expression over control, red—to twofold increase. e The RNA-seq analysis of BORIS KD, DZNep treatment, and the combination of both on the transcription of SVA elements that were apparently silent in control experiments (i.e., <10 normalized counts with over twofold increase by any treatment) shows a reproducible compound effect of BORIS KD and DZNep treatment on SVA activation (for the latter, the whole 2223 SVA sample’s paired two-tailed t test p value <0.001). The corresponding means of the distributions with standard errors of the mean: KD 1.03 ± 0.03, DZNep 1.37 ± 0.03, KD and DZNep 1.55 ± 0.02. f Dot-plot of RNA-seq normalized counts of BORIS KD versus DZNep treatment of K562, expressed as fold enrichment over the empty vector control. Only the SVA elements that were silent in the control are included. The blue lines correspond to cutoffs with no change in expression. g SVA elements that show concordant activation by BORIS KD and DZNep treatment do not belong to a preferential SVA class. The pie diagrams show the breakdown of SVA classes among all 2223 elements included in the analysis and among 471 elements co-activated by both treatments