Substrate recognition and ®delity of strand joining by an archaeal DNA ligase Masaru Nakatani 1,2 , Satoshi Ezaki 1,2 , Haruyuki Atomi 1,2 and Tadayuki Imanaka 1,2 1 Department of Synthetic Chemistry and Biological Chemistry, Graduate School of Engineering, Kyoto University; 2 Core Research for Evolutional Science and Technology Program of Japan Science and Technology Corporation, Japan We have previously identi®ed a DNA ligase (Lig Tk )froma hyperthermophilic archaeon, Thermococcus kodakaraensis KOD1. The enzyme is the only characterized ATP-depend- ent DNA ligase from a hyperthermophile, and allows the analysis of enzymatic DNA ligation reactions at tempera- tures above the melting point of the substrates. Here we have focused on the interactions of Lig Tk with various DNA substrates, and its speci®cities toward metal cations. Lig Tk could utilize M g 2+ ,Mn 2+ ,Sr 2+ and Ca 2+ as a m etal cation, but not Co 2+ ,Zn 2+ ,Ni 2+ ,orCu 2+ . The enzyme displayed typical Michaelis±Menten steady-state kinetics with an apparent K m of 1.4 l M for nicked DNA. The k cat value o f t he enzyme was 0.11ás )1 . U sing various 3¢ hydroxyl group donors (L-DNA) and 5¢ phosphate group donors (R-DNA), we could detect ligation products as short as 16 nucleotides, the products of 7 + 9 nucleotide or 8 + 8 nucleotide combinations at 40 °C.Anelevationintemper- ature led to a decrease in reaction eciency when short oligonucleotides were used, suggesting that the formation of a nicked, double-stranded DNA substrate precede d enzyme- substrate recognition. Lig Tk was not inhibited by the addition of excess duplex DNA, implying that the enzyme did not bind strongly to the double-stranded ligation prod- uct after nick-sealing. In terms of reaction ®delity, Lig Tk was found to ligate various substrates with mismatched base- pairing at the 5¢ end of the nick, but did not show activity towards the 3¢ mismatched substrates. Lig Tk could not seal substrates with a 1-nucleotide o r 2-nucleotide g ap. Small amounts of ligation products were detected with DNA substrates containing a single nucleotide insertion, relatively more with the 5¢ insertions. The results revealed the impor- tance of proper base-pairing a t the 3¢ hydroxyl side of the nick for the ligation reaction by Lig Tk . Keywords: archaea; DNA ligase; hyperthermophile; Thermococcus. DNA ligases (EC 6.5.1.1 and EC 6.5.1.2) are universally found in bacteria, eukaryotes and archaea. In addition, they are a lso found i n viruses and bacteriophages [ 1±5]. DNA ligases catalyse the phosphodiester bond formation between adjacent 3¢ hydroxyl and 5 ¢ phosphate groups at a single- strand break in double-stranded DNA [5,6]. They are essential enzymes for maintaining the integrity of the genome during DNA replication [7], DNA excision rep air [8] a nd DNA recombination [9]. DNA strand breaks are commonly generated as reaction intermediates in these events, and the sealing of these breaks depends solely on the proper function of DNA ligase [2]. Therefore DNA ligases are indispensable enzymes in all organisms. DNA ligases fall into two groups, ATP-dependent DNA ligases and N AD + -dependent DNA ligases, on the basis o f the required cofactor for ligase±adenylate formation [2,5,6]. ATP-dependent enzymes have been found in viruses, bacteriophages, eukaryotes, a rchaea and, recently, i n bac- teria, whereas NAD + -dependent enzymes h ave been f ound exclusively in bacteria [2,3,5]. There is high similarity among the ligases within the ATP-dependent groups [10] or NAD + -dependent groups [11,12]. However, enzymes between the two groups show no similarity, with the exception of the AMP-binding site [10]. I t is now accepted that both ATP-dependent and NAD + -dependent DNA ligases catalyse their reactions through a common mecha- nism [13]. The ligation reaction proceeds through t hree steps. In the ®rst step, a ttack on ATP or N AD + by the enzyme results in release of PPi or NMN from the cofactor and formation of enzyme±adenylate through t he covalent addition of AMP to the conserved