Metagenomic approach for the isolation of a novel low-temperature-active lipase from uncultured bacteria of marine sediment

Abstract

A novel lipase was isolated from a metagenomic library of Baltic Sea sediment bacteria. Prokaryotic DNA was extracted and cloned into a copy control fosmid vector (pCC1FOS) generating a library of >7000 clones with inserts of 24–39 kb. Screening for clones expressing lipolytic activity based on the hydrolysis of tributyrin and p-nitrophenyl esters, identified 1% of the fosmids as positive. An insert of 29 kb was fragmented and subcloned. Subclones with lipolytic activity were sequenced and an open reading frame of 978 bp encoding a 35.4-kDa putative lipase/esterase h1Lip1 (DQ118648) with 54% amino acid similarity to a Pseudomomas putida esterase (BAD07370) was identified. Conserved regions, including the putative active site, GDSAG, a catalytic triad (Ser148, Glu242 and His272) and a HGG motif, were identified. The h1Lip1 lipase was over expressed, (pGEX-6P-3 vector), purified and shown to hydrolyse p-nitrophenyl esters of fatty acids with chain lengths up to C₁₄. Hydrolysis of the triglyceride derivative 1,2-di-O-lauryl-rac-glycero-3-glutaric acid 6′-methylresorufin ester (DGGR) confirmed that h1Lip1 was a lipase. The apparent optimal temperature for h1Lip1, by hydrolysis of p-nitrophenyl butyrate, was 35°C. Thermal stability analysis showed that h1Lip1 was unstable at 25°C and inactivated at 40°C with t_1/2 <5 min.

Introduction

The rapidly expanding approach of metagenomics is attractive for understanding and accessing microbial genomes and their functions. Based on the direct cloning of environmental DNA (the metagenome) (Rondon et al., 2000) for the construction of large clone libraries, metagenomics allows access to novel sequences, genes, complete pathways and their products by multiple screening possibilities. The major screening approaches include expression of a function (Healy et al., 1995; Henne et al., 1999, 2000; Brady & Clardy, 2000; Entcheva et al., 2001; Gillespie et al., 2002; Knietsch et al., 2003), identification of interesting genes based on sequence homology through PCR or random sequencing (Rondon et al., 2000) and large-scale shotgun sequencing (Tyson et al., 2004; Venter et al., 2004; Tringe et al., 2005), hybridization (Stein et al., 1996), or the detection of substrate-induced gene expression (Uchiyama et al., 2005). The power of metagenomics is the access, without prior sequence information, to the so far uncultured majority, which is estimated to be more than 99% of the prokaryotic organisms (Amann et al., 1995; Rappe & Giovannoni, 2003). Through metagenomic cloning, novel enzymes have been isolated from soil (Henne et al., 1999, 2000; Rondon et al., 2000; Lee et al., 2004b; Yun et al., 2004; Kim et al., 2006), water (Cottrell et al., 1999; Ranjan et al., 2005) and thermal environments (Rhee et al., 2005).

An increasing number of diversity studies of marine sediments (for example Ravenschlag et al., 1999 and Lee et al., 2004a), including some of Baltic Sea soft-bottom sediments (F. Hårdeman and S. Sjöling, unpublished; Edlund et al., 2006), are reporting a high bacterial diversity. A highly diverse bacterial community potentially reflects a high diversity of metabolic enzyme genes. From a biotechnological perspective, these enzymes, and the knowledge derived from their structures and functions, could be valuable in industrial applications, making marine sediments suitable for enzyme mining. Given that the marine sediments largely represent low-temperature environments, the presence of low-temperature-adapted bacteria with low-temperature-active enzymes would be expected (Russell et al., 1998; Feller, 2003; D'Amico et al., 2006; Siddiqui & Cavicchioli, 2006). Low-temperature-active enzymes typically have maximal catalytic activity at temperatures below 40°C, are thermolabile, and are inactivated at moderate temperatures (>40–50°C) (Russell et al., 1998; Siddiqui & Cavicchioli, 2006). Other features could be increased flexibility around the active site, a decrease in disulfide bridges, an increased core hydrophobicity, and decreased ionic electrostatic interactions (Aghajari et al., 1996; Russell et al., 1998; Alquati et al., 2002; D'Amico et al., 2006; Siddiqui & Cavicchioli, 2006). These features may contribute to a catalytic activity, which at low temperatures can be up to 10 times higher than that of enzymes of mesophilic organisms (Feller, 2003).

