The 2B domain of the Escherichia coli Rep protein is not required for DNA helicase activity

Abstract

The Escherichia coli Rep protein is a 3′ to 5′ SF1 DNA helicase required for replication of bacteriophage φX174 in E. coli, and is structurally homologous to the E. coli UvrD helicase and the Bacillus stearothermophilus PcrA helicase. Previous crystallographic studies of Rep protein bound to single-stranded DNA revealed that it can undergo a large conformational change consisting of an ≈130° rotation of its 2B subdomain about a hinge region connected to the 2A subdomain. Based on crystallographic studies of PcrA, its 2B subdomain has been proposed to form part of its duplex DNA binding site and to play a role in duplex destabilization. To test the role of the 2B subdomain in Rep-catalyzed duplex DNA unwinding, we have deleted its 2B subdomain, replacing it with three glycines, to form the RepΔ2B protein. This RepΔ2B protein can support φX174 replication in a rep⁻ E. coli strain, although the growth rate of E. coli containing the repΔ2B gene is ≈1.5-fold slower than with the wild-type rep gene. Pre-steady-state, single-turnover DNA unwinding kinetics experiments show that purified RepΔ2B protein has DNA helicase activity in vitro and unwinds an 18-bp DNA duplex with rates at least as fast as wild-type Rep, and with higher extents of unwinding and higher affinity for the DNA substrate. These studies show that the 2B domain of Rep is not required for DNA helicase activity in vivo or in vitro, and that it does not facilitate DNA unwinding in vitro.

DNA helicases are a ubiquitous class of enzymes that use the binding and hydrolysis of nucleoside triphosphates to catalyze unwinding of the two strands of duplex DNA to transiently form the single-stranded DNA (ssDNA) intermediates required for replication, recombination, and repair (1). Mutations in DNA helicases are also linked to several human genetic disorders (2). Although a large number of enzymes with helicase activity have been identified, our understanding of the mechanisms by which these enzymes unwind DNA is still at an early stage.

The Escherichia coli Rep protein is a 3′ to 5′ SF1 superfamily DNA helicase that has been characterized both structurally (3) and biochemically (reviewed in ref. 4). In solution, this protein is monomeric in the absence of DNA, whereas binding of either ssDNA or duplex DNA induces the protein to dimerize (5). Biochemical studies indicate that Rep dimerization influences both its DNA binding and ATPase activities (6–9), which has led to proposals that Rep dimerization may be important for its helicase and/or translocation functions (10, 11). Indeed, it has been demonstrated that Rep monomers are unable to initiate DNA unwinding, and that Rep oligomerization is required for its helicase activity in vitro (12).

The E. coli Rep protein is structurally homologous to both the Bacillus stearothermophilus PcrA protein (13) and the E. coli UvrD protein (S.K., N. K. Maluf, T.M.L., and G.W., unpublished results), both of which are also 3′ to 5′ SF1 helicases. Each of these proteins is composed of two domains, 1 and 2, that are each further composed of two subdomains (A and B; Fig. 1A). As shown in Fig. 2, structures of Rep complexed with ADP and ssDNA [(dT)₁₆] indicate that the ADP is bound in the cleft formed between the 1A and 2A subdomains, and that ssDNA interacts with amino acids contained within domains 1A, 1B, and 2A (3). The 2B domain is the only domain that does not contain any of the conserved amino acids that comprise the seven so-called “helicase motifs” that define the SF1 helicase superfamily. In the Rep crystal structure (3), the asymmetric unit is composed of two molecules of Rep bound to (dT)₁₆. Although the two Rep monomers display the same major contacts with ssDNA, they exist in two conformations, referred to as the “open” and “closed” forms, which differ in the orientation of the 2B subdomain, relative to the 2A domain. These two conformations are shown superimposed in Fig. 2 and can be interconverted by an ≈130° rotation of the 2B domain about a hinge region connected to the 2A domain. Domains 1A, 2A, and 1B remain essentially unchanged between the two conformations. Interestingly, early studies indicated that on ssDNA binding, a site within Rep becomes hypersensitive to cleavage by trypsin (14), and we know now that this site lies within the hinge region (3). This finding suggested that some motion of the 2B domain occurs on ssDNA binding, leading to the hypothesis that movement of the 2B domain might be functionally relevant for its helicase activity. It was further hypothesized that the 2B domain might be involved in dimerization of the Rep protein (3).

