Identification of a Chloroplast Coenzyme A-Binding Protein ...Spinach (Spinacea olevacea cv Melody, Stocks Seeds, Buf- falo, NY) was grown in a greenhouse at 22°C under cool fluorescent

Plant Physiol. (1 996) 1 12: 1641-1 647

Identification of a Chloroplast Coenzyme A-Binding Protein Related to the Peroxisomal Thiolases'

Li-Ming Yang2 and Cayle Lamppa* Department of Molecular Genetics and Cell Biology, The University of Chicago, 920 East 58th Street,

Chicago, lllinois 60637

A 30-kD coenzyme A (COA)-binding protein was isolated from spinach (Spinacea oleracea) chloroplast soluble extracts using affinity chromatography under conditions in which 95% of the total protein was excluded. The 30-kD protein contains an eight-amino- acid sequence, DVRLYYCA, that i s identical to a region in a 36-kD protein of unknown function that i s encoded by a kiwifruit (Actin- idia deliciosa) cDNA. Southern blotting also detected a spinach gene that i s related to the kiwifruit cDNA. The kiwifruit 36-kD protein that was synthesized in Escherícbia coli was imported into chloroplasts and cleaved to a 30-kD form; it was processed to the same size in an organelle-free assay. Furthermore, the kiwifruit protein specifically bound to COA. The kiwifruit protein contains a single cysteine within a domain that i s related to the peroxisomal 6-ketoacyCCoA thiolases, which catalyze the COA-dependent degradative step of fatty acid P-oxidation. Within 50 amino acids surrounding the cysteine, considered to be part of the thiolase active site, the kiwifruit protein shows approximately 26% sequence identity with the mango, cucumber, and rat peroxisomal thiolases. N-terminal alignment with these enzymes, relative to the cysteine, indicates that the 36-kD protein i s cleaved after serine-58 during import, agreeing with the estimated size (approximately 6 kD) of a transit peptide. The 30-kD protein is also related to the E. coli and mitochondrial thiolases, as well as t o the acetoacetylSoA thiolases of prokaryotes. Features distinguish it from members o f the thiolase family, suggesting that it carries out a related but nove1 function. The protein i s more distantly related to chloroplast P-ketoacyl-acyl carrier protein synthase I I I , the initial condensing enzyme of fatty acid synthetase that utilizes acetyl-COA.

Plastids are multifunctional organelles that carry out numerous biochemical reactions in different organs in addition to photosynthesis in green tissues. The role of plastids as the primary source of fatty acids in the plant cell has been well documented. The major products of fatty acid synthesis are palmitic acid and oleic acid. Fatty acids are either transferred to glycerol-3-phosphate via an acyltrans- ferase, directing them along a "prokaryotic" pathway within the plastid, or released by a thioesterase for organel-

' This work was supported in part by U.S. Department of Ag- riculture (USDA) grant no. 91-37304-6466 and the USDA North Central Biotechnology Initiative, Purdue University subcontract no. 94-34190-1204 to G.K.L., and a Sigma Xi grant-in-aid of research to L.-M.Y.

Present address: Plant Gene Expression Center, 800 Buchanan Street, Albany, CA 94710.

* Corresponding author; e-mail [email protected]. edu; fax 1-312-702-3172.

lar export. The latter "eukaryotic" pathway is mediated by an acyl-COA synthetase of the outer membrane, leading to the production of complex lipids in the cytosol (for re- views, see Stumpf, 1987; Ohlrogge and Browse, 1995).

We have been investigating the biosynthetic pathway of ACP, to which fatty acid chains are attached during their elongation. During these studies (L.-M. Yang and G. Lamppa, unpublished results) we developed a protocol to purify the chloroplast holoACP synthase, identified as a soluble enzyme (Fernandez and Lamppa, 1990, 1991) that transfers the functional prosthetic group to apoACP from COA. COA-affinity chromatography was employed as a biochemical screen to rapidly search for a small group of chloroplast proteins with a high affinity for COA. One of these proteins, which was equal to 30 kD and showed the same affinity for COA as the holoACP synthase but has not been shown to possess this activity, was subjected to an interna1 sequence analysis. An eight-amino-acid sequence was identical to a region within a 36-kD protein that was encoded by a kiwifruit (Actinidia deliciosa) cDNA clone (pKIWI502) of unknown function, prepared from RNA differentially expressed in fruit (Ledger and Gardner, 1994).

Because of the strong affinity the spinach (Spinacia oleracea) 30-kD protein showed for COA in our chromatography experiments, the primary sequence of the kiwifruit 36-kD protein was carefully examined for motifs that might be important for the activity of proteins known to utilize COA or its derivatives. One characteristic of some of these enzymes that are needed either during fatty acid synthesis or P-oxidation is a special Cys residue that permits acyl-group transfer from one substrate to another. Our comparison revealed that the kiwifruit protein contains a single Cys that lies within a domain that shows significant sequence similarity to the degradative 0-ketoacyl-COA thiolases of the peroxisome, mitochondria, and the homolog in Escke- rickia coli (Igual et al., 1992). In plants the fatty acid P-oxidation pathway occurs primarily in peroxisomes (Kindl, 1993). The thiolases, in the final step of P-oxidation, transfer acyl groups to COA via a ping-pong mechanism from acyl-COA chains, releasing acetyl-COA. The proposed mechanism includes a thioester linkage to the special Cys, producing an acyl-S-enzyme intermediate, followed by

Abbreviations: ACP, acyl carrier protein; KAS, P-ketoacyl-acyl carrier protein synthase; pKIWI502, kiwifruit cDNA plasmid clone #502.

