当前位置: 首页 > 期刊 > 《核酸研究》 > 2004年第2期 > 正文
编号:11371361
Recognition of DNA substrates by T4 bacteriophage polynucleotide kinas
http://www.100md.com 《核酸研究医学期刊》
     Fred Hutchinson Cancer Research Center and the Graduate Program in Molecular and Cellular Biology, University of Washington, 1100 Fairview Avenue North, A3-025, Seattle, WA 98109, USA and 1 New England Biolabs, 32 Tozer Road, Beverly, MA 01915, USA

    *To whom correspondence should be addressed. Tel: +1 206 667 4031; Fax: +1 206 667 5894; Email: bstoddar@fhcrc.org

    PDB nos 1RC8, 1RRC, 1RPZ.

    ABSTRACT

    T4 phage polynucleotide kinase (PNK) displays 5'-hydroxyl kinase, 3'-phosphatase and 2',3'-cyclic phosphodiesterase activities. The enzyme phosphorylates the 5' hydroxyl termini of a wide variety of nucleic acid substrates, a behavior studied here through the determination of a series of crystal structures with single-stranded (ss)DNA oligonucleotide substrates of various lengths and sequences. In these structures, the 5' ribose hydroxyl is buried in the kinase active site in proper alignment for phosphoryl transfer. Depending on the ssDNA length, the first two or three nucleotide bases are well ordered. Numerous contacts are made both to the phosphoribosyl backbone and to the ordered bases. The position, side chain contacts and internucleotide stacking interactions of the ordered bases are strikingly different for a 5'-GT DNA end than for a 5'-TG end. The base preferences displayed at those positions by PNK are attributable to differences in the enzyme binding interactions and in the DNA conformation for each unique substrate molecule.

    INTRODUCTION

    T4 polynucleotide kinase (PNK) has become a widely used enzyme in the manipulation of polynucleotides since its discovery in 1965 (1–6). The most widely exploited function of PNK is its ability to catalyze the transfer of the -phosphate of a nucleoside triphosphate to the 5'-hydroxyl of a polynucleotide (1,3,7). However, PNK also possesses a 3'-phosphatase activity (8,9) and a 2',3'-cyclic phosphodiester hydrolysis activity (10). These activities together allow PNK to carry out its only known biological function: the repair of bacterial tRNA anticodon loops that are cleaved as a defense against T4 phage infection (10). PNK processes the cleaved phosphodiester ends, converting them into the 5'-phosphate and 3'-hydroxyl substrates required for T4 RNA ligase, which then repairs the lesion.

    T4 PNK can phosphorylate a wide variety of DNA and RNA substrates, including single-stranded (ss) and double-stranded (ds)DNA and RNA, with the minimum substrate being a 3'-phosphate mononucleoside (1). ssDNA molecules, as well as duplexes with 5'-overhangs, are phosphorylated more efficiently than are duplexes with blunt ends or 3'-overhangs (11). All four nucleotide bases at the 5' substrate end can be phosphorylated, although the efficiency of phosphorylation varies between bases (12,13). SsDNA substrates beginning with 5' guanine are phosphorylated 6-fold more efficiently than similar length ssDNA beginning with 5' cytosine; the enzyme also appears to slightly disfavor cytosine at the second position in potential ssDNA substrates (13). A variety of chemically modified bases and non-nucleosidic moieties are also phosphorylated when attached to the 5' end of an oligonucleotide (14,15).

    The kinase activity of PNK requires the presence of Mg2+ and has a pH optimum of 7.4 to 8.0 (1). No kinetic cooperativity is observed, despite the enzyme’s tetrameric structure (12). The reaction proceeds by an ordered sequential mechanism, with the DNA substrate binding first (12). ATP is generally used as the phosphate donor, although other nucleotides can substitute (7). Inversion of configuration is observed at the phosphorus atom following transfer, indicating a direct in-line transfer of the phosphate without the formation of a phosphoryl-enzyme intermediate (16). The ATP analog ?,-imidoadenylyl 5'-triphosphate is a competitive inhibitor of PNK with respect to ATP and a non-competitive inhibitor with respect to the DNA substrate (17). Mutational analyses have revealed five residues essential to PNK activity (18,19). Four of these residues, K15, D35, R38 and R126, do not permit even conservative substitutions, while the fifth, S16, does tolerate a change to threonine. Mutation of these residues leaves the phosphatase activity intact.

