Biochemical Properties of Naturally Occurring Human Bloom Helicase Variants

Abstract

Bloom syndrome helicase (BLM) is a RecQ-family helicase implicated in a variety of cellular processes, including DNA replication, DNA repair, and telomere maintenance. Mutations in human BLM cause Bloom syndrome (BS), an autosomal recessive disorder that leads to myriad negative health impacts including a predisposition to cancer. BS-causing mutations in BLM often negatively impact BLM ATPase and helicase activity. While BLM mutations that cause BS have been well characterized both in vitro and in vivo, there are other less studied BLM mutations that exist in the human population that do not lead to BS. Two of these non-BS mutations, encoding BLM P868L and BLM G1120R, when homozygous, increase sister chromatid exchanges in human cells. To characterize these naturally occurring BLM mutant proteins in vitro, we purified the BLM catalytic core (BLM_core, residues 636–1298) with either the P868L or G1120R substitution. We also purified a BLM_core K869A K870A mutant protein, which alters a lysine-rich loop proximal to the P868 residue. We found that BLM_core P868L and G1120R proteins were both able to hydrolyze ATP, bind diverse DNA substrates, and unwind G-quadruplex and duplex DNA structures. Molecular dynamics simulations suggest that the P868L substitution weakens the DNA interaction with the winged-helix domain of BLM and alters the orientation of one lobe of the ATPase domain. Because BLM_core P868L and G1120R retain helicase function in vitro, it is likely that the increased genome instability is caused by specific impacts of the mutant proteins in vivo. Interestingly, we found that BLM_core K869A K870A has diminished ATPase activity, weakened binding to duplex DNA structures, and less robust helicase activity compared to wild-type BLM_core. Thus, the lysine-rich loop may have an important role in ATPase activity and specific binding and DNA unwinding functions in BLM.

Introduction

Bloom syndrome (BS) is an autosomal recessive disorder that is characterized by shorter than average stature, sensitivity to sunlight, weakened immune response, and a predisposition to cancer (1). Patients with BS are more likely to encounter age-associated complications early in life, including adult-onset diabetes, various cancers, and chronic obstructive lung disease (1–3). BS can be diagnosed by assessing the level of damage and instability in chromosomes. BS is characterized by elevated chromosomal breakage, increased sister chromatid exchanges (SCEs), and quadriradial chromosomal structures (4). Previous research has found that BS-causing mutations lead to a 10–12 fold increase in SCEs in human lymphocytes, and similar phenotypes have been demonstrated in DT40 lymphoma cells (5,6).

BS is caused by mutations to BLM, which encodes the Bloom Syndrome DNA helicase (BLM) (4). BLM is a member of the RecQ family of helicases and can unwind a variety of DNA secondary structures with 3′ to 5′ polarity, including double-stranded DNA (dsDNA), forked DNA substrates, G-quadruplex DNA (G4s), and Holliday junctions (HJs) (3,7). There are five human RecQ-family helicases, which include BLM, WRN, RECQL1, RECQL4, and RECQL5. Mutations in four of these RecQ helicases, including BLM, are associated with autosomal recessive disorders (8,9). As such, RecQ-family helicases are often defined as “caretakers” of the genome (10).

BLM is essential for promoting genome stability in the cell and has functions in DNA replication, DNA repair, transcription, and telomere maintenance (3). In keeping with its role as a genome maintenance protein, BLM localizes to DNA replication and DNA damage sites, and to telomeres (3). Further, BLM interacts with a large network of DNA replication and repair proteins and is thought to act in concert with a larger complex of proteins for many of its core functions (3,11). For example, BLM is part of the dissolvasome, which is composed of BLM, topoisomerase IIIα (Top3α), RecQ-mediated genome instability protein 1 (RMI1), and RMI2 (11). The dissolvasome is involved in dissolution of double Holliday junctions (dHJs) generated during homologous recombination. This function of the dissolvasome prevents crossover during recombination, suppressing SCEs (12,13). Thus, defects in BLM cause many adverse effects on chromosome stability due to the multifaceted nature of BLM.

The BLM catalytic core includes a Helicase domain, a RecQ C-terminal domain (RQC), and a Helicase and RNase D C-terminal domain (HRDC) (Fig 1A) (14,15). The Helicase domain of BLM is composed of two RecA-like subdomains, which are important for ATP hydrolysis and helicase activity. The RQC domain is composed of Zinc-binding (ZN) and Winged Helix (WH) subdomains. The ZN subdomain includes four cysteine residues that coordinate a Zn²⁺ ion, and the WH is important for interactions with various DNA structures. The HRDC is likely involved in binding single-stranded DNA (ssDNA) and increasing BLM processivity (7). The N-terminal and C-terminal regions of BLM are highly disordered and expected to be important for protein-protein interactions (7,16).

Fig 1. Structure of BLM.

(A). Structure of BLM catalytic core (PDBID: 4O3M (14)), colored according to domain map below structure. ADP is shown as sticks and magnesium and zinc ions are shown as green and grays spheres, respectively. Positions of substitutions P868L and G1120R are highlighted on domain map. (B). P868, K869, and K870 residues, shown as sticks. (C). G1120 residue, shown as sticks.

Mutations in BLM that lead to BS include missense mutations, nonsense mutations, frameshifts, and splicing defects (4). Many of the missense mutations interfere with helicase and ATPase activity of BLM (4,17). While the impact of homozygous mutations in BLM have been extensively characterized clinically, the health impacts of heterozygous BS-causing mutations is less well-studied. Limited studies have found that heterozygous BS mutations can lead to increased cancer risk in humans and in mice (18,19). Thus, although BS is recessive and manifests only when both copies of BLM are mutated, disrupting activity of one BLM allele may still have detrimental impacts on human health and longevity.

