ABSTRACT
Limited proteolysis of bacteriophage T7 primase/helicase with endoproteinase Glu-C produces several proteolytic fragments. One of these fragments, which is derived from the C-terminal region of the protein, was prepared and shown to retain helicase activity. This result supports a model in which the gene 4 proteins consist of functionally separable domains. Crystals of this C-terminal fragment of the protein have been obtained that are suitable for X-ray diffraction studies.
The replication of DNA requires a number of proteins in addition to DNA polymerase (1 ). Amongst these are DNA primase, which initiates DNA replication by synthesis of RNA primers, and at least one DNA helicase. DNA helicases utilize the energy of NTP hydrolysis to unwind the DNA duplex (2 ). Helicases have been characterized biochemically according to their nucleic acid substrate specificity. There are two broad classes of enzyme based on the requirement for either a 3' or a 5' flanking region of single-stranded nucleic acid adjacent to the duplex to be unwound. The proteins are known as 3' -> 5' or 5' -> 3' helicases respectively. Helicases have been classified further into at least five families based on sequence alignment (3 ).
The bacteriophage T7 chromosome provides a simple in vitro model system for DNA replication, requiring just five proteins: the phage encoded T7 DNA polymerase (gene 5) and T7 primase/helicase (gene 4, gene products 4A and 4B), single-stranded DNA binding protein (gene 2.5) and Escherichia coli thioredoxin (4 -6 ).
Gene 4 of bacteriophage T7 encodes two polypeptides of 56 and 63 kDa. The 56 kDa product (termed 4B) is translated from an internal initiation codon (Met64) in-frame within the coding sequence for the 63 kDa product (termed 4A; 7 ). Both proteins are 5' -> 3' helicases, but 4A also has primase activity (8 ,9 ). The difference between the two proteins is the presence of a Cys4 zinc binding motif in 4A, which is proposed to function in recognition of priming sites (8 ). Mutation of Met64 to Leu in 4A (termed 4A') demonstrated that 4A alone is sufficient to support bacteriophage T7 DNA replication (10 ). In contrast, in E.coli the primase and helicase activities are two separate proteins (DnaG and DnaB respectively). The primase is 64 kDa and the replicative helicase is 52 kDa (1 ). However, there is an association of the two proteins (11 ,12 ).
Alignment of the protein sequence of 4A with those of E.coli DnaB and DnaG is suggestive of two functional domains in 4A (13 ). The N-terminus of 4A shows homology with DnaG and the C-terminal region with DnaB. Moreover, mutation of the `Walker A' nucleotide binding site in 4A destroys helicase activity but does not completely abolish primase activity (14 ), suggesting a degree of separation of helicase and primase activities. Low resolution structural data, obtained using electron microscopy, have recently shown that there are two structural domains for each monomer in a protein complex that consists of a hexameric ring (15 ). DnaB, which is a member of the same family of helicases as the gene 4 proteins, is also a hexameric ring consisting of monomers with a bilobal structure (16 ). Limited proteolysis of DnaB with trypsin released a C-terminal fragment (17 ). Interestingly, all the motifs suggested as being necessary for helicase activity are in this fragment (3 ,18 ). This suggests that there may be a common structural domain for helicases of the DnaB family.
In order to understand the differences between the structures and mechanisms of both classes of DNA helicase we have undertaken high resolution structural studies. We have recently solved the structure of a 3' -> 5' helicase, PcrA, from Bacillus stearothermophilus (19 ). In this paper we have used limited proteolysis to investigate whether there are functional domains in the bacteriophage T7 gene 4A protein and thus define a minimal 5' -> 3' helicase for high resolution structural studies. We describe the subsequent cloning and expression of fragments of 4A that are helicases and crystallization of one of these proteins. The results obtained support a model where the 4A protein does indeed consist of functionally discrete domains of defined biochemical function.
Escherichia coli MC1061 F-, araD139, [Delta](ara-leu)7696, galH, E15, galK16, [Delta](lac)X74, rpsL, (Strr), hsdR2 (rK-, mK+), mcrA, mcr B1 (20 ) was used for cloning. Escherichia coli HMS174(DE3) F-, recA, hsdR, (rK12-, mK12+), RifR (21 ) and B834(DE3) F-, ompT, hsdSB, (rB-, mB-), gal, met, dcm (22 ) were used for protein production.