    T4 PNK is active in solution as a tetramer (20–22). The crystal structure of both the isolated kinase domain (23) and of full-length PNK tetramer (24) indicate that the enzyme displays an unusual quaternary structure. Two separate protein interfaces contribute to the formation of the tetramer, one between N-terminal kinase domains only, and the other formed primarily by contacts between C-terminal phosphatase domains (24). The structure of the PNK domain is similar to 6-phosphofructo-2-kinase/fructose-2,6-bisphosphatase (25) and adenylate kinase (ADK) (26). The phosphatase domain is part of the large haloacid dehalogenase (HAD) structural family, of which L-2-haloacid dehalogenase and phosphoserine phosphatase are two of the most similar. In this paper, the mechanism by which T4 PNK acts on a broad spectrum of DNA substrates, while demonstrating base preferences at their 5' end, was investigated by solving the structures of three separate complexes of PNK with different DNA substrates of variable length and different 5' sequence.

    MATERIALS AND METHODS

    Selenomethionine T4 PNK was overexpressed, purified and crystallized as previously described (24). Gel-purified oligodeoxyribonucleotides d(GTC), d(GTCAC) and d(TGCAC) were purchased from Oligos Etc. Crystals were soaked overnight in a 5 μl drop of artificial mother liquor (5 mM MES, pH 6.5, 10% PEG 4000) containing 10 mM Ca2+ and 0.5 mM DNA. Soaked crystals were transferred into a cryo-solution of artificial mother liquor plus 30% (v/v) dimethyl sulfoxide (DMSO) in several steps of increasing DMSO concentration. Crystals were suspended in a fiber loop and flash frozen in liquid nitrogen. Data were collected on beamline 5.0.1 at the ALS (Advanced Light Source, Lawrence Berkeley Laboratory) using an ADSC CCD area detector. Data were processed and scaled using HKL2000 (27).

    The structures of the complexes were solved by molecular replacement with the program EPMR (28) using the structure of the PNK holoenzyme (24) (with ADP coordinates removed) as the search model. Refinement was performed using the CNS package (29) with 10% of reflections set aside for Rfree (30). Visual inspection and manual adjustment of structures was done in XtalView (31). Geometric quality of the refined models was evaluated using PROCHECK (32). Additional modeling was done using QUANTA97 (1997, Molecular Simulations, Inc., San Diego, CA) and Swiss PDB Viewer (33). Figures were prepared using PyMOL (34).

    The coordinates of the three structures have been deposited in the PDB as 1RC8 (GTCAC), 1RRC (GTC) and 1RPZ (TGCAC).

    RESULTS

    Previously grown crystals were soaked with DNA oligonucleotides (5'-GTC-3', 5'-GTCAC-3' and 5'-TGCAC-3') and CaCl2 to generate substrate complexes. Resolution of the structures ranged from 2.46 to 2.9 ? and all were in the I222 space group (Table 1). Unit cell dimensions varied in length by up to 1.5 ?, necessitating molecular replacement for structure determination. Attempts to soak or co-crystallize the enzyme with dsDNA duplexes of any length and/or overhang structure have thus far been unsuccessful. Bound DNA is visible in all three structures. The r.m.s.d. between any of the DNA-bound PNK structures and ‘free’ (only ADP-bound) PNK is 0.7 ?.

    Table 1. X-ray diffraction data and refinement statistics

    In the structure with the shortest (5'-GTC-3') oligonucleotide, the first two bases are visible in the kinase active site. In both structures containing longer DNA substrates (5'-GTCAC-3' and 5'-TGCAC-3'), the first three bases are visible, although the quality of the difference density for the third base is relatively poor (Fig. 1). All DNA bases are in the anti configuration. The kinase active site also contains an ADP molecule, generated by hydrolysis of ATP in the crystallization buffer. No bound metal ions are seen in the kinase active site in any of the structures, presumably indicating that the scissile -phosphate group or an appropriate analog is required for coordination of divalent cations.

