Molecular Cell

Promoter-proximal RNA Polymerase II (RNAPII) pausing and the processivity are controlled by distinct modules of the Integrator complex, which together fine-tune transcription and protect against the accumulation of defective RNAPII complexes. Compromised activity of individual Integrator modules has been linked to human disease including cancer and developmental disorders, caused by defective transcription of protein-coding or small-nuclear RNAs. Despite extensive characterisation of the Integrator complex both genetically and structurally, the role of smallest member of the complex, INTS12, has remained enigmatic. Here, we uncover that INTS12 loss acts to stabilise the association between NELF and Integrator via its PHD domain and N-terminus, respectively, thus safeguarding against the release of defective RNAPII complexes. Acute degradation of INTS12 results in the selective dissociation of Integrator from the NELF-RNAPII complex which subsequently convert to their canonical paused form from which they can be released by CDK9. In the absence of INTS12 excess release of defective RNAPII via P-TEFb/SEC, loss of the ARMC5 salvage pathway and deletion of the catalytic and core Integrator subunits is toxic to cells. These findings demonstrate that there is interconversion between canonical paused RNAPII and paused-Integrator, and highlight the critical interplay between these processes and P-TEFb mediated pause-release to ensure that only transcription competent complexes are released into elongation. O_LIINTS12 degradation confers CDK9 inhibitor resistance and triggers cellular stress through a phosphatase module-independent mechanism. C_LIO_LIINTS12 stabilizes the Integrator-NELF complex through its N-terminus and PHD domain. C_LIO_LIAcute INTS12 degradation promotes aberrant release of promoter-proximal RNA polymerase II complexes. C_LIO_LIRNA polymerase II complexes released upon INTS12 loss exhibit defective elongation and reduced processivity. C_LIO_LIINTS12 loss removes Integrator from RNAPII resulting in aberrant paused-state from which it can be released by CDK9. C_LIO_LIExcess CDK9 activity and ARMC5 loss are synthetically lethal with INTS12 deficiency. C_LI

Hydrogen Peroxide (H2O2) stress activates transcription factors (TFs) in a dose-dependent manner, with distinct TFs activated in response to low versus high H2O2. Here, we show that high H2O2 imposes a translational constraint that prevents accumulation of TFs requiring de novo protein synthesis. Under low H2O2 conditions, TFs including p53, NRF2, and ATF4, accumulate and drive stress-responsive gene expression. In contrast, high H2O2 induces coordinated inhibition of translation initiation and elongation through activation of the integrated stress response (ISR), suppression of mTORC1 signaling, and activation of eEF2K, thereby blocking accumulation of these TFs. Inhibition of translation and repression of p53, NRF2, and ATF4 coincides with nuclear shuttling of pre-existing TFs, including FOXO1, NFAT1, and NF-{kappa}B. We propose that shuttling TFs provide a backup mechanism to respond to severe oxidative stress while translation is inhibited. Together, these findings identify translational control as a central switch governing transcription factor response to H2O2 stress.

Post-transcriptional regulation is critical for mammalian embryogenesis yet has been underexplored. We previously showed that RNA-binding Pumilio proteins (Pum1/2) are essential for early mouse embryogenesis and embryonic stem cell (ESC) functions. Here, using acute protein degradation systems combined with time-resolved RNA-seq and eCLIP, we delineate a two-phase regulatory hierarchy modulated by Pum1/2 in mouse ESCs. The first phase, occurring within 10 hours of Pum1/2 depletion, is predominantly the stabilization of over 100 Pum1/2-target mRNAs, while the second phase, occurring in subsequent 66 hours, propagates to over 1,000 mRNAs mostly through indirect regulatory effects. Functionally, Pum1/2 depletion delays transition from naive to formative pluripotency, impairs neuroectoderm differentiation, and enhances germline specification. Mechanistically, Pum1/2 directly repress mRNAs encoding PRC2 subunits, including Suz12, thereby constraining H3K27me3 deposition at neuroectodermal gene loci. These findings establish Pum1/2 as biphasic post-transcriptional regulators of pluripotency and lineage balance and link RNA stability control to chromatin-mediated silencing. HIGHLIGHTSO_LIAcute Pum1/2 degradation provides temporal resolution for profiling post-transcriptional regulation. C_LIO_LIPum1/2 destabilize more than 100 direct targets and modulate a biphasic network of over 1,000 mRNAs. C_LIO_LIPum1/2 loss delays pluripotency transition, suppresses neuroectoderm, and promotes germline fate. C_LIO_LIPum1/2 directly regulate Suz12 mRNA decay to modulate PRC2-mediated repression. C_LI

