mutations at the transit peptide-mature protein junction separate two cleavage events during

THE JOURNAL OF BIOLOGICAL CHEMISTRY 0 1989 by The American Society for Biochemistry and Molecular Biology, Inc.

vel. 264, No. 29, Issue of October 15, pp. 17544-17550,1989 Printed in U.S.A.

Mutations at the Transit Peptide-Mature Protein Junction Separate Two Cleavage Events during Chloroplast Import of the Chlorophyll a/b-binding Protein*

(Received for publication, April 28, 1989)

Steven E. Clark, Mark S. Abad, and Gayle K. Lamppa$ From the Department of Molecular Genetics and Cell Biology, Uniuersity of Chicago, Chicago, Illinois 60637

We have shown previously that during in vitro import into chloroplasts, the precursor of the major light- harvesting chlorphyll a/b-binding protein (LHCP) generated from a wheat gene gives rise to two mature forms (25 and -26 kDa) which are inserted into the thylakoids. However, during incubation of the LHCP precursor with a chloroplast-soluble extract in an organelle-free processing reaction, the NH2 terminus is cleaved, yielding only a 25-kDa peptide. In the present study, mutations at the transit peptide-mature protein junction were introduced in the LHCP precursor to investigate the relationship between the two peptides and the determinants of proteolytic processing. Mutant pA3 lacks 3 amino acids including Met34 at the primary cleavage site thought to give rise to the 26-kDa peptide. It is still processed during import and in the organelle- free reaction yielding in both assays only a 25-kDa peptide. Mutant p+4 has 4 amino acids inserted immediately after Met34 and a proline that disrupts the a-helix predicted by the Garnier-Osguthorpe-Robson method (Garnier, J., Osguthorpe, D. J., and Robson, B. (1978) J. Mol. Biol. 120,97-120) to extend through this region. Although p+4 is imported, it is inefficiently processed; both a 25- and 26-kDa peptide are found, but at least 60% of the imported precursor remains uncleaved. Less than 5% is processed in the organelle- free assay. Replacement of the predicted a-helix in the mutant p+4a restores processing upon import into the chloroplast, but this mutant, which also has a 4-amino acid insert, yields only a 26-kDa peptide. p+4a is not processed in the organelle-free reaction. These results provide evidence that the two forms of LHCP obtained during import are the result of independent processing at two cleavage sites: the first site at Met34, and a second -10 amino acids downstream within what has been designated the NH2 terminus of the mature protein. Whereas pA3 has the first site removed but retains a functional second site, in p+4a only the first site, or one very near it, is accessible to the processing enzyme during import. The conditions of the organelle-free reaction are specific for processing at only the secondary site. We discuss the implications of these findings in terms of the heterogeneity of LHCP in vivo.

* This research was supported in part by National Institutes of Health Grant GM36419 and United States Department of Agriculture Grant 8701037 (awarded to G. K. L.) and by Sigma Xi grants-in-aid of research (to S. C. and M. A.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “aduertisernent” in accord- ance with 18 U.S.C. Section 1734 solely to indicate this fact.

$ To whom correspondence should be addressed Dept. of Molecu- lar Genetics and Cell Biology, University of Chicago, 920 E. 58th St., Chicago, IL 60637.

The major light-harvesting chlorophyll a/b-binding protein (LHCP)’ of the chloroplast is nuclear encoded, synthesized as a precursor with a cleavable NH2-terminal transit peptide, and post-translationally imported into the organelle (Schmidt et ul., 1981). LHCP functions in the thylakoid membranes, where it is primarily associated with photosystem I1 (PSII) to maximize photosynthetic efficiency (see Anderson and An- dersson, 1988). LHCP of PSII is composed of two major polypeptides: one is 26-27 kDa, depending on the plant species; and the other is 25 kDa. The sequence at the NH2 terminus of the -26-kDa form has been determined (Mullet, 1983) and was used to deduce that the proteolytic processing of the LHCP precursor occurs 33-34 amino acids from its NH2 terminus. The relationship of the two major forms of LHCP has not been resolved previously. Common epitopes have been identified in the 26- and 25-kDa polypeptides using monoclonal antibodies, but the NH, terminus of the 26-kDa form is absent in the 25-kDa protein (Darr et al., 1986). These results raise the possibility that in addition to removal of the transit peptide, another post-translational proteolytic processing occurs upon import of the LHCP precursor (pLHCP) into the organelle or that removal of the transit peptide can occur at two separate locations, producing two subpopulations of LHCP with different functional roles. Further evidence for post-translational processing comes from in vitro import studies. Upon chloroplast import of pLHCP, which has been synthesized in vitro from transcripts of a single wheat gene, both a -26- and 25-kDa form of LHCP are found associated with the thylakoids (Lamppa and Abad, 1987; Lamppa, 1988). Multiple forms of mature protein have also been obtained recently whether the LHCP gene originates from pea (Cline, 1988), tomato (Pichersky et al., 1987), corn (Dietz and Bogo- rad, 1987) or lemna (Kohorn et al., 1986).

In order to investigate the determinants of pLHCP cleavage, we have recently developed and optimized an organelle- free processing reaction specifically for pLHCP, in which radiolabeled precursor is incubated with a soluble chloroplast extract (Lamppa and Abad, 1987; Abad et al., 1989). Previ- ously, Robinson and Ellis (1984a, 198413) described an analo- gous in vitro reaction for the processing of the precursor of the small subunit (pS) of ribulose-1,5-bisphosphate carboxylase/oxygenase. We have shown that a soluble enzyme with an apparent M, = 240,000 cleaves pLHCP, removing the transit peptide, and that the properties of the enzyme are nearly identical to the enzyme that cleaves pS (Abad et ul.,

The abbreviations used are: LHCP, light-harvesting chlorophyll a/b-binding protein; pLHCP, LHCP precursor; pS, small subunit of ribulose-1,5-bisphosphate carboxylase/oxygenase precursor; SDS, so- dium dodecyl sulfate; PAGE, polyacrylamide gel electrophoresis; PSI and PSII, photosystems I and 11; bp, base pair(s); Hepes, 4-(2- hydroxyethy1)-1-piperazineethanesulfonic acid.

