6
THE JOURNAL OF BIOLOGIOAI. CHEMISTRY Q 1994 by The American Society for Biochemistry and Molecular Biology, Inc Vol. 269, No. 28, Issue of July 8, pp. 17918-17923, 1994 Printed in U.S.A. Synthesis and Turnover of Photosystem XI Reaction Center Protein Dl RIBOSOME PAUSING INCREASES DURING CHLOROPLAST DEVELOPMENT" (Received for publication, January 24, 1994, and in revised form, March 15, 1994) Jungmook Kim, Patricia Gamble Klein$, and John E. Mulletg From the Department of Biochemistry and Biophysics, Texas A & M University, College Station, Ikxas 77843-2128 and the $Department of Horticulture and Landscape Architecture, University of Kentucky, Lexington, Kentucky 40517-0091 The chloroplast photosystem 11 reaction center pro- tein Dl contains five membrane-spanning helices and binds chlorophyll, carotenoid, quinone, iron, and prob- ably manganese. Turnover of pulse-labeled Dl in iso- lated plastids was found to involve cleavage between helix IV and helix V, which releases a 23-kDa N-terminal peptide and two C-terminal peptides of 10 and 8 kDa. Ribosomes pause at specific sites during translation of Dl, which results in the accumulation of Dl translation intermediates. Pulse-labeling assays followed by poly- some isolation and immunoprecipitation identified paused Dl translation intermediates of 9,12.5,15-18,20, 21,24, and 28-32 kDa. Ribosome pausing was not altered when dark-grown seedlings were illuminated for up to 1 h, even though this treatment stimulated accumulation of chlorophyll and Dl. However, illumination of plants for 16-72 h resulted in increased ribosome pausing and the build-up of Dl translation intermediates. We hypoth- esize that ribosome pausing during synthesis of Dl im- proves the efficiency of chlorophyll binding to Dl nas- cent chains and enhances accumulation of Dl in mature chloroplasts, which have reduced rates of chlorophyll biosynthesis. Photosystem I1 is one of four large multisubunit protein com- plexes found in the chloroplast thylakoid membrane. The func- tion of this complex is to extract electrons from water using photon energy and to transfer them to plastoquinone in the photosynthetic electron transport chain. Primary charge sepa- ration in photosystem TI (PSII)' occurs within the reaction cen- ter heterodimer, Dlm2. Dl and D2 are encoded by the chloro- plast genes psbA and psbD, respectively. Dl shares amino acid sequence homology with D2 and the bacterial reaction center protein L. D l D 2 binds chlorophyll, pheophytin, quinone, ca- rotenoid, iron, and probably manganese (Mattoo et ul., 1989). Based on functional and sequence homology with the L subunit (Allen et al., 1987; Deisenhofer and Michel, 1989; Deisenhofer et at., 1985; Feher et al., 1989;Mattoo et aL, 1989;Trebst, 1986; Yeates et al., 1987)and immunolo~cal data (Sayer et aZ., 19861, Dl and D2 are predicted to contain five transmembrane cy-hel- ices. Some of the cofactors associated with the photosynthetic reaction centers are positioned within the heterodimer (Deisen- hofer and Michel, 1989). Therefore, it seems likely that these GM37987. The costs of publication of this article were defrayed in part * This research was supported by National Institutes of Health Grant by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. 5 To whom correspondence should be addressed. Tel.: 409-845-0722; Fax: 409-845-9274. The abbreviation used is: PSII, photosystem 11. cofactors associate with the individual subunits of the reaction center prior to protein dimerization. In mature chloroplasts, the majority of Dlm2 is located in grana stacks within hnc- tioning photosystem I1 complexes (Callahan et al., 1989). In contrast, synthesis of Dl and D2 occurs in unstacked mem- brane regions (Adir et al., 1990). Therefore, it has been pro- posed that Dl and D2 are synthesized on stroma lamellae, bind cofactors, associate with each other and other photosystem I1 proteins, and subsequently move into grana stacks (Adir et al., 1990). During this process, the N terminus of Dl is modified (Michel et al., 1988),the D l C terminus is truncated (Marder et al., 1984), and the protein undergoes reversible palmitoylation (Mattoo and Edelman, 1987). Dark-grown plants lack Dl (Klein and Mullet, 1987), even though psbA mRNA accumulates and is associated with poly- somes (Klein et al., 1988333 Mullet et al., 1990). When plastids from dark-grown plants are pulse-labeled, D l translation in- termediates and degradation products are radiolabeled, but little full-length Dl accumulates (Mullet et at., 1990).Accumu- lation of Dl is dependent on light-induced chlorophyll synthe- sis (Eichacker et al., 1990; Kleinand Mullet, 1986, 1987; Klein et al., 1988a). Chlorophyll stimulates the accumulation of Dl and other chlorophyll proteins by increasing chlorophyll ap- oprotein stability (Mullet et al., 1990; Kim et al., 1994). D l stability is also dependent on other cofactors, such as carote- noids (Herrin et al., 1992). Pulse labeling and toeprint analysis showed that ribosomes pause a t specific sites during synthesis of D l (Kim et al., 1991).Ribosome pausing results in the accu- mulation of polysome-associated D l translation intermediates of discrete sizes (Kim et al., 1991). Plastid protein synthesis is high during the first 6 1 2 h of light-induced chloroplast development in barley (Klein and Mullet, 1987). However, once mature chloroplasts are formed, the rate of synthesis of most plastid proteins declines (Klein and Mullet, 1987).In contrast, Dl and D2 synthesis in mature barley chloroplasts remains high (Klein and Mullet, 1987). Continued high rates of Dl and D2 translation in mature chlo- roplasts are related tolight-induced damage of these proteins, which necessitates their turnover and resynthesis if photosys- tem SI is to remain functional (for review, see Mattoo et al. (19891, Barber and Anderson (1992), and Christopher et al. (1994)).Analysis of D l turnover revealed degradation products of 23, 17, 16.5, 14, and 8-12 kDa and a protease cleavage site between helix IV and helix V (Greenberg et al., 1987; Shipton and Barber, 1991; Wettern and Galling, 1985; Marder et al., 1984). In this study we further investigated the relationship be- tween ribosome pausing, D l translation intermediates,and Dl turnover. We report that light-induced chlorophyll synthesis does not alter ribosome pausing on psbA mRNA. However, ri- bosome pausing increased during chloroplast maturation in parallel with increased abundance of Dl translation interme- 17918

