Báo cáo Y học: Engineering and mechanistic studies of the Arabidopsis FAE1 b-ketoacyl-CoA synthase, FAE1 KCS pot

Engineering and mechanistic studies of the Arabidopsis FAE1 b-ketoacyl-CoA synthase, FAE1 KCS Mahin Ghanevati and Jan G. Jaworski Department of Chemistry and Biochemistry, Miami University, Oxford, OH, USA The Arabidopsis FAE1 b-ketoacyl-CoA synthase (FAE1 KCS) catalyzes the condensation of malonyl-CoA with long- chain acyl-CoAs. Sequence analysis of FAE1 KCS predicted that this condensing enzyme is anchored to a membrane by two adjacent N-terminal membrane-spanning domains. In order to characterize the FAE1 KCS and analyze its mech- anism, FAE1 KCS and its mutants were engineered with a His 6 -tag at their N-terminus, and expressed in Saccharomyces cerevisiae. The membrane-bound enzyme was then solubi- lized and purified to near homogeneity on a metal affinity column. Wild-type recombinant FAE1 KCS was active with several acyl-CoA substrates, with highest activity towards saturated and monounsaturated C16 and C18. In the absence of an acyl-CoA substrate, FAE1 KCS was unable to carry out decarboxylation of [3– 14 C]malonyl-CoA, indica- ting that it requires binding of the acyl-CoA for decarb- oxylation activity. Site-directed mutagenesis was carried out on the FAE1 KCS to assess if this condensing enzyme was mechanistically related to the well characterized soluble condensing enzymes of fatty acid and flavonoid syntheses. A C223A mutant enzyme lacking the acylation site was unable to carry out decarboxylation of malonyl-CoA even when 18:1-CoA was present. Mutational analyses of the conserved Asn424 and His391 residues indicated the importance of these residues for FAE1-KCS activity. The results presented here provide the initial analysis of the reaction mechanism for a membrane-bound condensing enzyme from any source and provide evidence for a mechanism similar to the soluble condensing enzymes. Keywords: very long chain fatty acids; fatty acid elongation; condensation mechanism. Fatty acids with greater than 18 carbon atoms (very long chain fatty acids, VLCFA) are precursors of many biolo- gically important compounds such as sphingolipids [1,2], waxes [3], and triacylglycerols in many seed oils [4]. Biosynthesis of VLCFA in plants and animals, is dependent on the activity of a membrane-bound fatty acid elongation system which consists of four component reactions similar to fatty acid synthase. The first reaction of elongation involves condensation of malonyl-CoA with a long chain acyl substrate producing a b-ketoacyl-CoA. Subsequent reactions are reduction to b-hydroxyacyl-CoA, dehydration to an enoyl-CoA, followed by a second reduction to form the elongated acyl-CoA [5]. In both animals and plants, the initial condensation reaction is believed to be the rate- limiting step [6,7]. In Arabidopsis, FAE1 codes for a b-ketoacyl-CoA synthase (FAE1 KCS) which is expressed exclusively in the seed and catalyzes the initial condensation step in the elongation pathway [8]. Based on fatty acid profiles of transgenic plants and yeast, it has been reported that FAE1 KCS has a substrate preference for C18:1, producing eicosenoic (C20:1) acid as the major product and erucic acid (C22:1) as a minor product [7]. A prominent feature of b-ketoacyl-CoA synthases involved in VLCFA biosynthesis is their membrane-bound nature. This makes them different from all other condensing enzymes studied to date, which are soluble enzymes. These include, for example, those involved in fatty acid and polyketide synthesis. Amino-acid sequence analysis of the Arabidopsis KCS1 [9] indicated that elongase KCS enzymes, including FAE1 KCS, have two transmembrane-spanning domains close to their N-terminus, thus suggesting that these enzymes are anchored to the membrane. Although fatty acid elongases and their b-ketoacyl-CoA synthase component have been partially purified from a number of sources [10–13] and studied using cellular fractions [14,15], the information about KCS enzymes and their kinetic properties is very limited. This is mainly due to the complexity of the membrane fractions used as the enzyme source and the presence of a high level of background activities. Currently, there is only one report of an extensively purified membrane-bound KCS, jojoba KCS [13], and its characterization was limited to the substrate specificity. Despite their membrane-bound nature, some domains of the elongase condensing enzymes have limited homology to two soluble condensing enzymes: plant chalcone synthases [16] and 3-ketoacyl-ACP synthase III (KAS III) from plants [17] and Escherichia coli [18]. The reaction mechanism of the soluble condensing enzymes has been extensively studied, and recently the crystal structures of chalcone synthase [19] and all isoforms of KASs from E. coli [20–23] have been published. Site-directed mutagenesis studies, as well as crystal structures, indicate that these soluble condensing enzymes all utilize the same general reaction mechanism (Fig. 1). This involves, successively, transfer of the acyl Correspondence to J. G. Jaworski, Donald Danforth Plant Science Center, 975 North Warson Road, St Louis, MO 63132, USA. Fax: + 314 587 1721, Tel.: + 314 587 1621, E-mail: jjaworski@danforthcenter.org Abbreviations: VLCFA, very long chain fatty acid; FAE1 KCS, fatty acid elongase 1 b-ketoacyl-CoA synthase; FAS, fatty acid synthase. (Received 20 February 2002, revised 9 May 2002, accepted 10 June 2002) Eur. J. Biochem. 