One group of interesting enzymes, for a variety of biotechnological purposes, is that of lipases (EC 3.1.1.3) and carboxylesterases (EC 3.1.1.1), which belong to carboxylhydrolases. Whereas carboxylesterases preferentially hydrolyse water-soluble esters and triacylglycerols with fatty acids shorter than C₆, lipases prefer water-insoluble substrates, typically triacylglycerols with medium- to long-chain fatty acids (≥10 carbon atoms) (Jaeger et al., 1999; Pandey et al., 1999; Jaeger & Eggert, 2002). Industrially, lipases are used in a variety of applications, for example paper processing, food manufacturing, as food additives (Jaeger et al., 1999), and in the production of biofuel catalysing the conversion of vegetable oil to methylalcohol ester (Jaeger & Eggert, 2002). Features such as enantio-/stereoselectivity, discriminating between enantiomers in racemic pairs (Reetz, 2001), broad substrate specificity, and activity in organic solvents (Gupta et al., 2004) make lipases useful in synthetic organic chemistry and in other industrial processes, such as the production of pharmaceuticals (Reetz, 2001).

Thus, the construction of metagenomic libraries of prokaryotes of low-temperature environments for the identification of novel lipases should be valuable. In this study, we constructed a metagenomic fosmid library of high-molecular-weight prokaryotic DNA from low-temperature Baltic Sea sediment for the isolation, by functional expression screening, of fosmid clones with lipolytic activity. Several clones with lipolytic activity were detected, and one was selected for further characterization. A novel lipase was identified, overexpressed, purified, and characterized for thermal stability, temperature and substrate specificity.

Materials and methods

Sampling marine sediment

Soft-bottom sediment samples were collected on 31 March 2004 close to Askö Marine Research Centre (58°48′73 N, 17°34′28 E) in the Stockholm southern archipelago of the western Baltic Sea, Sweden. Samples of 0–5 cm of sediment depth were taken in triplicate at 39 m of water depth from a research vessel using Kajak cores (50 cm²). Samples were taken to the laboratory at the station, aseptically fractionated, and immediately frozen at −20°C. Sediment physicochemical parameters were determined according to previous studies (Edlund et al., 2006) and were as follows: salinity 6.4 psu, temperature 0.8°C and oxygen 12.4 mg L⁻¹.

Construction of a metagenomic fosmid library

Extraction of environmental DNA

Prokaryotic cells were isolated from the sediment in order to avoid a library dominated by eukaryotic DNA. The cells were extracted by a modified method based on low-speed centrifugation in order to extract the sediment prokaryotes prior to cell lysis (Bakken & Lindahl, 1995). One hundred grams of sediment and 100 mL of TE-pyrophospate buffer (10 mM Tris-HCl, 1 mM EDTA, 50 mM sodium-pyrophosphate, pH 8.0) were mixed under gentle agitation for 30 min. One hundred millilitres CTAB buffer (50 mM Tris-HCl, 20 mM EDTA, 0.15 M NaCl, 27 mM Hexadecylmethylamoniumbromide, pH 8.0) was added, and the sediment–buffer mixture was blended in a sterile Waring-type blender for 30 s at maximum speed. The sample was divided in two and centrifuged at low speed at 500 g for 10 min at 10°C. The supernatant, containing the cells, was recovered, and the pellet was resuspended in 200 mL of CTAB buffer. The blending and low-speed centrifugation step was repeated twice. The three supernatants obtained were pooled and centrifuged at 10k g for 30 min in order to enable collection of the prokaryotic cells. The cell pellets were resuspended in TE (10 mM Tris-HCl, 1 mM EDTA, pH 8.0) to a final volume of about 10 mL. The cells were lysed, and DNA was extracted according to the following method: aliquots of 750 μL of resuspended cells were incubated with 30 μL of 10% SDS and 20 μL of proteinase K (20 mg mL⁻¹) for 1 h at 37°C. One hundred microlitres of 5 M NaCl was added and mixed thoroughly. Eighty microlitres of CTAB/NaCl solution (10% CTAB, 0.7 M NaCl) was added and mixed again before incubating at 65°C for 10 min. DNA was then extracted by phenol/chloroform/isoamylalcohol (25 : 24 : 1) and ethanol precipitation.