Bacteriophage T7 genomic DNA and endoproteinase Glu-C were purchased from Sigma. Single-stranded M13mp18 was purified as described (23 ). Restriction enzymes were purchased from New England Biolabs and Gibco BRL. Taq polymerase was purchased from Promega. Oligonucleotides were synthesized using an Applied Biosystems 381A DNA synthesizer. Radiolabelled nucleotides were purchased from Amersham.
All chemicals were obtained from Sigma unless stated otherwise. Protein purification resins were obtained from Pharmacia.
HMS174(DE3)pAR5018 (10 ) was grown, induced and harvested as described previously (24 ). The cell pellet was resuspended in buffer A [50 mM Tris-HCl, pH 7.5, 2 mM EDTA, 1 mM dithiothreitol (DTT)] + 10% w/v sucrose and frozen at -80oC until required. The cells were lysed by sonication and the supernatant was clarified by centrifugation at 20 000 g for 10 min. 4A' was precipitated from the supernatant using polyethyleneimine and ammonium sulphate as described previously (24 ). The pellet from the ammonium sulphate precipitation was resuspended in buffer A to bring the conductance of the solution to that of buffer A + 200 mM NaCl, prior to loading onto a Mono-Q 10/10 column. The column was washed with buffer A + 300 mM NaCl and the protein eluted with a 160 ml 300-600 mM NaCl gradient in buffer A. The fractions containing the protein were pooled and concentrated using a Centricon 30 (Amicon) to 10 mg/ml. The protein was applied to a Superdex 200HR 10/30 gel filtration column pre-equilibrated with buffer A + 100 mM NaCl. The protein elutes as a single peak and is >99% pure as estimated by Coomassie stained SDS-PAGE (Fig. 1 ).
Proteolytic cleavage of purified 4A' (1 mg/ml) was carried out at room temperature (22oC) for 15 min. A number of proteases were tried, but the best results were obtained using endoproteinase Glu-C (V8 protease) with a buffer containing 50 mM sodium phosphate, pH 7.8, + 100 mM NaCl. The reactions were terminated by adding SDS-PAGE gel loading buffer and heating to 95oC for 5 min. The samples were loaded immediately onto a 15% SDS-polyacrylamide gel for analysis.
N-Terminal sequencing of the proteolytic fragments was carried out using an Applied Biosystems model 470A protein sequenator connected to an on-line 120A high pressure liquid chromatograph.
To enable overexpression of 4C, PCR was used to amplify the coding sequence. The DNA primers were based on the published T7 genomic sequence (7 ; GenBank accession no. G15572) and on the amino acid sequence of the proteolytic fragments of 4A'. The 5' primer had a NcoI site and the 3' primer had a HindIII site after the stop codon. This allowed cloning into T7 promoter-based pET expression vectors (Novagen). The 5' primer was 5'-GATATACC ATG GCT GCA CAG GTT CTA CCT G-3' and the 3' primer was 5'-GCA TCC AAGCT TCA GAA GTC AGT GTC GTT G-3'. PCR was performed in 50 [mu]l using these primers and T7 genomic DNA. The PCR reaction was electrophoresed in 1.0% agarose (Gibco BRL) and the product excised from the gel and purified using a QiaexII DNA extraction kit (Qiagen). The product was digested with NcoI and HindIII and further gel purified. The PCR products were ligated to dephosphorylated, NcoI/HindIII-cleaved pET21d, transformed into E.coli MC1061 and plated onto LB agar plates containing 100 [mu]g/ml ampicillin.
Positive clones were identified by restriction digestion. Clones containing the correct insert were screened for expression by transforming into B834(DE3)[pLysS], growing selected colonies in 5 ml Luria broth (LB) containing ampicillin and chloramphenicol at 37oC until the A600 reached 0.6, induction with 1 mM isopropyl-1-thio-[beta]-D-galactopyranoside (IPTG) and growing for a further 3 h. The level of induction was monitored using SDS-PAGE. This clone was designated pET214C. Cloning and expression of 4D was carried out similarly to 4C; the 5' primer was 5'-GATATA CC ATG GCT GCA CAG GTT CTA CCT G-3', the 3' primer was as for 4C. The expression vector used was pET28a. This clone was designated pET284D.