    Figure 1. Composite omit maps of the DNA binding region and kinase active site. (a) Observed density for the 5'-GTCAC-3' substrate. The second base (thymine) is stacked against the 5' guanine and forms a polar aromatic interaction with Y52. The edges of both bases are solvent-exposed, and the opposite face of the guanine base is flanked by non-polar residues in the base of the active site. (b and c) The density for the 5'-TGCAC-3' substrate . The 5' thymine base and its ribose sugar are located in a similar position as the 5' guanine on the previous panel. The second base (guanine) has swung out of the active site cleft and is stacked against the third base (cytosine). The structural roles of several side chains in DNA binding are different between the two complexes.

    The coordination and position of the terminal ribose sugar in the three structures are very similar (Figs 1 and 2), with an r.m.s.d. value of 0.4 ?. The 5' hydroxyl of the oligonucleotide points towards the ?-phosphate of the ADP. The distance between the two atoms ranges from 5.3 ? for the GTC oligo to 6.0 ? for the TGCAC oligo. The space between these atoms does not contain any ordered density in any of the structures, and is sufficient to accommodate the -phosphate group of ATP.

    Figure 2. Schematic of interactions to the first three ordered bases in the PNK–DNA crystal structures. The most conserved protein–DNA contacts between the two sequences of DNA substrates are to Val 131 (which contacts the 5' base in both structures), D35 and T86 (which contact the 5' ribose sugar and its adjoining phosphate non-binding oxygen). Both residues are known to be essential for catalysis in PNK mutagenesis studies, with the exception of a T86S substitution. One methionine residue, which is actually a selenomethionine in these structures, is observed to make a long contact between its terminal methyl group and the thymine nucleotide in the TGCAC complex. However, the specific activity of the selenomethionyl enzyme is not reduced relative to the wild-type enzyme (data not shown).

    The first two of the DNA oligonucleotides, 5'-GTC-3' and 5'-GTCAC-3', differ only in their length. This variation has little effect on the conformation of the visible DNA. The deoxyguanosine and deoxythymidine nucleotides bind in essentially the same conformation (r.m.s.d. of 0.3 ? for comparable atoms), mostly through contacts to the sugar-phosphate backbone (Figs 1 and 2). The bases themselves point away from the surface of the protein, and their outer edges are exposed to the solvent. Contacts to the DNA backbone are made by the side chains of residues R34, D35, R38, D85, T86 and N89. The backbone amide and carbons of D85 and T86 also contact a phosphate oxygen. The guanine base is wedged in a shallow hydrophobic pocket, with its edge exposed to solvent and an aromatic face in contact with three aliphatic residues: V131, P132 and V135. The thymine base is more exposed, and is contacted by Y52 and R38. The two ordered bases are stacked against one another, and tyrosine 52 is positioned such that one of its edges is forming a polar aromatic interaction with the thymine base at an angle of 30° from a parallel stacking arrangement.

    Beyond the first two nucleotides the quality of electron density decreases significantly in both structures. In the GTC structure only the third deoxyribose sugar is observable. In contrast, in the GTCAC structure the complete deoxycytidine nucleotide is observable, modeled and refined, although the base does display elevated B-values. This final visible nucleotide is flipped out and away from the first two bases and lies along a distal groove on the surface of the protein (Figs 1 and 3A). The 3' hydroxyl of the sugar points away from the enzyme. The cytosine base is contacted by G58, T61 and R92, while its ribose contacts the hydroxyl of Y52. It is not clear why the third (Cyt) base is more ordered in the longer DNA complex; there are no crystal lattice contacts in this region, or additional observed contacts between protein side chains and that base. However, even in the structures of longer DNA substrates the density for the third base is relatively poor, indicating that the enzyme primarily forms interactions with the first two bases of ssDNA substrates, regardless of length.

    Figure 3. Conformation of bound 5'-GTC-3' and ADP (A) or 5'-TGC-3' and ADP (B) to the PNK domain. The nucleotide phosphate donor is bound on the opposite side of the tunnel formed by the interactions of two loops that extend over the 5' nucleotide sugar. Protein orientation is the same in both panels.

    To examine the effect of a different 5' DNA sequence on substrate recognition, the sequence of the first two bases was inverted (from 5'-GT to 5'-TG) and a third structure was solved. At the actual site of phosphoryl transfer, the structures of PNK bound to a 5'-TGCAC-3' and a 5'-GTCAC-3' DNA oligonucleotide are similar. The 5' ribose sugar is bound in a similar conformation and position in both structures, and the corresponding 5' thymine base in TGCAC is located in the same hydrophobic pocket as the 5' guanine base in GTCAC. In contrast to the previous structure, the face of the 5' thymine base is only in contact with two aliphatic residues: V131 and the C carbon of M139 (Fig. 2).

