Wei Yang

DNA polymerase ζ (Polζ) is essential for replication across damaged DNA, yet the mechanisms governing its regulation in mammalian cells remain incompletely understood. The catalytic subunit REV3L is nearly twice the size of its yeast counterpart owing to a large vertebrate-specific insertion, and recent studies reported that REV3L undergoes proteolytic cleavage by the endopeptidase TASP1. Here, we show that REV3L cleavage generates two stable fragments, Nter-p70 and Cter-p300, which reassociate into a long-lived heterodimeric complex with enhanced polymerase activity. This interaction is stabilized by key residues within the catalytically inactive EXO domain. The cleaved Polζ complex associates with chromatin, binds nascent DNA, and undergoes ATR-dependent phosphorylation in response to replication stress. Importantly, preventing REV3L cleavage markedly impairs Polζ activity, reduces cisplatin-induced mutagenesis, and increases RAD51 nuclear foci and sister chromatid exchange, consistent with a compensatory shift toward homologous recombination-mediated repair. Together, our findings suggest that the vertebrate-specific expansion of REV3L necessitated cleavage to restore structural organization and optimize catalytic function, representing an evolutionary adaptation in DNA damage tolerance pathways.

Nucleotide excision repair (NER) removes bulky adducts from genomic DNA and prevents the ultraviolet light-sensitivity disease xeroderma pigmentosum, cancer and premature ageing¹. After initial lesion recognition by XPC in global genome repair or by stalled RNA polymerases in transcription-coupled repair, a lesion and surrounding DNA duplex are unwound by TFIIH, which includes the ATPases XPB and XPD, and additional NER factors XPA, XPF, XPG and RPA, to form a DNA bubble² comprising around 27 nucleotides. The double strand–single strand (ds-ss) junction-specific endonucleases XPF and XPG cleave DNA on the 5′ and 3′ sides of the lesion, respectively. Here we report the functional steps and atomic structures of the ATPase-driven and lesion-dependent DNA bubble formation and arrangement of the complete NER factors for dual incision. The unwinding of nearly 30 base pairs of DNA depends mainly on the double strand DNA translocase XPB and the duplex dividers XPA and XPF. XPD binds the lesion strand with XPF at the 5′ ds-ss junction. XPF cuts the lesion strand only after XPG binds the 3′ ds-ss junction. The ERCC1 subunit of XPF facilitates DNA strand separation and recruitment of RPA to the non-lesion strand. These findings provide insights on the causes of human diseases and potential targets for enhancing chemotherapeutic efficacy.

The leukemia fusion gene CBFB-MYH11 requires RUNX1 for leukemogenesis, but the underlying mechanism is unclear. By in vitro studies, we found that CBFβ-SMMHC, the chimeric protein encoded by CBFB-MYH11 , could enhance the binding affinity between RUNX1 and its target DNA. Increased RUNX1-DNA binding was also observed in myeloid progenitor cells from mice expressing CBFβ-SMMHC. Moreover, only CBFβ-SMMHC variants able to enhance the DNA binding affinity by RUNX1 could induce leukemia in mouse models. Marked transcriptomic changes, affecting genes associated with inflammatory response and target genes of CBFA2T3, were observed in mice expressing leukemogenic CBFβ-SMMHC variants. Finally, we show that CBFβ-SMMHC could not induce leukemia in mice with a Runx1 -R188Q mutation, which reduces RUNX1 DNA binding but does not affect its interaction with CBFβ-SMMHC or its sequestration to cytoplasm by CBFβ-SMMHC. Our data suggest that, in addition to binding RUNX1 to regulate gene expression, enhancing RUNX1 binding affinity to its target DNA is an important mechanism by which CBFβ-SMMHC contributes to leukemogenesis, highlighting RUNX1-DNA interaction as a potential therapeutic target in inv(16) acute myeloid leukemia.