Onelitre cultures of LB containing 100 [mu]g/ml ampicillin and 34 [mu]g/ml chloramphenicol or 30 [mu]g/ml kanamycin and 34 [mu]g/ml chloramphenicol were inoculated with a 5 ml culture of HMS174(DE3)[pLysS pET214C] or B834(DE3)[pLysS pET- 284D]. The cultures were grown with shaking at 37oC until the A600 reached 0.6-0.7. The cultures were induced by addition of 1 mM IPTG. Growth was continued for 3 h, before the cells were harvested by centrifugation at 5000 g. The cell pellets were resuspended in 20 ml buffer A + 10% sucrose and frozen at -80oC until required. The cells were lysed by sonication and the supernatant was clarified by centrifugation at 20 000 g, prior to precipitation by addition of an equal volume of saturated ammonium sulphate. The precipitate was harvested by centrifugation at 20 000 g. The pellet was resuspended in buffer A in a volume such that the conductivity of the solution was equal to or less than the conductivity of buffer A + 100 mM NaCl. This was applied to a 40 ml Q-Sepharose column equilibrated with buffer A + 100 mM NaCl. The column was washed with buffer A + 300 mM NaCl. The protein was eluted with a gradient of 300-800 mM NaCl in buffer A. The peak fractions were pooled and diluted to buffer A + 200 mM NaCl and loaded onto a Mono-Q 10/10 column. The column was washed with buffer A + 300 mM NaCl and the protein eluted with a 160 ml 300-700 mM NaCl gradient in buffer A. Purity of the sample was monitored using SDS-PAGE (Fig. 1 ).
Protein samples were analysed by SDS-PAGE in 10, 12 and 15% gels with 4% stacking gels (25 ). Gels were stained with Coomassie Brilliant Blue and destained in 10% acetic acid and 25% methanol.
Gel filtration was performed at 22oC in the absence of nucleotide. The protein samples were concentrated using a Centricon 30 (Amicon). Gel filtration was performed using a Superdex 200 HR 10/30 column. The column was equilibrated with buffer A + 100 mM NaCl and calibrated with protein standards. The protein sample was applied to the column in 100 [mu]l. The apparent molecular weights (Mapp) were calculated from an interpolation of a semi-log plot of partition coefficient (Kav) of the protein standards versus molecular weights.
dTTP hydrolysis was monitored by following the production of inorganic phosphate using acidic ammonium molybdate with malachite green (26 ). The reaction mixtures (50 [mu]l) contained 50 mM Tris-acetate, pH 7.5, 50 mM sodium acetate, 0.1 mM EDTA, 10 mM magnesium chloride, 0.1 mg/ml BSA, 1.5 [mu]g single-stranded M13mp18 DNA, dTTP and enzyme. The reactions were started by addition of enzyme and were incubated at 37oC for 10 min. Acidic malachite green/ammonium molybdate solution (0.8 ml) was used to stop the reaction, followed by addition of 0.1 ml 34% (w/v) sodium citrate 1 min later. The colour was left to develop for 20 min and the absorbance was read at 660 nm using a reaction minus enzyme for each concentration of substrate as a blank. All assays were performed in duplicate. The amount of phosphate produced was quantified using a standard curve.
A 68mer oligonucleotide with 22 bases homologous to M13mp18 and 5' and 3' tails of single-stranded DNA (27 ) was 5'-end-labelled using T4 polynucleotide kinase (New England Biolabs). The oligonucleotide was annealed to single-stranded M13mp18 DNA. The unannealed oligonucleotide and free label were removed by centrifugation through a 1 ml Sepharose CL-6B spin column.