The RNA exosome is known to participate in transcription, but the contribution of its ribonuclease activities to this process remains unclear. Here we investigated the role of DIS3, one of the exosome ribonucleases, in transcription by RNA polymerase II (RNAPII). Rapid depletion of DIS3 reduced RNA synthesis and induced RNAPII elongation defects that were exacerbated by UV irradiation, a treatment that generates transcription-blocking DNA lesions and promotes RNAPII backtracking. Notably, DIS3 itself was redistributed following UV irradiation in a manner that closely paralleled RNAPII dynamics, which suggested that DIS3 acts in concert with the transcription machinery. We also investigated whether RNA degradation by DIS3 was required for transcription elongation and found that the 3-5 exoribonucleolytic activity of DIS3, but not its endonucleolytic activity, is essential for efficient transcription elongation. More specifically, DIS3 degrades the 3 ends of backtracked RNA, as shown by sequencing of RNA fragments released by TFIIS-induced transcript cleavage in vitro. This DIS3-dependent degradation of backtracked RNA is critical for resolving stalled RNAPII complexes and enabling productive transcription elongation. Graphical abstract O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=169 SRC="FIGDIR/small/703186v1_ufig1.gif" ALT="Figure 1"> View larger version (27K): org.highwire.dtl.DTLVardef@c080dorg.highwire.dtl.DTLVardef@1e4b375org.highwire.dtl.DTLVardef@1c1f575org.highwire.dtl.DTLVardef@d9d640_HPS_FORMAT_FIGEXP M_FIG C_FIG HighlightsO_LIDIS3 depletion leads to severely reduced rates of RNA synthesis in human cells. C_LIO_LITranscription-blocking lesions reveal a critical role for DIS3 in RNA polymerase II elongation. C_LIO_LIThe 3 to 5 exoribonuclease activity of DIS3 is necessary for transcription elongation. C_LIO_LIDIS3-mediated degradation of backtracked RNA is required for the resolution of stalled transcription complexes. C_LI

Transcription factors (TFs) bind to enhancers and recruit H3K4me1 methyltransferase KMT2D, chromatin remodeler BAF, and H3K27 acetyltransferase p300 to activate transcription. However, the role of chromatin modifiers in regulating de novo binding of TFs on enhancers remains unclear. Using a robust nuclear translocation system, we show that the muscle lineage-determining TF MyoD binds to chromatin pervasively within one hour, with half of induced MyoD binding sites co-occupied by KMT2D, BAF, and p300. On the majority of these MyoD+ enhancers, acute depletion of KMT2D or short-term inhibition of BAF or p300 enzymatic activity markedly reduces de novo binding of MyoD as well as that of KMT2D, BAF, and p300. On enhancers with intact MyoD binding despite these perturbations, we observe a cooperative recruitment among chromatin modifiers. Similar interdependent relationships are observed between the signal-dependent TF Glucocorticoid Receptor and KMT2D, BAF, and p300. Together, our findings show that chromatin modifiers are not only downstream effectors but also required for de novo binding of TFs on enhancers, refining a model of enhancer establishment as a process governed by functional cooperation rather than a strict hierarchy. Bullet pointsO_LIAcute KMT2D depletion disrupts de novo binding of MyoD, BAF, and p300 on enhancers. C_LIO_LIBAF and p300 enzymatic activities are required for de novo binding of MyoD, BAF, KMT2D, and p300 on enhancers. C_LIO_LICooperative binding of KMT2D, BAF, and p300 on MyoD+ enhancers. C_LIO_LIGR displays interdependencies with chromatin modifiers KMT2D, BAF, and p300 on enhancers. C_LI

