Mutational analysis of a type II thioesterase associated with nonribosomal peptide synthesis Uwe Linne, Dirk Schwarzer, Gunnar N. Schroeder* and Mohamed A. Marahiel Philipps Universita ¨ t Marburg, Fachbereich Chemie/Biochemie, Germany Recent studies on type II thioesterases (TEIIs) involved in microbial secondary metabolism described a role for these enzymes in the removal of short acyl-S- phosphopantetheine intermediates from misprimed holo-(acyl carrier proteins) and holo-(peptidyl carrier proteins) of polyketide synthases and nonribosomal peptide synthetases. Because of the absence of structural information on this class of enzymes, we performed a mutational analysis on a prototype TEII essential for efficient production of the lipopeptide antibiotic surfactin (TEII srf ), which led to identification of catalytic and structural residues. On the basis of sequence alignment of 16 TEIIs, 10 single and one double mutant of highly conserved residues of TEII srf were constructed and biochemically investigated. We clearly identified a catalytic triad consisting of Ser86, Asp190 and His216, suggesting that TEII srf belongs to the a/b-hydrolase superfamily. Exchange of these residues with residues with aliphatic side chains abolished enzyme activity, whereas replacement of the active-site Ser86 with cysteine produced an enzyme with marginally reduced activity. In contrast, exchange of the second strictly con- served asparagine (Asp163) with Ala resulted in an active but unstable enzyme, excluding a role for this residue in catalysis and suggesting a structural function. The results define three catalytic and at least one structural residue in a nonribo- somal peptide synthetase TEII. Keywords: catalytic triad; fatty acid synthases; nonribosomal peptide synthesis; peptide synthetases; type II thioesterase polyketide synthases. Enzymes that cleave thioesters are ubiquitous in prokary- otes and eukaryotes, as thioesters appear in many different metabolic processes. For example, thioesterases have been reported to cleave formate from formylated glutathione [1], which is an intermediate in formaldehyde detoxification in plants, and fatty acids from cysteines of lipidated proteins [2]. Most common thioesterases are involved in 4¢-phospho- pantetheine (4¢-Ppant) metabolic processes, such as the synthesis of fatty acids, polyketides, or nonribosomal peptides. Many polyketides and nonribosomal polypeptides produced by bacteria and filamentous fungi are of great pharmacological interest. Among these are molecules that exhibit antibiotic (penicillin, cephalosporin, erythromycin and vancomycin), immunosuppressive (cyclosporin) and cytostatic (bleomycin and epothilone) activities. A common feature is that they are biosynthesized by large modular enzymes, the so called nonribosomal peptide synthetases (NRPSs) and the polyketide synthases (PKSs) [3,4]. During synthesis, all substrates and intermediates are covalently tethered to the enzymatic templates through a thioester linkage [5]. The thiol group of this thioester belongs to 4¢-Ppant, the prosthetic group of the peptidyl carrier proteins (PCPs) and acyl carrier proteins. The post-trans- lational modification (priming) of the carrier proteins is carried out by dedicated 4¢-phosphopantetheine transferases such as Sfp [6–8]. Two types of thioesterase are associated with NRPSs and PKSs: the well-studied integrated type I thioesterase domains (TE domains), which are responsible for the release of the synthesized products from the enzymatic templates [9–11], and the external stand-alone type II thioesterases (TEIIs). Disruption of the corresponding TEII genes in the producer strains inhibited product formation by 80–90% [12–14]. Recently, biochemical studies on TEIIs in polyke- tide synthesis suggested a role in the removal of short acyl chains originating from aberrant decarboxylation of chain extender units from the thiol moiety of the 4¢-Ppant cofactors of acyl carrier proteins [15,16]. In NRPSs there is no such decarboxylation process during product synthe- sis. However, we recently demonstrated that TEIIs associ- ated with NRPSs are also involved in the regeneration of misprimed PCPs by removing short acyl chains from the 4¢-Ppant cofactors [17]. These acyl chains are thought to be transferred to NRPSs during the priming process, because acyl-CoAs, which are present in significant concentration in Correspondence to M. A. Marahiel, Philipps Universita ¨ t Marburg, Fachbereich Chemie/Biochemie, Hans-Meerwein-Strasse, 35032 Marburg, Germany. Fax: + 49 6421 2822191, Tel.: + 49 6421 2825722, E-mail: marahiel@chemie.uni-marburg.de Abbreviations: DTNB, 5,5¢-dithiobis(2-nitrobenzoic acid); NRPS, nonribosomal peptide synthetase; PCP, peptidyl carrier protein; Ppant, phosphopantetheine; PKS, polyketide synthase; T, thiolation domain, referring to the same thing as PCP, but used for the des- cription of proteins (Ôone letter–one domainÕ nomenclature of NRPSs); TEII, type II thioesterase; TNB, 5-thio-2-nitrobenzoic acid; ESI, electrospray ion. *Present address: Institute of Microbiology, ETH-Zu ¨ rich, Schmelzbergstr.7, Zu ¨ rich, Switzerland. (Received 5 January 2004, revised 16 February 2004, accepted 2 March 2004) Eur. J. Biochem. 271, 1536–1545 (2004) Ó FEBS 2004 doi:10.1111/j.1432-1033.2004.04063.x the cells under most growth conditions [18], can also be utilized as substrates by 4¢-phosphopantetheinyl trans- ferases [6]. Interestingly, TEIIs associated with microbial secondary metabolism also show remarkably high sequence similarities (> 20%) and length ( 240–260 residues) to a specialized mammalian (rat) TEII (TEII rat ), which is involved in fatty acid biosynthesis in specialized tissues such as the mammary gland [12,19]. There, it catalyzes the release of short-chain fatty acids which are ingredients of milk [20]. In the case of TEII rat , two residues (Ser101, His237) have been suggested to be part of a catalytic triad [21–23]. Asp236 has been reported to enhance the catalytic activity of the enzyme, although its role in catalysis remains unclear as the catalytic efficiency of the Ala mutant was only marginally reduced (by  40%). Therefore, it is not clear if a catalytic triad or a catalytic diad consisting of only Ser and His is required for catalysis. In this study, we describe a mutational analysis and mechanistic investigations of