In addition to the lack of study on the impacts of heterozygous BLM mutations, there has been very little clinical or in vitro study of other naturally occurring BLM mutations that do not result in BS. There are over 1400 BLM mutations that exist in the human population, with 62 encoding for pathogenic mutations, 118 mutations considered benign, and the rest of uncertain or conflicting human health relevance (20). In 2012, Mirzaei and Schmidt (21) investigated the effects of 41 BLM single nucleotide polymorphisms from the Short Genetic Variations database map by assessing the effects of these mutations in Saccharomyces cerevisiae. Interestingly, two substitutions, P868L and G1120R, led to a partial loss in function, demonstrated by increased sensitivity to hydroxyurea compared to wild-type (WT) BLM. These mutations also cause a 4–5 fold increase in SCEs and increased sensitivity to hydroxyurea in human cells (22). While these BLM mutations have less severe phenotypes than BS-causing mutations, P868L and G1120R substitutions still lead to intermediate genomic instability in cells. Mutations in BLM that cause this intermediate phenotype could lead to increased risk of cancer or other adverse health impacts in humans. Interestingly, the BLM P868L allele has a 5.13% frequency in the human population, and the impacts of this heterozygous (or homozygous) BLM allele on human health and longevity are unknown (22). Residue P868 resides in the helicase domain of BLM (Fig 1B) and is part of a lysine-rich loop (868-PKKPKK) that appears to be important for interacting with DNA at the ssDNA-dsDNA junction in BLM crystal structures (14,15). Residue G1120 is part of the BLM WH and terminates an alpha helix preceding a loop proximal to dsDNA in the BLM-dsDNA structure (Fig 1C).

To better characterize naturally occurring non-BS BLM mutations, we tested BLM_core constructs (residues 636–1298) with P868L or G1120R substitutions using a battery of biochemical assays. We also measured activity of a BLM_core K869A K870A mutant protein to assess the role of lysines within the BLM lysine-rich loop. The BLM_core retains helicase, DNA binding, and ATPase activity, making it the simplest model for characterizing the impacts of these naturally occurring mutant proteins in vitro (14,15). Previous in vitro studies investigating the impacts of BS-causing mutations utilized a nearly identical BLM catalytic core construct (residues 642–1290) and found that BS-causing mutations had severe impacts on the helicase and ATPase activity of BLM (17). Additionally, defects detected with BLM P868L and BLM G1120R in yeast used a construct that included residues 648–1417 of BLM (21). Therefore, we focused our experiments on BLM_core constructs to foster comparison with prior reports and with the reasoning that changes to the BLM_core could allow measurement of moderate impacts on the in vitro function of BLM. In addition to our biochemical experiments, we compared molecular dynamics (MD) simulations of the WT BLM_core and BLM_core P868L bound to DNA and ADP.

While we predicted that the P868L and G1120R substitutions would impact the in vitro function of BLM_core, these mutant proteins had similar DNA-dependent ATPase activity, DNA binding activity, and helicase activity to WT BLM_core. In contrast, BLM_core K869A K870A requires ~4-fold more DNA to stimulate ATPase activity and has less robust helicase activity compared to WT BLM_core. Additionally, DNA binding experiments suggest that BLM_core K869A K870A may be deficient in recognizing ssDNA-dsDNA junctions. Since the naturally occurring BLM_core mutant proteins maintained activity levels that were very similar to WT BLM_core, our results indicate that these mutations lead to moderate genome instability through impairment of other cellular functions of BLM. In alignment with these results, the MD simulations suggest that the P868L substitution could subtly weaken interaction of DNA with BLM’s winged-helix domain and alter the orientation of the N-terminal lobe of the ATPase domain, which could shed light on previously identified defects of the hypomorphic P868L mutant protein in vivo (22).

Materials and Methods

Protein expression and purification

The BLM_core (residues 636–1298) and BLM_core mutant proteins were overexpressed as previously described (14). Briefly, overexpression was carried out in Rosetta 2 (DE3) E. coli transformed with pLysS and BLM_core or BLM_core mutant protein overexpression plasmids. Cells were grown at 37 °C in Terrific Broth supplemented with 50 μg/mL kanamycin and 50 μg/mL chloramphenicol until cultures reached an OD₆₀₀ ~1.8. BLM overexpression was induced using 0.5 mM IPTG and cells were grown overnight at 18 °C. Cells were then pelleted and stored at − 80 °C.

Cells pellets were resuspended in lysis buffer (20 mM Tris-HCl, pH 8.0, 0.5 M NaCl, 10% (v/v) glycerol, 0.1% Triton X-100, 20 mM imidazole, 1 Pierce Protease inhibitor tablet/100 mL buffer (Thermo Fisher), and 0.1 mM phenylmethylsulfonyl fluoride), lysed using sonication, and cleared via centrifugation. Clarified lysate was loaded onto a 5 mL HisTrap FF column (Cytiva) that was equilibrated with Buffer A (20 mM Tris-HCl, pH 8.0, 0.5 M NaCl, 10% (v/v) glycerol, and 20 mM imidazole). Protein was eluted with Buffer B (20 mM Tris-HCl, pH 8.0, 0.5 M NaCl, 10% (v/v) glycerol, 250 mM imidazole). Fractions containing BLM_core were diluted to 0.25 M NaCl and loaded onto a HiPrep 16/10 Heparin FF Column (Cytiva). The column was washed with Buffer C (20 mM Tris-HCl, pH 8.0, 0.25 M NaCl, 10% (v/v) glycerol) and then BLM_core was eluted from the column using Buffer D (20 mM Tris-HCl, pH 8.0, 1 M NaCl, 10% (v/v) glycerol). Fractions containing BLM_core were concentrated in Vivaspin-20 10 kDa concentrator (Sartorius) to 2 mL and loaded onto HiPrep 16/60 Sephacryl S-300 HR column (Cytiva) using Sizing Buffer (20 mM Tris-HCl, pH 8.0, 0.5 M NaCl, 5% (v/v) glycerol). Fractions containing BLM_core were concentrated to ~10 mg/mL then flash frozen in liquid nitrogen and stored at −80 °C. BLM_core and mutants proteins were tested for nuclease activity at the highest concentration used in this study and found to have no detectable nuclease activity (Fig S1).

