Purification, Characterization, and Molecular Cloning of Lactonizing Lipase from Pseudomonas Species*

An extracellular lipase catalyzing the synthesis of macrocyclic lactones in anhydrous organic solvents was purified to homogeneity from Pseudomonas nov. sp. 109, and characterized. The lipase showed a PI of 6.3 on isoelectric focusing and a M, of 29,000 f 1,000 on sodium dodecyl sulfate-polyacrylamide gel electro- phoresis. With respect to substrate specificity, optimum chain length for acyl moiety varied depending on the type of reaction catalyzed: C 18 in monomer lactone formation, C 11 or shorter in dimer lactone formation, and C8 in ester hydrolysis. The amino-terminal 19 amino acid residues of the purified lipase were deter- mined as Ser-Thr-Tyr-Thr-Gln-Thr-Lys-Tyr-Pro-Ile-Val-Leu-Ala-His-Gly-Met-Leu-Gly-Phe, and the gene encoding the lipase was identified by hybridization to a synthetic 20-nucleotide probe, cloned, and sequenced. Nucleotide sequence analysis predicted a 311-amino acid open reading frame, a putative ribo-some-binding site, and a 26-amino acid sequence at the amino terminus of the sequence that is not found in the mature protein. This 26-amino acid sequence has many of the characteristics common to known signal peptides. The lipase gene encoded a sequence of Val-Asn-Leu-Ile-Gly-His-Ser-His-Gly-Gly absorbance at 400 nm originating from p-nitrophenol (t4wnm = 14,200) was monitored. For determination of substrate specificity, p-nitrophenyl esters of various chain length were emulsified completely by sonication in the presence of 0.5% Triton X-100 in 0.05 M potassium phosphate (pH 6.5). After preincubation at 37 "C for 3 min, reactions were initiated by addition of an equal volume of enzyme solution (1.5 ml) in the same buffer, followed by a further incubation at 37 "C for 10 min. Reactions were stopped by addition of acetone (3 ml), the mixtures were clarified by centrifugation, and the absorbance at 400 nm was measured. One unit of esterase activity was defined as the amount of enzyme which liberates 1 pmol of p-nitrophenol or fatty acid/min. Lactonizing activity was assayed routinely by measuring the gel electrophoresis;

Purification, Characterization, and Molecular Cloning of Lactonizing Lipase from Pseudomonas Species* (Received for publication, May 14, 1991) Fumio Ihara, Yukari Kageyama, Mitsuyo Hirata, Takuya Nihira, and Yasuhiro YamadaS From the Department of Biotechnology, Faculty of Engineering, Osaka University, 2-1 Yamada-oka, Suita-shi, Osaka 565, Japan An extracellular lipase catalyzing the synthesis of macrocyclic lactones in anhydrous organic solvents was purified to homogeneity from Pseudomonas nov. sp. 109, and characterized. The lipase showed a PI of 6.3 on isoelectric focusing and a M, of 29,000 f 1,000 on sodium dodecyl sulfate-polyacrylamide gel electrophoresis. With respect to substrate specificity, optimum chain length for acyl moiety varied depending on the type of reaction catalyzed: C 18 in monomer lactone formation, C 11 or shorter in dimer lactone formation, and C8 in ester hydrolysis. The amino-terminal 19 amino acid residues of the purified lipase were determined as Ser-Thr-Tyr-Thr-Gln-Thr-Lys-Tyr-Pro-Ile-Val-Leu-Ala-His-Gly-Met-Leu-Gly-Phe, and the gene encoding the lipase was identified by hybridization to a synthetic 20-nucleotide probe, cloned, and sequenced. Nucleotide sequence analysis predicted a 311-amino acid open reading frame, a putative ribosome-binding site, and a 26-amino acid sequence at the amino terminus of the sequence that is not found in the mature protein. This 26-amino acid sequence has many of the characteristics common to known signal peptides. The lipase gene encoded a sequence of Val-Asn-Leu-Ile-Gly-His-Ser-His-Gly-Gly which is very well conserved among lipases, and showed 38-40% overall homology to the amino acid sequences of lipases from Pseudomonas fragie and Pseudomonas cepacia, but showed little homology to those of other lipases, suggesting that some structural features are required for catalyzing macrocyclic lactone synthesis in organic solvents and are restricted to lipases of the Pseudomonas origin.