Fig 1.

Construction of the RepΔ2B mutant and purified RepΔ2B proteins. (A) Ribbon diagrams of the wtRep crystal structure in the “open” conformation (3) and the hypothetical structure of RepΔ2B. The four subdomains within wtRep are 1A (yellow), 1B (green), 2A (red), and 2B (blue). In RepΔ2B, amino acids Thr-375 to Arg-542, constituting the 2B subdomain, were deleted and replaced by three Gly residues (shown in blue). For the hexahistidine-tagged RepΔ2B, the location of the N-terminal histidine tag is marked in black in the hypothetical RepΔ2B structure. The organization of the subdomains within the primary structure is shown below each structure. (B) SDS-polyacrylamide (12%) gel of the purified wtRep and RepΔ2B proteins (1 μg each).

Fig 2.

Ribbon diagrams showing the superposition of the two conformations (“open” and “closed”) of the wtRep protein in complex with ssDNA and ADP. Six nucleotides are shown in purple, and the ADP is shown in orange [modified from Korolev et al. (3)]. The two conformations differ in the orientation of the 2B domain, shown in light blue for the open form and deep blue for the closed form. The hinge region connecting the 2B domain to the 2A domain, and about which the 2B domain rotates by ≈130° to convert from one form to the other, is shown in yellow for the open form and red for the closed form.

Crystal structures of the PcrA protein in complex with a short ss/dsDNA junction (3′-7 nucleotide ssDNA attached to a 10-bp duplex; ref. 15) showed a monomer bound to the junction, with the ssDNA in the same position as was observed for the ssDNA in the Rep structure (3). The orientation of the 2B domain in the crystal structure of the apo PcrA protein (13) was in the “open” conformation, whereas the orientation of the 2B domain in the PcrA–DNA complex was closer to that of the “closed” Rep conformation (3). Based on these structural studies and subsequent site-directed mutagenesis studies (16), it was proposed that the 2B domain forms part of the duplex DNA binding site within a PcrA monomer, and that the 2B domain plays an essential role in duplex destabilization during DNA unwinding by the PcrA helicase (16).

To examine the functional importance of the 2B domain in Rep, we have constructed a deletion mutant, RepΔ2B, in which amino acids Thr-375 to Arg-542, comprising the 2B domain in wtRep, were replaced with three glycines as shown schematically in Fig. 1. We have examined the effect of this deletion on the growth rate of rep⁻ E. coli and its ability to support φX174 phage replication, and have examined the duplex DNA unwinding (helicase) activity of purified RepΔ2B in vitro.

Materials and Methods

Buffers.

Buffers were made with reagent-grade chemicals, using glass-distilled water that was further deionized using a Milli-Q System (Millipore). Buffer U is composed of 20 mM Tris⋅HCl, 6 mM NaCl, 1.7 mM MgCl₂, 5 mM 2-mercaptoethanol (2-ME), and 10% (vol/vol) glycerol (Aldrich), titrated to pH 7.5 at 25°C. Descriptions of the other buffers can be found in Supporting Methods, which is published as supporting information on the PNAS web site, www.pnas.org.

E. coli and Phage Strains and Plasmids.

E. coli CK11Δrep, has the chromosomal copy of the rep gene deleted (17). E. coli BL21(DE3) and the plasmid pET28a were from Novagen. ΦX174 bacteriophage (strain K9) was from P. Burgers (Washington University). The plasmid pRepO expresses wtRep protein under control of the temperature-inducible λP_L promoter (18). The plasmid pIWcI encodes the temperature-sensitive λcI857 repressor. The plasmid pGroESL (from C. Frieden, Washington University) overproduces the chaperonin GroES and GroEL.