269, 3531–3539 (2002) Ó FEBS 2002 doi:10.1046/j.1432-1033.2002.03039.x primer substrate to an active-site cysteine forming an acyl thioester intermediate, decarboxylation of the donor malo- nyl substrate to yield an acetyl carbanion intermediate, and finally, nucleophilic attack of the carbanion on the carbonyl carbon atom of the thioester intermediate, resulting in the formation of the product. The FAE1 KCS mechanism has not been characterized and is known to use malonyl-CoA instead of malonyl-ACP. Nonetheless, the mechanism of the fatty acid synthase condensing enzymes should serve as an appropriate model for the FAE1 KCS. In addition to an active-site cysteine, at least one histidine residue is directly involved in catalysis by soluble condensing enzymes. Crystal structures of both KAS I [21] and KAS II [20] from E. coli reveal the presence of two histidines in close proximity to their active site cysteine of which at least one is assumed to be important for enzyme catalysis. In addition, crystal structures of E. coli KAS III [22,23] and alfalfa chalcone synthase [19] show a histidine and an asparagine residue in the active site architecture. The role of these residues in catalysis were subsequently confirmed in both KAS III [23] and chalcone synthase [24] by in vitro mutagenesis and were shown to be important catalytic residues in the decarboxylation of the malonyl substrate. Based on its limited sequence similarity to resveratrol synthase, a closely related condensing enzyme to chalcone synthase, Lassner et al. [13] have tentatively identified the active-site cysteine of the jojoba elongase KCS. The corresponding cysteine of FAE1 KCS is Cys223. Recently, we confirmed this hypothesis by site-directed mutagenesis of FAE1 KCS [25]. Furthermore, of all conserved histidine and asparagine residues of FAE1 KCS only His391 and N424 align with the active-site histidine and asparagine of both KAS III and chalcone synthase-related enzymes [25]. It is likely that these residues play similar role to those involved in chalcone synthase and KAS III. In order to characterize the mechanism of an elongase KCS, an expression system that allowed facile purification of enzyme was required. We report here one approach to the expression and purification of FAE1 KCS. We engineered a His 6 -tag at N-terminus of FAE1, expressed it in yeast and isolated the recombinant protein from yeast microsomal pellet using a metal affinity column. Partially purified recombinant FAE1 KCS was assayed for conden- sation, decarboxylation, and substrate specificity. Addition- ally, this provided an opportunity to analyze the effect of mutagenesis of His391 and Asn424 in FAE1 KCS. To our knowledge, this is the first report of the analysis of the mechanism of a membrane-bound condensing enzyme from any source. EXPERIMENTAL PROCEDURES Materials S. cerevisiae (InvSc1) and pYES2 plasmid were purchased from Invitrogen (Carlsbad, CA, USA). pBluescript was from Stratagene (La Jolla, CA, USA). Oligonucleotide primers were purchased from Integrated DNA Technol- ogies (Coralville, IA, USA). Vent DNA polymerase, nucleotide triphosphates, and restriction endonucleases were purchased from New England Biolabs (Beverly, MA, USA). T4 DNA ligase was from Gibco BRL (Grand Island, NY, USA). Ni 2+ -PDC was purchased from Affiland (Affinity Methodology in Biotechnology, Belgium). Throm- bin, acyl-CoAs, and galactose were purchased from Sigma. All other chemicals for media culture were obtained from Fisher. [3– 14 C]malonyl-CoA was prepared as described by Roughan [26]. Engineering of FAE1 KCS and its mutants The overlap extension method as described by Ho et al.[27] was used to introduce a thrombin cleavage site into the Arabidopsis FAE1 KCS. Four constructs containing the thrombin cleavage sites near the membrane-spanning domain were prepared. LVPRGS was inserted at residues 74, 101, 115, and in one construct, at residue 106, LVPRGS was substituted for RKADTS. This method requires two gene-flanking primers and two internal overlapping oligo- nucleotides containing the sequence encoding the thrombin cleavage site. As our starting template was FAE1 gene subcloned in pYEUra-3 (Clontech, Palo Alto, CA, USA), flanking primers used were an antisense 5¢-CGTCAAG GAGAAAAAACCTCTAGCCGAAT-3¢ primer and an universal T7 sense primer. To insert a thrombin cleavage site at amino-acid residue 115 in FAE1 KCS (T115-FAE1 KCS) a sense primer 5¢-GAACGTG TTGGTTCCGC GTGGTAGCGCATGTGATGATCCGTCCTCG-3¢ and an antisense primer 5¢-CACATGC GCTACCACGCGG AACCAACACGTTCCGTGAAGAAGTATC-3¢ were used (underlined sequence encodes for thrombin cleavage site). This led to a 32-bp overlap in the second round of PCR amplification. A similar approach was taken to make the three other thrombin cleavage site constructs. Each of these constructs was subcloned in the yeast expression vector, pYES2, in which expression is under control of the GAL1 promoter. The constructs were sequenced to verify the presence of thrombin cleavage sequence. To generate FAE1 KCS construct with a N-terminus His 6 -tag the sense primer 5¢-CGCGGATCCGCGATG (CAT) 6 ACTTCCGTTAACGTTAAGCTCCTTTAC-3¢ and the antisense primer 5¢-CGCGGATCCGCGTTAG Fig. 1. Scheme of fatty acid synthase condensation reaction. The com- mon reaction scheme of fatty acid synthase b-ketoacyl-ACP synthases (KAS) involves: (1) acylation of an active-site cysteine; (2) binding of malonyl-ACP followed by decarboxylation; and (3) attack on the acyl group by the carbanion, producing a b ketoacyl-ACP. 3532 M. Ghanevati and J. G. Jaworski (Eur. J. Biochem. 