the microbial TEII Srf (242 amino acids,  28 kDa), which is associated with the production of the secondary metabolite surfactin [12,24]. A total of 11 TEII Srf mutants, including mutations of all strictly conserved residues with functionalized side chains among 16 TEIIs of diverse origin, were generated by site- directed mutagenesis and biochemically characterized to define important catalytic and structural residues and to determine if these enzymes are mechanistically related to TEII rat of primary metabolism. Experimental procedures Sequence alignment to identify highly conserved residues The sequences of 16 TEIIs were retrieved from the pub- licly accessible NCBI database (http://www.ncbi.nih.gov/). Sequences used were derived from NRPS and PKS biosynthetic clusters, in addition to that of the mammalian TEII rat . The sequences ( 250 amino-acid stretches) were aligned using the program MEGALIGN from the DNA Star package, applying the method of Jotun Hein with default parameters. Construction of TEII srf and TEII srf mutant expression plasmids and protein purification Cloning and overproduction of wild-type TEII srf has been described previously [17]. All mutants generated in this study were introduced into plasmid pTEII srf [17]. Mutants were obtained using the Quick Change Site Directed Mutagenesis Kit TM (Stratagene, Heidelberg, Germany) as described by manufacturer. The mutants constructed and the primers used with plasmid pTEII srf as template are summarized in Table 1. All mutant plasmids generated were confirmed by DNA sequencing using the ABI Prism dRhodamine Terminator Cycle Sequencing Ready Reac- tion Kit (v3.0) and an ABI 310 DNA sequencer (both Applied Biosystems, Darmstadt, Germany) according to the manufacturer’s protocols. Escherichia coli M15/pREP4 was transformed separately with all mutant plasmids constructed. Expression and purification of the His 6 -tagged enzymes (except D163A) were performed as previously described for wild-type TEII srf [17]. In the case of D163A, an overnight culture of the expression strain was grown at 22 °C and induced with iso- propyl thio-b- D -galactoside (0.2 m M ). After an additional 3 h incubation at 22 °C, the cells were centrifuged at 4 °C, and D163A was then purified, dialyzed against assay buffer using HiTrap TM Desalting columns (Amersham Biosciences, Freiburg, Germany) according to the manufacturer’s pro- tocols, and assayed for activity. Overproduction and purification of the enzymes after single-step Ni 2+ -affinity chromatography were confirmed by SDS/PAGE [25]. Protein concentrations were assigned Table 1. TEII srf mutants constructed and primers used (mutations introduced are underlined). Mutation Primer Primer sequences (5¢ and 3¢) TEII srf C18A 5¢ CA CAG CTC ATC GCT TTT CCG TTT GCC GGC GGC 3¢ GCC GCC GGC AAA CGG AAA AGC GAT GAG CTG TG TEII srf H85A 5¢ GTG CTG TTC GGA GCC AGT ATG GGC GGA ATG ATC AC 3¢ GT GAT CAT TCC GCC CAT ACT GGC TCC GAA CAG CAC TEII srf S86A 5¢ G CTG TTC GGA CAC GCT ATG GGC GGA ATG ATC ACC 3¢ GGT GAT CAT TCC GCC CAT AGC GTG TCC GAA CAG CAC TEII srf S86C 5¢ GTG CTG TTC GGA CAC TGT ATG GGC GGA ATG ATC AC 3¢ GT GAT CAT TCC GCC CAT ACA GTG TCC GAA CAG CAC TEII srf S112A 5¢ G GCG GTT ATC ATT GCT GCA ATC CAG CCG CC 3¢ GG CGG CTG GAT TGC AGC AAT GAT AAC CGC C TEII srf D163A 5¢ G CCT TCT TTC CGA TCA GCT TAC CGG GCT CTT G 3¢ C AAG AGC CCG GTA AGC TGA TCG GAA AGA AGG C TEII srf D189A 5¢ C TTT AAC GGG CTT GCT GAT AAA AAA TGC ATA CGA GAT GCG G 3¢ C CGC ATC TCG TAT GCA TTT TTT ATC AGC AAG CCC GTT AAA G TEII srf D190A 5¢ C TTT AAC GGG CTT GAT GCT AAA AAA TGC ATA CGA GAT GCG G 3¢ C CGC ATC TCG TAT GCA TTT TTT AGC ATC AAG CCC GTT AAA G TEII srf 5¢ C TTT AAC GGG CTT GCT GCT AAA AAA TGC ATA CGA GAT GCG G DD189/190AA 3¢ C CGC ATC TCG TAT GCA TTT TTT AGC AGC AAG CCC GTT AAA G TEII srf H216L 5¢ CAA TTT GAC GGC GGG CTC ATG TTC CTG CTG TC 3¢ GA CAG CAG GAA CAT GAG CCC GCC GTC AAA TTG Ó FEBS 2004 Mechanistic studies on TEII srf (Eur. J. Biochem. 271) 1537 using UV spectroscopy (A 280 ). The calculated absorption coefficients of wild-type TEII srf and TEII srf mutants are very similar. The value of the wild-type TEII srf (16 860 M )1 Æcm )1 ) was used for calculation of all enzyme concentrations. After addition of 10% (v/v) glycerol, enzymes were shock-frozen in liquid nitrogen and could be stored at )80 °C over several weeks without significant loss of activity, except TEII srf - D163A, which seems to be unstable and was directly assayed after expression and purification. Enzymatic preparation of acetyl- S -4¢-Ppant-PCP The preparation of acetyl-S-Ppant-PCP, which was used as substrate for TEII srf and TEII srf mutants, was carried out according to previously developed protocols [17]. 5,5¢-Dithiobis(2-nitrobenzoic acid) (DTNB)-HPLC assay TEII srf or TEII srf mutants at a concentration of 25 n M were incubated at 37 °C with various concentrations of acetyl- S-Ppant-PCP (2–35 l M inthecaseofactivemutants;initial concentration for activity test of all mutants was 25 l M )in the presence of DTNB (8 lLofa10m M stock solution) in assay buffer (50 m M Hepes, 100 m M NaCl, 1 m M EDTA, 10 m M MgCl 2 ,pH7.0)inatotalvolumeof400lL. At defined time points, 50 lL samples were collected, and the reaction was subsequently stopped by the addition of 100 lL methanol/trifluoroacetate (500 : 1, v/v). The ratio of acetyl-S-Ppant-PCP to 5-thio-2-nitrobenzoic acid (TNB)- S-Ppant-PCP was analyzed by the HPLC method described previously [17]. Assay of [1- 14 C]acetyl- S -Ppant-enzyme hydrolysis Apo-PCP [17] was incubated at a concentration of 1 m M in a total volume of 1.3 mL with 20 l M [1- 14 C]acetyl-CoA (50 mCiÆmmol )1 ;ICN,Eschwege,Germany),50n M Sfp, and 10 m M MgCl 2 in assay buffer (50 m M Hepes, 100 m M NaCl, 1 m M EDTA, pH 7.3) at 37 °C. After complete modification, the reaction mixture was split and TEII srf or TEII srf mutants were added to one aliquot to a final concentration of 500 n M inthecaseofPCP.Samples (100 lL) were taken at defined time points and mixed with 800 lL trichloroacetic acid (10%, w/v) and 15 lLBSA solution (25 mgÆmL )1 ). Denatured proteins were collected by centrifugation, washed with 800 lL 10% (v/v) trichloro- acetic acid, and dissolved in 400 lL formic acid. Enzyme- bound radioactivity was analyzed by liquid scintillation counting (Tri-Carb 2100 TR, Packard, Germany). Proline quench assay