RepΔ2B Plasmids.

Plasmids were constructed by established cloning methods (19). DNA primer sequences used in PCR are given in Supporting Methods. The N-terminal portion of the wtRep ORF from plasmid pRepO was amplified using primers GHG12 and GHG21, digested with NdeI and HindIII, and ligated into pET28a to generate subclone 1. The C-terminal portion of the wtRep ORF of pRepO was amplified using primers GHG8 and GHG25, and digested with HindIII and XhoI. This HindIII-XhoI restriction fragment was ligated into subclone 1 to create pGG206, encoding wtRep possessing an N-terminal hexa-histidine tag (+HRep). This histidine tag can be removed by cleavage with thrombin, yielding wtRep with an additional three amino acids (GSH) attached to its N terminus. A PCR overlap extension (20) was used to remove the codons encoding amino acids Thr-375 to Arg-542 in pGG206. Briefly, we generated two PCR fragments from the wtRep ORF: fragment 1 encodes the N-terminal portion of wtRep including domains 1A and 1B, and the N-terminal half of the 2A domain, and was amplified using primers GHG25 and GHG12; fragment 2 encodes the C-terminal half of Rep containing domain 2A, and was amplified using primers GHG13 and GHG8. Fragments 1 and 2 were then used in a third round of PCR to generate fragment 3, using primers GHG25 and GHG8. Fragment 3 was digested with HindIII and XhoI to yield a 714-bp fragment, which was ligated into pGG206, generating plasmid pGG205. Plasmid pGG205 encodes for +HRepΔ2B in which the 2B domain (Thr-375 to Arg-542) has been deleted and replaced with three Gly residues. The +HRepΔ2B has an N-terminal hexa-histidine tag with a thrombin cleavage site (indicated by the arrow): MGSSHHHHHHSSGLVPR↓GSH. After thrombin cleavage, three extra residues (GSH) remain on the N terminus of −HRepΔ2B protein. To generate an untagged RepΔ2B ORF, pGG205 was digested with BstXI and MluI to yield a 683-bp fragment, which was cloned into pRepO digested with BstXI and MluI, generating plasmid pRepOΔ2B, which expresses RepΔ2B protein with an N terminus identical to wtRep, under the control of λP_L promoter. All ORFs were confirmed by DNA sequencing.

Protein Purification.

Rep protein was purified to >99% homogeneity as described (18) and its concentration determined spectrophotometrically (ɛ₂₈₀ = 7.68 × 10⁴ M⁻¹⋅cm⁻¹) (21). His-tagged RepΔ2B (+HRepΔ2B) was purified from BL21(DE3) cells carrying pGG205 and pGroESL. The purification protocol is given in Supporting Methods. +HRepΔ2B protein concentration was determined spectrophotometrically in 6 M Guanidine HCl/20 mM Tris⋅Cl/0.1 M NaCl (pH 7.5) at 25°C, using an extinction coefficient, ɛ₂₈₀ = 4.58 × 10⁴ M⁻¹⋅cm⁻¹, calculated from its amino acid composition (4 Trp, 18 Tyr) (22). The extinction coefficient of native +HRepΔ2B protein is ɛ₂₈₀ = (4.88 ± 0.18) × 10⁴ M⁻¹⋅cm⁻¹ in 10 mM Tris⋅Cl/500 mM NaCl/20% vol/vol glycerol (pH 8.3) at 25°C. The histidine tag of +HRepΔ2B protein was cleaved to yield −HRepΔ2B by using restriction grade thrombin (Novagen), as described in Supporting Methods. RepΔ2B protein was purified from CK11Δrep/pIWcI/pRepOΔ2B by using a slight modification of the wtRep purification (18) and is described in Supporting Methods.

DNA Substrates.