    Beyond the 5' nucleotide, the bound conformations of GTCAC and TGCAC differ significantly (Figs 3 and 4). In the GTCAC structure the first two bases (5'-GT) are stacked against each other with the third (C) rotated away into a more distal protein surface groove (Fig. 3A). In contrast, in the TGCAC structure the second base (G) is rotated away from the active site, into the same distal surface groove, where it stacks against the third base (C) (Fig. 3B). Contacts to the first phosphate group and the second ribose sugar remain much the same between the two structures, although R38 makes different contacts in the TGCAC structure due to the shift in base placement. Overall, the third nucleotide is involved in significantly fewer contacts in the TGCAC structure, due to its more remote and solvent-exposed position.

    Figure 4. Superposition of bound 5'-GTC-3'(blue) and 5'-TGC-3' (gray) and surrounding protein residues (green and magenta, respectively). Note the flip of the second base in the 5'-TGC-3' structure, and its position relative to the third (cytosine) base in the 5'-GTC-3' structure.

    The different conformations of 5'-GTCAC-3' and 5'-TGCAC-3' cause many of the protein side chains to exhibit significantly different contacts to the two bound DNA substrates (Fig. 2). For example, the guanine base G2 in TGCAC is contacted by R34, Y52, E57, T61 and R92. In the GTCAC structure, two of these same residues (T61 and R92) instead contact cytosine C3, while R34 and E57 do not contact any bases. The edge-to-face ring interactions seen between Y52 and the thymine base T2 in the GTCAC structure are not present in the TGCAC structure, and density quality for Y52 is quite low in that structure. Finally, the most dramatic shift in conformation and contacts is seen for R34. The conformation of that side chain swings around such that it makes extensive contacts with the guanine base instead of contacting the second phosphate, which has shifted position by 4.6 ?.

    DISCUSSION

    To our knowledge, this is the first structure of a non-specific ssDNA binding enzyme in complex with DNA targets. In contrast, the structures of several ssDNA binding proteins that contain oligosaccharide-binding (OB) folds in complex with DNA targets have also been solved; these proteins display varying amounts of sequence specificity and a different mode of interaction with bound ssDNA as compared to PNK. These structures include Oxytricha telomere end binding protein (TEBP) (35), Schizosaccharomyces pombe telomeric ssDNA binding protein Pot1 (36), Escherichia coli ssDNA-binding protein (SSB) (37), human replication protein A (RPA) (38), the DNA-binding region of BRCA2 (39) and the T4 phage GP32 ssDNA binding protein (40). The BRCA2 protein appears to be the least sequence-specific of those proteins. In each case, the DNA is bound in an extended conformation spanning two or more OB domains, the nucleotide bases do not interact with one another, most bases are buried in the protein–DNA interface in complementary binding pockets, and the phosphate backbone is exposed to solvent. Sequence readout is accomplished through a combination of base-specific hydrogen bonds and substantial van der Waals complementarity. The conformation and position of the ssDNA, relative to the OB domain interfaces, varies significantly between those structures. In contrast, ssDNA is bound to PNK such that the phosphate backbone is buried and the bases point away from the enzyme, making limited contacts to a series of hydrophobic side chains while exposing their edges to solvent.

    Structures of several non-specific and partially specific dsDNA binding proteins and phosphoryl transferase enzymes in complex with DNA have also been solved, including bacterial DNA packing proteins Sso7d (41) and Sac7d (42), the nucleosome core particle (43), the chromosomal high mobility group proteins NHP6A (44) and HMG-D (45), the nucleases Vvn (46) and DNase I (47) and Taq polymerase (48). DNA binding interactions of these proteins are often localized to the minor groove, usually through contacts with the DNA backbone and additional non-specific contacts to bases (49). The endonuclease Vvn deviates from this pattern in interacting only with the backbone phosphates of its DNA substrate through both direct and water-mediated hydrogen bonds (46). In contrast, PNK interacts with its ssDNA substrate through a combination of contacts to backbone phosphates and sugars and the bases themselves. The contacts to the bases are for the most part hydrophobic contacts to their aromatic faces. Many of the protein side chains make significantly different contacts depending on the exact sequence of the bases at the 5' end of the various substrates.