The BLM_core K869A K870A construct was purified with the following modifications to remove a contaminating nuclease. After collecting fractions from the HisTrap FF column, fractions were assessed for nuclease activity. Five μL of each fraction was incubated for 30 minutes at ambient temperature with 40 nM fluorescein labeled dT₃₀ in 50 mM Tris-HCl, pH 7.5, 50 mM KCl, 1 mM DTT, 0.1 mg/mL Bovine Serum Albumin (BSA), and 5 mM MgCl₂. Five μL of stop buffer (2% SDS, 5 μg/mL proteinase K, 20% (v/v) glycerol, 0.1 ethylenediaminetetraacetic acid (EDTA)) was added to each sample, and 5 μL of each sample was loaded onto a 15% acrylamide 1.5-mm gel in Tris-Borate-EDTA (TBE) buffer supplemented with 100 mM KCl. Gels were run at 75 V for 1 hour at 4 °C in 1xTBE running buffer with 100 mM KCl and imaged on an Azure c600 (Azure Biosystems). Fractions that contained significant nuclease activity were discarded, and fractions containing BLM_core K869A K870A without significant nuclease activity were diluted to 0.15 M NaCl and loaded onto the HiPrep 16/10 Heparin FF Column, washed with Buffer C2 (20 mM Tris-HCl, pH 8.0, 0.15 M NaCl, 10% (v/v) glycerol), and eluted over a 20-column volume gradient up to 0.5 M NaCl. Fractions containing BLM_core were tested for nuclease activity. Fractions that contained BLM_core and no detectable nuclease activity were concentrated in Vivaspin-20 10 kDa concentrator (Sartorius) to 2 mL and loaded onto the HiPrep 16/60 Sephacryl S-300 HR column (Cytiva) using Sizing Buffer (20 mM Tris-HCl, pH 8.0, 0.5 M NaCl, 5% (v/v) glycerol). Fractions were tested for nuclease activity, and fractions with no detectable nuclease activity were concentrated then flash frozen in liquid nitrogen and stored at −80 °C.

DNA-dependent ATPase assay

WT BLM_core, BLM_core P868L, or BLM_core G1120R (5 nM) were incubated with either no DNA or dT₂₀ serially diluted from 10 nM to 9.8 × 10⁻³ nM in ATPase buffer (20 mM Tris-HCl, pH 8.0, 50 mM NaCl, 5% (v/v) glycerol, 0.1 mM DTT, 5 mM MgCl₂, 0.1 mg/mL BSA, 2 mM 2-phosphoenolpyruvate, 3 U/mL Pyruvate Kinase/Lactate Dehydrogenase, 0.2 mM NADH). BLM_core K869A K870A at 5 nM was incubated with no DNA or with dT₂₀ serially diluted from 100 nM to 0.98 nM in ATPase buffer. Reactions were initiated by addition of 1 mM ATP and A_{340 nm} was monitored for 1 hour at 25 °C. Data were analyzed as previously described (23) and plotted using Prism Version 9.3.1. ATPase assays were done in triplicate. Significance was determined using Welch’s two-tailed t-test using the default settings on Prism Version 9.3.1.

Electrophoresis Mobility Shift Assays

Electrophoresis mobility shift assays (EMSAs) were performed as previous described (24) with the following modifications. Partial duplex DNA (oRC32/FAM-oAV320, Table 1) or G4 DNA (FAM-oRC96, Table 1) constructs were folded by incubating 5 μM DNA in 10 mM Tris-HCl, pH 7.5 and 100 mM KCl at 95 °C for 10 minutes and slowly cooling the sample to room temperature over several hours. DNA constructs were stored at 4 °C. Serial dilutions of BLM_core or BLM_core mutant proteins were incubated with 40 nM partial duplex DNA, G4 DNA, or FAM-dT₃₀ in 50 mM Tris-HCl, pH 7.5, 50 mM KCl, 1 mM DTT, 0.1 mg/mL BSA, and 5 mM MgCl₂ for 30 minutes at room temperature. 3.3% (v/v) glycerol was added to samples, and 5 μL of each sample was loaded onto a 5% acrylamide 1.5-mm gel in TBE buffer supplemented with 100 mM KCl. Gels were pre-run at 75 V for 20 minutes before loading protein/DNA complexes and running at 75 V for 1 hour at 4 °C in 1xTBE running buffer supplemented with 100 mM KCl. Gels were imaged on the Azure c600 (Azure Biosystems). Experiments were done in triplicate.

Table 1.

Oligonucleotides used in this study.

FAM-oRC96	5’-GGGTTAGGGTTAGGGTTAGGGTTTTTTTTTT-FAM-3’
oRC32	5’-TGGCGACGGCAGCGAGGCTTAGGGTTAGCGTTAGCGTTAGGGTTTTTTTTTTTTTTT-3’
oRC75	5’-TGGCGACGGCAGCGAGGCTTAGGGTTAGGGTTAGGGTTAGGGTTTTTTTTTTTTTTT-3’
oAV320	5’-GCCTCGCTGCCGTCGCCA-FAM-3’
oAV322	5’-GCCTCGCTGCCGTCGCCA-3’

Helicase assay

Helicase assays were performed as previous described (24) with the following modifications. Partial dsDNA or G4-dsDNA constructs were folded by incubating 5 μM DNA in 10 mM Tris-HCl, pH 7.5 and 100 mM KCl at 95 °C for 10 minutes and slowly cooling the sample to room temperature over several hours, then stored at 4 °C. Serial dilutions of BLM_core or BLM_core mutant proteins were incubated with 40 nM oRC32/FAM-oAV320 (partial duplex, Table 1) or oRC75/FAM-oAV320 (G4-dsDNA, Table 1) in helicase assay buffer (50 mM Tris-HCl, pH 7.5, 50 mM KCl, 1 mM DTT, 0.1 mg/mL BSA, 5 mM MgCl₂, 5 mM ATP, 4 μM oAV322) for 15 minutes at 37 °C in 25 μL reactions. Folded DNA control was obtained by incubating reaction mixture without BLM_core, and melted control was obtained by omitting BLM_core and heating reaction to 95 °C for 10 minutes. Five μL of stop buffer (2% SDS, 5 μg/mL proteinase K, 20% (v/v) glycerol, 0.1 mM EDTA) was added to each reaction and 5 μL of each sample was loaded onto a 15% acrylamide 1.5-mm gel in TBE buffer supplemented with 100 mM KCl. Gels were run at 75 V for 1 hour at 4 °C in 1x TBE running buffer with 100 mM KCl. Gels were imaged on an Azure c600 (Azure Biosystems). BLM_core dsDNA, BLM_core P868L dsDNA and G4-dsDNA, and BLM_core K869A K870A dsDNA and G4-dsDNA were done in triplicate. BLM_core G4-dsDNA, BLM_core G1120R dsDNA and G4-dsDNA were done in quadruplicate. DNA unfolding was quantified using ImageJ and analyzed using Prism Version 9.3.1 and fitting data to Equation (1). Significance was determined using Welch’s two-tailed t-test using the default settings on Prism Version 9.3.1.

