Novel Transmembrane Lipases of Alpha/Beta Hydrolase Fold

FEBS Letters 585 (2011) 870–874

journal homepage: www.FEBSLetters.org

Novel transmembrane lipases of alpha/beta hydrolase fold ⇑ Michal Lazniewski a,b, Kamil Steczkiewicz a, Lukasz Knizewski a, Iwona Wawer b, Krzysztof Ginalski a, a Laboratory of Bioinformatics and Systems Biology, Interdisciplinary Centre for Mathematical and Computational Modelling, University of Warsaw, Pawinskiego 5a, 02-106 Warsaw, Poland b Department of Physical Chemistry, Faculty of Pharmacy, Medical University of Warsaw, Banacha 1, 02-097 Warsaw, Poland article info abstract

Article history: Processing of exogenous glycerol esters is an initial step in energy derivation for many bacterial Received 9 November 2010 cells. Lipid-rich environments settled by a variety of organisms exert strong evolutionary pressure Revised 4 February 2011 for establishing enzymatic pathways involved in lipid metabolism. However, a certain number of Accepted 14 February 2011 enzymes involved in this process remain unknown since they do not share detectable sequence sim- Available online 17 February 2011 ilarity with any known protein domains. Using distant homology detection and fold recognition we Edited by Takashi Gojobori predict that bacterial transmembrane proteins belonging to the uncharacterized domain of unknown function 2319 (DUF2319) family possess the alpha/beta hydrolase fold domain together with the catalytic triad critical for hydrolysis. A detailed analysis of sequence/structure features Keywords: Domain of unknown function 2319 and genomic context indicates that DUF2319 proteins may be involved in lipid metabolism. There- Lipase fore, these enzymes are likely to serve as extracellular lipases. Alpha/beta hydrolase Ó 2011 Federation of European Biochemical Societies. Published by Elsevier B.V. All rights reserved. Fold recognition Function assignment

1. Introduction bacterial proteins belonging to uncharacterized domain of unknown function 2319 (DUF2319) family. Exogenous lipids, consisting mainly of glycerol esters are, next to carbohydrates, one of the main sources of energy and carbon 2. Materials and methods for living cells [1]. In order to be utilized in respective metabolic pathways they need to undergo several enzymatic modifications. Proteins belonging to DUF2319 family (PF10081 accession First, the hydrolysis of ester bond is catalyzed by lipases. In this number in PFAM database [4]) were collected with exhaustive process free fatty acids (FFA) are released and if that reaction is car- PSI-BLAST [5] searches (E-value threshold 0.005) performed ried out outside the cell membrane, FFA can be further transported against NCBI non-redundant protein sequence database using into the cell by the fatty acid-transport proteins. Finally, during DUF2319 consensus sequence as a query. Multiple sequence align- beta-oxidation, long-chain fatty acids are split into acetyl-CoA ment was generated using PCMA program [6] and then manually molecules that enter citric acid cycle responsible for producing adjusted. For every sequence the secondary structure was pre- ATP and therefore, the energy for the cells. Since a number of spe- dicted with PSIPRED [7]. Distantly related families and structures cies have adapted to living in a lipid-rich environment, like the were identified with the meta-profile comparison method Meta- lungs (alveoli are padded with surfactant rich in phospholipids) BASIC [8] and fold recognition server 3D-Jury [9] using the consen- or the skin (coated with sebum composed of 41% triglycerides), sus sequence of DUF2319 and several family members. The se- there are many genes coding for enzymes involved in exogenous quence-to-structure alignment between DUF2319 proteins and lipids metabolism [1–3]. For example, in Mycobacterium tuberculo- selected structures in the fold core region was built using consen- sis more than five individual genes are known to encode lipases [3]. sus alignment approach and 3D assessment based on the results of In this work, using distant homology detection and fold recognition Meta-BASIC, 3D-Jury and secondary structure predictions, as well methods followed by a detailed sequence analysis, we have identi- as conservation of critical active site residues and hydrophobic pat- fied a novel family of transmembrane lipases among hypothetical terns [10]. The 3D model of M. tuberculosis H37Rv DUF2319 protein (gi|15608209) was built with MODELLER [11] using selected struc- Abbreviations: DUF, domain of unknown function; NOG, non-supervised tures of alpha/beta hydrolases as templates: Fusarium solani cutin- orthologous group; FFA, free fatty acids ase (pdb|1cex) [12], Aspergillus oryzae cutinase (pdb|3gbs) [13], Bos ⇑ Corresponding author. Fax: +48 22 5540801. taurus thioesterase (pdb|1eh5) [14] and Pseudomonas aeruginosa E-mail address: [email protected] (K. Ginalski).

