RQC helical hairpin in Bloom's syndrome helicase regulates DNA unwinding by dynamically intercepting nascent nucleotides

Abstract

The RecQ family of helicases are important for maintenance of genomic integrity. Although functions of constructive subdomains of this family of helicases have been extensively studied, the helical hairpin (HH) in the RecQ-C-terminal domain (RQC) has been underappreciated and remains poorly understood. Here by using single-molecule fluorescence resonance energy transfer, we found that HH in the human BLM transiently intercepts different numbers of nucleotides when it is unwinding a double-stranded DNA. Single-site mutations in HH that disrupt hydrogen bonds and/or salt bridges between DNA and HH change the DNA binding conformations and the unwinding features significantly. Our results, together with recent clinical tests that correlate single-site mutations in HH of human BLM with the phenotype of cancer-predisposing syndrome or Bloom's syndrome, implicate pivotal roles of HH in BLM's DNA unwinding activity. Similar mechanisms might also apply to other RecQ family helicases, calling for more attention to the RQC helical hairpin.

Keywords: Analytical chemistry instrumentation; Biochemistry; Biological sciences; Structural biology.

Graphical abstract

Figure 1

hBLM has multiple DNA binding conformations (A and B) Two crystal structures of the hBLM-DNA complex (PDB ID: 4O3M and 4CGZ). The zoom-ins display interactions between the helical hairpin and the ssDNA overhang according to the X-ray structures. (C) A DNA nanotensioner used in the assay. Cy3 was labeled at the 10th nucleotide on the ssDNA overhang, which is elongated upon BLM binding. (D) A typical FRET trace of hBLM binding to DNA. (E) Selected DNA binding bursts in the traces show multiple conformations. (F) Histogram with a multiple peak fitting to the binding signals. The histogram was built from 325 states of 52 traces. See also Figures S1 and S2.

Figure 2

Single-site mutations of the helical hairpin affect the DNA binding of hBLM (A–C) Typical single-molecule FRET traces of three hBLM mutants binding to DNA in the presence of ATPγS. (D–F) Histograms with multiple peak fittings of the binding signals. The histograms were built from 249 states of 53 traces for R1000A, 406 states of 67 traces for R1003A, and 201 states of 49 traces for K1009A, respectively. See also Figures S3 and S7.

Figure 3

Single-site mutations of the helical hairpin modulate the DNA unwinding pattern (A–E) Typical unwinding traces obtained by using the DNA construct in which Cy3 was labeled at the 10th nt on the overhang. (F–J) Typical unwinding traces when Cy3 was at the sixth nt on the overhang. (K–O) Typical unwinding traces when Cy3 was at −third nt on the overhang. The ATP concentration was 50 μM. See also Figures S4 and S5.

Figure 4

Single-site mutations at the helical hairpin modulate the DNA unwinding length (A–E) Histograms of the DNA unwinding length in terms of ΔFRET for the four hBLM helicases when Cy3 was labeled at the 10th nt on the overhang. (F–J) Histograms when Cy3 was labeled at the sixth nt. (K–O) Histograms when Cy3 was labeled at the −third nt. The start positions of the unwinding bursts (the green boxes) were estimated to be at the ends of the binding signals (the orange boxes). The error bars are proportional to the square root of the number of points for each bin. The number of bursts used in the statistics were 782–1,330. The corresponding number of traces were 383–564 for R1000A and 81–130 for other mutants.

Figure 5

A model of interception of nascent nucleotides by the helical hairpin during DNA unwinding by different hBLM mutants The DNA-interaction sites (filled cycles) in the helical hairpin affect the interception pattern and hence the binding and unwinding mode. The mutations are marked by open circles. The two RecA-like domains are denoted as D1 and D2. The dyes were labeled on the DNA.