Properties of a Chloroplast Enzyme that Cleaves Binding ... · PROPERTIES OF CHLOROPLAST PROCESSING ENZYME THAT CLEAVES pLHCP ColumnChromatography The crude soluble extract was centrifuged

Plant Physiol. (1989) 90, 117-1240032-0889/89/90/011 7/08/$01 .00/0

Received for publication November 21, 1988and in revised form December 21, 1988

Properties of a Chloroplast Enzyme that Cleaves theChlorophyll a/b Binding Protein Precursor1

Optimization of an Organelle-Free Reaction

Mark S. Abad, Steven E. Clark, and Gayle K. Lamppa*

Department of Molecular Genetics and Cell Biology, The University of Chicago, Chicago, Illinois 60637

ABSTRACT

The major light-harvesting chlorophyll a/b binding protein(LHCP) of higher plant chloroplasts is nuclear-encoded, synthe-sized as a precursor, and processed upon import. We havepreviously (GK Lamppa, M Abad [1987] J Cell Biol 105: 2641-2648) identified a soluble enzyme that cleaves the LHCP precur-sor (pLHCP). In this study, we describe the conditions for optimalrecovery of the processing activity and provide evidence that theN terminus of pLHCP is indeed cleaved, removing the transitpeptide. Two pLHCP deletions were made from a cloned pLHCPgene removing 13 and 21 amino acids, respectively, from thecarboxy terminus of the protein. After organelle-free processing,the cleavage products showed a shift in mobility during SDS-PAGE proportional to the size of the precursor truncations, aspredicted for N-terminal processing. Unexpectedly, a third trun-cated precursor lacking 91 residues of the C-terminus was notcleaved although the transit peptide domain was intact, suggest-ing that this deletion disrupted conformational features of theprecursor necessary for processing. The pLHCP processing en-zyme is inhibited by 2 millimolar EDTA and the metal chelator1,10 phenanthroline at 0.4 millimolar, while being inhibited byEGTA only at high concentrations and insensitive to iodoacetate.Optimal processing occurs at pH 8 to 9, and 260C. Gel filtrationchromatography shows that the pLHCP processing enzyme hasan apparent molecular weight of about 240,000. The identicalcolumn fractions that process pLHCP also convert the precursorof the small subunit of ribulose-1,5-bisphosphate carboxylase toits mature form.

The majority of chloroplast proteins are nuclear-encodedand synthesized as precursor polypeptides with N-terminaltransit peptides that are proteolytically cleaved upon importinto the organelle (for a review see ref. 26). The diversity ofchloroplast processing enzymes required for the maturationof the large number of imported proteins has not been estab-lished. A stromal protease has been partially purified frompea chloroplasts that cleaves the precursor ofthe small subunit

'This work was supported in part by a U.S. Department ofAgriculture grant (8701037) and a National Institutes ofHealth award(GM36419-01) awarded to G. K. L., and 1988 Sigma Xi Grants-in-Aid for Research awarded separately to both M. S. A. and S. E. C.

of pS,2 a stromal protein, to its mature form (24, 25) as wellas the precursor of plastocyanin, a thylakoid lumen protein,to an intermediate size. The intermediate is processed to itsfinal size by a thylakoid membrane-associated protease as itis translocated into the lumen (11, 14). A major questionregarding import has been whether the same enzyme is re-quired for the initial processing of all precursors as they enterthe stroma, or if there are classes of proteases that recognizedifferent subsets depending on their structure and final desti-nation. In mitochondria, a matrix metalloprotease, responsi-ble for cleavage of a large number of imported precursorsdestined for different compartments, has recently beenpurified ( 12).Our studies have focused on the import and processing of

the major LHCP, an integral thylakoid membrane proteinprimarily associated with PSII in higher plant chloroplasts(27). LHCP is synthesized with a 34 to 37 amino acid transitpeptide (5, 10, 15, 18) that facilitates LHCP import, but doesnot contain the signal for routing to the thylakoids (16). Wehave recently identified a soluble processing enzyme in bothpea and wheat chloroplasts that processes the LHCP precursor(pLHCP), releasing a 25 kD peptide that comigrates duringSDS-PAGE with the smaller form of mature LHCP producedupon import of pLHCP into intact chloroplasts (17). Thoseresults indicated that the processing enzyme is an endopepti-dase, but it remained to be determined if cleavage occurredat the N- or C-terminus. Multiple forms of LHCP are foundin vivo in the thylakoids and their origin has not been entirelyresolved. Precedence for COOH terminal processing has beenobtained for the chloroplast encoded, herbicide binding, 32kD protein (Dl) as it enters the granal thylakoids (20). SinceLHCP is also found mainly in the grana, the possibility wasinvestigated that, in addition to the removal of the transitpeptide, pLHCP was cleaved by a special class of COOHterminal processing enzymes.

In this study, we provide evidence that the soluble enzymeindeed cleaves the N terminus ofpLHCP. A region ofpLHCPis removed that is sufficient to code for the transit peptide.These results suggest that the processing enzyme functions asa pLHCP transit peptidase in vivo, or plays an important role

2 Abbreviations: pS, ribulose-1,5-bisphosphate carboxylase/oxy-genase small subunit precursor; pLHCP, light-harvesting Chl a/bbinding protein precursor; HSM, 50 mm Hepes-KOH (pH 8), 0.33 Msorbitol, 8 mM methionine.

