RepID represses megakaryocytic differentiation by recruiting CRL4A-JARID1A at DAB2 promoter

Background Megakaryocytes (MKs) are platelet precursors, which arise from hematopoietic stem cells (HSCs). While MK lineage commitment and differentiation are accompanied by changes in gene expression, many factors that modulate megakaryopoiesis remain to be uncovered. Replication origin binding protein (RepID) which has multiple histone-code reader including bromodomain, cryptic Tudor domain and WD40 domains and Cullin 4-RING ubiquitin ligase complex (CRL4) recruited to chromatin mediated by RepID have potential roles in gene expression changes via epigenetic regulations. We aimed to investigate whether RepID-CRL4 participates in transcriptional changes required for MK differentiation. Methods The PCR array was performed using cDNAs derived from RepID-proficient or RepID-deficient K562 erythroleukemia cell lines. Correlation between RepID and DAB2 expression was examined in the Cancer Cell Line Encyclopedia (CCLE) through the CellMinerCDB portal. The acceleration of MK differentiation in RepID-deficient K562 cells was determined by estimating cell sizes as well as counting multinucleated cells known as MK phenotypes, and by qRT-PCR analysis to validate transcripts of MK markers using phorbol 12-myristate 13-acetate (PMA)-mediated MK differentiation condition. Interaction between CRL4 and histone methylation modifying enzymes were investigated using BioGRID database, immunoprecipitation and proximity ligation assay. Alterations of expression and chromatin binding affinities of RepID, CRL4 and histone methylation modifying enzymes were investigated using subcellular fractionation followed by immunoblotting. RepID-CRL4-JARID1A-based epigenetic changes on DAB2 promoter were analyzed by chromatin-immunoprecipitation and qPCR analysis. Results RepID-deficient K562 cells highly expressing MK markers showed accelerated MKs differentiation exhibiting increases in cell size, lobulated nuclei together with reaching maximum levels of MK marker expression earlier than RepID-proficient K562 cells. Recovery of WD40 domain-containing RepID constructs in RepID-deficient background repressed DAB2 expression. CRL4A formed complex with histone H3K4 demethylase JARID1A in soluble nucleus and loaded to the DAB2 promoter in a RepID-dependent manner during proliferation condition. RepID, CRL4A, and JARID1A were dissociated from the chromatin during MK differentiation, leading to euchromatinization of the DAB2 promoter. Conclusion This study uncovered a role for the RepID-CRL4A-JARID1A pathway in the regulation of gene expression for MK differentiation, which can form the basis for the new therapeutic approaches to induce platelet production.

This study uncovered a role for the RepID-CRL4A-JARID1A pathway in the regulation of gene expression for MK differentiation, which can form the basis for the new therapeutic approaches to induce platelet production.
CRL4 also interacts with core components of histone methylation complexes, including EED, Polycombgroup protein, and the core subunit WDR5 of the human MLL histone H3K4 methyltransferase that are essential for epigenetic and transcription controls. Therefore, CRL4 inactivation impairs histone modi cations (28). The Cullin-4B anchored CRL4B can promote chromatin silencing through its association with the polycomb-repressive complex PRC2, the histone methyltransferase SUV39H1, the heterochromatin protein 1 (HP1) and the DNA methyltransferase 3A (DNMT3A), leading to H2AK119 monoubiquitination and subsequently to H3K27 trimethylation, H3K9 trimethylation and DNA methylation (29,30). These molecular characteristics suggest that RepID and CRL4 are crucial epigenetic controllers that trigger gene transcription or silencing, but the mechanistic details of those interactions remain to be fully elucidated.
Here, we report that RepID represses MK differentiation by recruiting CRL4A to the DAB2 promoter, resulting in heterochromatin formation. RepID dissociates from chromatin in response to megakaryocytic differentiation signals, thereby preventing the loading of CRL4A as well as the histone H3K4 demethylase JARID1A, a novel interaction partner of CRL4A, to the DAB2 promoter. These molecular events correlated with increased transcription of megakaryocyte markers and megakaryocytic morphological changes in a subpopulation of RepID-depleted cells. We also observed that megakaryocytic differentiation is accelerated in RepID-depleted cells upon stimulation by phorbol 12-myristate 13-acetate (PMA). These ndings provide an example of gene transcription regulation during megakaryocytic differentiation with the RepID-dependent recruitment of an epigenetic modi er.