BIOLOGIOAI. No. July 8, pp. OF Biology, Inc for U.S.A. in … · helix IV and helix V, which releases a 23-kDa N-terminal peptide and two C-terminal peptides of 10 and 8 kDa

  • Upload
    lydung

  • View
    215

  • Download
    1

Embed Size (px)

Citation preview

THE JOURNAL OF BIOLOGIOAI. CHEMISTRY Q 1994 by The American Society for Biochemistry and Molecular Biology, Inc

Vol. 269, No. 28, Issue of July 8, pp. 17918-17923, 1994 Printed in U.S.A.

Synthesis and Turnover of Photosystem XI Reaction Center Protein Dl RIBOSOME PAUSING INCREASES DURING CHLOROPLAST DEVELOPMENT"

(Received for publication, January 24, 1994, and in revised form, March 15, 1994)

Jungmook Kim, Patricia Gamble Klein$, and John E. Mulletg From the Department of Biochemistry and Biophysics, Texas A & M University, College Station, Ikxas 77843-2128 and the $Department of Horticulture and Landscape Architecture, University of Kentucky, Lexington, Kentucky 40517-0091

The chloroplast photosystem 11 reaction center pro- tein D l contains five membrane-spanning helices and binds chlorophyll, carotenoid, quinone, iron, and prob- ably manganese. Turnover of pulse-labeled D l in iso- lated plastids was found to involve cleavage between helix IV and helix V, which releases a 23-kDa N-terminal peptide and two C-terminal peptides of 10 and 8 kDa. Ribosomes pause at specific sites during translation of Dl, which results in the accumulation of D l translation intermediates. Pulse-labeling assays followed by poly- some isolation and immunoprecipitation identified paused D l translation intermediates of 9,12.5,15-18,20, 21,24, and 28-32 kDa. Ribosome pausing was not altered when dark-grown seedlings were illuminated for up to 1 h, even though this treatment stimulated accumulation of chlorophyll and Dl. However, illumination of plants for 16-72 h resulted in increased ribosome pausing and the build-up of D l translation intermediates. We hypoth- esize that ribosome pausing during synthesis of D l im- proves the efficiency of chlorophyll binding to Dl nas- cent chains and enhances accumulation of Dl in mature chloroplasts, which have reduced rates of chlorophyll biosynthesis.

