Fig. 2: Drug resistance regulated by UCEs.
a, UCE–drug interactions in K562 cells. MAGeCK MLE algorithm was used to identify UCEs and NCREs involved in imatinib resistance on the basis of dual-CRISPR screens. Three different cell populations were used: the day-0 population, the day-15 imatinib treatment population and the day-15 non-treatment control population. The beta score indicates the degree of selection upon UCE or NCRE removal relative to the day-0 initial population. The y axis represents the beta scores of the day-15 imatinib treatment. The x axis shows beta scores of the day-15 non-treatment condition. The horizontal and vertical dashed lines indicate the mean ± 1s.d. of the day-15 imatinib treatment and the day-15 non-treatment control beta score, respectively. The diagonal dashed line indicates the mean ± 1s.d. of the differential beta scores, which can be calculated by subtracting the day-15 non-treatment control from the day-15 imatinib treatment beta score. The orange group shows UCEs or NCREs conferring imatinib resistance upon removal; the purple group shows UCEs or NCREs sensitizing cells to imatinib treatment upon removal. Selected UCEs for downstream analyses are marked. b,c, Knockout ZNF503_Op (b) and QKI_Jo (c) from K562 cells using the dual-CRISPR system. K562 cells were transfected with the respective guide RNA pairs that target the indicated UCEs. Single-cell clones with the respective UCE deletions were selected. The blue and red arrowheads indicate the intact genomic regions and the UCE deletions, respectively. d, Drug resistance conferred by ZNF503_Op and QKI_Jo knockouts. The ZNF503_Op and QKI_Jo KO cells were treated with 0.4 μM imatinib for 3 days. Cell viability was measured using CellTiter-Blue (n = 3 independent biological samples; bars show mean ± s.d.; ZNF503_Op_KO_#1 *P = 0.0366, ZNF503_Op_KO_#2 *P = 0.0415 and QKI_Jo_KO_#1 *P = 0.0108, QKI_Jo_KO_#2 *P = 0.0102, calculated using two-tailed unpaired t-test). e, Enhancer activities were measured by luciferase assay. The respective UCEs were cloned by PCR into the enhancer reporter plasmid pGL4.23 with a minimal promoter. The empty pGL4.23 plasmid was used as the control for the baseline luciferase activities. The y axis represents the relative unit of luciferase activity compared to that of pGL4.23 empty plasmids (n = 3 independent biological samples; bars show mean ± s.d.; ZNF503_Op NSP = 0.9705, QKI_Jo **P = 0.0029, calculated using two-tailed unpaired t-test). f, Silencer activities were measured by luciferase assay. The respective UCEs were cloned by PCR into the silencer reporter plasmid pGL4.53 with a PGK promoter. The empty pGL4.53 plasmid was used as the control for the baseline luciferase activities. The y axis represents the relative unit of luciferase activity compared to that of pGL4.53 empty plasmids (n = 3 independent biological samples; bars show mean ± s.d.; ZNF503_Op ***P = 0.0002 and QKI_Jo **P = 0.0029, calculated using two-tailed unpaired t-test). g,h, Potential genes regulated by the UCEs. TADs identified by Hi-C surrounding ZNF503_Op (g) and QKI_Jo (h) are shown and the locations are indicated by vertical blue bars. Horizontal yellow and blue bars indicate distinct TADs. Transcriptions of nearby genes of the different KO clones were quantified by qPCR (n = 3 independent biological samples; bars show mean ± s.d.; ZNF503_Op_KO_#1: SAMD8 **P = 0.0049, VDAC2 **P = 0.007, ZNF503 ***P = 0.0008; ZNF503_Op_KO_#2: SAMD8 ****P < 0.0001, VDAC2 NSP = 0.1242, ZNF503 ***P = 0.0003. QKI_Jo_KO_#1: PACRG *P = 0.0111, QKI **P = 0.0011; QKI_Jo_KO_#2: PACRG *P = 0.0126, QKI NSP = 0.7127, calculated using two-tailed unpaired t-test).