Methods
Cell culture and chemicals

RT2-pro ler PCR array and qRT-PCR analysis
Proliferating K562 RepID WT and KO cells were harvested, and RNA was extracted using the RNeasy R Mini Kit (Qiagen, 74104) and converted to cDNA by using PrimeScript RT Master Mix (Takara, RR036A) according to the manufacturer's instructions. An RT 2 Pro ler PCR array (Qiagen, PAMM-043Z) with 84 genes associated with the Wnt signaling pathway was used.
Total RNA extraction, cDNA synthesis, and qRT-PCR were performed according to the manufacturer's instructions provided by Qiagen ( Proximity ligation assay (PLA) The Duolink in situ Red starter mouse/rabbit PLA uorescence assay (Sigma, DUO92101) was performed according to the manufacturer's instructions. Brie y, proliferating or PMA-induced differentiating K562 cells were attached to slides using cytospin. Cells were washed with 1×PBS and xed for 15 min at 4°C in 4% paraformaldehyde in PBS and permeabilized with 0.1% Triton X-100 in PBS for 5 min at 4°C. The coverslips were blocked with Duolink blocking solution and incubated with CUL4A (Abcam, ab92554, 1:200) and JARID1A (Abcam, ab78322, 1:200) antibodies in Duolink antibody diluent overnight, followed by incubation with PLUS and MINUS PLA probes, ligation, and ampli cation. The coverslips were then washed and mounted using mounting medium containing DAPI. Images were captured using a wide-eld microscope, processed, and analyzed using Zeiss Zen 3.7 and ImageJ software.

Cell morphology analysis
Proliferating or differentiating K562 cells were attached to slides using TXT3 cytospin (Nasco Korea) at 900×g for 5 min and stained with Wright-Giemsa staining solution (Sigma, 45253 and 32884). Cell morphology, including cell size and polynucleation, was analyzed by microscopy and ImageJ software. At least 290 cells in randomly separated elds were counted and expressed as relative cell size or percentage of the total cell number in these elds.

Database analysis
Pearson's correlation R values according to the gene expression levels between CRL4 components (RepID, CUL4A, CUL4B, and DDB1) and histone H3K4 methyltransferase/demethylase were analyzed using the Cancer Cell Line Encyclopedia (CCLE) dataset from the CellMiner cross-database (CellMinerCDB, https://discover.nci.nih.gov/cellminercdb/) (31). The Pearson correlation R-value was represented as a heatmap. Interaction pro les between CRL4 components and histone H3K4 methylation modi ers were obtained from the Human BioGRID Interaction Database (version 4.4.214 Oct 2022).

RepID negatively correlates with DAB2 expression
To determine whether RepID participates in transcriptional changes, we performed a PCR array assay using cDNAs derived from RepID-pro cient or RepID-de cient K562 erythroleukemia cell lines. Interestingly, the highest observed gene deregulation involved the Wnt/b-catenin pathway with Frizzled 3 receptor (FZD3), Disabled-2 (DAB2), and Van Gogh-like 2 (VANGL2) being upregulated in RepID-de cient K562 cells more than 20 fold (Fig. 1A, left panel). Conversely, CXXC Finger Protein 4 (CXXC4) increased 16.62 folds in RepID WT K562 cells (Fig. 1A, right panel). Interestingly, the Wnt signaling has been shown to have a profound role in platelet production (32). Of interest, DAB2 was shown to be upregulated when both hematopoietic stem cells and the human K562 cells undergo megakaryocytic differentiation (33). In silico analyses of cancer cell lines from the Cancer Cell Line Encyclopedia (CCLE) via the Cellminer database showed that the expression of RepID transcripts negatively correlates with DAB2 expression (Fig. 1B, C). Subsequent quantitative reverse transcription PCR and immunoblot analysis at a subset of cancer cell lines, such as DMS114 or K562 cells, induced transcription as well as protein expression levels of DAB2 in RepID-depleted conditions (Fig. 1D, E). Altogether, our results imply that RepID may transcriptionally repress DAB2 expression in some cancer cells.