1641 www.plantphysiol.orgon May 28, 2020 - Published by Downloaded from

Copyright © 1996 American Society of Plant Biologists. All rights reserved.

http://www.plantphysiol.org

1642 Yang and Lamppa Plant Physiol. Vol. 11 2, 1996

transfer of the acyl group to COA (Gehring and Harris, 1970; Miyazawa et al., 1981). The result is the gradual reduction in length of the fatty acid chain, two carbons per cycle, yielding acyl-CoA(,-,, and acetyl-COA.

In this study we provide evidence that the spinach CoA- binding protein and the protein encoded by the kiwifruit cDNA are homologs located in the chloroplast, and we asses their relationship to the family of thiolases.

MATERIALS AND METHODS

Plant Growth, Chloroplast Isolation, and Extract Preparation

Spinach (Spinacea olevacea cv Melody, Stocks Seeds, Buf- falo, NY) was grown in a greenhouse at 22°C under cool fluorescent lights (14 h of 1ight:lO h of dark). Plants (4 weeks old) were transferred to the dark 1 d before harvest- ing to reduce the starch content. Leaves were homogenized with a Polytron at 4°C in GR buffer (50 mM Hepes-KOH, pH 8.0, 0.33 M sorbitol, 2 mM EDTA, 1 mM MgCl,, 5 mM sodium ascorbate, and 0.25% BSA) at a ratio of 5 mL/g of tissue (fresh weight). Intact chloroplasts recovered from the interface of 40 and 80% Perco11 step gradients were washed once in GR buffer, collected at 15008 for 5 min, and lysed in 20 mM Hepes-KOH, pH 8.0, and 1 mM PMSF to make a crude chloroplast extract.

Expression of Recombinant Precursors in Escherichia coli Cells

A pKIWI502 cDNA clone (a 1.2-kD insert in the vector pSPORT) was kindly provided by Drs. R. Gardner and S. Ledger (University of Auckland, New Zealand). To ex- press the gene in E. coli, oligonucleotides overlapping the initiation codon and immediately 3' of the termination codon were used to amplify the coding region of the KIWI502 gene by PCR. A 1.0-kb fragment was then sub- cloned into the vector pET-3d at the NcoI and BamHI sites, and the plasmid pET-3d:KIWI502 was transformed into BL21(DE3)pLysS (Studier et al., 1990). The plasmid pET-3d::preACP that was used for the synthesis of spinach preACP was constructed previously (Yang et al., 1994). An overnight culture of BL21(DES)pLysS, carrying either pET-3d:KIWI or pET3d::preACP, was inoculated at a 1000-fold dilution into 10 mL of M9 medium containing 50 pg / mL ampicillin, 25 pg / mL chloramphenicol (Sig- ma), and 6 kg/mL sulbactum (Plizer, Groton, CT), and incubated at 37°C until AGO0 reached approximately 0.2. Cells were harvested, resuspended in 1 mL of M9 medium containing 1 mM isopropyl P-D-thiogalactopyranoside, and incubated at 37°C. Thirty minutes later, 200 pg of rifampicin was added and the cells were incubated for 90 min. After adding 10 pL of [35S]Met (100 mCi/mL, Dupont-NEN) the cells were incubated for 30 min, harvested, and lysed, and the insoluble materials were washed twice with a detergent buffer (0.5y0 Triton X-100 and 1 mM EDTA). The final pellet was dissolved in 8 M urea. Large amounts of unlabeled kiwifruit (Actinidia deliciosa) 36-kD protein were prepared according to Abad et al. (1991) and used for the production of antibodies in

rabbits. After recovery in 6 M urea, they were sequentially dialyzed for 2 h against 4.5 M, 3 M, and 1.5 M urea with 1 mM PMSF, followed by 20 mM Tris-HC1, pH 8.0, to en- hance refolding.

Preparation of COA-Affinity Matrix

Cyanogen bromide-activated Sepharose 4B (8 g, Pharma- cia) was hydrated in 80 mL of 1 mM HC1 for 15 min while shaking. The resin was washed four times with 200 mL of 1 mM HC1 and once with 200 mL of the cross-linking buffer (0.1 M NaHCO, and 0.5 M NaC1, pH 8.3). COA (Sigma) was added to the resin at approximately 2 pmol/mL, i.e. 70 pL of 25 mM COA per ml of resin, and the cross-linking reaction (via the amino group) was carried out at 4°C for 16 h. The remaining active sites on the resin were blocked by incubating with 0.5 M Tris-HC1, pH 8.0, at room temperature for 1 h. Finally, the resin was washed three times with 200 mL of the cross-linking buffer and 200 mL of 0.1 M

acetic acid. The COA-Sepharose was rinsed with distilled water before packing a glass column (18 cm X 2.5 cm) for affinity chromatography, or was stored at 4°C in 0.2% sodium azide.

