The cold-active lipase of Pseudomonas fragi Heterologous expression, biochemical characterization and molecular modeling Claudia Alquati, Luca De Gioia, Gianluca Santarossa, Lilia Alberghina, Piercarlo Fantucci and Marina Lotti Dipartimento di Biotecnologie e Bioscienze, Universita ` degli Studi di Milano-Bicocca, Milano, Italy A recombinant lipase cloned from Pseudomonas fragi strain IFO 3458 (PFL) was found to retain significant activity at low temperature. In an attempt to elucidate the structural basis of this behaviour, a model of its three-dimensional structure was built by homology and compared with homologous mesophilic lipases, i.e. the Pseudomonas aeruginosa lipase (45% sequence identity) and Burkholderia cepacia lipase (38%). In this model, features common to all known lipases have been identified, such as the catalytic triad (S83, D238 and H260) and the oxyanion hole (L17, Q84). Structural modifications recurrent in cold-adaptation, i.e. a large amount of charged residues exposed at the protein surface, have been detected. Noteworthy is the lack of a disulphide bridge conserved in homologous Pseudomonas lipases that may contribute to increased conformational flexibility of the cold-active enzyme. Keywords: lipase; Pseudomonas; cold-active enzymes; modeling; selectivity. Enzymes from psychrotrophic and psychrophilic microor- ganisms have recently received increasing attention, due to their relevanceforboth basic and applied research. This effort has been stimulated by the recognition that cold-adapted enzymes might offer novel opportunities for biotechnological exploitation based on their high catalytic activity at low temperatures, low thermostability and unusual specificities. These properties are of interest in different fields such as detergents, textile and food industry, bioremediation and biocatalysis under low water conditions [1,2]. Furthermore, fundamental issues concerning the molecular basis of cold activity and the interplay between flexibility and catalytic efficiency areof importance inthe study ofstructure–function relationships in proteins. Such issues are often approached through comparison with the mesophilic or thermophilic counterparts, if available, and/or mutagenesis [3,4]. In this context, the recent cloning of a few lipases (acylglycerol ester hydrolases, EC 3.1.1.3) active at low temperature is relevant [5–7]. Because of their metabolic and industrial role, lipases have been thoroughly investigated by studies encompassing sequence, structure, regulation of expression, activity and specificity [8]. Among bacterial lipases, a focus has been on enzymes produced by members of the genus Pseudomonas, some of which have been recently reclassified as Burkholderia. A dozen of lipase-encoding genes have been cloned from different species, and the corresponding proteins have been classified into families I.1 and I.2 of bacterial lipases according to their molecular properties and to the requirement of helper proteins for correct folding and secretion [9]. Crystal structure determinations of Pseudomonas lipases have been reported including B. glumae (BGL) [10], P. aeruginosa (PAL) [11], and B. cepacia (BCL) [12–14]. In this paper, we describe the characterization of a cold- active lipase cloned from P. fragi, the main spoiling agent of refrigerated meat and raw milk [15]. This enzyme shares high sequence similarity with Pseudomonas lipases of known three-dimensional structure, therefore providing a new tool to study the molecular bases of cold-adaptation. EXPERIMENTAL PROCEDURES P. fragi strain IFO3458 (LMG2191 T ) was obtained from the BCCM TM /LMG bacteria collection (Universiteit Gent, Belgium). As the cloning host, E. coli JM101 (Promega Co, Madison, Wisconsin) was used. Heterologous expression was performed in the E. coli strain SG13009[pREP4] (Qiagen). Cloning and expression DNA manipulations were according to Sambrook et al. [16] and according to manufacturer’s instructions for the enzymes and materials employed. Chromosomal DNA was extracted as described previ- ously [17] with minor modifications from a P. fragi culture grown to the late exponential phase at 25 °Cin1%bacto tryptone, 0.5% bacto yeast extract and 0.5% NaCl. The lipase-encoding gene was amplified from chromoso- mal DNA by PCR with oligonucleotide primers designed based on the sequence of the homologous lipase from strain IFO 12049 (AC X14033). Forward primer: 5¢-CACCCTG CGAGATTGAACATG-3¢ (nucleotides )18 to+3); reverse primer: 5¢-AAGCTTGATTACAGGCTACAAG-3¢ (+938 Correspondence to M. Lotti, Dipartimento di Biotecnologie e Bioscienze, Universita ` degli Studi di Milano-Bicocca, Piazza della Scienza 2, 20126 Milano, Italy. Fax: + 39 02 64483565, Tel.: + 39 02 64483527, E-mail: marina.lotti@unimib.it Abbreviations:BCL,Burkholderia cepacia lipase; BGL, Burkholderia glumae lipase; MM, molecular mechanics; PAL, Pseudomonas aeruginosa lipase; PFL, Pseudomonas fragi lipase; PCR, polymerase chain reaction; SRC, structural conserved regions. Enzyme: lipase (EC 3.1.1.3). (Received 18 February 2002, revised 17 May 2002, accepted 23 May 2002) Eur. J. Biochem. 