DOC

Bacterial lipolytic enzymes classification and properties

By Don Freeman,2014-03-09 08:55
5 views 0
Bacterial lipolytic enzymes classification and properties

Biochem. J. (1999) 343 (177183) (Printed in Great Britain)

Bacterial lipolytic enzymes: classification and properties

    1Jean Louis ARPIGNY and Karl-Erich JAEGER

Lehrstuhl für Biologie der Mikroorganismen, Ruhr-Universität Bochum, Universitätsstrasse 150,

    D-44780 Bochum, Germany

    Knowledge of bacterial lipolytic enzymes is increasing at a rapid and exciting rate. To obtain an overview of this industrially very important class of enzymes and their characteristics, we have collected and classified the information available from protein and nucleotide databases. Here we propose an updated and extensive classification of bacterial esterases and lipases based mainly on a comparison of their amino acid sequences and some fundamental biological properties. These new insights result in the identification of eight different families with the largest being further divided into six subfamilies. Moreover, the classification enables us to predict (1) important structural features such as residues forming the catalytic site or the presence of disulphide bonds, (2) types of secretion mechanism and requirement for lipase-specific foldases, and (3) the potential relationship to other enzyme families. This work will therefore contribute to a faster identification and to an easier characterization of novel bacterial lipolytic enzymes.

    Abbreviations used: 3D, three-dimensional; HSL, hormone-sensitive lipase; PAF-AH, platelet-activating-factor acetylhydrolase.

    1 To whom correspondence should be addressed (e-mail karl-erich.jaeger@ruhr-uni-bochum.de). Key words: alignment, carboxylesterase, classification, lipase, structure.

INTRODUCTION

    Bacteria produce different classes of lipolytic enzyme, including carboxylesterases (EC 3.1.1.1), which hydrolyse small ester-containing molecules at least partly soluble in water, true lipases (EC 3.1.1.3), which display maximal activity towards water-insoluble long-chain triglycerides, and various types of phospholipase. This paper deals with the former two classes of enzyme. For a description of phospholipases we refer the reader to recent review articles [1,2].

    Our knowledge of the structure of lipases and esterases has increased considerably in recent years through the elucidation of many gene sequences and the resolution of numerous crystal structures [3,4]. Efforts accomplished the classification of this large set of data and the identification families and subfamilies of lipolytic enzymes [5,6] (see also the ESTHER database at

    http://meleze.eusam.inra.fr/cholinesterase). Many attempts have been made to identify sequence motifs conserved in lipolytic enzymes originating from a broad variety of organisms, including higher and lower vertebrates, invertebrates, fungi and bacteria, and to relate them to three-dimensional (3D) structural elements involved in substrate recognition and catalysis, and therefore being essential for the enzyme's function.

    The structural superfamily of a/b-hydrolases defined by Ollis et al. [7] comprises a wide variety of

    enzymes whose activities rely mainly on a catalytic triad usually formed by Ser, His and Asp

    residues. This triad is functionally (but not structurally) identical with that of trypsin and subtilisin. In the amino acid sequences of a/b-hydrolases the three residues follow the order Ser-Asp-His. The serine residue usually appears in the conserved pentapeptide Gly-Xaa-Ser-Xaa-Gly. a/b-Hydrolases notably include lipolytic enzymes, among which true lipases demand special attention because their peculiar catalytic properties make them very attractive for industrial applications. Their marked preference for water-insoluble substrates and their adsorption on the oil/water interface before hydrolysis involve substantial conformational changes of the enzyme's architecture during catalysis; these have been particularly well documented for lipases of eukaryotic origin [8].

    Lipolytic enzymes are currently attracting enormous attention because of their biotechnological potential [911]. Most of the lipases used in industry are microbial enzymes, of both fungal and bacterial origin. The great versatility of fungal lipases (from genera such as Candida, Geotrichum,

    Rhizopus and Thermomyces) in biotechnology is illustrated extensively by Gandhi [12], Benjamin

    and Pandey [13] and Pandey et al. [14]. Among bacterial lipases, attention has usually been

    focused on particular classes of enzymes such as the lipases from the genus Pseudomonas, which

    are especially interesting for biotechnology [15,16], or esterases possibly involved in bacterial

    pathogenicity [17]. Unfortunately, information on the relatedness of the numerous bacterial lipases and esterases studied so far is incomplete and scattered in the literature.