17544

Import and Processing of the Chlorophyll a/b Protein Precursor 17545

1989). Unexpecte~y, when pLHCP was used as the substrate in the organelle-free reaction, only one peptide was released, and it co-migrated with the smaller 25-kDa peptide obtained during in vitro import of pLHCP. No evidence has been obtained for the 26-kDa form being a transient intermediate in the organelle-free reaction.

The determinants of pLHCP processing which reside within the structure of the precursor itself have not been established. Recently it has been shown that residues near the COOH terminus of the transit peptide of pS are required for cleavage (Wasmann et aL, 1988; Reiss et ai., 1989; Ostrem et at., 1989). To approach this question for pLHCP, in the present study, modifications were made at the transit peptide- mature protein junction. The effects of these modifications were assayed using both in vitro import and the organelle- free processing reaction. We report that the cleavage events giving rise to a 25-kDa and 26-kDa form during import can be separated by small changes in the junction domain of pLHCP. Insertion of 4 amino acids results in the production of only a 26-kDa peptide during import and a complete loss of processing in the organelle-free reaction. On the other hand, if the 3 amino acids containing the cleavage site at position 34 are deleted in a second mutant, processing upon import yields only the 25-kDa peptide. In the organelle-free reaction, this mutant substrate was cleaved as efficiently as wild type, producing again only the 25-kDa peptide. Our results indicate that there are two functionally independent cleavage sites in pLHCP. We discuss the hypothesis that the processing enzyme preferentially cleaves at the second site in the organelle-free reaction due to the absence, most notably, of a membrane component that may alter the conformation of the precursor. We also present the possibility that the relative-abundance of the 25- and 26-kDa polypeptides associated with PSI1 in vivo may reflect selective cleavage of pLHCP in response to environmental changes that influence the physiology of the chloroplast.

EXPERIMENTAL PRO~EDURES

Plusmid Construction---A wheat genomic clone coding for pLHCP was characterized previously (Lamppa et al., 1985) and suhcloned into the transcription vector SP65 to yield the plasmid SP65-pLHCP (Lamppa and Abad, 1987). In this study, the DNA construct coding for pA3 was made by deleting nine nucleotides, including the ATG codon for Met34, SP65-pLHCP was digested with EcoRI and Hind111 releasing a -1.6-kilobase insert. The insert was digested with HgaI, which produces two fragments (300 and 1300 bp), the smaller coding for the transit peptide. This fragment, with a staggered end beginning 5 hp upstream of the ATG, was isolated and treated with mung bean nuclease followed by BamHI. It was then inserted into an SP65- pLHCP plasmid from which the transit peptide plus one nucleotide immediately 3' of the ATG had been removed by digestion with BamHI and FspI. Upon ligation (T4 DNA ligase, Boehringer Mann- heim, as prescribed), the construct lacked codons for 3 amino acids at the junction of the transit peptide and mature protein, and a valine at position 33 had been converted to a glycine. Competent Escherichia coli HBlOl cells were transformed following a standard protocol (Maniatis et aL, 1982), and the plasmid construct SP65-pLHCPA3 was isolated.

The DNA template coding for p+4 was made by first constructing the intermediate plasmid SP65-pLHCP-CL containing a IO-bp ClaI linker (d(pCCATCGATGG)). SP65-pLHCP was partially digested with FspI, which linearized the plasmid at one of two sites in the plasmid. One site is at the junction of the transit peptide and mature protein coding regions for LHCP, and the other is in the gene coding for ampicillin resistance. The linearized plasmid was treated with calf intestine phosphatase (0.5 unit/pg DNA), and then 250 ng of DNA was ligated with 0.5 pg of the ClaI linker (New England BioLabs) using T4 ligase (3 units). E. coli HBlOl cells were transformed with 20% of the reaction mixture. To remove multiple linkers that may have been inserted into the resulting plasmid, the DNA was digested to completion with ClaI, and the linear form was isolated from an

agarose gel, ligated, and used to transform HBlOl cells. SP65- pLHCP-CL was linearized with C l d , isolated hy gel electrophoresis, and treated with the large fragment of DNA polymerase 1 (Klenow enzyme). Upon ligation, the ClaI site was lost, and an NruI site was formed, producing the construct SP65-pLHCP+4, containing four additional codons, coding for the polypeptide p+4. The DNA template coding for p i - 4 ~ ~ was constructed by again partially digesting SP65- pLHCP with FspI, and a 12-bp PuuI linker (d(pTCGCGATCGCGA)) (New England BioLabs) was inserted using essentially the same strategy as described above for insertion of the ClaI linker. The mutations were confirmed by digestion with the appropriate restric- tion enzyme based on the site created by the insertion of the DNA linker and dideoxy chain termination sequence analysis (Sanger et aL, 1977) using an oligonucleotide primer complementary to codons 55-59 of the precursor.

In Vitro Transcription and Translation-The DNA templates (1 pg) were linearized 600 bp 3' to the pLHCP stop codon with HirzdIII and transcribed using SP6 polymerase following the procedure of the vendor (Promega-Fisher). The synthetic RNA (1-2 pg; about 1/10 of the transcription products) was translated in a standard 30-p1 reticulocyte lysate reaction (Bethesda Research Laboratories) using 1 p l of [3sS]methionine (1000 Ci/mmol, Du Pont-New England Nuclear) as the labeled amino acid.