COA-Affinity Column Chromatography, Fraction Analysis, and Protein Blotting

Chloroplasts were isolated from 2 kg of spinach leaves to prepare 200 mL of a high-speed extract. They were lysed in 20 mM Hepes-KOH, pH 8.0, and 1 mM PMSF at a concen- tration of 1 to 2 mg chlorophyll/mL. The lysate was left on ice for 1 h. Membranes were removed by centrifugation first at 10,OOOg for 20 min and then at 96,0008 for 1 h. The extract was brought to 60% saturation by gradually adding 70 g of ammonium sulfate over 30 min while stirring at 4°C. The precipitate was collected at 10,OOOg for 20 min and dissolved in 20 mL of buffer A (20 mM Hepes-KOH, pH 8.0, 5 mM MgCl,, 100 mM NaC1, and 1 mM DTT) and 1 mM PMSF. Particulates were removed by centrifugation at 10,OOOg for 20 min. The supernatant was desalted twice with Sephadex G-25 columns (Pharmacia), pooled, and diluted 5-fold with buffer A and 1 mM PMSF. This protein sample was loaded onto a 30-mL COA-Sepharose 4B column at 4°C at a flow rate of 0.25 mL/min and the flow- through was collected. The column was then washed with 500 mL of buffer A and 0.1 mM PMSF at a flow rate of 0.5 mL/min. Bound proteins were eluted with 30 mL of 1 mM COA at 0.5 mL/min.

Fractions were analyzed by SDS-PAGE and proteins were blotted onto a nitrocellulose filter. The filter was washed twice with TTBS buffer (20 mM Tris-HC1, pH 7.5,0.5 M NaC1, and 0.3% Tween 20) for 20 min, three times in distilled water for 1 min, and soaked in colloidal gold staining solution (Bio-Rad) for approximately 30 min on a rotary platform. For protein sequencing, proteins were blotted onto a PVDF- PsQ membrane (Millipore) in 10 mM 3-(cyclohexylamino)-l- propanesulfonic acid and 10% methanol, pH 11.0, for 20 h at 100 mA. The membrane was stained by 1% Ponceau S in 5% acetic acid, and the membrane with the 30-kD protein was

www.plantphysiol.orgon May 28, 2020 - Published by Downloaded from Copyright © 1996 American Society of Plant Biologists. All rights reserved.


Chloroplast CoA-Binding Protein and Peroxisomal Thiolases 1643

excised for microsequencing (Rockefeller University Se-quencing Facility, New York, NY).

CoA-Affinity Batch Chromatography

CoA-Sepharose was resuspended in 10 mL of buffer B(equal to buffer A, except 10 mM Hepes-KOH, pH 8.0). Theprotein encoded by pKIWI502 (400 ;u,L at 2 mg/mL) recov-ered from £. coli was added to the matrix (15 mL) andincubated with gentle shaking for 2 h at 4°C. The matrixwas spun at 1300g for 5 min and the supernatant was savedas the flow-through. The matrix was washed three separatetimes with 30 mL of buffer B and spun at 1300g. The matrixwas then incubated with 3 mL of 1 mM CoA in buffer B for2 h at 4°C and spun at 1300g for 10 min, and the eluate wascollected. Fractions were analyzed by SDS-PAGE followedby immunoblotting using antibodies that were generatedagainst the purified kiwifruit protein that was synthesizedin £. coli.

Chloroplast Import and Organelle-Free Processing Assays

Import reactions were previously described (Fernandezand Lamppa, 1990). Chloroplasts were incubated with 3/nL of radiolabeled recombinant precursor for 30 min at26°C and stopped at 4°C. One-half of the reaction wastreated with 100 |ug/mL thermolysin, 8 mM CaCl2 in 50mM Hepes-KOH, pH 8.0, 0.33 M sorbitol, and 10 mM Metfor 30 min at 4°C and stopped by making it 20 mM EGTAand 2 mM PMSF. The chloroplasts were pelleted and lysedin 1 mM PMSF. Membranes were removed by centrifuga-tion at 16,000g for 15 min. Soluble proteins were lyophi-lized and resuspended in 0.5 M Tris-HCl, pH 6.8, 5%j3-mercaptoethanol, 2% SDS, and 10% glycerol for SDS-PAGE, followed by autoradiography. Organelle-free pro-cessing reactions with radiolabeled substrates were car-ried out as described previously (Abad et al., 1988).

Southern Blot Analysis

Spinach, kiwifruit, and wheat genomic DNAs (approxi-mately 50 ;ug) were digested with 20 units of BamHI, Hindlll,and EcoRl at 37°C for 16 h, run on a 0.8% agarose gel,transferred to a Hybond-N membrane (Amersham), andcross-linked to the membrane at 80°C for 2 h under a vac-uum. pKIWI502 was digested with Ncol and labeled withdigoxigenin by random priming. Reagents and protocolswere from Boehringer Mannheim. The labeling reaction (20JJ.L), containing 100 ng of pKIWI502, 10 X hexanucleotidemixture, 10 X deoxyribonucleotide triphosphate labelingmixture, and 2 units of Klenow enzyme, was performed at37°C for 20 h and then boiled. Before hybridization, themembrane was incubated with 25 mL of blocking buffer (450mM NaCl, 45 mM sodium citrate, pH 7.0, 1% blocking re-agent, 0.1% N-lauroylsarcosine, and 0.02% SDS) at 50°C for3 h. The digoxigenin-labeled pKIWI502 probe (15 ng) wasadded in 15 mL of the above buffer and hybridized at 50°Cfor 16 h. The membrane was washed twice with 2X SSC (300mM NaCl and 30 mM sodium citrate, pH 7.0) and 0.1% SDSat room temperature for 10 min, and twice with 0.5 X SSC (75mM NaCl and 7.5 mM sodium citrate, pH 7.0) and 0.1% SDS

at 65°C for 15 min. To detect the digoxigenin-labeled DNAby chemiluminescence, the membrane was blocked againand then incubated with an antidigoxigenin IgG. After add-ing 0.5 mL of Lumi-Phos 530 (Boehringer Mannheim) to thesurface, the membrane was exposed to x-ray film for 30 min.