Regulation of protein synthesis is essential for maintaining cellular homeostasis during stress. The integrated stress response (ISR) is a conserved signaling pathway that modulates global mRNA translation through four eIF2 kinases--GCN2, PKR, PERK, and HRI. However, how these kinases are selectively activated and tuned to distinct stress signals to direct appropriate cell fate decisions remains poorly understood. Here, we employ ultra-deep mutagenesis screens to systematically map regulators of protein synthesis across diverse stress perturbations in human cells. This comparative approach identifies stress-specific translational control factors, including a previously unrecognized role for the E3 ubiquitin ligase RNF25 in selectively sustaining translation following UV irradiation and other RNA-damaging treatments. In this context, we demonstrate that RNF25 operates independently of its partner RNF14, and that its ubiquitin ligase activity, as well as its RWD-domain, is required to restrain excessive activation of the eIF2 kinase GCN2. Accordingly, loss of RNF25 results in hyperactivation of GCN2, exacerbated translation shutdown, and impaired cell proliferation following RNA damage--phenotypes that can be fully reversed by genetic or pharmacological inhibition of GCN2. Together, these findings uncover a previously unappreciated RNF25-GCN2 signaling axis and identify ISR-driven toxicity as a potential vulnerability in combination with RNA-damaging chemotherapeutics.

The DNA damage response (DDR) is a complex network of interconnected pathways and sub-pathways that safeguards genome integrity. Deciphering the coordinated and complementary interactions among these pathways remains a major challenge. In this study, we employed CRISPR screening to systematically map the genetic interactions required for different sub-pathways of homologous recombination in human cells following PARP inhibitor treatment. Our approach recapitulated known interactions and uncovered several previously unrecognized connections. We identified RAD54L, in addition to ATRX, as a factor promoting the double Holliday junction (dHJ) pathway and demonstrated that RAD51AP1 and RAD54B function in synthesis-dependent strand annealing (SDSA). We provide evidence that loss of TOP3A induces a switch in HR sub-pathway usage from SDSA to the dHJ pathway. Furthermore, TOP3A deficiency abolishes the requirement for ATRX and the histone variant H3.3 in the dHJ pathway, while maintaining strict dependence on RAD54L. We further observed that H3.3 is involved in both HR sub-pathways, whereas its depositing chromatin remodelers HIRA and ATRX play pathway-specific roles in SDSA and dHJ, respectively. Together, our findings define the architecture underlying HR sub-pathway choice and reveal a key role for TOP3A in regulating pathway balance.

PDS5A, a regulatory subunit of the cohesin complex, and topoisomerase IIB (TOP2B), an enzyme resolving DNA topological problems, interact with CTCF and regulate transcription, chromatin loops, and genome organization. Yet, how PDS5A and TOP2B are recruited to chromatin to exert their function is not well-understood. Here, we studied the functional relationship between PDS5A and TOP2B and the resultant impact on genome organization and gene expression. Interestingly, TOP2B-PDS5A cooperate for their recruitment to CTCF-bound chromatin sites. The presence of catalytically active TOP2B increased PDS5A occupancy genome-wide. Notably, a novel PDS5A-CTCF interaction region in the CTCF N-terminal 95-116aa was required for CTCF-PDS5A-TOP2B interaction in vitro as well as for active TOP2B-mediated enrichment of PDS5A chromatin occupancy in vivo. The loss of CTCF(95-116aa) led to a reduced number of chromatin loops and dysregulated gene expression. In gliomas, PDS5A and TOP2B expression levels are variable and correlated, contributing to apparent heterogeneity in gene expression. Indeed, inducible knockdown of PDS5A led to reduced TOP2B occupancy and altered gene expression in the glioma genome. Importantly, PDS5A mediated sensitivity to TOP2 cancer drugs in glioma cells. This newly recognized functional interaction between PDS5A and TOP2B at chromatin boundaries clarifies the mechanisms fostering gene regulation through genome organization, with implications for glioma therapeutics.