269, 3321–3328 (2002) Ó FEBS 2002 doi:10.1046/j.1432-1033.2002.03012.x to +924), where a restriction site for HindIII was inserted. Reaction was carried out in a total volume of 100 lLandwas catalyzed by 2.5 U of Pfu TurboTM polymerase (Stratagene, CA, USA). The amplification program was as follows: 3min94°C followed by 25 cycles of 30 s 94 °C, 45 s 55 °C, 1min 72°C, the final elongation step was 5 min 72 °C and 15 min 10 °C. The amplified fragment, purified from a 0.8% agarose gel, was automatically sequenced by M-Medical (Firenze, Italy). The lipase-encoding gene was inserted in the BamHI and HindIII cloning sites of pQE 30 (Qiagen), a plasmid designed for the regulated expression in E. coli of foreign genes. Recombinant proteins are produced as fusion with a His 6 tag at the N-terminus. The gene was amplified as before and modified by: (a) introducing a BamHI site and deleting the starting ATG codon and (b) introducing a HindIII cleavage site at the 3¢ end. Primers were as follows: forward primer: 5¢-GGATCCGACGATTCGGTAAAT-3¢; reverse primer: 5¢-AAGCTTGATTACAGGCTACAAG-3¢. E. coli SG13009 transformed with the expression plasmid was grown overnight at 27 °C in Luria–Bertani medium supplemented with 100 lgÆmL )1 ampicillin and 25 lgÆmL )1 kanamycin. Isopropyl thio-b- D -galactoside was added to a concentration of 0.4 m M and cultivation was continued for an additional 4 h. Cells were collected by centrifugation (30 min, 8000 g,4°C) and resuspended in 50 m M NaH 2 PO 4 pH 8.0, 300 m M NaCl and 10 m M imidazole. After the addition of 1 mgÆmL )1 lysozyme, the suspension was incubated on ice for 30 min and then sonicated six times for 10 s. The cell lysate was centrifuged for 20 min at 4 °C, 25 000 g.TheHis 6 -tagged lipase was purified at 4 °Con 50% Ni-nitrilotriacetic acid resin (Qiagen). Two mililiters of clear lysate were added to 1 mL resin, mixed by gentle shaking at 4 °C for 60 min and loaded on a column. After washing with 4 mL of 50 m M NaH 2 PO 4 pH 8.0, 300 m M NaCl and 20 m M imidazole, the recombinant lipase was eluted at pH 7.5 with 2 mL of 50 m M NaH 2 PO 4 ,300m M NaCl and 250 m M imidazole. Purification of the recombin- ant protein was monitored spectrophotometrically follow- ing the increase in absorbance at 410 nm due to hydrolysis of p-nitro-phenylpalmitate. The reaction mixture (1 mL) contained 0.8 m M p-nitro-phenylpalmitate in 50 m M Na 2 HPO 4 Æ2H 2 O, KH 2 PO 4 with 2% arabic gum. Protein characterization SDS PAGE [18] was performed using 12% acrylamide gels with GELCODE staining (Pierce). The protein concentra- tion was determined according to Bradford [19] with BSA as the standard. Lipase activity was determined in a pH-stat assay by titrating fatty acids released from triacylglycerols with 0.01 M sodium hydroxide using a 718 STAT TITRINO (Metrhom). Emulsions of 20 m M triacylglycerols with 2% arabic gum were used as the substrate. Tricaprylin was the substrate for activity determination, if not otherwise stated. The pH and temperature optima were investigated in the range of 4–9 and 25–40 °C, respectively. Substrate speci- ficity was determined at 29 °C, pH 8.0 on triacylglycerol substrates with chain lengths ranging from C 4 to C 18 .The effect of temperature on stability was determined by preincubating samples of purified rPFL at 10°,27° and 50 °C before determining residual activity. The effect of calcium ions on activity was investigated under standard assay conditions by measuring enzymatic activity in the presence of 5 m M EDTA at free calcium concentrations varying between 0 and 50 m M . Molecular modeling Multiple sequence alignments were carried out with the program CLUSTALX [20] using the following lipase sequences: P. aeruginosa (PAL, Swissprot code: P26876), B. glumae (BGL, Swissprot code: O05489), B. cepacia (BCL, Swissprot code: P22088), P. fragi (PFL, this work). The alignment featuring the highest score was obtained using the Blosum matrix [21] and standard CLUSTALX parameters. The atomic coordinates of PAL and BCL in their open conformation [11,13] were obtained from the Brookhaven Protein Data Bank. The PFL