AMP-binding site lysine of the protein. In the second step, the AMP i s transferred from the protein to the 5¢ phosphate group of the nick on the DNA t o form DNA ±adenylate . In t he third step, the enzyme catalyses phosphodiester bond formation with concomitant release of free AMP from the DNA±adenylate [2,5,6,13]. Catalytic activity of DNA ligase is dependent on appropriate divalent cations and DNA substrates. In general, DNA ligases can utilize M g 2+ and s everal o ther divalent cations that belong to the fourth period of the elements [14±20]. The interaction between DNA ligase and its DNA substrates has been examined from various viewpoints, such as substrate length, and a ctivity towards substrates with gaps, mismatches, or insertions. S everal reports have shown that some enzymes can catalyse the Correspondence to T. Imanaka, Department of Synthetic Chemistry and Biological Chemistry, Graduate School of Engineering, Kyoto University, Yoshida-Honmachi, Sakyo-ku, Kyoto 606-8501, Japan. Fax: +81 75 753 4703, Tel.: +81 75 753 5568, E-mail: imanaka@sbchem.kyoto-u.ac.jp Abbreviations:Lig Tk , DNA ligase from Thermococcus kodakaraensis KOD1; L-DNA, oligonucleotide as 3¢ hydroxyl group donor; R-DNA, oligonucleotide as 5¢ phosphate group donor; T-DNA, complementary oligonucleotide to L-DNA and R-DNA. Enzymes: DNA ligase (EC 6.5.1.1 and EC 6.5.1.2). (Received 25 July 2001, revised 20 November 2001, accepted 21 November 2001) Eur. J. Biochem. 269, 650±656 (2002) Ó FEBS 2002 ligation reaction with gapped, inserted or mismatched substrates [15±17,19±27]. Although a signi®cant amount of k nowledge o n DNA ligases from bacteria, eukaryotes and viruses has accumu- lated, archaeal enzymes h ave only recently been reported. Shuman and coworkers have identi®ed an ATP-dependent DNA ligase from Methanobacterium therm oautotrophicum [18]. W e have report ed the exa mination of a DNA ligase (Lig Tk )fromThermococcus kodakaraensis KOD1, a hyper- thermophilic archaeon [3]. Lig Tk displayed two uniq ue features. One was t hat the enzyme, although belonging to the family of ATP-dependent DNA ligases, could utilize NAD + as a c ofactor. The other was the extreme thermo- stability of Lig Tk : nick-sealing was observed at temperatures up to 100 °C. The thermostability of the enzyme provide s a means to examine DNA ligation reactions at temperatures above the melting point of the DNA substrates. As little is known about DNA ligases from archaea o r f rom hyper- thermophiles, we have examined Lig Tk focusing on the following aspects: (a) its divalent cation speci®city; (b) the effect of temperature on the interaction between enzyme and DNA substrate; (c) t he ability of the enzyme to discriminate gapped, inserted and mismatched ends at the nick. MATERIALS AND METHODS Puri®cation of recombinant Lig Tk The DNA ligase gene (lig Tk )fromT. kodakaraensis KOD1 was s ubcloned into an expression vector, pET-21a(+) (Novagen) [3]. The resulting plasmid pET-lig was intro- duced into Escherichia coli BL21-CodonPlus(DE3)-RIL (Stratagene). The t ransformants were cult ivated in Luria± Bertani medium [ 28] containing 50 lgámL )1 ampicillin at 37 °C until the optical density at 660 nm reached 0.8. Isopropyl- D -thiogalactopyranoside was added at a ®nal concentration of 1 m M to in duce lig Tk gene expression for 7h. Cells were harvested by centrifugation (5000 g,15min, 4 °C), washed with buffer A (50 m M Tris/HCl pH 7.5) , and then resuspended in buffer A. The cells were disrupted by sonication and the super natant was obtained by centrifu- gation (12 000 g,30min,4°C). T he soluble fraction of cell- free extract was heat-treated at 80 °C for 30 min and the precipitate was removed by centrifugation (12 000 g, 30 min, 4 °C) to obtain thermostable proteins. The super- natant was applied to a ResourceQ column (Amersham Pharmacia Biotech) equilibrated with buffer A. As Lig Tk did not bind to the r esin, the ¯ow-through fractions were collected, dialysed with buffer B (50 m M Mes/KOH pH 6.0) and applied to a ResourceS column (Amersham