269) Ó FEBS 2002 GACCGACCGTTTTGGACATGAGTCTT-3¢ were used. To facilitate subcloning both primers were designed with a terminal BamHI restriction site. A similar approach was usedtopreparetheC-terminusHis 6 tag FAE1 KCS. Previously prepared mutants of FAE1 KCS, C223A, H391A, and H391K [25] were used as templates to generate the His 6 -tagged recombinant mutants. To generate H391Q, N424D, and N424H mutants, a set of overlapping muta- genic oligonucleotides along with the flanking primers listed above were used according to method by Ho et al. [27]. The sense and antisense mutagenic primers were as follows (the mutagenized codons are underlined): H391Q, sense 5¢-ATTTCTGTATTC AAGCTGGAGGCAGAGCCGTG AT-3¢,antisense5¢-CTGCCTCCAGC TTGAATACAG AAATG-3¢; N424D, sense 5¢-AGATTTGGG GATAC TTCATCTAGCTCAATTT-3¢,antisense5¢-AGATGAAGT ATCCCCAAATCTATGTAACG-3¢; N424H, sense 5¢-AGATTTGGG CATACTTCATCTAGCTCA-3¢,anti- sense 5¢-AGATGAAGT ATGCCCAAATCTATGTAA CG-3¢. PCR reactions were carried out with Vent DNA polymerase, and the amplified PCR products were sub- cloned into the yeast expression vector, pYES2. These constructs were sequenced to confirm the presence of mutations and that no errors were introduced during the PCR amplification or subcloning. Expression and microsomal preparation S. cerevisiae strain InvSc1 (Invitrogen) was transformed with the pYES2 vector or the pYES2 constructs described above using a lithium acetate procedure [28]. The trans- formants were selected on synthetic complete media lacking uracil (Cm-ura). Transformed yeast cells were grown overnight in YPDA at 30 °C. The overnight cultures were used to inoculate Cm-ura culture supplemented with 2% galactose to give an initial D 600 ¼ 0.02, and the cultures were grown to D 600 ¼ of 1.5. Yeast microsomes were prepared as previously described [25]. The microsomal pellet was resuspended in ice-cold IB (80 m M Hepes/KOH, pH 7.2, 5 m M EGTA, 5 m M EDTA, 10 m M KCl, 320 m M sucrose, 2 m M dithiothreitol) contain- ing 20% glycerol to give a final protein concentration of 2.5 mgÆmL )1 . Protein concentrations were determined according to the Bradford method using bovine serum albumin as standard [29]. Solubilization and purification of recombinant His 6 -FAE1 KCS Microsomal proteins were solubilized in the presence of 0.32 M NaCl and 0.5% Triton X-100 and a final protein concentration of 2 mgÆmL )1 . This yielded a detergent to protein ratio of 2.5 : 1 (w/w), which was the optimal ratio for solubilization of the FAE1 KCS protein. After incuba- tion on ice for 2 h, the samples were centrifuged at 100 000 g for 60 min, and the supernatant fractions were collected. The supernatant fractions were diluted three fold with buffer A (50 m M sodium phosphate buffer, pH 8.0, 0.5% Triton X-100, 0.15 M NaCl, 10% glycerol). A sample was loaded onto a 200-lLNi 2+ -PDC column that had been equilibrated with buffer A. The column was then washed with 1.0 mL of buffer A followed by 1.0 mL of buffer B (50 m M sodium phosphate buffer, pH 8.0, 0.5% Triton X-100, 0.5 M NaCl,10%glycerol,20m M imidazole), and finally with 1.0 mL of buffer A. His 6 -FAE1 KCS and its mutantswerethenelutedwith300lL of buffer A containing 300 m M imidazole, and dithiothreitol was added to a final concentration of 2 m M . The isolated recombinant FAE1 KCS and its mutants were stored at )80 °Cand remained stable. Immunoblot analysis and silver staining To a protein sample, trichloroacetic acid was added to a final concentration of 10% (w/v). The sample was frozen at )80 °C for 10 min, thawed and centrifuged, and the pellet washed twice with 1% trichloroacetic acid followed by one wash with 80% acetone. Precipitated protein was then resuspended in sample buffer, and the sample run on a 10% SDS/PAGE gel [30]. For Western blot analysis, proteins were transferred to poly(vinylidene difluoride) membrane by semidry transfer [31]. Western blot analysis was performed according to standard protocols [32], and the protein bands were detected using rabbit anti-(FAE1 KCS) Ig (a gift from L. Kunst, University of British Columbia, Canada) followed by alkaline phosphatase-conjugated goat antirabbit IgG and color development. Silver staining of the SDS/PAGE was carried out according to method by Hochstrasser et al.[33]. Enzyme assays FAE1 KCS condensation activity was routinely deter- mined by the method of Garwin et al.[34].Theassay contained 40 m M sodium phosphate buffer, pH 7.2, 15 l M 18 : 1-CoA, 20 l M [1- 14 C]malonyl-CoA (35.7 lCiÆlmol )1 ), and FAE1 KCS in a 25-lL reaction volume at 30 °C. Reactions were stopped by addition of 0.5 mL of 0.1 M K 2 HPO 4 ,0.4 M KCl, 30% tetrahydrofuran and 5mgÆmL )1 NaBH 4 ,heatedat37°C for 30 min, and extracted twice with 0.8 mL of petroleum ether. The extract was dried under N 2 gas, and 14 C product was quantified by liquid scintillation counting. The decarboxylation activity of FAE1 KCS and its mutants was determined by measuring the release of radiolabeled CO 2 from [3- 14 C]malonyl-CoA. Decarboxyla- tion assays were carried out in a 15 · 45 mm glass vial, sealed with a Mininert valve (Pierce). To capture the released radiolabeled CO 2 ,a6· 30-mm tube containing a filter paper was placed in the 15 · 45 mm glass vial. A 50-lL reaction mixture, containing 40 m M sodium phos- phate buffer, pH 7.2, 15 l M 18 : 1-CoA, and 20 l M [3- 14 C]malonyl-CoA (30.5 lCiÆlmol )1 ), was placed in the 15 · 45 mm glass vial. The reaction was started by addition of protein, and the mixture was incubated at 30 °C. The reaction was stopped by the addition of trichloroacetic acid to the reaction mixture to give a final concentration of 10%. Immediately after trichloroacetic acid addition, 200 lLof the CO 2 trapping solution (20% triethylamine in methanol) was added to the 6 · 30-mm tube containing the filter paper and incubated for 1 h at room temperature. After comple- tion of 14 CO 2 absorption, the tube containing the trapping solution was analyzed by liquid scintillation counter. An absorption efficiency factor of 50% for the system was determined using 14 C-labeled sodium bicarbonate. Ó FEBS 2002 Mechanistic studies of FAE1 KCS (Eur. J. Biochem. 