three-dimensional model was built according to the following procedure: (a) protein regions charac- terized by high similarity, as identified by sequence alignment and featuring very similar secondary structure, as derived from experimental data or as predicted by the PHD algorithm [22], were chosen as structurally con- served regions (SCR); (b) the atomic coordinates of the backbone atoms inside the SCR regions were transferred from the reference X-ray structure to the model; (c) fragments connecting the scaffold elements (usually loops) were modeled scanning the Brookhaven Protein Databank for protein structures of a predefined length that would fit properly into the model protein between two SCRs. The search was carried out by comparing the a-carbon distance matrix of the flanking SCR peptides with a precalculated matrix for all known proteins that have the same number of flanking residues and an intervening peptide segment of the given length. The following strategy has been adopted to optimize the structure of the model: (a) all atoms, except those corresponding to the fragments connecting SCRs, were kept fixed during the first Molecular Mechanics (MM) energy minimization. This allowed the nonSCR elements to re-arrange their conformation without affecting the global folding of the more conserved regions. (b) In a second MM optimization step, only the backbone atoms of the SCRs were kept fixed. (c) Only the a-carbons of the SCR regions were constrained to their initial position and (d) as the final step, the whole model was subjected to MM optimization without constraints. The optimized structure was subjected to the program PROCHECK [23], which allowed confirming its structural reliability. RESULTS AND DISCUSSION Sequence analysis Lipases have been cloned from a few P. fragi strains [24–26]. Out of them, the short lipase sequence (135 residues) from strain IFO3458 present in the database ([24]; AC: M14604) is probably truncated at its 3¢ half as it lacks functional sites. The coding sequence was therefore amplified from chro- mosomal DNA with oligonucleotide primers designed based on the sequences immediately upstream and down- stream the coding sequence of the lipase from the related strain IFO12049 (S02005) [25]. 3322 C. Alquati et al. (Eur. J. Biochem. 269) Ó FEBS 2002 Sequencing revealed an ORF of 879 bp encoding a polypeptide of 293 residues. Comparison with M14604 evidences a sequencing error at bp 354 where a missing nucleotide causes a reading frameshift followed by an early stop codon after few amino acids. The revised sequence has been assigned accession number AJ250176. Total GC content (59.3%) and the predicted codon usage, with 72.2% of the codons ending with G or C, are characteristic of Pseudomonas genes. Analysis of the deduced amino-acid sequence is consistent with a protein of M r 32.086 and an isoelectric point of 9.33. A single cysteine residue is present at position 39, thus ruling out the presence of a disulphide bridge, characteristic of most Pseudomonas lipases. A leader sequence for secretion could not be unambiguously identi- fied at the N-terminal end. The amino-acid sequence shares 97% identity with IFO 12049 lipase from which it only diverges in its C-terminal 22 residues (highlighted in Fig. 1). Scanning of protein sequence databases by FASTA [27] revealed a high degree of similarity with the P. fluorescens strain C9 (O68310, 48.1% identity over 297 amino acids), Proteus vulgaris (Q52614, 47.9% over 286 amino acids), Pseudomonas sp. (Q9X512, 47.2% over 290 amino acids), P. aeruginosa (Q9L6C7, 45.1% over 288 amino acids) and Vibrio cholerae (P15493, 48,2% over 288 amino acids) lipases. High identity is also shared with other Pseudomonas lipases of known three-dimensional structure, i.e. B. cepacia (37.7% over 318 amino acids), and B. glumae (37.9% over 322 amino acids). The similarity with P. fluorescens lipases is restricted to a 30-kDa enzyme recently isolated from milk [15] and does not extend to the 50-kDa P. fluorescens lipases described so far. Surprisingly, the remarkable similarity to the lipase from Proteus vulgaris K80 [28], that in addition shares with PFL the absence of cysteine residues possibly involved in intramolecular disulphide bond formation. On the other hand, the lipase sequence does not display obvious sequence similarity with either a lipase cloned from an alaskan psychrotrophic Pseudomonas [5] or with other cold-adapted lipases from different microorganisms [6,7]. Expression and purification Expression of rPFL carrying a His 6 tag at its N-terminus was obtained as described above. Culture