Fig 2: MLL-ELL transforms hematopoietic progenitors via association with AF4/ENL/P-TEFb (AEP), but not with EAF1 or p53.(A) Mutations of ELL selectively abrogate interaction with EAF1. Immunoprecipitation (IP)-western blotting of the chromatin fraction of HEK293TL cells transiently expressing FLAG-tagged GAL4-ELL proteins with or without S600A/K606T substitutions (SA/KT) along with xAF4 and EAF1h was performed as in Figure 4D. (B) Recruitment of exogenously expressed AF4 or EAF1 by ELL mutant proteins. Chromatin immunoprecipitation (ChIP)-qPCR analysis of HEK293TL cells transiently expressing FLAG-tagged GAL4-ELL proteins with or without the SA/KT mutation along with xAF4 and EAF1h was performed as in Figure 4C. (C) Effects of the SA/KT mutation on MLL-ELL-mediated leukemic transformation ex vivo. Various MLL-ELL constructs with or without the SA/KT mutation were examined for the transformation of myeloid progenitors as in Figure 1A. (D) Effects of the SA/KT mutation on MLL-ELL-mediated leukemic transformation in vivo. MLL-ELL or its SA/KT mutant was transduced to c-Kit-positive hematopoietic progenitors and transplanted into syngeneic mice (n = 20). Primary leukemia cells were harvested from the bone marrow and transplanted into secondary recipient mice (n = 20). (E) Enhanced interaction of ELL with AEP mediated by the trithorax homology domain 2 (THD2) domain. IP-western blotting of the chromatin fraction of virus-packaging cells, transiently expressing various HA-tagged MTMT-ELL fusion constructs (with or without the SA/KT mutation), was performed. Endogenous proteins co-purified with MLL-ELL proteins were visualized by specific antibodies.

Fig 3: AF4 and EAF1 form distinct SL1/MED26-containing complexes.(A) A schema of the structures of AF4, ELL, and EAF1. ALF is responsible for interaction with ELL, pSER is responsible for interaction with SL1, and DLXLS is responsible for interaction with MED26. EAF family homology domain (EHD) is responsible for homodimerization and interaction with ELL. NKW:NKW motif responsible for transcriptional activation. (B) Alignment of the amino acid sequences of the SDE, DLXLS, and NKW motifs of human AF4 family proteins and EAF family proteins. Conserved residues are indicated by asterisks. (C) Association of EAF1, ELL, and AF4 domains with various associating factors on chromatin. Immunoprecipitation (IP)-western blotting of the chromatin fraction of HEK293TL cells transiently expressing various FLAG-tagged GAL4 fusion proteins, with or without indicated HA-tagged constructs, was performed as in Figure 4D. Endogenous proteins were detected by specific antibodies for each protein, while exogenous proteins were detected by antibodies for FLAG or HA tag. (D) Recruitment of various transcriptional regulatory proteins by EAF1, ELL, and AF4 domains. Chromatin immunoprecipitation (ChIP)-qPCR analysis of HEK293TL cells transiently expressing various combinations of FLAG-tagged GAL4 fusion proteins along with indicated HA-tagged constructs was performed as in Figure 4C. (E) Putative complex recruitment mediated by ELL-AF4 interaction. (F) Putative complex recruitment mediated by ELL-EAF1 interaction. (G) Structural requirement of MTM-AF4 pSER fusion and MTMT-EAF1 fusion. Various MTMh/MTMTh constructs fused with EAF1 domains were examined for the transformation of myeloid progenitors as in Figure 1A.