Pharmacia Biotech) equilibrated with buffer B. After washing with buffer B, the enzyme was eluted with a linear gradient of 0±1.0 M KCl in buffer B. The peak fractions containing Lig Tk , which eluted between 0.10 and 0.14 M KCl, were concentrated by using C entricon-30 (Millipore). The enzyme solutio n was applied to a gel ® ltration column (Superdex 200 HR 10/30, Amersham Pharmacia B iotech) equilibrated with buffer C (50 m M Mes/KOH pH 6.0, 100 m M KCl) and eluted with t he same buffer. The active fractions were dialysed with buffer A and used as puri®ed Lig Tk in following experiments. The protein c oncentration was determined with the Bio-Rad protein assay system with BSA as a standard. DNA substrates DNA ligase a ctivity measurements w ere c arried out with synthesized oligonucleotides. The substrate used in most activity measurements was composed of two oligonucleo- tides (L-DNA and R-DNA) and a complementary oligonucleotide (T-DNA). A phosphate group was present at the 5¢ terminus of R-DNA. Deletions, i nsertions and m utations were introduced to the L-DNA(40), R-DNA(30) andT-DNA(80) whennecessary. T he sequences of the oligonucleotides are listed in F ig. 1. In addition, a complete duplex DNA added to the reaction in Fig. 3B consists of 50-mer DNA-A (5¢-CCACTCGACGAGC TTCTTGCCTTCACAGACGAGGACTTGGGAAGCT CACG-3¢) and 50-mer DNA-B (5¢-CGTGAGCTTCCCA AGTCCTCGTCTGTGAAGGCAAGAAGCTCGTCGA GTGG-3¢). Radiolabelling of oligonucleotides In the case of DNA ligase assays with labelled substrates, R-DNA was radiolabelled. A nonphosphorylated R-DNA Fig. 1. Schematic r epresentation of oligonucleotides used for DNA ligase assays. The 5¢ phosphate at the nick is indicated by P. The hyphens in T-DNA were inserted i n the se quence sole ly for a lignment. DN A ligation re actions we re performed u sing th ese oligonucle otides or th eir derivat ives indicated in the respective ®gures. Ó FEBS 2002 Properties of an archaeal DNA ligase (Eur. J. Biochem. 269) 651 was synthesized and phosphorylated at its 5¢ terminus using [c- 32 P]ATP. The oligonucleotide (10 pmol) was phosphor- ylated and radiolabelled b y incubation with 1.85 MBq [c- 32 P]ATP (Amersham Pha rmacia Biotech) and 1 0 U T 4 polynucleotide k inase (MEGALABEL TM , T akara Shuzo, Kyoto, Japan) at 37 °C for 30 min. The reaction product was puri®ed by centrifugation through a CENTRI-SEP Spin Column (Perkin-Elmer Applied Biosystems). DNA ligase assays Ligation activity was measured by using the DNA sub- strates described above. Unless otherwise stated, ligation reaction mixtures (20 lL) contained 20 m M Bicine/KOH pH 8 .0, 15 m M MgCl 2 ,20m M KCl, 1 m M ATP, 10 l M L-DNA, 10 l M R-DNA, 5 l M T-DNA, and 200 n M Lig Tk . The enzyme and other constituents of the reaction mixture were incubated separately at the desired temperature, and reactions were initiated by mixing the two solutions. Standard reactions were carried out at 80 °C for 2 h or at 40 °C for 4 h . The reactions were stopped by addition of 30 lL loading buffer [98% (v/v) formamide, 10 m M EDTA, 0.05% (w/v) xylene cyanol FF] and cooling in ice water. The products (12 lL) were heated at 95 °C for 3 min and then electrophoresed on a denaturing 6% polyacrylamide/7 M urea gel. Super Reading DNA Sequence PreMix Solution (6%) (Toyobo, Osaka, Japan) and Gel-Mix Running Mate Tris/borate/EDTA buffer (Gibco BRL) w ere used for electrophoresis. The gel was stained with ethidium b romide. In experiments determining the kinetic parameters of Lig Tk , ligation reaction mixtures (20 lL) contained 20 m M Bicine/ KOH pH 8.0, 15 m M MgCl 2 ,20m M KCl, 1 m M ATP, and 50 n M Lig Tk . DNA substrate [L-DNA(40), R-DNA(30), and T-DNA(80)] were added at various concentrations in the range 0.5±4 l M . With radiolabelled DNA substrates, 0.1 l M L-DNA, 0.1 l M T-DNA and 0.1 l M labelled R-DNA were used. After electrophoresis, the gel was dried and labelled oligonu- cleotides were detected by autoradiography. The ligation products were quanti®ed by densitometric analysis and QUANTITYONE software (pdi, Huntington Station, NY,U