Chloroplast Isolation and in Vitro Import-Chloroplasts were isolated from 8-day-old pea (Pisum satiuum, Laxton's Progress) seedlings and used for in vitro import reactions as described previously (Bartlett et al., 1982; Lamppa and Ahad, 1987). Typically, the chloroplasts were resuspended in 50 mM Hepes-KOH, pH 8, 0.33 M sorbitol, 8 mM methionine (HSM) at 700 pg of ch~orophyll/ml, and 50 p1 was added to a 300-pl sample containing HSM, plus 10 mM ATP, 10 mM MgCl2, and one translation reaction. Following a 30-60 min import reaction, the chloroplasts were either treated prior to lysis with the protease thermolysin (100 pg/ml) at 4 "C or not treated, and then were washed gently with 2 ml of HSM, pelleted, and resuspended in 400 pl of 1 mM phenylmethylsulfonyl fluoride for lysis. The membrane and soluble fractions were separated by centrifugation at 16,000 X g for 10 min. For direct comparison of the amount of protein localized to each organelle compartment, the entire membrane and soluble fractions from each import reaction were isolated as described (Lamppa and Abad, 19871, resuspended in loading buffer (0.2% SDS, 10% glycerol, 5% ~-mercaptoethanol, 0.5 M tris, pH 6.8, 0.01% bromophenol blue), and analyzed by SDS-PAGE using 15% acrylamide gels.

~ r g ~ e ~ - f r e e Processing ~eaction-Organelle-free processing reactions were carried out as optimized (Abad et al., 1989). Isolated chloroplasts in HSM were centrifuged at 2,500 X g for 60 s and gently resuspended in 5 mM Hepes-KOH, pH 8, at 700-900 pg of chlorophyll/ml. Lysis was at 4 "C for 30 min without further mixing. The membranes were pelleted by centrifugation at 16,000 X g for 10 min, and the supernatant was then spun at 137,000 X g for 60 min to remove membrane vesicles. A 25-pl processing reaction contained 10 p1 of the 137,000 X g supernatant, 5-10 pl of radiolabeled reticulocyte translation products, 22 mM Hepes-KOH, pH 8, and 2 pg/ml ehlor- amphenicol. The sample was incubated at 26 "C for 90 min and stopped by the addition of SDS-PAGE loading buffer.

~hylakoid Insertion Reaction-The thylakoid insertion reaction was essentially as described by Cline (1988). Chloroplasts were resuspended in 10 mM Hepes-KOH, pH 8, at 400-500 pg of chlorophyll/ ml, stored on ice for 5 min, and then 200 pl of the chloroplast lysate was used in a 300-pl reaction including 25 pl of reticulocyte translation products, 10 mM ATP (magnesium salt), 22 mM Hepes-KOH, pH 8, 0.11 M sorbitol, and 2.5 mM methionine. The reaction was incubated at 26 "C for 45 min, whereupon it was diluted 3-fold with 0.33 X HSM and centrifuged for 10 min at 16,000 X g. The washed membranes were resuspended in 10 mM Hepes-KOH, pH 8, at 300 pg of chlorophyll/ml, and then half was pelleted at 16,000 X g for 10 min and resuspended in SDS-PAGE loading buffer. The other half was treated with trypsin (100 pg/ml) a t 25 "C for 15 min, washed with a IO-fold excess of 1 mM phenylmethylsulfonyl fluoride, and again pelleted. The wash was repeated, and the sample was then resuspended in gel loading buffer for analysis.

S e c o ~ a ~ Structure AnaLys~ of Precursor Po~ypeptides-The full length amino acid sequences of wheat and pea wild-type pLHCP and the mutant wheat LHCP precursors were used to acquire secondary structure predictions by the method of ~ a r n i e r - O s ~ t h o ~ e - R o h s o n (Garnier et al., 1978). The first 100 amino acids of petunia and corn precursors were used to predict their secondary structures by the same method. The Peptidestructure and PlotStructure Programs

17546 Import and Processing of the ~ h ~ ~ r o p h y ~ l a/b Protein Precursor

pea:pLHCP MAASSSSSMALSSPTLAGKQJ~K~~PSSQELGAARFTMKK~ATTKKVASSGSPWYG

w h e : p L i i C P MAA----TTMSLSSSSFAGKA~~KNLPSSAVI~DAR~~K?AAKAKQVSSSSPWY~

00000 00 0 + t 30 - t 0 ti0 R o t + 00 0

00 0 0030 t t 00 ~ t d t 0 t * 0000

whe:pA3 MAR---TTMSLSSSSFAGKAV~~IL?SSAVIGDAR~K~~KAKQVSSSSF~~~G

whe:p+4

00 0 0000 + + 00 - + i 0 * + 0000

00 0 0000 t t 00 - +, ,+a t t 0000

00 G 0000 t + 00 - + ~ ?+0 t + 0000 MAA---?f’MSLSSSSCAGKAVK~~L?SSAVIGDAR~K?AAK.~KQVSSSSP~~‘~G

whe:p+4alpha MAR---TTMSZSSSSFAGKAVKNL?SSAV~GDA~~~~~ThAKAKQ~‘~SSS~’~YG

FIG. 1. Amino acid sequences of the wheat (whe) pLHCP mutants comparing the transit peptide- mature protein junctions. By using the single letter amino acid code, the NH2 termini of the precursors are shown including the full sequence of the transit peptide and the first 17 residues of the mature protein (26-kRa form). The NH2 terminus of the wheat wild-type precursor is presented on the second tine with the VNMR at the cleavage site bracketed, including Met34, and the pea pLHCP sequence is given above for comparison. The position of modification (insertion, deletion, substitution) is also bracketed for each mutant. The open circles indicate the positions of the polar uncharged amino acids (S, Ser, T, Thr), the basic amino acids are indicated by aplus symbol ( R , Arg, K , Lys), and the acidic residues by a minus (D, Asp, E, Glu). The dashes indicate the absence in the wheat protein of residues found in the pea transit peptide.

(Wolf et at., 1988) from the University of Wisconsin Genetics Com- puter Croup collection were used to calculate and diagram the predictions.