269) 3533 RESULTS Structural analysis of FAE1 KCS protein Hydropathy analysis (Kyte–Doolittle) of amino-acid sequence of FAE1 KCS revealed several hydrophobic domains, which constituted potential membrane spanning domains (Fig. 2A). However, alignment of KCS1 and several other putative KCSs [25] with FAE1 KCS and analysis with the TMAP algorithm [35] predicted only two N-terminal transmembrane domains. The first transmem- brane domain corresponds to amino-acid residues 9–36, and the second one spans residues 48–76 (Fig. 2B), suggesting that the FAE1 KCS is anchored to the membrane. In addition, FAE1 KCS and other elongase condensing enzymes lack any known signal targeting sequence for plant enzymes [36], and might suggest that these micro- somal membrane proteins are targeted to the endoplasmic reticulum. Engineering FAE1 KCS Earlier work in our laboratory to express FAE1 KCS in E. coli was unsuccessful, resulting in inclusion bodies. In contrast, expression of this protein in yeast yielded an active enzyme and proved to be a reliable system for analysis of the FAE1 KCS activity [7,9,25]. Two approaches were taken to engineer FAE1 KCS to facilitate its purification. One approach was to engineer in a thrombin cleavage site just downstream from the putative transmembrane domains with the aim to release an active soluble protein after the thrombin cleavage. The second approach was to entail a His-tagatC-orN-terminusofFAE1KCS,allowing purification on a metal affinity column after solubilizing the enzyme. Out of the four FAE1 KCSs engineered with thrombin cleavage site at different locations, only T115-FAE1 KCS retained wild-type activity after expression in yeast (data not shown). However, thrombin digestion of microsomal T115- FAE1 KCS resulted in complete loss of activity (data not shown). Although this approach failed to yield an active soluble enzyme, it provided useful information regarding the structure of FAE1 KCS. Immunoblot analysis revealed that thrombin treatment of the microsomal T115-FAE1 KCS produced a fragment corresponding to the expected size of 43 kDa (Fig. 3). However, this fragment was not released from the membrane as it was still associated with the pellet fraction after centrifugation at 100 000 g for 1 h (Fig. 3, Lane 5 and 6). This indicated that there were additional interactions, beyond amino-acid residue 115, between this protein and the membrane. To determine the nature of this interaction, microsomal pellet samples were treated with 0.5% Triton X-100, 0.5% Triton X-100 plus 0.32 M NaCl, or 2 M NaCl after thrombin digestion. As we had determined earlier for the native enzyme, treatment with 0.5% Triton X-100 alone did not solubilize the cleaved fragment completely (Fig. 3, Lane 7 and 8). However, treatments with Triton X-100 in combination with 0.32 M NaCl or treatment with 2 M NaCl alone resulted in complete release of the cleaved fragment (Fig. 3). Engineering the His-tag at C-terminus of FAE1 KCS led to a significant loss of activity of the recombinant protein (data not shown). However, the microsomal pellet contain- ing the N-terminus His-tagged FAE1 KCS retained the same level of condensation activity (1.06 ± 0.04 nmolÆmin )1 Æmg )1 ) as microsomal pellet containing wild- type protein (1.06 ± 0.03 nmolÆmin )1 Æmg )1 ). Solubilization and purification of N-terminus His-tagged FAE1 KCS The optimal detergent to protein ratio for solubilization of N-His 6 -FAE1 KCS protein was 2.5 : 1 (0.5% w/v Triton X-100 with 2 mg proteinÆmL )1 ), and the presence of 0.32 M salt was required for solubilization of recombinant protein. After 2 h of treatment, microsomes were centrifuged at 100 000 g for 1 h, and the supernatant fractions were assayed for FAE1 KCS activity. All of the activity was Fig. 2. Hydropathy analysis of FAE1 KCS. (A) Hydropathy plot of FAE1 KCS indicating the presence of several hydrophobic regions. The position of the active-site cysteine, Cys223, is indicated by an arrow. (B) Schematic representation of the putative transmembrane domains of FAE1 KCS amino-acid sequence as predicted by TMAP analysis [35]. Numbers shown inside the boxes correspond to the residues of each domain in FAE1 KCS. Fig. 3. Immunoblot analysis of thrombin-treated microsomal T115- FAE1 KCS. Microsomal T115-FAE1 KCS was treated overnight at 4 °C with thrombin. After thrombin digestion, sample was divided into four aliquots. Each aliquot was treated separately with 0.5% Triton X-100, 0.5% Triton X-100 plus 0.32 M NaCl, 2 M NaCl, or no treat- ment for 2 h at 4 °C. Samples were then centrifuged at 100 000 g for 60 min. The pellet (P) and supernatant (S) fractions of each sample were separated on 10% SDS/PAGE gel, followed by immunoblot analysis. Lanes 1: control yeast microsomes; Lanes 2, and 3: T115- FAE1 KCS, and Thrombin-treated T115-FAE1 KCS microsomes, respectively. Lanes 4 and 5, respectively, P and S fractions of untreated thrombin digested microsomes. Lanes 6 and 7, respectively, P and S fractions of 0.5% Triton X-100 treatment of thrombin digested microsomes. Lane 8 and 9, respectively, P and S fractions of 0.5% Triton X-100 plus 0.32 M NaCl treatment of thrombin digested microsomes. Lane 10 and 11, respectively, P and S fractions of 2 M NaCl treatment of thrombin digested microsomes. 3534 M. Ghanevati and J. G. Jaworski (Eur. J. Biochem. 269) Ó FEBS 2002 recovered in the supernatant fraction indicating that the enzyme has been solubilized. In all experiments, microsomes from yeast transformed with the empty vector were used as a negative control. The supernatant fractions (0.4 mg protein) were purified on a Ni 2+ -PDC column, and the eluants were assayed for FAE1 KCS activity using 18 : 1-CoA substrate. The yield for the purified recombinant FAE1 KCS was 3.5–4 lg, and its activity was close to 100% of the activity loaded onto the Ni 2+ -PDC column, indicating no loss of activity. No condensation activity was detected in the Ni 2+ -PDC purified control. Silver stain analysis of the eluant for purified recombinant protein indicated the presence of a major distinct band with the apparent molecular mass of 56 kDa (Fig. 4). The identity of this band as FAE1 KCS was confirmed by Western blot analysis (data not