RESULTS

Affinity chromatographic methods were developed toidentify proteins of the chloroplast that selectively bindCoA and, thus, presumably depend on an interaction withCoA for their function. A spinach chloroplast extract wasclarified by centrifugation, precipitated in ammonium sul-fate to remove small molecules, including endogenousCoA and its acyl forms, and then applied to a CoA-Sepharose column (see "Materials and Methods"). Afterwashing the column until A280 reached baseline and at least95% of the loaded protein flowed through (e.g. the largesubunit of Rubisco; Fig. 1, lane 2), 1 mM CoA was appliedand 1-mL fractions (Fig. 1, lanes 3-11) were collected. Theprofiles of proteins eluted with CoA were analyzed bySDS-PAGE and colloidal gold staining. The fractions con-taining a 30-kD protein—one of the last to elute (starting at0.6-0.7 mM CoA) from the matrix and, thus, with a verystrong affinity for CoA—were combined, run on a secondSDS-PAGE gel, and transferred to a PVDF membrane forsequencing. The 30-kD protein was digested with trypsin,and one peptide that was clearly resolved by HPLC yieldedthe amino acid sequence DVRLYYGA. A database searchidentified a kiwifruit cDNA clone (pKIWI502, accession no.L27809; Ledger and Gardner, 1994) encoding a 36-kD pro-tein that contained a block of eight residues, that are iden-tical to the tryptic peptide released from the spinach 30-kDprotein. Assuming random amino acid usage, the proba-bility of this match is 20 ~8, and considering the composi-tion of the peptide it is considerably less. An additional

HSFT CoA Eluate

-«30kD

-«18kD

1 2 3 4 5 6 7 8 9 1 0 1 1

Figure 1. Affinity chromatography identifies a 30-kD protein thatstrongly binds CoA. A chloroplast high-speed extract (HS, lane 1) wasapplied to a CoA-Sepharose column and the flow-through (FT, lane2) was collected. The column was washed until the absorbancereached baseline, and then 1 mM CoA was applied. One-milliliterfractions were collected. The fractions were analyzed by SDS-PACEand colloidal gold staining. Lanes 3 to 11 show the profile of proteinsthat bound to the column and selectively eluted with CoA. Thepositions of three proteins are indicated: the large subunit of Rubisco(LSU), the 30-kD protein, and an 18-kD protein. www.plantphysiol.orgon May 28, 2020 - Published by Downloaded from



1644 Yang and Lamppa Plant Physiol. Vol. 112, 1996

search of the database with the full amino acid sequence ofthe kiwifruit protein did not reveal extensive similarity toany known protein. However, one feature of the kiwifruit36-kD protein was noted as significant. It contained anN-terminal region with characteristics of a transit peptide,which we predicted should be present given that the spin-ach 30-kD CoA-binding protein originated from chloro-plast extracts.

To investigate whether pKIWI502 coded for a polypep-tide that could be imported into chloroplasts, a 1.0-kbfragment of the cDNA that was equivalent to the entireopen reading frame was cloned into a T7 expression vector(Studier et al., 1990), and the recombinant [35S]Met-radiolabeled 36-kD protein was synthesized in £. coli. Itwas recovered from inclusion bodies by solubilization in 8M urea (Fig. 2, lane 3). Import reactions were carried out byincubating the labeled recombinant protein with spinachchloroplasts. The 36-kD protein was found primarily in themembrane fraction at the end of the reaction, and it wassensitive to thermolysin treatment, indicating that it hadnot crossed the envelope (Fig. 2, compare lanes 6 and 7). Incontrast, a thermolysin-resistant product of approximately30 kD was found in the soluble fraction, indicating success-ful import and processing (Fig. 2, lanes 8 and 9). The originof the minor, more slowly migrating bands that we ob-served in the soluble fraction is unknown, but they couldbe the result of either incomplete precursor cleavage orpost-import modification of the 30-kD protein. When the36-kD protein was added to an organelle-free reaction(Abad et al., 1988) optimized for cleavage by the chloro-plast processing enzyme, i.e. transit peptidase, a 30-kDprotein was also produced (Fig. 2, lanes 4 and 5). In theprocessing reactions the precursor for ACP was included totest for nonspecific proteolysis, which can occur if prepa-ration of the chloroplast extract inadvertently severely dis-rupts the thylakoid membranes (Abad et al., 1988). Only

Organelle-free rxn Import"~- + '• + 7 'M MT s ST'

"KIWIS02

preACP •>

holoACP*-

apoACP »-

Washs

1 2 3 4 5 6 7 8 9

Figure 2. Organelle-free processing and import by spinach chloro-plasts of the polypeptide encoded by pKIWI502. Lanes 1 and 2,Products of processing of preACP. Lane 3, The radiolabeled kiwifruitpolypeptide recovered from E. coli. Lanes 4 and 5, Products oforganelle-free processing (duplicates) of the kiwifruit polypeptide. +and - indicate the addition or absence of a chloroplast solubleextract. Lanes 6 and 8, Membrane (M) and soluble (S) products ofimport from chloroplasts untreated with thermolysin. Lanes 7 and 9,Products of import from chloroplasts treated with thermolysin (MTand ST, respectively) following the reaction. Unlabeled arrowheadindicates the 30-kD import product.