SA). RESULTS AND DISCUSSION In our previous study, w e identi®ed an ATP-dependent DNA ligase (Lig Tk ) from a hyperthermophilic archaeon, T. kodakaraensis KOD1 [3]. It was shown that Lig Tk was able to: (a) catalyse DNA nick-sealing at temperatures up to 100 °C; (b) utilize NAD + as a cof actor; and ( c) form a n Fig. 3. Turnover of Lig Tk . The reactions were performed with nonla- belled oligonucleotides as described in Materials and metho ds. (A) The relationship between template DNA con centratio n and the prod uction of 70-mer DNA. Reaction mixtures (20 lL) containing 20 m M Bicine/ KOH pH 8.0, 15 m M MgCl 2 ,20m M KCl, 1 m M ATP, 5 l M L-DNA(40), 5 l M R-DNA(30), 200 n M Lig Tk and the indicated amount of T-DNA(80) were incubated at 80 °Cfor2h.(B)Theeects of addition of excess duplex DNA on the ligation reaction. Reaction mixtures (20 lL) contained 20 m M Bicine/KOH pH 8.0, 15 m M MgCl 2 ,20m M KCl, 1 m M ATP, 10 l M L-DNA(40), 10 l M R-DNA(30), 5 l M T-DNA(80) and 200 n M Lig Tk , with (right side) or without (left side) exce ss duplex D NA. The d uplex DN A was a m ixture of 10 l M of 50 -mer DNA-A and 10 l M of 50-mer DNA-B. T hese mixtures were incubated at 80 °C and then the products were sampled at 5, 15, 30, 60 and 120 min after the start o f the reaction. Fig. 2. Divalent ca tion speci®city of Lig Tk . Ligation reaction s were performed with dierent divalent cations. R eaction mixtures (20 lL) containing 20 m M Bicine/KOH pH 8.0, 1 m M ATP, 0.1 l M L-DNA(40), 0.1 l M T-DNA(80), 0.1 l M labelled R-DNA(30), 2 00 n M Lig Tk and 15 m M of the indicated divalent cation were incubated at 80 °Cfor2h. 652 M. Nakatani et al. (Eur. J. Biochem. 269) Ó FEBS 2002 enzyme±AMP complex in the presence of ATP and absence of DNA substrate. As still little i s known about DNA ligases from archaea [3,18] or from hyperthermophiles [3,19], we have carried out biochemical and kinetic charac- terization of Lig Tk . Effects of divalent metal cations on the ligation reaction We have previously shown that Mg 2+ supported the DNA ligase activity of Lig Tk [3]. Here, we substituted various divalent metal ca tions for Mg 2+ at a concentration of 15 m M (Fig. 2 ). In comparison to Mg 2+ (100%), Lig Tk could use Mn 2+ (65%) and Sr 2+ (40%) as an alternative cation cofactor to support ligase a ctivity. The enzyme was less active with Ca 2+ (9%), whereas Co 2+ and Zn 2+ failed to support ligation. The optimal cation concentration for Mg 2+ was 15 m M [3], and those for Mn 2+ ,Sr 2+ and Ca 2+ were 25 m M ,25m M and 5 m M , respectively (data not shown). As little difference was found in activity levels between concentration s of 5 m M and 25 m M , t he data in Fig. 2 accurately re¯ect the cation preference of L ig Tk .We observed i nhibitory effects on activity only in the case of Ca 2+ at concentrations above 4 0 m M .Lig Tk could not use Ni 2+ and Cu 2+ , which have been reported not to support activity in previously repo rted DNA ligases (data not shown) [15±20]. The results suggest that Lig Tk preferred alkaline earth metal ions as a cation cofactor. All previously reported DNA ligases have been sh own to use Mg 2+ and Mn 2+ [14±20]. Utilization of Ca 2+ and Co 2+ differ among DNA ligases. It has been reported that theenzymefromThermus thermophilus [20] used Ca 2+ , but not Co 2+ , that the enzymes from Chlorella virus PBCV-1 [16], V accinia v irus [15] and M. thermoautotrophicum [18] could use Co 2+ , but not Ca 2+ , and that the enzymes from Haemophilus in¯uenzae [17] and Aquifex aeolicus [19] could use neither Ca 2+ nor Co 2+ .Thereseemstobenocommon tendency a mong DNA ligases in terms o f divalent cation speci®city. The use of Sr 2+ has not been examined for other enzymes. Interaction between Lig Tk and DNA substrates We have reported previously that Lig Tk displayed DNA ligase activity at temperatures up to 100 °C [3]. However, we observed that the ligation reaction ceased before the complete consumption of the substrates, raising the possi- bility that Lig Tk could not turnover. We addressed this possibility