Holo-ProCAT [17] was incubated at a concentration of 1 l M in a total volume of 1.5 mL with 10 m M MgCl 2 ,5 m M ATP and 4.1 l M [ 14 C]Pro (246 mCiÆmmol )1 ; Hartmann Analytics, Braunschweig, Germany) in assay buffer at 37 °C. After complete aminoacylation, 15 lL of a 100 m M unlabeled proline solution was added, and the reaction mixture was split. TEII srf or TEII srf mutants were added to one aliquot to a final concentration of 500 n M . At defined time points, 100 lL samples were removed and prepared and analyzed as described above. TNB modification of TEII srf and TEII srf mutants DTNB (4 lLofa10m M solution in dimethyl sulfoxide) was added to 46 lLofa25l M solution of TEII srf , TEII srf C18A, TEII srf S86C or TEII srf S86A and incubated for 30 min at room temperature. Subsequently, the yellow reaction mixture was applied to MicroSpin TM G-50 col- umns (Amersham Biosciences), which were pre-equilibrated with assay buffer (2 · resuspension in 400 lL assay buffer followed by centrifugation at 735 g). The colorless enzyme solutions were collected in fresh tubes by centrifugation of the columns at 735 g. Samples were desalted by application to a 30/2 mm Nucleosil C8 column (Macherey-Nagel, Du ¨ ren, Germany) using an Agilent 1100 Series HPLC system (Agilent Technologies, Waldbronn, Germany). The following gradient was applied at a flow rate of 0.1 mLÆmin )1 [buffer A: 0.1% (v/v) trifluoroacetate in water; buffer B: 0.1% (v/v) trifluoroacetate in acetonitrile]: holding buffer B constant (10%) for 4 min, followed by a linear gradient to 95% buffer B in 1 min and elution of the enzymes at 95% buffer B for 10 min. Samples were directly transferred to an electrospray ion (ESI) source connected to a Qstar Pulsar i mass spectrometer (Applied Biosystems, Darmstadt, Germany). ESI-TOF spectra (300–3000 amu) were recorded with the following parameters: curtain gas 25; nebulizer gas 35; DP1 90 V; DP2 15 V; FP 220 V; ion spray voltage 5500 V. Deconvolution was performed using the supplied ANALYST TM software. CD spectroscopy of TEII srf mutants For CD spectroscopy, TEII srf and TEII srf mutants were dialyzed against 50 m M sodium phosphate buffer (pH 7; 19.5 m M NaH 2 PO 4 and 30.5 m M Na 2 HPO 4 ) and diluted in the same buffer to a final concentration of 10 l M . Spectra were recorded on a J-810 spectrapolarimeter (Jasco, Grob-Umstadt, Germany). For this, 300 lL enzyme solution was pipetted into a cuvette of 1 mm layer thickness. The ellipticity was then measured at a constant temperature of 25 °C in the wavelength range 260–190 nm and a scan speed of 50 nmÆmin )1 (one data point per nm). The bandwidth was set to 4 nm and the response time was 4 s. Each sample was measured 10 times, and a final spectrum was calculated by the supplied Jasco software. The data were evaluated by the method of Yang [26]. Results Homology searches to identify the target residues for site-directed mutagenesis of TEII srf To identify strictly conserved residues among TEIIs involved in microbial secondary metabolism, 15 amino-acid sequences of NRPS and PKS TEIIs as well as the mammalian TEII rat were aligned. A catalytic role for Ser101, Asp236 and His237 in TEII rat has been reported [21]. The relative identity scores between TEII srf and the other TEIIs range from 20.3% (MegH [27] and TEII rat [19]) to 28.9% (TycF [28]) (Table 2). Two exceptions are LchA-TE (59.1%) [29] and LicTE (58.3%) [30], which are closely related to TEII srf . Interestingly, the overall identity 1538 U. Linne et al.(Eur. J. Biochem. 271) Ó FEBS 2004 compared with type I thioesterases (TE domains) [31] of NRPSs is only about 10% (data not shown). From the TEIIs investigated, we identified 18 absolutely invariant residues, of which six (His85, Ser86, Ser112, Asp163, Asp190 and His216) carry functionalized side chains (carboxyl, amino, amine, guanidino, thiol or hydroxy groups). Interestingly, Asp236 of TEII rat ,whichwas previously thought to be involved in catalysis in the case of the mammalian TEII associated with primary metabo- lism, is not conserved among the microbial TEIIs of secondary metabolism (Fig. 1). Mutants of these six invari- ant residues were generated, in which the functionalized residues were exchanged with nonfunctionalized residues (H85A, S86A, S112A, D163A, D190A and H216L). In addition, Ser86, thought to be the active site Ser because it is embedded in a typical core sequence found for many hydrolases (GxSxG) [32] and because of comparison with TEII rat , was replaced with Cys (S86C). In the case of active- site serines of catalytic triads, it is known that such a substitution slows down reactions, possibly allowing detec- tion of covalent reaction intermediates [22]. His216, which is the corresponding residue in TEII srf to the active-site His237 of TEII rat [23] and is located in a thioesterase core motif (GxxHxF), was also replaced by Arg (H216R). Directly adjacent to the invariant Asp190, a second Asp residue was identified (Asp189). Therefore, a single (D189A) and double (D189A/D190A) mutant were generated, ensuring that these two residues cannot replace each other. Finally Cys18, which is conserved in 14 out of the 16 TEIIs [in the other two cases (NrpT and YbtT), a Cys is found in the close neighborhood of this position], was changed to Ala (C18A), because Witkowski & Smith [33] showed inhibition of TEII rat with DTNB. DTNB is a reagent that modifies free thiol groups [34]. They postulated that this modified Cys residue is probably involved in substrate binding. Three Cys residues are found in TEII srf and four in TEII rat , whereby only the one mutated (Cys18 in TEII srf ) is conserved. Furthermore, the assay used for the kinetic characterization of TEII srf depends on DTNB [17]. The most important parts of the alignment showing all the conserved regions in which mutations were introduced are illustrated in Fig. 1. Generation, expression and purification of the TEII srf mutants We constructed a set of 10 single TEII srf mutants (C18A, H85A, S86A, S86C, S112A, D163A, D189A, D190A, H216L and H216R; Fig. 1) and one double mutant (DD189/190AA) by site-directed mutagenesis. Mutations other than to alanine or leucine were designed to show residual activity for similar functional groups (S86C and H216R). In the case of a catalytic triad (Asp-His-Ser) similar functionalized groups were expected to exhibit