Interstrand crosslinks, or ICLs, are among the most cytotoxic forms of DNA damage because they covalently tether the two strands of the double helix to one another, blocking both DNA replication and transcription. Cisplatin, a platinum-based drug, induces these crosslinks as part of its antitumor arsenal, and the resulting suppression of transcription contributes significantly to the drug’s ability to kill rapidly dividing cancer cells. However, cells possess a dedicated pathway known as transcription-coupled repair that can remove such lesions from actively transcribed genes, and this repair capacity can mitigate cisplatin’s effects, contributing to treatment resistance in some tumors. Understanding exactly how the crosslink stalls the polymerase, and how that stall is recognized by the repair machinery, is therefore of direct clinical relevance.

The Zhejiang team assembled defined RNA polymerase II elongation complexes on DNA templates containing a site-specific cisplatin-induced interstrand crosslink, positioning the lesion at various distances from the enzyme’s active site. By vitrifying these complexes and imaging them with cryo-electron microscopy, the researchers were able to reconstruct a molecular movie of sorts, following the polymerase as it advanced toward the crosslink and ultimately ground to a halt. The eight captured states span the journey from the initial approach of the lesion through the final stalled configuration, offering a step-by-step structural narrative that has never before been available for this class of DNA damage.

The most striking discovery emerged when the polymerase reached the crosslink itself. Rather than simply blocking progress like a physical roadblock, the crosslinked guanine bases underwent an unexpected conformational transformation. The enzyme actively remodeled the crosslink, prying the damaged guanines into a novel configuration positioned above the bridge helix, a structural element of the polymerase that plays a central role in translocating DNA and RNA through the enzyme. This rearrangement had not been observed for any other type of transcription-blocking lesion and reveals RNA polymerase II in an unanticipated role: that of a structural remodeler of the very damage it encounters.

The consequence of this remodeling is functionally decisive. By relocating the crosslinked base above the bridge helix, the polymerase prevents the damaged guanine from entering the active site, the catalytic chamber where templated nucleotide addition normally occurs. With the lesion physically excluded from the position where it would need to reside to serve as a template, transcription cannot proceed in the conventional manner. The structural data show that the remodeled crosslink creates an empty nucleotide-binding site poised in an abnormal configuration, setting the stage for a secondary failure mode that the researchers characterized in detail.

That secondary failure is ATP misincorporation followed by backtracking. Because the active site is left in an aberrant state, the polymerase can misinsert adenosine triphosphate opposite the position where the crosslinked guanine should have been read. Following this misincorporation event, the enzyme undergoes backtracking, a process in which the polymerase slides backward along the DNA and RNA, extruding the freshly synthesized 3′ end of the RNA transcript from the active site. Backtracking is a well-known mechanism of transcriptional arrest, and the new structures demonstrate how the remodeled crosslink actively drives the enzyme into this arrested configuration rather than merely obstructing forward motion.

The study also examined oxaliplatin, another clinically important platinum chemotherapy drug, and found that its interstrand crosslinks stall polymerase II through a conserved mechanism. Notably, the structural analyses revealed that when the platinum crosslink sits several nucleotides downstream of the active site, the polymerase drives the platinum adduct from the minor groove of the DNA helix into the major groove, and the crosslinked guanine on the non-template strand migrates toward the upstream side of the complex as the lesion approaches the catalytic center. These observations, together with experiments using alternative DNA sequences, demonstrated that the remodeling behavior is not an artifact of a particular template but a general property of how polymerase II handles platinum crosslinks.

A central question in transcription-coupled repair concerns what happens after the polymerase stalls. Cockayne syndrome protein B, or CSB, is an ATP-dependent translocase that is recruited to stalled polymerase II and is considered the initiating factor of transcription-coupled repair. One might expect CSB to either help the polymerase bypass the lesion or actively displace the stalled enzyme to clear the way for repair enzymes. The new data show that CSB does neither. In the reconstituted system, CSB neither promoted bypass of the cisplatin crosslink nor displaced the stalled polymerase, and the transcription factor TFIIS, which can stimulate RNA cleavage and restart arrested polymerases, likewise failed to rescue transcription through the lesion. This failure to resolve the blockage effectively commits the crosslinked site to the transcription-coupled repair pathway, ensuring that the damage is flagged for processing rather than erroneously bypassed.

The structural comparisons revealed additional mechanistic elegance. The fork loop-2 region of the polymerase was observed to separate the template and non-template DNA strands at the downstream fork of the transcription bubble, and this separation appears to facilitate the migration of the crosslinked non-template guanine as the lesion is remodeled. Furthermore, alignment of the cisplatin-stalled complexes with previously determined structures of polymerase II bound to CSB homologs, including yeast Rad26 and the human CSB-CSA-UVSSA complex, showed that the crosslink-stalled enzyme adopts a conformation competent for recruiting these repair factors, mirroring the way polymerase II arrested by bulky lesions on the template strand presents itself to the repair machinery.

The implications of this work extend from basic molecular biology to the clinic. By identifying RNA polymerase II as an active remodeler of cisplatin-induced interstrand crosslinks and defining the structural basis of the resulting stall, the study fills a long-standing gap in our understanding of how platinum drugs exert their transcription-blocking effects at atomic resolution. Because the efficiency of transcription-coupled repair influences cellular sensitivity to cisplatin, with repair-deficient cells such as certain testis tumor cells showing heightened drug sensitivity, the molecular details uncovered here could inform strategies to modulate repair pathways and improve therapeutic outcomes. The cryo-electron microscopy maps and atomic models have been deposited in public databases, ensuring that the structural framework established by the Zhejiang team will serve as a foundation for future studies of crosslink repair, polymerase stalling, and the cellular response to platinum-based chemotherapy.

Subject of Research: Structural mechanism of transcription stalling by cisplatin-induced DNA interstrand crosslinks

Article Title: Transcription remodeling and stalling by cisplatin-induced interstrand crosslinks

Article References: Zhu, L., Zhao, W., Ye, L., Liu, Y., Deng, W., Qian, B., Yang, X., & Xu, J. (2026). Transcription remodeling and stalling by cisplatin-induced interstrand crosslinks. Nature Chemical Biology. https://doi.org/10.1038/s41589-026-02338-3

Image Credits: AI Generated

DOI: 10.1038/s41589-026-02338-3

Keywords: cisplatin, interstrand crosslink, RNA polymerase II, transcription-coupled repair, cryo-electron microscopy, DNA damage, Cockayne syndrome protein B, oxaliplatin, backtracking, platinum chemotherapy, bridge helix, DNA repair