High-molecular-weight DNA purification and fractionation

The extracted high-molecular-weight (HMW) DNA was size-separated on a 1% low-melting-point agarose gel (GE Healthcare) using pulse-field gel electrophoresis (PFGE) (CHEF-DR II, BIO-RAD Laboratories) at 6 V cm⁻¹, 5–15 s switch for 13 h. In addition to the sample DNA, each gel contained a PFGE DNA Molecular marker (Low Range PFG Marker, New England Biolabs), fosmid control DNA (Epicentre), and a small amount of the extracted DNA. All except the sample DNA were stained in ethidium bromide DNA of a size suitable for fosmid cloning (30–50 kb) was cut out and extracted with β-agarase (New England Biolabs) following the manufacturer's protocol.

Cloning of HMW metagenomic DNA

Four hundred-nanograms size-separated DNA of 30–50 kb was end-repaired to blunt end and cloned into the copy control pCC1FOS vector using the CopyControl Fosmid Library Production Kit, according to the manufacturer's protocol (Epicentre). The advantage of this vector was the possibility of cloning into a stable low-copy-number vector, which during screening was induced to high copy number by arabinose. The vector also accepts rather large DNA fragments, and thus the library could be screened not only for small genes, such as lipases, presented here, but also for whole operons. The fosmid clones were manually arrayed with toothpicks into 96-well microtitre plates and grown in Luria–Bertani (LB) media (Sambrook & Russell, 2001) supplemented with 7% glycerol and 12.5 μg mL⁻¹ chloramphenicol at 37°C for 24 h before replication and long-term storage at −80°C. One uninoculated well was left on each library plate to ensure the absence of contaminations. Fifteen randomly selected fosmid clones were used for determining the average insert size of the library. The fosmid was isolated using alkaline minipreps according to the protocol for Bacterial Artificial Chromosomes minipreparations on the California Institute of Technology homepage (http://www.tree.caltech.edu). Positive controls contained Fosmid control DNA (Epicentre), and negative controls contained Escherichia coli host not carrying any fosmid. Isolated fosmids were digested with Not1 and analysed by PFGE under the conditions: 1% agarose in 0.5 × TBE buffer, 6 V cm⁻¹, 1–12 s switch for 11.5 h.

Library screening

End sequencing

Ten randomly selected fosmid clone inserts were end-sequenced using T7 promoter primer (Epicentre) in order to confirm that inserts were of prokaryotic origin. The obtained sequences were searched for homologies in public databases using the tools BlastN, BlastX (Altschul et al., 1997) and orf-finder (NCBI) (http://www.ncbi.nlm.nih.gov).

Expression screening for lipolytic activity

The metagenomic library was screened for fosmid clones expressing lipolytic activity using LB-agar plates containing 1% tributyrin, 1% gumarabic and chloramphenicol 12.5 μg mL⁻¹ (Kok et al., 1993) and by chromogenic p-nitrophenyl ester microtitreplate assay (Wilkinson, 2000). Arabinose (0.001%) was included in the medium in order to induce the fosmids to high copy number using the oriV/trfA system (Wild et al., 2002). Escherichia coli EPI300™-T1^R (Epicenter) carrying pCC1FOS with cloned fosmid control DNA provided with the kit was used as a negative control. The library was replicated onto tributyrin agar plates and incubated at 22°C for the host E. coli to grow until colonies had formed. The plates were further incubated at 4°C for up to 7 days, allowing the detection of low-temperature lipolytic activity by means of the formation of clear zones around colonies expressing any lipolytic activity. The activity of the clones was reconfirmed by isolation of the fosmid clone, retransformation and repeated screening to eliminate any false positives. The retransformation was performed into EPI300™-T1^R cells by electroporation using the settings for transformation of large circular DNA molecules suggested by Sheng et al. (1995).

Subcloning of fosmid insert with lipolytic activity

Out of all fosmids expressing lipolytic activity, one fosmid clone, h1L, was selected for further analysis. It was subcloned in order to identify the gene encoding the lipolytic activity out of the 29-kb fosmid insert. The h1L fosmid clone was cut into smaller fragments using HaeIII, and the fragments were size-separated by 1% agarose TAE gel electrophoresis (Sambrook & Russell, 2001). The fragments (1–8 kb) were cut out, extracted and purified from the gel (Geneclean Turbo kit, Qbiogene) and then ligated with T4 DNA ligase (New England Biolabs) into pUC18 (Invitrogen Life Technologies), which had been isolated (Plasmid Maxi Kit, Qiagen), cut with Sma1 (Fermentas), dephosphorylated by Shrimp Alkaline Phosphatase (New England Biolabs) and purified (Geneclean Turbo Kit, Qbiogen) prior to ligation. The ligated vector was transformed into E. coli TOP-10 cells (Invitrogen Life Technologies) according to the manufacturer's protocol. The transformants were selected by blue/white screening, and white colonies were arrayed in 96-well plates as described above but with ampicillin 100 μg mL⁻¹. The subclone library was replicated onto tributyrin screening plates containing ampicillin 100 μg mL⁻¹, incubated and analysed for clones expressing lipolytic activity, as described above.