residual activity, whereas nonfunctionalized groups would have none. The integrity of all the mutants was confirmed by DNA sequencing, and all were individually expressed as C-terminal His 6 tag fusions in the heterologous host E. coli. Table 2. Similarities (%) between the 16 aligned TEIIs. Associated biosynthesis operons: TEII srf , surfactinA [24]; BacT, bacitracin [39]; Ery ORF5, erythromycin [40]; GrsT, gramicidin S [41]; LchA-TE, lichenysin D [29]; LicTE, lichenysin D [42]; MegH, megalomicin [27]; NrpT [43], NysE, nystatin [44]; PchC, pyochelin [45]; PikAV, pikromycin [46]; PimI, piramicin [47]; RifR, rifamycin [48]; TycF, tyrocidin [28]; YbtT, yersinabactin [49]; TEII rat , fatty acid synthase [19]. Part 1 TEII srf BacT Ery ORF5 GrsT LchA-TE LicTE MegH TEII srf 26.5 23.6 24.0 59.1 58.3 22.3 BacT 23.1 31.6 25.2 24.4 23.5 Ery ORF5 31.6 26.7 25.9 77.7 GrsT 20.8 20.7 31.6 LchA-TE 97.3 25.5 LicTE 24.7 MegH Part 2 NrpT NysE PchC PikAV PimI RifR TycF YbtT TeII rat TEII srf 23.1 24.8 26.9 24.0 28.5 25.6 28.9 27.7 22.3 BacT 23.9 25.2 27.8 29.5 27.8 31.2 34.2 23.1 27.8 Ery ORF5 21.9 38.9 39.7 36.8 41.3 43.7 25.4 29.1 24.7 GrsT 22.7 32.7 27.1 31.6 31.4 32.4 29.9 25.8 27.0 LchA-TE 21.6 26.3 28.3 27.5 28.2 26.7 28.3 23.5 21.6 LicTE 21.4 23.9 27.5 24.5 26.3 25.7 26.6 23.7 21.0 MegH 22.7 36.8 38.1 38.5 36.4 41.3 26.6 27.1 24.3 NrpT 21.9 25.5 26.5 23.5 24.1 24.6 38.9 25.3 NysE 39.4 43.8 69.3 50.6 28.7 25.9 23.1 PchC 37.1 37.8 41.8 27.9 29.5 25.5 PikAV 48.6 42.9 29.9 32.6 22.4 PimI 49.8 29.9 27.8 25.9 RifR 29.1 27.0 26.3 TycF 24.6 24.2 YbtT 22.1 TeII rat Ó FEBS 2004 Mechanistic studies on TEII srf (Eur. J. Biochem. 271) 1539 Expression and purification was carried out as described previously for wild-type TEII srf [17]. Because D163A was unstable when expressed at 30 °C under standard condi- tions and precipitated when dialyzed against 50 m M sodium phosphate buffer (the others did not; see the paragraph about CD spectroscopy of the TEII srf mutants), the expression was performed at a lower temperature (22 °C). As judged by SDS/PAGE, all His 6 -tagged recombinant proteins were purified to near-homogeneity by single-step Ni 2+ -affinity chromatography (data not shown). Determination of the catalytic activities of the mutants by the DTNB-HPLC assay For the initial activity test, the previously developed DTNB-HPLC assay was used [17]. In this assay, acetyl- S-Ppant-PCP was used as a model substrate for wild-type TEII srf , which is hydrolyzed very efficiently by the enzyme (K m ¼ 0.9 ± 0.4 l M ; k cat ¼ 95 ± 5 min )1 [17]). The products formed are HS-Ppant-PCP (holo-PCP) and acetic acid. The products were analysed by an HPLC method that requires modification of the free thiol of the HS-holo-PCP with DTNB, resulting in TNB-holo-PCP. Therefore, DTNB had to be added to the reaction mixture. TNB-holo-PCP, acetyl-holo-PCP and apo-PCP can be separated by an optimized HPLC method [17]. For the initial activity screen, the PCP substrate was used at a concentration of 25 l M , which is significantly higher than the K m of TEII srf (0.9 ± 0.4 l M [17]). Mutants C18A, H85A, S112A and D189A were still hydrolytically active, whereas S86A, D190A, DD189/190AA, H216L and H216R showed no hydrolytic activity in the 30 min reaction time. The S86C mutant was inactive in the DTNB-HPLC assay also, as shown in Fig. 2A. Further biochemical characterization of this mutant is described below. In initial assays directly after purification, mutant D163A seemed to be active. However, the results could not be reproduced with the same enzyme preparation stored for a few days at )80 °C. Furthermore, as mentioned above, it precipitated during dialysis against sodium phosphate buffer. Therefore, the enzyme was expressed at a lower temperature (22 °C), purified (standard procedure), and dialyzed against assay buffer directly after harvesting of the cells, avoiding the freezing step. The enzyme was then subjected to the DTNB-HPLC assay without storage for longer than 1 h on ice after the purification procedure was finished. The D163A mutant was then hydrolytically active towards its cognate sub- strate acetyl-S-Ppant-PCP. Interestingly, after 5–10 min of incubation at 37 °C, no further hydrolysis of the remain- ing substrate was observed, indicating clear instability of this enzyme under the assay conditions (data not shown). However, the same enzyme preparation was still active after storage of the stock solution for 24 h on ice. To determine the effect of the mutations on enzyme activity, a kinetic characterization of the active mutants was performed according to Michaelis–Menten. The kin- etic parameters obtained are summarized in Table 3 and represent the results of at least three independent meas- urements. As mentioned above, mutant S86C was not suitable for the DTNB-HPLC assay, and D163A was unstable during the assay. Therefore no kinetic parameters could be determined for these mutants. The K m values obtained for C18A (1.7 ± 0.4 l M ), S112A (0.2 ± 0.7 l M ) and D189A (0.6 ± 1.2 l M ) were in the same range as the wild-type K m (0.9 ± 0.4 l M [17]). The same was true for k cat (C18A 99 ± 4 min )1 , S112A 98 ± 10 min )1 , D189A 154 ± 18 min )1 , and wild-type 95 ± 5 min )1 ) and there- fore for k cat /K m (C18A 0.97 · 10 6 M )1 Æs )1 , S112A 8.2 · 10 6 M )1 Æs )1 , D189A 4 · 10 6 M )1 Æs )1 , and wild-type 1.75 · 10 6 M )1 Æs )1 ). Obviously, the single mutations intro- duced in these strictly conserved positions (C18A, S112A Fig. 1. Alignment of 16 TEIIs. TEII srf was aligned together with 14 other TEIIs of microbial secondary metabolism. The mammalian TEII rat was also added to the alignment. All the highly conserved regions are shown as well as the residues where mutations were introduced (marked by an arrow). 1540 U. Linne et al.(Eur. J. Biochem. 