Oligodeoxynucleotides were synthesized and purified to >99% homogeneity and their concentrations determined as described (12). Oligodeoxynucleotides were radiolabeled with ³²P at the 5′ end by using T4 polynucleotide kinase (United States Biochemical), and purified as described (23). DNA substrates contain an 18-bp duplex region with a 3′-(dT)₂₀ ssDNA tail (all oligothymidylates) and were prepared as described (12). The base sequence of the “top strand” of the 18-bp duplex is 5′-GCCTCGCTGCCGTCGCCA-3′ and is the same as used in previous experiments (12).

Single-Turnover (STO) DNA Unwinding Kinetics.

STO DNA unwinding experiments were carried out using a quenched-flow apparatus (RQF-3, Kintek, University Park, PA) at 25.0°C in Buffer U, essentially as described (12). Briefly, helicase protein was preincubated with ³²P-labeled DNA substrate (2 nM) in Buffer U containing BSA (0.1 mg/ml) on ice for 20 min, and then loaded in one loop of the quenched-flow apparatus. The other loop contained ATP (3 mM) and an excess of an unlabeled 10-bp hairpin DNA (8 μM) possessing a 3′-(dT)₄₀ tail (mock DNA substrate, HP10T40) in Buffer U that served as a trap to prevent rebinding of free protein to the DNA substrate. Control experiments indicated that a final concentration of 4 μM HP10T40 is sufficient to trap 800 nM free protein. This insures that these are STO DNA unwinding experiments (single round of DNA unwinding). The DNA and protein concentrations reported are the preincubation concentrations before mixing with ATP. Reactions were initiated by rapidly mixing the two solutions and then quenched by addition of 400 mM EDTA in 10% (vol/vol) glycerol at time intervals from 2 ms to 120 s after mixing. Quenched samples were analyzed after separation of the duplex substrate from the ssDNA product by electrophoresis in nondenaturing 10% polyacrylamide gels (PAGE). The radioactivity in each band was quantitated using a Storm 840 PhosphorImager (Molecular Dynamics). DNA unwinding time courses were analyzed using Eq. 1, derived from Scheme (ref. 12; A. L. Lucius, N. K. Maluf, C. J. Fischer, and T.M.L., unpublished work),

where F(t) is the fraction of DNA molecules unwound at time t, A_T is the total amplitude of DNA unwinding, n is the number of steps in the unwinding reaction, and k_obs = (k_u + k_d).

Scheme 1.

Results

RepΔ2B Protein Supports ΦX174 Phage Replication.

Although Rep protein is not essential for the viability of E. coli under laboratory growth conditions, Rep protein is the only E. coli helicase that functions to unwind the double-stranded replicative form of ΦX174 phage DNA required for replication of that phage (24–26). We therefore tested whether RepΔ2B protein can support ΦX174 DNA replication in vivo by performing ΦX174 phage plaque assays, using CK11Δrep/pIWcI, which carries a deletion of the wild-type rep gene, as the host. As summarized in Table 1, no plaques were formed when CK11Δrep/pIWcI was used, reflecting the inability of ΦX174 phage to replicate in a rep⁻ E. coli host. However, CK11Δrep/pIWcI carrying either of the plasmids, pRepO or pRepOΔ2B, encoding wtRep protein or the RepΔ2B protein, respectively, produced essentially the same numbers of plaque-forming units (pfu). The only difference was that the plaque size was slightly smaller when assayed with CK11Δrep/pIWcI containing pRepΔ2B than with the same host containing pRepO. This likely reflects the fact that the growth rate of CK11Δrep/pIWcI/pRepΔ2B is slower than the growth rate of CK11Δrep/pIWcI/pRepO (see below). Nonetheless, these results indicate that RepΔ2B can function as a helicase in ΦX174 DNA replication in vivo.

Table 1.

ΦX174 plaque assay

Phage dilution	No. of plaques (pRepO)	No. of plaques (pRepOΔ2B)
10−6	71  ± 10	60  ± 8
10−7	7  ± 2	6  ± 2

The plaque numbers are averaged from eight independent assays.