$$Fractionunwound=Fractionunwoun{d}_{max}\ast \frac{[BLM]}{\left(K_{BLM}+[BLM]\right)}+Background$$ 1

Molecular dynamics simulations

The initial conformation for WT BLM_core was taken from one of the X-ray structures (PDB ID: 4CGZ) (15). This structure contains a portion of the BLM protein (residues 637–1290) bound to DNA, an ADP molecule and a Zn²⁺ ion. Coordinates of the missing hydrogens were generated using the ‘pdb2gmx’ module of the GROMACS software (25), and the initial coordinates of a missing loop (residues 1194–1206) were constructed using MODELLER (26). All titratable amino acids were assigned their default protonation states at pH of 7.4, except for the four cysteines (residues 1036,1055,1063,1066) that coordinate the bound Zn²⁺ ion, which are expected to be deprotonated (27). Histidine protonation assignments (HD1 or HE2) were selected by carrying out a global optimization of the hydrogen bond network (28). After building all missing atoms, the complex was energy minimized and then placed in a 12×12×12 nm cubic box containing 54,282 water molecules, 97 K⁺ ions and 79 Cl⁻ ions, corresponding to an ionic strength of 100 mM. The difference in the numbers of K⁺ and Cl⁻ ions serves to counter balance the net negative change of −12 eu on the complex. Following addition of solvent, the system was energy minimized for 500 steps, after which it was subjected to molecular dynamics (MD) simulation. The initial conformation of the P868L mutant protein was constructed from the energy minimized structure of the WT complex. BLM P868L was placed in a unit cell identical to the WT complex and it contains the same numbers of water molecules and salt ions. The P868L mutant protein unit cell was also energy minimized for 500 steps, and then subjected to MD. MD simulations were carried out using GROMACS version 2020 (25). Integration was carried out using the leap-frog algorithm and with a time step of 2 fs. All bonds were constrained (29). Simulations were carried out under isothermal-isobaric conditions. Pressure was regulated at 1.01325 bar using the Parrinello-Rahman extended ensemble approach (30), coupling time constant of 1.0 ps, and a compressibility of 4.5 × 10⁻⁵. Temperature was regulated at 310 K using the velocity-rescale thermostat (31) and a coupling constant of 0.1 ps, which uses a stochastic term for the generation of a proper canonical ensemble. Periodic boundary conditions were set in all directions, and electrostatic interactions were computed using a particle mesh Ewald scheme (32) with a short-range cutoff of 10 Å, Fourier grid spacing of 1 Å, and a fourth-order interpolation. van der Waals interactions were computed explicitly for interatomic distances smaller than 10 Å. Neighbor lists were constructed using grid search and were updated every 5 steps. Less frequent updates to neighbor lists, which are typically implemented to accelerate simulations on GPUs, resulted in DNA deformation. Water is described using the SPC/E model (33), and ion-water interactions are described using Joung and Cheatham parameters (34). Protein, solvent ions, and DNA are described using the AMBER99SB-ILDN force field (35), along with NB-fix corrections for protein-DNA, ion-DNA and ion-ion interactions (36).

Results

BLM_core K869A K870A requires higher levels of ssDNA than WT BLM_core to stimulate ATPase activity

BLM requires ATP hydrolysis for translocation along DNA and helicase activity. Therefore, the human BLM mutant proteins BLM P868L and BLM G1120R could cause cellular defects due to a decrease in ATPase activity. To test this, we measured the DNA-dependent ATPase activity of mutant BLM_core proteins. We used a dT₂₀ ssDNA substrate for this assay to examine only ssDNA-dependent activity, avoiding complications that could arise from ATPase activity associated with dsDNA unwinding. Interestingly, both of the BLM_core mutant proteins had similar maximum ATPase rates to WT BLM_core (Fig 2A, Table 2), indicating that BLM_core P868L and BLM_core G1120R are able to efficiently hydrolyze ATP when bound to ssDNA. While BLM_core P868L had an apparent increase in maximum ATPase rate compared to WT BLM (P value = 0.0116), there is a relatively small difference (35%) between the WT BLM_core and BLM_core P868L maximum ATPase rates (Table 2). Since the concentration of BLM_core directly impacts measured maximal ATPase rates, small inaccuracies in concentration determination can lead to modest apparent differences that are difficult to interpret. We therefore do not ascribe biochemical significance to this difference. The DNA concentration-dependence of the ATPase activity was also similar for WT BLM_core, BLM_core P868L, and BLM_core G1120R. The DNA concentrations required for half-maximum ATPase rate (K_DNA) differed by less than 1-fold among all three proteins (Fig 2B, Table 2).

Fig 2. DNA-dependent ATPase assays.

(A). Maximum DNA-dependent ATPase rates for WT BLM_core, BLM_core P868L, BLM_core G1120R, and BLM_core K869A K870A. These data represent the mean of three replicates with error bars representing the standard error of the mean. P-values are indicated above the bars. (B). ssDNA concentrations required for half-maximum ATPase rate (K_DNA) for WT BLM_core, BLM_core P868L, BLM_core G1120R, and BLM_core K869A K870A. These data represent the mean of three replicates with error bars representing the standard error of the mean. P-values are indicated above the bars.

Table 2.