DNA replication initiation requires activation of the CMG helicase to establish the replisome. This process involves the extrusion of single-stranded DNA (ssDNA) from the central channel of MCM double hexamers, allowing the two CMG helicases to pass each other; however, the factors that mediate this process in human cells remain unclear. We show that degron-mediated depletion of either MCM10 or RECQL4 alone causes only mild replication defects, whereas simultaneous depletion of both proteins completely blocks CMG activation. ChIP-seq analyses demonstrate that RECQL4 localises to replication initiation zones (IZs) independently of MCM10, whereas MCM10 recruitment to IZs is enhanced upon RECQL4 depletion, suggesting RECQL4 primarily functions in CMG activation, and MCM10 acts as a backup or supporting factor. Rescue experiments further indicate that RECQL4 cooperates with MCM10 through direct interaction, and that their ssDNA-binding activity underlies their functional overlap. We propose MCM10 and RECQL4 act cooperatively and redundantly to promote CMG activation.

Eggs of many species accumulate thousands of dormant mRNAs that are translated after fertilization at specific times and locations to direct development. However, how embryos coordinate translation of these mRNAs remains unclear. In this study, we identified sequential waves of translation critical for proper development progression. The first wave occurred within 1 h and included translation of ewsr1b mRNA that harbored a short 3' untranslated region (UTR) comprising 16 nucleotides. The resulting Ewsr1b protein triggered the second translation wave through binding cytoplasmic mRNAs, including pou5f3, which encodes a transcription factor promoting zygotic genome activation. In contrast, HuR and Syncrip repressed translation until the first and second waves, respectively. ewsr1b mRNA that had a long 3'UTR was translated in the second wave, and the 3'UTRs length determined protein localization and function. Overall, our findings reveal previously unknown molecular principles that coordinate translation timings and protein functions to drive long-term, multilayered processes.

DNA inter-strand crosslinks (ICLs) are highly cytotoxic lesions that require coordinated processing to faithfully repair. Members of the RecQ4 helicase family, such as Saccharomyces cerevisiae Hrq1, are implicated in ICL repair yet remain poorly understood mechanistically. Hrq1 promotes ICL repair by stimulating the nuclease Pso2, but cells expressing a helicase-dead hrq1-K318A allele are more sensitive to ICL damage than hrq1{Delta} cells, indicating a dominant-negative defect. To define the basis of this toxicity, we used unbiased genetic suppressor screening, structural modeling, and biochemical analyses. Spontaneous suppressors of hrq1-K318A sensitivity were overwhelmingly intragenic second-site mutations in hrq1that either destabilized Hrq1-K318A or disrupted its ability to bind DNA. In all cases, these mutations relieved the dominant-negative repair defect. Biochemical characterization of representative mutants, including a rationally designed DNA-binding mutant, demonstrated that loss of DNA binding abolishes helicase activity and suppresses Hrq1-K318A toxicity despite retaining protein stability. Together, these findings identify persistent DNA binding by a catalytically inactive RecQ4-family helicase as the central driver of dominant-negative ICL repair defects and provide mechanistic insight into how incomplete loss-of-function alleles of human RECQL4 may actively interfere with genome maintenance pathways. GRAPHICAL ABSTRACT O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=81 SRC="FIGDIR/small/705743v1_ufig1.gif" ALT="Figure 1"> View larger version (15K): org.highwire.dtl.DTLVardef@4d93bcorg.highwire.dtl.DTLVardef@11e48eeorg.highwire.dtl.DTLVardef@14628cforg.highwire.dtl.DTLVardef@1216f50_HPS_FORMAT_FIGEXP M_FIG C_FIG