Photosystem I1 is one of four large multisubunit protein com- plexes found in the chloroplast thylakoid membrane. The func- tion of this complex is to extract electrons from water using photon energy and to transfer them to plastoquinone in the photosynthetic electron transport chain. Primary charge sepa- ration in photosystem TI (PSII)' occurs within the reaction cen- ter heterodimer, Dlm2. Dl and D2 are encoded by the chloro- plast genes psbA and psbD, respectively. Dl shares amino acid sequence homology with D2 and the bacterial reaction center protein L. DlD2 binds chlorophyll, pheophytin, quinone, ca- rotenoid, iron, and probably manganese (Mattoo et ul. , 1989). Based on functional and sequence homology with the L subunit (Allen et al . , 1987; Deisenhofer and Michel, 1989; Deisenhofer et at . , 1985; Feher et al., 1989; Mattoo et aL, 1989; Trebst, 1986; Yeates et al., 1987) and immunolo~cal data (Sayer et aZ., 19861, Dl and D2 are predicted to contain five transmembrane cy-hel- ices. Some of the cofactors associated with the photosynthetic reaction centers are positioned within the heterodimer (Deisen- hofer and Michel, 1989). Therefore, it seems likely that these

GM37987. The costs of publication of this article were defrayed in part * This research was supported by National Institutes of Health Grant

by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

5 To whom correspondence should be addressed. Tel.: 409-845-0722; Fax: 409-845-9274.

The abbreviation used is: PSII, photosystem 11.

cofactors associate with the individual subunits of the reaction center prior to protein dimerization. In mature chloroplasts, the majority of Dlm2 is located in grana stacks within hnc- tioning photosystem I1 complexes (Callahan et al . , 1989). In contrast, synthesis of Dl and D2 occurs in unstacked mem- brane regions (Adir et al . , 1990). Therefore, it has been pro- posed that Dl and D2 are synthesized on stroma lamellae, bind cofactors, associate with each other and other photosystem I1 proteins, and subsequently move into grana stacks (Adir et al . , 1990). During this process, the N terminus of Dl is modified (Michel et al., 1988), the D l C terminus is truncated (Marder et al . , 1984), and the protein undergoes reversible palmitoylation (Mattoo and Edelman, 1987).

Dark-grown plants lack D l (Klein and Mullet, 1987), even though psbA mRNA accumulates and is associated with poly- somes (Klein et al., 1988333 Mullet et al., 1990). When plastids from dark-grown plants are pulse-labeled, D l translation in- termediates and degradation products are radiolabeled, but little full-length Dl accumulates (Mullet et at., 1990). Accumu- lation of Dl is dependent on light-induced chlorophyll synthe- sis (Eichacker et al . , 1990; Klein and Mullet, 1986, 1987; Klein et al., 1988a). Chlorophyll stimulates the accumulation of Dl and other chlorophyll proteins by increasing chlorophyll ap- oprotein stability (Mullet et al . , 1990; Kim et al . , 1994). D l stability is also dependent on other cofactors, such as carote- noids (Herrin et al., 1992). Pulse labeling and toeprint analysis showed that ribosomes pause a t specific sites during synthesis of D l (Kim et al., 1991). Ribosome pausing results in the accu- mulation of polysome-associated D l translation intermediates of discrete sizes (Kim et al., 1991).

Plastid protein synthesis is high during the first 6 1 2 h of light-induced chloroplast development in barley (Klein and Mullet, 1987). However, once mature chloroplasts are formed, the rate of synthesis of most plastid proteins declines (Klein and Mullet, 1987). In contrast, D l and D2 synthesis in mature barley chloroplasts remains high (Klein and Mullet, 1987). Continued high rates of Dl and D2 translation in mature chlo- roplasts are related to light-induced damage of these proteins, which necessitates their turnover and resynthesis if photosys- tem SI is to remain functional (for review, see Mattoo et al. (19891, Barber and Anderson (1992), and Christopher et al. (1994)). Analysis of D l turnover revealed degradation products of 23, 17, 16.5, 14, and 8-12 kDa and a protease cleavage site between helix IV and helix V (Greenberg et al., 1987; Shipton and Barber, 1991; Wettern and Galling, 1985; Marder et al . , 1984).

In this study we further investigated the relationship be- tween ribosome pausing, D l translation intermediates, and Dl turnover. We report that light-induced chlorophyll synthesis does not alter ribosome pausing on psbA mRNA. However, ri- bosome pausing increased during chloroplast maturation in parallel with increased abundance of Dl translation interme-

17918

Synthesis of Reaction Center Protein D l 17919 diates. D l degradation products and D l translation interme- diates were identified providing a description of D l synthesis and turnover in barley chloroplasts.

MATERIALS AND METHODS Plant Growth and Plastid Isolation-Barley was planted in vermicu-

lite and grown at 23 "C in a dark chamber located in a light-tight room (Klein et al., 1988b). After 4.5 days in the dark, the seedlings were transferred to an illuminated chamber (350 microeinsteins/m%) for an additional 1,16, or 72 h. Plastids were isolated from the apical 3-cm leaf segment by Percoll gradient centrifugation as described previously (Klein and Mullet, 1986, 1987).