Fig 4: SPT6 is necessary for human epidermal differentiation.a Eight elongation factors were knocked down greater than 75% by siRNAs. Knockdown cells were placed in high confluence and calcium for 3 days to induce differentiation. RT-qPCR was used to quantify the knockdown levels of each elongation factor shown on the X-axis. Two separate siRNAs (SPT6-1i and SPT6-2i) targeting different regions of SPT6 were used. qPCR values were normalized to L32. n = 3 independent experiments. b Western blot for knockdown of SPT5 and ELL. N = 3 independent experiments which all showed similar results. c, d RT-qPCR quantifying the relative mRNA expression levels of epidermal differentiation genes TGM1 and GRHL1 in differentiated cells as described in a. n = 3 independent experiments. e Immunofluorescent staining of late differentiation markers FLG (red) and LOR (green) in day 6 regenerated human epidermis treated with control (CTLi) or two distinct siRNAs targeting SPT6 (SPT6-1i and SPT6-2i). Merged image includes Hoechst staining of nuclei. n = 3 independent experiments, scale bar = 50 μm. f Immunofluorescent staining of early differentiation marker K10 (red) and proliferation marker ki67 (green) in day 6 regenerated human epidermis treated with control (CTLi) or two distinct siRNAs targeting SPT6 (SPT6-1i and SPT6-2i). Merged image includes Hoechst staining of nuclei. n = 3 independent experiments. Scale bar = 50 μm. g Quantification of the %ki67 positive cells in the basal layer from images taken in f. n = 3 independent experiments. At least 100 basal layer cells were quantitated per experiment. h Hematoxylin and eosin staining of CTLi or SPT6i knockdown regenerated human epidermis (day 6). N = 3 independent experiments which all showed similar results. Scale bar = 50 μm. i Quantification of the tissue thickness from images taken from e, f. n = 3 independent experiments. j RT-qPCR quantifying the relative mRNA expression levels of epidermal differentiation genes in CTL and SPT6 knockdown regenerated human epidermis (day 6). N = 3 independent experiments. k RNA-Seq analysis of CTL and SPT6 knockdown regenerated human epidermis (day 6). N = 3 independent experiments. In total, 4275 genes were upregulated (red) and 1465 genes were downregulated (blue) upon SPT6 depletion. Heatmaps are shown in Log2 scale. l Gene ontology (GO) terms for the 4275 genes upregulated in SPT6i tissue using hypergeometric distribution. m GO terms for the 1465 genes reduced in expression upon SPT6 loss using hypergeometric distribution. n Overlap of the SPT6 knockdown gene expression signature with the differentiation signature. p value was calculated using hypergeometric distribution. Mean values are shown with error bars = SD. ****p < 0.0001 (two-sided t-test for c, d, g, i, j).

Fig 5: Trithorax homology domain 2 (THD2)-mediated functions promote MLL-fusion-dependent leukemic transformation.(A) Structure/function analysis of MLL-ELL. Various MLL-ELL constructs were examined for the transformation of myeloid progenitors. HA-tag (h: indicated as a red triangle) was fused to MTM and MTMT constructs. A schema of myeloid progenitor transformation assay is shown on top. Hoxa9 expression normalized to Gapdh in first-round and second-round colonies (left) is shown as the relative value of MLL-ELL (arbitrarily set at 100%) with error bars (mean ± SD of PCR triplicates). Colony-forming ability at third- and fourth-round passages (right) is shown with error bars (mean ± SD of ≥3 biological replicates). IBD: integrase-binding domain; hMBM: high-affinity MENIN-binding motif; LBD: LEDGF-binding domain; THD1/2: trithorax homology domains 1 and 2; SET: Suver3-9/enhancer-of-zeste/trithorax motif; TED: transcription elongation domain; OHD: occludin homology domain. (B) Requirement of THD2 for various MLL fusion proteins in leukemic transformation. Various MLL fusion constructs were examined for the transformation of myeloid progenitors as in (A). RA1/2: RAS association domains 1/2; PDZ: PSD-95/Dlg/ZO-1 domain; OM: octapeptide motif; LZ: leucine zipper. MISD: minimum interaction site of DOT1L: YEATS: Yaf9 ENL AF9 Taf14 Sas5 domain; AHD: ANC1 homology domain. (C) Colony morphologies of MTMTh or MTMh-ENL′ transformed cells. The colonies on day 5 of fourth passage are shown with a scale bar. The ratio of compact colonies (≥100 total colonies were counted in each experiment) is shown on the left (mean ± SD of six biological replicates). Representative images are shown on the right.