A single amino acid substitution of Leu130Ile in snake DNases I contributes to the acquisition of thermal stability A clue to the molecular evolutionary mechanism from cold-blooded to warm-blooded vertebrates Haruo Takeshita 1, *, Toshihiro Yasuda 2, *, Tamiko Nakajima 1 , Kouichi Mogi 1 , Yasushi Kaneko 1 , Reiko Iida 3 and Koichiro Kishi 1 1 Department of Legal Medicine, Gunma University School of Medicine, Maebashi, Japan; 2 Department of Biology and 3 Department of Legal Medicine, Fukui Medical University, Matsuoka, Japan We purified pancreatic deoxyribonucleases I (DNases I) from three snakes, Elaphe quadrivirgata, Elaphe climaco- phora and Agkistrodon blomhoffii, and cloned their cDNAs. Each mature snake DNase I protein comprised 262 amino acids. Wild-type snake DNases I with Leu130 were more thermally unstable than wild-type mammalian and avian DNases I with Ile130. After substitution of Leu130Ile, the thermal stabilities of the snake enzymes were higher than those of their wild-type counterparts and similar to mam- malian wild-type enzyme levels. Conversely, substituting Ile130Leu of mammalian DNases I made them more thermally unstable than their wild-type counterparts. Therefore, a single amino acid substitution, Leu130Ile, might be involved in an evolutionally critical change in the thermal stabilities of vertebrate DNases I. Amphibian DNasesIhaveaSer205insertioninaCa 2+ -binding site of mammalian and avian enzymes that reduces their thermal stabilities [Takeshita, H., Yasuda, T., Iida, R., Nakajima, T., Mori, S., Mogi, K., Kaneko, Y. & Kishi, K. (2001) Biochem. J. 357, 473–480]. Thus, it is plausible that the thermally stable wild-type DNases I of the higher vertebrates, such as mammals and birds, have been generated by a single Leu130Ile substitution of reptilian enzymes through molecular evolution following Ser205 deletion from amphibian enzymes. This mechanism may reflect one of the evolutionary changes from cold-blooded to warm-blooded vertebrates. Keywords: cDNA cloning; deoxyribonuclease I; molecular evolution; snake; thermal stability. Deoxyribonuclease I (DNase I, EC 3.1.21.1) is an enzyme that preferentially attacks, by Ca 2+ -and Mg 2+ -dependent endonucleolytic cleavage, double-stranded DNA to produce oligonucleotides with 5¢-phospho and 3¢-hydroxy termini [1]. It is considered to play a major role in digestion in the alimentary canal, because, in mammals, it is secreted by exocrine glands such as the pancreas and/or parotid gland [2–7]. However, DNase I also exists outside the alimentary tract [8–11], raising a doubt as to whether its major role in DNA metabolism in vivo is merely digestion. Recently, DNase I was postulated to be responsible for the removal of DNA from nuclear antigens at sites of high cell turnover and thus for the prevention of systemic lupus erythematosus (SLE) [12]. The gene product of human DNASE1*6 was more thermally unstable than that of the other alleles and subjects who were heterozygous for this allele had significantly low serum DNase I activity levels [13]. These findings indicate that the thermal stabilities of DNase I in vitro might reflect the enzyme activities in vivo.We found that amphibian DNases I are characterized by a C- terminal end with a unique cysteine-rich stretch and by insertion of a Ser residue into the Ca 2+ -binding site, resulting in thermal instability compared with DNases I from mammals and birds [14]. Fish DNase I also exhibited similar low thermal stability relative to amphi- bian DNases I (K. Mogi, H. Takeshita, T. Yasuda, T. Nakajima, E. Nakazato, Y. Kaneko, M. Itoi & K. Kishi, personal communication). In these contexts, it would be very interesting how the higher vertebrates, such as mammals and birds, which are also classified as warm- blooded vertebrates, have acquired thermal stability of their DNase I molecules through the evolutionary steps from the lower, cold-blooded, vertebrates, such as amphi- bia and fish. We have already reported the purification and biochemi- cal characterization of mammalian [4,5,7,14–18], avian [19] Correspondence to K. Kishi, Department of Legal Medicine, Gunma University School of, Medicine, Maebashi, Gunma 371–8511, Japan. Fax: + 81 27 220 8035, E-mail: kkoichi@med.gunma-u.ac.jp Abbreviations: aa, amino acid; Con A, Concanavalin A; nt, nucleotide; SLE, systemic lupus erythematosus; SRED, single radial enzyme diffusion. Enzymes: DNase I, (EC 3.1.21.1). Note: The nucleotide sequence data reported will appear in DDBJ, EMBL and GenBank Nucleotide Sequence Databases under accession nos. AB046545, AB050701 and AB058784. *Note: These authors contributed equally to this research and listed in alphabetical order. (Received 23 September 2002, revised 18 November 2002, accepted 25 November 2002) Eur. J. Biochem. 