RESULTS

To investigate the determinants of proteolytic processing of the precursor of LHCP, mutations were made at the junction of the transit peptide and mature protein. A wheat gene coding for pLHCP (Lamppa et al., 1985) was r e s t ~ c t u r ~ , creating the three mutants described below and transcribed in uitro using sP6 polymerase. The synthetic RNA was translated in a reticulocyte lysate in the presence of [35S]methionine, and the radiolabeled mutant forms of pLHCP were analyzed in two separate assays. First, the pLHCP mutants were tested for their competence for import into isolated pea chloroplasts. As shown previously, the import of wild-type wheat pLHCP into either pea or wheat chloroplasts gives the same result, i.e. two forms of mature LHCP with apparent sizes of 25 and 26 kDa were found associated with the thylakoids. We have addressed previously the point that although there are also two minor forms of the precursor synthesized in the reticulocyte lysate, the 31-kDa is the most abundant, cross-reacts with LHCP-specific antibody, and alone will yield the two mature forms in an import reaction (Lamppa and Abad, 1987). In a second assay, the wheat precursors were incubated with a soluble chloroplast extract, which we described as an organelle-free reaction (Lamppa and Abad, 1987; Abad et al., 19891, containing a “transit peptidase” activity, i.e. the NH2 terminus of pLHCP containing the transit peptide is removed. In this assay, pLHCP generated only a 25- kDa peptide. Thus, it was possible to monitor this processing event independent of translocation into the chloroplast or insertion into the thylakoids.

The first objective of these experiments was to remove the cleavage site of the major form of LHCP, which was deduced from amino acid sequence analysis of a -26-kDa peptide from pea (Mullet, 1983). Although this cleavage site has not been unequivocally established, it has been shown that it begins either immediately before or after the methionine that is 34 residues from the NH2 terminus of the precursor. For ease of discussion and because the 26-kDa peptide is the predominant form found in the chloroplast, we define this as the primary cleavage site of pLHCP. Mutant pb3 was constructed in which 9 bp coding for 3 amino acid~-Asn~~-Met~~-Arg3“ (numbered from the NH2 terminus of the precursor)-encom- passing this site were deleted. In addition, the valine at position 32 was converted to a glycine. The amino acid sequences at the transit peptide-mature protein junction of pA3 and the other precursors used in this study are shown in Fig. 1.

After incubation of pA3 with intact chloroplasts in an import reaction, the organelles were either immediately lysed and separated into membrane and soluble fractions or treated with thermolysin prior to lysis to remove proteins that had not successfully crossed the envelope membrane. Products were analyzed by SDS-PAGE. In parallel reactions, wild-type pLHCP was analyzed to control for the efficiency of import and processing. As typically observed, wild-type pLHCP yielded two peptides with sizes of approximately 25 and 26- kDa, which were membrane associated and thermolysin resistant (Fig. 2 A , kftpanel). No radiolabeled protein was found in the soluble fractions. We have demonstrated recently that the two forms are both produced by processing at the NHz terminus of pLHCP using COOH-terminally truncated polypeptides.2 To confirm that the production of two mature forms in the import assay was not restricted to wheat pLHCP, the precursor was also synthesized in uitro from a pea gene (Cashmore, 1985) and analyzed. Pea pLHCP gave rise to two forms of mature LHCP, and the 26-kDa form was the predominant (Fig. 2B). In contrast, upon import of pA3 into the chloroplast, only a single peptide was released, and it co- migrated with the lower of the two forms, i.e. the 25-kDa peptide, found upon import of the wild-type precursor (Fig. 2 A , right panel). It was also consistently observed that compared with wild-type, only about 50% of the cleaved protein was found resistant to thermolysin, suggesting either a lower efficiency of pA3 import or a less stable association of the processed form with the thylakoids for the duration of the import reaction. Since pb3 gave rise to only a 25-kDa peptide, we conclude that deletion of the primary cleavage site at Met34 did not affect processing at a downstream secondary site in pLHCP.

Incubation of ra~olabeled pA3 with a chloroplast-soluble extract in the organelle-free processing reaction produced only a 25-kDa peptide, and it was the same size peptide found upon processing of wild-type pLHCP (Fig. 2C). These processed forms co-migrated with the 25-kDa peptide seen upon import of pLHCP or pA3 into chloroplasts (not shown).

A second construct designated p+4, in which 4 amino acids-Ile-Ala-Met-Gly-were inserted into the precursor, was designed to ieave the sequence of the transit peptide intact while disrupting the NH2 terminus of the mature protein (see Fig. 1). The spacing of the cluster of positively charged amino acids, arginine and lysine, was altered in this domain by the insertion of a ClaI linker into the appropriate location of the wheat gene (see “Experimental Procedures”). In addition, the a-helical structure, predicted by the method of Gamier-Osguthorpe-Robson (1978) to extend for 15 amino

-2 S. E. Clark, J. E. Oblong, and G. K. Lamppa, unpublished results.

Import and Processing of the Chlorophyll alb Protein Precursor 17547

A Wheat

31 -

26 - 2 5-

B

PA3

3

TP M M+ S S+ TP M M+ Th Th Th

Pea C

1 2 3 4 5 kDa

4 5

s s+ Th

1 2

25-

TP M M+ S S+ Th Th

FIG. 2. Import and processing of the deletion mutant p A 3 lacking the primary cleavage site. Analysis was by SDS-PAGE, and estimated sizes are given in kilodaltons. Panel A shows the results of precursor import into isolated chloroplasts for both wild-type wheat pLHCP (wheat, left) and the mutant (pA3, right). Panel B shows a similar import experiment for pea pLHCP. In panels A and B, the [3sS]methionine-labeled translation products (2 pl of a 30-pl reticulocyte reaction) are shown (TP; lanes 1 ) ; also presented are the proteins associated with the total membrane fraction after import (M; lanes 2), the membrane proteins resistant to treatment of the organelles with thermolysin (Th) after import but before lysis (lanes 3 ) , the total soluble proteins after import, i.e. nonprotease-treated (S; lanes 4) , and the soluble proteins resistant to thermolysin treatment, representing primarily the stromal fraction (lanes 5). The samples in panel A were analyzed on the same gel in adjacent lanes to compare migration patterns of the radiolabeled protein. Panel C represents the results of an organelle-free processing reaction using the soluble chloroplast extract prepared as described under “Experi- mental Procedures.” Lane 1 shows processing of radiolabeled wild- type pLHCP, and lane 2 shows processing of pA3. Approximately 2 times more radiolabeled pA3 than pLHCP was used in the reaction.