The recombination activating genes 1 and 2 (RAG1/2) recombinase, which initiates V(D)J recombination in jawed vertebrates, evolved from RNaseH-like transposases such as Transib and ProtoRAG. However, its postcleavage transposase activity is strictly suppressed. Previous structural studies have focused only on the conserved core domains of RAG1/2, leaving the regulatory mechanisms of the noncore regions unclear. To investigate how RAG1/2 suppresses transposition and regulates DNA cleavage, we determined cryo-electron microscopy (cryo-EM) structures of nearly full-length RAG1/2 complexed with cleaved recombination signal sequences (RSS) in a signal-end complex (SEC) at resolutions up to 2.95 Å. Two key structures, SEC-0 and SEC-Plant Homeodomain (PHD), reveal distinct regulatory roles of RAG2, which is absent in Transib transposase. SEC-0 displays a closed conformation, revealing that the core RAG2 facilitates sequential DNA cleavage by stabilizing the RSS-cleaved states in a “spring-loaded” mechanism. SEC-PHD reveals how RAG2’s noncore PHD and Acidic Hinge (AH), which are absent in ProtoRAG, inhibit target DNA binding in transposition. Histone H3K4me3, which recruits RAG1/2 to RSS sites, does not influence RAG1/2 binding to V, D, or J gene segments bordered by RSS. In contrast, the suppressed transposition can be activated by H3K4me3 peptides that dislodge the inhibitory PHD. To achieve this derepression in vivo, however, would require an unlikely close placement of two nucleosomes flanking a target DNA bent by nearly 180°. Our structural and biochemical results elucidate how RAG1 has acquired RAG2 and utilizes its core and noncore domains to enhance V(D)J recombination and suppress transposition.

Non-homologous end joining (NHEJ) is the main repair pathway of double-strand DNA breaks in higher eukaryotes^1,2. Here we report reconstitution of the final steps of NHEJ and structures of DNA polymerase μ and ligase IV (LIG4) engaged in gap filling and end joining. These reactions take place in a flexible ω-shaped framework composed of XRCC4 and XLF. Two broken DNA ends, each encircled by Ku70–Ku80 internally, are docked onto the ω frame, mediated by LIG4. DNA polymerase and ligase attached to each ω arm repair only one broken strand of a defined polarity; the final steps of NHEJ requires coordination and toggling of a pair of such enzymes. The facilitators XLF and PAXX additively stimulate NHEJ reactions. As DNA-end sensor and protector, LIG4 replaces DNA-PKcs for end joining and bridges the two DNA ends for polymerase to fill remaining gaps. These assemblies present new targets for NHEJ inhibition to enhance efficacy of radiotherapy and accuracy of gene editing.

The RAG1/2 recombinase, which initiates V(D)J recombination in jawed vertebrates, evolved from RNaseH-like transposases such as Transib and ProtoRAG ¹. However, its post-cleavage transposase activity is strictly suppressed. Previous structural studies have focused only on the conserved core domains of RAG1/2, leaving the regulatory mechanisms of the non-core regions unclear. To investigate how RAG1/2 suppresses transposition and regulates DNA cleavage, we determined cryo-EM structures of nearly full-length RAG1/2 complexed with cleaved Recombination Signal Sequences (RSS) in a Signal-End Complex (SEC), at resolutions up to 2.95 Å. Two key structures, SEC-0 and SEC-PHD, reveal distinct regulatory roles of RAG2, which is absent in Transib transposase. SEC-0 displays a closed conformation, revealing that the core RAG2 facilitates sequential DNA cleavage by stabilizing the RSS-cleaved states in a "spring-loaded" mechanism. SEC-PHD reveals how RAG2’s non-core PHD and Acidic Hinge (AH) domains, which are absent in ProtoRAG, inhibit target DNA binding in transposition. Histone H3K4me3, which recruits RAG1/2 to RSS sites, does not influence RAG1/2 binding to V, D or J gene segments bordered by RSS ². In contrast, the suppressed transposition can be activated by H3K4me3 peptides that dislodge the inhibitory PHD domain ^3,4. To achieve this de-repression in vivo, however, would require an unlikely close placement of two nucleosomes flanking a target DNA bent by nearly 180°. Our structural and biochemical results elucidate how RAG1 has acquired RAG2 and utilizes its core and non-core domains to enhance V(D)J recombination and suppress transposition.