0014-5793/$36.00 Ó 2011 Federation of European Biochemical Societies. Published by Elsevier B.V. All rights reserved. doi:10.1016/j.febslet.2011.02.016 M. Lazniewski et al. / FEBS Letters 585 (2011) 870–874 871 lipase (pdb|1ex9) [15]. Genomic context analysis was performed cal proteins from various prokaryotic organisms belonging to using STRING [16] in order to detect possible functional associa- four phyla: Actinobacteria, Proteobacteria, Planctomycetes and tions. Operon organization was investigated using Operon Predic- Gemmatimonadetes. Among them some anxious pathogens can tions Tool [17] and OperonDB [18]. Structure visualization was be found, including M. tuberculosis and Pseudomonas stutzeri,in carried out with Pymol (http://www.pymol.org). addition to nitrogen fixating Rhizobium or industrially utilized Corynebacterium glutamicum. Application of other standard 3. Results and discussion sequence comparison methods such as CDD [19] and SMART [20] did not reveal any significant similarity to other known protein do- 3.1. DUF2319 belongs to the superfamily of alpha/beta hydrolase fold mains or structures. Interestingly, TOPCONS [21] predicted five proteins transmembrane a-helices at the N-terminus of DUF2319 proteins, with the C-terminal region predicted to be located outside the cell. Exhaustive PSI-Blast [5] searches against NCBI nr database initi- Using Meta-BASIC [8], an advanced method for distant homology ated with DUF2319 consensus sequence identified 136 hypotheti- detection, we confidently mapped the C-terminal region of DUF2319 consensus sequence onto the PF01083 family of cutinas- Table 1 es with a confidence score >40 (predictions with scores over 40 are Results of distant homology detection (Meta-BASIC) and fold recognition (3D-Jury) for considered to be 90% correct according to various rigorous M. tuberculosis DUF2319 sequence (gi|15608209). benchmarks [22]). Further Meta-BASIC searches started with PDB ID/PFAM Meta-BASIC 3D-Jury Function M. tuberculosis sequence (gi|15608209) pointed to two additional ID score score PFAM families: DUF900 and DUF1023 [23] with Meta-BASIC scores PF05990 48.12 –* Alpha/beta hydrolase of of 48 and 36, respectively. All identified families belong to the (DUF900) unknown function alpha/beta hydrolase superfamily. Meta-BASIC predictions were PF01083 41.51 –* Cutinase further confirmed by the consensus fold recognition server 3D-Jury pdb|3gbs 40.10 48.29 Cutinase [9] that provided consistent matches with reliable scores >40 [24] PF06259 36.08 –* Potential lipase of alpha/beta (DUF1023) hydrolase fold to several fungal cutinase structures from A. oryzae (pdb|3gbs) and pdb|1cuw 35.43 54.29 Serine esterase F. solani (pdb|1cex and pdb|1cuw) (see Table 1). pdb|1cex 35.40 53.86 Serine esterase SCOP database [25] divides the large and diverse superfamily of alpha/beta hydrolases into 41 distinct families including (i) carb- * 3D-Jury score can be calculated only for PDB structures. oxylesterases, taking part in hydrolysis of short chain aliphatic