271) Ó FEBS 2004 and D189A) had only a minor effect on the catalytic efficiency of the enzyme. The K m of mutant H85A (0.5 ± 0.03 l M ) was also very close to that of the wild- type. In contrast, its k cat and therefore the catalytic efficiency was significantly lower (k cat 13.85 ± 0.09 min )1 , k cat /K m 0.46 · 10 6 M )1 Æs )1 ). Biochemical characterization of S86C Mutant S86A was inactive in the DTNB-HPLC assay (Fig. 2A). Therefore, we decided to use other assays developed in previous work. The removal of [1- 14 C]acetate or [ 14 C]proline from the corresponding acyl-S-Ppant-PCPs is detected by the decrease in radioactivity covalently attached to the protein fraction on addition of TEII srf [17]. AsshowninFig.2,intheabsenceofTEII,averyslow ÔbackgroundÕ hydrolysis (decreasing enzyme-bound radio- activity) occurred over the time scale observed. However, the enzyme-bound radioactivity decreased rapidly on addition of S86C to an assay mixture containing [1- 14 C]ace- tyl-S-Ppant-PCP (Fig. 2B), confirming the enzyme’s func- tionality. On the other hand, with [ 14 C]Pro-S-Ppant-PCP (Fig. 2C) as substrate, hydrolysis was only slightly above background rates. Although the radioactive assay is not suitable for absolute quantification of reaction rates, in the case of the [ 14 C]Pro-S-Ppant-PCP substrate it became obvious that the enzymatic activity of S86C is reduced compared with the wild-type enzyme [17]. In the case of a catalytic triad, one would expect a reaction intermediate in which the acyl group is covalently attached to the active-site Ser or Cys of the enzyme. However, all attempts to detect such an enzyme-bound intermediate with the S86C mutant by using the radioactive assays in combination with SDS/PAGE analysis followed by autoradiography of the gels failed. In no case was thioesterase-bound radioactivity observed (data not shown). Obviously, the reaction was still too fast to capture such an intermediate. Fig. 2. Biochemical characterization of TEII srf mutant S86C. (A) DTNB-HPLC assay: acetyl-4¢-S-Ppant-PCP is used as a model substrate for TEII srf [17]. The products formed are HS-4¢-S-Ppant- PCP (holo-PCP) and acetic acid. Product analysis is performed with an HPLC method, which requires modification of the free thiol of the HS- holo-PCP with DTNB, resulting in TNB-holo-PCP. Therefore, DTNB had to be added to the reaction mixture. TNB-holo-PCP, acetyl-holo- PCP and apo-PCP can be separated from each other by an optimized HPLC method [17]. S86C hydrolyses the substrate directly at the beginning of the reaction very efficiently compared with the wild-type enzyme. However, after less than 1 min, the enzyme becomes com- pletely inactivated. This inhibition of S86C is due to the covalent modification of the active-site Cys with TNB as judged by ESI-MS experiments. (B) Radioactive assay: [1- 14 C]acetyl-4¢-S-Ppant-PCP is used as substrate for TEII srf as previously described [17]. Therefore, apo-PCP was converted into [1- 14 C]acetyl-4¢-S-Ppant-PCP by the action of Sfp6 with [1- 14 C]acetyl-CoA as substrate. The assay mixture was then divided; S86C was added to one part and omitted from the other. Hydrolytic activity was observed in the absence of DTNB. The enzyme-bound radioactivity decreased rapidly on addition of S86C. Therefore, the catalytic efficiency of the mutant appeared to still be very high. (C) Proline-quench assay: a recombinant NRPS module (ProCAT) was allowed to activate and covalently load 14 C-labeled proline. The assay mixture was then split; S86C was added to one part and omitted from the other. The loss of radioactivity seems to be slightly increased in the presence of S86C. However, the hydrolytic rate is significantly decreased compared with previous results gained with wild-type TEII srf [17]. Ó FEBS 2004 Mechanistic studies on TEII srf (Eur. J. Biochem. 271) 1541 Secondary-structure validation of the inactive mutants by CD spectroscopy CD spectroscopy is an easy and fast method to determine the relative values of a-helices, b-sheets and loops within a secondary structure of a protein, although it gives no detailed structural information. To investigate the folding of the inactive or unstable mutants S86A, D163A, D190A, H216L and H216R, they were dialyzed against 50 m M sodium phosphate buffer and subjected to CD spectro- scopy. Under these conditions, D163A and H216L preci- pitated and could not be measured. This indicated that the enzyme structures became unstable as a result of the mutation of the strictly conserved Asp163 to Ala and His216 to Leu. The spectra obtained for mutants S86A, D190A and H216R looked very similar to that of wild-type TEII srf and to each other (data not shown). The computer-aided evaluation of the spectra resulted in relative numbers for the secondary-structure elements, which are presented in Table 4. These values are in the same ranges for all four enzymes, indicating no significant destruction of the enzyme structures caused by the intro- duction of the mutations. TNB modification of TEII srf and TEII srf mutants DTNB, which was used as reagent in the DTNB-HPLC assay for the determination of the kinetic parameters, reacts with free thiol groups. As TEII srf contains three Cys residues, of which Cys18 is highly conserved, and mutant S86C showed no activity in the DTNB-HPLC assay, we were interested to determine if, and how many, Cys residues will be modified by the reagent. MS methods were used to address this question. Wild-type TEII srf , C18A, S86A, and S86C were incubated at room temperature in the presence of DTNB for 30 min. For ÔquenchingÕ of the assays, the reaction mixtures were purified very quickly on small gel filtration columns. The DTNB-free enzyme fractions were subsequently applied to ESI-TOF mass analysis, and the number of TNB molecules covalently attached to the proteins was determined. The results of these modification studies are summarized in Table 5. Wild-type TEII srf and S86A mutant were both modified with one molecule of TNB. A portion of these enzymes remained unmodified, indicating a slow reaction with the reagent DTNB. In contrast, two TNB molecules were exclusively observed to be covalently attached to mutants C18A and S86C. For mutant S86C, the result explained our biochemical data. Obviously one molecule of TNB binds to the active-site Cys86 and thereby inactivates the enzyme in the DTNB-HPLC assay. However, mutant C18A showed the modification of both remaining Cys residues, while the parent enzyme, containing three cyste- ines, was modified by only one molecule of TNB in a slow reaction. This leads to the conclusion that the mutant’s structure may be changed to some extent. Therefore, the small increase in K m observed for mutant C18A could be due to this conformational change rather than to direct involvement of this residue in substrate binding. Discussion For a long time, the biochemical role of TEIIs, which are encoded by distinct