Type IA topoisomerases (TopoIAs) are present in all living organisms. They resolve DNA/RNA catenanes, knots and supercoils by breaking and rejoining single-stranded DNA/RNA segments and allowing the passage of another nucleic acid segment through the break. Topoisomerase III-β (TOP3B), the only RNA topoisomerase in metazoans, promotes R-loop disassembly and translation of mRNAs. Defects in TOP3B lead to severe neurological diseases. We present a series of cryo-EM structures of human TOP3B with its cofactor TDRD3 during cleavage and rejoining of DNA or RNA, thus elucidating the roles of divalent metal ions and key enzyme residues in each step of the catalytic cycle. We also obtained the structure of an open-gate configuration that addresses the long-standing question of the strand-passage mechanism. Our studies reveal how TOP3B catalyzes both DNA and RNA relaxation, while TOP3A acts only on DNA.

DNA repair protein MutSβ promotes CAG•CTG repeat expansions, which cause ∼20 untreatable neurodegenerative disorders including Huntington’s disease (HD). However, how MutSβ is recruited onto chromatin to enable repeat expansion is unknown. Here, we show that H3K56 interacts with MutSβ and brings it to chromatin. The H3K56-MutSβ interaction is regulated by H3K56 acetylation (H3K56ac) conditions, with acetylation inhibiting but deacetylation facilitating MutSβ chromatin recruitment. Blocking the H3K56-MutSβ interaction by either disrupting the MutSβ aromatic packet or increasing the H3K56ac level by removing histone deacetylases stabilizes CAG repeats, while depleting histone acetyltransferase p300/CBP promotes the H3K56-MutSβ interaction and CAG•CTG repeat instability, including expansion of CAG repeats of the huntingtin gene. Therefore, our study suggests a novel mechanism, by which MutSβ induces triplet repeat expansion, and new strategies of blocking the H3K56-MutSβ interaction to prevent HD and other triplet repeat expansion-caused disorders.

Summary: Expansion of trinucleotide repeats causes Huntington’s disease, Fragile X syndrome and over twenty other monogenic disorders ¹ . How mismatch repair protein MutSβ and large repeats of CNG (N=A, T, C or G) cooperate to drive the expansion is poorly understood. Contrary to expectations, we find that MutSβ prefers to bind the stem of an extruded (CNG) hairpin rather than the hairpin end or hairpin-duplex junction. Structural analyses reveal that in the presence of MutSβ, CNG repeats with N:N mismatches adopt a B form-like pseudo-duplex, with one or two CNG repeats slipped out forming uneven bubbles that partly mimic insertion-deletion loops of mismatched DNA ² . When the extruded hairpin exceeds 40-45 repeats, it can be bound by three or more MutSβ molecules, which are resistant to ATP-dependent dissociation. We envision that such MutSβ-CNG complexes recruit MutLγ endonuclease to nick DNA and initiate the repeat expansion process ^3,4 . To develop drugs against the expansion diseases, we have identified lead compounds that prevent MutSβ binding to CNG repeats but not to mismatched DNA.

Nucleotide excision repair removes DNA lesions caused by ultraviolet light, cisplatin-like compounds and bulky adducts¹. After initial recognition by XPC in global genome repair or a stalled RNA polymerase in transcription-coupled repair, damaged DNA is transferred to the seven-subunit TFIIH core complex (Core7) for verification and dual incisions by the XPF and XPG nucleases². Structures capturing lesion recognition by the yeast XPC homologue Rad4 and TFIIH in transcription initiation or DNA repair have been separately reported^3,4,5,6,7. How two different lesion recognition pathways converge and how the XPB and XPD helicases of Core7 move the DNA lesion for verification are unclear. Here we report on structures revealing DNA lesion recognition by human XPC and DNA lesion hand-off from XPC to Core7 and XPA. XPA, which binds between XPB and XPD, kinks the DNA duplex and shifts XPC and the DNA lesion by nearly a helical turn relative to Core7. The DNA lesion is thus positioned outside of Core7, as would occur with RNA polymerase. XPB and XPD, which track the lesion-containing strand but translocate DNA in opposite directions, push and pull the lesion-containing strand into XPD for verification.