Fig. 1. Multiple sequence alignment for DUF2319 family representatives and selected alpha/beta hydrolase fold structures. Sequences are labeled according to the NCBI gi number or PDB code and an abbreviation of the species name: Mt, Mycobacterium tuberculosis; Sm, Sinorhizobium medicae; Nm, Nakamurella multipartita; Ps, Pseudomonas stutzeri; Rb, Rhodopirellula baltica; Ga, Gemmatimonas aurantiaca; Gb, Gordonia bronchialis; Mm, Mycobacterium marinum; Re, Rhodococcus erythropolis; Cj, Corynebacterium jeikeium, Cg, Corynebacterium glutamicum; Ck, Corynebacterium kroppenstedtii; Pa, Propionibacterium acnes. The numbers of residues that are not shown are specified in parentheses. Residue conservation in DUF2319 family is presented with the following scheme: uncharged, highlighted in yellow; charged or polar, highlighted in grey; small, letters in red; invariant catalytic triad (S, D, H), highlighted in red. Locations of predicted (gi|15608209) and observed (pdb|1cex) secondary structure elements (E, b-strand; H, a-helix) are marked above the corresponding sequences. Locations of two insertion regions, INS1 and INS2, are also shown. For each sequence its respective genomic context is provided. 872 M. Lazniewski et al. / FEBS Letters 585 (2011) 870–874 and aromatic esters, (ii) gastric, bacterial and fungal lipases, which are responsible for releasing FFA from lipids, and (iii) cutinases that hydrolyze cutin, a polymer of hydroxy-fatty acids. These enzymes catalyze the hydrolysis of different substrates and play an important role in various metabolic pathways involved in acquiring energy for the cell, as well as in the synthesis and degradation of a variety of compounds. The structure of these proteins is defined by a common fold consisting of three layers (a/b/a) with the cen- tral b-sheet composed of eight strands with 12435678 order (the second b-strand is antiparallel to the rest). Yet in a number of structures this structural core can be reduced; for example, in F. solani cutinase (pdb|1cex) and in Thermus thermophilus HB8 TTHA1544 protein (pdb|2dst) [6] it embraces only six parallel b- strands with 243567 and 124356 strand order, respectively. The enzymes retaining alpha/beta hydrolase fold are characterized by the presence of the catalytic triad responsible for the cleav- age reaction [26]. The triad consists of three conserved residues: (i) ‘nucleophile’ residue (serine, cysteine or aspartate) that in the canonical version of the fold is placed at the sharp turn called ‘nucleophile elbow’ after the fifth b-strand, (ii) the active site ‘acid’ residue (glutamate or aspartate) that is positioned after the seventh or the eighth b-strand, and (iii) histidine located in the loop after the last core b-strand. Although histidine is an invariant superfamily residue, the position and length of the loop on which Fig. 2. 3D model for DUF2319 C-terminal catalytic domain from M. tuberculosis this amino acid is located may vary significantly. Alpha/beta H37Rv (gi|15608209). Catalytic triad (S403, H547 and D534) is shown in red, hydrolases bind a variety of ligands: lipids, peptides and ethers, whereas conserved hydrophobic residues forming hydrophobic pocket, responsible for substrate binding, are highlighted in green. Strands numbering is consistent which entails a significant diversity of their structures [26]. Sub- with the nomenclature commonly used for eight-stranded alpha/beta hydrolases. strate specificity is usually shaped by the presence and the length DUF2319 proteins lack b1 strand of the canonical fold. INS1 corresponds to the of loops between the third and the eighth core b-strands. Interest- insertion between b2 and b3, whereas INS2 represents larger insertion after b6, ingly, several lipases form a labile lid after the sixth core strand which embraces amino acids responsible for ‘lid’ formation. that switches between closed and opened conformations. Its role is to shield the active site region from the access of solvent molecules and further, it is believed to be important for lipase activation mae (pdb|1tah) [29] and P. aeruginosa (pdb|1ex9) it functions as a [26]. The hydrophobicity of catalytic triad environment is an lid preventing the solvent from access to the active site and is in- important feature defining ligand specificity. For instance, in cutin- volved in enzymes interfacial activation. In the latter case the lid ases no lid is observed, yet the ‘gateway’ leading to the active site is consists mostly of hydrophobic residues forming short a-helices hydrophobic [13]. separated with long loops. Although a few DUF2319 family mem- As shown in Fig. 1, DUF2319 proteins conserve the core struc- bers have shorter insertion, in most cases it embraces nearly 100 tural elements and all amino acids critical for the enzymatic activ- amino acids, which are mostly hydrophobic, thus they may serve ity of alpha/beta hydrolases. Catalytic triad is present in DUF2319 as a lid or may be involved in protein aggregation. In addition, in its most canonical form, including S403, D534 and H547 resi- DUF2319 proteins retain hydrophobic ‘gateway’ directly flanking dues (positions from M. tuberculosis gi|15608209). Furthermore, the active site. W361, F364, L365, V541, and P542 (numbering DUF2319 proteins preserve the GXSXG sequence motif (where X from M. tuberculosis gi|15608209) form a hydrophobic pocket stands for any amino acid), which is characteristic of alpha/beta potentially responsible for substrate specificity of DUF2319 en- hydrolases. The sequence variation within this motif is one of the zymes that are thus expected to bind hydrophobic ligands (Fig. 2). factors distinguishing various lipolytic families [27]. For example, Genomic context analysis provides additional hints as concerns lipolytic enzymes called true lipases (from I.1 and I.2 families) the possible functional role of DUF2319 family proteins. Using conserve the GHS(H/Q)G sequence, whereas the GDSLS motif is STRING database [16] we identified two non-supervised ortholo- characteristic of GDSL family [28]. In DUF2319 this motif is repre- gous group (NOG) families, NOG43037 and NOG06582, as sharing sented by the G(E/L)SLG sequence, which might suggest that these neighborhood and co-occurrence with genes encoding DUF2319 proteins are closely related to GDSL enzymes and may share a sim- proteins. The confidence score of this genomic association is mod- ilar broad substrate specificity toward triglycerides or lysophos- erate (0.612 and 0.492), yet the results of our analyses show that pholipids. However, two catalytic residues: histidine and aspartic these NOGs together with DUF2319 may be involved in a common acid that are located in close proximity to each other within GDSL metabolic pathway. NOG43037 is classified as glyoxylase enzyme family sequences (DXXH motif), in DUF2319 are separated by a that is part of the glyoxalase system responsible for detoxification much longer loop. Therefore, DUF2319 is likely to constitute a of methylglyoxal and other reactive aldehydes. Methylglyoxals are new family of lipolytic enzymes. by-products during carbohydrate and lipid metabolism [30,31]. NOG06582 embraces yet uncharacterized proteins. Nevertheless, 3.2. DUF2319 family proteins are potential extracellular lipases using Meta-BASIC we detected a significant similarity between NOG06582 and human Sterol Carrier protein type 2 (also known DUF2319 family is characterized by the presence of a large as non-specific lipid transfer protein), responsible for transporting insertion between the sixth and the seventh b-strands (marked sterols and various other lipids through the cell membrane [32]. as INS2 in Figs. 1 and 2). A similar insertion can be observed in Interestingly, in microbial genomes containing DUF2319, genes the interfacial activated lipases or thioesterases. For instance, in located in the proximity to DUF2319 are often involved in the B. taurus thioesterase (pdb|1eh5) such insertion serves as a sub- beta-oxidation pathway, i.e., they encode enoyl-CoA hydratase or strate binding domain, whereas in lipases from Burkholderia glu- acyl-CoA dehydrogenase (Fig. 1). This situation is quite common, M. Lazniewski et al. / FEBS Letters 585 (2011) 870–874 873