Cellular stressors often cause widespread repression of RNA polymerase II (RNAP II) activity, which is thought to facilitate a focused transcriptional output towards stress resolution. In many cases, however, the underlying regulatory mechanisms remain unknown. Here, we demonstrate that stress-induced downregulation of the general transcription factor TFIIB tempers expression of specific stimulus response genes. Following a variety of stressors, TFIIB is proteolytically cleaved between its cyclin folds at conserved aspartic acid residue D207 by caspases- 3 and 7. Cleavage in this portion of the protein significantly reduces the ability of TFIIB to form a TBP-TFIIB-DNA promoter complex in vitro. Using both overexpression and endogenous base-editing, we find that B and T cells that are unable to cleave TFIIB upregulate expression of a select gene set during apoptosis. These TFIIB-sensitive genes are primarily short, stimulus-responsive and proto-oncogenic loci, and cleavage of TFIIB temporally restricts their expression. Failure to cleave TFIIB during stress leads to aberrant lymphocyte proliferation during chemical perturbation. Hence, caspase targeting of TFIIB destabilizes transcription to tune gene expression, allowing for proper stress resolution.

The cytosolic iron-sulfur cluster assembly (CIA) pathway maturates essential nuclear and cytosolic Fe-S proteins required for genome maintenance and cellular metabolism. Nar1 (also called CIAO3 or IOP1) is a conserved Fe-S protein that connects the early and late steps of the CIA pathway, yet the molecular basis for its proposed function as a metallocluster carrier remains poorly defined. In particular, the interactions responsible for Nar1 recruitment to the CIA targeting complex (CTC) during cluster delivery remain unknown. Here, we define the molecular basis for Nar1 recruitment to the CTC using biochemical reconstitution, quantitative protein-protein interaction assays, and AlphaFold modeling. Our data reveal that Nar1 binds the CTC through two distinct interfaces. A primary interface comprises an electrostatic interaction that anchors Nar1 to a conserved acidic surface on the Cia1 subunit of the CTC and a secondary interface involves binding of Nar1s divergent targeting complex recognition peptide at the Cia1-Cia2 interface. Thus, Nar1 engages a conserved CTC surface that serves as a recruitment platform for multiple binding partners, including CIA clients. Computational structural models position the putative Fe-S cluster donor site of Nar1 adjacent to a proposed acceptor site on Cia2, suggesting that this bipartite binding mechanism positions Nar1 for transfer of an Fe-S cluster to the targeting complex. Together, these findings resolve conflicting models for Nar1 recruitment and establish a mechanistic framework for understanding how the CTC engages multiple binding partners during cytosolic iron-sulfur protein maturation.

During development, cellular identity is ultimately determined by transcriptional output: lineage-specific genes must be activated, while genes associated with alternative fates must be repressed. This process depends on the activity of chromatin remodelling complexes, which regulate the accessibility of transcription factors to chromatin regulatory elements. In addition, cellular identity is shaped by exposure to intercellular signals. Understanding the mechanisms by which extracellular signals are translated into changes in the transcriptional program is essential for understanding cell fate decisions during development, as well as in disease conditions such as cancer. Here we describe a rapid and widespread enhancer resetting event in response to ERK signalling in mouse ES cells. This process occurs in two distinct phases: an immediate, genome-wide alteration in transcription factor binding dynamics at regulatory regions which is dependent on the release of paused RNA Polymerase II, followed by the re-establishment of a context appropriate, stable chromatin state. We demonstrate that the chromatin remodelling complex NuRD is required for this reestablishment phase and for the appropriate transcriptional response to ERK signalling. We propose that enhancer resetting places genomic regulatory regions in a state which is permissive to the exchange of transcription factors in order to establish a new, stable enhancer topology enabling rapid yet precise transcriptional response to extracellular signals.