    Many new bacterial lipolytic enzymes have been studied since the publication of a comprehensive review article in 1994 [18]. However, no attempt has been undertaken to organize this information.

    2+Some biochemical properties (such as the dependence of activity on Ca ions, pH and

    temperature) of the best studied families of lipases (from the genera Bacillus, Pseudomonas and

    Staphylococcus) have been summarized previously [15,16,18,19]. Usually, lipolytic enzymes are

    characterized by their ability to catalyse a broad range of reactions. Unfortunately, the wide diversity of methods used for lipase assays (such as the hydrolysis of p-nitrophenyl esters, the

    pH-stat method and the monolayer technique) prevents a direct comparison of results on substrate specificities. In an effort to standardize the measurements, comparative studies have been performed [20,21]. However, only a limited number of bacterial lipases [16] were investigated in

    these studies.

    In the present paper, 53 sequences of bacterial lipases and esterases are compared and classified according to conserved sequence motifs and the biological properties of these enzymes. Relevant information obtained from the 3D structures is also highlighted when available. This work presents an overview of bacterial lipases and esterases currently known and permits the classification of newly isolated lipolytic enzymes, thereby giving a hint about their general characteristics as a starting point to their investigation.

    DATA SEARCH AND ANALYSIS

    Sequences were retrieved from protein and nucleotide databases by means of the Entrez server at NCBI (http://www.ncbi.nlm.nih.gov/Entrez/), by using the keywords 'bacteria, archaea, lipase, esterase, carboxylesterase'. Sequence similarity searches were performed with the BLAST 2.0 program [22]. Sequence comparison, sorting and alignment were obtained with the help of the Match-Box server [23] and the CLUSTAL W program [24]. The final presentation of results was

    prepared with the MEGALIGN program from the Lasergene software package (DNASTAR, Madison, WI, U.S.A.).

    RESULTS AND DISCUSSION

True lipases

    Pseudomonas lipases

    Bacterial true lipases were formerly ordered in the so-called Pseudomonas groups 1, 2 and 3

    because Pseudomonas lipases were probably the first to be studied and have a preponderant role in

    industry. Because some Pseudomonas species that produce important lipases have recently been

    renamed Burkholderia [25] and because many lipases originate from various other genera, we

    propose a revised classification of true lipases on the basis of six subfamilies (Table 1).

Table 1 Families of lipolytic enzymes

    Amino acid sequence similarities were determined with the program MEGALIGN (DNASTAR),

    with the first member of each family (subfamily) arbitrary set at 100%. Abbreviations: OM, outer

    membrane; PHA, polyhydroxyalkanoate.

     Similarity (%)

    Family Subfamily Enzyme-producing Accession Family Subfamily Properties

    strain no.