    In the structures of both 5-base DNA oligonucleotides, the final visible 3' hydroxyl is pointing out away from the surface of the protein into the solvent, and is poorly ordered. This suggests that beyond the first three nucleotides the DNA substrate is not in contact with the protein, a conclusion supported by the lack of substrate sequence preference exhibited by PNK beyond the second base (12). Modeling a dsDNA substrate, based on the observed positions of ssDNA, indicates that at least two and possibly three nucleotides at the 5' end of the substrate must be dissociated from their Watson–Crick partners and distorted from their B-form conformation in order to bind in the active site. Beyond the first three bases, dsDNA structure can be accommodated. This observation could explain the decreased kinase efficiency towards dsDNA with 3'-overhangs or blunt ends, or 5' overhangs less than three bases in length (11).

    PNK displays preference for certain bases at the first two positions of potential kinase substrates as described in the Introduction. The structures of the ssDNA complexes described here indicate that the 5' DNA sequence plays a role in determining substrate preference by effecting both the substrate conformation and the resulting set of protein–DNA contacts. The most obvious difference in the two bound DNA conformations is the pattern of stacking exhibited by the three ordered bases of the substrate. Thermodynamic studies of the folding of single-stranded dinucleotide DNA molecules have shown that purine–purine stacking interactions are more stable than purine–pyrimidine stacks, which in turn are more stable than a pyrimidine–pyrimidine stack (50). These differences are slight, however, with measured free energy difference between the purine–purine and pyrimidine–pyrimidine stacks of no more than 1 kcal/mol. This suggests that the observed conformations of the bound DNA substrates are primarily dictated by individual sets of optimized contacts to surrounding protein side chains, combined with the additional favorable stacking interactions between adjacent bases. In both structures, the observed DNA base stacking interactions follow a previously documented structural pattern of placing an extracyclic polar group of one base directly over the aromatic ring of the next base, in order to induce favorable charge dipole interactions (50). In the PNK structures, the N4 of thymidine or the O4 of cytosine is located within 3.5 ? of the nearest atoms of the aromatic ring of the guanosine base.

    Additional modeling experiments were performed in an attempt to further understand the preference of the enzyme for 5' guanine, relative to its reduced kinase activity towards substrates with 5' cytosine. In these analyses, the 5' base in the two structures were replaced with a cytosine. The main structural difference inferred from these models, relative to the crystal structures, is the loss of most non-polar van der Waals contacts between the base and the surrounding binding pocket. In particular, the removal of the methyl group from the pyrimidine base (when thymine is replaced with cytosine) eliminates contact with Val 31, which provides the primary protein contact to the 5' base in the bound TGCAC complex (Fig. 2). This would leave a cytosine base at the 5' position with the fewest van der Waals interactions of any of the four bases, assuming that each base is located in the same hydrophobic pocket described in this study.

    The structures reported here correlate well with previously published mutational data. All five of the residues essential for kinase activity make contact with either the ADP or DNA molecule. Residues K15, S16 and R126 contact the ?-phosphate of the ADP molecule in a similar manner to that described previously (24). In addition, R126 is positioned such that it has the potential to interact with the -phosphate of ATP, presumably acting to stabilize the expected phosphoanion transition state. On the other side of the active site, R38 contacts the phosphate group of the 5' base of the bound oligonucleotide. The final essential residue, D35, contacts both R38 and the 5' hydroxyl of the DNA oligonucleotide. This placement allows it to assist in proper positioning of R38 as well as the phosphate acceptor end of the DNA substrate. The position of D35 is similar to one of the conformations reported previously for the enzyme–ADP complex (24); however the alternate conformation is not seen in the complex structures. This could be a result of the DNA stabilizing one conformation in the absence of the catalytic Mg2+. Although Ca2+ was part of the soak conditions, no bound metal was found in the active site. The complex structures do contain enough room for D35 to assume the alternate conformation and possibly bind both the 5' hydroxyl and a metal ion.