117

www.plantphysiol.orgon February 21, 2020 - Published by Downloaded from Copyright © 1989 American Society of Plant Biologists. All rights reserved.

http://www.plantphysiol.org

Plant Physiol. Vol. 90,1989

in modifying the N-terminal structure of LHCP. Key prop-erties of the pLHCP processing enzyme are described andcompared with those of the stromal enzyme identified byRobinson and Ellis (24, 25) that processes pS. The samefractions of the soluble extract isolated by gel filtration chro-matography process both pLHCP and pS. We discuss thehypothesis that a general protease may be responsible for theinitial cleavage of precursors that enter the organelle andassemble into different functional complexes.

MATERIALS AND METHODS

Plant Growth and Purification of Chloroplasts

Pea (Pisum sativum, Laxton's Progress) plants were grownin a greenhouse (Department of Ecology and Evolution, Uni-versity ofChicago) and transferred to a growth chamber (26°C,cool fluorescent lights) 1 to 6 d after planting. Chloroplastswere isolated from leaf tissue of 8 d old seedlings on Percollgradients (1). A typical chloroplast isolation used 20 g of freshweight leaf tissue and yielded 3 to 4 mL of chloroplastssuspended in HSM buffer (50 mM Hepes [pH 7.5], 0.33 Msorbitol, 10 mm methionine) at a Chl content of 600 to 900jsg/mL.

Plasmid Construction

A cDNA sequence (W9) coding for wheat pS (4) wasreleased from pBR322 with PstI and inserted into pSP65 in a5'-3' orientation downstream of the SP6 promoter, yieldingthe plasmid pSP65-pS. A hybrid pea pS gene was constructedfrom an intron-containing genomic clone, E9 (8) and a cDNAclone, pSS 15 (7), in a two-step process. A HindIII-SphI restric-tion fragment containing the E9 transit peptide was insertedinto the transcription vector pIBI31 (International Biotech-nologies Inc.) 3' of the T7 promoter to produce the clone T7-E9T. The vector was linearized with SphI and PstI, and aSphI-PstI restriction fragment encoding the mature proteinfrom pSS 15 was inserted into T7-E9T in frame and creatingno new codons, producing the construct T7-E9/SS 15. A SphI-PstI restriction fragment coding for the mature protein fromthe wheat cDNA W9 was similarly introduced into T7-E9Tyielding T7-E9/W9.

Deletions were made in SP65-pLHCP, a wheat genomicclone coding for pLHCP (18, 19) previously inserted into thetranscription vector pSP65 (17), that progressively removedthe carboxy terminal coding sequences. The SmaI site in thepolylinker was removed by digestion with BamHI and EcoRI,followed by treatment with mung bean nuclease (0.5 units/,ug DNA) and blunt end ligation to create the constructp(ASma)SP65-pLHCP. The A13 deletion ofpLHCP was con-structed from SmaI linearized p(ASma)SP65-pLHCP, whichcuts 6 nucleotides (18) 5' to the translation terminationcodon. The linearized plasmid was treated with TaqI methy-lase (New England Biolabs) leaving only one sensitive HincIlsite, 45 nucleotides 5' of the termination codon. The SmaIlinearized plasmid was then digested with HincIl and the large(vector containing) fragment was separated from the smaller39 nucleotide fragment by agarose gel electrophoresis and wasisolated by electroelution. The purified, blunt-ended fragment

was ligated with T4 DNA ligase (Boehringer Mannheim Bio-chemicals) and used to directly transform Escherichia colistrain HB101. This deletion resulted in the loss of 13 aminoacids from position 252 through 264 in the precursor leavingthe C-terminal glycine-lysine (residues 265 and 266) ofLHCP(18). To generate the A27 deletion, SmaI linearizedp(ASma)SP65-pLHCP was digested with ApaI, which cuts 87nucleotides upstream of the termination codon, followed bymung bean nuclease digestion and ligation. This removed an81 nucleotide fragment, resulting in the loss of amino acids238 to 264. The A91 construct was made by digesting theXmaI linearized p(ASma)SP65-pLHCP with KpnI, which cuts277 nucleotides 5' to the termination codon. A 272 nucleotidefragment coding for amino acid residues 174 to 264 wasremoved. After mung bean nuclease treatment, the sampleswere prepared for ligation and transformation as describedabove. In these three constructs, no new amino acid codonswere added or created at the sites of deletion.

In Vitro Transcription and Translation

Plasmid constructs were prepared for in vitro transcriptionby linearization at a restriction site 3' of the precursor gene.pSP65-pLHCP and pSP65-pS were digested with HindIIl andused to program in vitro transcription reactions with SP6RNA polymerase (Promega Biotech) to generate syntheticprecursor mRNAs, whereas T7-E9/SS 15 and T7-E9/W9 weredigested with EcoRI and incubated with T7 RNA polymerase.In vitro translation reactions using either a wheat germ lysate(Amersham Corp.) for E9/SS1 5 and pS, or a rabbit reticulo-cyte lysate (Bethesda Research Laboratories) for pLHCP andE9/W9, were programmed with the synthetic mRNA in thepresence of [35S]methionine to synthesize labeled precursorsaccording to the supplier's directions.