acids through this region (see “Discussion” and Fig. 6), was disrupted by the conversion of arginine a t position 35 to a proline. In an import experiment, as shown in Fig. 3, p+4 was transported into the chloroplast and localized to the thylakoid membranes; and similar to pLHCP, both the 25- and 26-kDa peptides were found. However, based on densitometer analysis of the corresponding autoradiogram, approximately 60% of the radiolabeled protein found in the thylakoid membranes was the precursor that remained unprocessed. Chitnis et ul. (1988) have observed recently that pLHCP is transported into barley etioplasts form dark-grown seedlings apparently lacking an active processing enzyme. Our results further demon- strate using greened chloroplasts with a fully functional processing enzyme that translocation across the envelope membrane is not obligately coupled to processing of pLHCP.

To determine if the imported precursor was inserted into the thylakoids and not peripherally associated, the membranes were treated with trypsin (100 rg/ml), which selec- tively cleaves the NH2 terminus of the polypeptide that normally extends into the stroma (Mullet, 1983). A comparison

of the mutant and wild-type pLHCP shows (Fig. 3B) that both gave the same size trypsin-resistant fragment migrating a t -24 kDa, agreeing well with earlier studies. The amount of trypsin-resistant protein seen on analysis of the mutant (Fig. 3B, lane 3) , however, is markedly higher than the amount of mature protein seen upon import (lane Z), an indication that the precursor itself contributes to the amount of trypsin- resistant product. Although our previous studies (Lamppa, 1988) indicated that the cognate transit peptide of LHCP is not required for thylakoid localization, these results show that the mutant precursor can also insert into the thylakoids in the correct orientation without removal of the transit peptide. The temporal relationship between pLHCP maturation and thylakoid insertion has not yet been established.

When p+4 was used in an organelle-free processing reaction, essentially no mature protein was produced. At most, 5% of the precursor was converted to a 25-kDa peptide, whereas a t least 50% of pLHCP was cleaved under identical conditions, and a COOH-terminally truncated mutant, with 13 amino acids from positions 252-264 of pLHCP deleted, was also cleaved (Fig. 3C). Taken together with the import results, we conclude that the 4 amino acids inserted into the precursor and the substitution of proline for arginine in the mutant p+4 significantly reduce the efficiency of substrate recognition and/or cleavage by the processing enzyme at both sites that produce the 25- and 26-kDa peptides. An additional assay supports this conclusion. The three substrates, pLHCP, pA3, and p+4, were incubated with chloroplast lysates under conditions in which insertion into the thylakoids and processing can both occur independently of translocation into the organelle (Cline, 1988). Although processing is somewhat less efficient than in the import reaction, we found that both pLHCP and pA3 inserted into the thylakoids and were cleaved to a mature form that was resistant to trypsin treatment (Fig. 4). In contrast, p+4 inserted into the thylakoids, but almost no mature protein was found. Once again though, the unprocessed precursor upon trypsin treatment yielded a 24-kDa peptide, as did pLHCP and pA3 (Fig. 4, lanes 3 ) .

To investigate whether the inefficient cleavage of p+4 was due to the change in spacing of the positively charged amino acids or to the disruption of the predicted a-helical structure in this region, the insertion mutant p+4a was constructed (see Fig. 1). In this mutant, the arginine at position 35 was converted to a leucine, and 4 additional uncharged amino acids-Ala-Ile-Ala-Ser-were inserted at the same location as in p+4. These changes, however, did not alter the a-helix predicted to extend through the transit peptide-mature protein junction domain (see Fig. 6). Import of p+4a into chloroplasts was as efficient as the wild type, and the precursor was completely processed, i.e. none was found associated with the thylakoids at the end of the import reaction (Fig. 5A). Unexpectedly, however, only a -26-kDa peptide was observed in the membrane fraction. The products of p+4, p+4a, and wild-type pLHCP import reactions are compared in Fig. 5B, in which the membrane fractions from thermolysin-treated chloroplasts were analyzed side by side. Clearly, the peptide released from p+4a (Fig. 5B, lune 2) migrates with the larger form of LHCP produced from the wild-type precursor (lane 3) , or p+4 (lane I ) . In addition, this experiment shows con- vincingly that under identical import conditions, mutant p+4 is not efficiently processed compared with p+4a and pLHCP. It appears that replacement of an a-helix in the transit peptide-mature protein junction domain of p+4a restores a conformational feature essential for processing; however, only one cleavage event occurs during import. That is, only one site is accessible to the processing enzyme. As indicated by

17548 Import and Processing of the Chlorophyll alb Protein Precursor

FIG. 3. Analysis of import and processing of the insertion mutant p+4. Panel A shows the import of both wild-type pLHCP (wt) and p+4 compared for the relative efficiency of import using the same chloroplast preparation. Lanes 1-5 are as described in Fig. 2, panels A and B, and abbreviations are the same. Panel B shows an analysis of trypsin treatment of the proteins associated with the thylakoids after an import reaction. The membrane fraction was divided, and half was either not treated (lanes 2 ) or treated with trypsin (Tr) at 100 pg/ml (lanes 3) before SDS- PAGE. The total soluble fraction after import was simultaneously analyzed (lanes 4 ) . The reticulocyte lysate translation products used in the import reaction are shown (lanes 1). Panel C presents the results of an organelle-free processing reaction for p+4 (lane I), pLHCP (lane 2) , and a control mutant, pA13, in which 13 amino acids were deleted from the COOH terminus of the precursor (lane 3). In this case, electrophoresis of the products was on a 15% polyacrylamide gel for 12 h, thus, the two bands that normally migrate above the 31-kDa polypeptide were well separated species.