270, 307–314 (2003) Ó FEBS 2003 doi:10.1046/j.1432-1033.2003.03387.x and amphibian [14] DNases I. As the primary structures of amphibian DNases I differ considerably from those of other vertebrate DNases I, comprehensive characterization, including determination of thermal stabilities, of reptilian DNases I is required, not only to elucidate the molecular evolutional aspect of the DNase I family but also to address the queries about how the higher vertebrate DNases I acquired their thermal stabilities described above. In this study, we purified DNases I from the pancreases of three snake species, Elaphe quadrivirgata (Shima-hebi in Japan- ese) and Elaphe climacophora (Aodaisho) of the Colubridae OPPEL and Agkistrodon blomhoffii (Nihon-mamushi) of the Viperidae Laurenti, which are widely distributed in Japan, and cloned the cDNA of each. A single amino acid (aa) substitution was confirmed to affect the thermal stabilities of vertebrate DNases I and, furthermore, one of the postulated mechanisms whereby thermal stability is acquired by a DNase I family at the evolutional step from cold-blooded vertebrates, such as snakes, to warm-blooded ones, such as mammals, is discussed. Materials and methods Materials and biological samples Three different species of snake, E. quadrivirgata, E. climacophora and A. blomhoffii weighing about 210 g (110 cm long), 270 g (130 cm long) and 120 g (70 cm long), respectively, were obtained from the Japan Snake Institute, Gunma, Japan. Phenyl Sepharose CL-4B, DEAE Seph- arose CL-6B and Superdex 75 were purchased from Amersham Pharmacia Biotech; Concanavalin A (Con A)- agarose was from Seikagaku Kogyo (Tokyo, Japan); rabbit muscle G-actin and salmon testis DNA were from Sigma; Superscript II reverse transcriptase (RT), all the oligonu- cleotide primers used, and the RACE systems were from Life Technologies; the Expanded High Fidelity PCR system was from Roche Diagnostics. All the other chemicals used were of reagent grade and available commercially. The snakes and Japanese white rabbits were acquired, main- tained and used in accordance with the Guidelines for the Care and Use of Laboratory Animals (NIH, USA; revised 1985). Analytical methods DNase I activity was assayed by the previously described test tube [15] or single radial enzyme diffusion (SRED) [2] methods, except that 50 m M Tris/HCl buffer, pH 7.5, containing 10 m M MgCl 2 and 1 m M CaCl 2 was substi- tuted for the reaction buffer. Protein concentrations were measured using a protein assay kit (Bio-Rad) with BSA as the standard. The enzymological properties of the snake enzymes and the inhibitory effects of specific antibodies on their activities were examined as described previously [4,15,20]. Samples of 15 different tissues were obtained from each snake as soon as possible after it had been killed by exsanguination under general anesthesia with diethyl ether. Preparation of the samples for the assays and determination of enzyme activity were performed as described previously [7,21–22]. The N-terminal aa se- quences of the purified enzymes were determined by Edman degradation [4]. The presence of DNase I-specific mRNA was verified by RT-PCR amplification of the total RNA extracted from each snake tissue using sets of primers corresponding to the N- and C-terminal aa sequences of the respective enzymes [23]. Purification of DNases I from snake pancreases All the procedures described below were carried out at 0–4 °C. Pancreas samples, weighing approximately 3.4, 1.2, and 0.7 g, were obtained from five individuals each of the species E. quadrivirgata, E. climacophora and A. blomhoffii, respectively. The samples were minced separately and homogenized in 5–10 mL 25 m M Tris/HCl buffer, pH 7.5 (buffer I), containing 1 M ammonium sulfate and 1 m M phenylmethane sulfonyl fluoride. After centrifugation (10 000 g, 20 min), the supernatant (crude extract) was applied to a first phenyl Sepharose CL-4B column (1.6 · 15 cm) pre-equilibrated with the same buffer and the adsorbed materials were eluted with a 300-mL linear reverse ammonium sulfate concentration gradient (1.0–0 M ) in buffer I. The active fractions eluted with about 500 m M ammonium sulfate were collected and solid ammonium sulfate was added to give a concentration of 1.0 M .Thiswas then applied to a second phenyl Sepharose CL-4B column (1 · 10 cm) pre-equilibrated with the same buffer. The DNase I was eluted with 100 mL of the same gradient. The active fractions were collected, dialyzed against buffer I, applied to a DEAE Sepharose CL-6B column (1 · 15 cm) pre-equilibrated with buffer I and the adsorbed materials were eluted with a 100-mL linear NaCl concentration gradient (0–1.0 M ) in buffer I. The active fractions eluted over the NaCl concentration range of 250–300 m M were concentrated using polyethylene glycol 6,000, then subjec- tedtogelfiltrationusingtheA ¨ KTA FPLC system (Amersham Pharmacia Biotech) equipped with a Superdex 75 column (1.6 · 60 cm) with buffer I containing 150 m M NaCl as the eluent. The active fractions were collected and then applied to a Con A-agarose column (1 · 2cm)pre- equilibrated with buffer I containing 150 m M NaCl. The column was washed well with the same buffer and then DNase I was eluted with 300 m M methyl-a- D -mannopyr- anoside in the same buffer. The active fractions were collected and used as the purified enzymes for the subsequent experiments. A specific rabbit antibody against purified DNase I from E. quadrivirgata was prepared as described previously [15]. Construction of the cDNA species encoding E. quadrivirgata , E. climacophora and A. blomhoffii DNases I Total RNA was isolated from each snake pancreas by the acid guanidinium thiocyanate/phenol/chloroform method [24] and any DNA contamination was removed by treat- ment with RNase-free DNase I (Stratagene). The 3¢-end region of cDNAs of all the snakes were obtained by 3¢-RACE method using two degenerate primers based on aa sequences that are highly conserved in vertebrate DNase I, the Tyr97–Cys104 and Met166–Cys173 sequences of human DNase I [25]. Next, the 5¢-end regions of the cDNAs were amplified by the 5¢-RACE method using 308 H. Takeshita et al. (Eur. J. Biochem. 