    Overlaying the DNA-bound structures of PNK with its closest structural homolog, ADK bound to its adenylate substrate and a non-hydrolyzable ATP analog (26), reveals substantial differences in substrate interaction mechanisms. The PNK substrates are sitting in clefts on the surface of the protein, with only the catalytic phosphoryl transfer site and the 5' ribose and hydroxyl secluded from solvent. The DNA bases appear accessible to the solvent, especially on their outer edges, with most binding contacts occurring to the backbone. The phosphates of the ADP molecule are also specifically bound and more secluded from solvent as they enter the active site tunnel, but the remainder of the ADP is solvent exposed. In contrast, the substrates of ADK are almost completely secluded in the interior of the protein. This is accomplished in part by the significant movement of a lid domain absent in PNK. In the absence of substrate, the ADK lid domain is open to permit access to the substrate binding sites. In the presence of bound substrate, the lid domain has swung down and almost completely encloses the substrates within the protein. Only a portion of the base on the AMPPNP molecule in the structure appears to remain accessible to solvent. This results in more specific contacts between protein residues and the adenine bases, especially for the AMP molecule when compared with the DNA in the PNK structure.

    ACKNOWLEDGEMENTS

    We thank Geoffrey Wilson of New England Biolabs, and Eric Galburt, Betty Shen and other Stoddard laboratory members for advice and assistance; and Adrian Ferre-D’Amare for discussions and advice. This research was supported by NIH grant GM49857 (B.L.S.) and NIH training grant T32GM07270 (J.H.E.).

    REFERENCES

    Richardson,C.C. (1965) Phosphorylation of nucleic acid by an enzyme from T4 bacteriophage-infected Escherichia coli. Proc. Natl Acad. Sci. USA, 54, 158–165.

    Novogrodsky,A. and Hurwitz,J. (1966) The enzymatic phosphorylation of ribonucleic acid and deoxyribonucleic acid. I. Phosphorylation at 5'-hydroxyl termini. J. Biol. Chem., 241, 2923–2932.

    Novogrodsky,A., Tal,M., Traub,A. and Hurwitz,J. (1966) The enzymatic phosphorylation of ribonucleic acid and deoxyribonucleic acid. II. Further properties of the 5'-hydroxyl polynucleotide kinase. J. Biol. Chem., 241, 2933–2943.

    Berkner,K.L. and Folk,W.R. (1980) Polynucleotide kinase exchange as an assay for class II restriction endonucleases. Methods Enzymol., 65, 28–36.

    Chaconas,G. and van de Sande,J.H. (1980) 5'-32P labeling of RNA and DNA restriction fragments. Methods Enzymol., 65, 75–85.

    Sambrook,J., Fritsch,E.F. and Maniatis,T. (1989) Enzymes used in molecular cloning. Molecular Cloning: A Laboratory Manual, Vol. 1, 2nd Edn. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.

    Richardson,C.C. (1981) Bacteriophage T4 Polynucleotide Kinase. In Boyer,P.D. (ed.), The Enzymes. Academic Press, San Diego, CA, Vol. 14, pp. 299–314.

    Sirotkin,K., Cooley,W., Runnels,J. and Snyder,L.R. (1978) A role in true-late gene expression for the T4 bacteriophage 5' polynucleotide kinase 3' phosphatase. J. Mol. Biol., 123, 221–233.

    Cameron,V. and Uhlenbeck,O.C. (1977) 3'-Phosphatase activity in T4 polynucleotide kinase. Biochemistry, 16, 5120–5126.

    Amitsur,M., Levitz,R. and Kaufmann,G. (1987) Bacteriophage T4 anticodon nuclease, polynucleotide kinase and RNA ligase reprocess the host lysine tRNA. EMBO J., 6, 2499–2503.

    Lillehaug,J.R., Kleppe,R.K. and Kleppe,K. (1976) Phosphorylation of dsDNAs by T4 polynucleotide kinase. Biochemistry, 15, 1858–1865.

    Lillehaug,J.R. and Kleppe,K. (1975) Kinetics and specificity of T4 polynucleotide kinase. Biochemistry, 14, 1221–1225.

    vanHouten,V., Denkers,F., van Dijk,M., van den Brekel,M. and Brakenhoff,R. (1998) Labeling efficiency of oligonucleotides by T4 polynucleotide kinase depends on 5'-nucleotide. Anal. Biochem., 265, 386–389.

    van de Sande,J.H. and Bilsker,M. (1973) Phosphorylation of N-protected deoxyoligonucleotides by T4 polynucleotide kinase. Biochemistry, 12, 5056–5062.