Summary of ATPase, DNA binding, and helicase activity

ATPase Data Summary
	WT BLMcore	BLMcore P868L	BLMcore G1120R	BLMcore K869A K870A
Maximum ATPase Rate (min−1)	1010 ± 48.2	1370 ± 62.0	1130 ± 48.0	811 ± 25.0
KDNA(nM)	0.839 ± 0.137	1.13 ± 0.163	1.45 ± 0.185	3.57 ± 0.444
ATPase rate at 0 nM DNA (min−1)	110 ± 5.4	78 ± 17	45 ± 3.7	Not detected
DNA Binding Data Summary
	WT BLMcore	BLMcore P868L	BLMcore G1120R	BLMcore K869A K870A
	79 nM, distinct	62.5 nM, distinct	62.5 nM, distinct	62.5 nM, smear
	125 nM, smear	125 nM, smear	250 nM, smear	625 nM, smear
	125 nM, smear	62.5 nM, smear	62.5 nM, smear	312.5 nM, smear
Helicase Activity Summary
	WT BLMcore	BLMcore P868L	BLMcore G1120R	BLMcore K869A K870A
KBLM for dsDNA substrate unwinding (nM)	15 ± 2.4	12 ± 2.4	13 ± 2.4	22 ± 6.6
Maximum fraction of dsDNA unwound	0.99 ± 0.038	1.1 ± 0.058	0.92 ± 0.043	0.76 ± 0.060
KBLM for G4-dsDNA substrate unwinding (nM)	17 ± 2.4	6.2 ± 1.4	23 ± 3.1	46 ± 8.8
Maximum fraction of G4-dsDNA unwound	0.96 ± 0.031	0.95 ± 0.051	0.99 ± 0.031	0.72 ± 0.032

Summary of ATPase, DNA binding, and helicase assay data described in this study. For DNA binding data, values indicated are the estimated apparent K_D which either manifested as a smear or distinct band shift.

BLM_core K869A K870A was found to have a modestly different maximum ATPase rate compared to WT BLM_core (P value = 0.0348). More interestingly, this mutant protein had a ~4-fold higher K_DNA value (P value = 0.0183), indicating that higher concentrations of ssDNA were necessary to stimulate its half-maximum ATPase activity (Fig 2, Table 2). Unlike maximal ATPase measurements, K_DNA values are independent of protein concentration, indicating that this difference is biochemically significant. These data show that disrupting the charge of the lysine-rich loop impacted the concentration of ssDNA required to stimulate DNA-dependent ATP hydrolysis in BLM_core whereas the P868L and G1120R substitutions did not. This result is consistent with a reduced DNA binding affinity for BLM_core K869A K870A relative to the other BLM_core constructs tested.

Summary of ATPase, DNA binding, and helicase assay data described in this study. For DNA binding data, values indicated are the estimated apparent K_D which either manifested as a smear or distinct band shift.

We also assessed the ATPase rate for each mutant protein in the absence of DNA. WT BLM_core, BLM_core P868L, and BLM_core G1120R were capable of hydrolyzing ATP in the absence of DNA, with rates of 4–10% of that observed with saturating levels of ssDNA. In contrast, BLM_core K869A K870A had no detectable DNA-independent ATPase activity (Table 2). This finding aligns with the higher DNA concentration requirement for this mutant protein and may point to a defect in its overall ATPase functions.

BLM_core K869A K870A has less specificity for binding ssDNA-dsDNA junctions

BLM is able to bind and unwind a variety of DNA structures (3). To test the ability of the BLM_core proteins to bind to different DNA structures, we used Electrophoretic Mobility Shift Assays (EMSAs) to assess binding to ssDNA, G4 DNA with a 3’ ssDNA overhang, and dsDNA with a 3’ ssDNA overhang (partial dsDNA), each done in triplicate. WT BLM_core bound to ssDNA and the G4 substrate, as evidenced by slower migration of the labeled DNA on a gel (Fig 3). This shift manifested as “smearing” for these substrates, which could indicate that these complexes are dynamic, dissociating and re-associating in the assay. In contrast, when tested with partial dsDNA, WT BLM_core was able to form a distinct band shift, consistent with BLM_core forming a stable complex with the substrate. The same trend was observed for BLM_core P868L and BLM_core G1120R, which were both able to shift ssDNA and G4 DNA, indicated by smearing, and were able to shift partial dsDNA, indicated by distinct band shifts (Fig 3). From estimating qualitative binding affinities from the gel shift assays, WT BLM_core, BLM_core P868L, and BLM_core G1120R all appear to have similar affinities for each of these DNA substrates (Table 2).

Fig 3. DNA binding assays for WT BLM and mutant proteins.

Electrophoretic mobility shift assays for WT BLM_core, BLM_core P868L, BLM_core G1120R, and BLM_core K869A K870A (left to right) for binding ssDNA (top), G4 DNA (middle), and partial-duplex DNA (bottom). Lowest and highest concentrations are indicated below each gel, and the estimated apparent K_D is indicated with an arrow. Concentrations for WT BLM_core with G4 and with ssDNA, BLM_core P868L with G4, ssDNA, and partial-dsDNA, and BLM_core G1120R with G4, ssDNA, and partial-dsDNA are 1.00 μM, 500 nM, 250 nM, 125 nM, 62.5 nM, 31.2 nM, 15.6 nM, 7.81 nM, 3.91 nM, and 1.95 nM. WT BLM with partial-dsDNA concentrations are 600 nM, 400 nM, 267 nM, 178 nM, 118 nM, 79.0 nM, 52.7 nM, 35.1 nM, 23.4 nM, and 15.6 nM. Concentrations for BLM_core K869A K870A with G4, ssDNA, and partial-dsDNA are 2.5 μM, 1.25 μM, 625 nM, 312 nM, 156 nM, 78.1 nM, 39.1 nM, 19.5 nM, 9.77 nM, and 4.88 nM. DNA concentration for each assay is 40 nM. EMSAs were run in triplicate with the gel shown here as a representative example.

BLM_core K869A K870A was also able to bind to ssDNA and G4 DNA (Fig 3) but required higher protein concentrations to shift the DNA, consistent with a lower affinity for these substrates (Table 2). This binding also resulted in a smear above the substrate migration site. Whereas WT BLM_core, BLM_core P868L, and BLM_core G1120R were able to bind the partial dsDNA causing a distinct band shift, the BLM_core K869A K870A mutant protein did not have this effect. Instead, BLM_core K869A K870A binding to partial duplex DNA resulted in a smear, similar to what was observed for binding to ssDNA and G4 DNA. Additionally, a band corresponding to the labeled ssDNA from the substrate was consistently observed with the highest concentration of BLM_core K869A K870A (1 μM). This could be due to modest ATP-independent unwinding at high concentrations of this mutant protein. As described in the Materials and Methods, this effect is not due to a contaminating nuclease in the BLM_core K869A K870A purification. Thus, BLM_core K869A K870A appears to have reduced overall DNA binding stability and reduced ability to discriminate among different DNA structures.