shown). Furthermore, Western blot analysis demonstrated the purified FAE1 KCS comigrated with the membrane-bound FAE1 KCS (as shown in Fig. 3) and thus confirmed that FAE1 KCS had not undergone degradation during solubilization and puri- fication. In addition to FAE1 KCS, several other minor protein bands were present, indicating the sample was highly enriched for the FAE1 KCS. The expression and accumulation of FAE1 KCS was very low in these samples as evidenced by the lack of a distinguishable FAE1 KCS band in the solubilized microsomes prior to purification (Fig. 4, lane 2). This one step purification resulted in approximately 100-fold purification of the recombinant FAE1 KCS with the specific activity increasing from 1.0 to 2.0 nmolÆmin )1 Æmg protein )1 to 150–200 nmolÆmin )1 Æ mg protein )1 . Attempts to further purify the His-tagged FAE1 KCS to homogeneity were not successful due to the loss of activity in subsequent steps. Optimization of assay conditions of the wild-type recombinant FAE1 KCS Measurement of the condensation activity of the isolated recombinant FAE1 KCS in the pH range between 4.5 and 8.5 in sodium phosphate buffer indicated a pH optimum in the range of 6.6–7.5. Addition of cofactors such as CoA, NADPH and ATP had no effect on the condensation activity of the recombinant FAE1 KCS. Condensation activity, as measured by the incorporation of [1– 14 C] malonyl-CoA, was linear for at least 15 min at low concentration (35 ng) of protein (Fig. 5). All subsequent condensation assays for FAE1 KCS were carried out at low protein concentration for 10 min. Substrate specificity of wild-type recombinant FAE1 KCS Analysis of the substrate preference of isolated recombinant FAE1 KCS showed that 18:1-CoA is the preferred substrate for this enzyme (Fig. 6). However, FAE1 KCS was nearly as active with 16:0, 16:1, and 18:0 and had 35% activity with 20:1. In contrast with its high activity with 18:0 and 18:1-CoAs, FAE1 KCS had no activity with polyunsatu- rated C18:2 and C18:3. Little or no activity was detected with acyl-CoAs having 22 carbons or longer in chain length. Decarboxylation activity In order to assay the second partial reaction of the condensation mechanism (Fig. 1), the decarboxylation of malonyl-CoA was monitored by the release of 14 CO 2 .Yeast microsomes exhibited high rates of decarboxylation activity, such that the yeast control activity was equal to the decarboxylation activity of the microsomal FAE1 KCS (data not shown). Furthermore, these high rates of decarb- oxylation were observed in the solubilized fraction of the control microsomes, and this activity was 18 : 1-CoA independent. The purification of the recombinant FAE1 KCS on the Ni 2+ -PDC column eliminated nearly all of this background decarboxylation activity (Fig. 7). In addition, decarboxylation of malonyl-CoA by the isolated recombin- ant FAE1 KCS was reduced to the background activity Fig. 4. SDS/PAGE analysis of isolated recombinant FAE1 KCS. Lane 1: 15 lg of the supernatant fraction of solubilized control microsomes; Lane 2: 15 lg of supernatant fraction of solubilized microsomes con- taining recombinant His-tag FAE1 KCS; Lane 3: 0.3 lgofNi 2+ -PDC purification of solubilized control microsomes; Lane 4: 0.3 lgofNi 2+ - PDC purified recombinant FAE1 KCS. The position of FAE1 KCS is indicatedbyanarrow. Fig. 5. Time course for condensation activity of FAE1 KCS. Isolated recombinant FAE1 KCS was assayed for condensation activity as described under Experimental procedures using either 35 ng (open circle)or70 ng(closedcircle)ofproteinina25-lL reaction mixture for indicated times. Ó FEBS 2002 Mechanistic studies of FAE1 KCS (Eur. J. Biochem. 269) 3535 when 18 : 1-CoA substrate was excluded from the reaction mixture (Fig. 7). Site-directed mutagenesis of the conserved residues To investigate the role of several conserved residues in the reaction mechanism of the FAE1 KCS, several FAE1 KCS mutants (C223A, H391A, H391K, H391Q, N424D, N424H) were made with N-terminus His 6 -tag and expressed in yeast cells. The His-tagged proteins were isolated on a Ni 2+ -PDC column and analyzed for overall condensation and decarboxylation activity. An initial progress curve was established for both activities for all mutant proteins. All subsequent measurements were carried out in quadruple at a fixed time point in the linear region of progress curve. Of the His391 mutants, the H391Q mutant remained the most active, with 25% activity compared to the wild-type for both condensation and decarboxylation reaction (Table 1). Replacement of His391 with Ala abolished condensation activity and Lys substitution resulted in retaining of only 1% of condensation activity. Decarboxy- lation activity of His391A mutant protein was at the background level and H391K had decarboxylation activity that was slightly above that of the background (Table 1). Substitution of Cys223 with Ala abolished overall condensation activity, as expected based on our earlier study [25]. In addition, decarboxylation activity was reduced to background activity for this mutant, indicating that decarboxylation of malonyl-CoA is dependent on binding of acyl-CoA substrate (Table 1). Substitution of Asn424 with His produced inactive enzyme, while its substitution with Asp led to only modest 80% and 70% reduction in activity for condensation and decarboxylation reactions, respectively (Table 1). DISCUSSION Site-directed mutagenesis and crystal structure analysis of soluble condensing enzymes involved in fatty acid and polyketide biosynthesis have demonstrated that the reaction catalyzed by these enzymes is tripartite and involves Cys, His, His [20,21] or Cys, His, Asn [19,23] as catalytic triad. It is now well documented that the active site cysteine acts as the nucleophile and provides an attachment site for the acyl substrate. Studies of both chalcone synthase [24] and KAS III [23] have demonstrated the importance of active site histidine and asparagine residues in decarboxylation of malonyl substrate by stabilizing the carbanion intermediate derived from decarboxylation. Unlike soluble condensing enzymes, which have been well characterized, little information is available on the structure and mechanism of the membrane-bound condensing enzymes. This is mainly due to the difficulties associated in solubilization and purification of these enzymes. Nearly all enzymatic studies of these membrane-bound condensing enzymes have been carried out using microsomal membrane or solubilized membranes, which precluded any analysis of reaction mechanism [10–12,14,15]. To overcome this shortcoming, we attempted to engineer the FAE1 KCS so that it could be rapidly isolated. In Fig. 7. Decarboxylation activity of isolated N-His 6 -FAE1 KCS. Time course decarboxylation of purified control and recombinant FAE1 KCS in the presence and absence of 15 l M 18:1-CoA. Decarboxylation activity was measured by release of CO 2 from [3- 14 C]malonyl-CoA as described under Experimental procedures. (d) FAE1 KCS with 18:1-CoA; (s) FAE1 KCS without 18:1-CoA; (r)yeastcontrolwith 18:1-CoA; (m)yeastcontrolwithout18:1-CoA. Fig. 6. Substrate specificity of recombinant FAE1 KCS. Substrate preference of FAE1 KCS was determined as measurement of con- densation activity using indicated acyl-CoA substrates at a final con- centration of 15 l M in 25 lL reaction mixture. The condensation assay was carried out as described in Experimental procedures. Reactions were started by addition of protein and carried out for 10 min. The activities are expressed as nmolÆmin )1 Æmg protein )1 , and they represent amean±SDforn ¼ 3. Table 1. Condensation and decarboxylation activity of purified mutant proteins. The activities are expressed as nmolÆmin )1 Æmg protein )1 and they represent a mean ± SD for n ¼ 4. ND; not detectable. Condensation Decarboxylation Vector ND 3.44 ± 0.87 FAE1-KCS 158 ± 23 67.0 ± 9.0 H391Q 37 ± 4.2 18.8 ± 3.8 N424D 36 ± 6.0 24.2 ± 3.5 H391K 1.0 ± 0.17 5.69 ± 0.83 C223A ND 2.24 ± 0.46 H391A ND 2.01 ± 0.52 N424H ND 1.04 ± 0.14 3536 M. Ghanevati and J. G. Jaworski (Eur. J. Biochem. 269) Ó FEBS 2002 addition, expression of this enzyme in yeast provided an opportunity to further analyze the FAE1 KCS using site- directed mutagenesis. In so doing, comparison of this membrane-bound condensing enzyme to soluble conden- sing enzymes became feasible. KCSs are predicted by the TMAP algorithm to have two transmembrane spanning domains close to their N-terminus [35]. Our results presented here for T115- FAE1 KCS confirmed this prediction. Treatment of the thrombin-digested microsomal T115-FAE1 KCS by 2 M salt alone was sufficient to solubilize the cleaved fragment, suggesting that the interaction of FAE1 KCS beyond its transmembrane domains with the membrane is mainly ionic. These results therefore support a model in which FAE1 KCS is anchored to the membrane by its trans- membrane domains, and the region beyond the transmem- brane domains constitutes the globular portion of this enzyme. Elongation of acyl substrates by fatty acid elongase system has been shown to be dependent on the presence of ATP and CoA [37,38]. However, it has not been demonstrated whether the b-ketoacyl-CoA synthase com- ponent of this elongase system requires cofactors for its activity. We found that there was no requirement for ATP, CoA, and NADPH for the activity of FAE1 KCS. FAE1 KCS showed high activity with monounsaturated and saturated C16 and C18 and no activity with polyunsaturated C18:2 and C18:3. In addition, consistent with previous observations [7], the level of activity on saturated and monounsaturated C20 was substantially lowerthanonC18. Wild-type recombinant FAE1 KCS was unable to carry out decarboxylation of malonyl-CoA in the absence of 18 : 1-CoA, thus suggesting that binding of the acyl-CoA to the active-site cysteine is required for decarboxylation of malonyl-CoA. Similarly, C223A recombinant FAE1 KCS protein was unable to carry out the decarboxylation of malonyl-CoA substrate, indicating that decarboxylation activity is dependent on acylation of the enzyme. Replace- ment of Cys223 with an alanine eliminates the binding site required for covalent attachment of the acyl group, therefore making this protein incapable of carrying the decarboxylation reaction. These results are consistent with the observations for decarboxylation activity of soluble condensing enzymes involved in fatty acid biosynthesis [39] and the b-ketoacyl synthase domain of the multifunctional animal fatty acid synthase [40] in which decarboxylation of malonyl substrate is dependent on the binding of the acyl substrate to the active-site cysteine. It is suggested that these enzymes follow a ping pong mechanism, in which after binding acyl-CoA, CoA is released before binding the second substrate, malonyl-CoA. In contrast, recent muta- tional studies of chalcone synthase have demonstrated that decarboxylation of malonyl-CoA is independent of acyla- tion of the active site cysteine [24]. In these studies, substitution of the active-site cysteine to alanine did not significantly reduce the decarboxylation activity of the chalcone synthase, thus indicating that acylation of the active-site cysteine is not essential for decarboxylation of malonyl-CoA substrate. Therefore, despite its higher degree of homology to chalcone synthase than to other condensing enzymes, FAE1 KCS appears to be more similar to soluble condensing enzymes involved in fatty acid biosynthesis with regard to the effect of acylation on decarboxylation activity. To further analyze the relation of structure and activity of FAE1 KCS, site-directed mutagenesis was also carried out on the histidine and asparagine residues that were conserved with chalcone