Identification of the h1Lip1 lipase ORF and promoter region

Eight lipolytically active subclones, with insert sizes ranging from 1800 to 4000 bp, were sequenced from both ends using m13-forward and -reverse primers (Uppsala Genome Center). The subclone sequences were assembled using the CAP contig function (Huang, 1992) of the BioEdit program (http://www.mbio.ncsu.edu/BioEdit/bioedit.html). The complete assembled sequence was analysed by sequence and translated sequence comparison using the tools BlastN, BlastX and orf-finder (NCBI) (http://www.ncbi.nlm.nih.gov). The identified translated gene was aligned with similar proteins to establish phylogenetic relationship. Comparisons with known low-temperature-active proteins were made. The assembled sequence was analysed for promoter sequences using the program Promoter prediction (http://www.fruitfly.org/seq_tools/promoter.html) (Reese, 2001).

Overexpression and purification of h1Lip1

In order to introduce BamH1 restriction sites (underlined in the primer sequences below) allowing the cloning of h1Lip1 into the expression vector pGEX-6P-3 (GE Healthcare), the gene was amplified with the primers LE1f (5′) GAGAACCGGATCCATGCCCAGTCAACGTGCC and LE2r (5′) GGATCCTCGGAACTAGGCGGTGAGGC. A standard PCR reaction (Sambrook & Russell, 2001) with 54°C annealing temperature was performed. In order to ensure that the correct restriction sites were flanking the amplified h1Lip1 gene it was first cloned into the pCR 2.1-TOPO vector (Invitrogen Life Technologies) according to the manufacturer's recommendations. The BamHI (Fermentas) restriction-digested h1Lip1 fragment was gel-purified (Geneclean Turbo Kit, Bio101), ligated into the pGEX-6P-3 vector (GE Healthcare) according to the manufacturer's recommendations (using T4-DNA ligase, New England Biolabs), and transformed into E. coli TOP-10 (Invitrogen Life Technologies), because E. coli BL21(DE3) cells, used for overexpression, are less suitable for plasmid isolation required for end-sequencing. The cloning of the h1Lip1 gene was confirmed by lipolytic activity assay and end-sequencing (MWG Biotech, Germany).

The h1Lip1-pGEX-6P-3 construct was then transformed into E. coli BL21(DE3) cells (Invitrogen Life Technologies), grown in 200 mL of LB medium at 37°C until an OD₆₀₀ of 1.0, and induced by isopropyl-β-D-thiogalactopyranoside (IPTG), 0.5 mM. The induced culture was further incubated at 25°C for 4 h to increase the ratio of soluble h1Lip1. The culture was harvested by centrifugation at 5k g for 10 min, and the pellet was frozen overnight. The pellet was resuspended in 5 mL of phosphate-buffered saline (PBS) pH 7.3 and lysed by sonication for 2 min at medium output (Branson Sonifier 250). The supernatant was centrifuged at 13k g for 15 min to pellet cell debris. The supernatant was ultracentrifuged at 139k g for 30 min. The h1Lip1-glutathione S-transferase fusion protein was purified from the supernatant using a 1-mL GSTrap affinity column and the PreScission™ Protease (GE Healthcare), which cleaves off and binds to the GST-tag that stays bound to the column, according to the manufacturer's protocol. The pure, soluble h1Lip1 was eluted in 50 mM Tris-HCl, 150 mM NaCl, 1 mM EDTA, 1 mM dithreitol, pH 7.5. The protein content was determined using the Bradford protein assay kit (BIO-RAD Laboratories), with bovine serum albumin as a standard.