1 2 3 4 5 6

Figure 3. The recombinant kiwifruit 36-kD protein synthesized in E.coli binds to CoA-Sepharose. The polypeptide encoded by pKIWI502was recovered from inclusion bodies in 6 M urea and then graduallydialyzed to promote refolding (see "Materials and Methods"). It wasincubated with CoA-Sepharose, and then the unbound protein frac-tion (flow through, FT, lane 1), three washes (#1-#3, lanes 3-5), andthe eluate released by 1 mM CoA (El, lane 2) were analyzed byimmunoblotting, along with the recombinant 36-kD protein control(Kiwi 502, lane 6). The polyclonal antibodies were prepared againstthe purified recombinant kiwifruit protein.

apoACP and holoACP (produced by an active chloroplastholoACP synthase) were observed (Fig. 2, lane 2), showingthe specificity of the assay. We conclude that pKIWI502codes for a 36-kD polypeptide that is the precursor of achloroplast soluble protein of 30 kD.

To determine if the protein encoded by pKIWI502 is ableto bind CoA, the unlabeled 36-kD protein was synthesizedin E. coli, isolated from inclusion bodies, and then incu-bated with CoA-Sepharose using a batch chromatographyprotocol. The conditions were similar to those used topurify the spinach 30-kD protein by column chromatogra-phy (see "Materials and Methods"). Rabbit antibodiesraised against the kiwifruit recombinant protein were usedin immunoblotting experiments to analyze the chromato-graphic fractions. The immunoblot in Figure 3 included thekiwifruit protein recovered from E. coli (lane 6), the fractionthat did not bind to the CoA-Sepharose, referred to as theflow-through (lane 1), and the protein collected in each ofthree buffer B washes (lanes 3-5) of the matrix. In the thirdwash no protein was detected. However, when 1 mM CoAwas added to the elution buffer, the kiwifruit protein wasreleased from the CoA-Sepharose matrix (Fig. 3, lane 2).Therefore, the kiwifruit protein recognizes and specificallybinds CoA. The seemingly low efficiency of binding mayresult from the presence of the transit peptide and/or thefact that the protein was recovered from inclusion bodiesand, thus, may not be entirely refolded. Taken together, theevidence strongly indicates that the kiwifruit 30-kD "ma-ture" protein encoded by pKIWI502 and the spinach 30-kDprotein isolated by CoA-affinity chromatography are chlo-roplast homologs.

We examined whether a gene corresponding to the ki-wifruit 36-kD protein could be identified in spinach. www.plantphysiol.orgon May 28, 2020 - Published by Downloaded from



Chloroplast CoA-Binding Protein and Peroxisomal Thiolases 1645

Digoxigenin-labeled pKIWI502 was used to probe a South-ern blot of spinach genomic DNA digested with BamHI,Hindlll, and EcoRl. A single band of 2.2 kb was detected inthe spinach digest under conditions in which the probehybridized to an approximately 15-kb fragment of kiwi-fruit genomic DNA and with the expected fragments frompKIWI502 (Fig. 4). To see if a monocot contains a relatedgene, wheat genomic DNA was also analyzed. A high-molecular-weight band hybridized with the probe. Boththe wheat and kiwifruit DNAs were only partially di-gested. Nevertheless, the Southern blot shows that besideskiwifruit, a gene(s) related to pKIWI502 is present in spin-ach as well as in wheat.

Figure 1 demonstrates that the conditions for purifica-tion of the spinach 30-kD protein are highly selective. Onlya small number of proteins (eight) remain bound to theCoA-affinity column and were specifically eluted withCoA, suggesting that they require interaction with a CoAmoiety for function. Similarly, the kiwifruit protein exhib-its an affinity for CoA. Enzymes in this category that areneeded for either fatty acid synthesis or /3-oxidation oftenhave a special Cys residue at the active site, where itparticipates in acyl group transfer (Siggaard-Andersen,1993). Therefore, the region surrounding the single Cysresidue that is found in the 30-kD protein encoded bypKIWI502 was compared with the active sites of knownCoA-dependent enzymes. Surprisingly, we found (Fig. 5)that the Cys residue lies within a 189-amino-acid domainwith significant conservation among the peroxisomalj8-ketoacyl-CoA thiolases, including the enzyme frommango (Bojorquez and Gomez-Lim, 1995), cucumber(Preisig-Muller and Kindl, 1993), and rat (Hijikata et al.,1990), which carry out the degradative step in fatty acidjS-oxidation. Although amino acid sequence similarity istoo low (approximately 35% in the 50 amino acids imme-diately surrounding the Cys, with 26% identity; approxi-

1 £ 2US Q.

i4.0kb2.2kb

0.7kb

1 2 3 4Figure 4. Southern blot analysis detects a related gene sequence inspinach and wheat. A digoxigenin-labeled probe, pKIWI502, wasused in the hybridization reaction. Cenomic DNAs (Spinach, lane 1;Wheat, lane 2; and Kiwi, lane 3) were cut with Ba/nHI, H/ndlll, andfcoRI. The pKIWI502 DNA was cut with fcoRI and Ncol (lane 4).

46kiwi SAILQDTAIRQDTYIWTPVPISRVLPAAAESLFKVIVDLSRSPD LVcucum. CAAGDSASYQRTSVFGDDWIVAAYRTAICK SKRGGFKDTYPDDLLmango CAAGDSASYHRASVYGDDWIVAAHRTALCK SKRGGFKDTYPDDLLrat p. C SAGFPQAS ASDWWHGQRTPIGR AGRGGFKDTTPDELLE . coli MEQWIVDAIRTPMGRSKGGAFRNVRAEDLSArat m. MALLRGVFIVAAKRTPFG AYGGLLKDFTATD LT