BLM_core K869A K870A is a less efficient helicase than WT BLM_core, BLM_core P868L, and BLM_core G1120R

We next tested the ability of the BLM mutant proteins to unwind dsDNA and G4-dsDNA. This experiment was done using a gel-based helicase assay with a substrate containing a 3’ ssDNA 15 nucleotide overhang, either the human telomere G4-forming sequence or a control sequence that cannot form a G4, and 18 base pairs of dsDNA (Fig 4A). The annealed strand contains a fluorescent label on its 3’ end and unwinding of this substrate can be observed by tracking migration of fluorescent label in a gel (Fig 4B). The reaction is initiated by the addition of BLM_core or mutant protein, ATP, and an unlabeled trap oligo that has the same sequence as the fluorescently labeled ssDNA. The total fraction of DNA unwound at various BLM_core concentrations is measured (Fig 4C), allowing a determination of the fraction of DNA that BLM_core or mutant protein can unwind and the concentration of BLM_core required for half-maximum unwinding (K_BLM).

Fig 4. WT BLM_core and BLM_core mutant protein helicase assays unwinding duplex DNA and G4 DNA.

(A) Schematic of experimental system. (B). (Top to bottom): WT BLM_core, BLM_core P868L, BLM_core G1120R, and BLM_core K869A K870A gel-based helicase assays assessing duplex DNA unwinding (left) and G4-dsDNA unwinding (right). For WT BLM_core and BLM_core G1120R, protein concentrations are 1.00 μM, 500 nM, 250 nM, 125 nM, 62.5 nM, 31.2 nM, 15.6 nM, 7.81 nM, 3.91 nM, and 1.95 nM. For BLM_core P868L, protein concentrations are 100 nM, 50 nM, 25 nM, 12.5 nM, 6.25 nM, 3.12 nM, 1.56 nM, 0.781 nM, 0.391 nM, and 0.195 nM. For BLM_core K869A K870A, protein concentrations are 2.5 μM, 1.25 μM, 625 nM, 312 nM, 156 nM, 78.1 nM, 39.1 nM, 19.5 nM, 9.77 nM, and 4.88 nM. DNA concentrations are 40 nM for each assay. Gels are representative examples of helicase assays run in triplicate or quadruplicate. (C). Quantification of fraction of substrate unwound for dsDNA unwinding (black) and G4-dsDNA unwinding (blue). Datapoints are the mean values of helicase assays with error bars representing the standard error of the mean.

WT BLM_core was able to efficiently unwind both the dsDNA and the G4-dsDNA substrates (Figs 4B and 4C), with over 90% maximum unwinding for both. The K_BLM value for WT BLM was 15 ± 2.4 nM and 17 ± 2.4 nM for dsDNA and G4-dsDNA, respectively (Table 2). Both BLM_core G1120R and BLM_core P868L were able to unwind over 90% of the dsDNA substrate and G4-dsDNA substrate as well, indicating that these mutant proteins have similar maximum activity levels to WT BLM_core. BLM_core P868L had K_BLM values of 12 ± 2.4 nM and 6.2 ±1.4 nM for dsDNA and G4-dsDNA, respectively (Table 2), indicating that BLM_core P868L requires similar amounts of protein to unwind dsDNA as WT BLM_core but that it was somewhat more efficient at unwinding G4-dsDNA substrates (P values = 0.36 and 0.012). BLM_core G1120R had K_BLM values of 13 ± 2.4 nM and 23 ± 3.1 nM for dsDNA and G4-dsDNA unwinding, indicating that BLM_core G1120R unwound dsDNA and G4-dsDNA with a similar efficiency to that of WT BLM_core (P values = 0.56 and 0.22).

BLM_core K869A K870A was unable to unwind the same percentage of substrate, with lower than 80% total unwinding for both dsDNA and G4-dsDNA (Table 2). At high concentrations of BLM_core K869A K870A, we observed inhibition of unwinding (Fig 4C), which is also observed at high concentrations of WT BLM_core. The K_BLM values obtained from BLM_core K869A K870A unwinding dsDNA and G4-dsDNA were 22 ± 6.6 nM and 46 ± 8.8 nM, respectively, however standard error of the mean (SEM) was much higher for this mutant protein (Table 2). These values were determined to not be statistically significant when compared to the WT BLM_core dsDNA and G4-dsDNA K_BLM values (P values = 0.42 and 0.072), but this is likely due to the large SEM for BLM_core K869A K870A.

Molecular Dynamics simulations of BLM_core P868L

The BLM P868L allele has a 5.13% frequency in the human population and causes an increased frequency of sister chromatid exchanges in cell culture, making it essential to better understand how this substitution impacts BLM function. To examine effects of the P868L substitution on the structure and dynamics of the BLM_core complexed with partial dsDNA and ADP, we carried out six 1-μs long MD simulations, three of the WT BLM_core (control) and three of the BLM_core P868L mutant protein. First, we analyzed these simulations to examine the association of DNA with BLM. In the X-ray structure of WT BLM_core (PDB ID: 4CGZ (15)), the dsDNA (12 bp) interacts directly with the RecA-like lobe 2 of the ATPase domain (RecA-D2) and with the ZN and WH subdomains of the RQC domain of BLM. The single-stranded portion of the DNA (5 nucleotides) interacts directly only with RecA-D2. This general pattern of association between BLM and DNA is maintained in the MD simulations of the WT BLM_core. The one difference we note between the X-ray structure and MD simulation is that the distance between DNA and ZN is larger in MD simulations (Fig 5A). This is unlikely to be an artifact of the employed force field because we note such increased DNA-ZN distances even when the complex is simulated with an older version of the DNA-protein force field (amber-99sb-ILDN without DNA-protein NB-fix corrections (35)) in which protein-DNA binding is stronger. P868L substitution leads to one discernible change in DNA-BLM binding: while DNA-ZN and DNA-D2 distances are similar to those in the simulations of the WT form, the DNA-WH distance is larger (Fig 5A), implying that the P868L substitution could weaken interaction between the WH domain and DNA. This was not detected in our biochemical experiments, but this could be due to the qualitative nature of the assay or less of a reliance on WH-DNA interactions than other domains interactions with DNA for BLM_core DNA binding.

Fig 5. Effects of the P868L substitution on the structure and dynamics of the BLM_core protein complexed with DNA and ADP.