Organelle-Free Processing Assay

Labeled precursor proteins were incubated with a pea chlo-roplast soluble extract optimized for processing (see "Re-sults"). To prepare the extract, chloroplast suspensions inHSM buffer were spun at 1,600g for 1 min, gently resuspendedin a volume of ice cold 5 mm Hepes-KOH (pH 8.0) equal tothe original HSM volume, and kept at 4°C for 30 min. Thelysed chloroplasts were centrifuged at 16,000g for 10 min toremove the bulk membranes and the total soluble extract waseither assayed directly or after centrifugation at 137,000g toremove membrane vesicles. A typical 25 ,uL processing reac-tion contained 20 mM Hepes (pH 8.0), 1 ,ug/mL chloram-phenicol, 3,L ofa 30,uL in vitro translation reaction (25,000-30,000 cpm), and 15,L of soluble extract. The reactions wereincubated for 90 min at 26°C, and stopped by bringing theproducts to 60 mm DTT, 60 mm sodium carbonate, 12%sucrose, 0.04% bromophenol blue, and 2% SDS. The sampleswere then boiled for 1 min, and analyzed on SDS-polyacryl-amide gels as described (17). Processing was indicated by theappearance of a labeled species of 25 kD cleaved from theprecursor.

118 ABAD ET AL.



PROPERTIES OF CHLOROPLAST PROCESSING ENZYME THAT CLEAVES pLHCP

Column Chromatography

The crude soluble extract was centrifuged at 137,000g toremove membrane vesicles and was chromatographed on aSephacryl S-300 superfine (Pharmacia) gel filtration column(1.5 x 120 cm) in 5 mm Hepes (pH 8.0) after preequilibrationof the resin with excess bovine serum albumin to saturatesites of nonspecific protein adsorption. The bed volume ofthe column equalled 198 mL and the void volume wasdetermined to be 71 mL using Blue dextran (Sigma ChemicalCo.). The elution volumes of the following mol wt standardswere measured: thyroglobulin (Mr = 667,000), urease(545,000 and 272,000), ferritin (450,000), catalase (240,000),fl-amylase (200,000), alcohol dehydrogenase (150,000), phos-phorylase b (96,000), bovine serum albumin (66,000), andovalbumin (44,000). A calibration curve was obtained byplotting the Kay value [(Velution - VV1id)/( Vb-d - Vvoid)] againstthe log M, of each standard. The sample applied to thecolumn, which contained 2.7 mg of protein as determinedusing a Coomassie dye binding assay (Bio-Rad), in a volumeof 0.9 mL, was eluted using 5 mM Hepes (pH 8.0) at a flowrate of 10 mL/h at 4°C for 16 h. The fractions (1 mL/fraction)were assayed for processing activity by adding 15 ,uL aliquotsto the 25 ,uL processing reaction as described above.

RESULTS

Recovery of the pLHCP Processing Activity

Conditions ofchloroplast lysis were investigated to optimizethe recovery of the pLHCP processing enzyme. Labeled pre-cursor was synthesized in a reticulocyte lysate using syntheticRNA made from the pSP65-pLHCP template (see "Materialsand Methods"), and then incubated with a pea chloroplastsoluble extract. To prepare the extract, chloroplasts wereresuspended at a concentration of about 800 gtg Chl/mL inosmotic buffer (HSM), pelleted at 1600g for 1 min, and gentlylysed by resuspension in ice cold 5 mM Hepes (pH 8.0) atratios of 1: 1(800,ug Chl/mL of lysate), 2:1 (400,g CHL/mLof lysate), or 4:1 (200 ,ug Chl/mL of lysate) of the originalHSM volume. The chloroplasts were kept at 4°C for 30 min,centrifuged at 16,000g for 10 min to remove the bulk mem-branes and the crude soluble extract was assayed for process-ing activity. Extracts from the 1:1 and 2:1 resuspensionsprocessed pLHCP as indicated by the appearance of the 25kD peptide; however, when the chloroplasts were lysed by a4:1 resuspension in hypotonic buffer, there was a dramaticreduction in the amount of processing activity recovered (Fig.IA). To compensate for the dilution effect during hypotoniclysis, proportionally more extract was added to the appropri-ate processing reactions, and thus, the lower activity was notdue to the limiting concentration of a factor in the extract.Examination of the chloroplast resuspensions by light mi-croscopy showed that the chloroplasts are completely dis-rupted at a 4:1 dilution. In contrast, in the 1:1 and 2:1resuspensions the chloroplast thylakoid membranes appearedswollen, but still exhibited a loose organization (data notshown). We found it essential to resuspend the organelles onlybriefly and gently before removing the membranes by centrif-ugation or the activity was greatly diminished, even with the

A

p _

1 2 3 4

B

1 2 3 4 5Figure 1. Processing of wheat pLHCP by extracts of hypotonicallylysed pea chloroplasts. A, Synthetic pLHCP mRNA was translatedusing a reticulocyte lysate and labeled precursor was incubated withextracts of chloroplasts lysed with increasing amounts of hypotonicbuffer (5mM Hepes, pH 8.0). Radiolabeled pLHCP (lane 1) wasincubated with extract from chloroplasts lysed at 800 Mg Chl/mL (lane2), at 400 Ag Chl/mL (lane 3), and at 200 Mg Chl/mL (lane 4). B,Labeled precursor (lane 1) was incubated with a crude soluble extractfrom chloroplasts lysed at 800 Mg Chl/mL (lane 2). This extract wascentrifuged at 137,000g, the supernatant was removed, and thevesicle pellet was resuspended to the volume of the original lysatewith 5 mm Hepes (pH 8.0). The precursor was incubated with theclarified supernatant (lane 3), with the resuspended membrane vesi-cles (lane 4), and with an undiluted aliquot of the dispersed vesiclepellet (lane 5). The arrows point to the precursor (p) and mature (m)forms of pLHCP.