94kiwi YNFVSPGQYVQIRIPEAIVNPPPRPAYFYIASPPSLVKKNLEFEFLIRcucum. APVLKALIEKTNLNPSEVGDIVVGSVLAPGSQRASECRMAAFYAGFPEmango APVLKALIEKTNLDPSEVGDIWGTVLAAGSQRASECRMAAFYAGFPDrat p. SAVLTAVLQDVKPKPECLGDISVGNVLQPGAGAA MARIAQFLSGIPEE . coli HLMRSLLARNPALEAAALDDIYWGCVQQTLEQGFNIARNAALLAEVPHrat m. EFAARAALSAGKVPPETIDSVIVGNVMQSSSDAAYLARHVGLRVGVPT

* 141kiwi SVPGTTSEVLCSLKEGDWDLTQIIGRGF DIEQILPPEDYPTVLISVcucum. TVPVRTVNRQCSSGLQAVADVAAAIRAGFYDIGIGAGLESMTTNPMAWmango TVPLRTVNRQCSSGLQAVADVAAAIKAGFYDIGIGAGLESMTANPMAWrat p. TVPLSAVNRQCSSGLQAVANIAGGIRNGSYDIGMACGVESMTLSERGNE.coli SVPAVTVNRLCGSSMQALHDAARMIMTGDAQACLVGGVEHMGHVPMSHrat m. ETGALTLNRLCGSGFQSIVSGCQEICSKDAEWLCGGTESMSQSPYSV

189kiwi TGYGMSAGRSFIEEGFGANKRSDVRLYYGAENLETMGYOERFKDWE..cucum. EGSVNPKVKEFEQARNCLLPMGITSENVAHRFGVTRQEQDQAAVES..mango EGSVNPRVKSIENAQNCLLPMGVTSENVAQRFGVSREKQDQAAIES..rat p. PGNISSRLLENEKARDCLIPMGITSENVAERFGISRQKQDAFALAS..E.coli GVDFHPGLSRNV AK AAGMMGLTAEMLARMHGISREMQDAFAARS . .rat m. RNVRFGTKFGLDLKLEDTLWAGLTDQHVKLPMGMTAENLAAKYNIS . .

Figure 5. N-terminal alignment of the sequence of the 30-kD kiwi-fruit "mature" protein deduced from pKIWI502 with the family ofdegradative thiolases. Alignment is relative to the active-site Cys,which is acylated (*). The eight amino acids that are identical to thetryptic peptide recovered from the spinach 30-kD protein are under-lined. The peroxisomal thiolase sequences are from cucumber (cu-cum.; Preisig-Muller and Kindl, 1993), mango (Bojorquez andComez-Lim, 1995), and rat (rat p.; Hijikata et al., 1990). The E. colithiolase (gene fadA; Yang et al., 1990) and rat mitochondrial (rat m.;Arakawa et al., 1987) sequences are also given. The single-letteramino acid code is used. Conserved residues are shown in boldface.The amino acid number is given at the right, relative to the kiwifruitprotein from the -1 position, relative to the predicted cleavage site(T) of the precursor.

mately 22% in a 189-amino-acid region; and thereafter morescattered) to be recognized by a general database search, thenearly exact alignment of the conserved residues, with al-most no gaps, strongly suggests that the chloroplast 30-kDprotein is related to this group of enzymes. Similarity is alsofound with /3-ketoacyl-CoA thiolase (the fadA gene product)from E. coli (Yang et al., 1990) and the rat mitochondrialdegradative thiolase (Fig. 5) (Arakawa et al., 1987). It ex-tends to other members of the thiolase family, including thebiosynthetic acetoacetyl-CoA thiolases (see "Discussion")(Igual et al., 1992). A comparison with the KAS used in fattyacid synthesis showed that the chloroplast 30-kD proteinshares some sequence relatedness around the Cys residue(PLSYDITAACSGFMLGLVSAACHVRGG; residues shownin boldface are conserved) with spinach KAS III (Tai and www.plantphysiol.orgon May 28, 2020 - Published by Downloaded from



1646 Yang and Lamppa Plant Physiol. Vol. 11 2, 1996

Jaworski, 1993), which utilizes acetyl-COA in the first condensation step.

The active-site Cys for acyl group transfer in the mango and cucumber peroxisomal P-ketoacyl-COA thiolases is at position 104, inferred from biochemical and mutational studies of the closely related biosynthetic thiolases (Raaka and Lowenstein, 1979; Thompson et al., 1989). Alignment of the kiwifruit chloroplast protein with the plant peroxi- soma1 thiolases places the N terminus of the mature protein 103 amino acids upstream of the conserved Cys, therefore suggesting that cleavage of the 36-kD precursor polypeptide occurs between Ser5' and Ala59 during import and in the organelle-free processing assay. That is, an approximately 6-kD N-terminal sequence would be removed, in agreement with the predicted size of the transit peptide cleaved from the 36-kD precursor in these reactions. The processed, mature 30-kD protein has a predicted pI of 4.8.

Dl SCU SSl ON

We have presented evidence that a 30-kD COA-binding protein purified from spinach chloroplast extracts corre- sponds to a protein encoded by a kiwifruit cDNA gener- ated (Ledger and Gardner, 1994) from RNA that was synthesized during early fruit development, a period of active cell division. We show that the kiwifruit protein synthesized in E. coli contains a functional transit peptide, which targets it to the chloroplast, where it is cleaved, yielding a 30-kD protein. Significantly, it was also able to bind COA in a specific manner. The deduced amino acid sequence of the kiwifruit protein shows significant sequence relatedness to the peroxisomal P-ketoacyl-COA thiolases in an N-terminal domain. Most importantly, it contains an active-site Cys surrounded by conserved residues found in these thiolases that are nearly identically spaced. Acyl groups are thought to be attached to the Cys via a thioester linkage during the final cleavage step of fatty acid P-oxidation before transfer to COA (Masamune et al., 1989).