Fig. 3. Potential role of DUF2319 proteins. DUF2319 may hydrolyze triglycerides or lysophospolipids. In consequence, FFA are released and transported to the bacterial cell, where they undergo further modifications. Subsequently, this leads to the production of acetyl-CoA, which can be converted to a number of other metabolites. R1 indicates several polar head groups of phospholipids, whereas R2 and R3 correspond to carbon chains of various fatty acids. especially for microorganisms from Actinobacteria and Proteobac- belonging to Nocardia possessing DUF2319 proteins, are em- teria, although a number of organisms (e.g., Gemmatimonas auran- ployed in purification of water contaminated with oil. DUF2319 tiaca) lack functional annotation of genes adjacent to DUF2319 is also widely found among lipolytic Corynebacteria. Therefore, genes. Nevertheless, we identified the proteins encoded by these proteins belonging to DUF2319 family are likely to be an essential genes to contain transmembrane regions, which is consistent with part of lipid metabolism in the above organisms by catalyzing the our hypothesis that DUF2319 functions together with other mem- release of FFA from glycerol esters (see Fig. 3). DUF2319 lipases brane-anchored proteins. may thus be an attractive target for new antibacterial com- To verify whether genes encoding DUF2319 family members pounds, since blocking the initial step of lipid processing would belong to operons involved in lipid metabolism we used Operon interrupt the overall cell metabolism leading to microorganism Predictions Tool [17] and OperonDB [18]. M. tuberculosis strain F11 starvation and death. TBFG_11087 gene (DUF2319) is predicted to be part of the operon (mtf_TBFG_11089) together with two other genes, TBFG_ 11089 Acknowledgements and TBFG_11088, encoding enoyl-CoA hydratases, which are involved in beta-oxidation of FFA [33]. Similarly, Rv1069c This work was supported by EMBO Installation, FNP, MNiSW (N (DUF2319) from M. tuberculosis strain H37Rv is likely to belong to N301 159435), Innovative Economy 2007–2013 Programme the same operon with enoyl-CoA hydratases encoded by Rv1070c (POIG.01.01.02-14-054/09-00, POIG.02.02.00-14-024/08-00), Med- and Rv1071c genes. Other Mycobacteriaceae species such as ical University of Warsaw (FW/28/N/2010) and European Social M. marinum or M. bovis seem to have a similar operon structure. Fund (UDA-POKL.04.01.01-00-072/09-00) grants. The calculations Altogether, this supports our hypothesis that DUF2319 proteins were performed in the Interdisciplinary Centre for Mathematical are involved in beta-oxidation of FFA (at least in Mycobacteriaceae). and Computational Modelling under computational Grants G14-6 and G30-2.