P+4 WT ”

1 2 3 1 2 3

P A 3 - 1 2 3

TP M M+ TP M M+ Tr Tr

TP M M+ Tr

FIG. 4. Thylakoid insertion assay using a chloroplast lysate. Analyses of p+4 (left), wild-type pLHCP (ut, middle), and pA3 (right) are shown. The radiolabeled translation products ( T P ) are in lanes 1. The precursors were incubated with lysed chloroplasts (see “Ex- perimental Procedures” and Cline, 1988), and the thylakoid membranes were washed with 0.33 X HSM and isolated by centrifugation at 16,000 X g. Shown are the total radiolabeled precursor localized to the membranes ( h e s 2 ) and the remaining protein after treatment of the thylakoids with trypsin (Tr ) at 100 pg/ml (lanes 3).

the size of the peptide cleaved from p+4a, the processing site is distinct from the downstream site cleaved when pA3 is the substrate (Fig. 2). Significantly, when p+4a was incubated with a chloroplast-soluble extract in the organelle-free processing reaction, no cleavage occurred (Fig. 5C).

DISCUSSION

In the present study, we show that two cleavage events can be separated during the import of the LHCP precursor into the chloroplast by mutations at the transit peptide-mature protein junction. Thus, we conclude that the heterogeneity of LHCP associated with PSI1 arises in part from proteolytic processing of a single precursor substrate as well as from a multigene family found in many higher plants (Coruzzi et al., 1983; Lamppa et al., 1985; Dunsmuir, 1985; Sheen and Bogo-

A WT P+4 r I I I

1 2 3 4 5 1 2 3 4 5

0 0.

TP M M+ S S+ TP M M+ S S + Th Th

B WT P+4 I I t I

1 2 3 4 1 2 3 4 -7.-

Th T t i

C

t 2

A B kDa 1 2 3 1

C 2 3 1 2

TP M M+ Th Th Th Th

M+ M+ M+

FIG. 5. Analysis of import and processing of the insertion mutant p+4a. Panel A shows the results of an import reaction. The translation products (lane I ) , the total membrane fraction after import (lane Z), and the membrane fraction of chloroplasts treated with thermolysin before lysis (lane 3) are presented. Abbreviations are as described in Fig. 2. Panel B compares the membrane fractions isolated from chloroplasts that were treated with thermolysin after an import reaction. The precursor used in the reaction was either p+4 (lane 1 ), p+4a (lane 2) , or wild-type pLHCP (lane 3). Panel C presents the results of an organelle-free processing reaction for p+4a before (lane 1 ) and after (lane 2 ) incubation with the chloroplast- soluble fraction.

rad, 1986). When the wild-type precursor is imported into the chloroplast in vitro, a 25-kDa peptide and 26-kDa peptide are formed. Although the 26-kDa form is usually the most abundant, the relationship between these peptides has been un- clear. The results presented here indicate that they arise from independent processing events that remove the transit peptide at two sites approximately 1 kDa apart. Deletion of the primary cleavage site a t Met34 in mutant pA3 affects only the production of the 26-kDa peptide, which is then absent upon import. However, the precursor is still imported and cleaved at a secondary site, downstream about 8-10 residues, giving rise to a 25-kDa peptide. Furthermore, pA3 also is processed in the organelle-free reaction by a chloroplast-soluble enzyme, again yielding a 25-kDa peptide. Conversely, when the transit peptide-mature protein domain is disrupted in mutant p+4a by the insertion of 4 amino acids, only a 26-kDa peptide is

Import and Processing of the Chlorophyll alb Protein Precursor 17549

released, indicating that the secondary cleavage site becomes inaccessible to the processing enzyme. This precursor is not a cleavable substrate in the organelle-free reaction, nor was the insertion mutant p+4.

The correlation between a mutation that affects cleavage at the secondary site during import and a loss of processing in the organelle-free reaction emphasizes that the conditions of this reaction are specific for cleavage at the secondary site of pLHCP only, However, as we have shown previously, pS is cleaved under identical conditions (Abad et al., 1989), and the processing enzyme has properties that are very similar to the enzyme that cleaves the precursors of both the small subunit and plastocyanin (Robinson and Ellis, 1984a, 1984b), indicating that it is a bona fide transit peptidase. Although a number of parameters have been varied, production of the 26-kDa peptide from pLHCP in the organelle-free reaction has not yet been observed. Importantly, there is no accumulation of the 26-kDa peptide in the organelle-free reaction when cleavage at the secondary sit.e is impaired, and thus we conclude that it is not an intermediate that normally gives rise to the 25-kDa peptide. Using pea pLHCP as a substrate, we also find that only a 25-kDa peptide is released! The organelle- free reaction may lack an essential component that confers upon the precursor a conformation favorable for cleavage at both sites upon import into the chloroplast. In regard to pLHCP processing, most obviously, the reaction lacks a thylakoid membrane component. We have shown here using the precursor mutant p+4 and it has been observed using barley etioplasts with an inactive processing enzyme (Chitnis et al., 1988) that the precursor can be translocated across the envelope without cleavage and can insert into the thylakoids in the correct orientation as judged by resistance to trypsin digestion. The precursor will also insert into the thylakoids in a chloroplast lysate (see Fig. 4 and Cline, 1986, 1988). However, it is not yet resolved whether the precursor in vivo normally inserts into the thylakoids before, concurrent with, or after processing. An interesting possibility is that interac- tion with the thylakoid lipid bilayer orients the precursor for preferential cleavage at one site, and thus, the organelle-free reaction, which reconstitutes processing in the absence of membranes, can only yield the 25-kDa peptide. Lending sup- port to this idea, which remains to be established, is the o~servation by ourselves2 and others (Cline, 1988; Viitanen et al., 1988) that direct insertion of pLHCP into the thylakoids, using a chloroplast lysate, results in processing, and both mature forms are produced under some conditions. Finally, we cannot rule out the possibility that there are in fact two enzymes involved in pLHCP maturation, and in the organelle- free assay only one is active or found.