Lactone, an intramolecular ester, has generally been found in natural compounds. Especially, 14-membered or larger membered lactones (macrocyclic lactones) constituted key structures of many useful compounds, such as macrolide antibiotics (1). Although 5-or 6-membered lactones can be formed automatically and exclusively from corresponding hydroxy acids (2), the formation of macrocyclic lactones, especially the monomeric lactones, is not straightforward the * This work was supported by the Nagase Science and Technology Foundation. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. formation of monomer lactone should compete with the formation of linear oligomers and also with that of multimeric lactones, such as dimeric lactones (diolides). Therefore, several chemical methods have been developed to facilitate an intramolecular esterification, mainly by activating acyl carbonyl groups weakly with special reagents, such as 2,2'dipyridyl disulfide (3), 2-chloro-1-methylpyridinium iodide (4), and carboxylic 2,4,6-trichlorobenzoic anhydride (5). However, the necessity to use very complicated, often very expensive, reagents or drastic conditions constitutes the main disadvantages of these chemical syntheses.
Recently, we have established that crude preparation of an extracellular lipase ("lipase P") produced by Pseudomonas nov. sp. 109 catalyzed in anhydrous organic solvents an intramolecular transesterification of methyl 16-hydroxyhexadecanoate leading to the efficient formation of cyclohexadecanolide ( Fig. 1, n = 15) (6). Furthermore, when racemic mixtures of (w-1)hydroxy acid methyl ester was used as substrate, the lactonizing lipase utilized only R-isomer, indicating the highly stereospecific nature of the lactonization. However, the lactonizing activity was not common to lipases. Our preliminary screening revealed that only some lipases of Pseudomonas origin and porcine pancreas lipase possessed the lactonizing activity (6), which suggest that special structural features are necessary to catalyze the synthesis of macrocyclic lactones.
To initiate the detailed three-dimensional analysis of the lactonizing lipase, large amounts of highly purified enzyme would be necessary. To that goal, we described herein the purification of the lactonizing lipase from Pseudomonas nov. sp. 109, and cloning of the gene encoding the lactonizing lipase (lipL). Sequence analysis revealed the high homology to amino acid sequences of lipases from Pseudomonas cepacia ( 7 ) and Pseudomonas fragi (8,9), which in turn predicted that those lipases may also have lactonizing activity.

Bacteria, Bacteriophage, and Plasmid
Pseudomonas nov. sp. 109, the lipase P producer, was obtained from Nagase Biochemicals Ltd. and has been deposited to the Fermentation Research Institute, Agency of Industrial, Science and Technology, Tsukuba (FERM-P No. 3025). Escherichia coli JM105 was used as a host for recombinant plasmids, bacteriophage M13mp18 and M13mp19 for nucleotide sequencing, and plasmid pUC19 as a vector for cloning (10).

Purification of Lactonizing Lipase
Crude powder of lipase P was obtained from Nagase Biochemicals Ltd. (Osaka). All the purification procedures were done at 4 "C unless otherwise specified.
fi) Preparation of Crude Extract-Crude lipase P (5 g) was suspended in 30 ml of 0.05 M potassium phosphate buffer (pH 6.5) containing 1 mM MgCL, 5 mM 2-mercaptoethanol, and 20% glycerol. After standing for 1 h a t 4 "C with mild stirring, the solution was Lactonizing Lipase from Pseudomonas Species dialyzed for 8 h against 3 liters of the same buffer with two changes, and centrifuged for 10 min at 8,000 X g.
