Human Pif1 helicase unwinds synthetic DNA structures resembling stalled DNA replication forks

Abstract

Pif-1 proteins are 5'-->3' superfamily 1 (SF1) helicases that in yeast have roles in the maintenance of mitochondrial and nuclear genome stability. The functions and activities of the human enzyme (hPif1) are unclear, but here we describe its DNA binding and DNA remodeling activities. We demonstrate that hPif1 specifically recognizes and unwinds DNA structures resembling putative stalled replication forks. Notably, the enzyme requires both arms of the replication fork-like structure to initiate efficient unwinding of the putative leading replication strand of such substrates. This DNA structure-specific mode of initiation of unwinding is intrinsic to the conserved core helicase domain (hPifHD) that also possesses a strand annealing activity as has been demonstrated for the RecQ family of helicases. The result of hPif1 helicase action at stalled DNA replication forks would generate free 3' ends and ssDNA that could potentially be used to assist replication restart in conjunction with its strand annealing activity.

Figure 1.

Purification and enzymatic activities of hPifHD. (A) Domain diagram of hPif. Amino acids 206–620 correspond to the conserved helicase domain. Relative positions of the seven conserved SF1 helicase motifs (I–VI) are indicated, and also motifs A, B and C (unknown function) characteristic of the same helicase family. (B) hPifHD fractions from a Superdex 75 column analysed on a 10% SDS-PAGE gel. The molecular weight of the purified protein corresponded to that predicted for hPifHD (45.5 kDa). The peak protein concentration was in fraction A11. (C) ATPase activity of fractions was determined in the presence of a 55-base poly T oligonucleotide and correlated with protein concentration. (D) Helicase activity of the peak fractions was determined using a ³²P labeled substrate with a 55 base T tail and a 20 bp duplex portion (PST55). S, native substrate; P, ssDNA product as determined by boiling the substrate. (E) Helicase activity was completely abolished by a mutation, E307Q, in the Walker B ATPase motif (S, substrate; P, single-stranded product), but the mutant retained wild-type ssDNA-binding activity (PD, protein–DNA complex; D, free DNA). (F) hPifHD unwinds DNA in the 5′→3′ direction; statistical data for 3 repeats.

Figure 2.

hPifHD DNA unwinding and DNA binding as a function of 5′ tail length. (A) ³²P-end-labeled substrates were generated with a 20-bp dsDNA portion and 5′ T tails of increasing length. The position of the label is indicated with a circle on the depiction of the substrate as in all subsequent figures. Reaction products (0.1 nM substrate; 2.5, 10 and 40 nM hPifHD) were resolved on poly-acrylamide gels and the extents of unwinding determined from quantified phosphorimages. Boil indicates the heat-denatured substrates. The graph on the right shows that unwinding increases with tail lengths up to 55 bases (data for 40 nM hPifHD, average of three repeats). (B) Binding of hPifHD to partially single-stranded substrates, as in (A) in the absence of ATP/Mg²⁺ Binding reactions (0.1 nM probe; 0.1, 2.5 and 5 nM hPifHD) were resolved on native poly-acrylamide gels. Two predominant complexes (C1 and C2) formed on probes with ssDNA tails of 15 bases or more, and a third formed at higher concentrations (C3). As the size of the probes increased, complexes became more difficult to resolve, particularly C1 from free probe. However, careful comparison of the lanes with no protein with those containing protein reveal distinct shifts, for example, lane 21 compared with lane 22, PST45.

Figure 3.

hPifHD interactions with ssDNA. (A) Binding to ³²P-end-labeled poly T probes of increasing length (0.1 nM probe; 0.1, 0.5 and 2.5 nM hPifHD) was assesed by gel-shift assay. Significant binding was detected only with probes of 35 bases or greater (lanes 9–28). (B) Stimulation of hPifHD ATPase activity as a function of probe length and concentration. hPifHD (50 nM) was incubated with oligonucleotides T8, T16, T30 and T55 at three concentrations, 25, 50 and 200 nM, and product release (from 5 µl of reaction) determined after 10 min. During this time period ATPase activity was in the linear range.

Figure 4.