    I 1 Pseudomonas D50587 100 True lipases

    aeruginosa*

     Pseudomonas AF031226 95

    fluorescens C9

     Vibrio cholerae X16945 57

     Acinetobacter X80800 43

    calcoaceticus

     Pseudomonas fragi X14033 40

     Pseudomonas U88907 39

    wisconsinensis

     Proteus vulgaris U33845 38

     2 Burkholderia glumae* X70354 35 100

     Chromobacterium Q05489 35 100

    viscosum*

     Burkholderia cepacia* M58494 33 78

     Pseudomonas luteola AF050153 33 77

     3 Pseudomonas D11455 14 100

    fluorescens SIK W1

     Serratia marcescens D13253 15 51

     4 Bacillus subtilis M74010 16 100

     Bacillus pumilus A34992 13 80

     5 Bacillus U78785 15 100

    stearothermophilus

     Bacillus X95309 14 94

    thermocatenulatus

     Staphylococcus hyicus X02844 15 29 Phospholipase

     Staphylococcus aureus M12715 14 28

     Staphylococcus AF090142 13 26

    epidermidis

     6 Propionibacterium X99255 14 100

    acnes

     Streptomyces U80063 14 50

    cinnamoneus

    II Aeromonas hydrophila P10480 100 Secreted (GDSL) acyltransferase

     Streptomyces scabies* M57297 36 Secreted esterase

     Pseudomonas AF005091 35 OM-bound esterase

    aeruginosa

     Salmonella AF047014 28 OM-bound esterase

    typhimurium

     Photorhabdus X66379 28 Secreted esterase

    luminescens

    III Streptomyces M86351 100 Extracellular lipase

    exfoliatus*

     Streptomyces albus U03114 82 Extracellular lipase

     Moraxella sp. X53053 33 Extracellular

    esterase 1 IV Alicyclobacillus X62835 100 Esterase (HSL) acidocaldarius

     Pseudomonas sp. AF034088 54 Lipase

    B11-1

     Archaeoglobus AE000985 48 Carboxylesterase

    fulgidus

     Alcaligenes eutrophus L36817 40 Putative lipase

     Escherichia coli AE000153 36 Carboxylesterase

     Moraxella sp. X53868 25 Extracellular

    esterase 2 V Pseudomonas M58445 100 PHA-depolymerase

    oleovorans

     Haemophilus U32704 41 Putative esterase

    influenzae

     Psychrobacter X67712 34 Extracellular

    immobilis esterase

     Moraxella sp. X53869 34 Extracellular

    esterase 3

     Sulfolobus AF071233 32 Esterase

    acidocaldarius

     Acetobacter AB013096 20 Esterase

    pasteurianus

    VI Synechocystis sp. D90904 100 Carboxylesterases

     Spirulina platensis S70419 50

     Pseudomonas S79600 24

    fluorescens*

     Rickettsia prowazekii Y11778 20

     Chlamydia AE001287 16

    trachomatis

    VII Arthrobacter oxydans Q01470 100 Carbamate

    hydrolase

     Bacillus subtilis P37967 48 p-Nitrobenzyl

    esterase

     Streptomyces CAA22794 45 Putative

    coelicolor carboxylesterase

    VIII Arthrobacter AAA99492 100 Stereoselective

    globiformis esterase

     Streptomyces CAA78842 43 Cell-bound esterase

    chrysomallus

     Pseudomonas AAC60471 40 Esterase III

    fluorescens SIK W1

    * Lipolytic enzyme with known 3D structure.

The Burkholderia glumae lipase was for a long time the only bacterial lipase with a known 3D

    structure [26], until the publication of the crystal structures of the lipases from Chromobacterium viscosum [27] and from Burkholderia cepacia [28,29]. All these enzymes belong to family I.2 of true lipases. Very recently the crystal structure of the Ps. aeruginosa lipase was solved (D. Lang, K. E. Jaeger and B. W. Dijkstra, unpublished work) providing the first structure in the lipase

    family I.1.

    Since the publication of comparative studies on Pseudomonas lipases [15,16], the sequences of lipases from several bacterial genera were reported that are obviously related to families I.1 and

    I.2 on the basis of amino acid sequence comparison (Table 1 and Figure 1). Lipases from Vibrio

    cholerae, Acinetobacter calcoaceticus, Ps. wisconsinensis and Proteus vulgaris have molecular masses in the range 3032 kDa and display a higher sequence similarity to the Ps. aeruginosa

    lipase. Enzymes from subfamily I.2 are characterized by a slightly larger size (33 kDa) owing to

    an insertion in the amino acid sequence forming an anti-parallel double b-strand at the surface of

    the molecule [26,28]. The Ps. luteola lipase possesses this insertion (residues 254272 in the

    preprotein) and shows a high similarity to the Burkholderia enzymes, notably in this region (Figure 1).