1:1 resuspension that yielded optimal processing of pLHCP.It is not clear at present why extensive lysis of the organellescauses a loss of processing activity in the extract. One possi-bility is that the complete disruption of the thylakoids, result-ing in the release of their contents, creates a change in thephysiological properties (e.g. pH, ion concentration) of thesoluble fraction sufficient to inhibit the processing.When the crude soluble extract was clarified of membrane

vesicles by ultracentrifugation, the major fraction of the proc-essing activity was recovered in the supernatant (17). Todetermine ifany ofthe processing activity remained associatedwith membrane vesicles the crude soluble extract was centri-fuged at 137,000g for 1 h and the supernatant was assayed ina typical processing reaction (Fig. 1B, lanes 2 and 3). Inaddition, an aliquot of the membrane pellet was resuspendedin a volume of 5 mm Hepes (pH 8.0) equal to the originalvolume of the extract, and this resuspension was assayed forprocessing activity of pLHCP. No processing activity wasfound associated with the vesicle pellet after resuspension(Fig. 1B, lane 4). In fact, addition of the dispersed pelletresidue directly to the organelle-free reaction resulted in deg-radation of the precursor (Fig. 1B, lane 5). We conclude thatthe processing enzyme is not associated with the membranesof the chloroplast.

Evidence for N-Terminal Processing of pLHCP by aSoluble Enzyme

The major product synthesized in the reticulocyte systemfrom RNA generated by pSP65-LHCP is a 31 kD polypeptide(Fig. 1, lane 1), which comigrates with the major pLHCPpolypeptide synthesized from wheat poly(A+) RNA, andcross-reacts with LHCP-specific antibody (17). During in vitroimport, the 31 kD precursor alone gives rise to two peptides

119



Plant Physiol. Vol. 90, 1989

of 25 and 26 kD, indicating that 5 to 6 kD are removed by atransit peptidase (16, 17). (The size ofthe wheat transit peptideas calculated from its amino acid composition is about 4 kD.)In the organelle-free assay, the processing enzyme cleavedapproximately 6 kD from pLHCP and produced only a 25kD peptide (Fig. 1A, lane 2). A priori there was no reason toconclude that the soluble enzyme cleaved at the N-terminusof the precursor since there are multiple forms of LHCPfound in vivo, and it was possible that a subpopulation ofLHCP was formed byCOOH terminal cleavage. To determineif the processing activity actually removes the N-terminus ofpLHCP, mutant precursor proteins lacking their C-terminalamino acids were synthesized in vitro and incubated in aprocessing reaction. The pLHCP gene was modified at its 3'end using convenient restriction enzyme sites (see "Materialsand Methods"). Three deletions were made in which 13, 27,and 91 amino acids were removed (A 13, A27, A9 1; Fig. 2)resulting in a loss of approximately 1.5, 2.9, and 10 kD fromthe precursor, respectively. We predicted that if cleavageoccurs at the N-terminus of pLHCP, the products of thereaction would show a shift in mobility during SDS-PAGEthat reflects the size of the deletion. Alternatively, if cleavageby the processing enzyme occurs at the C-terminus, the A 13and A27 precursors would yield identically sized peptidessince the cleavage must occur internal to the position of thedeletion to remove the predicted 4 to 6 kD. In contrast, theA9 1 precursor would not be processed by a C-terminal activity(Fig. 2B). The results are shown in Figure 2C. Both the A13and A27 precursors were processed and the products migratedwith a shift proportional to the size of their carboxy-terminaldeletions. This provides strong evidence that the soluble proc-essing enzyme cleaves the N-terminus of pLHCP, removingthe transit peptide. Processing of A9 1 has not been observed.We interpret this result to indicate that conformational fea-tures of the precursor essential for processing were disruptedby the A91 deletion, which removed almost one-third ofpLHCP.

Properties of the pLHCP Processing Enzyme

We have previously showed that basic properties of thepLHCP processing activity in chloroplast extracts indicate itis a proteolytic enzyme (17). In this study, to determine theeffect of pH on the efficiency ofpLHCP cleavage, 1 M Hepeswas added to the organelle-free reaction to a final concentra-tion of20 mm over a pH range of 6 to 9.5. Optimal processingoccurred at pH 8 to 9 (Fig. 3A). The temperature optimumwas determined by incubating the reactions at 4, 16, 26, 37,and 42°C. Maximum processing occurred at 26°C, with littleactivity at either extreme (Fig. 3B).The sensitivity of the processing enzyme to a number of

known protease inhibitors was investigated. Iodoacetate, athiol protease inhibitor, was assayed over a range of 1 to 50mm and no inhibition was observed (not shown). EDTA, ageneral divalent cation chelator, inhibited cleavage beginningat 2 mM (Fig. 3C, lane 3). Free ATP, which strongly bindscations, also blocked processing, and EGTA, which prefer-entially binds calcium ions, significantly affected processingonly at a concentration of 50 mm (data not shown). Toinvestigate whether the enzyme is a metalloprotease, the metal

A4

ocJ n.