Protein Synthesis in Isolated Intact Plastids and Immunoprecipita- tion-ATP-driven protein synthesis by intact chloroplasts was con- ducted as previously described (Klein and Mullet, 1986). Intact plastids were added at a final concentration of 1.375 x lo7 plastids in a 75-pl reaction mixture containing ["Slmethionine and incubated a t 23 "C. In some experiments, pulse-labeled polypeptides were chased by the ad- dition of 5 mM unlabeled methionine. Following the labeling period, radiolabeled plastids were fractionated into membrane and soluble phases (Klein and Mullet, 1987), solubilized in SDS, and electropho- resed on 12-186 polyacrylamide gels containing 8 M urea. Gels were fixed and fluorographed (Mullet et al., 1986). In some cases, samples were immunoprecipitated with antibodies against the N terminus (amino acids 37-84) or C terminus (amino acids 154-320) of D l as previously described (Mullet et al., 1990).

Extension and Inhibition Analysis (Toeprintingl-Toeprinting reac- tions were done as previously described (Kim et al., 1991). Oligonucle- otides, which are complementary to nucleotides 825-844 (probe 5 in Kim et al. (1991)), numbered relative to the psbA mRNA 5'-end were used for determination of ribosome pausing. Fifty pmol of each syn- thetic oligonucleotide probe was 5'-end labeled with [y-32P]ATP in a 50-pl reaction using T4 polynucleotide kinase (Life Technologies, Inc.) under standard conditions. For toeprint analysis, membrane pellets or membrane-bound polysomes were prepared after lysis of 1 x 10" plas- tids as previously described (Kim et al., 1991). Samples were resus- pended in standard assay buffer containing 5'-end labeled probe (1.6 x lofi cpm) to make annealing mixtures (Kim et al., 1991). In vitro syn- thesized psbA mRNA (0.4 pmol) and plastid RNA extracted with phenol were used per 10 pl of annealing mixtures for sequencing reactions and toeprint control reactions. Extension inhibition reactions and sequenc- ing reactions of psbA mRNA were performed as previously described (Hartz et al., 1988; Kim et al., 1991). Samples were analyzed on 86 denaturing polyacrylamide gels and autoradiographed.

RESULTS

Ribosome Pausing on psbA mRNA versus Chloroplast Development-Toeprint assays were previously used to docu- ment that ribosomes pause during translation of D l helix IV (Kim et al., 1991). The location of the previously mapped ribo- some pause sites (numbered 23-36) and proposed chlorophyll binding sites in this region of D l are marked on the diagram in Fig. 1. Our earlier study of ribosome pausing utilized mature chloroplasts isolated from barley plants that had been illumi- nated for 72 h. In the present study, we analyzed ribosome pausing onpsbAmRNAusing plastids isolated from 4.5-day-old dark-grown plants and plants illuminated for up to 16 h in order to determine if chlorophyll or chloroplast development influenced ribosome pausing. The primer used for the toeprint assays allowed analysis of ribosomes translating the fourth membrane-spanning helix of the D l protein (Fig. 1). Lane 9 shows that protein-free psbA mRNA caused the reverse tran- scriptase to terminate at several locations, most likely due to RNA secondary structure (Fig. 1, representative bands are marked with asterisks). These same bands were observed when membrane-associated psbA mRNA was analyzed (Fig. 1, lanes 5-8). The intensity of the RNA-induced signals was similar in plastids isolated from dark-grown plants and those illuminated for up to 1 h, but it increased approximately 2-fold in plastids isolated from plants illuminated for 16 h (Fig. 1, lanes 5-7 versus lane 8). The increase in the intensity of the RNA-in- duced signals in plants illuminated for 16 h most likely reflects

Light,hr RNA Sequence1 ,,, IC

'U A C G ' O O .--o N. o w 0

"

*4 31

2 3 2 4 3 3

-35

-36

1 2 3 4 5 6 7 8 9

Helix IV

Thylakoid Membrane

D l proteln Lumen

FIG. 1. Toeprints showing ribosome pausing on psbA mRNA during chloroplast development. Intact plastids were isolated from seedlings grown in the dark for 4.5 days and then illuminated for 0, 0.25, 1.0, and 16 h. Plastids were fractionated into membrane and soluble phases, and membranes were isolated. psbA mRNA associated with membranes from each developmental stage and phenol-extracted plastid RNA (Con) were toeprinted using probe 5 (Kim et al., 1991). Toeprints of plastid membranes from plants illuminated for 0, 0.25, 1, and 16 h are shown in lanes 5, 6, 7, and 8, respectively. Asterisks indicate representative termination events most likely caused by RNA secondary structure (Kim et al., 1991). Toeprint signals that are en- hanced in the membrane ribosome fractions and not observed in pro- tein-free RNA samples are designated by arrows numbered from the 5'-end ofpsbA(as in Kim et al. (1991)). The dideoxy-RNAsequence of in vitro synthesized psbA RNA performed with toeprint probe 5 is shown in lanes 14. The diagram at the right of the toeprint gel shows the location of ribosome pause sites on the folding structure of t he D l polypeptide (Kim et al., 1991). CHL,, chlorophyll involved in the reac- tion center "special pair"; CHL,, accessory chlorophyll (Mattoo et al., 1989).