270) Ó FEBS 2003 gene-specific primers based on the nucleotide (nt) sequences determined in this study. These RACE procedures were carried out using the 3¢-and5¢-RACE systems described above, according to the manufacturer’s instructions. The RACE products were subcloned into the pCR II TA cloning vector (Invitrogen, San Diego, CA, USA) and sequenced. The nt sequences were determined by the dideoxy chain-termination method using a Dye Terminator Cycle sequencing kit (Applied Biosystems, Urayasu, Japan). The sequencing run was performed on a Genetic Analyzer (model 310, Applied Biosystems) and all the DNA sequences were confirmed by reading both strands. Construction of expression vectors and transient expression of the constructs in COS-7 cells A DNA fragment containing the entire coding sequence of E. quadrivirgata DNase I cDNA was prepared from the total RNA derived from the pancreas by RT-PCR amplification using an Expanded High Fidelity PCR system with a set of two primers, 5¢-GAATTCGAGGCC ATGAAGACCATCTTG-3¢ (sense) and 5¢-CTCGAGG GGCTCAGGTGGATTTTAGG-3¢ (antisense), corres- ponding to the nt sequences of the cDNA from positions 28–48 and 867–885, respectively. The amplified fragment was ligated into the pcDNA3.1 (+) vector (Invitrogen) to construct the expression vector. Six other expression vectors with cDNA inserts encoding E. climacophora, A. blomhoffii, Xenopus laevis, human, rat and mouse DNases I, were prepared in the same manner. Two substitution mutants for surveying the G-actin binding site at aa position 67, E. quadrivirgata (Ile67Val) and A. blomhoffii (Val67Ile), in which an Ile or Val residue was substituted with Val or Ile, respectively, at aa position 67 in E. quadrivirgata and A. blomhoffii DNases I, respectively, were constructed using the splicing by overlap extension method [26] with the corresponding wild-type construct as a template. Ten mutants for surveying the aa sites responsible for thermal stability at positions 130 and 166, E. quadrivirgata (Leu130Ile), A. blomhoffii (Leu130Ile), human (Ile130Leu), rat (Ile130Leu), mouse (Ile130Leu), E. quadrivirgata (Leu166Met), A. blomhoffii (Leu166Met), human (Met166Leu), rat (Met166Leu) and mouse (Met166Leu), and four double substitution mutants, E. quadrivirgata (Leu130Ile/Leu166Met), rat (Ile130Leu/Met166Leu), mouse (Ile130Leu/Met166Leu) and human (Ile130Leu/Met166Leu), were prepared in the same manner. All the constructs had their sequences confirmed and were purified for transfection using the Plasmid Midi kit (Qiagen). Transient expression of the constructs in COS-7 cells followed by analysis of the enzyme was performed as described previously [27]. All transfections were performed in triplicate with at least two different plasmid preparations. Phylogenetic analysis The nt sequences of the DNase I cDNA of human [28], mouse [17], rat [29], rabbit [5], pig [16], fish (Oreochromis mossambicus) [30], cow [18], chicken [19], two frogs, toad and newt [14] with the following respective database accession numbers EMBL M55983, EMBL U00478, EMBL X56060, EMBL D82875, EMBL AB048832, EMBL AJ001305, EMBL AJ001538, EMBL AB013751, EMBL AB030958, EMBL AB038776, EMBL AB045037 and EMBL AB041732 were obtained. Phylogenetic trees with the nt sequence of the open reading frame (ORF) of their cDNAs and the corresponding aa sequences, in Table 1. Summary of the purification of DNases I from the pancreases of three species of snake. The results of the sequential enzyme purification procedure, using pancreases obtained from five individuals of each species as starting material, are summarized. Species and purification step Protein (mg) Total activity (U) Specific activity (UÆmg )1 ) Purification (fold) Yield (%) E. quadrivirgata Crude extract 600 1100 1.8 1.0 100 Phenyl Sepharose CL-4B 120 970 8.1 4.5 88 28 950 34 20 86 DEAE Sepharose CL-6B 5.4 780 140 80 71 Superdex 75 0.55 770 1400 780 70 Con-A agarose 0.3 750 2500 1400 68 A. blomhoffii Crude extract 220 400 1.8 1.0 100 Phenyl Sepharose CL-4B 45 350 7.8 4.3 88 10 330 33 18 83 DEAE Sepharose CL-6B 2.0 210 110 58 53 Superdex 75 0.23 200 870 480 50 Con-A agarose 0.05 190 3800 2100 48 E. climacophora Crude extract 110 290 2.6 1 100 Phenyl Sepharose CL-4B 18 280 16 6 97 5 270 51 20 88 DEAE Sepharose CL-6B 0.5 160 320 122 55 Superdex 75 0.1 120 1200 460 41 Con-A agarose 0.02 100 5000 1900 34 Ó FEBS 2003 A mechanism from cold-blooded to warm-blooded vertebrates (Eur. J. Biochem. 