Unwinding of fork substrates by hPifHD. (A) Fork-like substrates were generated comprising of a 20 bp duplex and a 55 base 5′ T tail (PST55/PST55-C0) and a 3′ poly C tail of increasing length (10, 20, 30 and 55 residues, from PST55-C10 to PST55-C55 respectively). Assays were performed as in Figure 1A. (B) Unwinding of gapped substrates by hPifHD. Unwinding substrates were designed with two 20 bp duplex regions of equivalent T_m, one with a 55 base 5′ T-tail, annealed to a complementary strand; an increasing length of unpaired T residues was introduced between the duplex portions on the bottom strand to generate gapped molecules. Assays were performed and analyzed as described for Figure 1A. Lanes 26–29 are markers for the single-stranded forked products produced. Statistical data show fork formation as a function of gap size (10 nM hPifHD, graph on the left) and a comparison of products formed for PST55-Gap20 (graph on the right). (C) The apparent unwinding reactions on the fork substrate are described as R1 (engagement of the free T55 5′ tail) and R2, engagement of the ssDNA of the gapped duplex.

Figure 5.

Strand annealing activity of hPifHD. Lanes 1–3, labeled component strands of the flap substrate PST55-Gap20. Lanes 4–8, reaction products for hPifHD action on the substrate PST55-Gap20. Lanes 9–12, annealing of the component oligonucleotides of substrate PST55-Gap20 catalysed by hPifHD. The long strands of the substrate are preferentially annealed. Lanes 13–17 and 18–22, annealing of individual pairs of the component oligonucleotides of PST55-Gap20. Lanes 23–27 and 28–32, hPifHD action on the substrate lacking the ssDNA flap oligonucleotide and the reference substrate PST55, respectively.

Figure 6.

Action of hPifHD on substrates with an extended duplex. The ability of hPifHD to unwind progressively longer stretches of duplex DNA was assessed using a population of substrate molecules with a 55-base 5′ T tail and a duplex portion increasing in 20 base increments. (A) Schematic diagram of substrate generation. Two pairs of phosphorylated oligonucleotides were annealed as illustrated. The oligonucleotide annealed to the strand with a 55 base T extension was phosphorylated with (γ-³²P)ATP. Double stranded oligonucleotides were ligated with a 5-fold molar excess of the non-tailed duplex oligo generating a series of products with a 55 base 5′ T tail and duplex portions increasing in 20 bp increments. (B) Unwinding products were resolved on a 6% (19 : 1) native poly-acrylamide gel. Lanes 1 and 2 show the substrates heat-denatured with and without formamide present. Lane 3 is the native substrate and lanes 4–6 unwinding product generated by hPifHD (2.5, 10 and 40 nM hPifHD). Lanes 7 and 8 are markers for the minimal substrate (55 bases 5′ ssDNA, 20 bp duplex), native and denatured respectively. The fractions (%) unwound for the 20, 40, 60 and 80 bp substrate molecules were 18, 8, 1.3 and 0% at 2.5 nM hPifHD; 41, 27, 8.5 and 1.8% at 10 nM hPifHD and 47, 43, 12.4 and 5% at 40 nM hPifHD.

Figure 7.

Binding and unwinding of stalled replication fork-like substrates by hPifHD. (A) A substrate (ds55-Gap20) was generated with two 20 bp duplex portions separated by 20 unpaired T residues with one extended to form a 55 bp dsDNA flap as illustrated. Each component oligonucleotide was end-labeled with ³²P (lanes 1–4), and all possible combinations of annealed oligonucleotide generated (lanes 5–10). The only products of hPifHD action on the substrate ds55-Gap20 resulted from displacement of the 20 base oligonucleotide (lanes 10–14). Lanes 15–19, substrate PST55. (B) Fork-like substrates were generated and analysed, as described in (A), except that the number of unpaired T residues on the bottom strand was varied from 0 (a nick) to 30 residues. Substrates with 20 and 30 unpaired T residues (ds55-Gap20 and ds55-Gap30) were unwound to similar extents. The substrate with 10 unpaired T residues was less efficiently unwound at lower protein concentrations, while the substrate with a nick was not unwound by the enzyme. (C) Direct comparison of ds55-Gap20 with the substrate lacking the 55 bp flap (Gap20), or the substrate with a 20 base T overhang (PST20), reveals the dependence of the dsDNA arm of the substrate for efficient unwinding. (D) Binding of hPifHD to gapped DNA molecules and comparison to other substrates. Binding of hPifHD (0.1, 0.5 and 2.5 nM) to labeled DNA probes (0.1 nM) was analyzed by gel-shift. hPifHD bound the probe PST0-Gap20 with 20 unpaired T residue gap at high protein concentrations more effectively than the 60 bp probe (ds60) or the 40 bp duplex with a 20 base 5′ T tail (T20-ds40; lanes 9–12 compared with 1–4 and 5–8). The addition of 15 5′ T residues to the gapped substrate to generate a fork-like molecule (PST15-Gap20) further increased hPifHD-binding activity to levels comparable with PST55 (lanes 17–20 compared with lanes 9–16). Addition of larger ss- or dsDNA flaps resulted in complexes that were difficult to resolve.