A model for the organization of LHCP associated with PSII has been presented recently (Larsson et at., 1987a, 1987b; Anderson and Andersson, 1988) in which the two major forms of LHCP in pea, estimated to be 25 and 27 kDa four estimate is 25 and -26 kDa), make up the peripheral antennae of the PSII light-harvesting complex. The ratio of these two peptides and their relative location in the thylakoids change during development and in response to the light environment (Lars- son et al., 1987a, 1987b). Most interestingly, the 25-kDa peptide appears to be preferentially phospho~lated under certain light regimes (Larsson and Andersson, 1985; Kuhl- brandt and Barber, 1988) and migrates more rapidly from the grana1 stacks (Larsson et ai., 1987a), theoretically to redis- tribute light energy between PSII and PSI (see Bennett, 1984; Anderson and Andersson, 1988 for reviews). The identifica- tion of two separable processing sites in pLHCP using an in

M. Abad, unpublished results.

vitro assay provides insight into the physiological role of the 25- and 26-kDa peptides associated with PSII and how the ratio of the two may be regulated in vivo. Knowing that the primary cleavage site occurs at Met3*, we estimate that the secondary sites occurs 8-10 amino acids (-1 kDa) downstream based on the relative mobility of the two mature forms. Therefore, the 25-kDa peptide would lack the 4 positively charged amino acids ( A r p , L Y S ~ ~ , Lys40, and LyP; see Fig. 1) at the NH2 terminus, the major feature of LHCP thought to be involved in thylakoid stacking (Mullet, 1983). The absence of the positively charged residues would also make the 25-kDa peptide more susceptible to the charge repulsion effect of‘ phosphorylation which accompanies peripheral

A - + .5) + + D-

B

I+ Random Coll A : Wheat pLHCP

B : PA3 1 4 Beta-Sheet

c : p+4 p i Alpha-Helix D : p+4alpha

FIG. 6. Secondary structure analysis of amino acid sequence at the junction of the transit peptide and mature protein for pLHCP and the mutant precursors. The entire amino acid sequence of each precursor was analyzed using the Garnier- Osguthorpe-Robson method (Garnier et al., 1978; see “Experimental Procedures”). Shown are the amino acids at the transit peptide- mature protein junction: A, residues 22-48 for wild-type pLHCP; B, residues 22-45 for pb3; C , 22-52 for p+4; and D, 22-52 for p-t-4~~. The positions of predicted random coil, p-sheet, and or-helix are indicated schematically using the key at the bottom of the figure. No turns were predicted in this region. Amino acids in the single letter code are shown above the region of predicted structure they compose. A closed circle is above methionine 34, and the basic (+) and acidic (- 1 amino acids are indicated.

17550 Import and P r o ~ e s s ~ ~ of the ~ h l o r o ~ ~ y l l a/b Protein Precursor

LHCP migration (Kyle et al., 1984). We propose as a working hypothesis that the selective cleavage of pLHCP at either the primary or secondary cleavage site plays an important role in the relative contribution of the 25- and -26-kDa forms found in the peripheral population of LHCP surrounding PSII.

An analysis of the secondary structure of the transit peptide-mature protein junction of wheat pLHCP is shown in Fig. 6, using the method of Gamier-Osguthorpe-Robson which takes into account neighboring residue interactions (Garnier et al., 1978). This domain is predicted to form a strong a- helix with the primary cleavage site at Met34 centrally located in not only wheat pLHCP but also in pLHCP from pea (Pongor et al., 1985), petunia, and corn (not shown). Based on these predictions, our data suggest that a-helicity may be essential for efficient processing of pLHCP at the primary cleavage site. Disruption of the a-helix in the mutant p+4 resulted in a marked reduction in processing a t both sites during import, but cleavage at the primary site was restored in p+4a when the predicted a-helix was replaced. On the other hand, the 4 amino acids inserted into the junction domain of p+4a abolished processing at the downstream site. The secondary structure predictions allow us to design further experiments to test the importance of specific residues and the conformation they favor in the junction domain.

The loss of processing at the secondary cleavage site of p+4a as well as the poor processing of p+4 may be due to the change in the relative distribution of the basic amino acids at the junction, or alternatively, to the separation of the site from processing determinants located in the transit peptide. Transit peptide modifications have been identified in pS which severely disrupt processing during import. The residues Ile-Thr-Ser near the carboxyl terminus of the transit peptide, minus 10 amino acids from the cleavage site, seem to be particularly critical (Wasmann et at., 1988; Ostrem et al., 1989). Although this sequence is not found in the transit peptide of pLHCP, amino acids with similafproperties are located in this region. For example, Ser-Ser-Ala-Val occur at positions 23-26 (see Figs. 1 and 6). Mutations in the leader peptides of proteins destined for the mitochondria have also been described which result in a loss of processing by the matrix protease (Nguyen et al., 1987; Kraus et aL, 1988). The mutations described in the present study left the transit peptide intact, yet they too had a significant effect on both the efficiency of processing and the site selected for cleavage, suggesting that interactions between the transit peptide and the NH2 terminus of the mature protein confer a conformational motif essential for processing. Successful cleavage of the precursor depends on both recognition by the processing enzyme, i.e. transit peptidase, and hydrolysis of the peptide bond, each step undoubtedly determined by the overall conformation of the precursor polypeptide.