The degradative thiolases utilized for fatty acid P-oxidation actually belong to a Iarger group that includes the biosynthetic thiolases, which catalyze the reversible condensation of two acetyl-COA molecules to produce acetoacetyl-COA. Acetoacetyl-COA is used in some prokaryotes to produce poly-p-hydroxybutyrate (see Peoples and Sinskey, 1989). In plants it is also the precursor for 3-hydroxy-3-methylglutaryl-COA during isoprenoid synthesis, but the enzymes that are necessary for the first steps of this pathway have not been well characterized (Mc- Garvey and Croteau, 1995). More distantly related enzymes to the thiolases include KAS family members that catalyze the initial condensation and elongation steps of fatty acid synthesis (Siggaard-Andersen, 1993), the production of polyketide antibiotics (Bibb et al., 1989), and the acylation of oligosacharrides that are used in signaling during legume nodulation (Fisher and Long, 1992).

A phylogenetic analysis of members of the thiolase family from animals, fungi, and prokaryotes shows strong conservation, extending from 32 to 86% sequence identity (Igual et al., 1992). Overall, the mango, cucumber, and rat peroxisomal degradative thiolases show 55% identity. That

the kiwifruit 30-kD protein is considerably less conserved-only 22% similarity with the plant peroxisomal enzymes over 189 amino acids at the N terminus-suggests that it participates in an acyl transfer reaction distinct from these thiolases, with a different substrate specificity. In addition, the pI of the thiolases is very basic (pI = 9), whereas the pl of the kiwifruit 30-kD protein equals 4.8. A11 members of the thiolase family have a very basic C-terminal domain, which is absent in the chloroplast 30-kD protein. This C-terminal domain is highly conserved (Igual et al., 1992) and contains a second Cys that is needed for the thiolase ping-pong reaction, producing first the acyl-S-enzyme intermediate and then acyl-COA (see the introduction). Mutation of Cys378 in a prokaryotic thiolase results in a 50,000-fold decrease in thiolysis compared with the wild-type enzyme (Masamune et al., 1989; Palmer et al., 1991). Since the cDNA coding for the kiwifruit 36-kD protein was prepared by priming poly(A)+ RNA with oli- go(dT) and contains a 3' nontranslated region (Ledger and Garner, 1994), the primary sequence deduced from it includes the C terminus. The size of the mature protein after import of the 36-kD precursor into chloroplasts also agrees with that of the spinach 30-kD protein that was purified by COA-affinity chromatography. Therefore, if the chloroplast 30-kD protein from kiwifruit, and its homolog in spinach, performs a thiolase-related role, as predicted from (a) the alignment of the conserved residues surrounding the acyl- accepting Cys shown in Figure 5 and (b) its ability to bind COA, another subunit would be required to supply the second critica1 Cys for full activity. The recent production of antibodies against the kiwifruit protein should now allow an examination of its native structure.

In summary, a nove1 chloroplast protein has been identified in spinach by its ability to strongly bind COA; a homolog has been found in kiwifruit. The N-terminal domain of the kiwifruit protein, including a special Cys, resembles the peroxisomal degradative thiolases, as well as related members of this family. The chloroplast protein in both plants most probably participates in a CoA- dependent acyl transfer reaction that is distinct from, although mechanistically similar to, the well-characterized degradative thiolases for the reasons outlined above. It is interesting to note that in recent experiments using the affinity chromatography approach described in this study, we have isolated another spinach chloroplast COA-binding protein (L.-M. Yang and G. Lamppa, unpublished results) that contains a 20-amino-acid block, which shows 50% identity with E . coli 3-hydroxyacyl-COA dehydrogenase (DiRusso, 1990), the third enzyme-out of the four-required for fatty acid P-oxidation. The question arises, is there a plastid P-oxidation pathway used to reduce the chain length of a special subpopulation of acyl chains under certain conditions? An important future objective is to understand the biochemical role that the 30-kD protein plays in the plastid by the development of enzymatic assays to establish the nature of its substrates. This study using kiwifruit, showing maximal expression of the 30-kD protein gene in young fruit (Ledger and Richards, 1994), may provide a clue to the source of tissue that is required to develop these assays.



Chloroplast COA-Binding Protein and Peroxisomal Thiolases 1647

ACKNOWLEDGMENTS

We thank Drs. Richard Gardner and Susan Ledger (Auckland University) for generously providing us with the kiwifruit cDNA clone (pKIWI502), and the National Plant Lipid Cooperative for supplemental funds. In addition, the excellent research assistance of Hong Ma and the greenhouse staff (Department of Ecology and Evolution) was greatly appreciated.

Received April 25, 1996; accepted August 22, 1996. Copyright Clearance Center: 0032-0889/96/112/1641/07.