270) 309 which the region of the putative signal peptide was not included, aligned by the neighbor-joining algorithm using the CLUSTALW program were constructed [31,32]. Results and discussion Purification and characterization of snake pancreatic DNases I The purification results for each snake pancreatic DNase I are summarized in Table 1. This procedure, using four different types of column chromatography, allowed all three snake DNases I to be easily and reproducibly isolated and purified to electrophoretic homogeneity (Fig. 1, A1), representing 1400- to 2100-fold purification. Tandem phenyl Sepharose chromatographies without loss of total enzyme activity were a particularly effective step for purifying each snake enzyme. Both gel filtration and SDS/PAGE analysis showed that E. quadrivirgata DNaseIhadamolecularmassof35kDa.ItspH optimum of 7.5 was higher than those of mammalian enzymes [4,5,7,15–17] and lower than those of amphibia (pH 8.0) [14]. Edman degradation of the purified E. quad- rivirgata and A. blomhoffii enzymes revealed the same N-terminal aa sequences over the first 10 cycles: Leu1- Arg-Ile-Gly-Ala-Phe-Asn-Ile-Arg-Ala10. Although G-actin is known to be a potent inhibitor of human [15], cow [33] and mouse [17] DNases I, the activities of rat [4], porcine [16], chicken [19] and amphibian [14] DNases I were unaffected by G-actin. G-actin (3.2 n M ) abolished the enzyme activity of human DNase I (0.3 units) and reduced that of A. blomhoffii wild-type DNase I to 50% of its initial level, but did not affect the activities of E. quadrivirgata or E. climacophora wild-type DNases I. It has been suggested that two aa residues (Tyr65 and Val67) are mainly responsible for actin binding in human and bovine DNases I [34,35]. The latter residue (Ile67) in both E. quadrivirgata and E. climacophora DNase I was substituted. In comparison with the susceptibility of each wild-type enzyme to G-actin, the susceptibility of the E. quadrivirgata (Ile67- Val) enzyme was increased to the level of the wild-type A. blomhoffii enzyme, whereas the A. blomhoffii (Val67Ile) enzyme and E. quadrivirgata wild-type enzymes were equally susceptible. These findings show that the presence of Val67 is one of the essential factors responsible for the actin-binding capacity of these snake DNases I. Tissue distribution of snake DNases I The DNase I activities in the 15 different tissues from each snake were determined. The activity detected in the pancreas of each snake was over three orders of magni- tude greater than that in the small intestine. However, other tissues listed in Fig. 1 exhibited no DNase I activity under our assay conditions. These enzyme activities were abolished by 20 m M EDTA, 5 m M EGTA and the appropriate specific anti-DNase I Ig, which confirmed they were due to DNase I. The presence of DNase I-specific mRNA was verified by RT-PCR analysis of the total RNA extracted from each snake tissue (Fig. 1B). Specific PCR products were amplified only from the pancreatic and small intestinal total RNAs of the three snakes. No amplified products were obtained from the RNAs of the other tissues. The restriction of DNase I gene expression to only two tissues, the pancreas and small intestine, in snakes and amphibia [14] contrasts with the situation in mammals and birds, in which more widespread expression has been observed in various tissues, including the kidney, liver and stomach, as well as the pancreas and small intestine [2,7,19]. cDNA structures encoding three snake DNases I and expression of the DNase I cDNAs in COS-7 cells The total RNAs isolated from the pancreases of the three snake species were amplified separately by the 3¢-and 5¢-RACE methods to construct cDNAs encoding DNases I. The use of two primers based on aa sequences Fig. 1. Electrophoretic patterns of snake purified DNases I and recombinant enzymes (A), and RT-PCR analysis of the total RNA from several tissues of E. quadrivirgata (B). (A1) Purified DNase I (about 1 lg) from human urine (lane 1), E. quadrivirgata pancreas (lane 2), E. climacophora pancreas (lane 3) and A. blomhoffii pancreas (lane 4) were subjected to SDS/PAGE using a 12.5% gel [20], followed by silver staining. (A2) Recombinant E. quadrivirgata, E. climacophora and A. blomhoffii DNases I expressed in COS-7 cells were subjected to activity staining for DNase I using a DNA-cast PAGE method [21,22]: aliquots containing about 0.2 units of activity of the purified (lanes 1, 3 and 5) and recombinant (lanes 2, 4 and 6) enzymes were used. Lanes 1 and 2, E. quadrivirgata enzyme; lanes 3 and 4, E. climacophora enzyme; lanes 5 and 6, A. blomhoffii enzyme. The cathode is at the top. (B) The total RNA isolated from several tissues, including the pancreas (lane 2), small intestine (lane 3), liver (lane 4), kidney (lane 5), large intestine (lane 6), stomach (lane 7) and parotid gland (lane 8) of E. quadrivirgata was reverse-transcribed and PCR-amplified with a set of specific primers. A unique 850-bp fragment corresponding to the region encoding E. quadrivirgata DNase I was amplified from only the pancreas and small intestine: the cerebrum, heart, lung, spleen, skin, muscle, esophagus and Harder’s gland gave no amplified fragment. Also, RT-PCR analysis of the corresponding set of tissues from E. climacophora and A. blomhoffii exhibited the same results as E. quadrivirgata (data not shown). Lane 1 contains a DNA marker derived from /X174 DNA digested with HaeIII. 310 H. Takeshita et al. (Eur. J. Biochem. 