BMature

~~~~~Afr~~~~~~~~

Figure 2. Evidence of N-terminal processing of pLHCP by the solubleactivity. A, A restriction map of the pLHCP gene showing sites usedin constructing modified templates. The junction (Jct) between thetransit peptide and mature protein coding regions and the translationinitiation (AUG) and termination (TAA) codons are indicated. The Pstl,Kpnl, Apal, HinciI, Sma/Xmal restriction enzyme recognition sites areshown. B, Diagramatic representation of the wild type and truncatedpLHCP encoded by their respective templates. The polypeptide des-ignations used are WT (wild type) pLHCP, Al 3, A27, and A91 (mutantproteins lacking 13, 27, and 91 amino acids, respectively, from thecarboxy terminus of the wild-type pLHCP). The question mark (?)denotes the theoretical cleavage site if processing were to occur atthe carboxy terminus of pLHCP. C, Autoradiogram showing thereticulocyte translation products expressed from the wild-type pLHCPand mutant templates, minus (left) and plus (right) incubation with theprocessing extract.

ion chelator 1,10-phenanthroline was added to the reactionover a concentration range of 0.1 to 3 mm. This compoundstrongly inhibited processing ofpLHCP at 0.4 mM (Fig. 3D).These results suggest that the soluble processing enzyme re-quires divalent metal ions for efficient processing of pLHCP.We previously observed that 10 mm Mg2' slightly reduced thelevel of cleavage ( 17); in light of these more extensive results,it is probable than an optimal cation concentration wasexceeded in those experiments.

rs -,,i

120 ABAD ET AL.

-'. :- ,: .: .., -?

-A-1-1




pH B temperature

I p_,.O

.._: _o WA

1 2 3 4 5 6

1 2 3 4 5 6

EDTA D 1,1 O-phenanthrolir

1 2 3 4 5 6 71 2 3 4 5

Figure 3. Physical characterization of the pLHCPprocessing activity. A, The pH optima of the process-ing reaction was investigated. Radiolabeled pLHCP(lane 1) was used in an organelle-free reaction at pH6.0 (lane 2), pH 7.0 (lane 3), pH 8.0 (lane 4), pH 9.0(lane 5), and pH 9.5 (lane 6). B, The temperatureoptima of the reaction was investigated. The radiola-beled precursor was incubated with the soluble extractat 40C (lane 1), 160C (lane 2), 260C (lane 3), 370C(lane 4), 420C (lane 5), and 260C (the control temper-ature) (lane 6). C, The inhibition of the processingactivity by ethylenediaminetetraacetic acid (EDTA)was investigated. Radiolabeled pLHCP translationproducts (lane 1) were incubated with the processing

ie extract in the absence of EDTA (lane 2), and at con-centrations of 2 mm EDTA (lane 3), 5 mm EDTA (lane4), 10 mm EDTA (lane 5), 20 mm EDTA (lane 6), and50 mM EDTA (lane 7). D, The inhibition of processingby 1,1 0-phenanthroline was investigated. The radio-labeled precursor was incubated with the processingreaction with no phenanthroline added (lane 1), or inthe presence of phenanthroline at 0.1 mm (lane 2), 0.2mm (lane 3), 0.4 mM (lane 4), 0.8 mm (lane 5), 1.0 mM

6 7 8 (lane 6), 2.0 mm (lane 7), and 3.0 mm (lane 8). Thearrows point to the precursor (p) and mature (m) formsof pLHCP.

Size Estimation of the pLHCP Processing Enzyme by GelFiltration Chromatography

To estimate the molecular mass of the processing enzyme,

chloroplasts were lysed in 5 mM Hepes (pH 8.0) in a 1:1resuspension volume (800 ,g Chl/mL lysate) and centrifugedat 137,000g to remove vesicles. The clarified soluble extractwas applied to a Sephacryl S-300 superfine column at 4°Cand eluted fractions were assayed for processing activity. Theelution profile of the processing enzyme as indicated by theappearance of the 25 kD peptide cleaved from pLHCP isshown in Figure 4A. One symmetrical peak of activity was

found from fractions covering a size range of 50 to 700 kD.Calibration of the column revealed that the peak of activityeluted with a Kav value corresponding to a calculated apparentmol wt of about 242,000 (Fig. 5). Under these conditions thepeak always eluted with catalase, a protein standard with a

Mr of 240,000. Based on the amount of protein (2710 ,ug) inthe soluble fraction applied to the column and the amount(48 ,ug) recovered in the peak fraction of processing activity,we estimate a 57-fold enrichment of the processing enzyme

in this single, chromatography step. Using the estimate ofDouce et al. (9) that the soluble phase of the chloroplastmakes up 50% of the total organellar protein, there has beena 114-fold enrichment of the enzyme starting from purifiedorganelles. The maximum extent of processing we have ob-served using the partially purified enzyme corresponds to 85%conversion of the labeled precursor to the mature form,although 50% conversion is most common, based on densi-tometry scans of autoradiograms (data not shown). Becausecomplete processing of the precursor has not yet been ob-served, and a linear increase in processing has not been foundwith the addition of more extract to the organelle-free reac-

tion, the number of units of enzymatic activity in the extracthas not been calculated.

Processing of pS

The small subunit of ribulose- 1,5-bisphosphate carboxyl-ase/oxygenase is the major nuclear encoded protein targetedto the chloroplast stroma. The enzyme responsible for thematuration of the small subunit precursor (pS) has beenpartially purified from pea chloroplasts and has an estimatedmol wt of 180,000 (24). Although this enzyme cleaves theplastocyanin precursor to an intermediate form (11, 14), noreports have described processing of pLHCP by this enzymeor any other chloroplast targeted precursor. In order to deter-mine if the soluble extract prepared under the conditionsdescribed here also contained the protease that cleaves pS, wefirst assayed the total soluble extract using pS generated froma wheat cDNA (4) that we cloned into pSP65. Very littleprocessing, if any, of wheat pS by the pea extract was found(Fig. 6A). A comparison ofpS (see review, ref. 26) from wheatand pea reveals considerable sequence divergence, especiallyin the transit peptide which may contain the determinants forcleavage. The wheat pS transit peptide lacks a block of 11