growth and induction was performed at 27 °C to cope with both enzyme thermolability and the optimal growth temperature for the E. coli expression system. Under these conditions, rPFL was partly obtained in a soluble, active form suitable for further characterization. About 2 mg of pure recombinant rPFL per g wet weight cells were recovered by a one-step purification method involving metal-chelating chromatog- raphy (Fig. 2). In order to exclude any influence of the His 6 tag on the recombinant lipase activity, a control plasmid was constructed where the tag was followed by a recognition site for Tev protease. Enzyme obtained by protease digestion did not show any difference in activity (not shown). Several Pseudomonas lipases have been reported to require a chaperone (or helper) protein for efficient secretion and folding of the active lipase [9]. Therefore, coexpression of the lipase- and helper-encoding genes has been success- fully exploited as a tool to obtain high levels of recombinant active lipase in heterologous bacterial hosts [29]. However, coexpression of rPFL with the foldase of P. aeruginosa did Fig. 1. Alignment of the PFL sequence with Pseudomonas lipases of known three-dimensional structure. The sequence is identical to that of the lipase from strain IFO12049 except for the C-terminal 22 amino acids which are shown in bold. ˆ ¼ amino acid forming the catalytic triad, ° ¼ amino acid forming the LID, § ¼ amino acid involved in calcium binding. PAL ¼ P. aeruginosa lipase, PFL ¼ P. fragi lipase, BCL ¼ B. cepacia lipase and BGL ¼ B. glumae lipase. Ó FEBS 2002 Cold-active Pseudomonas fragi lipase (Eur. J. Biochem. 269) 3323 not produce significant improvements in the fraction of soluble lipase. The apparent ineffectiveness of foldase together with the lack of a signal peptide at the N-terminal suggests the hypothesis that PFL might be secreted by a signal peptide-independent pathway and calls for further investigation on this subject. Enzyme activity and specificity In the standard tricaprylin hydrolysis assay, rPFL displayed highest activity at 29 °C and pH 8.0, in good agreement with results reported for the wild type enzyme [30]. Interestingly, the enzyme lost most of its activity at 50 °C but retained about 60% of its specific activity at 10 °C (Table 1). Under the same conditions, the activity of the homologous lipase from Burkholderia cepacia was reduced to 10% (not shown). The stability of rPFL was further investigated. Incubation of the enzyme at different temper- atures showed that at 10 °C rPFL retains for several hours most of its activity, whereas its half-life at 27 °C is about 5 h. Activity dramatically drops after a few minutes at 50 °C (Fig. 3). This is an important characteristic for PFL identification as a cold-active enzyme and well fits with the structural features evidenced by the three-dimensional model, as reported below. Specificity towards triglycerides was tested in a pH-stat assay using six triacylglycerol substrates of different chain length (C 4 to C 18 ). As shown in Fig. 4, rPFL activity decreases going from short- to long-chain substrates and triolein is a poor substrate. Therefore, PFL differs in chain length selectivity from both PAL, with broad substrate specificity on triglycerides [31] and BCL which shows a high preference for the hydrolysis of triglycerides with a chain length ‡ 8 [32]. No activity was detected in a phospholipase assay using phosphatidylcholine as the substrate. Finally, experiments have been carried out to elucidate the influence of Ca 2+ ions on the catalytic activity. rPFL activity measured in the presence of 5 m M EDTA was by 55% lower than the control. Variation of the calcium concentration resulted in a saturation curve with a plateau between 10 m M and 20 m M calcium (Fig. 5). Fig. 3. Effect of temperature on the enzyme stability as measured at 10 °C(d), 27 °C(r)and50°C(m). Fig. 4. Activity of rPFL on triacylglycerols with different chain length. C4, C8, C12, C16, C18 and C18* are tributyrin, tricaprylin, trilaurin, tripalmitin, tristearin and triolein, respectively. Fig. 5. Enzymatic activity of rPFL as a function of calcium concentra- tion. Fig. 2. Electrophoretic analysis showing the purification of recombinant PFL. (1) Molecular mass standards; (2) 2 lg rPFL after purification by metal-chelating chromatograpy; (3) total soluble extract of E. coli expressing the recombinant protein. Table 1. Rate of hydrolysis of tricaprylin by rPFL at different temper- atures. Temperature of hydrolysis (°C) 10 29 50 Relative hydrolysis (%) 59 100 15 3324 C. Alquati et al. (Eur. J. Biochem. 