270) Ó FEBS 2003 that are highly conserved in vertebrate DNases I allowed successful amplification of specific RACE products from the total RNA of each species. The full-length cDNA encoding E. quadrivirgata DNase I (accession number AB046545) comprised 1071 bp, including an ORF of 849 bp coding for 283 aas, a 33-bp 5¢-untranslated region (UTR) and a 189-bp 3¢-UTR. The sequence flanking the first ATG (positions 34 –36) was in accordance with the Kozak consensus for a translation start site [36]. We also cloned and sequenced the cDNA species encoding the DNases I of A. blomhoffii (accession number AB050701) and E. climacophora (accession number AB058784) and found full-length sequences of 1050 and 1071 bp, respect- ively. The entire nt sequences of both the ORF and 5¢-UTR regions of E. climacophora DNase I cDNA were identical to those of E. quadrivirgata, but 8.4% (16/189) of the entire nt sequence in the 3¢-UTR of the former was different from that of the latter. The N-terminal aa sequences determined chemically from the purified enzymes exactly matched those deduced from the cDNA data of E. quadrivirgata and A. blomhoffii, indicating that each putative upstream signal sequence containing the first Met residue was 20 aa residues long. About 12% (33/283) of the aa residues in the entire sequence of A. blomhoffii deduced from its cDNA data differ from those of E. quadrivirgata and E. climacophora. Each expression vector containing the entire coding region of E. quadrivirgata, E. climacophora or A. blomhoffii DNase I cDNA was transiently transfected into COS-7 cells. The snake enzyme activities expressed in the COS-7 cells were abolished by the appropriate specific antibodies and, furthermore, they migrated to positions corresponding to the purified pancreatic enzymes on the DNA-cast PAGE gel (Fig. 1A2), confirming that the isolated cDNAs did indeed encode the expected snake DNases I. Comparison of the predicted primary structure (Fig. 2) with human, chicken and frog sequences allowed us to demonstrate several common structural features unique to the snake enzymes. The four residues responsible for the catalytic activity of the other vertebrate DNases I, Glu78, His134, Asp212 and His252 [5,14–19,25,28,33], were conserved in all the snake enzymes. Cys173 and Cys209, which form the disulfide bond responsible for the stability of the enzyme [37], and also Arg41 and Tyr76, that mediate DNase I– DNA contact in the other vertebrate DNases I and orientate the scissile phosphate relative to the enzyme [38], were also found in all the snake enzymes. The unique cysteine-rich-terminus and inserted Ser in the Ca 2+ -binding site of amphibian DNases I [14], were not observed in the snake enzymes. As in mammalian, but not amphibian DNases I [14–19,25,28,33], two potential N-linked glycosy- lation sites, Asn18 (Asn-Gln-Thr) and Asn106 (Asn-Gly/ Thr-Thr), were well present in the snake enzymes, which also showed high affinity for Con A–lectin. These findings indicate that, with respect to their structural relationships, snake DNases I are far from amphibian enzymes, but close to mammalian and avian DNases I. Thermal stabilities of wild-type and substitution mutant snake DNases I The thermal stabilities of the wild-type and mutant enzymes of the snakes were examined by measuring the activities remaining after incubation for 40 min at various temperatures (Fig. 3). Wild-type snake and amphibian DNases I are more thermally unstable than those of higher vertebrates, such as the human, rabbit, rat, mouse and chicken [4,5,15,17,19]. When the primary structures of the three snake enzymes were compared with those of other vertebrates, only two aa residues, Leu130 and Leu166, of Fig. 2. Alignment of the amino acid sequences of snake DNases I with those of human, chicken and X. laevis DNases I. The aa sequences of the snake DNases I were deduced from their respective cDNAs and compared with those published for human, chicken and X. laevis DNases I. Position 1 aa was assigned by comparison with the N-terminal aa sequences determined chemically from the purified enzymes. The aa sequences of E. climacophora DNase I are not shown because they were identical to those of the E. quadrivirgata enzyme. Alignment of the sequences was performed using the Genedoc program (available at http://www.psc.edu/biomed/genedoc/). Ó FEBS 2003 A mechanism from cold-blooded to warm-blooded vertebrates (Eur. J. Biochem. 270) 311 the former were observed in all the lower vertebrates studied, i.e. four amphibia [14] and one fish [37], which are all classified as cold-blooded vertebrates. These two residues were replaced by Ile130 and Met166, respectively, in warm-blooded vertebrates, i.e. the human [28], cow [18], pig [16], rabbit [5], rat [4], mouse [17] and chicken [19]. These findings furnished us with a clue to a mechanism whereby DNases I have evolved, i.e. Leu130Ile and/or Leu166Met might convert the thermally unstable DNases I of cold-blooded vertebrates to the thermally stable ones of warm-blooded vertebrates. Therefore, we constructed a series of substitution mutant enzymes (Fig. 3). In brief, E. quadrivirgata, A. blomhoffii and X. laevis wild-type DNases I were all less thermally stable than human, rat and mouse wild-type enzymes. The mutant enzymes E. quadrivirgata (Leu130Ile), E. quadrivirgata (Leu130Ile/ Leu166Met) and A. blomhoffii (Leu130Ile) were all more thermally stable than the corresponding wild-type DNases I, whereas E. quadrivirgata (Leu166Met) and A. blomhoffii (Leu166Met) were as thermally unstable as their wild-type counterparts. These results suggest that Leu130Ile conferred