4. Conclusions References

In this work we have identified novel potential extracellular li- [1] Azizan, A., Sherin, D., DiRusso, C.C. and Black, P.N. (1999) Energetics pases among previously uncharacterized transmembrane proteins underlying the process of long-chain fatty acid transport. Arch. Biochem. of unknown function. These enzymes may catalyze the initial Biophys. 365, 299–306. [2] Chroneos, Z.C., Midde, K., Sever-Chroneos, Z. and Jagannath, C. (2009) steps in the modification of various glycerol esters (triglycerides, Pulmonary surfactant and tuberculosis. Tuberculosis (Edinb) 89 (Suppl. 1), lysophospholipids or phospholipids), which are essential for ATP S10–S14. production in the cell. Remarkably, DUF2319 proteins can be [3] Cole, S.T. et al. (1998) Deciphering the biology of Mycobacterium tuberculosis from the complete genome sequence. Nature 393, 537–544. found mostly in microorganisms that live in lipid-rich environ- [4] Bateman, A. et al. (2004) The Pfam protein families database. Nucleic Acids ment. M. tuberculosis is a well studied human pathogen that in- Res. 32, D138–D141. fests the lungs rich in surfactant (a mixture of proteins and [5] Altschul, S.F., Madden, T.L., Schaffer, A.A., Zhang, J., Zhang, Z., Miller, W. and Lipman, D.J. (1997) Gapped BLAST and PSI-BLAST: a new generation of protein lipids), whereas Propionibacterium acnes thrives in the sebum database search programs. Nucleic Acids Res. 25, 3389–3402. covering the skin. Interestingly, DUF2319 proteins share over [6] Pei, J., Sadreyev, R. and Grishin, N.V. (2003) PCMA: fast and accurate multiple 99% sequence identity among known M. tuberculosis strains sug- sequence alignment based on profile consistency. Bioinformatics 19, 427– gesting that this protein is unlikely to be responsible for the 428. [7] McGuffin, L.J., Bryson, K. and Jones, D.T. (2000) The PSIPRED protein structure observed differences in virulence of these strains. Some species prediction server. Bioinformatics 16, 404–405. 874 M. Lazniewski et al. / FEBS Letters 585 (2011) 870–874