(ii) Isoelectric Focusing-Ampholine (pH 3.5-10, Pharmacia LKB Biotechnology Inc.) was added to the supernatant from step i to a final concentration of 2%, and the mixture was fractionated on Rotofor (Bio-Rad) for 4 h with 12 W constant power. Fractions 8-10 (corresponding to pH 4-5.3) were mixed.
(iv) Gel Filtration HPLCI-HPLC was performed with a JASCO model Tri Rotor-V equipped with a UV detector (JASCO UVIDEC-100-V) and a fluorescence detector (JASCO FP-20). The concentrated sample from step iii was separated at ambient temperature in a portion of 80 p1 each on a TSK G2000SWxL column (0.78 X 30 cm, Tosoh, Osaka) equilibrated with 0.05 M potassium phosphate buffer (pH 6.5) containing 10% CH3CN at a flow rate of 0.3 ml/min. Protein was monitored by UV at 280 nm and fluorescence (excitation 280 nm, emission 340 nm). Activity appeared at the elution volume of 7.0 ml, and fractions ranging from 6.8 to 7.2 ml of the elution volume were pooled. Chemicals p-Nitrophenyl acetate was obtained from Tokyo Kasei Co. p -Nitrophenyl butyrate, p-nitrophenyl hexanoate, p-nitrophenyl octanoate, p-nitrophenyl dodecanoate, p-nitrophenyl tetradecanoate, pnitrophenyl hexadecanoate, p-nitrophenyl octadecanoate, and trilaurin were obtained from Nacalai Tasque Ltd. Triacetin, tributyrin, tricaprylin, tricaprin, tripalmitin, and tristearin were obtained from Wako Pure Chemicals Ltd. Tricaproin and trimiristin were obtained from Fluka Chemie AG and Sigma, respectively. p-Nitrophenyl decanoate was synthesized from decanoyl chloride and p-nitrophenol. Several chain lengths of w-hydroxy acid methyl esters were synthesized as described previously (6).

SDS-PAGE
SDS-PAGE was done with a ready-made 4-20% linear gradient gel (Daiichi Pure Chemicals Co. Ltd.) with a minigel apparatus (Daiichi Pure Chemicals Co. Ltd.) according to Laemmli ( l l ) , and stained with Coomassie Brilliant Blue G-250. Silver staining was done with Ag stain (Daiichi Pure Chemicals Co. Ltd) according to the manufacturer's protocol. Activity staining was done with a-naphthyl lactetate and fast blue RR salt according to Gabriel (12).
Assay of Esterase and Lactonizing Activity Esterase activity was routinely assayed by measuring the amount of p-nitrophenol formed from p-nitrophenyl acetate. Reaction mixtures (3.0 ml) contained 0.01 M potassium phosphate (pH 6.5), 2.5 mM p-nitrophenyl acetate, 5% CH&N, and enzyme. Reactions were initiated by addition of p-nitrophenyl acetate (50 mM, 0.15 ml) dissolved in dry CH3CN to a solution of other components which had been preincubated at 37 "C for 2 min. During the reaction, absorbance at 400 nm originating from p-nitrophenol (t4wnm = 14,200) was monitored. For determination of substrate specificity, p-nitrophenyl esters of various chain length were emulsified completely by sonication in the presence of 0.5% Triton X-100 in 0.05 M potassium phosphate (pH 6.5). After preincubation at 37 "C for 3 min, reactions were initiated by addition of an equal volume of enzyme solution (1.5 ml) in the same buffer, followed by a further incubation at 37 "C for 10 min. Reactions were stopped by addition of acetone (3 ml), the mixtures were clarified by centrifugation, and the absorbance at 400 nm was measured. One unit of esterase activity was defined as the amount of enzyme which liberates 1 pmol of p-nitrophenol or fatty acid/min.