The reactions (10 [mu]l) contained 20 mM Tris, pH 7.5, 50 mM NaCl, 3 mM MgCl2, 5 mM DTT, 10% glycerol, 9.5 nM substrate, enzyme and 1 mM dTTP. The reactions were started by addition of enzyme and incubated at 37oC for 30 min. One fifth volumes of 2% SDS, 200 mM EDTA and 50% glycerol were used to stop the reaction. The mixture was electrophoresed through a non-denaturing 6% polyacrylamide gel containing TBE (23 ). After electrophoresis, the gel was fixed and dried under vacuum. Autoradiography was carried out overnight at -80oC using Fuji RX X-ray film. Alternatively, quantification of the degree of displacement of the oligonucloetide was carried out using a phophorimager and ImageQuant software (Molecular Dynamics). The percentage displacement was calculated from the amount of radioactivity (adjusted for background) in the annealed and unannealed bands and corrected for the background of unannealed product in the protein minus control.
The protein was concentrated to 7 mg/ml using a Centricon 30 (Amicon). The protein was exchanged into 10 mM Tris, pH 7.5, 50 mM NaCl, 1 mM DTT during this process. Crystals of 4D were obtained by vapour diffusion at 20oC in hanging drops (equal volumes of protein and well solution) over a 1 ml well solution of 50 mM CHES-NaOH, pH 9.5, 0.8 M sodium acetate, 1 mM sodium azide, polyethylene glycol (PEG) of molecular weight 4000 at a concentration of 20-30% (w/v). The crystals were either mounted in capillaries at room temperature or flash frozen, in the same solution as the well solution with the addition of 15% glycerol (v/v), at 100 K. Data were collected using a Rigaku rotating anode X-ray generator and a MAR image plate detector and processed with DENZO and SCALEPACK (28 ).
4A' was digested under limiting conditions with a range of proteases, including trypsin, chymotrypsin, clostripain, elastase, endoproteinase Glu-C and thermolysin. Despite the differences in specificity, all enzymes produced a major product of between 33 and 36 kDa (data not shown). However, endoproteinase Glu-C gave rise to discrete fragments that were resistant to further proteolysis (Fig. 2 a). Fragments 1 and 2 were produced rapidly and were cleaved further. N-Terminal sequencing was used to identify the cleavage sites (summarized schematically in Fig. 2 b). Fragment 1 has the same N-terminal sequence as full-length 4A', showing that cleavage occurs near the C-terminus, resulting in the loss of ~3 kDa. We did not carry out C-terminal sequencing and so were unable to determine the cleavage site at the C-terminus. However, we have indicated an approximate position (Fig. 2 b) based on the apparent molecular weight of fragment 1 as determined by SDS-PAGE (Fig. 2 a). Following the initial cleavage, the 51 N-terminal amino acids were also removed by proteolysis (fragment 2). The 33 kDa fragment 3 was the result of proteolysis between residues 218 and 219. The position of this cleavage site is in the same region of the protein as the trypsin-sensitive site at amino acid 275 reported recently by Washington et al. (29 ). This suggests that there may indeed be a helicase structural domain with a flexible region of protein linking it to the N-terminus. Sequence comparison of 4A' from amino acid 219 (Fig. 3 ) with a similar tryptic fragment of DnaB (17 ) showed that the C-terminal fragment has all the motifs identified in the DnaB family of helicases (13 ). The fragment derived from 4A' was also similar both in size and sequence composition to the 30 kDa hexameric 5' -> 3' helicase from the incompatibility group Q plasmid RSF1010 (30 ,31 ), suggesting that the C-terminal fragment derived from 4A' may be a helicase.A
Purified 4C, expressed in B834(DE3)[pLysS], was visualized using SDS-PAGE and was observed to be a mixture of two proteins, with the majority having a lower molecular weight than predicted (Fig. 1 , lane 5). 4C purified from HMS174- (DE3)[pLysS] was >99% pure (Fig. 5 , lane 3). However, when lower concentrations of 4C were analysed by SDS-PAGE (data not shown) it can be seen that, although the majority of the protein is equivalent to the expected molecular weight, a low level of the lower molecular weight product seen previously was still observed. Consequently, HMS174[pLysS] was used for production of 4C, since the purified protein is a more homogeneous population.