In summary, the mutations described in this report have allowed us to establish that pLHCP contains two functional, separately mutable, cleavage sites. Although both sites are accessible to proteolytic processing during in uitro import into the chloroplast, only the secondary site is cleavable in the organelle-free reaction lacking most notably a membrane component. A prediction from our results is that pLHCP may enter one of two pathways upon import into the chloroplast, thereby giving rise to either the usually more abundant 26- kDa peptide or the 25-kDa form, depending on the physiological state of the organelle. We are currently performing ex-

periments to address this question.

Acknowledgment-We wish to thank Roben Buell for her excellent technical assistance.

REFERENCES Abad, M., Clark, S., and Lamppa, G. (1989) Plant Physiol. 90, 117-

124 Anderson, J. A., and Andersson, B. (1988) Trends Biochem. Sci. 13 ,

Bartlett, S., Grossman, A. R., and Chua, N. H. (1982) in Methods in Chloroplast Biology (Edelman, M., Hallick, R. B., and Chua, N.-H., eds) pp. 1081-1091, Elsevier Science Publishing Co., New York

Bennett, J. (1984). Physio~. Plant. 60,583-590 Cashmore, A. (1985) Proc. Natl. Acad. Sci. U. S. A. 8 1 , 2960-2964 Chitnis, P. R., Morishige, D. T., Nechushtai, R., and Thornber, J. P.

Cline, K. (1986) J. Biol. Chem. 261,14804-14810 Cline, K. (1988) Plant Physiol. 86, 1120-1126. Coruzzi, G., Broglie, R., Cashmore, A,, and Chua, N.-H. (1983) J.

Darr, S. C., Somerville, S. C., and Arntzen, C. d. (1986) J. Cell. Biol.

Dietz, K.-J., and Bogorad, L. (1987) Plant Physiol. 85,816-822 Dunsmuir, P. (1985) Nucleic Acids Res. 13, 2503-2518 Garnier, J., Osguthorpe, D. J., and Robson, B. (1978) J. MOL BioL

Kohorn, B. D., Harel, E., Chitnis, P. R., Thornber, J. P., and Tobin,

Kraus, J. P., Novotny, J., Kalousek, F., Swaroop, M., and Rosenberg,

Kuhlbrandt, W., and Barber, J. (1988) Biochim. Biophys. Acta 9 3 4 ,

Kyle, D. J., Kuang, T.-Y., Watson, J. L., and Arntzen, C. J. (1984)

Lamppa, G. K. (1988) J. Biot. Chem. 263,14996-14999 Lamppa, G. K., and Abad, M. S. (1987) J . Cell. Biol. 105,2641-2648 Lamppa, G. K., Morelli, G., and Chua, N.-H. (1985) Mol. Cell. Biol.

Larsson, U. K., and Andersson, B. (1985) Biochim. Biophys. Acta

Larsson, U. K., Sundby, C., and Andersson, B. (1987a) Biochim.

Larsson, U. K., Anderson, J. M., and Andersson, B. (198713) Biochim. Biophys. Acta 894,69-75

Maniatis, T., Fritsch, E. F., and Sambrook, J. (1982) Molecular Cloning: A Laboratory Manual, Cold Spring Harbor Laboratory, Cold Spring Harbor, NY

351-355

(1988) Plant. Mol. Biol. 11,95-I07

Bioi. Chem. 258,1399-1402

103,733-740

120,97-120

E. M. (1986) J. Cell Biol. 102,972-981

L. E. (1988) Proc. Natl. Acad. Sci. U. S. A. 86,8905-8909

118-122

Biochim. Biophys. Acta 765,89-96

5,1370-1378

809,396-402

Biophy~. Acta 894, 59-68

Mullet, J. E. (1983) J. Bid. Chem. 258,9941-9948 Nguyen, M., Argan, C., Sheffield, W. P., Bell, A. W., Shields, D., and

Shore, G. C. (1987) J . Cell Biol. 104 , 1193-1198 Ostrem, J. A., Ramage, R. T., Bohnert, H. J., and Wasmann, C. C.

(1989) J. Biol. Chem. 264,3662-3665 Pichersky, E., Hoffman, N. E., Malik, V. S., Bernatzky, R., Tanksley,

S. D., Szabo, L., and Cashmore, A. J. (1987) Phnt Mol. Biol. 9, 109-120

Pongor, S., Gutteiri, M., Cohen, L. M., and Szalay, A. (1985) DNA

Reiss, B., Wasmann, C. C., Schell, J., and Bohnert, H. J. (1989) Proc.

Robinson, C., and Ellis, R. J. (1984a) Eur. J. Biochem. 142,337-342 Robinson, C., and Ellis, R. J. (1984b) Eur. J. Biochem. 142,343-346 Sanger, F., Nicklen, S., and Coulsen, A. R. (1977) Proc. Natl. Acad.

Schmidt, G. W., Bartlett, S. G., Grossman, A. R., Cashmore, A. R.,

Sheen, J.-Y., and Bogorad, L. (1986) Proc. Natl. Acad. Sci U. S. A.

( N Y ) 4,319-326

Natl. Acad. Sci. U. S. A. 86,886-890

Sci. U. S. A. 74,5463-5467

and Chua, N.-H. (1981) J. Cell Bid. 9 1 , 468-478

83.7811-7815 Viitanen, P. V., Doran, E. R., and Dunsmuir, P. (1988) J. BioZ. Chem.

Wasmann, C. C., Reiss, B., and Bohnert, H. J. (1988) J. Biol. Chem. 263, 15000-15007

263,617-619 Wolf, H., Modrow, S., Motz, M., Jameson, B. A., Hermann, G., and

Foertsch, B. (1988) Comput. Appl. Biosci. 4, 187-192

mutations at the transit peptide-mature protein junction separate two cleavage events during

Documents