Lactonizing activity was assayed routinely by measuring the The abbreviations used are: HPLC, high performance liquid chromatography; SDS-PAGE, sodium dodecyl sulfate-polyacrylamide gel electrophoresis; kb, kilobase pair. amount of cyclohexadecanolide formed from methyl 16-hydroxyhexadecanoate. After dialysis of enzyme sample against 0.05 M triethanolamine-HC1 (pH 6.5) containing 1 mM MgClz, the dialyzed sample was lyophilized to remove water in a 50-ml round-bottom evaporation flask. The lactonization reaction was initiated by suspending the lyophilized sample with 1 mM methyl 16-hydroxyhexadecanoate in dry hexane. During the reaction with vigorous stirring at 40 "C, aliquots were withdrawn intermittently, mixed with an equal volume of octacosane (0.2 g/liter in dry hexane) as protein concentration was determined by the procedure of Lowry et al. (13) using crystalline bovine serum albumin as standard.
Amino Acid Sequence Analysis A sample of purified protein (HPLC fraction) was passed through an TSK G2000SWx~ column (0.78 X 30 cm) equilibrated with 0.01 M ammonium acetate (pH 6.5) containing 0.1% Noigen (Daiichi Kogyo Seiyaku, Kyoto), protein fractions were pooled, and lyophilized to remove ammonium acetate. A sample was then analyzed using a pulsed-liquid protein sequenator 477A (Applied Biosystems) equipped with an on-line phenylthiohydantoin-derivative analyzer 120A (Applied Biosystems). This analysis was carried out with the assistance of Dr. Hiroshi Kataoka (University of Tokyo, Dept. of Agriculture).

DNA Manipulations
Chromosomal DNA of Pseudomonas nov. sp. 109 was obtained from cells grown in LB medium at 30 "C for 16 h by the modified method of Marmur as described by Coleman et al. (14). Plasmid DNAs were prepared by alkaline extraction (15). Restriction, ligation, end filling, and exonuclease treatment were done with enzymes from Takara Shuzo (Kyoto) or Toyobo Co. Ltd. (Osaka) under conditions recommended by the supplier. E. coli JM105 was transformed by the CaC12 method (16).

Southern Blot Analysis
High molecular weight genomic DNA was digested completely with PstI, SalI, or SmaI, and the digested DNAs were fractionated electrophoretically on a 1.0% agarose gel. The DNA was transferred to a nylon membrane (Hybond-N, Amersham Corp.), and hybridized with oligonucleotide probe end-labeled with [y-3ZP]ATP (>3000 Ci/mmol, ICN Biomedicals Inc.) and T4 polynucleotide kinase. Hybridization was done in 5 X SSPE containing 0.5% SDS, 5 X Denhardt's solution, and 0.1 mg/ml of salmon sperm DNA for 20 h at 45 "C and washed twice with 0.1 X SSC containing 0.1% SDS for 10 min each at 45 "C (15,17).
DNA Sequencing and Sequence Analysis Nucleotide sequences were determined by the dideoxynucleotide chain termination method (18) using [a-32P]dCTP (>3000 Ci/mmol, ICN Biomedicals Inc.) and M13 phage single stranded DNAs as template with Sequenase (United States Biochemical Corp.). The sequence was determined from both strands, and most nucleotides were sequenced several times in different overlapping clones.
Homology comparisons with EMBL, GenBank, NBRF, and SWIS-SPROT sequences (release April 1991) were performed on a personal computer with GENETYX software package (Software Development Co. Ltd., Tokyo).

RESULTS AND DISCUSSION
Purification of Lactonizing Lipase-We have previously shown that crude powder of lipase from Pseudomonas nov. sp. 109 (lipase P) catalyzed, in anhydrous organic solvents, intramolecular transesterification of methyl 16-hydroxyhexadecanoate leading to cyclohexadecanolide ( n = 15, Fig. 1).
However, when purified lipase constituting the major lipase activity in the crude powder was used instead, no lactonization was observed (data not shown). Therefore, Pseudomonas nov. sp. 109 should produce at least 2 lipases, and the lactonizing lipase would be a minor component. Actually, our previous screening with commerical lipases demonstrated that only lipases from Pseudomonas species and porcine pancreas lipase (Sigma type 11) possessed the lactonizing activity (6). Again,  for porcine pancreas lipase, further purified enzyme (Sigma type VI) did not show any lactonizing activity, indicating that contaminating minor enzyme present in the rather crude preparation was responsible for the lactonization (data not shown).