A sample of 4C purified from B834(DE3)[pLysS] was subjected to electrospray mass spectrometry. Two peaks were observed (data not shown). Comparisons of the observed molecular weights (38 329 +- 4 and 35 994 +- 3 Da, a difference of 2335 Da) with calculated molecular weights for truncations of either the N- or C-terminus suggested that the larger protein is the 4C protein lacking its N-terminal methionine (calculated molecular weight 38 318 Da) and that the smaller protein lacks the 24 N-terminal amino acids (calculated molecular weight 35 983 Da, 2335 Da smaller than 4C). The gene for the smaller protein (designated 4D) was cloned by PCR. 4D was over-expressed using pET28a (designated pET284D) and purified (Fig. 1 , lane 4).
The purifed proteins were >99% pure as assessed by SDS-PAGE (Fig. 1 ).
Nucleotide hydrolysis is essential for the activity of helicases (2 ). 4A is able to hydrolyse both dNTPs and rNTPs using single-stranded DNA as an effector (4 ). However, the preferred substrate is dTTP (34 ). For most of the nucleotides the concentration dependence of the NTPase activity is hyperbolic, however, the ATPase activity is sigmoidal (24 ). Since dTTP is the preferred substrate, the dTTPase activity of 4C and 4D, using single-stranded M13 as an effector, was compared with 4A'.
4C and 4D were able to hydrolyse dTTP. A plot of the primary data is hyperbolic (data not shown) and thus the activities obey Michaelis-Menten kinetics as reported previously for 4A' (24 ,34 ). The kinetic parameters were determined by plotting v/[dTTP] against v (Fig. 5 ) and are shown compared with those of 4A' determined at the same time in Table 1 . The values of Km for 4C and 4D are of a similar order of magnitude to 4A' and are in good agreement with the value reported previously for 4A' (24 ,34 ). However, while 4A' and 4C have similar values of kcat, the value for 4D is 5-fold lower.
Table 1
The ability of 4C and 4D both to hexamerize and to hydrolyse dTTP suggests that they may be helicases. The helicase activity of 4C and 4D compared with 4A' was analysed as a function of protein concentration. The displacement of an oligonucleotide substrate annealed to single-stranded M13 was used as an assay system. The substrate had a 22 bp duplex and both 5' and 3' tails of single-stranded DNA. dTTP was used as an energy source.
4C and 4D are both helicases active upon a tailed substrate (Fig. 6 ). However, the activities of 4C and 4D were markedly less than that of 4A'. At a concentration of 0.21 [mu]M 4A' displaced 89% of the annealed oligonucleotide, whereas, even at 3.4 [mu]M, 4C only displaced 40% of the oligonucleotide. 4D is an even poorer helicase than 4C; at a concentration of 16.9 [mu]M 4D displaced only 18% of the oligonucleotide.
Despite the reduction in activity of the fragments, we conclude that we have isolated a minimal helicase domain and that this is likely to be a structural domain. In order to obtain high resolution structural data we have obtained crystals of 4D by the hanging drop method (Fig. 7 ). Preliminary characterization of these crystals reveals that they are tetragonal and belong to the space group P41212 (or P43212) with unit cell dimensions a = b = 122 Å, c = 285 Å. The crystals diffract to 3.5 Å using a rotating anode source. Assuming one hexamer/asymmetric unit, the Vm is 2.5 Å3/Da and the approximate solvent content is 50%. However, it is possible that there may be half a hexamer in the asymmetric unit if the hexamer is sitting on a crystallographic 2-fold axis. A heavy atom derivative search is underway.