increased thermal stabilities to the snake enzymes, but Leu166Met did not. Conversely, the human (Ile130Leu) and human (Ile130Leu/ Met166Leu) mutants were more thermally unstable than their wild-type counterparts, whereas human (Met166Leu) was not. The same was true for these mutants of the rat and mouse DNase I (Fig. 3). These findings demonstrate that the nature of the amino acid at position 130 may generally and markedly affect the thermal stabilities of vertebrate DNases I. The 3D structure of DNase I based on X-ray structure analysis of the bovine enzyme has demonstrated that the central core of DNase I is formed by two tightly packed six-stranded b-sheets and that the extended hydrophobic core is mainly responsible for the structural stability and rigidity of DNase I [37,39]. The aa residue at position 130 is located in the central core, whereas that at position 166 is not. Accordingly, it could be predicted that a substitution of the former residue might induce some alterations in the structural stability of DNase I, whereas that of the latter would not. These predictions were found to be compatible with the experimental results described above. Therefore, these facts suggest that the aa residue at position 130 may be responsible for the thermal stability of DNase I. We have reported another mechanism of generating a thermally stable enzyme form from a thermally unstable one: frog, toad and newt DNases I all have a Ser205 insertion in a domain that contains an essential Ca 2+ - binding site in the mammalian enzymes and are thermally Fig. 3. Comparison of the thermal stabilities of wild-type and mutant DNases I derived from snakes (A), human (B) and other vertebrates (C). Each wild-type and mutant DNase I sample (1.0 unit) was incubated in 50 m M Tris/HCl buffer, pH 7.5, for 40 min at various temperatures, as indicated in the figure, using a Dry Thermo Unit DTU-2B (TAI- TEC, Saitama, Japan), and then its residual activity was measured by the SRED method (2). The temperature of thermal denaturation (T 1/2 ) is defined as that at which the DNase I activity is halved and shown in the figure. The value for each enzyme represents triplicate determina- tions and the assay precision was estimated to be within 10%. (A) E. quadrivirgata wild-type (s), E. quadrivirgata (Leu130Ile) (d), E. quadrivirgata (Leu130Ile/Leu166Met) (j)andE. quadrivirgata (Leu166Met) (h) DNases I. The thermal stabilities of A. blomhoffii wild-type, A. blomhoffii (Leu130Ile) and A. blomhoffii (Leu166Met) enzymes were similar to those of the corresponding E. quadrivirgata enzymes. (B) Human wild-type (s), human (Ile130Leu) (d), human (Ile130Leu/Met166Leu) (j) and human (Met166Leu) (h)DNasesI. (C) Rat wild-type (s), rat (Met166Leu) (h), rat (Ile130Leu) (d), rat (Ile130Leu/Met166Leu) (j)andX. laevis wild-type (m)DNasesI. The thermal stabilities of the mouse wild-type, mouse (Ile130Leu), mouse (Ile130Leu/Met166Leu) and mouse (Met166Leu) enzymes were very similar to those of the corresponding rat enzymes. 312 H. Takeshita et al. (Eur. J. Biochem. 270) Ó FEBS 2003 unstable [14]. Insertion of a corresponding Ser residue between Ala204 and Thr205 of human and rat DNases I reduced their thermal stabilities to levels similar to those of amphibian enzymes. These findings led us to conclude that there are at least two mechanisms that might be involved in changing the thermally stable characteristics of vertebrate DNases I, substitution of Ile130Leu in snakes and insertion of Ser205 in amphibia. It is interesting that DNases I of warm-blooded vertebrates, such as humans, pigs, rabbits, rats and chicken, are all thermally stable, while those of cold-blooded vertebrates, such as snakes, frogs, toads and newts, are all thermally unstable. It could be postulated that the thermally stable DNases I of the higher vertebrates must have been produced principally by the Leu130Ile substitu- tion first in avian enzymes at the evolutionary stage from reptiles to birds after deletion of Ser205 from the enzymes of amphibians as they evolved into reptiles. Thermal stability of the enzyme might be evaluated as one of the factors that reflect the DNase I activity levels in vivo [13].Asalackof,or decrease in, DNase I activity has been suggested to be a critical factor in the initiation of human and rat SLE [12,40], the acquisition of thermally stable characteristics during DNase I evolution may provide a clue to the etiology of SLE in humans and mice, which are classified as warm- blooded vertebrates. Phylogenetic analyses of interspecies variations in the DNase I family Based on both the aa and nt sequences of 14 vertebrate DNases I, the phylogenetic trees for the DNase I family were constructed (Fig. 4). The bootstrap values calculated in the tree based on the aa sequence were lower than those in the latter tree, and an alignment of nt sequences was found to be more adequate for molecular evolutionary analysis of the DNase I family. The mammalian group formed a relatively tight cluster, while the snake (E. quadrivirgata, E. climacophora and A. blomhoffii), amphibian (X. laevis, Rana catesbeiana, Bufo vulgaris japonicus and Cynops pyrrhogaster), avian (chicken) and fish (O. mossambicus) DNases I were individually situated at independent posi- tions far from the mammalian DNase I cluster. The