synthase and KAS III. Both of these latter enzymes have been crystallized and the effect of mutagenesis on these conserved residues analyzed [23,24]. For both chalcone synthase and KAS III, a histidine to alanine substitution led to complete loss of condensation and decarboxylation activity, whereas a histidine to glutamine mutant of chalcone synthase retained approximately 15% of both its condensation and decarboxylation activity [24]. In the present study, very similar results were obtained, with complete loss of activity with the H391A mutant and retention of 25% of condensation and decarboxylation activities by the H391Q mutant. Similar to chalcone synthase, high retention of activity for H391Q mutant suggests that this residue is not involved in proton abstraction from the active site Cys223. Recently, kinetic studies of histidine mutants of chalcone synthase have demonstrated the existence of a thiolate-imidazolium ion pair at the chalcone synthase active site [41]. It is reported that due to its potential to form hydrogen bond, glutamine residue is still capable of stabilizing the thiolate of theactivesitecysteine.Theloweractivity,comparedtothe wild-type, of the histidine to glutamine mutant of chalcone synthase has been attributed to an increase in pKa value of the active site cysteine for this mutant. It is very likely, that the slight decrease in activity for H391Q mutant of FAE1 KCS is due to a similar effect. Furthermore, as FAE1 KCS is still very active at low pH of 4.5 it might suggest the presence of a thiolate-imidazolium ion pair at its active site similar to chalcone synthase. The effect, on activity, of amino-acid substitutions for the conserved asparagine residue in FAE1 KCS was also similar to the effect of the same substitutions in chalcone synthase. The chalcone synthase mutant N336H was completely inactive, whereas the N336D mutant retained 0.06% condensation activity and 0.3% of the decarboxylation activity [24]. The N424H mutant of FAE1 KCS was also completely inactive, whereas N424D mutant retained a surprising 20% and 30% of the condensation and decarb- oxylation activities, respectively. Although the N424D mutant was much more active than the corresponding mutant of chalcone synthase, it may be more significant that in the case of both mutants, the substitution of an acidic residue resulted in an active enzyme, whereas substitution of basic histidine for the asparagine resulted in inactive enzyme. Taken together, the analysis of the decarboxylation activity and characterization of the mutants of the putative catalytic triad strongly support the hypothesis that the membrane-bound FAE1 KCS shares the same basic mechanism with the soluble condensing enzymes. Addi- tional studies will determine the full extent of this similarity. ACKNOWLEDGEMENTS This work was supported by National Science Foundation Grant MCB-9728786. Ó FEBS 2002 Mechanistic studies of FAE1 KCS (Eur. J. Biochem. 269) 3537 REFERENCES 1. Lynch, D.V. (1993) Sphingolipids. In Lipid Metabolism in Plants (Thomas, S, Moore, J., eds), pp. 286–308. CRC Press, Inc, Boca Raton, FL. 2. Merrill, A.H.J., Schmelz, E.M., Wang, E., Dillehay, D.L., Rice, L.G., Meredith, F. & Riley, R.T. (1997) Importance of sphingol- ipids and inhibitors of sphingolipid metabolism as components of animal diets. J. Nutrition 127, 830S–833S. 3. Post-Beittenmiller, D. (1996) Biochemistry and molecular biology of wax production in plants. Ann. Rev. Plant Phys. Plant Mol. Biol. 47, 405–430. 4.Downey,R.K.&Ro ¨ bbelen, G. (1989) Brassica species. In Oil Crops of the World (Ro ¨ bbelen,G.,Downey,R.KandAshri,A., eds) , pp. 339–362. McGraw-Hill, Inc, New York, NY. 5. Fehling, E. & Mukherjee, K.D. (1991) Acyl-CoA elongase from a higher plant (Lunaria annua): metabolic intermediates of very- long-chain acyl-CoA products and substrate specificity. Biochim. Biophys. Acta. 1082, 239–246. 6. Suneja, S.K., Nagi, M.N., Cook, L. & Cinti, D.L. (1991) Decreased long-chain fatty acyl coenzyme A elongation activity in quaking and jimpy mouse brain: deficiency in one enzyme or multiple enzyme activities? J. Neurochem. 57, 140–146. 7. Millar, A.A. & Kunst, L. (1997) Very-long-chain fatty acid biosynthesis is controlled through the expression and specificity of the condensing enzyme. Plant J. 12, 121–131. 8. James, J.D.W., Lim, E., Keller, J., Plooy, I., Ralston, E. & Dooner, H.K. (1995) Directed tagging of the Arabidopsis fatty acid elongation1 (FAE1) gene with maize transposon activator. Plant Cel. 7, 309–319. 9. Todd, J., Post-Beittenmiller, D. & Jaworski, J.G. (1999) KCS1 encodes a fatty acid elongase 3-ketoacyl-CoA synthase affect- ing wax biosynthesis in Arabidopsis thaliana. Plant J. 17, 119–130. 10. Lessire, R., Bessoule, J J. & Cassagne, C. (1985) Solubilization of C18-CoA and C20-CoA elongases from Allium porrum L. epi- dermal cell microsomes. FEBS Lett. 187, 314–320. 11. Fehling, E., Lessire, R., Cassagne, C. & Mukherjee, K.D. (1992) Solubilization and Partial Purification of Constituents of Acyl- CoA Elongase from Lunaria annua. Biochim. Biophys. Acta. 1126, 88–94. 12. Imai, H., Hlousek-Radojcic, A., Matthis, A. & Jaworski, J. (1994) Elongation system involved in the biosynthesis of very long chain fatty acids in Brassica napus seeds: characterization and solubili- zation. In Plant Lipids Metabolism (Kader, J C & Mazliak, P., eds), pp. 118–120. Kluwer Academic Publishers, Dordrecht, the Netherlands. 13. Lassner, M.W., Lardizabal, K. & Metz, J.G. (1996) A jojoba b-ketoacyl-CoA synthase cDNA complements the canola fatty acid elongation mutation in transgenic plants. Plant Cell 8, 281–292. 14. Evenson, K.J. & Post-Beittenmiller, D. (1995) Fatty acid-elonga- ting activity in rapidly expanding leek epidermis. Plant Physiol. 109, 707–716. 