269) Ó FEBS 2002 Model building In order to derive a three-dimensional model of the PFL structure, its sequence was aligned with those of Pseudo- monas lipases of known three-dimensional structure, namely B. cepacia (BCL), B. glumae (BGL) and P. aeruginosa (PAL). As can be observed in Fig. 1, sequence similarity extends along all the protein sequence with the exception of the 177–213 region (PFL numbering) where the BCL and BGL sequences are characterized by peptide insertions. Use of secondary structure data allowed to better define SCR in regions where sequence similarity alone provided ambigu- ous results. Using this approach, 11 SCRs were defined that constitute the structural scaffold upon which the PFL three- dimensional model was built. These SCRs span the peptide segments 5–23, 31–49, 54–68, 76–120, 129–148, 155–181, 191–201, 214–218, 237–242, 245–267 and 277–293 of PFL. It is interesting to note that most nonSCRs are peptide fragments forming loop regions in the experimentally derived three-dimensional structures, supporting our choice of SCRs. On the other hand, regions forming important structural and/or functional portions in the selected lipases, such as the catalytic triad, the oxyanion hole and the calcium binding site are all located in SCRs. In this work, PAL was chosen as structural reference as this enzyme shows the highest degree of similarity to PFL. Furthermore, the computed structure was validated against the BCL structure. We observed that our PFL model might overlap to a high degree with both structures, as expected from the overall structural similarity relating PAL and BCL [11]. On these bases, the structural features of the PFL were analysed in deeper detail (Figs 6–8). Fig. 6. Overall three-dimensional structure of PFL, as obtained by homology modeling. For clarity, only arginine side chains are explicitly shown. Fig. 7. Schematic representation of the Ca-coordination environment in PFL. The conserved residues D217 and D262 are coordinated to the metal ion by one carboxylic oxygen atom belonging to the side chain. The Ca ion is coordinated also by the backbone carbonyl groups of H266 and R269. Bond distances and angles involving the metal atom and the coordinating amino acids are very similar to the corresponding values observed in PAL (data not shown). Fig. 8. Schematic representation of the two aromatic residues of PFL (W184 and F244) that structurally correspond to the disulphide bridge in PAL, shown in grey. Ó FEBS 2002 Cold-active Pseudomonas fragi lipase (Eur. J. Biochem. 269) 3325 Amino acids forming the catalytic center in PAL and BCL are conserved in PFL and correspond to S83, D238 and H260. However, some subtle structural differences between the active sites are worth noting. PFL is charac- terized by a hydrogen bond between D259 and H260, which is not present either in PAL nor BCL; the proximal carboxylic acid suggested to function as an alternative proton acceptor in other Pseudomonas lipases (E289, BCL numbering) [33] is not conserved in PFL nor does the analysis of its model reveal other glutamate or aspartate residues in that region. Another feature characterizing lipases is the so-called oxyanion hole, an arrangement of amino acids which plays an important role in the stabilization of the tetrahedral intermediate transiently formed in the hydrolysis reaction through hydrogen bonding to main chain donors [34]. Even in this case, two residues that might contribute to the formation of the stabilizing configuration (L17 and Q84, PFL numbering) are conserved and located in SCRs. In most lipases, the catalytic site is occluded by a surface amphiphilic structure, named the lid, that makes the active site inaccessible to the substrate, unless it is displaced by the interaction with aggregated substrates. The region forming the lid is located in a SCR and appears to be well conserved in PFL (128–148). A further target for investigation was the presence of a binding site for Ca 2+ ions, which are reported to enhance the activity of several bacterial lipases. Accordingly, a calcium binding pocket is present both in PAL and BCL, where its function seems to be rather related to structure stabilization [11,13]. The crucial features of the calcium binding site are well conserved in PFL, where D217 and D262 are favourably located to interact with the metal by their side chain carboxylic groups (Fig. 7). Moreover, two other amino acids (H266 and R269, PFL numbering), even if different from the corresponding residues of BCL and PAL, have a proper structural orientation of the backbone carbonyl groups to bind the metal