Effect of RepΔ2B on E. coli Growth.

Although the exact roles of Rep helicase in E. coli have not been fully defined, it is clear that Rep plays a role in DNA replication (17, 27, 28), possibly in replication restart (29). We therefore compared the growth rates of E. coli expressing wtRep vs. RepΔ2B. Growth rates were determined for CK11Δrep/pIWcI/pRepOΔ2B, CK11Δrep/pIWcI/pRepO, and CK11Δrep/pIWcI at 30°C in LB media, conditions under which overexpression of Rep and RepΔ2B has not been induced. CK11Δrep/pIWcI/pRepOΔ2B has a generation time of 58 ± 4 min, which is slower than that of CK11Δrep/pIWcI/pRepO (40 ± 4 min), but the same as that of CK11Δrep/pIWcI alone (see Fig. 4, which is published as supporting information on the PNAS web site). The slower growth rate of cells expressing RepΔ2B as compared with wtRep suggests that the 2B domain of wtRep does play some role in E. coli, although we can conclude little beyond this.

DNA Helicase Activity of RepΔ2B in Vitro.

To characterize the effect of removal of the 2B domain on Rep helicase activity in vitro, we purified and characterized three versions of the RepΔ2B protein: +HRepΔ2B with a histidine tag added to the N terminus, −HRepΔ2B obtained after thrombin cleavage of the histidine tag from +HRepΔ2B (which retains three additional amino acids, GSH, fused to the N terminus), and RepΔ2B without any tag. All three were purified to >98% purity as described in Materials and Methods, and their helicase activities were compared with wtRep protein by performing STO DNA unwinding kinetics experiments (30), using an 18-bp DNA substrate with a 3′-(dT)₂₀ tail, as depicted in Fig. 3. As described in Materials and Methods, in these STO experiments, only protein that is prebound to the DNA can participate in DNA unwinding. Although the quenched-flow assay is an “all or none” DNA unwinding assay, detecting only fully unwound DNA, the time courses of these experiments can yield the rates of the rate-limiting steps that are repeated during the course of DNA unwinding (30).

Fig 3.

STO kinetic studies of Rep-catalyzed DNA unwinding of an 18-bp DNA substrate with a 3′-(dT)₂₀ tail examined using rapid chemical quenched-flow methods. (A) Comparisons of the fraction of DNA molecules unwound as a function of time for wtRep (○), +HRepΔ2B (▪), and RepΔ2B (□). All time courses were performed in Buffer U at 25°C, with preincubation concentrations of 30 nM (protein) and 2 nM (DNA). The time courses were fit individually to Eq. 1, constraining the number of steps n = 4, to obtain estimates of k_obs and k_NP, and the extent of unwinding, A_T. The solid lines are simulations using Eq. 1 and the best fit parameters determined from nonlinear least squares analysis. (B) The total unwinding amplitudes, A_T, as a function of protein concentration for wtRep (○), +HRepΔ2B (▪), and RepΔ2B (□).

Fig. 3A shows individual time courses for wtRep, +HRepΔ2B, and RepΔ2B at a protein concentration of 30 nM. The results with −HRepΔ2B were identical to those obtained with RepΔ2B (data not shown). All time courses were biphasic, as observed previously for both UvrD (23, 30) and wtRep (12). The lag phase reflects the transient formation of partially unwound DNA intermediates along the path to complete DNA unwinding, and depends on the duplex DNA length. The presence of the second, slower phase suggests that the helicase protein is prebound to the DNA substrate in at least two different states, consistent with our previous experiments (23). These time courses were analyzed using Scheme , which assumes that the helicase can be bound to the DNA in two forms, a productive from, (P-DNA)_L, and a nonproductive from, (P-DNA)_NP. The productively bound form unwinds DNA in a series of rate-limiting steps with rate constant k_u. The nonproductively bound helicase must first isomerize with rate constant k_NP to form the productive form, which can then initiate DNA unwinding. Because we have not yet performed experiments as a function of duplex length with these helicases, which is necessary to estimate the number of intermediate steps, n, needed to fully unwind the 18-bp duplex (see Scheme ), we have analyzed each time course by assuming a value of n = 4, based on our studies of DNA unwinding by E. coli UvrD (23, 30), which is structurally homologous to wtRep.