[8] Ginalski, K., von Grotthuss, M., Grishin, N.V. and Rychlewski, L. (2004) [21] Bernsel, A., Viklund, H., Hennerdal, A. and Elofsson, A. (2009) TOPCONS: Detecting distant homology with Meta-BASIC. Nucleic Acids Res. 32, W576– consensus prediction of membrane protein topology. Nucleic Acids Res. 37, W581. W465–W468. [9] Ginalski, K., Elofsson, A., Fischer, D. and Rychlewski, L. (2003) 3D-Jury: a [22] Rychlewski, L. and Fischer, D. (2005) LiveBench-8: the large-scale, continuous simple approach to improve protein structure predictions. Bioinformatics 19, assessment of automated protein structure prediction. Protein Sci. 14, 240– 1015–1018. 245. [10] Ginalski, K. and Rychlewski, L. (2003) Protein structure prediction of CASP5 [23] Zheng, M., Ginalski, K., Rychlewski, L. and Grishin, N.V. (2005) Protein domain comparative modeling and fold recognition targets using consensus alignment of unknown function DUF1023 is an alpha/beta hydrolase. Proteins 59, 1–6. approach and 3D assessment. Proteins 53 (Suppl. 6), 410–417. [24] Ginalski, K. and Rychlewski, L. (2003) Detection of reliable and unexpected [11] Fiser, A. and Sali, A. (2003) Modeller: generation and refinement of homology- protein fold predictions using 3D-Jury. Nucleic Acids Res. 31, 3291–3292. based protein structure models. Methods Enzymol. 374, 461–491. [25] Murzin, A.G., Brenner, S.E., Hubbard, T. and Chothia, C. (1995) SCOP: a [12] Longhi, S., Czjzek, M., Lamzin, V., Nicolas, A. and Cambillau, C. (1997) Atomic structural classification of proteins database for the investigation of sequences resolution (1.0 Å) crystal structure of Fusarium solani cutinase: stereochemical and structures. J. Mol. Biol. 247, 536–540. analysis. J. Mol. Biol. 268, 779–799. [26] Nardini, M. and Dijkstra, B.W. (1999) Alpha/beta hydrolase fold enzymes: the [13] Liu, Z. et al. (2009) Structural and functional studies of Aspergillus oryzae family keeps growing. Curr. Opin. Struct. Biol. 9, 732–737. cutinase: enhanced thermostability and hydrolytic activity of synthetic ester [27] Arpigny, J.L. and Jaeger, K.E. (1999) Bacterial lipolytic enzymes: classification and polyester degradation. J. Am. Chem. Soc. 131, 15711–15716. and properties. Biochem J. 343 (Pt. 1), 177–183. [14] Bellizzi 3rd, J.J., Widom, J., Kemp, C., Lu, J.Y., Das, A.K., Hofmann, S.L. and [28] Akoh, C.C., Lee, G.C., Liaw, Y.C., Huang, T.H. and Shaw, J.F. (2004) GDSL family Clardy, J. (2000) The crystal structure of palmitoyl protein thioesterase 1 and of serine esterases/lipases. Prog. Lipid Res. 43, 534–552. the molecular basis of infantile neuronal ceroid lipofuscinosis. Proc. Natl. [29] Noble, M.E., Cleasby, A., Johnson, L.N., Egmond, M.R. and Frenken, L.G. (1993) Acad. Sci. USA 97, 4573–4578. The crystal structure of triacylglycerol lipase from Pseudomonas glumae [15] Nardini, M., Lang, D.A., Liebeton, K., Jaeger, K.E. and Dijkstra, B.W. (2000) reveals a partially redundant catalytic aspartate. FEBS Lett. 331, 123–128. Crystal structure of Pseudomonas aeruginosa lipase in the open conformation. [30] Thornalley, P.J. (2003) Glyoxalase I – structure, function and a critical role in The prototype for family I.1 of bacterial lipases. J. Biol. Chem. 275, 31219– the enzymatic defence against glycation. Biochem. Soc. Trans. 31, 1343–1348. 31225. [31] Dmitriev, L.F. and Titov, V.N. (2010) Lipid peroxidation in relation to ageing [16] Jensen, L.J. et al. (2009) STRING 8 – a global view on proteins and their and the role of endogenous aldehydes in diabetes and other age-related functional interactions in 630 organisms. Nucleic Acids Res. 37, D412–D416. diseases. Ageing Res. Rev. 9, 200–210. [17] Taboada, B., Verde, C. and Merino, E. (2010) High accuracy operon prediction [32] Szyperski, T., Scheek, S., Johansson, J., Assmann, G., Seedorf, U. and Wuthrich, method based on STRING database scores. Nucleic Acids Res. 38, e130. K. (1993) NMR determination of the secondary structure and the three- [18] Pertea, M., Ayanbule, K., Smedinghoff, M. and Salzberg, S.L. (2009) OperonDB: dimensional polypeptide backbone fold of the human sterol carrier protein 2. a comprehensive database of predicted operons in microbial genomes. Nucleic FEBS Lett. 335, 18–26. Acids Res. 37, D479–D482. [33] Bahnson, B.J., Anderson, V.E. and Petsko, G.A. (2002) Structural mechanism of [19] Marchler-Bauer, A. et al. (2005) CDD: a conserved domain database for protein enoyl-CoA hydratase: three atoms from a single water are added in either an classification. Nucleic Acids Res. 33, D192–D196. E1cb stepwise or concerted fashion. Biochemistry 41, 2621–2629. [20] Letunic, I., Copley, R.R., Pils, B., Pinkert, S., Schultz, J. and Bork, P. (2006) SMART 5: domains in the context of genomes and networks. Nucleic Acids Res. 34, D257–D260.