To characterize in more detail the lactonizing lipase, the lipase was purified from the crude powder of lipase P. Fraction was monitored both by esterase activity towardp-nitrophenyl acetate and formation of cyclohexadecanolide from methyl 16-hydroxyhexadecanoate. On isoelectric focusing of the crude lipase P, two esterase peaks were observed at PI values of 5.3 and 7.0, respectively, and the lactonizing activity coincided with the former peak. The lactonizing lipase did not adsorb to CM-Sephadex C-25 at pH 7, naturally from its PI value of 5.3, and activity recovery was less than 50% with DEAE-Sephadex, thus eliminating the use of ion exchange chromatography for further purification. Furthermore, on hydrophobic chromatography with C. , , C8, or phenyl resin, the lipase was strongly adsorbed and no activity was recovered

TABLE I1
Lactonizing activity of the lipase toward various w-hydroxy acid methyl esters Lactonization was performed with crude lipase P (200 mg) at a substrate concentration of 1 m M in dry hexane (20 ml) at 40 "C. The amount of corresponding monomer lactone and diolide were determined by gas chromatography as described (6). Reaction progress was followed intermittently, and catalytic rate was calculated from the linear part of each reaction.  (Table I). However, during purification the yield in lactonizing activity exceeded 100%. This is due to the presence of polysaccharide or Ampholine, which makes it difficult to remove water before the assay of lactonizing activity. In lactonization, the anhydrous condition is necessary to afford intramolecular transesterification. The presence of water facilitated hydrolysis, rather than transesterification, of substrate, thus resulting in underestimation of the lactonizing activity. Therefore, actual purification should be around 300-fold. Electrophoresis of 0.5 pg of protein showed that the HPLC fraction is greater than 95% pure (Fig.  2) and migrated with an apparent molecular weight of 29,000 k 1,000. Silver staining and activity staining also revealed a single band (data not shown). The mobility of samples run in gels in the absence of reducing agents did not change. However, apparent molecular weight estimated by gel filtration differed significantly (Mr = 290,000 on Spehadex G-200 with 10 mM potassium phosphate, pH 7, and M , = 133,000 on TSK gel G2000SWxL with 50 mM potassium phosphate, pH 7, containing 10% (v/v) CH3CN), which indicated that the lipase existed as several aggregates and the aggregation state changed by the hydrophobicity of the eluent.
Substrate Specificity of the hctonizing Lipase-Chain Esterase activity of lactonizing lipase toward several p-nitrophenyl esters and triglycerides Esterase activities were measured with purified lipase. Hydrolysis of p-nitrophenyl esters were measured at a fixed concentration (2.5 mM) of esters having an indicated chain length emulsified completely by sonication in the presence of 0.5% Triton X-100. Hydrolyses of triglycerides were measured at a fixed concentration (25 mM) of triglycerides emulsified as above. For triglycerides, after 60 min reaction a t 37 "C with vigorous stirring, reaction was stopped by the addition of a 10-fold volume of ether/ethanol (l:l), and the amount of free fatty acid was determined by titration with 0.01 N KOH. Triton X-100 is added to ascertain micell formation for all the substrates. In the absence of Triton X-100, the hydrolysis of triacetin correlated well with the degree of micellar formation measured by iodine method (19), indicating that the lactonizing lipase preferentially hydrolyzes the micellar form of substrates.   (lanes 3 and 6 ) , and SnaI (lanes 4 and 7) together with size markers (200 ng, XHindIII digests) were electrophoresed on 1.0% agarose gel and stained with ethidium bromide (lanes 1-4). After transfer to nylon membrane (Hybond N), the blot was hybridized with '"P-end labeled oligonucleotide probes (lanes 3-7) as described under "Materials and Methods," and exposed to x-ray film (Kodak X-Omat, -80 "C, 4 h). B, the strategy for sequencing is indicated length dependence of the lactonizing lipase toward lactone synthesis was investigated with several w-hydroxy acid methyl esters (Table 11). Lactone synthesis was clearly dependent on the chain length of the substrates, and the formation of monomer lactone was maximum with methyl 18-hydroxyoctadecanoate. With shorter substrates, monomer lactone decreased and the formation of diolide increased. To characterize this phenomenon, we measured the chain length dependence of its esterase activity toward p-nitrophenyl esters and triglycerides (Table 111). To our surprise, the lactonizing lipase showed preference to much shorter chain length, i.e. an ester of C8 acyl group was optimum for both kinds of substrates. Therefore, the lipase seems to have a binding site suitable to retain octanoyl or shorter acyl group. In monomer lactone formation, intramolecular hydroxyl groups should reach close proximity of the acyl carbonyl group. It is reasonable, therefore, that in monomer lactone formation twice the chain length was necessary than for simple hydrolysis. After formation of enzyme acyl intermediates, the C8 portion near the acyl moiety should be retained tightly by the enzyme and the rest of the molecule containing the w-hydroxy group can move freely and fold back toward acyl carbonyl and attack it to form the lactone. The shorter the substrates, then, the more difficult it was for the w-hydroxyl group to attack the acyl carbonyl. Instead, probability that the w-hydroxyl group on a different substrate molecule attacks the acyl carbonyl and increased, resulting in the increase of diolide formation. The ratio of monomer lactone synthesis to that of diolide was the same for crude enzyme and purified enzyme (formation ratio for monomer/diolide = 218 for crude and 215 for pure lipase, respectively), eliminating the possibility that the synthesis of with arrows representing the directions and length of the sequencing runs. A composite map of the overlapping clones with noncoding (white) and coding (striped) sequences is given at the top of the figure with the initiation (ATG) and stop (TAG) codons together with restriction endonuclease sites. E, ECoRI; S, SalI; P, PstI. The scale is shown in kilobase pairs. monomer lactone and diolide were catalyzed by different lipases.
NH2-terminal Amino Acid Sequence of the Lactonizing Lipase-To isolate the lipase gene by hybridization to oligonucleotide probe, the NH2-terminal region of the purified lipase was sequenced by automated Edman degradation. Only 1 amino acid was detected at each step of the degradation, confirming the homogeneity of the purified protein. By this procedure, the sequence of STYTQTKYPIVLAHGMLGF was obtained, and a mixture of the 20-nucleotide probe, (C/T)CC-3', was synthesized. Isolation and Sequencing of a Gene Encoding the Lactonizing Lipase-The probe was hybridized to restriction fragments generated from genomic DNA of Pseudomonas nov. sp. 109 (Fig. 3A), and 3.8-, 0.5-, and 6.0-kb bands were detected for digests by PstI, SalI, and SmaI, respectively. After enrichment of the 6.0-kb SmaI fragment by preparative agarose gel electrophoresis, the fragment was inserted into the SmaI site of pUC19, and the ligated DNA was transformed into E. coli JM105 on the L-plate containing 0.5% tributyrin (8) and ampicillin (50 gglml). Among 300 transformants, two colonies formed a clear zone due to lipase activity. Recombinant plasmids recovered from the two colonies were found identical by restriction analyses. To further localize the lipase gene (lipL) in the 6.0-kb insert, various plasmids containing deletions at either end of the 6.0-kb fragment were generated by exonuclease I11 and mung bean nuclease, and a plasmid containing the 2.2-kb insert (pUY45) was found sufficient for lipase activity. Up to 3.3 kb could be removed from the left end of the 6.0-kb fragment and up to 0.5 kb could be removed from the right end without influencing activity (Fig. 3B).