    The expression in an active form of lipases belonging to subfamilies I.1 and I.2 depends on a chaperone protein named lipase-specific foldase ('Lif'). However, such specific helper proteins have yet not been described for Ps. fluorescens C9, Ps. fragi, Ps. vulgaris and Ps. luteola. Both

    subfamilies also share important structural features, which are shown in Figure 1. Apart from the

    2+residues forming the catalytic triad, two aspartic residues involved in the Ca-binding site

    described in the crystal structures are found at homologous positions in all sequences. Two cysteine residues forming a disulphide bridge are conserved in a majority of sequences. Because

    2+the residues involved in the formation of both the Ca-binding site and the disulphide bridge are

    located in the vicinity of the catalytic His and Asp residues, they are believed to be important in the stabilization of the active centre of these enzymes [28]. The two Cys residues of the Ps.

    fluorescens C9 lipase do not lie at equivalent positions and no information is available on the possible existence of a disulphide bridge in this molecule. Ps. fragi and Ps. vulgaris lipases

    contain only one Cys residue.

    Subfamily I.3 contains enzymes from at least two distinct species: Ps. fluorescens and Serratia

    marcescens. These lipases have in common a higher molecular mass than lipases from subfamilies I.1 and I.2 (Ps. fluorescens, 50 kDa; S. marcescens, 65 kDa) and the absence of an N-terminal

    signal peptide and of Cys residues. The secretion of these enzymes occurs in one step through a three-component ATP-binding-cassette transporter system [30,31].

    Lipases from Gram-positive organisms

    The various Bacillus lipases known have in common that an alanine residue replaces the first glycine in the conserved pentapeptide: Ala-Xaa-Ser-Xaa-Gly. However, the lipases from the two mesophilic strains B. subtilis and B. pumilus stand apart because they are the smallest true lipases

    known (approx. 20 kDa) and share very little sequence similarity (approx. 15%) with the other Bacillus and Staphylococcus lipases.

    B. thermocatenulatus and B. stearothermophilus produce lipases with similar properties. Their

    molecular mass is approx. 45 kDa and they display maximal activity at approx. pH 9.0 and 65 ?C [32,33].

    Staphylococcal lipases are larger enzymes (approx. 75 kDa) that are secreted as precursors and cleaved in the extracellular medium by a specific protease, yielding a mature protein of approx. 400 residues. The propeptide (207267 residues) presumably acts as an intramolecular chaperone and facilitates the translocation of the lipase across the cell membrane [19]. Interestingly, the

    lipase from Staphylococcus hyicus also displays a remarkable phospholipase activity [34], which

    is unique among true lipases.

    Other lipases

    The lipases from Propionibacterium acnes (339 residues) [35] and from Streptomyces

    cinnamoneus (275 residues) [36] show significant similarity to each other (39% identity, 50% similarity). The central region of these proteins (residues 50150) is approx. 50% similar to

    lipases from B. subtilis and from subfamily I.2. No similarity was found between the Strep.

    cinnamoneus lipase and other Streptomyces lipases known so far.

    The GDSL family

    The enzymes grouped in family II do not exhibit the conventional pentapeptide Gly-Xaa-Ser-Xaa-Gly but rather display a Gly-Asp-Ser-(Leu) [GDS(L)] motif containing the active-site serine residue (Figure 2). In these proteins this important residue lies much closer to the N-terminus than in other lipolytic enzymes [17]. We included in this family the esterase from

Strep. scabies because of its significant similarity to the Aeromonas hydrophila esterase (30%).

    Convincingly, this similarity is not restricted to the vicinity of functionally important residues but is distributed over the entire sequence. As shown by its crystal structure [37], the catalytic centre

    of Strep. scabies esterase has a particular architecture in that it forms a catalytic dyad instead of a triad. The acidic side chain, which usually stabilizes the positive charge of the active-site histidine residue, is replaced by the backbone carbonyl of the residue located three positions upstream of the histidine itself, namely Trp-315. Interestingly, a second enzyme displaying the GDSL motif, the a1 subunit of the platelet-activating-factor acetylhydrolase (a1PAF-AH) from bovine brain, shows a catalytic triad in which an aspartic residue also lies three positions upstream of the active-site histidine [38]. Both enzymes have an a/b tertiary fold substantially different from that of the a/b-hydrolase family and share conserved sequence blocks with at least four other bacterial esterases, as shown in Figure 2.