(A) Effect of the P868L substitution on the structure and dynamics of bound DNA. Top panels: Fluctuations in distances between centers of mass (CoMs) of DNA and BLM domains (ZN-zinc-binding domain, WH-winged helix domain, D2-RecA-like lobe 2 of the ATPase domain). The dashed vertical lines indicate CoM distances in the X-ray structure. Bottom panels: Sixty superimposed snapshots of DNA and CoMs of WH, ZN and D2 domains taken at regular intervals from MD simulations. DNA is shown using three line consisting of 5 nucleotides and two double stranded segments, each consisting of six base pairs. The CoMs of WH, ZN and D2 domains are connected by solid lines. The CoMs in each snapshot are determined after fitting the backbone atoms of D2 in that snapshot on to the D2 in the X-ray structure. To give the CoMs context, the X-ray structure is overlaid in cartoon form. (B) Effect of P868L substitution on the angle between the CoMs of ZN, D2, and D1 domains. The dashed vertical line on the figure on the left is the angle observed in the X-ray structure. The figure on the right compares the distribution of the domain CoMs (spheres) in 60 equally spaced snapshots extracted from MD simulations. The grey spheres are CoMs of domains in the WT BLM and those in other colors are CoMs of domains in the P868L mutant protein. (C) P868L induced shift (η) in conformational ensemble of the D2 domain. The X-ray structure of the D2 domain is shown as cartoon putty where the width of the putty as well as color of the putty denote the magnitude of η. Higher values of η imply larger effects of P868L induced shifts in conformational ensembles.

Other than BLM-DNA interactions, the P868L substitution also changes the orientation of the RecA-D1 relative to the rest of BLM in MD simulations. The P868L substitution reduces the angle between the centers of mass (CoMs) of the D1, D2 and ZN domains (Fig 5B). Since the sites for ATP-binding and hydrolysis are at the interface between RecA-D1 and RecA-D2, this result suggests potential effects of P868L on ATP hydrolysis and, thus, BLM activity. Changes were not observed in our biochemical assays, but this may be due to use of ssDNA instead of partial dsDNA. However, interpretation of ATPase rate data with partial duplex DNA in the assay is not simple, since ATPase activity would be coupled to DNA unwinding that would change the structure of the stimulating DNA from partial duplex to ssDNA during steady-state measurements.

To understand how P868L substitution alters RecA-D1 orientation, we examined the effect of P868L substitution on the conformational ensemble of RecA-D2. We used Direct Comparison of Ensembles (DiCE) (37–39), which computes the physical overlap between two conformational ensembles in high-dimensional space. It yields the difference in ensembles in terms of a quantity ‘η’ that is normalized to vary between 0 and 1. The larger the difference in ensembles, the closer the value of η is to 1. η is normalized with respect to amino acid size, which allows comparison of η between different amino acids. We found that for most amino acids η < 0.5 (Fig 5C). If conformational ensemble shifts were purely in mean positions, then η = 0.5 would correspond roughly to a shift in mean position by 1 Å. Therefore, amino acids with η < 0.5 were considered to have small induced shifts. We found that amino acids with η > 0.5 (the most affected with η ~ 0.7) are close to RecA-D1 (Fig 5C). This suggests that the signal that originates at the P868 substitution site at the periphery of RecA-D2 (21) could allosterically reach the catalytic cleft and the linker between the two lobes, explaining the potential effect of P868L on RecA-D1 orientation.

Discussion

BLM P868L and BLM G1120R both exist in the human population and are not associated with BS. However, both of these mutations cause increased SCEs and decreased recovery from hydroxyurea in human cells (22). While these phenotypes are not as severe as what is observed in BS-causing mutations, these mutations still lead to genome instability and could cause an increased risk of cancer or other negative health outcomes. Many BS-causing BLM mutant proteins have been investigated in vitro and have been found to have decreased helicase activity and ATPase activity (17). We hypothesized that these human BLM mutant proteins would lead to a moderate decrease in ATPase activity, DNA binding activity, and/or helicase activity.

Interestingly, BLM_core P868L and BLM_core G1120R are both active ATPases that have similar maximum ATPase rates to WT BLM_core and require similar concentrations of ssDNA for ATPase activity stimulation (Fig 2, Table 2). Additionally, both mutant proteins have ATPase activity even in the absence of DNA, similar to WT BLM_core. When assessing the ability of these mutant proteins to bind to ssDNA, G4 DNA, and partial dsDNA, both were able to bind substrates with similar affinities to WT BLM BLM_core While binding to ssDNA and G4 DNA was indicated by a smeared complex, binding to the partial duplex DNA induced a distinct band shift. The smearing observed with binding to ssDNA and G4 DNA substrates could indicate weaker overall binding and be caused by dissociation of BLM_core from the DNA while running in the gel. The distinct band shift observed from the partial duplex is consistent with WT BLM_core, BLM_core P868L, and BLM_core G1120R binding more stably to the partial duplex than the other DNA constructs.

We also assessed the ability of BLM_core P868L and BLM_core G1120R to unwind dsDNA and G4-dsDNA substrates. WT BLM_core, BLM_core P868L, and BLM_core G1120R were able to unwind dsDNA efficiently and required similar amounts of BLM to reach half-maximum unwinding (Table 2). Interestingly, BLM_core P868L was able to unwind G4-dsDNA more efficiently than WT BLM_core, indicating that this mutant protein is slightly better at unwinding G4s than WT BLM_core. This could indicate that BLM_core P868L has a tighter binding affinity for G4s than WT BLM_core, causing more proficient unwinding. However, direct binding studies did not demonstrate a higher affinity of BLM_core P868L for G4s than WT BLM_core (Table 2), so other properties such as processivity may be the source for this modest difference. BLM_core G1120R is less efficient at unwinding the G4-dsDNA substrate than the dsDNA substrate alone, requiring more BLM_core G1120R for half-maximum unwinding. This could be because this mutant protein is less efficient at unwinding G4s than dsDNA or because the BLM_core G1120R mutant protein is less processive on longer substrates.

