TTF2 promotes replisome eviction from stalled forks in mitosis

Abstract

When cells enter mitosis with under-replicated DNA, sister chromosome segregation is compromised, which can lead to massive genome instability. The replisome-associated E3 ubiquitin ligase TRAIP mitigates this threat by ubiquitylating the CMG helicase in mitosis, leading to disassembly of stalled replisomes, fork cleavage, and restoration of chromosome structure by alternative end-joining. Here, we show that replisome disassembly requires TRAIP phosphorylation by the mitotic Cyclin B-CDK1 kinase, as well as TTF2, a SWI/SNF ATPase previously implicated in the eviction of RNA polymerase from mitotic chromosomes. We find that TTF2 tethers TRAIP to replisomes using an N-terminal Zinc finger that binds to phosphorylated TRAIP and an adjacent TTF2 peptide that contacts the CMG-associated leading strand DNA polymerase ε. This TRAIP-TTF2-pol ε bridge, which forms independently of the TTF2 ATPase domain, is essential to promote CMG unloading and stalled fork breakage. Conversely, RNAPII eviction from mitotic chromosomes requires the ATPase activity of TTF2. We conclude that in mitosis, replisomes undergo a CDK- and TTF2-dependent structural reorganization that underlies the cellular response to incompletely replicated DNA.

Fig. 1:. TRAIP is activated by B-CDK1 phosphorylation in mitotic egg extracts.

(A) Domain organization and conserved CDK sites (red) of Xenopus laevis TRAIP, which is a dimer (21). (B) TRAIP is phosphorylated. The indicated TRAIP proteins were translated in wheat germ extracts and added to TRAIP-depleted Xenopus nucleoplasmic extract optionally supplemented with B-CDK1. Samples were treated with λ phosphatase, as indicated, separated on a Phos-tag gel, and blotted for TRAIP protein. (C) S295 and T325 are phosphorylated. The indicated TRAIP proteins were added to TRAIP-depleted NPE that was optionally supplemented with B-CDK1. Samples were blotted with the indicated antibodies, including two phospho-specific TRAIP antibodies, whose specificity was validated by the absence of signal in the appropriate phospho-site mutations (lanes 8 and 9). 2A, TRAIP S295A/T325A double mutant. (D) TRAIP’s CDK sites are required for ARP formation. A plasmid with 48 tandem lacO sites bound to LacR (“LacR plasmid”) was replicated in the indicated egg extracts supplemented with [α−³²P]dATP and different TRAIP mutants expressed in TTE. Samples were withdrawn at 30, 45, 60, 90, and 180 minutes after replication was initiated and subjected to gel electrophoresis and autoradiography. The structures and positions of ARP, θ, and supercoiled plasmid (SC, enriched when fork cleavage fails) are indicated. (E) TRAIP’s CDK sites are required for CMG unloading from stalled forks. LacR plasmid was replicated in the indicated egg extracts supplemented with different TRAIP mutants expressed in TTE, geminin (to prevent licensing and replication initiation), and B-CDK1, as noted. At 30, 45, and 90 minutes, chromatin was recovered and analyzed by blotting with the indicated antibodies. Two chromatin samples that were run on the same gel and blotted for TRAIP and GINS were inadvertently switched (red and blue asterisks). Here and in other complicated gels, samples with related conditions are grouped by dotted lines. (F) TRAIP’s CDK sites are required for ubiquitylation of stalled CMGs. LacR plasmid was replicated and recovered as in (E), except extracts contained NMS873 (p97-i) and MLN4924 (Cul-i; to inhibit CRL2^Lrr1-dependent ubiquitylation), and chromatin was recovered at 45 minutes.

Fig. 2:. TTF2’s zinc finger domain is required for mitotic CMG ubiquitylation by TRAIP.

(A) Mass spectrometry analysis of chromatin-associated proteins. LacR plasmid was replicated in the indicated egg extracts (as in Fig. 1D), recovered, and analyzed by label free mass spectrometry. Three replisome components are shown (bottom three rows), as are proteins specifically enriched in the presence of both B-CDK1 and p97-i (red box). For full results and confidence levels, see fig. S4 and Table S1. (B) Domain organization and key elements of Xenopus TTF2. (C) TRAIP is required for mitotic TTF2 recruitment. LacR plasmid replicated in the indicated extracts (as in Fig. 1D) was recovered (as in Fig. 1E) after 30 minutes, and blotted for TTF2 and histone H3. (D) The TTF2 Zinc finger domain but not its ATPase function is required for ARP formation. ARP assay was performed as in Fig. 1D using the indicated extracts and TTF2 constructs. (E) The TTF2 Zinc finger domain but not its ATPase function is required for stalled CMG unloading in mitosis. CMG unloading was measured (as in Fig. 1E) using the indicated extracts and TTF2 constructs. (F) The TTF2 Zinc finger domain but not its ATPase function is required for CMG ubiquitylation. CMG ubiquitylation was performed (as in Fig. 1F) using the indicated extracts and TTF2 constructs. (G) TTF2(1–200) is sufficient to support ARP formation. ARP assay was performed as in Fig. 1D using the indicated extracts and TTF2 constructs.