The solid lines describing these time courses were generated using the best fit parameters obtained from nonlinear least squares fitting of each unwinding time course to Eq. 1. Fig. 3 shows that both the observed rates and the extents of unwinding for RepΔ2B and +HRepΔ2B are significantly greater than for wtRep. The values for k_obs (=k_u + k_d; assuming n = 4) are 118.8 ± 16 s⁻¹ for RepΔ2B, 42.0 ± 3.2 s⁻¹ for +HRepΔ2B, and 17.9 ± 5.4 s⁻¹ for wtRep. Interestingly, unwinding by RepΔ2B is also faster than +HRepΔ2B, which indicates that the histidine tag has a slightly inhibitory effect under these conditions. At protein concentrations above 100 nM, the observed rates of unwinding by RepΔ2B decrease and become similar to those observed for wtRep at 17 ± 5.4 s⁻¹ (see Fig. 5, which is published as supporting information on the PNAS web site). This effect of protein concentration is currently under study, but may indicate an effect of different oligomeric forms of RepΔ2B on DNA unwinding. Fig. 3B shows the dependence of the extent of DNA unwinding (fraction of DNA molecules unwound) on helicase concentration. The total amplitudes of DNA unwinding (fast and slow phases) for both +HRepΔ2B and RepΔ2B are much greater than for wtRep at all protein concentrations. The midpoint concentrations obtained from fitting the unwinding amplitudes to a hyperbola are also much lower for RepΔ2B (3.5 ± 0.6 nM) and +HRepΔ2B (8.5 ± 3.0 nM), compared with wtRep (57 ± 20 nM).

Discussion

We have examined the effect of removal of the 2B subdomain from the E. coli Rep protein on its DNA helicase activity by constructing a mutant, RepΔ2B, in which amino acids Thr-375 to Arg-542 that comprise the 2B subdomain were deleted and replaced with three glycines. Our studies show that the 2B subdomain of Rep is neither required for its DNA helicase activity in vitro, nor for its function in φX174 phage replication in vivo. In fact, the RepΔ2B protein can unwind DNA with rates that are faster than those catalyzed by wtRep. Unwinding by RepΔ2B is also more efficient, due at least in part to a higher apparent binding affinity for the DNA substrate. Based on the observed inhibitory effect of the 2B subdomain on Rep helicase activity in vitro, it is possible that this subdomain plays a role in regulating Rep helicase function in vivo or is involved in interactions with accessory proteins. The only observed effect of deletion of the 2B subdomain of Rep is that the growth rate of E. coli cells expressing the RepΔ2B mutant is reduced somewhat relative to cells expressing a wild-type rep gene. Hence, removal of the 2B subdomain of Rep clearly has some consequences for E. coli metabolism.

This result was unexpected for several reasons. First of all, crystal structures of the E. coli Rep protein bound to ssDNA indicate that the 2B subdomain of Rep can exist in two major orientations (“open” vs. “closed”; ref. 3). Furthermore, some movement of the 2B subdomain appears to accompany ssDNA binding of Rep, based on the ssDNA-binding-induced appearance of a hypersensitive site for trypsin hydrolysis (14) that maps near the 2B subdomain hinge region (3). Similarly, crystal structures of the apo form of the B. stearothermophilus PcrA protein show the protein in the open form, whereas when bound to a ss/dsDNA junction, the protein is in a conformation that more closely resembles the closed form (15).

Implications for Mechanisms of DNA Unwinding.