    For the Aeromonas hydrophila esterase, Brumlik and Buckley [39] proposed that the active-site

    aspartic residue is Asp-116 located in block III (Figure 2), arguing that the Asp-116 Asn

    mutant is completely inactive. However, both the absence of activity of the enzyme and its severely impaired secretion, which was also reported, might be due to the misfolding of the protein in the periplasm and to its subsequent proteolytic degradation without implying that Asp-116 belongs to the catalytic triad. However, it was shown in the above-mentioned bovine a1PAF-AH [38] that a second conserved aspartic residue located three positions upstream from the active histidine (Figure 2) can take part in the active site and that this third acidic residue is not essential for the enzyme's function, as in the Strep. scabies esterase [37]. Obviously, more

    structural information is needed to establish whether these enzymes share a common fold and a common architecture of their catalytic triad (or dyad) in addition to the similarity of their sequences.

    Another interesting feature of the GDSL esterases from Ps. aeruginosa, Salmonella typhimurium

    and Photorhabdus luminescens is an additional C-terminal domain that encompasses

    approximately one-third of their entire sequence and is similar to that of a newly identified family of autotransporting bacterial virulence factors ([40,41], and S. Wilhelm, J. Tommassen and K.-E.

    Jaeger, unpublished work). In these proteins the C-terminal domain is presumably folded into approx. 12 amphipathic b-sheets forming an aqueous pore in the outer membrane. The catalytic N-terminal domain transits through this pore and is in some instances released in the extracellular medium by a specific proteolytic process.

    Family III

    This family of lipases was identified primarily by Cruz et al. [43] and mentioned by Wei et al. [44],

    who solved the 3D structure of the Strep. exfoliatus (M11) lipase. This enzyme displays the

    canonical fold of a/b-hydrolases and contains a typical catalytic triad. It also shows approx. 20% amino acid sequence identity with the intracellular and plasma isoforms of the human PAF-AH. These PAF-AHs are monomer proteins, in contrast with the heterotrimeric PAF-AH from bovine brain. Their tertiary fold was modelled on the basis of the Strep. exfoliatus lipase structure [44].

    Their active-site aspartic residue, identified primarily by site-directed mutagenesis, was shown to be located in the sequence at a position non-equivalent to that found in the Strep. exfoliatus

    enzyme, again underlining the great functional versatility of the a/b-hydrolase scaffold. The hormone-sensitive lipase (HSL) family

    A number of bacterial enzymes (family IV) display a striking amino acid sequence similarity to

the mammalian HSL [45]. Figure 3 shows sequence blocks that are highly conserved in HSL and

    six lipolytic enzymes from distantly related prokaryotes. The proposed active-site residues, which were inferred from the three-dimensional model of the human HSL [46], are also highlighted. The

    mammalian HSL seems to derive from a catalytic domain, homologous with the bacterial enzymes, merged with an additional N-terminal domain and a regulatory module inserted in the central part of the sequence. The relatively high activity at low temperature (less than 15 ?C) retained by HSL and the lipase from Moraxella sp. [47] was once thought to derive from the conserved sequence

    motifs of these enzymes [48]. However, the marked sequence similarity (Table 1 and Figure 3)

    between esterases from psychrophilic (Moraxella sp., Psychrobacter immobilis), mesophilic

    (Escherichia coli, Alcaligenes eutrophus) and thermophilic (Alicyclobacillus acidocaldarius,

    Archeoglobus fulgidus) origins indicates that temperature adaptation is not responsible for such an extensive sequence conservation. A comparative study of these enzymes would therefore be very valuable for the determination of the distinctive properties of this family of hydrolases. Family V

    Like proteins in the HSL family, enzymes grouped in family V originate from mesophilic bacteria (Pseudomonas oleovorans, Haemophilus influenzae, Acetobacter pasteurianus) as well as from

    cold-adapted (Moraxella sp., Psy. immobilis) or heat-adapted (Sulfolobus acidocaldarius)

    organisms. They share significant amino acid sequence similarity (2025%) to various bacterial

    non-lipolytic enzymes, namely epoxide hydrolases, dehalogenases and haloperoxidase, which also possess the typical a/b-hydrolase fold and a catalytic triad [49,50]. On the basis of the crystal

    structure of the Xanthobacter autotrophicus dehalogenase [49], the tertiary fold and the active site

    residues of the Psy. immobilis lipase were predicted by molecular modelling [51]. The sequence

    patterns conserved around the active-site residues of family V enzymes are presented in Figure 3.