Chromatin is organized into topologically associating domains (TADs) that are critical for gene regulation. Transitions between cell states, such as proliferation and quiescence, involve genome-scale chromatin dynamics and transcriptional changes, but the underlying mechanisms are not well-understood. Here we show that the reversible shift between proliferation and quiescence is regulated by the opposing action of CTCF and the histone modification H4K20me3. CTCF is a boundary element that defines TADs, while H4K20me3 has previously been associated with heterochromatin, chromatin compaction, and repetitive elements. Using CUT&Tag, RNA-seq, and functional perturbations, we demonstrate that H4K20me3 and CTCF compete for chromatin, including at TAD boundary elements, with increased H4K20me3 in quiescent cells antagonizing CTCF activity in proliferating cells. Manipulation of H4K20me3 levels shows that in quiescent cells, elevated H4K20me3 alters the chromatin landscape, leading to more compact chromatin architecture, elongated nuclei, and induction of a quiescence gene expression program. Conversely, CTCF binding is associated with open chromatin conformation, rounder nuclei, expression of proliferation-associated genes, lower levels of quiescence marker p27/CDKN1B, and cell division in cultured cells and mice. Fibroblasts deficient for H4K20me3 methyltransferase KMT5C are more proliferative, with reduced expression of quiescence genes, and KMT5C-deficient mice are larger. Our findings reveal a reversible antagonistic interplay between H4K20me3 and CTCF that modulates the functional outcome of chromatin architecture to effect changes in cellular state. This new paradigm for regulation of the proliferation-quiescence transition suggests a molecular basis for common developmental transitions and disorders of proliferation. HighlightsO_LIH4K20me3, a mark associated with heterochromatin, and CTCF compete for chromatin boundary elements C_LIO_LICTCF and H4K20me3 play opposing roles, with increased CTCF in proliferating cells and elevated H4K20me3 in quiescent cells C_LIO_LIModulating H4K20me3 or CTCF reversibly shifts cells between proliferative and quiescent state C_LIO_LIMice with reduced H4K20me3 are larger, consistent with hyperproliferation C_LI

Multiple septin family proteins co-assemble with strict subunit stoichiometry into hetero-oligomers. In the absence of native septin partners, purified septins aggregate in vitro, and "orphan" septins are found in pathological aggregates associated with neurodegenerative diseases. Cytosolic chaperones bind the septin GTPase domain to promote on-pathway septin folding but it was unclear how cells manage orphan septins to maintain septin subunit stoichiometry. Most septins have C-terminal domains (CTDs) that form heteromeric coiled coils within or between septin complexes. Here we present evidence that orphan yeast septins are protected from proteasomal degradation by forming transient coiled-coil homodimers and trimers and, in parallel, by the disaggregase chaperone Hsp104. Septins unable to undergo CTD-mediated homo-oligomerization require Hsp104 to accumulate to super-stoichiometric levels. We show that the number of septin-encoding mRNAs per yeast cell is low and variable, creating opportunities for transient subunit imbalances. These findings reveal a novel role for coiled coils and the cellular proteostasis machinery in the fidelity of higher-order septin assembly.

Mitosis is triggered when the rising activity of CDK1-Cyclin B, amplified by the CDK1/Cdc25/Wee1 feedback loop, overcomes inhibitory signalling from Wee1 and counteracting phosphatases. CDK-opposing phosphatases PP1, PP2A-B55 and PP2A-B56 are regulators of mitosis. A screen for differentially phosphorylated sites in a{Delta} PP1dis2 genetic background in Schizosaccharomyces pombe identified phosphorylation of T73 or T75 in the regulatory B56Par1 subunit. The B56Par1.T73T75 phosphorylation is directly mediated by CDK1-Cyclin B, and a phospho-mimetic mutation increased PP2A-B56Par1 phosphatase activity. Blocking B56Par1.T73T75 phosphorylation reduced cell length in unperturbed divisions from 14 to 12 {micro}m, with no other detectable phenotypes. Therefore, blocking phosphorylation at T73T75 alone prematurely unlocked amplification of the CDK1/Cdc25/Wee1 feedback loop, advancing cells into mitosis. Signalling from T73T75 reveals for the first time that timely mitotic commitment in unperturbed cycles is mediated by PP2A-B56.