The fact that the 2B subdomain of Rep is not required for its DNA helicase functions has implications for some of the models that have previously been proposed to explain DNA unwinding by E. coli Rep and the structurally homologous B. stearothermophilus PcrA. In particular, two proposals have been made concerning the possible functions of the 2B subdomain. Korolev et al. (3) suggested that the 2B subdomain may mediate dimerization of Rep protein (3), which is needed for DNA unwinding in vitro (12). A second proposal, based on structural and mutagenesis studies of the B. stearothermophilus PcrA protein, is that a PcrA monomer is the active form of the helicase and that the 2B subdomain of PcrA forms a part of its duplex DNA binding site that is essential for DNA unwinding activity (16).

Cheng et al. (12) have shown that monomers of wtRep protein do not possess DNA helicase activity in vitro; in fact, Rep monomers are unable to even initiate DNA unwinding by themselves, and DNA unwinding in vitro requires Rep oligomerization (12). In fact, recent single-molecule experiments indicate that wtRep monomers are unable to continue unwinding when the Rep oligomer dissociates during the course of DNA unwinding (31). This supports the proposal that a Rep dimer is the minimal form required for helicase activity in vitro. Although an active, rolling model was first proposed as a mechanism for dimeric Rep helicase activity (10), all of our data are also consistent with a dimeric inch-worm model in which the same Rep subunit remains as the leading subunit (12). However, the regions of Rep involved in oligomerization are not known, and the hypothesis was made that the 2B subdomain might provide the site for oligomerization (3). Although the results of this study indicate that the 2B domain of Rep is not required for helicase activity, it still remains possible that this domain is involved in wtRep oligomerization. An important question that remains to be answered in this regard is whether the RepΔ2B monomer can function as a helicase, or whether oligomerization is also required for its helicase activity. For example, if the RepΔ2B monomer can function as a helicase, then it is possible that the 2B domain in the wtRep monomer blocks its ability to initiate DNA unwinding. In this case, interactions between the monomers (i.e., oligomerization) might function to relieve this inhibition by affecting the relative orientation of the 2B domain. However, if the helicase activity of the RepΔ2B protein also requires oligomerization, then such oligomerization clearly cannot involve the 2B domain. Experiments to investigate these alternatives are currently underway.

Our results also have important implications for the mechanisms of closely related helicases, such as the B. stearothermophilus PcrA and E. coli UvrD. Although the 2B subdomain of Rep is not required for its DNA unwinding activity, it is possible, in principle, that the 2B subdomain of the structurally homologous PcrA helicase is required for duplex DNA binding and unwinding as proposed by Soultanas et al. (16). However, it also may be that the 2B subdomain of PcrA does not play a functional role in duplex DNA binding and DNA unwinding, but rather inhibits productive DNA binding and initiation of DNA unwinding as we observe for wtRep in vitro. The proposal that the 2B domain of PcrA is essential for duplex DNA binding and unwinding was based partially on the interpretation that in the crystal structures of a complex of PcrA with a 3′-7 nucleotide ss/dsDNA (10 bp) junction, the duplex region appears to contact regions of the 2B and 1B subdomains (15). Subsequent mutagenesis of amino acids within the region of proposed contact in the 2B subdomain was found to inhibit DNA unwinding rates by PcrA in multiple-turnover (MTO) experiments (16). In this regard, it is important to recognize that it is not possible to conclude from such MTO unwinding experiments whether the PcrA mutations affect the actual steps involved in DNA unwinding and/or translocation. The interpretation of such MTO DNA unwinding experiments is not straightforward because an effect on the macroscopic rate of production of ssDNA could result from effects on any steps in the MTO pathway, including the rates of PcrA binding to or dissociation from the DNA or any protein oligomerization, if required, rather than effects on the actual unwinding steps. This is why a STO experiment is required to assess the actual effects on DNA unwinding.