genes associated with microbial NRPS and PKS operons, was a matter of speculation. Recently, however, biochemical studies have suggested a possible role Table 4. Percentage distribution of secondary-structure elements in wild-type TEII srf and the inactive mutants. Mutants D163A and H216L were insoluble under the required buffer conditions. TEII srf a-Helices b-Sheets Loops Undefined wt 58.3 38.6 0.0 3.1 S86A 53.3 38.9 0.0 7.7 D163A – – – – D190A 54.1 37.6 0.0 8.3 H216L – – – – H216R 52.0 42.5 0.0 5.5 Table 5. ESI-TOF results of TEII srf or TEII srf mutant reaction with DTNB. The calculated values are without the starting methionine, which was missing in all cases. wt, Wild-type; ND, not detected. Unmodified 1 molecule TNB 2 molecules TNB Calculated Measured Calculated Measured Calculated Measured wt TEII srf 28425.4 28424.7 28622.6 28623.1 28819.8 ND C18A 28393.3 ND 28590.5 ND 28787.7 28788.1 S86A 28409.4 28408.8 28606.6 28607.7 28803.8 ND S86C 28441.5 ND 28634.6 ND 28835.8 28834.1 Table 3. Summary of TEII srf and TEII srf mutant activities. TEII srf K m (l M ) k cat (min )1 ) k cat /K m ( M )1 Æs )1 ) Wild-type 0.9 ± 0.4 95 ± 5 1.75 · 10 6 C18A 1.7 ± 0.4 99 ± 4 0.97 · 10 6 H85A 0.5 ± 0.03 13.85 ± 0.09 0.46 · 10 6 S86C Active, but not suitable for HPLC-DTNB assay S86A Inactive S112A 0.2 ± 0.7 98 ± 10 8.2 · 10 6 D163A Active, but unstable during the HPLC-DTNB assay D189A 0.6 ± 1.2 154 ± 18 4.00 · 10 6 D190A Inactive DD189/190AA Inactive H216L Inactive H216R Inactive 1542 U. Linne et al.(Eur. J. Biochem. 271) Ó FEBS 2004 for these TEIIs in the regeneration of mis-acylated NRPSs and PKSs. They are obviously involved in removing short acyl-S-Ppant intermediates from acyl carrier proteins and PCPs associated with secondary metabolite biosynthesis [15–17]. Enabled by the discovery of the natural substrate of microbial secondary metabolism TEIIs [17], we set out to determine the mechanistic properties of these enzymes [15– 17]. Therefore, the TEII srf associated with surfactin biosyn- thesis in Bacillus subtilis was used as a prototype TEII for our mechanistic studies. Because of the surprisingly high identities (> 20%) between these TEII enzymes of micro- bial secondary metabolism and a mammalian TEII (TEII rat [19]), which has been intensively studied [19,21–23,33,35,36], the presence of a catalytic triad in microbial TEIIs was postulated [24]. In TEII rat , Ser101, Asp236 and His237 were reported to be involved in catalysis [21–23], although the D236A mutant showed a residual activity of 40% [21]. There are two possible reasons for this: (a) Asp236 of TEII rat is not part of the catalytic triad of this class of hydrolases; (b) a catalytic diad (Ser-His) is sufficient for catalytic activity. As our sequence alignment revealed that Asp236 of TEII rat is not conserved in microbial TEIIs of secondary metabolism, it was more likely that one residue of the proposed catalytic triad had not been identified so far. In agreement with the results for the mammalian TEII, we confirmed that the corresponding residues Ser86 and His216 are part of the catalytic triad in TEII srf of microbial secondary metabolism. Mutants S86A and H216L showed no hydrolytic activity. In addition, the S86C mutant was inhibited by DTNB, and active in its absence. As judged by ESI-TOF high-resolution MS, this inhibition was due to covalent modification of the active-site Cys with one molecule of TNB. Interestingly, H216R was also completely inactive in the DTNB-HPLC assay. Therefore it seems that Arg cannot functionally replace His216 in TEII srf .However, inthecaseofTEII rat H237R, the residual activity reported was reduced more than threefold compared with the parent enzyme, which was only slightly above the detection limit [23]. The second strictly conserved His residue found in TEII srf (His85) is located directly next to the active-site Ser86. The observed reduction in catalytic efficiency may be caused by repositioning of the Ser86 in mutant H85A resulting from the replacement of His with Ala rather than by direct involvement in catalysis. To determine the identy of the remaining residue of the proposed catalytic triad, the two strictly conserved Asp residues among all TEIIs aligned (Asp163 and Asp190 in TEII srf ) were separately exchanged with Ala (D163A and D190A mutants). Our results clearly indicate that Asp190 (Asp212 in TEII rat ) is the missing member of the catalytic triad, whereas Asp163 (Asp183 in TEII rat ) seems to be structurally important, as evidenced by the precipitation of the enzyme when dialyzed against phosphate buffer and the observed instability when assayed at 37 °C for several minutes or after storage at )80 °C. In contrast, wild-type TEII srf and the other mutants studied showed no instability under these conditions. Many such hydrolases that have a catalytic triad belong to the large class of a/b-hydrolases. They show a conserved characteristic fold, which was first described by Ollis et al. [37]. This fold was also recently reported for the thio- esterases of type I, which are located as internal domains at the C-termini of the termination modules of NRPS- biosynthetic and PKS-biosynthetic enzymes [31,38]. For the latter type I thioesterases, a catalytic triad was biochemically confirmed in the case of the TE domain of surfactin synthetase C [11]. The canonical a/b-hydrolase fold, which is illustrated in Fig. 3, consists of eight b-strands (1–8), which are positioned in plane. Above them are positioned two (A and F) and under them four (B, C, D, and E) a-helices. The nucleophile of the catalytic triad (Ser86 in TEII srf )isalways located at the Ônucleophile elbowÕ in a G-x-Nu-x-G core sequence (x, any amino acid; Nu, nucleophile). The nucleophile elbow is a loop directly after b-strand 5. The acidic residue of the triad, an Asp residue, is located on a loop following b-strand seven (Asp190 in TEII srf ). Finally, the catalytic triad is completed by the His residue, which is located on a longer loop between b-strand 8 and a-helix F (His216 in TEII srf ). Based on the relative positioning and distances between the three residues forming the catalytic triadinTEII srf (Ser86, Asp190, and His216) as well as the existence of the core motif ÔGHSxGÕ, which is always found in a/b-hydrolases (G-x-Nu-x-G, see above), there is strong evidence that the microbial TEIIs of secondary metabolism as well as the TEII rat belong to this large class of enzymes, too. However, no structural data are available on them. In summary, we have clearly identified a catalytic triad in the prototype TEII srf consisting of Ser86, His216 and Asp190. Moreover, because of the remarkably high simi- larities of the microbial TEIIs of secondary metabolism to Fig. 3. Schematic representation of the canonical a/b-hydrolase fold [50]. (A) Three-dimensional structure. (B) Two-dimensional repre- sentation. Ó FEBS 2004 Mechanistic studies on TEII srf (Eur. J. Biochem. 