ion. However, some subtle differences in the Ca 2+ coordination environment are observed in the three considered lipases. In PAL (and BCL), one of the water molecules coordinated to the Ca 2+ ion is also involved in hydrogen bonds with the side-chains of S211 and D212 (S244 and T245 in BCL). In PFL, these residues are substituted by L219 and H220. Due to the hydrophobic nature of leucine, the binding energy of the water molecule is expected to be smaller in PFL than in PAL and BCL. Moreover, in BCL a water molecule interacts with a large loop (A213–N239) and is coordinated to the Ca 2+ ion. In PAL the corresponding loop (S202– F207) is shorter and the water molecule in the coordination sphere of the metal is substituted by the OH group of T205. In PFL, the loop (A210–L215) has geometry features similar to PAL but T205 is substituted by the hydrophobic residue L213, suggesting a weaker coordination to the metal ion. The enzyme–substrate interaction has been studied in detail for PAL and BCL crystallized in complex with the same tryglyceride-like inhibitor [11,14] For both enzymes, the substrate binding pocket has been shown to consist of a cleft mostly lined by hydrophobic amino-acid side chains and residues relevant for substrate recognition have been identified. In this context, a comparison between PFL and BCL shows important differences in their preference towards short and long chain acyl substrates. A structural comparison suggests features that are responsible for these differences. Interesting substitutions are located at the entrance of the substrate binding funnel (V123 with L119 and F249 with R224) and within the binding pocket where T18, F119, A247 and T251 in BCL are substituted in PFL by F, L, N and V residues, respectively. Site-directed mutagenesis is required to elucidate the relevance of the observed aminoacid substitutions. PFL structure was then analysed with the aim of explaining PFL activity at low temperature observed in this study. The path followed during evolution to ÔdesignÕ cold adapted enzymes is still not completely understood. However, the number of disulphide bridges, the location of proline and arginine residues, and hydrophobic core interactions are generally recognized as key factors in determining the flexibility of the enzyme molecule and, as a consequence, its catalytic activity at different temperatures [2]. One of the most striking differences between PAL, BCL and PFL concerns the number of arginine residues, 11, 9 and 24, respectively. In PFL, 20 out of 24 arginine residues are almost evenly distributed on the surface of the protein (Fig. 6) and only two of them are involved in intramolecular salt bridges. Therefore, PFL is characterized by the abundance of charged residues on the surface which enhances flexibility and ability of interaction with the solvent [3]. Both factors are of importance for cold activity. On the other hand, the presence of disulphide bridges may increase the rigidity of the protein backbone and therefore its thermostability [35]. Interestingly, the only disulphide bridge present in BCL and PAL is missing in the PFL, where only one cysteine residue is present. Recently, it has been shown that CysfiSer mutants of PAL are more sensitive to heat denaturation, confirming a stabilizing role for the PAL disulphide bond [36]. Moreover, the analysis of the three-dimensional structure of PFL allow to observe that two aromatic residues (W184 and F244), that struc- turally correspond to the disulphide bridge in BCL, are involved in a stacking interaction (Fig. 8), with a moderate stabilizing effect on this portion of the protein. Finally, distribution of proline residues was considered, as it has been suggested that conformationally rigid proline residues in loops and turns might increase the rigidity of a protein and decrease its catalytic efficiency at low temperature [35]. PAL, BCL and PFL all contain 13 proline residues, eight of which are located in conserved regions. Other prolines are mainly located in loop regions of PAL and BCL, whereas only one is in a loop in PFL. Again, this suggests an increased conformational flexibility for the latter enzyme. CONCLUSIONS The P. fragi lipase characterized in this study is a new example of cold-active lipase and provides a model for studying the molecular basis of cold adaptation. In fact, comparison of proteins from organisms belonging to divergent evolutionary lineages might confuse the analysis of structural differences of relevance which are masked by substitutions arose from evolutionary divergence [37]. Other examples of psychrophilic lipases are reported in the literature, one of which has been cloned from Pseudomon- ads. However, such lipases are related in sequence to each other but not to any mesophilic lipase. In contrast, PFL can 3326 C. Alquati et al. (Eur. J. Biochem. 