an increase in membrane-associated psbA mRNA. Toeprint assays of membrane-associated psbA mRNA pro-

duced termination products that were not observed in protein- free RNA samples (Fig. 1, bands marked with arrows 3136). These bands were previously shown to be induced by paused ribosomes (Kim et al., 1991). Plastid membranes from dark- grown plants or those illuminated for up to 1 h showed weak ribosome-induced toeprint signals compared with plastid mem- branes from plants illuminated for 16 h (Fig. 1, lanes 5-7 versus lane 8). These results indicate that light-induced chlorophyll biosynthesis, which occurs within 15 min of plant illumination and induces D l accumulation, did not alter ribosome pausing. However, ribosome pausing was increased in more mature chlo- roplasts of plants illuminated for 16 h to levels similar to those observed earlier in chloroplasts isolated from plants illumi- nated for 72 h (Kim et al., 1991).

Changes in Plastid Protein Synthesis versus Chloroplast Development-The increase in D l ribosome pausing during chloroplast development might cause detectable changes in D l protein synthesis. To investigate this possibility, chloroplasts were isolated from plants illuminated for 1, 16, or 72 h, and protein synthesis was carried out for 10 min in the presence of

17920 Synthesis of Reaction Center Protein D l

Membrane

Light, hr 1 16 72 -

CP43- p u u c . .

d23* Y

1 2 3 4 5 6 7 8 9 10

~ d 1 2 . 5 d12.5- .L

1 2 3 FIG. 2. Autoradiogram showing membrane polypeptides syn-

thesized by plastids during chloroplast development. Plastids were isolated from seedlings grown for 4.5 days in the dark and then illuminated for an additional 1, 16, or 72 h (lanes 1 3 ) . Protein synthe- sis in intact plastids was carried out for 10 min in the presence of ["Slmethionine. Plastids were fractionated into membrane and soluble phases, and membrane proteins were electrophoresed, and fluoro- graphed. The numbers d24, d23, d21, and d l 5 4 1 8 designate Dl-re- lated proteins. Other plastid proteins are designated as follows: P700, P700 chlorophyll apoproteins of PSI; CP43, 43-kDa PSII chlorophyll apoprotein; 0 2 , 34-kDa PSII reaction center polypeptide; D l , 32-kDa PSII reaction center polypeptide; pD1, precursor of Dl.

[35Slmethionine. Chloroplast membranes were subsequently isolated, and radiolabeled proteins were analyzed on SDS-poly- acrylamide gels (Fig. 2). Radiolabel accumulated in several previously identified proteins including P700, CP43, 02, the precursor to Dl (pDl) , and Dl (Fig. 2). CP43, D2, and Dl were synthesized in all three plastid populations. Several lower mo- lecular weight proteins, designated d24, d23, d21, and dl5418 were radiolabeled in all three plastid populations, although the relative amount of radiolabel incorporation in these proteins changed with chloroplast development. For example, d23 was radiolabeled more in plastids from plants illuminated for 1 h, whereas radiolabel incorporation into d24, d21, and d l5418 increased in the older plants. The increase in radiolabel incor- poration into d24, d21, and d l5418 i s correlated with the increase in ribosome pausing during D l synthesis observed in plants illuminated for 16-72 h (Fig. 1) (Kim et al., 1991).

Identification of D l Degradation Products and D l Paused D-anslation Intermediates-In an earlier study we identified paused polysome-associated D l translation intermediates of 28, 24, 21, and 15-18 kDa (Kim et al., 1991). Other investiga- tors have shown that Dl can be degraded into protein frag- ments of 23,17,16.5,14, and 8-12 kDa (Greenberg et al., 1987; Shipton and Barber, 1991; Wettern and Galling, 1985; Marder et al., 1984). Therefore, the radiolabeled proteins observed in Fig. 2 could be D l translation intermediates or D l degradation products. The origin of these products was investigated by pulse labeling chloroplasts isolated from 72-h illuminated plants for 2 min (Fig. 3, lane 1 ) followed by membrane isolation, disruption with detergents, and immunoprecipitation with an- tibodies against D l (Fig. 3, lane 2). Antibodies against the C terminus of D l (amino acids 154-320) immunoprecipitated pD1, Dl, and proteins labeled d24, d23, d21, d20, d10, and d8 (Fig. 3, lane 2). These antibodies recognized d l 5 4 1 8 only very weakly (observed in long term exposures). Antibodies against