LITERATURE ClTED

Abad MS, Clark SE, Lamppa GK (1988) Properties of a chloroplast enzyme that cleaves the chlorophyll a / b binding protein precursor. Plant Physiol 90: 117-124

Abad MS, Oblong JE, Lamppa GK (1991) Soluble chloroplast enzyme cleaves preLHCP made in E. coli to a mature form lacking a basic N-terminal domain. Plant Physiol 96: 1220-1227

Arakawa H, Takiguchi M, Amaya Y, Nagata S, Hayashi H, Mori M (1987) cDNA-derived amino acid sequence of rat mitochondrial 3-oxoacyl-COA thiolase with no transient presequence: structural relationship with peroxisomal isozyme. EMBO J 6:

Bibb MJ, Biro S , Motanmedi H, Collins JF, Hutchison CR (1989) Analysis of the nucleotide sequence of the Strepfomyces glauce- scens tcml genes provides key information about the enzymol- ogy of polyketide antiobiotic biosynthesis. EMBO J 8: 2727-2736

Bojorquez G, Gómez-Lim MA (1995) Peroxisomal thiolase mRNA is induced during mango fruit ripening. Plant Mo1 Biol 28:

DiRusso CC (1990) Primary sequence of the Escherichia coli fadBA operon, encoding the fatty acid-oxidizing multieneyzme complex, indicates a high degree of homology to eucaryotic enzymes. J Bacteriol 172: 6459-6468

Fernandez MD, Lamppa GK (1990) Acyl carrier protein (ACP) import into chloroplasts does not require the phosphopantetheine: evidence for a chloroplast holo-ACP synthase. Plant Cell2:

Fernandez MD, Lamppa GK (1991) Acyl carrier protein import into chloroplasts. Both the precursor and mature forms are substrates for phosphopantetheine attachment by a soluble chloroplast holoACP synthase. J Biol Chem 266: 7220-7226

Fisher RF, Long SR (1992) Rhizobium-plant signal exchange. Na- ture 357: 655-660

Gehring U, Harris JI (1970) Active site cysteines of thiolase. Eur J Biochem 16: 492-498

Hijikata M, Wen J-K, Osumi T, Hashimoto T (1990) Rat peroxi- soma1 3-ketoacyl-COA thiolase gene. J Biol Chem 265: 4600-4606

Igual JC, Gonzalez-Bosch C, Dopazo J, Perez-Ortin JE (1992) Phylogenetic analysis of the thiolase family: implications for the evolutionary origin of peroxisomes. J Mo1 Evol 35: 147-155

1361-1366

811-820

195-206

Kindl H (1993) Fatty acid degradation in plant peroxisomes: function and biosynthesis of the enzymes involved. Biochimie 75:

Ledger SE, Gardner RC (1994) Cloning and characterization of five cDNAs for genes differentially expressed during fruit development of kiwifruit (Actinidia deliciosa var. deliciosa). Plant Mo1 Biol 25: 877-886

Masamune S, Palmer AJ, Gamboni R, Thompson S, Davis JT, Williams SF, Peoples OP, Sinskey AJ, Walsh CT (1989) Bio- claisen condensation catalyzed by thiolase from Zoogloea ramigera. J Am Chem SOC 111: 1879-1881

McGarvey DJ, Croteau R (1995) Terpenoid metabolism. Plant Cell

Miyazawa S, Furuta S , Osumi T, Hashimoto T, Ui N (1981) Properties of peroxisomal 3-ketoacyl-COA thiolase from rat liver. J Biochem 90: 511-519

Ohlrogge J, Browse J (1995) Lipid biosynthesis. Plant Cell 7:

Palmer MAJ, Differding E, Gamboni R, Williams SF, Peoples OP, Walsh CT, Sinskey AJ, Masamune S (1991) Biosynthetic thiolase from Zoogloea ramigera. J Biol Chem 266: 8369-8375

Peoples OP, Sinskey AJ (1989) Poly-hydroxbutyrate biosynthesis in Alcaligenes eutrophus H16. J Biol Chem 264: 15293-15297

Preisig-Miiller R, Kindl H (1993) Thiolase mRNA translated in uitro yields a peptide with a putative N-terminal presequence. Plant Mo1 Biol 2 2 59-66

Raaka BM, Lowenstein JM (1979) Inhibition of fatty acid oxidation by 2-bromooctanoate. J Biol Chem 254 6755-6762

Siggaard-Andersen M (1993) Conserved residues in condensing enzyme domains of fatty acid synthases and related sequences. Protein Sequences Data Ana1 5: 325-335

Studier FW, Rosenber AH, Dunn JJ, Dubendorff JW (1990) Use of T7 RNA polymerase to direct expression of cloned genes. Meth- ods Enzymol 185: 60-89

Stumpf PK (1987) The biosynthesis of saturated fatty acids. In PK Stumpf, EE Conn, eds, The Biochemistry of Plants, Vol 9. Aca- demic Press, New York, pp 121-136

Tai H, Jaworski JG (1993) 3-Ketoacyl-acyl carrier protein synthase 111 from spinach (Spinacia oleracea) is not similar to other condensing enzymes of fatty acid synthase. Plant PhysiollO3: 1361- 1367

Thompsom S, Mayer F, Peoples OP, Masamune S, Sinskey AJ, Walsh C (1989) Mechanistic studies on P-keto thiolase from Zoogloea ramigera: identification of the active site nucleophile as cys89, its mutation to ser89, and kinetic and thermodynamic characterization of wildtype and mutant enzymes. Biochemistry

Yang L-M, Fernandez MD, Lamppa GK (1994) Acyl carrier protein (ACP) import into chloroplasts: covalent modification by a stromal holoACP synthase is stimulated by exogenously added COA and inhibited by 3',5'-ADP. Eur J Biochem 224: 743-750

Yang S-Y, Yang X-YH, Healy-Louie G, Schulz H, Elizinga M (1990) Nucleotide sequence of the fadA gene. J Biol Chem 265:

225-230

7: 1015-1026

957-970

28: 5735-5742

10424-10429



Documents

Identification of a Chloroplast Coenzyme A-Binding Protein ...Spinach (Spinacea olevacea cv Melody, Stocks Seeds, Buf- falo, NY) was grown in a greenhouse at 22°C under cool fluorescent