15. Lessire, R., Bessoule, J J. & Cassagne, C. (1989) Involvement of a b-ketoacyl-CoA intermediate in acyl-CoA elongation by an acyl- CoA elongase purified from leek epidermal cells. Biochim. Biophys. Acta 1006, 35–40. 16. Martin, C.R. (1993) Structure, function, and regulation of the chalcone synthase. International Review of Cytology – a Survey of Cell Biology, Vol. 147 (Jeon, K.W. & Jarvik, J., eds), pp. 233–284. Academic Press Inc, San Diego, CA, USA. 17. Tai, H.Y. & Jaworski, J.G. (1993) 3-Ketoacyl-acyl carrier protein synthase-iii from spinach (Spinacia oleracea)isnotsimilartoother condensing enzymes of fatty acid synthase. Plant Physiol. 103, 1361–1367. 18. Tsay,J.T.,Oh,W.,Larson,T.J.,Jackowski,S &Rock,C.O. (1992) Isolation and characterization of the b-ketoacyl-acyl carrier protein synthase-III gene (fab H) from Escherichia coli K-12. J. Biol. Chem. 267, 6807–6814. 19. Ferrer, J.L., Jez, J.M., Bowman, M.E., Dixon, R.A. & Noel, J.P. (1999) Structure of chalcone synthase and the molecular basis of plant polyketide biosynthesis. Nat. Struct. Biol. 6,775– 784. 20. Huang,W.J.,Jia,J.,Edwards,P.,Dehesh,K.,Schneider,G.& Lindqvist, Y. (1998) Crystal structure of b-ketoacyl-acyl carrier protein synthase II from E. coli reveals the molecular architecture of condensing enzymes. EMBO J. 17, 1183–1191. 21. Olsen, J.G., Kadziola, A., von Wettstein-Knowles, P., Siggaard- Andersen, M., Lindquist, Y. & Larsen, S. (1999) The X-ray crystal structure of b-ketoacyl [acyl carrier protein] synthase I. FEBS Lett. 460, 46–52. 22. Qiu, X.Y., Janson, C.A., Konstantinidis, A.K., Nwagwu, S., Silverman, C., Smith, W.W., Khandekar, S., Lonsdale, J. & Abdel-Meguid, S.S. (1999) Crystal structure of b-ketoacyl-acyl carrier protein synthase III – a key condensing enzyme in bacterial fatty acid biosynthesis. J. Biol. Chem. 274, 36465–36471. 23. Davies, C., Heath, R.J., White, S.W. & Rock, C.O. (2000) The 1.8 _ crystal structure and active-site architecture of b-ketoacyl-acyl carrier protein synthase III (FabH) from Escherichia coli. Struc- ture 8, 185–195. 24. Jez, J.M., Ferrer, J L., Bowman, M.E., Dixon, R.A. & Noel, J.P. (2000) Dissection of malonyl-Coenzyme A decarboxylation from polyketide formation in the reaction mechanism of a plant poly- ketide synthase. Biochemistry 39, 890–902. 25. Ghanevati, M. & Jaworski, J.G. (2001) Active-site residues of a plant membrane-bound fatty acid elongase b-ketoacyl-CoA syn- thase, FAE1 KCS. Biochim. Biophys. Acta 1530, 77–85. 26. Roughan, G. (1994) A semi-preparative enzymic synthesis of malonyl-CoA from [C-14]acetate and (CO 2 )-C-14: labelling in the 1, 2 or 3 position. Biochem. J. 300, 355–358. 27. Ho, S.N., Hunt, H.D., Horton, R.M., Pullen, J.K. & Pease, L.R. (1989) Site-directed mutagenesis by overlap extension using the polymerase chain reaction. Gene 77, 51–59. 28. Gietz, R.D. & Woods, R.A. (1994) High efficiency transformation in yeast. In Molecular Genetics of Yeast: Practical Approaches. (Johnston, J.A., ed.), pp. 121–134. Oxford University Press, Oxford, UK. 29. Bradford, M.M. (1976) A rapid and sensitive method of quanti- tation of microgram quantities of protein utilizing the principle of protein-dye binding. Anal. Biochem. 72, 248–254. 30. Laemmli, U.K. (1970) Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227, 680–685. 31. Kyhse-Andersen, J. (1984) Electroblotting of multiple gels: a simple apparatus without buffer tank for rapid transfer of proteins from polyacrylamide to nitrocellulose. J. Biochem. Biophys. Methods 10, 203–209. 32. Harlow,E.&Lane,D.(1988)Antibodies: A Laboratory Manual, Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY, USA. 33. Hochstrasser, D.F., Patchornik, A. & Merril, C.P. (1988) Development of polyacrylamide gels that improve the separation of proteins and their detection by silver staining. Anal. Biochem. 173, 412–423. 34. Garwin, J.L., Klages, A.L. & Cronan, J.E. (1980) Jr Structural, enzymatic, and genetic studies of b-ketoacyl-acyl carrier protein synthases I and II of Escherichia coli. J. Biol. Chem. 255, 11949– 11956. 35. Persson, B. & Argos, P. (1994) Prediction of transmembrane segments in proteins utilising multiple sequence alignments. J. Mol. Biol. 237, 182–192. 3538 M. Ghanevati and J. G. Jaworski (Eur. J. Biochem. 269) Ó FEBS 2002 36. Emanuelsson, O., Nielsen, H., Brunak, S. & von Heijne, G. (2000) Predicting subcellular localization of proteins based on their N-terminal amino acid sequence. J. Mol. Biol. 300, 1005–1016. 37. Agrawal, V., Lessire, R. & Stumpf, P. (1984) Biosynthesis of very long chain fatty acids in microsomes from epidermal cells of Allium porrum L. Arch. Biochem. Biophys. 230, 580–589. 38. Agrawal, V. & Stumpf, P. (1985) Characterization and solubili- zation of an acyl chain elongation system in microsomes of leek epidermal cells. Arch. Biochem. Biophys. 240, 154–165. 39. Kresze, G B., Steber, L., Oesterhelt, D. & Lynen, F. (1977) Reaction of yeast fatty acid synthetase with iodoacetamide. 2. Identification of the amino acid residues reacting with iodoaceta- mide and primary structure of a peptide containing the peripheral sulfhydryl group. Eur. J. Biochem. 79, 181–190. 40. Witkowski, A., Joshi, A.K., Lindqvist, Y. & Smith, S. (1999) Conversion of a b-ketoacyl synthase to a malonyl decarboxylase by replacement of the active-site cysteine with glutamine. Bio- chemistry 39, 11643–11650. 41. Jez, J.M. & Noel, J.P. (2000) Mechanism of chalcone synthase: pKa of the catalytic cysteine and the role of the conserved histi- dine in a plant polyketide synthase. J. Biol. Chem. 275, 39640– 39646. Ó FEBS 2002 Mechanistic studies of FAE1 KCS (Eur. J. Biochem. 269) 3539 . activity. All of the activity was Fig. 2. Hydropathy analysis of FAE1 KCS. (A) Hydropathy plot of FAE1 KCS indicating the presence of several hydrophobic. Chemistry and Biochemistry, Miami University, Oxford, OH, USA The Arabidopsis FAE1 b-ketoacyl-CoA synthase (FAE1 KCS) catalyzes the condensation of malonyl-CoA