Mechanism of structure-specific DNA binding by the FANCM branchpoint translocase

Abstract

FANCM is a DNA repair protein that recognizes stalled replication forks, and recruits downstream repair factors. FANCM activity is also essential for the survival of cancer cells that utilize the Alternative Lengthening of Telomeres (ALT) mechanism. FANCM efficiently recognizes stalled replication forks in the genome or at telomeres through its strong affinity for branched DNA structures. In this study, we demonstrate that the N-terminal translocase domain drives this specific branched DNA recognition. The Hel2i subdomain within the translocase is crucial for effective substrate engagement and couples DNA binding to catalytic ATP-dependent branch migration. Removal of Hel2i or mutation of key DNA-binding residues within this domain diminished FANCM's affinity for junction DNA and abolished branch migration activity. Importantly, these mutant FANCM variants failed to rescue the cell cycle arrest, telomere-associated replication stress, or lethality of ALT-positive cancer cells depleted of endogenous FANCM. Our results reveal the Hel2i domain is key for FANCM to properly engage DNA substrates, and therefore plays an essential role in its tumour-suppressive functions by restraining the hyperactivation of the ALT pathway.

Graphical Abstract

Figure 1.

Translocase domain vs Full-length FANCM complex activity. (A) Overall domain structure of FANCM showing translocase domain. (B) Coomassie blue stained SDS-PAGE of purified full length FANCM complex or FANCM translocase (aa82–647). (C) Electromobility shift assay using oligonucleotide Holliday Junction (HJ) structures. ^=HJ bound by FANCMc, *=HJ bound by trFANCM, with graph of quantification. (D) Quantification of HJ-stimulated ATPase activity of FANCMc or trFANCM. (E) Branch migration assays showing conversion of 3-way (RF) and 4-way (HJ) DNA structures into dsDNA, with graph of quantification.

Figure 2.

Structural models of FANCM and its comparison to evolutionarily related proteins. (A) FANCM and related protein translocase domain structure, showing ATP-binding regions I–VI. (B) From left to right crystal structure of HEF (36), Homology modelled FANCM, Alpha Fold model of FANCM, TrRosetta model of the Hel2i domain. Only residues 82–600 are shown for clarity. (C) Crystal structures of dsRNA bound RIG-I (PDB:2YKG ( ref37) and MDA5 (PDB: 5JCH) (39) showing only equivalent regions to the above FANCM models.

Figure 3.

Structure specific DNA binding within the FANCM translocase domain. (A) Domain structure of FANCM variants used. (B) Coomassie stained WTc vs FANCMc_ΔHel2i, trFANCM and trFANCM_ΔHel2i, MBP-Hel2i and MBP. (C) Electromobility shift assays (EMSA) using recombinant FANCMc EMSA versus FANCMc_Δins and 4-way HJ DNA. (D) HJ EMSA with competition from increasing concentrations of unlabelled dsDNA. (E and F) As in (C) and (D), but using translocase only FANCM proteins. (G) EMSA of HJ using MBP-Hel2i domain or control MBP-only protein.

Figure 4.

The Hel2i domain is essential for branch migration of Holliday junctions (HJ) and R-loops, dependent upon its ATPase domain. (A) Branch migration activity of full length FANCMc and FANCMc_ΔHel2i on HJ substrate Lane 1 contains heat-denatured HJ (100%dsDNA product). (B) Branch migration activity of trFANCM, trFANCM_RIGI, trFANCM_ΔHel2i and ATPase-dead trFANCM_D214A translocase only proteins Lane 1 contains heat-denatured HJ. (C) R-loop resolution activity of FANCMc and FANCM_ΔHel2i on 140 bp R-loop structure within 2.1 kb plasmid. (D) HJ-dependent ATPase stimulation of trFANCM, trFANCM_RIGI and trFANCM_ΔHel2i.

Figure 5

Key residues in the FANCM-Hel2i domain govern engagement with, and activity on, branched DNA. (A) AlphaFold 3 model of FANCM translocase domain engaged with a splayed DNA molecule, colourized as per Figure 2, with residues C-terminal of the RecA2 fold shown in grey. (B) Zoom in of (A) showing key Hel2i residues predicted to bind at or near junction and mutated for testing in branch migration assays. (C) Effects of indicated single amino acid changes in Hel2i on HJ branch migration activity, (endpoint unwound product). (D) DNA stimulated ATPase activity and (E) HJ binding by EMSA for indicated mutants using 1, 4 or 16 nM protein. (F) Example DSF assays showing three replicate denaturation experiments of trFANCM in the absence or presence of ATP, DNA or DNA + ATP. (G) Change in overall stability of trFANCM or indicated mutants as measured by DSF. Only WT trFANCM shows increased thermal stability upon addition of DNA, while all variants show an ATP-dependent shift.

Figure 6.

FANCM binds the open form of HJ. Schematics of (A) open planar and stacked forms of HJ and (B) the set of 6 BXRH HJ of different arm lengths, showing their relative mobility in polyacrylamide gels. (C) Electromobility shift assay (EMSA) of BXRH HJ substrates with and without trFANCM in the absence of presence of 1 mM MgCl₂ in the sample and running buffer. (D) Cleavage of 10 nM static HJ by 1, 3, 10, 30, 90 or 270 nM GEN1 or nuclease dead GEN1-^D30N. Lane 1 is substrate only, lane 2 is control GEN1 cleavage (E) Cleavage of HJ by GEN1 (4 nM) in presence of 12.5, 25, 50, 100, 200 or 400 nM trFANCM or trFANCM^ΔHel2i.

Figure 7.

FANCM Hel2i domain integrity is essential for the survival of ALT + U2OS cells. (A) siFANCM targets endogenous but not exogenous FANCM mRNA for degradation in U2OS cells. (B) Western blot showing knockdown and re-expression of FANCM wildtype or mutants. (C) RPA/telomere FISH co-staining in example cells (scale bar = 10 mm), along with quantification of at least 100 nuclei in each line. Statistical analysis versus siCtrl by one way ANOVA (****P< 0.0001, ns = not significant). (D) Quantification of cell cycle profiles as measured by propidium iodide staining and flow cytometry. (E) Example clonogenic survival assay and quantification from n = 3 experiments performed in triplicate.