Characterization of h1Lip1

The overexpressed and purified h1Lip1 was used for substrate specificity, temperature stability and kinetic analysis. All experiments were performed in triplicate. The substrate specificity for nitrophenyl esters was analysed using p-nitrophenyl esters (Sigma) 0.5 mM, of butyrate (C₄), caprylate (C₈), decanoate (C₁₀), laurate (C₁₂), myristate (C₁₄) and palmitate (C₁₆) (dissolved in acetonitrile), respectively, in 0.1 M NaCl, 0.1 M NaH₂PO₄, 15% acetonitrile and 0.038 mM Triton X-100, pH 7.25 (Choo et al., 1998). The initial hydrolysis velocity of h1Lip1 (1.5 μg) at various concentrations of substrate was determined spectrophotometrically by monitoring the formation of p-nitrophenol at 400 nm at 22°C. To determine the presence of lipase activity, the triglyceride derivative 1,2-di-O-lauryl-rac-glycero-3-glutaric acid 6′-methylresorufin ester (DGGR) (Sigma Aldrich) (Panteghini et al., 2001) was used as a chromogenic substrate (Jaeger et al., 1999). According to a modified method by Zandonella et al. (1996) with 25 μM DGGR, 2.5 μg of h1Lip1 were incubated for 20 min at 25°C, and the formation of methylresorufine was analysed spectrophotometrically at 580 nm (Panteghini et al., 2001). Candida rugosa lipase (Sigma Aldrich) was used as a positive control.

The apparent optimal temperature of h1Lip1 and mesophilic porcine liver esterase (Sigma Aldrich) activity was analysed by incubating the lipase for 2 min at various temperatures (in 0.1 M NaCl, 0.1 M NaH₂PO₄, pH 7.25 supplemented with 0.5 mM p-nitrophenyl butyrate dissolved in acetonitrile) and monitoring the formation of p-nitrophenol. This substrate was selected because it remained water-soluble under the conditions employed (Laurell et al., 2000). The mesophilic porcine liver esterase (Sigma Aldrich) was chosen because no bacterial low-temperature lipase was accessible, and the substrate used was suitable for all esterases but not for all lipases. Relative activity at different temperatures was expressed as a percentage of the maximal activity of each respective enzyme. Thermostability was analysed by determining the residual specific activity of h1Lip1 towards p-nitrophenyl butyrate after incubation at 0, 10, 25 and 40°C for 0, 5, 15, 30, 90 and 130 min, respectively. Kinetic parameters of h1Lip1 at 10°C and 20°C were determined by analysing the initial velocity of hydrolysis at various p-nitrophenyl butyrate concentrations (0.1–1 mM) under standard assay conditions. Initial reaction velocities determined at different substrate concentrations were fitted to the Lineweaver–Burk transformation of the Michaelis–Menten equation.

The nucleotide sequence reported in this study has been deposited in GenBank, under the accession number DQ118648.

Results

Construction of a metagenomic library and screening

One hundred micrograms of prokaryotic DNA was extracted per gram of wet-weight sediment (Fig. 1a). 1.5 μg of size-selected, pulse-field gel-purified HMW DNA suitable for fosmid cloning was obtained (Fig. 1b). Four hundred nanograms of 30–50 kb HMW DNA was ligated into the copy control pCC1FOS vector and transfected to the E. coli strain EPI300-T1^R, with an efficiency of 1.7 × 10⁴ clones μg⁻¹ DNA, producing a library of more than 7000 fosmids with insert sizes ranging from 24 to 39 kb, with an average size of 31 kb (Fig. 1c), covering approximately 217 Mbp of the total metagenomic DNA. Given an average prokaryotic genome of approximately 5 Mbp, the library theoretically reached the size of over 40 prokaryotic genomes. The prokaryotic origin of the library was confirmed by end-sequencing of randomly selected fosmids and comparison with known ORFs in GenBank (NCBI).

Figure 1.  Pulse field gel electrophoresis (a) environmental high molecular weight of extracted DNA. Ethidium bromide stained 1% agarose gel with Lane 1: PFGE Ladder, Lanes 2 and 3: Extracted prokaryotic sediment DNA, ∼50 ng and ∼2 μg respectively, Lane 4: fosmid control DNA, 50 ng; (b) Environmental DNA that has been cut out from gel “a” Lane 1: PFGE Ladder, Lane 2: fosmid control DNA 50 ng, Lanes 3 and 4: prokaryotic sediment DNA ∼30 and ∼50 ng respectively, extracted and size separated for cloning; (c) Fosmid DNA insert size analysed by restriction with NotI. Lane 1; 12 and 22 low-range PFGE ladder, Lanes 2–11 and 13–18; 15 randomly analysed fosmid clones with sediment DNA inserts, Lane 19; positive control, Fosmid cloned control DNA, Lane 21: negative control, host E. coli, minipreparation.