Neural stem/progenitor cells (NSPCs) in the adult mouse brain reside in distinct niches, including the dentate gyrus (DG) and subventricular zone (SVZ). The contribution of cell intrinsic versus extrinsic factors to distinct fates of NSPCs and their neuronal progeny remains largely unknown. We here show that DG- and SVZ-derived NSPCs retain niche-specific chromatin accessibility states that predict neuronal subtype specification. Furthermore, metabolic profiling and gene expression analyses comparing DG- and SVZ-derived NSPCs revealed differences in carnitine synthesis pathways and lipid metabolism. Supplementation with carnitine or S-adenosylmethionine (SAM) induced chromatin remodeling via histone modifications, indicating that a metabolic-epigenetic cross-talk determines regional neuronal identity. Our findings show that NSPCs are intrinsically programmed by chromatin and metabolic states, and that metabolic interventions alter epigenetic landscapes and fate potential.

TDP-43 is a nuclear RNA-binding protein regulating numerous steps in RNA metabolism, including alternative splicing. It is a major pathological hallmark of several neurodegenerative diseases, where it forms cytoplasmic aggregates in affected brain regions. TDP-43 can undergo phase separation (PS) and this condensation behavior may be linked to aggregate formation. Whether and how PS governs TDP-43s RNA regulatory functions remains poorly understood. Here we utilized rationally designed mutations in the TDP-43 low complexity domain to tune TDP-43 PS, yielding a panel of TDP-43 variants with reduced propensity to form condensates ("PS-deficient"), and a panel of TDP-43 variants that form more irreversible, undynamic condensates ("solid-like") in vitro and in cells. Affinity proteomics coupled to mass spectrometry identified a set of interactors whose association with TDP-43 is PS-dependent. This includes multiple splicing regulators and the RNA helicase UPF1, which show increased interactions with solid-like TDP-43 variants. Consistently, we identified PS-dependent alternative splicing events that translate into measurable changes in RNA and protein abundance. Our results highlight that TDP-43 PS regulates RNA and protein homeostasis both directly, by altering a subset of TDP-43-dependent alternative splicing events, and indirectly, by changing interactions with other RNA regulatory factors.

The conserved multifunctional chromatin modulator and oncogene DEK exhibits context-dependent genomic binding and function, but how these activities are regulated in cancer remains poorly understood. Using multi-omics and biochemical approaches, we find that while DEK predominantly occupies promoter-proximal regions in HeLa cells and primary melanocytes, its chromatin binding is dramatically reduced in melanoma cell lines--despite DEK overexpression. We attributed this to CK2-mediated phosphorylation, which governs DEK chromatin association and transcriptional output in a cell-type-specific manner. Phosphoproteomics identified 34 phosphorylation sites, including S287 and S288 within the DEK C-terminal DNA-binding domain. Strikingly, CK2 inhibition and concomitant loss of phosphorylation at S287/S288 triggered DEK redistribution to promoter regions, coinciding with transcriptional repression of oncogenic pathways and global chromatin compaction. Melanoma subtypes showed divergent responses: NRAS-mutant cells displayed dynamic, phosphorylation-dependent DEK redistribution, whereas BRAF-mutant cells lacked detectable DEK binding. Our work establishes DEK as a phosphorylation-sensitive regulator of chromatin states, with CK2-mediated modification orchestrating its tumor-specific regulatory functions. These findings nominate phospho-DEK as a potential biomarker and therapeutic target in melanoma and possibly other cancers.