    Fontanel,M.L., Bazin,H. and Teoule,R. (1994) Sterical recognition by T4 polynucleotide kinase of non-nucleosidic moieties 5'-attached to oligonucleotides. Nucleic Acids Res., 22, 2022–2027.

    Jarvest,R.L. and Lowe,G. (1981) The stereochemical course of phosphoryl transfer catalysed by polynucleotide kinase (bacteriophage-T4-infected Escherichia coli B). Biochem. J., 199, 273–276.

    Lillehaug,J.R. (1978) Inhibition of T4 polynucleotide kinase by the ATP analog, beta, gamma-imidoadenylyl 5'-triphosphate. Biochim. Biophys. Acta, 525, 357–363.

    Wang,L.K. and Shuman,S. (2001) Domain structure and mutational analysis of T4 polynucleotide kinase. J. Biol. Chem., 276, 26868–26874.

    Wang,L.K. and Shuman,S. (2002) Mutational analysis defines the 5'-kinase and 3'-phosphatase active sites of T4 polynucleotide kinase. Nucleic Acids Res., 30, 1073–1080.

    Panet,A., van de Sande,J.H., Loewen,P.C., Khorana,H.G., Raae,A.J., Lillehaug,J.R. and Kleppe,K. (1973) Physical characterization and simultaneous purification of bacteriophage T4 induced polynucleotide kinase, polynucleotide ligase and deoxyribonucleic acid polymerase. Biochemistry, 12, 5045–5050.

    Lillehaug,J.R. (1977) Physicochemical properties of T4 polynucleotide kinase. Eur. J. Biochem., 73, 499–506.

    Soltis,D.A. and Uhlenbeck,O.C. (1982) Independent locations of kinase and 3'-phosphatase activities on T4 polynucleotide kinase. J. Biol. Chem., 257, 11340–11345.

    Wang,L.K., Lima,C.D. and Shuman,S. (2002) Structure and mechanism of T4 polynucleotide kinase: an RNA repair enzyme. EMBO J., 21, 3873–3880.

    Galburt,E.A., Pelletier,J., Wilson,G. and Stoddard,B.L. (2002) Structure of a tRNA repair enzyme and molecular biology workhorse: T4 polynucleotide kinase. Structure (Camb), 10, 1249–1260.

    Hasemann,C.A., Istvan,E.S., Uyeda,K. and Deisenhofer,J. (1996) The crystal structure of the bifunctional enzyme 6-phosphofructo-2-kinase/fructose-2,6-bisphosphatase reveals distinct domain homologies. Structure, 4, 1017–1029.

    Berry,M.B., Meador,B., Bilderback,T., Liang,P., Glaser,M. and Phillips,G.N.,Jr (1994) The closed conformation of a highly flexible protein: the structure of E.coli adenylate kinase with bound AMP and AMPPNP. Proteins, 19, 183–198.

    Otwinowski,Z. and Minor,W. (1997) Processing of X-ray diffraction data collected in oscillation mode. Methods Enzymol., 276, 307–326.

    Kissinger,C.R., Gehlhaar,D.K. and Fogel,D.B. (1999) Rapid automated molecular replacement by evolutionary search. Acta Crystallogr. D Biol. Crystallogr., 55 (Pt 2), 484–491.

    Brunger,A.T., Adams,P.D., Clore,G.M., DeLano,W.L., Gros,P., Grosse-Kunstleve,R.W., Jiang,J.S., Kuszewski,J., Nilges,M., Pannu,N.S. et al. (1998) Crystallography and NMR system: A new software suite for macromolecular structure determination. Acta Crystallogr. D Biol. Crystallogr., 54 (Pt 5), 905–921.

    Kleywegt,G.J. and Brunger,A.T. (1996) Checking your imagination: applications of the free R value. Structure, 4, 897–904.

    McRee,D.E. (1999) XtalView/Xfit–A versatile program for manipulating atomic coordinates and electron density. J. Struct. Biol., 125, 156–165.

    Laskowski,R.J., Macarthur,M.W., Moss,D.S. and Thornton,J.M. (1993) PROCHECK – a program to check the stereochemical quality of protein structures. J. Appl. Crystallogr., 26, 283–291.