Expression screening of the fosmid library for hydrolytic activity based on the hydrolysis of tributyrin resulted in the detection of c. 1% of positive fosmid clones. Out of all positive clones identified, one, h1L, expressing particularly strong hydrolytic activity compared with the other positive clones, was selected for further investigation and its activity was reconfirmed after retransformation. In order to identify the hydrolytic gene within a fragment of 29 kb, the insert was subject to further subcloning.

Subcloning and sequence analysis

The DNA insert (29 kb) of fosmid h1L was fragmented and cloned into pUC18, producing a subclone library of >10³ clones with an average insert size of 2 kb. Three hundred subclones were screened for lipolytic activity. The eight subclones that expressed extracellular lipolytic activity were sequenced from both ends and the sequences were assembled into a contig of 3274 bp (Fig. 2). An ORF of 978 bp encoding a putative esterase/lipase (named h1Lip1) of 325 amino acids was identified. Through amino acid sequence alignment, the h1Lip1 showed 54% similarity (over 302 amino acids) to a putative esterase of Pseudomonas putida (BAD07370). Interestingly, the highest amino acid similarity (57%, but only over 246 amino acids of the complete h1Lip1 protein) was to a putative protein of the Sargasso Sea (EAD76328) in the environmental database, NCBI (Venter et al., 2004) (Fig. 3). One additional putative ORF within the assembled contig encoding an alkylated DNA repair protein and one partial ORF encoding another esterase/lipase were identified, the latter lacking a start and stop codon (Fig. 2). A putative promoter region of h1Lip1 by the presence of a TATA-box, GCTATAAT, 36 bp upstream, and a putative Shine Dalgarno sequence, AAAGAG, 11 bp upstream of the start codon were identified.

Figure 2.  Sequencing of subclones (fosmid h1L) expressing lipolytic activity resulted in the assembly of a 3274 bp contig. Two major ORF with conserved domains were identified: h1Lip1, encoding the novel lipase; and AlkB, with homology to an alkylated DNA repair protein. The third ORF located closely to h1Lip1 lacks start and stop codons but shares sequence homology to known lipolytic proteins.

Figure 3.  Amino acid alignment of lipase h1Lip1 with the most closely related proteins in GenBank (NCBI): Sargasso: unknown predicted protein isolated from Sargasso Sea samples (EAD76328); Pseudomonas: putative esterase/lipase from Pseudomonas putida (BAD07370); Moraxella: lipase Lip2 from Moraxella sp (CAA37862); B11-1 LipP: lipase from Pseudomonas sp. B11-1 (AAC38151). Conserved regions are indicated by (^*), identical in all species; (:), similar in all species; and (.), similarities exist. The conserved HGG region is indicated by (□), and the active site and the other members of the catalytic triad are indicated by (#).

Various lipases and esterases contain the conserved active site motif of the pentapeptide GXSXG with a serine acting as the catalytic nucleophile, a conserved aspartate or glutamate and a histidine, together constituting a catalytic triad (Jaeger et al., 1994), organized in the α/β hydrolase fold (Ollis et al., 1992). The amino acid sequence alignment to the Pseudomonas putida putative esterase/lipase (BAD07370), the putative protein from the Sargasso Sea (EAD76328), and to two low-temperature-adapted esterase/lipases of Pseudomonas sp. B11-1 LipP (AAC38151) (Choo et al., 1998) and Moraxella sp. Lip2 (CAA37862) (Feller et al., 1991) identified the conserved motifs, including the putative active site GDSAG (Fig. 3). Thus, h1Lip1 probably uses a catalytic triad consisting of the serine (Ser148) in the GDSAG active site, the glutamate (Glu242) and the highly conserved histidine (His272) for catalysis. Moreover, h1Lip1 contains a HGG conserved motif (starting from His78), which corresponds to the HGGG motif that is in close proximity to the active site contributing to the formation of the oxyanion hole that is likely to participate directly in the catalytic process (Jaeger et al., 1994; Laurell et al., 2000).

The h1Lip1 contained a higher content of glycine residues than did the mesophilic Pseudomonas esterase (BAD07370) and the psychrophilic Moraxella Lip2 (CAA37862), a factor that potentially could contribute to low-temperature adaptation by increasing the protein flexibility.