271) 1543 the mammalian TEII rat , our results strongly suggest that Asp212 is the acidic residue of the proposed catalytic triad in TEII rat . With this knowledge, the reduction in catalytic efficiency of the TEII rat mutant D236A, which was observed by Tai et al. [21] is probably more likely to be due to the mutation directly adjacent to the catalytic His237 than to a direct involvement in catalysis. Furthermore, the relative positioning of the residues of the catalytic triad in TEII srf and TEII rat provides evidence that this class of enzymes belongs to the large family of a/b-hydrolases. Acknowledgements We thank Antje Scha ¨ fer for excellent technical assistance and protein purification. The CD spectroscopy was carried out in the Laboratory of Professor T. Carell with the assistance of Alexandra Mees. We also thank Mohammad R. Mofid for providing CD spectroscopy data on wild-type TEII srf . This work was funded by the Deutsche Forschung- sgemeinschaft and the Fonds der chemischen Industrie. References 1. Kordic, S., Cummins, I. & Edwards, R. (2002) Cloning and characterization of an S-formylglutathione hydrolase from Ara- bidopsis thaliana. Arch. Biochem. Biophys. 399, 232–238. 2. Das, A.K., Lu, J.Y. & Hofmann, S.L. (2001) Biochemical analysis of mutations in palmitoyl-protein thioesterase causing infantile and late-onset forms of neuronal ceroid lipofuscinosis. Hum. Mol. Genet. 10, 1431–1439. 3. Schwarzer, D. & Marahiel, M.A. (2001) Multimodular bio- catalysts for natural product assembly. Naturwissenschaften 88, 93–101. 4. Cane, D.E. & Walsh, C.T. (1999) The parallel and convergent universes of polyketide synthases and nonribosomal peptide synthetases. Chem. Biol. 6, R319–R325. 5. Stein, T., Vater, J., Kruft, V., Otto, A., Wittmann-Liebold, B., Franke, P., Panico, M., McDowell, R. & Morris, H.R. (1996) The multiple carrier model of nonribosomal peptide biosynthesis at modular multienzymatic templates. J. Biol. Chem. 271, 15428– 15435. 6.Reuter,K.,Mofid,M.R.,Marahiel,M.A.&Ficner,R.(1999) Crystal structure of the surfactin synthetase-activating enzyme Sfp: a prototype of the 4¢-phosphopantetheinyl-transferase superfamily. EMBO J. 18, 6823–6831. 7. Finking, R., Solsbacher, J., Konz, D., Schobert, M., Schafer, A., Jahn,D.&Marahiel,M.A.(2002)Characterizationofanewtype of phosphopantetheinyl transferase for fatty acid and siderophore synthesis in Pseudomonas aeruginosa. J. Biol. Chem. 277, 50293– 50302. 8. Mootz,H.D.,Finking,R.&Marahiel,M.A.(2001)4¢-Phospho- pantetheine transfer in primary and secondary metabolism of Bacillus subtilis. J. Biol. Chem. 276, 37289–37298. 9. Trauger,J.,Kohli,R.,Mootz,H.,Marahiel,M.&Walsh,C. (2000) Peptide cyclization catalysed by the thioesterase domain of tyrocidine synthetase. Nature (London) 407, 215–218. 10. Kohli, R.M., Trauger, J.W., Schwarzer, D., Marahiel, M.A. & Walsh, C.T. (2001) Generality of peptide cyclization catalyzed by isolated thioesterase domains of nonribosomal peptide synthe- tases. Biochemistry 40, 7099–7108. 11. Tseng, C.C., Bruner, S.D., Kohli, R.M., Marahiel, M.A., Walsh, C.T. & Sieber, S.A. (2002) Characterization of the surfactin syn- thetase C-terminal thioesterase domain as a cyclic depsipeptide synthase. Biochemistry 41, 13350–13359. 12. Schneider, A. & Marahiel, M.A. (1998) Genetic evidence for a role of thioesterase domains, integrated in or associated with peptide synthetases, in non-ribosomal peptide biosynthesis in Bacillus subtilis. Arch. Microbiol. 169, 404–410. 13. Geoffroy, V.A., Fetherston, J.D. & Perry, R.D. (2000) Yersinia pestis YbtU and YbtT are involved in synthesis of the siderophore yersiniabactin but have different effects on regulation. Infect Immun. 68, 4452–4461. 14. Butler, A.R., Bate, N. & Cundliffe, E. (1999) Impact of thioes- terase activity on tylosin biosynthesis in Streptomyces fradiae. Chem. Biol. 6, 287–292. 15. Heathcote, M.L., Staunton, J. & Leadlay, P.F. (2001) Role of type II thioesterases: evidence for removal of short acyl chains pro- duced by aberrant decarboxylation of chain extender units. Chem. Biol. 8, 207–220. 16. Kim, B.S., Cropp, T.A., Beck, B.J., Sherman, D.H. & Reynolds, K.A. (2002) Biochemical evidence for an editing role of thioes- terase II in the biosynthesis of the polyketide pikromycin. J. Biol. Chem. 277, 48028–48034. 17. Schwarzer, D., Mootz, H.D., Linne, U. & Marahiel, M.A. (2002) Regeneration of misprimed nonribosomal peptide synthetases by type II thioesterases. Proc.NatlAcad.Sci.USA99, 14083– 14088. 18. Vallari, D.S., Jackowski, S. & Rock, C.O. (1987) Regulation of pantothenate kinase by coenzyme A and its thioesters. J. Biol. Chem. 262, 2468–2471. 19. Randhawa, Z.I. & Smith, S. (1987) Complete amino acid sequence of the medium-chain S-acyl fatty acid synthetase thioester hydrolase from rat mammary gland. Biochemistry 26, 1365–1373. 20. Naggert, J., Williams, B., Cashman, D.P. & Smith, S. (1987) Cloning and sequencing of the medium-chain S-acyl fatty acid synthetase thioester hydrolase cDNA from rat mammary gland. Biochem. J. 243, 597–601. 21. Tai, M.H., Chirala, S.S. & Wakil, S.J. (1993) Roles of Ser101, Asp236, and His237 in catalysis of thioesterase II and of the C-terminal region of the enzyme in its interaction with fatty acid synthase. Proc.NatlAcad.Sci.USA90, 1852–1856. 22. Witkowski, A., Naggert, J., Witkowska, H.E., Randhawa, Z.I. & Smith, S. (1992) Utilization of an active serine 101–cysteine mutant to demonstrate the proximity of the catalytic serine 101 and histidine 237 residues in thioesterase II. J. Biol. Chem. 267, 18488–18492. 23. Witkowski, A., Naggert, J., Wessa, B. & Smith, S. (1991) A cata- lytic role for histidine 237 in rat mammary gland thioesterase II. J. Biol. Chem. 266, 18514–18519. 