The 2.2-kb insert encoded a 311-amino acid polypeptide  Cha, charged amino acid ?, not detected. B, sequence alignment was done according to Lipman and Pearson (31). Identical amino acids were represented by an asterisk (*), and similar amino acids were by dots.(:), which were grouped as follows: Pro, Gly; Ser, Thr; Lys, Arg; Glu, Gln, Asn, Asp; Phe, Trp, His, Try; Ala, Ile, Val, Leu, Met, Cys (32). (Fig. 4). A 26-amino acid residue sequence (presumably the signal peptide) beginning with methionine precedes the NHzterminal end of mature lactonizing lipase as obtained by protein sequencing. All the signal peptides of secretory proteins so far reported are 13-to 36-amino acid residues in length, contain positively charged amino acid residue(s) in its NHz-terminus, and have Ala residue at its COOH terminus (20). The presumptive 26-amino acid signal peptide of the lactonizing lipase contains 3 Lys residues at position 2-4, Ala at the COOH terminus, and is extremely hydrophobic as a whole, thus matching well with the general features of signal peptides. The molecular weight of the polypeptide, excluding the signal peptide, is 30,149 which agreed well with that determined for purified lipase by SDS-PAGE (Mr = 29,000 f 1,000, Fig. 2). Furthermore, amino acid composition of the purified lipase was essentially identical to that deduced from the nucleotide sequence (data not shown), thus confirming that the cloned sequence was encoding the lactonizing lipase from Pseudomonas nov. sp. 109.

-V -N -L -I -G -H -S -H -G -G --Y -N -L -I -G -H -S -Q -G -A --V -N -L -I -G -H -S -0 -G -A --V -N -L -V -G -H -S -Q -G -Q --V -H -L -V -G -H -S -M -G -G -. V . H . F -I -G -H -S -M -G -G -. V . M . I -F -G -E -S -A -G -
The deduced amino acid sequence contains a sequence, -Val-Asn-Leu-Ile-Gly-His-Ser-His-Gly-Gly-, which match almost exclusively to the consensus sequence of lipase (Fig.  5A). Comparison with known sequences of lipases, such as from P. fragi (8,9), P. cepuciu (7), Staphylococcus uureus (21), Staphylococcus hyicus (22), Geotrichum candidum (23), Candida cylindrucea (24), porcine pancreas (25), rat lingua (26), canine pancreas (27), human liver (28), and human pancreas (29), revealed that the lactonizing lipase have 38-40% homologies on the amino acid level with lipases of pseudomonads, but no overall homologies with amino acid sequences of other lipases. When the alignment of functionary similar amino acid residues are considered, similarities of lactonizing lipase toward lipases of pseudomonads increased (50% for lipase of P. fragi and 59% for P. cepacia, Fig. 5B). These high similarities on overall sequence together with the similarity in molecular weight suggest that those lipases from pseudomonads may have lactonizing activity, which lack in other lipases. For eucaryotic lipases, lipase activity has been reported to be inhibited by SH-blocking agents and the positions of Cys residues are well conserved (27,30). However, for procaryotic lipases, Cys residues are lacking or low in its content, and seems not to participate in the catalytic activity. The lactonizing lipase contains only 2 Cys residues and its esterase activity was not affected by treatment of iodoacetic acid or 5,5'-dithiobis(2-nitrobenzoic acid) (no inactivation after incubation at 5 mM reagent at 37 "C for 1 h). Thus, similar to other procaryotic lipases, Cys residues in lactonizing lipase are not involved in its catalytic function. No homologies were found when the nucleotide and predicted amino acid sequences were compared with the EMBL and GenBank nucleotide and National Biomedical Research Foundation and SWISSPROT protein sequences, except for the lipases mentioned above and one hypothetical protein from Vibrio cholerue (SWISSPROT, accession number P15493). Although the latter only covered the NHZ-terminal part of the hypothetical protein, homology with lactonizing lipase is very high (58%), and, therefore, may represent a lipase-like enzyme of Vibrio origin.
The availability of genomic DNA clones for lactonizing lipase will provide tools for studying the mechanism involved in regulating the synthesis of the enzyme, and also make it possible to provide large quantities of highly purified enzyme, which is a prerequisite to initiate detailed investigations on the three-dimensional structural features which allow macrocyclic lactone formation.