Overexpression and purification of h1Lip1

The overexpressed and pure h1Lip1 (2–4 mg L⁻¹ of culture) was visualized as a single protein band of c. 35 kDa by sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) (Fig. 4), consistent with the molecular weight of 35.4 kDa deduced from the h1Lip1 nucleotide sequence.

Figure 4.  SDS-polyacrylamide gel with overexpressed lipase h1Lip1; lane 1: molecular weight marker, lane 2: uninduced culture, total protein extract, lane 3: induced culture, total protein extract (GST-h1Lip1 65.4 kDa), lane 4: purified h1Lip1 (35.4 kDa).

Substrate specificity of the h1Lip1 lipase

Results showed that h1Lip1 was able to hydrolyse p-nitrophenyl esters with acyl chains up to 14 carbons (p-nitrophenyl myristate), with the highest activity towards four to eight carbon fatty acids (p-nitrophenyl butyrate and caprylate) (Fig. 5) at pH 7–9, with an apparent optimum at pH 8 (data not shown). The extinction coefficient of p-nitrophenol at pH 7.25 was determined to be 11 438 M⁻¹ cm⁻¹, consistent with what has been reported under similar assay conditions elsewhere (12 000 M⁻¹ cm⁻¹; Laurell et al., 2000). h1Lip1 was also able to hydrolyse the triglyceride derivative 1,2-di-O-lauryl-rac-glycero-3-glutaric acid 6′-methylresorufin ester (DGGR) (data not shown), releasing the chromogenic product methylresorufin when hydrolysed. These results indicate that h1Lip1 is a lipase and not an esterase (Jaeger et al., 1999).

Figure 5.  Substrate specificity of overexpressed and purified lipase h1Lip1. Specific activity of hydrolysis of different p-nitrophenyl esters; butyrate (C₄), caprylate (C₈), decanoate (C₁₀), laurate (C₁₂), myristate (C₁₄) and palmitate (C₁₆).

Temperature apparent optimum and thermostability

h1Lip1 showed the highest activity at 35°C, with activity rapidly decreasing at temperatures above 40°C. Kinetic analysis revealed a lower K_m (0.22 mM) at 10°C than at 20°C (0.32 mM). At low temperatures, 5°C and 10°C, 33% and 44%, respectively, of the maximal activity at 35°C remained (Fig. 6). The apparent optimal temperature of h1Lip1 was lower, and the relative activity at low temperatures was higher than the values for the mesophilic porcine liver esterase (Fig. 6). Thermostability analysis showed that h1Lip1 was unstable at 25°C and rapidly inactivated at 40°C with t_1/2 of less than 5 min (Fig. 7).

Figure 6.  Apparent temperature optimum of lipase h1Lip1. Relative activity of p-nitrophenyl butyrate hydrolysis at different temperatures by purified h1Lip1 lipase (◆) compared to mesophilic porcine liver esterase (□) expressed as percentage of maximal activity of respective enzyme.

Figure 7.  Thermostability of lipase h1Lip1. Residual specific activity of the lipase h1Lip1 after incubation at different temperatures, 0°C (◆), 10°C (□), 25°C (▴) and 40°C (○), towards p-nitrophenyl butyrate at 25°C.

Discussion

In this study, we have identified and isolated a novel low-temperature active lipase, h1Lip1, from low-temperature sediment of the Baltic Sea. This is, to our knowledge, the first reported low-temperature active lipase isolated from a large insert metagenomic library constructed from marine sediments. Previous studies, using metagenomic libraries for the detection of lipolytic activities, have identified esterases from thermal environments (Rhee et al., 2005), saline lake (Rees et al., 2003), or forest soil (Lee et al., 2004b), or lipases from field soil (Henne et al., 2000; Rondon et al., 2000) or drinking water (Voget et al., 2003). During preparation of the manuscript, an esterase with an apparent low-temperature optimum but with limited thermosensitivity (Kim et al., 2006) was reported.

It is only recently that the first extremophilic enzymes, esterases from soda lakes and thermal environments, have been isolated using a metagenomic approach (Rees et al., 2003; Rhee et al., 2005). These studies, together with the study presented here, support the potential of metagenomic cloning as a unique tool for the discovery of enzymes from uncultured extremophilic bacteria.