If, in fact, the 2B subdomain of PcrA does not play a functional role in duplex DNA binding and DNA unwinding, but rather inhibits productive DNA binding and initiation of DNA unwinding, then it is possible that the PcrA mutations studied by Soultanos et al. (16) actually make the 2B domain more inhibitory, yielding the observed reduction in the rate of ssDNA production in a MTO unwinding experiment. STO DNA unwinding kinetics experiments, such as those described here, are needed to examine the effects of these mutations on PcrA-catalyzed DNA unwinding. Because there is significant sequence and structural homology between E. coli Rep and B. stearothermophilus PcrA, it is possible that removal of the entire 2B domain from PcrA would also generate a more efficient helicase in vitro. If in fact the 2B domain of PcrA is inhibitory, then it is possible that the complex of PcrA bound to the ss/dsDNA junction observed in the crystal structure (15) may reflect an inhibited complex, rather than one that is competent to unwind DNA. This would be consistent with the observation that a monomer of Rep bound to a DNA substrate is not competent to initiate DNA unwinding in vitro (12).

Although our results with Rep indicate that the 2B subdomain is not required for DNA helicase activity, we note that several mutations have been identified within the 2B subdomain of E. coli UvrD, a structural homologue of both Rep and PcrA that functions in methyl-directed mismatch repair (32) and nucleotide excision repair (33), and these mutations may influence its helicase activity (34). Alteration of the amino acid sequence DDAAFER located within the 2B subdomain between residues 403 and 409 results in temperature sensitivity and dominant lethality, and affects conjugal recombination (34). In particular, the double mutant, D403AD404A, produces a more efficient helicase than wild-type UvrD in a MTO unwinding experiment in vitro (34). Based on these results, it was proposed that the region between amino acids 403 and 409 within the 2B subdomain of UvrD helicase serves to regulate its DNA unwinding activity (34). Interestingly, our attempts to generate a UvrDΔ2B mutant by making the same deletion as we made with Rep have been unsuccessful. After growth of E. coli transformed with a plasmid that encodes UvrDΔ2B, all plasmids that were recovered had additional sequences deleted from the ORF of the gene encoding for UvrDΔ2B. One of these plasmids was sequenced and found to encode a uvrDΔ2B gene with an additional deletion of amino acids 188–219. In wild-type UvrD, position 220 is the conserved aspartic acid residue of motif II that is essential for ATP hydrolysis. Although the exact reason for the inability to express a UvrDΔ2B in E. coli is not known, one possibility is that removal of the 2B subdomain of UvrD makes a more active and unregulated helicase that is lethal in E. coli. This result and those observed for Rep suggest that the 2B subdomains of these proteins may function to regulate helicase activity, perhaps through interactions with accessory proteins.

The 2B domain of Rep occurs as an insertion between motifs IV and V, two of the seven conserved helicase motifs defining the SF1 helicase superfamily (35). However, the 2B domain does not contain any of the conserved helicase motifs. It was reported for PcrA helicase that site-directed mutagenesis of several residues in the 2B domain affects the apparent affinity of the protein for duplex DNA (16); however, only one of the residues (T-426 in PcrA) is also present in both Rep and UvrD. In addition, a comparison of the crystal structures of Rep and the helicase domain of the NS3 RNA helicase encoded by hepatitis C virus indicates that NS3 does not contain an equivalent 2B domain (36). The size of the insertions between motifs IV and V for a number of SF1 and SF2 helicases varies significantly with no consensus sequence across the superfamilies (35). The 2B domain is generally larger than 200 aa in SF1 helicases, with the exception of E. coli helicase IV, which is only 87 aa, whereas it is generally much smaller for the SF2 helicases (see Fig. 6, which is published as supporting information on the PNAS web site). These observations are also consistent with the proposal that the 2B insertion plays a regulatory role and/or is involved in interactions with other proteins, rather than being necessary for DNA helicase activity, although it is possible that these enzymes unwind by using different mechanisms.

Finally, we have noted that the hexa-histidine tag placed at the N terminus of RepΔ2B partially inhibits its DNA unwinding activity. Although further experiments are required to address the mechanism of inhibition, this result emphasizes the need for caution in using these types of recombinant proteins.

Abbreviations

• ssDNA, single-stranded DNA

• dsDNA, double-stranded DNA

• STO, single turnover

• MTO, multiple turnover