amino acids present in the pS transit peptide of pea. Thus, itseemed probable that this was the reason for the absence orlow efficiency of processing. Furthermore, wheat pS was

inefficiently imported into and processed by pea chloroplasts,in contrast to wheat pLHCP which was readily incorporated(16). To determine if the soluble extract described here couldprocess pS containing the pea transit peptide, two additionalgene constructs were made to synthesize pS. In one of theseconstructs, the pea pS transit peptide was linked to the codingregion of mature S from wheat (T7-E9/W9) and in the other,both the pea transit peptide and mature protein (T7-E9/SS 15)were linked 3' to a T7 promoter (see "Materials and Meth-

A

C

121

& A-.

----- P

----OAmmftL-. m




Tr Cont r tO i' 82 64 86 88 90 92 94 96 98 100 '02" 104 -iUopL.H.

pLHCP

. .._""I':

-.. P

pS

_

a_-_,f*SON*

Figure 4. Fractionation of the soluble extract by gel filtration chromatography. A crude soluble extract (800 Ag Chl/mL lysate) was clarified bycentrifugation at 137,000g and the supernatant was chromatographed on a Sephacryl S-300 column. One mL fractions were collected andassayed for pLHCP processing activity (A). Translation products of pLHCP were made in the reticulocyte system (Tr.) and incubated with theclarified soluble extract (Cont.) or with fractions eluted from the column, designated 76 through 106 (mL of the eluate). (B) The column fractionswere assayed for pS processing activity. Translation products of E9/SS1 5 (pea pS) were made in the wheat germ system (Tr.) and incubatedwith the clarified extract (Cont.) and with the same fractions (76 through 106) assayed in (A). Compression of lanes 96 and 98 (an artifact ofelectrophoresis) occurred in panel B, falsely accentuating the density of processed S in that part of the elution profile.

A B CN

Gel Filtratlon Chromatography

C0.2d estimated MrK \ e of activity

av* 242,000

0.1I -

0.0 JE4.6 4.8 5.0 5.2 5.4 5.6 5.8 6.0

Log Mr

Figure 5. Estimation of the mol wt of the processing activity by gelfiltration chromatography. The calibration of the Sephacryl S-300column using a linear regression fit of a scatter plot of Kay (see"Materials and Methods" for definition) versus log Mr of 10 proteinstandards. The mol wt of the processing enzyme is estimated to be242,000 based on the elution volume of its peak of activity. Thestandards used were ovalbumin (a); bovine serum albumin (b); phos-phorylase b (c); alcohol dehydrogenase (d); f-amylase (e); catalase(f); urease (g); ferritin (h); ribulose-1,5-bisphosphate carboxylase/oxygenase (i); and thyroglobulin (j).

2

H<5....i.. "O W

M--

2

Figure 6. Processing of pS in the organelle-free assay. A, Translationproducts of the wheat pS template, W9, were made using a wheatgerm lysate (lane 1) and were incubated with the processing extract(lane 2); B, wheat germ translation products of the pea pS template,E9/SS15 (lane 1), were incubated with the processing extract (lane2); C, translation products of the pea/wheat pS template, E9/W9,were made in the reticulocyte system (lane 1) and incubated with theprocessing extract (lane 2). The arrows point to the precursor (p) andmature (m) forms of pS.

ods"). In contrast to the results with wheat pS, the pea solubleextract successfully processed both forms ofpS containing thepea transit peptide, producing a 14 kD peptide (Fig. 6, B andC) that comigrated with endogenous mature S as seen on

Coomassie stained gels (data not shown). Thus, in the organ-elle-free reaction the processing enzyme possesses a substratespecificity that is dependent on the origin of the transit

A

B

toqmw .,-..

Ill!"I 1101,4111111 111111111

122 ABAD ET AL.




peptide. The greater similarity between the pLHCP transitpeptides ofwheat and pea (5, 18) could explain why the wheatprecursor is readily processed by the pea extract.To establish if fractions eluted from the Sephacryl S-300

column contained the pS processing activity, the same frac-tions used to size the pLHCP processing enzyme were againassayed using pea pS (E9/SS 15) as a substrate, which was

cleaved by the activity in the crude extract. Figure 4 showsthat pS was converted to its mature form by the identicalfractions that processed pLHCP, and furthermore, the peakof both activities was identical, eluting at 88 mL.

DISCUSSION

In this study, we describe the optimal conditions for pLHCPprocessing in an organelle-free assay. We find that the pLHCPprocessing enzyme is inhibited by the metal binding agent1, 0-phenanthroline and the chelator EDTA, while beingessentially insensitive to EGTA and iodoacetate, and has a

pH optimum of 8 to 9, at 26°C. The partially purified stromalprotease that cleaves pS and the plastocyanin precursor hasvery similar properties (24). Most significantly, our experi-ments show that the identical fractions eluted from a gelfiltration column convert both pLHCP and pS to their matureforms. Taken together, two alternatives emerge from our

results: (a) the precursors ofLHCP and S are indeed processedby the same enzyme, or of equal significance; (b) a family ofprocessing enzymes is present in the soluble phase of thechloroplast that has evolved substrate specificity while main-taining very similar physical properties. Only when thepLHCP processing enzyme is purified to homogeneity will itbe possible to unequivocally establish if this enzyme is a