Overall, the in vitro activities of BLM_core P868L and BLM_core G1120R are strikingly similar to that of WT BLM_core, despite these mutant proteins causing moderate genome instability in cells. While it has been observed that BS-inducing BLM mutant proteins have decreased helicase and ATPase activity in vitro, this is not the case for these non-BS human BLM_core mutant proteins (17). However, there are several other ways that these mutations could cause genome instability in cells. Many RecQ helicases, including BLM, are regulated by posttranslational modifications (PTMs), such as phosphorylation, acetylation, and SUMOylation (3). For example, BLM is phosphorylated in response to replication associated DNA damage or dsDNA breaks. Mutations that impact phosphorylation of BLM can decrease the ability of cells to recover from hydroxyurea exposure and other DNA damage (1). Residues 868 and 1120 are not known sites of PTMs, making it unlikely that this is the reason that these mutant proteins cause increased SCEs and decreased response to hydroxyurea induced damage (40). However, residue K873, which is part of the lysine rich loop proximal to P868 is a reported site of SUMOylation and ubiquitination (40). Because P868 terminates a β-strand at the start of the lysine-rich loop, it is possible that the mutation to leucine changes the structure of the loop and disrupts recognition of K873 by cellular PTM enzymes. This could result in decreased BLM activity in cells but more limited impact on in vitro activity of BLM.

Another possibility is that these mutations lead to increased SCEs in cells by producing BLM mutant proteins with disrupted dHJ dissolution activity and/or altered interaction with other proteins. BLM interacts with over 20 different proteins, including other dissolvasome proteins that are involved in dissolving dHJs (1,3,12). If the BLM dissolvasome function is impaired due to weaker interactions between BLM and Top3α, RMI1, or RMI2 because of these mutations, this could lead to increased genome instability. Additionally, if BLM mutations lead to a decreased affinity for HJs, this could be detrimental for BLM function in vivo. The G1120 residue is important for maintaining the WH α2-α3 loop and is conserved among many RecQ family helicases (21). Interestingly, substitutions in the α2-α3 loop, S1121A and K1125A, have been shown to induce a modest decrease in binding affinity for HJs in vitro, indicating that this loop could be important for HJ binding and recognition (41). While BLM G1120R functions similarly to WT BLM in our assays, the G1120R substitution could cause a decrease in BLM binding affinity for HJs. Thus, BLM G1120R might lead to decreased dissolvasome activity, and subsequently, increased SCEs in cells. The weakened interaction between DNA and the WH domain observed in the MD simulations could contribute in a similar manner to the partial loss of cellular function of BLM P868L. Additionally, the change in orientation of the RecA-D1 lobe, which affects the catalytic cleft of the BLM helicase core, as well as subtle conformational changes in conserved helicase motifs V and VI (Supplemental Fig S2), could contribute to the functional deficits of BLM P868L in yeast and in human cells (21,22). However, the effects of these changes may be too small to be detected by the in vitro biochemical assays used in this study.

We also investigated the effects of substitutions to the lysine-rich loop proximal to the P868 residue in the BLM_core K869A K870A construct. This BLM_core mutant protein had a 4-fold increase in ssDNA concentration required for half-maximum stimulation of ATPase activity and a loss of ATPase activity in the absence of DNA. While DNA is known to stimulate ATPase activity of RecQ helicases, it was surprising that BLM_core K869A K870A had no ATPase activity in the absence of DNA. It is important for helicases to couple ATP hydrolysis to their functions of DNA translocation and DNA unwinding, and it has been found in the bacterial RecQ that the aromatic-rich loop (ARL) (residues 798–808 in BLM) is important for coupling ATP hydrolysis to the unwinding activity of RecQ. Substitutions to the RecQ ARL produce mutant proteins that either require more DNA for maximum ATP hydrolysis or with increased DNA-independent ATPase activity (42). While the BLM_core K869A K870A substitutions are not part of the ARL of BLM, this loop might have a similar effect on regulating ATPase activity and ATP-coupled functions of BLM. Indeed, the MD simulations suggest that P868L has allosteric effects that alter the RecA-D1/D2 interface that forms the catalytic cleft of the BLM helicase domain. There could be multiple allosteric signaling pathways that connect the P868 site, and likely the lysine-rich loop, to the RecA-D1 lobe (43). Identifying these signaling pathways and hotspots will require additional simulations (39), and we expect this to be a target area for future studies.

In addition to the diminished ATPase activity, BLM_core K869A K870A has a decreased apparent affinity for ssDNA and G4 substrates. Further, BLM_core K869A K870A is deficient in stably and selectively binding to ssDNA-dsDNA junctions. While WT BLM_core band shifts partial dsDNA, BLM_core K869A K870A only causes a smear above the substrate. This smear is consistent with an overall reduced DNA binding stability for BLM_core K869A K870A. Additionally, at high concentrations of BLM_core K869A K870A, there is the appearance of a band at the size of the unannealed fluorescent-labeled ssDNA. This band appears to be the result of some amount of ATP-independent unwinding of the substrate from BLM_core K869A K870A. If the lysine-rich loop of BLM is important for linking ATPase activity of BLM to its other functions, such as helicase activity, it is possible that substitutions to this loop are able to decouple these activities and generate a mutant protein that has weak ATP-independent helicase activity. Despite this, BLM_core K869A K870A is unable to unwind the same amount of dsDNA or G4-dsDNA as WT BLM_core. BLM_core K869A K870A does not reach over 80% unwinding of either substrate and requires higher concentrations of protein to unwind both dsDNA and G4-dsDNA. This may be due to enhanced strand annealing in the mutant protein, as has been observed with BLM previously (44). BLM_core K869A K870A requires approximately 2-fold more protein to unwind the G4-dsDNA compared to the dsDNA substrate. This could be because this mutant protein has reduced G4 binding or unwinding or that this mutant protein is less processive than WT BLM_core.

From this work, we conclude that BLM_core P868L and BLM_core G1120R are competent helicases in the in vitro activities tested herein. However, since both of these mutant proteins lead to genome instability phenotypes in human cells, these mutant proteins are clearly defective in some activities that are essential for genome stability. Future studies should focus on assessing the protein interaction network, dHJ dissolution activity, and cellular localization of these mutant proteins. We have also shown that the lysine-rich loop of BLM is an important regulator of both ATPase activity and helicase activity of BLM. This linker could play a role in coupling ATP hydrolysis with the unwinding action of BLM and can provide a rich area for future BLM studies.