24. Cosmina, P., Rodriguez, F., de Ferra, F., Grandi, G., Perego, M., Venema, G. & van Sinderen, D. (1993) Sequence and analysis of the genetic locus responsible for surfactin synthesis in Bacillus subtilis. Mol Microbiol. 8, 821–831. 25. Laemmli, U.K. (1970) Cleavage of structural proteins during the assembly of the head of the bacteriophage T4. Nature (London) 227, 491–493. 26. Yang, C.T. (1986) Calculation of protein conformation from cir- cular dichroism. Methods Enzymol. 130, 208–269. 27. Volchegursky,Y.,Hu,Z.,Katz,L.&McDaniel,R.(2000)Bio- synthesis of the anti-parasitic agent megalomicin: transformation of erythromycin to megalomicin in Saccharopolyspora erythraea. Mol. Microbiol. 37, 752–762. 28. Mootz, H.D. & Marahiel, M.A. (1997) The tyrocidine biosynth- esis operon of Bacillus brevis: complete nucleotide sequence and biochemical characterization of functional internal adenylation domains. J. Bacteriol. 179, 6843–6850. 29. Konz,D.,Doekel,S.&Marahiel,M.A.(1999)Molecularand biochemical characterization of the protein template controll- ing biosynthesis of the lipopeptide lichenysin. J. Bacteriol. 181, 133–140. 30. Yakimov, M.M., Kroger, A., Slepak, T.N., Giuliano, L., Timmis, K.N. & Golyshin, P.N. (1998) A putative lichenysin A synthetase 1544 U. Linne et al.(Eur. J. Biochem. 271) Ó FEBS 2004 operon in Bacillus licheniformis: initial characterization. Biochim. Biophys. Acta 1399, 141–153. 31. Bruner, S.D., Weber, T., Kohli, R.M., Schwarzer, D., Marahiel, M.A., Walsh, C.T. & Stubbs, M.T. (2002) Structural basis for the cyclization of the lipopeptide antibiotic surfactin by the thio- esterase domain SrfTE. Structure (Camb). 10, 301–310. 32. Dodson, G. & Wlodawer, A. (1998) Catalytic triads and their relatives. Trends Biochem. Sci. 23, 347–352. 33. Witkowski, A. & Smith, S. (1985) Inhibition of the functional interactionbetweenfattyacidsynthetaseandthioesteraseIIby modification of a single cysteine thiol on the thioesterase. Arch. Biochem. Biophys. 243, 420–426. 34. Riddles, P.W., Blakeley, R.L. & Zerner, B. (1983) Reassessment of Ellman’s reagent. Methods Enzymol. 91, 49–60. 35.Buchbinder,J.L.,Witkowski,A.,Smith,S.&Fletterick,R.J. (1995) Crystallization and preliminary diffraction studies of thioesterase II from rat mammary gland. Proteins 22, 73–75. 36. Witkowski, A., Witkowska, H.E. & Smith, S. (1994) Reengineering the specificity of a serine active-site enzyme. Two active-site mutations convert a hydrolase to a transferase. J. Biol. Chem. 269, 379–383. 37. Ollis, D.L., Cheah, E., Cygler, M., Dijkstra, B., Frolow, F., Franken, S.M., Harel, M., Remington, S.J., Silman, I., Schrag, J. et al. (1992) The alpha/beta hydrolase fold. Protein Eng. 5, 197–211. 38. Tsai, S.C., Miercke, L.J., Krucinski, J., Gokhale, R., Chen, J.C., Foster, P.G., Cane, D.E., Khosla, C. & Stroud, R.M. (2001) Crystal structure of the macrocycle-forming thioesterase domain of the erythromycin polyketide synthase: versatility from a unique substrate channel. Proc.NatlAcad.Sci.USA98, 14808–14813. 39. Konz, D., Klens, A., Schorgendorfer, K. & Marahiel, M.A. (1997) The bacitracin biosynthesis operon of Bacillus licheniformis ATCC 10716: molecular characterization of three multi-modular peptide synthetases. Chem. Biol. 4, 927–937. 40. Weber,J.M.,Leung,J.O.,Maine,G.T.,Potenz,R.H.,Paulus,T.J. & DeWitt, J.P. (1990) Organization of a cluster of erythromycin genes in Saccharopolyspora erythraea. J. Bacteriol. 172, 2372– 2383. 41. Kratzschmar, J., Krause, M. & Marahiel, M.A. (1989) Gramicidin S biosynthesis operon containing the structural genes grsA and grsB has an open reading frame encoding a protein homologous to fatty acid thioesterases. J. Bacteriol. 171, 5422–5429. 42. Yakimov, M.M., Kroger, A., Slepak, T.N., Giuliano, L., Timmis, K.N. & Golyshin, P.N. (1998) A putative lichenysin A synthetase operon in Bacillus licheniformis: initial characterization. Biochim. Biophys. Acta 1399, 141–153. 43. Gaisser, S. & Hughes, C. (1996) Proteus mirabilis NrpS (nrpS) gene, partial cds, NrpU (nrpU), NrpT (nrpT), NrpA (nrpA), NrpB (nrpB), NrpG (nrpG) and IrpP (irpP) genes, complete cds. NCBI Nucleotide Databank, accession number U46488. http:// www.ncbi.nlm.nih.gov. 44. Brautaset, T., Sekurova, O.N., Sletta, H., Ellingsen, T.E., StrLm, A.R., Valla, S. & Zotchev, S.B. (2000) Biosynthesis of the polyene antifungal antibiotic nystatin in Streptomyces noursei ATCC 11455: analysis of the gene cluster and deduction of the biosyn- thetic pathway. Chem. Biol. 7, 395–403. 45. Serino, L., Reimmann, C., Visca, P., Beyeler, M., Chiesa, V.D. & Haas, D. (1997) Biosynthesis of pyochelin and dihydroaeruginoic acid requires the iron- regulated pchDCBA operon in Pseudo- monas aeruginosa. J. Bacteriol. 179, 248–257. 46. Xue,Y.,Zhao,L.,Liu,H.W.&Sherman,D.H.(1998)Agene cluster for macrolide antibiotic biosynthesis in Streptomyces venezuelae: architecture of metabolic diversity. Proc.NatlAcad. Sci. USA 95, 12111–12116. 47. Aparicio, J.F., Fouces, R., Mendes, M.V., Olivera, N. & Martin, J.F. (2000) A complex multienzyme system encoded by five polyketide synthase genes is involved in the biosynthesis of the 26-membered polyene macrolide pimaricin in Streptomyces nata- lensis. Chem. Biol. 7, 895–905. 48. August, P.R., Tang, L. et al. e. (1998) Biosynthesis of the ansa- mycin antibiotic rifamycin: deductions from the molecular ana- lysis of the rif biosynthetic gene cluster of Amycolatopsis mediterranei S699. Chem. Biol. 5, 69–79. 49. Bearden, S.W., Fetherston, J.D. & Perry, R.D. (1997) Genetic organization of the yersiniabactin biosynthetic region and con- struction of avirulent mutants in Yersinia pestis. Infect. Immun. 65, 1659–1668. 50. Nardini, M. & Dijkstra, B.W. (1999) Alpha/beta hydrolase fold enzymes: the family keeps growing. Curr. Opin. Struct. Biol. 9, 732–737. Ó FEBS 2004 Mechanistic studies on TEII srf (Eur. J. Biochem. 271) 1545 . AGA AGG C TEII srf D18 9A 5¢ C TTT AAC GGG CTT GCT GAT AAA AAA TGC ATA CGA GAT GCG G 3¢ C CGC ATC TCG TAT GCA TTT TTT ATC AGC AAG CCC GTT AAA G TEII srf D19 0A 5¢ C TTT AAC GGG CTT GAT GCT AAA AAA. AAA TGC ATA CGA GAT GCG G 3¢ C CGC ATC TCG TAT GCA TTT TTT AGC ATC AAG CCC GTT AAA G TEII srf 5¢ C TTT AAC GGG CTT GCT GCT AAA AAA TGC ATA CGA GAT GCG G DD189/190AA 3¢ C CGC ATC TCG TAT GCA TTT. GGC AAA CGG AAA AGC GAT GAG CTG TG TEII srf H8 5A 5¢ GTG CTG TTC GGA GCC AGT ATG GGC GGA ATG ATC AC 3¢ GT GAT CAT TCC GCC CAT ACT GGC TCC GAA CAG CAC TEII srf S8 6A 5¢ G CTG TTC GGA CAC GCT ATG