general stromal protease responsible for the initial cleavage ofthe large diversity of precursors imported into the chloroplast.We estimate that the size of the processing enzyme which

cleaves pLHCP is about 242 kD. This is somewhat larger thanthe stromal protease (180 kD) initially described for pS mat-uration (24), but the two values are not strikingly differentgiven the difficulty in accurately resolving polypeptides in thissize range by gel filtration. For comparison, the matrix pro-tease of mitochondria is composed of multiple subunits andis estimated to be about 110 kD (3, 12, 31). The molecularcomposition of the pLHCP soluble processing enzyme isunknown, and it will be important to determine if it is a singlepolypeptide, or is comprised of multiple subunits that interactto determine its substrate specificity.

Processing of pLHCP in the organelle-free reactions yieldsa single peptide of about 25 kD, following cleavage of the N-terminus (Fig. IA). This peptide comigrates with the smalleroftwo forms (about 25 and 26 kD) ofmature LHCP producedduring in vitro import of the wheat 31 kD precursor (16, 17).Multiple forms of mature LHCP also have been observedupon import of translation products generated from a singlepLHCP mRNA from genes of Lemna, corn, tomato, and pea

(6, 15, 22, 29), providing strong evidence that these forms are

physiologically important, and reflect the heterogeneity ofLHCP usually found in vivo. Why does the soluble proteaseconvert pLHCP only to the 25 kD peptide, while two formsofLHCP are found after transport of the same precursor intoisolated chloroplasts? One possibility is that the 26 kD form

is transiently produced and immediately (in less than 1 min;see ref. 17) converted to the 25 kD form. It is also possiblethat the precursor is in a folded configuration in the organelle-free reaction, which has a low ionic strength and lacks bothan added energy source and a membrane component, thusexposing one site for preferential cleavage. Thus far, we havevaried only the pH and temperature of the reaction and otherchanges in parameters are currently being explored.The determinants for proteolytic processing that reside

within the imported precursors themselves are not known.Recently, the sequence ile-thr-ser in the pea pS transit peptidewas found to be required for complete processing (30). Thissequence is not present in the transit peptides of pS fromwheat (4) or from Chlamydomonas (28). This may explainthe inefficient processing of the wheat precursor by the peasoluble extract (Fig. 4A) and why pea chloroplasts can importthe algal precursor but only process it to an intermediate form(21). The ile-thr-ser sequence is also not found in the transitpeptides ofpLHCP from pea (5), wheat (18), Lemna (15), orpetunia (10). Thus, as yet unknown features must determinecorrect processing of these other precursors. All transit pep-tides are rich in serine and threonine and are basic, in partic-ular near the cleavage site, but the function of these residuesremains to be established. A common amino acid frameworkof transit peptides has been suggested (13); however, deletionof several key amino acids in this framework did not preventcleavage of pS during import (23). Using the organelle-freeprocessing assay, the functional determinants ofboth pLHCPand pS cleavage can now be established, compared, andseparated from the requirements for precursor translocationinto the chloroplast.Three COOH terminal deletions were constructed in

pLHCP to obtain evidence for N-terminal processing by thesoluble protease. Unexpectedly, the A91 deletion was notcleaved in the organelle-free reaction, whereas both A 13 andA27 showed a change in mobility during SDS-PAGE pre-dicted for removal of the N-terminus. We have also foundthat the A91 truncated precursor is not imported into chlo-roplasts in vitro (S Clark, G Lamppa, unpublished results).Although the A91 deletion removes about one-third of theprecursor, it begins at a considerable distance (141 aminoacids) from the cleavage site and the surrounding structuraldomains, which are still intact. While others have observed,as discussed above, that transit peptide deletions near thecleavage site can prevent complete processing of pS, evidencefor modifications in the mature protein that have a similarinhibitory effect have not been described. An analogous ob-servation, however, has been made for the LamB protein ofEscherichia coli. A deletion in the mature portion of theLamB precursor between residues 70 and 200 prevents proc-essing during export (2). It appears that in addition to theprimary sequence required for pLHCP maturation, the sec-ondary conformation of the protein establishes the accessibil-ity of the processing enzyme recognition and cleavage sites.We are currently investigating the determinants of pLHCPmaturation using chloroplast in vitro import and the organ-elle-free processing assay.

123




ACKNOWLEDGMENTS

We would like to thank Thomas Dorman for his excellent technicalassistance. We thank Dr. Nam-Hai Chua for providing the pea cDNAclone pSS 15 and Dr. Colleen Jacks for critically reading themanuscript.

LITERATURE CITED

1. Bartlett S, Grossman AR, Chua N-H (1982) In vitro synthesisand uptake ofcytoplasmically synthesized chloroplast proteins.In M Edelman, RB Hallick, N-H Chua, eds, Methods inChloroplast Biology. Elsevier Biomedical, New York, pp 1081 -1091

2. Benson SA, Silhavy TJ (1983) Information within the maturelamB protein necessary for localization to the outer membraneof E. coli K12. Cell 32: 1325-1335