269) Ó FEBS 2002 be classified in a group of bacterial lipases well character- ized from the biochemical and structural point of view. Comparison of the three-dimensional model with the structures of mesophilic homologous lipases accounts for PFL activity at low temperature, in the frame of the largely accepted assumption that relates cold adaptation to changes in the protein conformational flexibility. This study high- lights relevant features in the PFL structure, such as a reduced number of disulphide bridges and of prolines in loop structures. Arginine residues are distributed differently than in mesophilic enzymes, with only a few residues involved in stabilizing intramolecular salt bridges and a large proportion of them exposed at the protein surface and therefore able to interact with the solvent enhancing flexibility. ACKNOWLEDGEMENTS This work was supported by a grant of the Progetto Finalizzato Biotecnologie of the Italian National Research Council to L. A. We also acknowledge contribution from the Vigoni Program. The authors wish to thank K E. Jaeger for helpful discussion and for providing the P. aeruginosa foldase-encoding gene. REFERENCES 1. Gerday, C., Aittaleb, M., Bentahir, M., Chessa, J.P., Claverie, P., Collins, T., D’Amico, S., Dumont, J., Garsoux, G., Georlette, D., Hoyoux, A., Lonhienne, T., Meuwis, M.A. & Feller, G. (2000) Cold-adapted enzymes: from fundamentals to biotechnology. Trends Biotechnol. 18, 103–107. 2. Russell, N.J. (2000) Toward a molecular understanding of cold activity of enzymes from psychrophiles. Extremophiles 4, 83–90. 3. Narinx, E., Baise, E. & Gerday, C. (1997) Subtilisin from psy- chrophilic antarctic bacteria: characterization and site-directed mutagenesis of residues possibly involved in the adaptation to cold. Protein Eng. 10, 1271–1279. 4. Wintrode, P.L., Miyazaki, K. & Arnold, F.H. (2000) Cold adap- tation of a mesophilic subtilisin-like protease by laboratory evo- lution. J. Biol. Chem. 275, 31635–31640. 5. Choo, D W., Kurihara, T., Suzuki, T., Soda, K. & Esaki, N. (1998) A cold-adapted lipase on an alaskan psychrotroph, Pseu- domonas sp. strain B11-1: gene cloning and enzyme purification and characterization. Appl. Environ. Microbiol. 64, 486–491. 6. Feller, G., Thiry, M., Arpigny, J.L. & Gerday, C. (1991) Cloning and expression in Escherichia coli of three lipase-encoding genes from the psychrotrophic antarctic strain Moraxella TA144. Gene 102, 111–115. 7. Arpigny, J.L., Feller, G. & Gerday, C. (1993) Cloning, sequence and structural features of a lipase from the antarctic facultative psychrophilic Psychrobacter immobilis B10. Biochim. Biophys. Acta 1171, 331–333. 8. Rubin, B. & Dennis, E.A. (1997) Lipases. In Methods in Enzymology,Vol.284.AcademicPress,USA. 9. Jaeger, K E., Dijkstra, B.W. & Reetz, M.T. (1999) Bacterial biocatalysts: molecular biology, three-dimensional structures, and biotechnological applications of lipases. Annu. Rev. Microbiol. 5, 315–351. 10. Noble, M.E.M., Cleasby, A., Johnson, L.N., Egmond, M.R. & Frenken, L.G.J. (1993) The crystal structure of triacylglycerol lipase from Pseudomonas glumae reveals a partially redundant catalytic aspartate. FEBS Lett. 331, 123–128. 11. Nardini, M., Lang, D.A., Liebeton, K., Jaeger, K.E. & Dijkstra, B.M. (2000) Crystal structure of Pseudomonas aeruginosa lipase in the open conformation – the prototype for family I.1 of bacterial lipases. J. Biol. Chem. 275, 31219–31225. 12. Kim, K.K., Song, H.K., Shin, D.H., Hwang, K.Y. & Suh, S.W. (1997) The crystal structure of a triacylglycerol lipase from Pseu- domonas cepacia reveals a highly open conformation in the absence of a bound inhibitor. Structure 5, 173–185. 13. Schrag, J.D., Li, Y., Cygler, M., Lang, D., Burgdorf, T., Hecht, H J., Schmid, R., Schomburg, D., Rydel, T.J., Oliver, J.D. et al. (1997) The open conformation of a Pseudomonas lipase. Structure 5, 187–202. 14. Lang, D.A. & Dijkstra, B.W. (1998) Structural investigations of regio- and enantioselectivity of lipases. Chem. Phys. Lipids 93, 115–122. 15. Dieckelmann, M., Johnson, L.A. & Beacham, I.R. (1998) The diversity of lipases from psychrotrophic strains of Pseudomonas:a novel lipase from a highly lipolytic strain of Pseudomonas fluor- escens. J. Appl. Microbiol. 85, 527–536. 16. Sambrook,K.J.,Fritsch,E.F.&Maniatis,T.