FIG. 3. Autoradiograms showing D l degradation products and paused D l translation intermediates. Plastids were isolated from seedlings grown for 4.5 days in the dark and then illuminated for 72 h. Intact plastids were incubated with I"S1methionine for 2 min (lanes 1 - 5 ) . Plastids were fractionated, and radiolabeled membrane polypep- tides (lanes 1 and 3) were immunoprecipitated with antibodies against the C terminus of D l (lane 2 ) . Another set of radiolabeled plastids was lysed; membranes were isolated and dissolved in detergent. The solu- bilized membrane material was layered over a 1.5 M RNase-free sucrose cushion and centrifuged to pellet polysomes. Proteins associated with polysomes were immunoprecipitated with antibodies against the N ter- minus (lane 4 ) or C terminus (lane 5) of Dl. Intact plastids were pulse- labeled for 3 min with ["SJmethionine (lane 6 ) or preincubated without label for 10 min and then pulse-labeled for 3 min (lane 7). Plastids preincubated in translation buffer prior to labeling were pulse-labeled and lysed, and membranes were isolated (lane 8) and dissolved in detergent. Proteins associated with polysomes were isolated and immu- noprecipitated with the antibodies against the N terminus (lane 9 ) or the C terminus (lane 10) of Dl. Membranes or immunoprecipitated samples were electrophoresed and fluorographed. Dl-related proteins are designated d 9 4 3 2 . D l , 32-kDa PSII reaction center polypeptide; pD1, precursor of Dl .

the N terminus of D l (amino acids 37-84) immunoprecipitated d15418, d21424, pD1, and D l but not d l0 or d8 (data not shown). These results indicate that dl0 and d8 are Dl degra- dation products derived from the C terminus of Dl.

It has been shown previously that photodamage causes Dl to be proteolyzed into D l fragments of 23 and 8-12 kDa among other products. Therefore, d23, d10, and d8 observed in our experiments could be derived from proteolysis of newly trans- lated pDl/Dl. If d23, d10, and d8 are D l degradation products and d24, d21, and d l5418 a re paused translation intermedi- ates, then only the latter proteins will co-sediment with poly- somes through a sucrose pad. The result of this experiment is shown in Fig. 3, lanes 3-5. Polysomes and membrane-associ- ated proteins from pulse-labeled plastids (lane 3) were treated with detergent and centrifuged through a sucrose pad. The polysome pellet was denatured in SDS, diluted into Triton X-100, and immunoprecipitated with D l antibodies against the D l N terminus (Fig. 3, lane 4 ) or C terminus (Fig. 3, lane 5). The proteins d23, d10, and d8 did not co-sediment with poly- somes, consistent with their identification as proteolytic deg- radation products. In contrast, d24, d21, d l 5 4 1 8 and several additional proteins labeled d28432, d20, d12.5, and d9 co- sedimented with polysomes and were immunoprecipitated with D l antibodies. Based on this result, these latter proteins are tentatively identified as D l translation intermediates.

The D l proteolytic fragment d23 was abundant when chlo- roplasts from plants illuminated for 72 h were pulse-labeled for 2 min (Fig. 3, lane 1 ) but not when pulse labeling was carried out for 10 min or longer (Fig. 2, lanes 2 and 3) . This result suggested that d23 (and d l0 and d8) is generated in isolated plastids during the first several minutes of translation but not a t later times. To test this idea, plastids were either pulse- labeled for 3 min (Fig. 3, lane 6) or protein synthesis was

Synthesis of Reaction Center Protein D l 17921

- C A P

P 3 ' Chase ' Duration (rnin) 2 2 4 10

w

+CAP

' Chase 2 4 10

d23-

d10- d8-

1 2 3 4 5 6 7

tion of Dl proteolytic products, d23, dlO, and d8. Plastids from FIG. 4. Autoradiogram of pulse-chase assays showing degrada-

plants illuminated for 72 h were incubated with [3'S]methionine for 2 min (lane 1 ) and then chased with unlabeled methionine for 2 (lanes 2 and 5) , 4 (lanes 3 and 6) , and 10 min (lanes 4 and 7) in the presence (lanes 5-7) or absence (lanes 2-4) of chloramphenicol (CAP). Plastids were fractionated, and membrane proteins were electrophoresed and fluorographed.

carried out in the absence of radiolabel for 10 min and then plastids were pulse-labeled for 3 min (Fig. 3, lane 7). As de- scribed earlier, pD1, Dl , d24, d23, d21, d15418, d10, and d8 became radiolabeled when plastids were pulse-labeled without pretreatment. In contrast, pD1, Dl , d24, d21, and d l 5 4 1 8 were the primary D l translation products when plastids were allowed to translate for 10 min prior to pulse labeling (Fig. 3, lune 7). This result is consistent with d23, d10, and d8 being D l degradation products, which are generated during the first sev- eral minutes of protein synthesis in isolated chloroplasts. Fur- ther confirmation of this conclusion was obtained by radiola- beling plastid proteins after a 10-min preincubation, followed by separation of polysome-associated and polysome-free radio- labeled proteins by centrifugation through a sucrose cushion (Fig. 3, lanes 8 and 9). This experiment showed that d28432, d24, d21, d15418, d12.5, and d9 co-sedimented with poly- somes and were immunoprecipitated with antibodies against D l consistent with their designation as paused D l translation intermediates.