snake enzymes are placed closer to the avian than the amphibian forms. With regard to the evolutionary origin of birds, conflicting results of phylogenetic analysis supporting a bird–mammal or bird–reptile relationship have been repor- ted [41,42]. However, our data based on the nt sequences of DNase I molecules may provide evidence of a bird–reptile rather than bird–mammal relationship. Acknowledgements We thank Dr Atsushi Sakai, The Japan Snake Institute, Gunma, Japan, for providing us three kinds of snake. We thank Mrs Masako Itoi and Miss Emiko Nakazato for their excellent technical assistance. This work was supported in part by Grants-in-Aid from Japan Society for the Promotion of Science (12770216 to H. T., 12307011, 14657111 to K. K. and 12357003 to T. Y.) and grants from the Japan Science Society 2001 to K. M. (JSS-12-153), Uehara Memorial Foundation 2002 to H. T. and Daiwa Securities Health Foundation to K. K. References 1. Moore, S. (1981) Chapter 15. Pancreatic DNase. In The Enzymes (Boyer, P.D., ed), Vol. 14, 3rd edn. pp. 281–296. Academic Press, New York, USA. 2.Nadano,D.,Yasuda,T.&Kishi,K.(1993)Measurementof deoxyribonuclease I activity in human tissues and body fluids by a single radial enzyme-diffusion method. Clin. Chem. 39, 448–452. 3. Takeshita, H., Yasuda, T., Nadano, D., Tenjo, E., Sawazaki, K., Iida, R. & Kishi, K. (1994) Detection of deoxyribonuclease I and II (DNases I and II) activities in reproductive organs of male rabbits. Int. J. Biochem. 26, 1025–1031. 4. Takeshita, H., Yasuda, T., Nadano, D., Iida, R. & Kishi, K. (1995) Deoxyribonuclease I from rat urine: affinity purification, characterization, and immunochemical studies. J. Biochem. 118, 932–938. 5. Yasuda, T., Takeshita, H., Nakajima, T., Hosomi, O., Nakashima, Y. & Kishi, K. (1997) Rabbit DNase I: purification from urine, immunological and proteochemical characterization, nucleotide sequence, expression in tissues, relationships with other mamma- lian DNases I and phylogenetic analysis. Biochem. J. 325, 465–473. 6. Tsutsumi, S., Kaneko, Y., Asao, T., Kuwano, H., Kudo, S., Takeshita, H., Yasuda, T. & Kishi, K. (2001) DNase I is present in the chief cells of human and rat stomachs. Histochem. J. 33, 531–535. 7. Takeshita, H., Mogi, K., Yasuda, T., Nakajima, T., Nakashima, Y., Mori, S., Hoshino, T. & Kishi, K. (2000) Mammalian deoxy- ribonucleases I are classified into three types: pancreas, parotid, Fig. 4. Phylogenetic analysis of the vertebrate DNase I family based on their aa (A) and nt (B) sequences. The nt and aa sequences of the DNase I proteins of various species were subjected to phylogenetic analysis. The numbers in figure represent bootstrap values calculated using the CLUSTALW program. Frog, Rana catesbeiana; toad, Bufo vulgaris japonicus;newt,Cynops pyrrhogaster;fish(tilapia),Oreochromis mossambicus. Ó FEBS 2003 A mechanism from cold-blooded to warm-blooded vertebrates (Eur. J. Biochem. 270) 313 and pancreas-parotid (mixed), based on differences in their tissue concentrations. Biochem. Biophys. Res. Commun. 269, 481–484. 8. Zhivotovsky, B., Wade, D., Nicotera, P. & Orrenius, S. (1994) Role of nucleases in apoptosis. Int. Arch. Allergy Immunol. 105, 333–338. 9. Yasuda, T., Sawazaki, K., Nadano, D., Takeshita, H., Nakanaga, M. & Kishi, K. (1993) Human seminal deoxyribonuclease I (DNase I): purification, enzymological and immunological char- acterization and origin. Clin. Chim. Acta 218, 5–16. 10. Polzar, B., Zanotti, S., Stephan, H., Rauch, F., Peitsch, M.C., Irmler, M., Tschopp, J. & Mannherz, H.G. (1994) Distribution of deoxyribonuclease I in rat tissues and its correlation to cellular turnover and apoptosis (programmed cell death). Eur. J. Cell Biol. 64, 200–210. 11. Peitsch, M.C., Irmler, M., French, L.E. & Tschopp, J. (1995) Genomic organisation and expression of mouse deoxy- ribonuclease I. Biochem. Biophys. Res. Commun. 207, 62–68. 12. Napirei, M., Karsunky, H., Zevnik, B., Stephan, H., Mannherz, H.G. & Mo ¨ ro ¨ y, T. (2000) Features of systemic lupus erythematosus in DNase1-deficient mice. Nat. Genet. 25, 177–181. 13. Yasuda, T., Takeshita, H., Iida, R., Kogure, S. & Kishi, K. (1999) A new allele, DNASE1*6, of human deoxyribonuclease I poly- morphism encodes an Arg to Cys substitution responsible for its instability. Biochem. Biophys. Res. Commun. 260, 280–283. 14. Takeshita, H., Yasuda, T., Iida, R., Nakajima, T., Mori, S., Mogi, K., Kaneko, Y. & Kishi, K. (2001) Amphibian DNases I are characterized by a C-terminal end with a unique, cysteine-rich stretch and by the insertion of a serine residue into the Ca 2+ - binding site. Biochem. J. 357, 473–480. 15. Yasuda,T.,Awazu,S.,Sato,W.,Iida,R.,Tanaka,Y.&Kishi,K. (1990) Human genetically polymorphic deoxyribonuclease: pur- ification, characterization, and multiplicity of urine deoxy- ribonuclease I. J. Biochem. 108, 393–398. 16. Mori, S., Yasuda, T., Takeshita, H., Nakajima, T., Nakazato, E., Mogi, K., Kaneko, Y. & Kishi, K. (2001) Molecular, biochemical and immunological analyses of porcine pancreatic DNase I. Biochim. Biophys. Acta 1547, 275–287. 17. Takeshita, H., Yasuda, T., Nakajima, T., Hosomi, O., Nakashima, Y. & Kishi, K. (1997) Mouse deoxyribonuclease I (DNase I): biochemical and immunological characterizations, cDNA structure and tissue distribution. Biochem. Mol. Biol. Int. 42, 65–75. 