    Family VI

    With a molecular mass in the range 2326 kDa, the enzymes presented here are among the

    smallest esterases known. The 3D structure of the Ps. fluorescens carboxylesterase was solved

    [52]. The active form of this enzyme is a dimer. The subunit has the a/b-hydrolase fold and a classical Ser-Asp-His catalytic triad. This carboxylesterase hydrolyses small substrates with a broad specificity and displays no activity towards long-chain triglycerides [53]. Very little is

    known about the other enzymes in this family. Their amino acid sequences were derived from whole-genome sequences except that for the Spirulina platensis esterase, which was cloned

    specifically [54]. The enzymes in family VI display approx. 40% sequence similarity to eukaryotic

    2+lysophospholipases (Ca-independent phospholipases A). Their major conserved sequence 2

    motifs are shown in Figure 3.

    Family VII

    A number of rather large bacterial esterases (55 kDa) share significant amino acid sequence homology (30% identity, 40% similarity) with eukaryotic acetylcholine esterases and intestine/liver carboxylesterases. The esterase from Arthrobacter oxydans is particularly active

    against phenylcarbamate herbicides by hydrolysing their central carbamate bond [55]. It is

    plasmid-encoded and is therefore potentially transmissible to other strains or species. The B.

    subtilis esterase was found to efficiently hydrolyse p-nitrobenzyl esters. It can therefore be used to

    advantage in the final removal of the p-nitrobenzyl ester used as a protecting group in the

    synthesis of b-lactam antibiotics [56]. The genome sequencing project of Strep. coelicolor

    revealed a putative open reading frame corresponding to a carboxylesterase; however, this protein

has not yet been characterized.

    Family VIII

    The three enzymes forming this family are approximately 380 residues long and show a striking similarity to several class C b-lactamases. A stretch of 150 residues (from positions 50 to 200) is, notably, 45% similar to an Enterobacter cloacae ampC gene product [57]. This feature suggests

    that the esterases in family VIII possess an active site more reminiscent of that found in class C b-lactamases, which involves a Ser-Xaa-Xaa-Lys motif conserved in the N-terminal part of both enzyme categories [58,59]. In contrast, Kim et al. [60] proposed that the esterase/lipase consensus

    sequence Gly-Xaa-Ser-Xaa-Gly that appears in the Ps. fluorescens esterase would be involved in

    the active site of the enzyme. However, this motif, which is also present in the Strep. chrysomallus

    esterase, is not conserved in the Arthrobacter globiformis enzyme. Moreover, the motif lies near

    the C-terminus of the Ps. fluorescens and Strep. chrysomallus enzymes and no histidine residue

    follows it in the sequence. This implies that the order of the catalytic residues in the sequence (Ser-Asp-His) that is conserved throughout the entire superfamily of lipases and esterases would not be respected in this case. Obviously, more structural information is needed to describe unambiguously the catalytic mechanism of the family VIII esterases.

    Conclusions

    Despite a highly conserved tertiary fold and obvious sequence similarities, lipolytic enzymes display a wide diversity of properties and of relatedness to other protein families. In an attempt to help the microbiologist confronted with a new bacterial lipolytic enzyme, we have tried to distinguish between subgroups in this large family and to summarize the current knowledge available for each group.

    By consulting the protein and gene databases and using the keywords 'lipase, esterase, carboxylesterase' combined with 'bacteria, archaea' we found 217 entries, of which many turned out to be redundant, corrected or closely related sequences. We therefore restricted our analysis to the 53 sequences listed in Table 1. We are aware that some relevant sequences might have been

    overlooked in this procedure and that many others will appear, especially from the continuing genome-sequencing projects. Nevertheless we hope that this work will serve as a basis for a more complete and evolving classification of bacterial lipolytic enzymes as more structural and kinetic information becomes available.

    This work was supported by EC grant BIO4-CT97-5023.

Report this document

For any questions or suggestions please email
cust-service@docsford.com