Evidence for Stable and Unstable Dl-The stability of d23, d10, and d8 was investigated using the pulse-chase assays shown in Fig. 4. In this experiment, plastids were radiolabeled for 2 min (lane 1) and then chased by the addition of unlabeled methionine in the presence (lanes 5-7) or absence (lanes 2 4 ) of chloramphenicol, an inhibitor of plastid protein synthesis. Dur- ing the pulse-labeling period, radiolabel accumulated primarily in pD1, Dl, d23, d10, and d8 (Fig. 4, lane 1). Radiolabel con- tinued to accumulate in pDl/Dl for at least 4 min into the chase period in the absence of chloramphenicol (lanes 2 and 3). This is most likely due to read-out of translation products ra- diolabeled during the pulse and to the time it takes to equili- brate externally added unlabeled methionine with the pool of radiolabeled methionine in intact plastids. Radiolabel persisted

L

- pD1

- D l

-d23'(N)

>-d12.5

L -d9

-dlO*(C) + -d8*(C)

FIG. 5. Model showing steps in Dl translation and turnover. The pDlD1 proteins and the Dl intermediates are shown by bars. [pDl] indicates an unstable D l polypeptide that is not associated with chlorophyll (or other cofactors) or is abnormally folded. Numbers indi- cate the approximate molecular masses (kDa) of D l intermediates. Cleavage of D l generates a 23-kDa N-terminal ( N ) polypeptide (d23+) and a 10-kDa C-terminal ( C ) polypeptide (dlOx). The 8-kDa C-terminal proteolytic product (d8* ) is proposed to be derived from the cleavage or processing of 10-kDa D l degradation product.

in d23 for 4 min into the chase period and then declined rapidly (lanes 2 4 ) . Radiolabel in d23 decreased after 10 min of chase, even though pD1 and D l remained heavily radiolabeled. This suggests that d23 was not being derived from the pool of pD1/ Dl, which accumulated radiolabel during the pulse-chase pe- riod. This conclusion was further supported when turnover of proteins was followed in the presence of chloramphenicol after the pulse-labeling period (Fig. 4, lanes 5-7). Radiolabel incor- poration in pDl/Dl was relatively stable in the presence of chloramphenicol. In contrast, radiolabel in d23 (and d l0 and d8) rapidly decreased, indicating that the half-life of d23 is very short and that continued translation is required to maintain the level of d23 observed in pulse-labeled plastids. In addition, this experiment indicates that there is a pool of pDl/Dl that is not subject to rapid degradation into d23, d10, and d8. Several unidentified proteins of 15-24 kDa accumulate during the chase. Some of these proteins may be derived from d23 or the other higher molecular weight proteins.

DISCUSSION Synthesis and lhrnover of Dl-A model summarizing steps

in Dl translation and turnover is shown in Fig. 5. Pulse-label- ing experiments described in this paper revealed D l transla- tion intermediates of 9, 12.5, 15-18, 20, 21, 24, and 28-32 kDa associated with membrane polysomes. Our previous study demonstrated a correlation between ribosome pausing and the presence of these D l translation intermediates (Kim et al., 1991). The 9- and 12.5-kDa D l translation intermediates were immunoprecipitated by antibodies against the Dl N terminus (amino acids 37-84) but not by those against the Dl C terminus (amino acids 154-320) (Fig. 3, lanes 4 and 9 versus lanes 5 and 10). The 12.5-kDa intermediate contains the first a-helix of Dl , which in mature Dl traverses the thylakoid membrane. How- ever, because the N terminus of Dl is located on the outside (or stromal side) of the thylakoid and the C terminus of the 12.5- kDa intermediate at this stage in translation is anchored on the

17922 Synthesis of Reaction Center Protein D l

same side of the membrane by the ribosome, it is unlikely that this helix crosses the membrane until the second membrane- spanning helix is released from the ribosome. In contrast, the intermediates d15-d18, d21, and d24 could be anchored to the membrane by membrane-spanning helices.