b y examining the ligation reaction by Lig Tk with various amounts of template DNA. As shown in Fig. 3A, the a mount of the ligation product produced by Lig Tk depended s trictly on the amount of T-DNA(80) in the reaction mixture. When T-DNA(80) was present in the reaction mixture at a concentration of 20 l M , the substrates, L-DNA(40) and R-DNA(30), were consumed almost completely and ligated by 0.2 l M of Lig Tk . The concentra- tions of L-DNA(40) and R-DNA(30) were 5 l M each and considerably higher than that of Lig Tk , indicating that Lig Tk turned over. We further performed a kinetic analysis of Lig Tk using various concentrations of L-DNA(40), R-DNA(30) and T-DNA(80) as substrates. The enzyme displayed typical Michaelis±Menten steady-state kinetics with an apparent K m of 1.4 l M for nicked DNA. The k cat value of the enzyme was 0.11ás )1 .TheK m value of Lig Tk was slightly higher than those o f the NAD + -dependent DNA ligases from Pseudoalteromonas haloplanktis (0.296 l M ), E. coli (0.702 l M )andThermus scotoductus (0.465 l M ) [29], and a lso h igher t han t he ATP-dependent DNA ligase from Coprinus cinereus (0.024±0.100 l M ) [30]. The higher appar- ent K m value of Lig Tk is likely to be due to the lower population of double-stranded nicked DNA at higher temperatures, or the different s ubstrates used in each case. The k cat value of Lig Tk was also slightly higher than those of the DNA ligases from P. haloplanktis (0.0337ás )1 ), E. coli (0.0212ás )1 )andT. scotoductus (0.061 3ás )1 )[29]. We further investigated the effects of adding duplex DNA to the reaction mixture. The duplex DNA added to the reaction mixtures did not include nicks a nd were not complementary to any of the substrate oligonucleotides. No inhibition of the ligase reaction c ould b e observed in the presence of duplex DNA (Fig. 3B). Our results support the theory that Lig Tk does not bind strongly to double-stranded DNA and therefore after joining DNA substrates, the enzyme would promptly separate from the duplex DNA produced. Length of oligonucleotides recognized as DNA substrates We investigated the length of oligonucleotides recognized by Lig Tk as DNA substrates. At 80 °C, Lig Tk could ligate oligonucleotides of nine nucleotides or more as L-DNA with an R-DNA of 30 nucleotides, and an R-DNA of eight nucleotides or more with an L-DNA of 30 nucleotides (Fig. 4 A,B). When we performed the same experiments at 40 °C, the enzyme could ligate L-DNA and R-DNA of six or more nucleotides (Fig. 4C). The results o f Fig. 4B,C indicate that an elevation in temperature led to a decrease in ligation products when 6-nucleotide or 7-nucleotide sub- strates were examined. As the activity of Lig Tk itself is higher at 80 °C, it is likely that f ormation of a nicked, duplex DNA substrate, which i s temperature-dependent, is necessary for recognition by Lig Tk and subsequent initiation of the reaction. It has been reported that bacteriophage T7 DNA ligase, which represents one of the smallest known DNA ligases, binds asymmetrically to DNA nicks, extending 3±5 nucle- otides on the 3¢ hydroxyl side of the nick and 7±9 nucleotides on the 5¢ phosphate side [31]. Nick sealing was observed for oligonucleotides of six nucleotides on the 3¢ side to those of n ine nucleotides on the 5¢ side [32]. The enzyme from T. thermophilus could not join oligonucleo- tides of six or fewer nucleotides on the 3¢ side to an oligonucleotide of n ine nucleotides on the 5¢ side [32]. In thecaseofLig Tk at 40 °C, we could detect ligation products as short as 16 nucleotides, the products of a (7 + 9 nucleotide) or (eight nucleotide + 8 nucleotide) combination (Fig. 4D). We had observed previously that Lig Tk could ligate DNA fragments at temperatures above their melting point [3], and the results shown above with the use of short oligonucleo- tides, con®rmed this property. The former results tempted us to speculate that Lig Tk could enhance the formation and/ or stability of duplex DNA substrate a t high temperature [3]. However the results of this study indicate otherwise. Fig. 4B,C suggest that an enzyme-independent formation of a n icked duplex DNA substrate w as necessary for recog- nition by Lig Tk . Furthermore, experiments shown in Ó FEBS 2002 Properties of an archaeal DNA ligase (Eur. J. Biochem. 