Next, we tested whether RepID depletion could interfere with the induction of megakaryocyte markers and/or megakaryocytic differentiation. In RepID-expressing cells, DAB2 protein levels gradually increased after PMA treatment (Fig. 2E, F, WT panel). However, in RepID-depleted cells, the expression of the DAB2 protein reached maximum levels earlier than in RepID-pro cient cells (48 hr in KO cells vs. 96 hr in WT cells, Fig. 2E, F, KO panel). These results were consistent with the confocal microscopy observations, in which DAB2 clearly appeared in polynucleated cells in response to PMA treatment and was more strongly induced in RepID-de cient cells (Fig. 2G, PMA panel). Moreover, results showing similar intensity of DAB2 levels between larger cells with polynucleate nuclei in RepID-de cient cells and single-nuclei cells in the absence of PMA exposure con rmed that DAB2 was relatively more abundant in RepID-depleted cells (Fig. 2G, CTL panel). In line with DAB2 induction, RepID-de cient K562 cells showed more advanced morphological changes, regarded as megakaryocytic differentiation (Fig. 2H), together with rapidly reaching maximum transcription levels of megakaryocyte markers, compared to RepID-expressing K562 cells (Fig. 2I). These results demonstrate that RepID-de cient K562 cells exhibit accelerated megakaryocytic differentiation.
RepID-CRL4A dissociates from chromatin during megakaryocytic differentiation Since our observations suggest that RepID may be a crucial factor for DAB2 expression, we sought to identify the mechanism of action of RepID for altering DAB2 transcription. RepID contains three functional domains: a WD40-repeat containing domain that serves as scaffold for protein-protein interactions, the two bromodomain and a tudor chromatin binding domains (24). Because we previously show that RepID WD40 domain was crucial for its functional activity during DNA replication (16), we generated stable K562 cell lines expressing FLAG-tagged full length RepID and a RepID fragment lacking the WD40 domain in RepID-depleted cells (Fig. 3A). The results showed that DAB2 expression was compromised with the RepID full-length (FL), but not with the RepID fragment lacking the WD40 domain (ΔWD40) (Fig. 3B, C). Additionally, there were no substantial alterations in the protein levels of DAB2 in both RepID-expressing and RepID-depleted cells treated with the protein synthesis inhibitor cycloheximide or the proteasome inhibitor MG132. This result suggests that DAB2 is not a substrate for CRL4, but the WD40 domain of RepID is crucial for the transcriptional repression of DAB2 (Fig. 3D).
Together with RepID, CRL4 is also a possible factor that alters gene transcription by regulating heterochromatin formation (28-30). Therefore, we investigated whether the expression or chromatinbinding ability of RepID or CRL4 could be altered during megakaryocytic differentiation. Transcriptional intensity from published microarray datasets comparing treated (20 nM, 2 days) and untreated PMA K562 cells showed a reduction in CUL4A and RepID expression but a markedly increased of CUL4B and CUL7 transcripts, whereas CRL4 adaptor DDB1 was not changed (Fig. 3E). Consistent with these observations, protein expression levels of RepID, CUL4A or CUL4B changed slightly during PMA-mediated differentiation (Fig. 3F, G, WCL panel). Interestingly, chromatin-bound RepID decreased substantially in differentiating cells, leading to CUL4A dissociation from chromatin (Fig. 3F, G, RepID and CUL4A panel). In contrast, the chromatin binding of CUL4B and DDB1 increased slightly, even in RepID-depleted cells, suggesting that these proteins may be loaded onto chromatin by another mechanism (Fig. 3F, G, CUL4B and DDB1 panel). These results also suggest that CUL4A and CUL4B may have different functions in megakaryocytic differentiation and that RepID-CRL4A dissociation from chromatin seems to be correlated with the induction of DAB2 transcription. DAB2 promoter forms euchromatin by dissociation of the CRL4A-JARID1A complex during megakaryocytic differentiation A ChIP-PCR analysis was used to determine if RepID-CRL4A in uences epigenetic changes in the DAB2 promoter. The ratio of H3K4me3 to H3K9me3 clearly differed according to RepID expression, showing euchromatin formation in DAB2 promoter regions mediated by a decrease in H3K9me3 and an