The effect of temperature on the hydrolytic activity by h1Lip1 correlates well with what has been reported for other low-temperature-active lipases and esterases: lipase LipP (AF034088) of the psychrotrophic Pseudomonas sp. strain B11-1 showed apparent optimal activity at 45°C and was unstable >45°C (Choo et al., 1998); lipase KB-Lip (AB063391) of the psychrotrophic Pseudomonas sp. strain KB700A showed optimal activity at 35°C, which decreased by 70% after 5 min at 60°C (Rashid et al., 2001); lipase rPFL (AJ250176) of Pseudomonas fragi showed optimal activity at 29°C, which dramatically decreased at 50°C (Alquati et al., 2002); esterase PsyEst (AF260707) of psychrophilic Psychrobacter sp. Ant300 showed optimal activity at 35°C and t_1/2 of 16 min at 40°C (Kulakova et al., 2004). Given that h1Lip1 had an apparent optimum activity at 35°C with 44% of the activity remaining at 10°C, a lower K_m at 10°C than at 20°C and was thermosensitive, we concluded that h1Lip1 was a low-temperature-active lipase with low thermostability, probably derived from a low-temperature-active organism (Russell et al., 1998; Siddiqui & Cavicchioli, 2006).

Sequence comparison showed that h1Lip1 lipase is related to the group IV family of esterases/lipases and the hormone-sensitive lipase (HSL) family according to the classification suggested by Arpigny & Jaeger (1999). The conserved active site, GDSAG, located close to the N-terminus, and the HGG motif immediately upstream are characteristic of the group IV lipases (Jaeger et al., 1999). In this group, both low-temperature-adapted (from isolates) and thermophilic members can be found (Jaeger et al., 1999). Crystal structure analysis of mutants of esterase 2 of the thermophile Alicyclobacillus acidocaldarius belonging to the HSL family showed that the N-terminal domain (which is part of a ‘lid’ structure including the HGG motif) was important for thermostability and substrate specificity (Mandrich et al., 2005). Secondary structure prediction (Structure Prediction Meta Server, http://bioinfo.pl/meta/target) indicated that h1Lip1 lipase could have an α/β-hydrolase fold and did not exclude the possibility of the amino terminal as part of a so-called ‘lid’ formation (Mandrich et al., 2005).

The general limitations of metagenomic cloning and of applying heterologous expression screening to cloned unknown DNA (from many different prokaryotic organisms) in E. coli are several and have been discussed elsewhere (Daniel, 2005). Specifically for this study, the development and application of low-temperature-adapted vector and host systems may contribute to an increased throughput and the detection of more enzymes. The inclusion of a PCR step with degenerated primers towards lipases could also increase the screening throughput, although the possibility of detecting sequence-unique genes would be limited. A harsh, indirect DNA extraction technique could increase the coverage of the total sediment population but would reduce the DNA quality severely, both in terms of HMW DNA suitable for cloning and unwanted eukaryotic DNA (Gabor et al., 2003). The coverage of the fosmid library was, however, sufficient for the detection of lipolytic genes (hit rate 1%) and in concordance with other environmental libraries reported in the literature.

The potential applications for a lipase with a relatively high activity at low temperatures could be detergent additives, or in the processing of volatile substances, thereby making it possible to reduce temperature and thus bring down energy costs (Margesin & Schinner, 1994). A low-temperature-active lipase would be attractive for the bioremediation of low-temperature soils or waters. The low stability at moderate temperatures could, for example, be beneficial for terminating an enzyme reaction by heat-inactivation in biotransformation processes.

Future studies should include further characterization of the other lipolytic genes detected in this study. Site-directed mutagenesis and three-dimensional structure analysis will clarify whether the h1Lip1 contains a characteristic lid (Brzozowski et al., 1991) and whether any low-temperature adaptations in the protein structure can be detected.

Acknowledgements

We would like to acknowledge Anna-Ida Johnsson and Johannes Fritsch for laboratory assistance, Stockholm Marine Research Centre and staff for assistance in the field, Dr Maria Bergström, Dr Karin Hjort (Metacontrol QLK3CT200202068) and Professor Janet Jansson for discussions on DNA extraction, Dr Magnus Johansson for suggestions on protein expression, and Professor D. Cowan at the University of Western Cape, South Africa for fruitful discussions. The study was financially supported by the Baltic Sea Foundation, the Carl Tryggers Foundation, the Wenner-Gren foundations and Stockholm Marine Research Centre.