Structural insight into how the human helicase subunit MCM2 may act as a histone chaperone together with ASF1 at the replication fork

Abstract

MCM2 is a subunit of the replicative helicase machinery shown to interact with histones H3 and H4 during the replication process through its N-terminal domain. During replication, this interaction has been proposed to assist disassembly and assembly of nucleosomes on DNA. However, how this interaction participates in crosstalk with histone chaperones at the replication fork remains to be elucidated. Here, we solved the crystal structure of the ternary complex between the histone-binding domain of Mcm2 and the histones H3-H4 at 2.9 Å resolution. Histones H3 and H4 assemble as a tetramer in the crystal structure, but MCM2 interacts only with a single molecule of H3-H4. The latter interaction exploits binding surfaces that contact either DNA or H2B when H3-H4 dimers are incorporated in the nucleosome core particle. Upon binding of the ternary complex with the histone chaperone ASF1, the histone tetramer dissociates and both MCM2 and ASF1 interact simultaneously with the histones forming a 1:1:1:1 heteromeric complex. Thermodynamic analysis of the quaternary complex together with structural modeling support that ASF1 and MCM2 could form a chaperoning module for histones H3 and H4 protecting them from promiscuous interactions. This suggests an additional function for MCM2 outside its helicase function as a proper histone chaperone connected to the replication pathway.

Figure 1.

The histone H3-H4 interacting domain of MCM2 (A) Overlay of the ¹H-¹⁵N HSQC spectra of MCM2 (63–153) free (red) and in excess of full-length histones H3–H4 (purple) (the spectra were recorded at 20°C, in a Tris 10 mM buffer pH 8, 1.5 M NaCl). Signals disappearing in the bound form report for residues of MCM2 involved in the interaction with histones. (B) Histogram of the ¹³Cα chemical Shift Index (CSI) plotted with Wishart's reference sets for the unfolded state (see Methods for details) along the sequence of free MCM2 (62–153) highlights the propensity of the residues 107–121 to adopt a helical conformation (positive CSI > 1.0). (C) Intensity ratios of the ¹H-¹⁵N HSQC signals between bound and free MCM2 (63–153) indicate that residues 72–134 are involved in the interaction with histones H3-H4 except for the 97–104 segment.

Figure 2.

Structure of the ternary complex containing two MCM2 histone-binding domains (69–121) bound to the histone H3-H4 tetramer. (A) Ribbon representation of the MCM2⁶⁹–¹²¹−H3²³–¹⁰⁰−H4⁵⁷–¹³⁴ complex compared to the structure of the nucleosome (PDB code 1KX5 (1)) on the left and right panels, respectively. MCM2, H3 and H4 are in magenta, dark blue and light blue, respectively. H2A, H2B and DNA (right panel) are in light gray, dark gray and brown, respectively. H3′ and H4′ labels indicate the location of the second histones heterodimer. Three boxes in solid, dot-dashed and dashed lines (shown as close-up views in panels B–D) specify the zones anchoring MCM2 to the histones (left panel) and corresponding regions are boxed in right panel. In the close-up views, residues side-chain of MCM2 taking part to the interaction with H3−H4 heterodimer are shown as sticks while other side-chains are shown as lines. (B) Interactions made by MCM2 residues Leu72, Asp80 and Tyr81 with histone H4 helices. A rotation of the view shows that MCM2 helix (77–81) is incompatible with the binding of the αN-helix of histone H3 as in the nucleosome (right). No density for the αN residues of histone H3 could be observed in the structure of the MCM2−H3−H4 complex. (C) Close-up view of interactions made by MCM2 residues Asp88, Tyr90 and Leu95 with histone H3 helices. (D) The last helix (107–119) in the MCM2 histone-binding domain interacts with histone H4 (left panel) in the same region as the one involved in the binding of histone H2B in the entire nucleosome structure (right panel).

Figure 3.

Biophysical characterization and modeling of the quaternary complex formed by MCM2, ASF1 and the histone H3−H4 heterodimer. (A) Sec-MALS analysis at 20°C, in a Tris 50 mM buffer pH 8, 0.5 M NaCl of human ASF1a (1–156) (in green), ASF1a (1–156)-H3-H4 in dark green, ASF1a (1–156)-H3-H4-MCM2 (69–138) in light brown and ASF1a (1–156)-H3-H4-MCM2 (1–160) in dark brown. Relative optical density at 280 nm was plotted in arbitrary units in continuous lines as a function of the elution volume. The calculated molecular mass is reported as dashed line in the corresponding color with the secondary scale on the right. (B) SDS-PAGE analysis of Sec-MALS fractions for the different samples collected in A with a color code as in A. (C) Time course analysis of quaternary complex formation by gel filtration. Equimolar samples of the proteins ASF1, H3-H4 and MCM2 (69–138) were mixed at time zero and injected in the column at the indicated time. Composition of the three peaks is analyzed by SDS-PAGE on the right panel together with quantification of the histone fraction bond to MCM2 only or to ASF1and MCM2 over time. (D) Dissociation constant (Kdiss) of MCM2 (69–138) in complex with the preformed H3−H4−ASF1a (1–156) ternary complex as determined by ITC in 0.5 M NaCl, pH 8, at 20°C. (E) Model of the structural arrangement of the MCM2-H3-H4-ASF1 quaternary complex obtained by superimposing the structure of H3−H4 dimer in the H3−H4−ASF1a (1–156) ternary complex (PDB code 2IO5 (12)) with its structure in the MCM2 (69–121)-H3-H4 complex. ASF1 and MCM2 can accommodate around the histone H3−H4 heterodimer without any clash.

Figure 4.

Models for the role of the ternary (H3-H4)₂-MCM2 complex and the quaternary ASF1-H3-H4-MCM2 complex in handling histones during replication. (1) Several pathways can be envisaged for parental histone dissociation upstream the replication fork. After removal of H2A-H2B by dedicated chaperones (not represented on the figure), histone chaperones dedicated to H3-H4 could dissociate histones from DNA. Alternatively, the replicative MCM2–7 helicase could also destabilize interactions of parental histones H3-H4 with DNA via its mechanical force. The (H3-H4)₂-MCM2 ternary complex could then be transiently formed and protecting from re-association with DNA upstream the replisome machinery. (2) The quaternary Asf1–(H3-H4)–MCM2 complex observed in our study could constitute the next intermediate step further protecting and destabilizing the H3-H4 complex. (3) Parental histone reassembly could proceed by different mechanisms; the tetramer captured by MCM2 the N-terminus of MCM2 could be directly deposited on DNA without splitting or this histone tetramer it could be directly transferred to the assembly chaperone CAF-1. Alternatively, histone tetramers split by ASF1 could be reassembled by CAF-1 upon deposition on DNA. Besides, the human pathway dedicated to newly synthesized histones was also shown to involve ASF1 and CAF-1. The major histone fraction associated with MCM2 carries the specific modifications of newly synthesized histones suggesting that the quaternary complex could also transfer histones to CAF-1 for further assembly. All presented pathways remain hypothetical, may not be conserved in all species and may be restricted to some regions of the chromatin.