increase in H3K4me3 levels in RepID-depleted K562 cells (Fig. 4A). Therefore, we analyzed whether the expression of histone modi ers, including H3K4 methyltransferase and demethylase, was altered by the expression of CRL4 components, including RepID, CUL4A, CUL4B, and DDB1. However, our in-silico analyses of cancer cell lines from the CCLE of the Cellminer dataset did not show considerable expression correlations between the modi ers and CRL4 components (Fig. 4B, left panel). Next, we examined whether euchromatin formations of DAB2 promoter was due to interaction between the modi ers and CRL4 components. Although the protein-protein interaction database in BioGRID showed few modi ers including RBBP5 or WDR5 as CRL4 interactors (Fig. 4B, right panel), our immunoprecipitation analysis with antibodies directed against endogenous CUL4A or JARID1A found that CRL4A forms a complex with JARID1A (Fig. 4C, D), and their interaction decreased during megakaryocytic differentiation (Fig. 4E). In line with RepID-CRL4A dissociation from the chromatin, JARID1A chromatin binding and expression levels were also substantially reduced in response to PMA-mediated differentiation. Unlike the other modi ers we tested, high amounts of JARID1A left chromatin in RepID-de cient cells compared to RepIDexpressing cells (Fig. 4F, G). To validate the interaction between CRL4A and JARID1A, we performed a proximity ligation assay (PLA) that allows in situ detection of protein interactions. Without pre-extraction, we did not detect substantial differences in interaction signals between RepID-expressing and RepIDdepleted cells (Fig. 4H, left panel), whereas the signals for JARID1A-CRL4A interaction completely disappeared in pre-extracted RepID-de cient nuclei (Fig. 4H, middle panel), suggesting that most CRL4A-JARID1A interactions may occur in soluble nuclear fractions and then be recruited to chromatin mediated by RepID. However, similar amounts of chromatin-bound JARID1A in RepID-expressing and RepIDde cient K562 cells without PMA treatment suggest that a subset of JARID1A forms a complex with CRL4A. Compared to proliferating K562 cells, cells differentiating into megakaryocytes lost the CRL4A-JARID1A interaction faster in RepID-de cient cells than in RepID-pro cient cells (Fig. 4H, right panel), suggesting that CRL4A dissociates from JARID1A during megakaryocytic differentiation. These results were consistent with the ChIP-qPCR results, in which the fraction of JARID1A and CRL4A bound to the DAB2 promoter decreased substantially in response to PMA treatment in both cell types (Fig. 4I). In addition, DAB2 promoter-bound disappearance of JARID1A-CRL4A was faster in RepID-depleted cells. This observation could be correlated to the fact that these cells show very low amounts of CRL4A on the DAB2 promoter (Fig. 4I). Our observations also noted that a small fraction of JARID1A can still bind to the DAB2 promoter in RepID-depleted cells and would suggest that a small fraction of JARID1A can be loaded on the DAB2 promoter in a RepID-independent manner (Fig. 4I). Taken together, our observations suggest that RepID may modulate megakaryocytic differentiation by regulating the loading of a subset of CRL4A-JARID1A onto the DAB2 promoter.

Discussion
Megakaryocytes (MKs) are highly specialized large cells that are crucial precursors of blood platelets, small anucleate cell fragments with crucial roles in wound healing, in ammation, angiogenesis and hemostasis. MKs differentiation leads to polyploidy by endomitosis followed by a maturation process for platelet production, and these processes are governed by intensive changes in gene expression patterns. The data presented here demonstrate that megakaryocytic differentiation requires RepID dissociation from chromatin. The loss of chromatin-bound RepID compromises the loading of the CRL4A-JARID1A complex at the DAB2 promoter, leading to the initiation of DAB2 transcription, a key regulator of platelet signaling. MK differentiation was partially executed in K562 RepID-de cient cells, even in the absence of a differentiation signal, while induced MK differentiation by PMA led to faster megakaryocytes-like cell formation when using K562 RepID-depleted cells. Altogether, these observations identi ed the RepID-CRL4A-JARID1A complex as a putative factor in the maintenance of K562 proliferation.