3. Bohni PC, Daum G, Schatz G (1983) Import of proteins intomitochondria. Partial purification of a matrix-located proteaseinvolved in cleavage of mitochondrial precursor polypeptides.J Biol Chem 258: 4937-4943

4. Broglie R, Coruzzi G, Lamppa G, Keith B, Chua N-H (1983)Structural analysis of nuclear genes coding for the precursor tothe small subunit of wheat ribulose-1,5-bisphosphate carbox-ylase. Biotechnology 1: 55-61

5. Cashmore AR (1984) Structure and expression of a pea nucleargene coding a chlorophyll a/b binding polypeptide. Proc NatlAcad Sci USA 81: 2960-2964

6. Cline K (1988) Light-harvesting chlorophyll a/b protein: mem-brane insertion, proteolytic processing, assembly into LHC II,

and localization to appressed membranes occurs in chloroplastlysates. Plant Physiol 86: 1120-1126

7. Coruzzi G, Broglie R, Cashmore A, Chua N-H (1983) Nucleotidesequence of two pea cDNA clones encoding the small subunitof ribulose-1,5-bisphosphate carboxylase and the major chlo-rophyll a/b-binding thylakoid polypeptide. J Biol Chem 258:1399-1402

8. Coruzzi G, Broglie R, Edwards C, Chua N-H (1984) Tissue-specific and light-regulated expression of a pea nuclear geneencoding the small subunit of ribulose-1,5-bisphosphate car-boxylase. EMBO J 3: 1671-1679

9. Douce R, Holtz RB, Benson AA (1973) Isolation and propertiesof the envelope of spinach chloroplasts. J Biol Chem 248:7215-7222

10. Dunsmuir P (1985) The petunia chlorophyll a/b binding proteingenes: a comparison of Cab genes from different families.Nucleic Acids Res 13: 2503-2518

11. Hageman J, Robinson C, Smeekens S, Weisbeek P (1986) Athylakoid processing protein is required for complete matura-tion of the lumen protein plastocyanin. Nature 324: 567-569

12. Hawlitschek G, Schneider H, Schmidt B, Tropschug M, HartlF-U, Neupert W (1988) Mitochondrial protein import: iden-tification of processing peptidase and of PEP, a processingenhancing protein. Cell 53: 795-806

13. Karlin-Neumann GA, Tobin EA (1986) Transit peptides of nu-clear-encoded chloroplast proteins share a common aminoacid framework. EMBO J 5: 9-13

14. Kirwin PM, Elderfield PD, Robinson C (1987) Transport ofproteins into chloroplasts: partial purification of a thylakoidalprocessing peptidase involved in plastocyanin biogenesis. J BiolChem 262: 16386-16390

15. Kohorn BD, Harel E, Chitnis PR, ThornberJP, TobinEM (1986)Functional and mutational analysis of the light-harvestingchlorophyll a/b protein of thylakoid membranes. J Cell Biol102: 972-981

16. Lamppa GK (1988) The chlorophyll a/b binding protein insertsinto the thylakoids independent of its cognate transit peptide.J Biol Chem 263: 14996-14999

17. Lamppa GK, Abad M (1987) Processing of a wheat light-har-vesting chlorophyll a/b protein precursor by a soluble enzymefrom higher plant chloroplasts. J Cell Biol 105: 2641-2648

18. Lamppa GK, Morelli G, Chua N-H (1985) Structure and devel-opmental regulation of a wheat gene encoding the majorchlorophyll a/b binding polypeptide. Mol Cell Biol 5: 1370-1378

19. Lamppa GK, Nagy F, Chua N-H (1985) Light-regulated andorgan-specific expression of a wheat Cab gene in transgenictobacco. Nature 316: 750-752

20. Marder JB, Goloubinoff P, Edelman M (1984) Molecular archi-tecture of the rapidly metabolized 32-kilodalton protein ofphotosystem II. J Biol Chem 259: 3900-3908

21. Mishkind ML, Wessler SR, Schmidt GW (1985) Functionaldeterminants in transit sequences: import and partial matura-tion by vascular plant chloroplasts of the ribulose-l,5-bisphos-phate carboxylase small subunit of Chlamydomonas. J CellBiol 100: 226-234

22. Pichersky N, Hoffman V, Malik V, Bernatzky R, Tanksley S,Szabo L, Cashmore A (1987) The tomato Cab-4 and Cab-Sgenes encode a second type of CAB polypeptides localized inPhotosystem II. Plant Mol Biol 9: 109-120

23. Reiss B, Wasmann CC, Bohnert HJ (1987) Regions in the transitpeptide of SSU essential for transport into chloroplasts. MolGen Genet 209: 116-121

24. Robinson C, Ellis RJ (1984) Transport of proteins into chloro-plasts: partial purification of a chloroplast protease involved inthe processing of imported precursor polypeptides. Eur JBiochem 142: 337-342

25. Robinson C, Ellis RJ (1984) Transport of proteins into chloro-plasts: the precursor of small subunit of ribulose bisphosphateis processed to the mature size in two steps. Eur J Biochem142: 343-346

26. Schmidt GW, Mishkind ML (1986) The transport of proteinsinto chloroplasts. Annu Rev Biochem 55: 879-912

27. Schmidt GW, Bartlett SG, Grossman AR, Cashmore AR, ChuaN-H (1981) Biosynthetic pathways oftwo polypeptide subunitsof the light-harvesting chlorophyll a/b protein complex. J CellBiol 91: 468-478

28. Schmidt GW, Devillers-Thiery A, Desruisseaux H, Blobel G,Chua N-H (1979) NH2-terminal amino acid sequences ofprecursor and mature forms of the ribulose-1,5-bisphosphatecarboxylase small subunit from Chlamydomonas reinhardtii.J Cell Biol 83: 615-622

29. Sheen J-Y, Bogorad L (1986) Differential expression of six light-harvesting chlorophyll a/b binding protein genes in maize leafcell types. Proc Natl Acad Sci USA 83: 7811-7815

30. Wasmann CC, Reiss B, Bohnert HJ (1988) Complete processingof a small subunit of ribulose-1,5-bisphosphate carboxylase/oxygenase from pea requires the amino acid sequence ile-thr-ser. J Biol Chem 263: 617-619

31. Witte C, Jensen RE, Yaffe MP, Schatz G (1988) MAS 1, a geneessential for yeast mitochondrial assembly, encodes a subunitof the mitochondrial processing protease. EMBO J 7: 1439-1447

124 ABAD ET AL.



Properties of a Chloroplast Enzyme that Cleaves Binding ... · PROPERTIES OF CHLOROPLAST PROCESSING ENZYME THAT CLEAVES pLHCP ColumnChromatography The crude soluble extract was centrifuged

Documents