269) 653 Fig. 3B indicated that Lig Tk did not display af®nity towards double-stranded DNA, and deny a stabilization effect of duplex DNA by Lig Tk . Among the various steps in the r eaction mechanism of Lig Tk ,wehaveclari®edthe following: (a) substrate (nicked, duplex DNA) formation precedes recognition by Lig Tk ; (b) adenylation of the enzyme can occur before enzyme±DNA binding [3]; and (c) after nick-sealing, Lig Tk promptly detaches from the ligation product. Effect of single base mismatches at the nick on the ligation reaction. A mismatched base pair i s structurally distinct from a matched one. Therefore, a 3¢ or 5¢ mismatch at the nick may have drastic effects against the ligation reaction. We investigated the effect of single base mismatches at the nick of DNA substrates on the ligation reaction. In the case of 3¢ mismatched substrates (Fig. 5A), Lig Tk ef®ciently ligated Fig. 4. Length of oligonucleotides recognized by Lig Tk as DNA substrates. The oligonucleotides used in the ligation reactions are described in Fig. 1. The reactions were performed with nonlabelled oligonucleotides as described in Materials and methods. (A,B) Ligation of various oligonucleotides by Lig Tk at 80 °C. Reaction mixtures were incubated at 80 °C for 2 h. Lengths of the oligonucleotides are indicated above the gels. (C,D) Ligation of various oligonucleotides by Lig Tk at 40 °C. The reaction mixtures were incubated at 40 °C for 4 h. Lengths of the oligonucleotides are indicated above the gels. The bands i n dicated by  15-mer i n (D) represent the forefront of migration during electrophoresis and c orrespond to all oligonucleotides of 15 bases or fewer. Fig. 5. Ligation of mismatched substrates by Lig Tk . DNA substrates used in this experiment were derivatives of L-DNA(40), R-DNA(30) and T-DNA(80). The reactions were performed with labelled oligonucleotides as described in Materials and methods. (A) Ligation of 3¢ matched and 3¢ mismatched substrates. The substrates used are indicated at the top. (B) Ligation of 5¢ matched and 5¢ mismatched substrates. The substrat es used are indicated at the top. 654 M. Nakatani et al. (Eur. J. Biochem. 269) Ó FEBS 2002 only the matched substrates. Lig Tk was more tolerant towards 5¢ mismatched substrates (Fig. 5 B). Ef®cient liga- tion was observed with the mismatches, 5¢-T : T, 5¢-G : T, 5¢-T : G, 5¢-A : C, and 5¢-T : C. These results indicated that proper base p airing at the 3¢ side of the nick was necessary for ef®cient ligation by Lig Tk . The ability to discriminate mismatched ends has been investigated for DNA ligases from several organisms using synthetic duplex DNA substrates containing 3¢ or 5¢ mis- matches at their nicks [15,19,23±27]. Lig Tk could not ef®ciently ligate 3¢ mismatched substrates and was more tolerant towards 5¢ mismatched substrates. This tendency has been observed in all previously reported enzymes [15,19,23,26]. Ligation of gapped or inserted DNA substrates by Lig Tk One- or 2-nucleotide gapped substrates were formed by deleting one or two nucleotides from the 3¢ side of L-DNA(40) a nd two nucleotide-insert ed substrates were formed by adding two nucleotides (5¢-CA-3¢)atthe3¢ side of L-DNA(40) [see Fig. 1 for sequence of L-DNA(40)]. In 1- or 2-nucleotide gap ligation s or 2-nucleotide insert ligation, Lig Tk was incapable of catalysing the formation of ligation p roducts (data not shown). Substrates with a 1-nucleotide gap could be ligated by DNA ligases from bacteriophage T4 [21], Chlorella virus PBCV-1 [ 16], Lymantria dispar multicapsid nucleopolyhedrovirus [22] and H. in¯uenzae [17], but not by those from V accinia virus [15], T. thermophilus [20], A. aeolicus [19] and Saccharomyces cerevisiae [25], while no ligation was detectable with 2-nucleotide gapped substrates for all enzymes [15,16,19,20,22]. As for the 1-nucleotide insert ligation, Lig Tk was able to catalyse the ligation reaction under several conditions (Fig. 6 ). However, as expected, activities were small compared to the case of matched substrates. Ligation products were detected in all 5¢ insert ligations, whereas 3¢ insertions tended to i nhibit t he ligation reaction except when the overlapped nucleotides were identical (X  Y), thereby equivalent to a 5¢ insertion. A cytosine at location X also allowed the ligation r eaction to proceed. These results also support t hat the proper base pairing at the 3¢ end of the nick is important for nick-sealing by Lig Tk . It was not clear why reaction activities were detected in the case that r esidue X was cytosine. Ligation of 1-nucleotide i nsert substrates has been p artially investigated for T. thermophilus and A. aeolicus DNA ligases and displayed the same tendencies as Lig Tk [19,20]. Our results with mismatched and inserted DNA substrates display the importance of proper base pairing at the 3¢ hydroxyl side of the nick for the ligation reaction to proceed. DNA metabolism, which includes the replication, repair and recombination of DNA, has been well examined in eukaryotes and bacteria. As DNA ligase plays an important role in all of these events, many studies have been performed on the enzyme f rom various organisms. H owever, in t he case of archaea, knowledge on the me chanisms of DNA metabolism and the individual proteins involved, has yet to accumulate. Our biochemical studies on Lig Tk ,alongwith future studies of the enzyme in vivo, should c ontribute to a better understanding of the mechanisms of DNA metabo- lism in archae a. REFERENCES 1. Lindahl, T. & Barnes, D.E. (1992) Mam malian D NA ligas es. Annu.Rev.Biochem.61, 251±281. 2. Timson, D.J., Singleton, M.R. & Wigley, D.B. (2000) DNA ligases in the repair and replication of DNA. Mutat. Res. 460, 301±318. 3. Nakatani, M., Ezaki, S., Atomi, H. & Imanaka, T. (2000) A DNA ligase from a hyperthermophilic archaeon with unique cofactor speci®city. J. Bacteriol. 182, 6424±6433. 4. Tomkinson, A.E. & Mackey, Z.B. (1998) Structure a nd function of mammalian DNA ligases. Mutat. Res. 407, 1±9. 5. Wilkinson, A., Day, J. & Bowater, R. (2001) Bacterial DNA lig- ases. Mol. Microbiol. 40, 1241±1248. 6. Lehman, I.R. (1974) DNA ligase: structure, mechanism, and function. Science 186, 790±797. 7. Li, J.J. & K elly, T.J. (198 4) Simian virus 40 DNA re plication in vitro. Proc. Natl Acad. Sci. USA 81, 6973±6977. Fig. 6. Ligation of substrates with 1-nucleotide insertions by Lig Tk . DNA substrates used in this experiment were derivatives of L-DNA(40), R-DNA(30) and T-DNA(80). The reactions were per- formed with labelled oligonucleotides as described in Materials and methods. The substrates used are indicated at the top and the activity of each reaction was normalized to the speci®c activity observed with substrates without insertions (de®ned as 100%). Ó FEBS 2002 Properties of an archaeal DNA ligase (Eur. J. Biochem. 269) 655 8. Wood, R.D., Robins, P. & Lindahl, T. (1988) Complementation of the x erod erma pigmentosum DNA repair de fe ct in cell-free extracts. Cell 53, 97±106. 9. Jessberger, R. & Berg, P. (1991) Repair of deletions and double- strand gaps by homologous recombination in a m ammalian in vitro syst em . Mol. Cell. Biol. 11, 445±457. 10. Kletzin, A. (1992) Molecular characterisation of a DNA ligase gene of the extremely thermophilic arch aeon Desulfurolobus ambivalens shows close phylogenetic relationship to eukaryotic ligases. N ucleic Acids Res. 20, 5389±5396. 11. Kaczmarek, F.S., Zaniewski, R.P., Gootz, T.D., Danley, D.E., Mansour, M.N., Grior, M., Kamath, A.V., Cronan, M., Mueller, J., Sun, D., Martin, P.K., Benton, B., McDowell, L., Biek, D. & Sc hmid, M.B. (2001) Cloning and fu nctional charac- terization of an NAD + -dependent D NA ligase from Staphylo- coccus aureus. J. Bacteriol. 183, 3016±3024. 