Fig. 3:. TRAIP and POLE2 binding by TTF2 is required for mitotic CMG unloading.

(A) AlphaFold3-predicted complex between TRAIP (280–339) containing phosphorylated T325 and TTF2 (1–200). Top, ribbon diagram. Bottom, same orientation as top with TTF2 shown with surface charge representation. (B) The predicted binding of TTF2 to TRAIP and POLE2 is required for ARP formation. ARP assay was performed as in Fig. 1D using the indicated extracts and TTF2 constructs. Although deletion of the Zinc finger from TTF2(1–200) abolished ARP formation, this result is not interpretable because we could not assess this mutant’s expression (see fig. S5D). (C) The predicted binding of TTF2 to TRAIP and POLE2 is required for CMG unloading. CMG unloading was performed (as in Fig. 1E) using the indicated extracts and TTF2 constructs. (D) The predicted binding of TTF2 to TRAIP and POLE2 is required for CMG ubiquitylation. CMG ubiquitylation was performed (as in Fig. 1E) using the indicated extracts and TTF2 constructs. (E) TTF2 is predicted to interact with POLE2. AF-M-predicted complex of full length TTF2 and POLE2 (see predictomes.org). Only the interacting peptide of TTF2 (118–126) is shown. (F) TTF2 and pol ε interact via residues 120–125 of TTF2. Anti-Flag resin was optionally incubated with Flag-tagged pol ε (tagged on POLE1), incubated with TTE expressing TTF2(1–200)^WT or TTF2(1–200)^Δ120−125, recovered, and eluted proteins were blotted as indicated.

Fig. 4:. A TRAIP-TTF2 fusion bypasses the need for B-CDK1 phosphorylation.

(A) Schematic of TRAIP-TTF2 fusions. TTF2 amino acids 95–205 were inserted between residues 350 and 351 of TRAIP. Orange box, POLE2 binding region of TTF2. Red X, mutations. (B-D) A TRAIP-TTF2 fusion lacking CDK sites (2A) supports ARP formation (B), CMG unloading (C), and CMG ubiquitylation (D), and a fusion also deficient in POLE2 binding (2A-ΔPOLE2) fails to support CMG unloading (C) and ubiquitylation (D). ARP, CMG unloading, and CMG ubiquitylation assays were performed as in Figs 1D, E, and F, respectively, using mock-depleted extract or extract depleted of both TRAIP and TTF2. Depleted extract was supplemented with the indicated fusion proteins or a mixture of TRAIP and TTF2.

Fig. 5.. The TTF2 ATPase domain is sufficient for RNAPII eviction from mitotic chromosomes and cell proliferation.

(A) Top, schematic of the human TTF2 gene with its encoded structural domains and the engineered C-terminal dTAG and Smash degrons. Bottom, complementing constructs under doxycycline (DOX) inducible tetON promoter. (B) The degron-tagged TTF2 construct is degraded within 30 minutes of ligand addition. The TTF2-degron and parental cell lines were exposed to dTAGV-1 and ASV ligands for the indicated times, and lysates were blotted for TTF2. (C) The TTF2 ATPase domain is required for RNAPII eviction. Immunofluorescence microscopy detection of RNAPII (S2-phosphorylated C-terminal domain) and complementing TTF2 constructs (α-V5 tag). Endogenous TTF2 was degraded, and TTF2 construct expression was induced (0.2 ug/mL DOX). Mitotic chromosome compaction was visualized using Hoechst dye and indicated with yellow arrowheads. (D) Quantification of RNAPII-S2ph signal intensities on mitotic chromatin relative to interphase nuclei. Cells were complemented with wild-type TTF2 or the indicated mutants after endogenous TTF2 degradation. RNAPII-S2ph intensities were measured after background subtraction and normalized to interphase levels. Data are presented as mean ± SEM; *p < 0.05. (E) The TTF2 ATPase domain is sufficient for cell viability and proliferation. Quantification of colony sizes from clonogenic assays is presented. (F) Model of TRAIP and TTF2 interaction with the mitotic replisome. The Zinc finger (ZF) of TTF2 binds phosphorylated T325 of TRAIP (yellow P) and a conserved motif adjacent to the ZF binds POLE2. In this manner, TTF2 forms a bridge between TRAIP and the replisome. *, putative contact point between TRAIP and the replisome (see Discussion). The flexible attachment of TRAIP to the replisome allows both cis ubiquitylation of CMG (TRAIP with solid lines) and trans ubiquitylation of a barrier ahead of the fork (TRAIP with dotted lines). TTF2 engaged in RNAPII eviction via the ATPase domain is not shown.