Previous work has indicated that the gene 4 proteins have defined structural domains (15 ). These low resolution structural data are consistent with the idea proposed by Ilyina et al. (13 ) that gene 4 proteins evolved through fusion of a primase and helicase. It is therefore likely that there would be two major structural and functional domains. The limited proteolysis presented in this study has helped delineate the domain structure of the 4A' protein in terms of function. The susceptibility of the N-terminus of 4A', a region only found in 4A', to proteolysis is consistent with a discrete structural domain consisting of the Cys4 zinc binding motif that functions to recognize priming sites (8 ,9 ). Rapid cleavage of the C-terminus also suggests a discrete region of the protein. This region of the protein is highly acidic (8/24 residues) and may be involved in protein-protein interactions such as that observed with the T7 DNA polymerase (35 ). The remainder of the protein is likely to consist of the two structural domains observed by electron microscopy (15 ). The evidence presented in this paper suggests that these may also be functional domains, since the proteins derived from the C-terminal region of 4A' are both helicases. However, both 4C and 4D are poorer helicases than 4A'. The reduction in helicase activity is not due solely to an effect on the DNA-dependent dTTPase activity, because, while 4D has a lower kcat than 4A', 4C has very similar kinetic parameters to 4A'. The rate of the helicase reaction is dependent upon enyme concentration even at levels considerably in excess of the DNA substrate concentration (Fig. 6 ). This suggests that the reduction in helicase activity is likely to be due to a reduced affinity for DNA or to reduced processivity of the enzyme. With the present data we are unable to distinguish between these alternatives. The defect in helicase activity could possibly be due either to a difference in oligomerization relative to 4A or because of other sequences that are not present in the fragments. The C-terminal tryptic fragment of DnaB, which only possesses single-stranded DNA-dependent ATPase activity and no detectable helicase activity (36 ), is defective in oligomerization; only stable dimers and not hexamers are formed. In contrast, both 4C and 4D seem to form more stable oligomers than 4A'. The mechanism by which this could decrease helicase activity is not clear and needs further investigation. 4C and 4D are larger than the DnaB tryptic fragment, with most of the additional amino acids being at the N-terminus of the proteins. Since all of the conserved motifs thought be necessary for helicase activity in the DnaB family of proteins are present in both fragments, it is possible that sequences additional to the observed conserved motifs are necessary for helicase activity and could either lie just N-terminal of Motif I or within the N-terminal domain. If sequences just N-terminal of Motif I are important, this could explain why 4C is a better helicase than 4D. There is evidence for a degree of cooperativity between the two enzymatic activities of 4A, since mutation of the `Walker A' motif, which abolishes helicase activity, also reduces primase activity (14 ). The reduction in primase activity is due to a reduced affinity for DNA, the helicase activity in this case acting to tether the primase to the DNA. A similar situation is seen in E.coli, where DnaG primase requires DnaB to stimulate its activity (37 ,38 ) but hydrolysis of ATP by the helicase and therefore duplex unwinding is not necessary (39 ). However, it need not follow that the helicase requires the primase domain. Serial truncations of 4A from the N-terminus and addition of the purified N-terminus to the helicase reaction will help address this question.
The high resolution structure of a 3' -> 5' helicase, PcrA from B.stearothermophilus, has recently been reported (19 ). Interestingly, the structure revealed two sub-domains, both with homology to RecA. RecA has the same ATP binding fold as F1-ATPase (40 ), a pseudo-hexameric protein which has sequence homology with the bacterial hexameric RNA helicase Rho (41 ). Furthermore, an alignment of RecA with a secondary structure prediction of 4A (29 ) suggested that the C-terminus of 4A would have the same structure as both RecA and F1-ATPase. Since 3' -> 5' helicases have been shown to have a RecA fold we have predicted that all helicases have a common structural core, namely a RecA domain. Comparison of the residues that comprise the nucleotide binding site of RecA with those in PcrA shows that they are conserved spatially but not conserved in primary sequence (19 ,42 ). For example, the adjacent DE residues of Motif II in PcrA are found far apart in the primary sequence in RecA (D144 and E96) but adjacent in space in the structure. The Walker A and B motifs (B contains the D of the DE found in Motif II in PcrA) found in RecA align well with Motifs I and II in gene 4 protein (29 ), suggesting that the structure of hexameric helicases, in terms of the relative positions of amino acids in primary sequence, may in fact be more similar to RecA than the 3' -> 5' helicases. In order to answer this question, we have crystallized the helicase domain of the gene 4 protein and await the structural determination with interest.
We thank A.Rosenberg for the 4A' clone and data prior to publication, V.Cooper for synthesis of oligonucleotides, T.Willis for N-terminal sequencing of proteolytic fragments, K.Marians for DnaB and R.Aplin for mass spectrometry. This work was supported by the MRC and Wellcome Trust.
Protein
Km (mM)
kcat (per s)a
kcat/Km (per s/M)
4A'
0.36
0.9
2.5 * 103
4C
0.25
1.1
4.4 * 103
4D
0.22
0.2
0.9 * 103
REFERENCES