12. Thorbjarnardottir, S.H., Jonsson, Z.O., Andresson, O.S., Kristjansson, J.K., Eggertsson, G. & Palsdottir, A. (1995) Cloning and sequence analysis of the DNA ligase-encoding gene of Rho- dothermus m arinus, and overproduction, puri®cation and charac- terization of two thermophilic DNA ligases. Gene 161 , 1±6. 13. Doherty, A.J., A shford, S.R., Subramanya, H.S. & Wigley, D.B. (1996) Bacteriophage T7 DNA ligase. Overexpression, puri®ca- tion, crystallization, and characterization. J. Biol. Chem. 271, 11083±11089. 14. Takahashi,M.,Yamaguchi,E.&Uchida,T.(1984)Thermophilic DNA ligase. Puri®cation and properties of the enzyme from Thermus thermophilus HB8. J. Biol. Chem. 259, 10041±10047. 15. Shuman, S. (1995) Vaccinia virus DNA ligase: speci®city, ®delity, and inhibition. Biochemistry 34, 16138±16147. 16. Ho, C .K., Van Etten, J.L. & Shuman, S. (1997) Characterization of an ATP-dependent DNA ligase encoded b y Chlorella virus PBCV-1. J. Virol. 71, 1931±1937. 17. Cheng, C. & Shuman, S. (1997) Characterization of an ATP- dependent DNA ligase encoded by Haemophilus in¯uenzae. Nucleic Acids Res. 25 , 1369±1374. 18. Sriskanda, V., Kelman, Z., Hurwitz, J. & Shuman, S. (2000) Characterization of an ATP-dependent DNA ligase from the thermophilic archaeon Metha nobacterium thermoautotrophicum. Nucleic Acids Res. 28 , 2221±2228. 19. Tong, J., Barany, F. & Cao, W. (2000) Ligation reaction speci- ®cities of an NAD + -dependent DNA ligase from the hyp erther- mophile Aquifex aeolicus. Nucleic Acids Res. 28, 1447±1454. 20. Tong, J., Cao, W. & Barany, F. (1999) Biochemical properties of a high ®delity DNA ligase from Thermus species AK16D. Nucleic Acids Res. 27, 788±794. 21. Gon, C., Bailly, V. & Verly, W.G. (1987) Nicks 3¢ or 5¢ to AP sites or to mispaired bases, and one-nucleot ide gaps can be sealed by T4 DNA ligase. Nucleic Acids Res. 15, 8755±8771. 22. Pearson, M.N. & Rohrmann, G.F. (1998) Characterization of a baculovirus- encod ed ATP-dependent DNA ligase. J. Virol . 72, 9142±9149. 23. Sriskanda, V. & Shuman, S. (1998) Spe ci®city and ® delity of strand joining by Chlorella virus DNA ligase. Nucleic Acids Res. 26, 3536±3541. 24. Bhagwat, A .S., Sanderson, R.J. & L indahl, T. ( 1999) Delayed DNA joining at 3 ¢ mismatches by human DNA ligases. N ucleic Acids Res. 27, 4028±4033. 25. Tomkinson, A.E., Tappe, N.J. & Friedberg, E.C. (1992) DNA ligase I from Saccharomyces cerevisiae: physical and biochem ical characterization of the CDC9 gene product. Biochemistry 31, 11762±11771. 26. Luo, J., Bergstrom, D.E. & Barany, F. (1996) Improving t he ®delity of Thermus thermophilus DNA ligase. Nucleic A cids Res. 24, 3071±3078. 27. Husain, I., Tomkinson, A.E., Burkhart, W.A., Moyer, M.B., Ramos,W.,Mackey,Z.B.,Besterman,J.M.&Chen,J.(1995) Puri®cation and characterization of DNA ligase III from bovine testes. Homology w ith DN A ligase I I a nd vaccin ia DN A ligase. J. Biol. Chem. 270, 9 683±9690. 28. Sambrook, J. & Russell, D.W., eds (2001) Molecular Cloning: A Laboratory Manu al, 3rd edn. Cold Spring Harbor La boratory Press, Cold Spring Harbor, New York. 29. Georlette, D., Jo  nsson, Z.O., Van Petegem, F., Chessa, J P., Van Beeumen, J., Hu È bscher, U. & Gerday, C. (2000) A DNA ligase from the p sychrophile Pseudoalteromonas haloplanktis gives insights into the adaptation of proteins to low temperatures. Eur. J. Biochem. 267, 3502±3512. 30. Matsuda, S., Sakaguchi, K., T sukada, K . & T eraoka, H . (1996) Characterization of DNA ligase from the fungus Coprinus cine- reus. Eur. J. Biochem. 237, 691±697. 31. Doherty, A.J. & Daorn, T.R. (2000) Nick recognition by DNA ligases. J. Mol. Biol. 296, 43±56. 32. Pritchard, C.E. & Southern, E.M. (1997) Eects of base mis- matches on joining of short oligodeoxynucleotides by D NA ligases. N ucleic Acids Res. 25, 3403±3407. 656 M. Nakatani et al. (Eur. J. Biochem. 269) Ó FEBS 2002 . Substrate recognition and ®delity of strand joining by an archaeal DNA ligase Masaru Nakatani 1,2 , Satoshi Ezaki 1,2 , Haruyuki Atomi 1,2 and Tadayuki. function of DNA ligase [2]. Therefore DNA ligases are indispensable enzymes in all organisms. DNA ligases fall into two groups, ATP-dependent DNA ligases and