18. Chen,C Y.,Lu,S C.&Liao.T H.(1998)Cloning,sequencing and expression of a cDNA encoding bovine pancreatic deoxy- ribonuclease I in Escherichia coli: purification and characterization of the recombinant enzyme. Gene 206, 181–184. 19. Nakashima, Y., Yasuda, T., Takeshita, H., Nakajima, T., Hosomi, O., Mori, S. & Kishi, K. (1999) Molecular, biochemical and immunological studies of hen pancreatic deoxyribonuclease I. Int. J. Biochem. Cell Biol. 31, 1315–1326. 20. Laemmli, U.K. (1970) Cleavage of structural proteins during the assembly of the head of bacteriophage T 4 . Nature (London) 227, 680–685. 21. Nakajima, T., Yasuda, T., Nakashima, Y., Hosomi, O., Takesh- ita, H. & Kishi, K. (1998) Two novel screening methods for selecting monoclonal antibodies which specifically inhibit DNase I enzyme activity. Immunol. Invest. 27, 145–152. 22. Nakajima, T., Yasuda, T., Takeshita, H., Nakashima, Y., Mori, S., Mogi, K. & Kishi, K. (2000) Rapid purification of human DNase I using mouse monoclonal anti-DNase I antibodies and characterization of the antibodies. Exp. Clin. Immunogenet. 17, 71–76. 23. Yasuda, T., Takeshita, H., Iida, R., Nakajima, T., Hosomi, O., Nakashima, Y. & Kishi, K. (1998) Molecular cloning of the cDNA encoding human deoxyribonuclease II. J. Biol. Chem. 273, 2610–2616. 24. Chomczynski, P. & Sacchi, N. (1987) Single-step method of RNA isolation by acid guanidinium thiocyanate-phenol-chloroform extraction. Anal. Biochem. 162, 156–159. 25. Yasuda, T., Kishi, K., Yanagawa, Y. & Yoshida, A. (1995) Structure of the human deoxyribonuclease I (DNase I) gene: identification of the nucleotide substitution that generates its classical genetic polymorphism. Ann. Hum. Genet. 59, 1–15. 26. Ho, S.N., Hunt, H.D., Horton, R.M., Pullen, J.K. & Pease, L.R. (1989) Site-directed mutagenesis by overlap extension using the polymerase chain reaction. Gene 77, 51–59. 27. Takeshita, H., Yasuda, T., Nakazato, E., Nakajima, T., Mori, S., Mogi, K., Kaneko, Y., Iida, R. & Kishi, K. (2001) Use of human recombinant DNase I expressed in COS-7 cells as an immunogen to produce a specific anti-DNase I antibody. Exp. Clin. Immunogenet. 18, 226–232. 28. Shak, S., Capon, D.J., Hellmiss, R., Marsters, S.A. & Baker, C.L. (1990) Recombinant human DNase I reduces the viscosity of cystic fibrosis sputum. Proc. Natl Acad. Sci. USA 87, 9188–9192. 29. Polzar, B. & Mannherz, H.G. (1990) Nucleotide sequence of a full length cDNA clone encoding the deoxyribonuclease I from the rat parotid gland. Nucleic Acids Res. 18, 7151. 30. Hsiao, Y M., Ho, H C., Wang, W Y., Tam, M.F. & Liao, T H. (1997) Purification and characterization of tilapia (Oreochromis mossambicus) deoxyribonuclease I–primary structure and cDNA sequence. Eur. J. Biochem. 249, 786–791. 31. Saitou, N. & Nei, M. (1987) The neighbor-joining method: a new method for reconstructing phylogenetic trees. Mol. Biol. Evol. 4, 406–425. 32. Thompson, J.D., Higgins, D.G. & Gibson, T.J. (1994) CLUSTAL W: improving the sensitivity of progressive multiple sequence alignment through sequence weighting, position-specific gap penalties and weight matrix choice. Nucleic Acids Res. 22, 4673– 4680. 33. Paudel, H.K. & Liao, T H. (1986) Comparison of the three pri- mary structures of deoxyribonuclease isolated from bovine, ovine, and porcine pancreas. Derivation of the amino acid sequence of ovine DNase and revision of the previously published amino acid sequence of bovine DNase. J. Biol. Chem. 261, 16012–16017. 34. Mannherz, H.G. (1992) Crystallization of actin in complex with actin-binding proteins. J. Biol. Chem. 267, 11661–11664. 35. Ulmer, J.S., Herzka, A., Toy, K.J., Baker, D.L., Dodge, A.H., Sinicropi,D.,Shak,S.&Lazarus,R.A.(1996)Engineeringactin- resistant human DNase I for treatment of cystic fibrosis. Proc. NatlAcad.Sci.USA93, 8225–8229. 36. Kozak, M. (1987) At least six nucleotides preceding the AUG initiator codon enhance translation in mammalian cells. J. Mol. Biol. 196, 947–950. 37. Oefner, C. & Suck, D. (1986) Crystallographic refinement and structure of DNase I at 2 A ˚ resolution. J. Mol. Biol. 192, 605– 632. 38.Pan,C.Q.,Ulmer,J.S.,Herzka,A.&Lazarus,R.A.(1998) Mutational analysis of human DNase I at the DNA binding interface: implications for DNA recognition, catalysis, and metal ion dependence. Protein Sci. 7, 628–636. 39. Suck, D., Oefner, C. & Kabsch, W. (1984) Three-dimensional structure of bovine pancreatic DNase I. EMBO J. 3, 2423–2430. 40.Yasutomo,K.,Horiuchi,T.,Kagami,S.,Tsukamoto,H., Hashimura, C., Urushihara, M. & Kuroda, Y. (2001) Mutation of DNASE1 in people with systemic lupus erythematosus. Nat. Genet. 28, 313–314. 41. Hedges, S.B., Moberg, K.D. & Maxson, L.R. (1990) Tetrapod phylogeny inferred from18S and 28S ribosomal RNA sequences and a review of the evidence for amniote relatioships. Mol. Biol. Evol. 7, 607–633. 42. Hedges, S.B. (1994) Molecular evidence for the origin of birds. Proc.NatlAcad.Sci.USA91, 2621–2624. 314 H. Takeshita et al. (Eur. J. Biochem. 270) Ó FEBS 2003 . A single amino acid substitution of Leu130Ile in snake DNases I contributes to the acquisition of thermal stability A clue to the molecular evolutionary. close to mammalian and avian DNases I. Thermal stabilities of wild-type and substitution mutant snake DNases I The thermal stabilities of the wild-type and