RepID, known as the replication origin binding protein as well as a structural receptor for CRL4 ubiquitin E3 ligase, initiates a subpopulation of DNA replication origins and prevents over-replication via CRL4 recruitment on chromatin to ubiquitinate CDT1 (16, 17,[37][38][39]. RepID shows direct chromatin binding ability using its three histone-code reader bromodomain, cryptic Tudor domain and WD40 domains. Genome-wide ChIP-Seq results reveal that most of the RepID binding sites are concentrated in promoter regions (16,22,24,40,41). Moreover, reduced heterochromatin formation due to compromised interaction between CRL4A and the epigenetic silencing factors EED, a subunit of the PRC2 methyltransferase, or the histone H3K4me3 demethylase JARID1C suggest that RepID-CRL4 may regulate both DNA replication and transcription landscapes (28, 42). We identi ed JARID1A as a novel interaction protein of CRL4A, with the JARID1A-CRL4 complex being prominently present in the soluble nuclear fraction. The JARID1A-CRL4 complex seems to be recruited to the DAB2 promoter in a RepID dependent manner to repress DAB2 transcription. Upon PMA-induced megakaryocytic differentiation of K562, we observed that the CRL4A-JARID1A complex dissociates from chromatin genome wide, including at the DAB2 promoter. This dissociation might re ect the absence of chromatin recruitment by RepID. RepID is a crucial player for cell proliferation in driver-negative tumors, including melanoma, breast, and lung cancers which lose HER2, EGFR, ALK, or BRAF expression (43). The observations reported here suggest that RepID may also play a role in hematopoietic proliferation by inhibiting CRL4A-JARID1A-mediated repression of genes correlated with megakaryocytic differentiation. For example, megakaryocytic differentiation is closely related to the canonical Wnt signaling pathway, which plays an essential role in cell fate determination, differentiation, and proliferation. Previous reports have shown that platelet function is negatively regulated by canonical Wnt signaling (44). Interestingly, RepID-repressed DAB2 is a negative regulator of Wnt/b-catenin pathway. DAB2 induction has also been reported during erythrocyte differentiation of mouse embryonic stem cells and K562 cells (45), suggesting that DAB2 may be an initial factor for cell fate decisions. These studies can lead to a better understanding of RepID's role in erythropoiesis.
Our results revealed that the dynamics of chromatin loading differ between CUL4A (decreased loading) and CUL4B (increased loading) during induced megakaryocytic differentiation. Human CUL4A and CUL4B share approximately 82% sequence identity, 90% homology and contribute to redundant functions in cell cycle progression and embryonic development by mediating the ubiquitination of CRL4 targets (16, 46-48). However, distinctive functions of CUL4B have also been revealed in the assembly of atypical CRL4B ubiquitin ligase complexes integrated with dioxin receptors to target sex steroid receptors for degradation (49). Speci c CUL4B mutations have been identi ed as genetic defects responsible for Xlinked mental retardation (50,51). Therefore, in line with previous reports, we propose that CUL4A and CUL4B may have speci c functions on chromatin leading to the alterations of undiscovered targets and/or gene expression, that, in turn, would modulate proliferation and differentiation. It would therefore be interesting to determine which genes or proteins are differentially regulated by CRL4A or CRL4B during megakaryocytic differentiation.
We propose the following model for the role of RepID in repressing the expression of genes required for megakaryocytic differentiation by recruiting the CRL4A-JARID1A complex (Fig. 5). In proliferating condition, CRL4A forms a complex with JARID1A in the nucleus, that is recruited to the DAB2 promoter by RepID, that, in turn, leads to heterochromatization of DAB2 promoter and DAB2 reduced expression. A subset of JARID1A is also recruited to the DAB2 promoter in a RepID-independent manner. In response to induced megakaryocytic differentiation, CRL4A-JARID1A dissociates from the DAB2 promoter due to the three mechanisms; 1. the reduction in chromatin-bound RepID, 2. downregulation of CRL4A, JARID1A and RepID expression, 3. complex dissociation between CRL4A and JARID1A. RepID-independent JARID1A is also dissociated from the DAB2 promoter, leading to a high ratio of H3K4me3 required for the upregulation of DAB2 transcription. In RepID-de cient cells, a subset of JARID1A that interacts with CRL4A cannot be loaded onto the DAB2 promoter because CRL4A loading onto chromatin is compromised. Therefore, genes required for megakaryocytic differentiation may be partially transcribed that is translated to increased number of RepID-depleted cells with megakaryocytic traits of with an accelerated response to PMA treatment. These ndings provide novel mechanistic insights into the dynamics of RepID-mediated complex-based megakaryocytic differentiation, which can form the basis for the development of new therapeutic approaches to induce platelet production. increased in a RepID-depleted cancer cell lines. Transcription level of DAB2 was examined by qRT-PCR in K562 or DMS114 cell lines in RepID WT or RepID-depleted condition. Average of transcription level relative to GAPDH was indicated, and error bars represent the standard deviation (**P-value < 0.01, ***P < 0.001). (E) DAB2 protein is highly expressed in RepID-de cient cancer cell lines. Immunoblot analysis with antibodies against DAB2 was performed and Histone H3 was used as loading control. Relative intensity of DAB2 protein was indicated as fold changes of RepID KO to RepID WT cells.    JARID1A loading on DAB2 promoter maintains its epigenetic con guration as heterochromatin status,