(1989)Molecular Cloning: a Laboratory Manual, 2nd edn. Cold Spring. Harbor Laboratory Press, Cold Spring Harbor, New York. 17. Coleman, K., Dougan, G. & Arbuthnott, J.P. (1983) Cloning, and expression in Escherichia coli K-12, of the chromosomal hemoly- sin (phospholipase C) determinant of Pseudomonas aeruginosa. J. Bacteriol. 153, 909–915. 18. Laemmli, U.K. (1970) Cleavage of structural proteins during the assembly of the head of the bacteriophage T4. Nature 227, 680– 685. 19. Bradford, M. (1976) A rapid and sensitive method for the quan- titation of microgram quantities of protein utilizing the principle of protein-dye binding. Anal. Biochem. 72, 248–254. 20. Thompson, J.D., Gibson, T.J., Plewniak, F., Jeanmougin, F. & Higgins, D.G. (1997) The Clustal–X windows interface – flexible strategies for multiple sequence alignment aided by quality ana- lysis tools. Nucleic Acids Res. 25, 4876–4882. 21. Henikoff, S. & Henikoff, J.G. (1992) Amino acid substitution matrices from protein blocks. Proc. Natl Acad. Sci. USA 89, 10915–10919. 22. Rost, B., Sander, C. & Schneider, R. (1994) Redefining the goals of protein secondary structure prediction. J. Mol. Biol. 235, 13–26. 23. Laskowski, R.A., MacArthur, M.W., Moss, D.S. & Thornton, J.M. (1993) PROCHECK: a program to check the stereochemical quality of protein structures. J. Appl. Cryst. 26, 283–291. 24. Kugimiya, W., Otani, Y., Hashimoto, Y. & Takagi, Y. (1986) Molecular cloning and nucleotide sequence of the lipase gene from Pseudomonas fragi. Bioch. Biophys. Res. Comm. 141, 185–190. 25. Aoyama, S., Yoshida, N. & Inouye, S. (1988) Cloning, sequencing and expression of the lipase gene from Pseudomonas fragi IFO- 12049 in E. coli. FEBS Lett. 242, 36–40. 26. Okada, Y., Kamio, Y., Isaki, K. & Shigyo, T. (1993) Patent JP 1993076372-A1. Sapporo Breweries Ltd. ‘New lipase B gene’. 27. Pearson, W.R.E. & Lipman, D.J. (1988) Improved tools for biological sequence comparison. Proc. Natl Acad. Sci. USA 85, 2444–2448. 28. Kim, H.K., Lee, J.K., Kim, H. & Oh, T.K. (1996) Characteriza- tion of an alkaline lipase from Proteus vulgaris K80 and the DNA sequence of the encoding gene. FEMS Microbiol. Lett. 135, 117–121. 29. Quyen, D.T., Schmidt-Dannert, C. & Schmid, R.D. (1999) High- level formation of active Pseudomonas cepacia lipase after heterologous expression of the encoding gene and its modified chaperone in Escherichia coli and rapid in vitro refolding. Appl. Environ. Microbiol. 65, 787–794. 30. Mencher, J.R. & Alford, J.A. (1967) Purification and character- ization of the lipase of Pseudomonas fragi. J. General Microbiol. 48, 317–328. 31. Jaeger, K E., Schneidinger, B., Liebeton, K., Haas, D., Reetz, M.T., Philippou, S., Gerritse, G., Ransac, S. & Dijkstra, B.W. (1996) Lipase of Pseudomonas aeruginosa: molecular biology and biotechnological applications. In Molecular Biology of Ó FEBS 2002 Cold-active Pseudomonas fragi lipase (Eur. J. Biochem. 269) 3327 Pseudomonads (Nakazawa, T., ed.), pp. 319–330. ASM Press, Washington. 32. Bornscheuer, U., Reif, O W., Lausch, R., Freitag, R., Scheper, T., Kolisis, F.N. & Menge, U. (1994) Lipase of Pseudomonas cepacia for biotechnological purposes: purification, crystallization and characterization. Biochim. Biophys. Acta 1201, 55–60. 33. Lang, D., Hofmann, B., Haalck, L., Hecht, H.J., Spener, F., Schmid, R.D. & Schomburg, D. (1996) Crystal structure of a bacterial lipase from Chromobacterium viscosum ATCC 6918 refined at 1.6 A ˚ resolution. J. Mol. Biol. 259, 704–717. 34. Ollis, D.L., Cheah, E., Cygler, M., Dijkstra, B., Frolow, F., Franken, S.M., Harel, M., Remington, S.J., Silman, I., Schrag, J., Sussman, J.L., Verschueren, K.H.G. & Goldman, A. (1992) The alpha/beta hydrolase fold. Prot. Eng. 5, 197–211. 35. Aghajari, N., Feller, G., Gerday, C. & Haser, R. (1998) Structures of the psychrophilic Alteromonas haloplanctis amylase give insights into cold adaptation al a molecular level. Structure 6, 1503–1516. 36. Liebeton, K., Zacharias, A. & Jaeger, K E. (2001) Disulfide bond in Pseudomonas aeruginosa lipase stabilizes the structure but is not required for interaction with its foldase. J. Bacteriol. 183, 597–603. 37. Sheridan, P.P., Panasik, N., Coombs, J.M. & Brencheley, J.E. (2000) Approaches for decifering the structural basis of low tem- perature enzyme activity. Biochim. Biophys. Acta 1543, 417–433. 3328 C. Alquati et al. (Eur. J. Biochem. 269) Ó FEBS 2002 . The cold-active lipase of Pseudomonas fragi Heterologous expression, biochemical characterization and molecular modeling Claudia Alquati,. bridges may increase the rigidity of the protein backbone and therefore its thermostability [35]. Interestingly, the only disulphide bridge present in BCL and