    Guex,N. and Peitsch,M.C. (1997) SWISS-MODEL and the Swiss-PdbViewer: an environment for comparative protein modeling. Electrophoresis, 18, 2714–2723.

    DeLano,W.L. (2002) DeLano Scientific, San Carlos, CA.

    Horvath,M.P., Schweider,V.L., Bevilacqua,J.M., Ruggles,J.A. and Schultz,S.C. (1998) Crystal structure of the Oxytricha nova telomere end binding protein complexed with single strand DNA. Cell, 95, 939–950.

    Lei,M., Podell,E.R., Baumann,P. and Cech,T.R. (2003) DNA self-recognition in the structure of Pot1 bound to telomeric single-stranded DNA. Nature, 426, 198–203.

    Raghunathan,S., Kozlov,A.G., Lohman,T.M. and Waksman,G. (2000) Structure of the DNA binding domain of E. coli SSB bound to ssDNA. Nat. Struct. Biol., 7, 648–652.

    Bochkarev,A., Pfuetzner,R.A., Edwards,A.M. and Frappier,L. (1997) Structure of the single-stranded DNA binding domain of replication protein A bound to DNA. Nature, 385, 176–181.

    Yang,H., Jeffrey,P.D., Miller,J., Kinnucan,E., Sun,Y., Thoma,N.H., Zheng,N., Chen,P.L., Lee,W.H. and Pavletich,N.P. (2002) BRCA2 function in DN binding and recombination from a BRCA2-DSS1-ssDNA structure. Science, 297, 1837–1848.

    Shamoo,Y., Friedman,A.M., Parsons,M.R., Konigsberg,W.H. and Steitz,T.A. (1995) Crystals structure of a replication fork single-stranded DNA binding protein (T4 gp32) complexed to DNA. Nature, 376, 362–366.

    Agback,P., Baumann,H., Knapp,S., Ladenstein,R. and Hard,T. (1998) Architecture of nonspecific protein-DNA interactions in the Sso7d-DNA complex. Nat. Struct. Biol., 5, 579–584.

    Robinson,H., Gao,Y.G., McCrary,B.S., Edmondson,S.P., Shriver,J.W. and Wang,A.H. (1998) The hyperthermophile chromosomal protein Sac7d sharply kinks DNA. Nature, 392, 202–205.

    Luger,K., Mader,A.W., Richmond,R.K., Sargent,D.F. and Richmond,T.J. (1997) Crystal structure of the nucleosome core particle at 2.8 ? resolution. Nature, 389, 251–260.

    Masse,J.E., Wong,B., Yen,Y.M., Allain,F.H., Johnson,R.C. and Feigon,J. (2002) The S. cerevisiae architectural HMGB protein NHP6A complexed with DNA: DNA and protein conformational changes upon binding. J. Mol. Biol., 323, 263–284.

    Murphy,F.V.T., Sweet,R.M. and Churchill,M.E. (1999) The structure of a chromosomal high mobility group protein-DNA complex reveals sequence-neutral mechanisms important for non-sequence-specific DNA recognition. EMBO J., 18, 6610–6618.

    Li,C.L., Hor,L.I., Chang,Z.F., Tsai,L.C., Yang,W.Z. and Yuan,H.S. (2003) DNA binding and cleavage by the periplasmic nuclease Vvn: a novel structure with a known active site. EMBO J., 22, 4014–4025.

    Suck,D., Lahm,A. and Oefner,C. (1988) Structure refined to 2? of a nicked DNA octanucleotide complex with DNase I. Nature, 332, 464–468.

    Eom,S.H., Wang,J. and Steitz,T.A. (1996) Structure of Taq ploymerase with DNA at the polymerase active site. Nature, 382, 278–281.

    Luscombe,N.M. and Thornton,J.M. (2002) Protein-DNA interactions: amino acid conservation and the effects of mutations on binding specificity. J. Mol. Biol., 320, 991–1009.

    Saenger,W. (1984) Forces stabilizing associations between bases: hydrogen bonding and base stacking. Principles of Nucleic Acid Structure. Springer-Verlag, New York, NY, pp. 116–158.(Jennifer H. Eastberg, John Pelletier1 an)