Pulse labeling and immunoprecipitation experiments re- vealed D1-related proteins of 23, 10, and 8 kDa that were not associated with polysomes. The 23-kDa polypeptide was immu- noprecipitated by antibodies against the N terminus and C terminus of D l (Fig. 3, lane 2 ) (Mullet et al., 1990). In contrast, dl0 and d8 were recognized only by antibodies against the C terminus of D l (Fig. 3, lane 2) . Therefore, these proteins most likely originate from cleavage of pDl/Dl between helix IV and V to release the 23-kDa N-terminal D l fragment d23 and the C-terminal D l peptides dl0 and d8 (Fig. 5). The D l C terminus is known to be processed to remove approximately nine amino acids (Diner et al. (1988); for review, see Barber and Anderson (1992)). The difference between dl0 and d8 may be the result of this proteolytic cleavage.

The D l degradation products described here are similar in molecular weight to those observed when D l is proteolyzed after photodamage. Light-induced D l degradation products of 23, 17, 16.5, 14, and 8-12 kDa have been reported (Greenberg et al., 1987; Shipton and Barber, 1991; Wettern and Galling, 1985; Marder et al., 1984). It is difficult to determine if the lower molecular weight turnover products observed in this study correspond exactly to those generated in response to pho- todamage. The apparent molecular weight of the radiolabeled D l fragments generated in our studies varied between 8 and 12 kDa depending on the gel system used (presence or absence of urea, percent acrylamide). Even so, it is clear that turnover of a portion of photodamaged D l and newly synthesized D l in- volves cleavage between helix IV and V.

When plastids from chlorophyll-deficient dark-grown plants are pulse-labeled, the D l turnover product, d23, accumulates radiolabel but not pD1 or D l (Mullet et al., 1990). Synthesis of chlorophyll in plastid membranes triggers the accumulation of D l and other chlorophyll proteins (Klein et al., 1988a; Eichacker et al., 1990). These and other results have led to the conclusion that chlorophyll binds to the D l apoprotein co-trans- lationally or shortly after release from the ribosome, thus sta- bilizing the apoprotein and allowing its accumulation (Mullet et al., 1990; Kim et al., 1994). The D l degradation product, d23, was observed in this study when plastids isolated from plants illuminated for 16 or 72 h were pulse-labeled for 2-3 min. However, when chloroplasts from illuminated plants were al- lowed to translate for 10 min prior to labeling or if the labeling period was 10 min in duration, much less d23 was observed. The half-life of d23 radiolabeled during the first 2-3 min of translation was very short (Fig. 4). These results suggest that during the first few minutes of translation in isolated plastids, a portion of the Dl synthesized is not associated with chloro- phyll (or other cofactors) or abnormally folded and, as a conse- quence, is rapidly degraded to d23, d10, and d8. At later times, the primary D l products labeled are pDl/Dl and D l transla- tion intermediates still associated with polysomes. The prote- ase labile D l products generated during the first few minutes of translation are derived by read-out of ribosomes that were in the process of translating D l i n vivo prior to plastid isolation. These ribosomes and associated nascent chains are trapped when plant leaves are quickly cooled prior to plastid isolation. Leaf chilling is followed by a 1-2-h plastid isolation procedure prior to translation assays. The process of plastid isolation could have caused a portion of the D l nascent polypeptides to fold abnormally or to become incompetent to bind chlorophyll. When these nascent chains are completed during the first few

minutes of the plastid translation assays, they would be rapidly degraded to yield d23, d10, and dB. The proteolytic products d23, d10, and d8 generated in this way are very unstable. Proteolytic products generated from functional D l during pho- todamage may have a much longer half-life due to correct pro- tein folding and cofactor association.

Modulation of Ribosome Pausing During Chloroplast Deuelopment-Ribosome pausing occurs during the translation of several plastid proteins (Kim et al., 1991; Kim and Hollings- worth, 1992). Ribosome pausing could be induced by RNA sec- ondary structure, RNA binding proteins, codon usage, interac- tion of nascent chains with ribosomes, or interaction between chaperoninshascent chains and the ribosome. Transient arrest of ribosome elongation by the signal recognition particle pro- vides one example of pausing induced by protein-protein inter- action (Lipp et al., 1987). In a previous study of Dl , ribosome pausing sites were mapped using toeprint analysis and corre- lated with polysome-associated D l translation intermediates. In this paper, we show that ribosome pausing during transla- tion of D l is not altered by light-induced chlorophyll biosyn- thesis, which allows accumulation of D l through protein sta- bilization. Because chlorophyll modulates protein stability late in translation and/or post-translationally (Mullet et al., 1990; Kim et al., 19941, it is not surprising that chlorophyll biosyn- thesis has little influence on ribosome pausing.

Physiological Significance of Ribosome Pausing-The in- crease in ribosome pausing between 1 and 16 h of illumination was correlated with a decrease in the abundance of the Dl turnover product, d23, and an increase in the abundance of paused translation intermediates d15-dl8, d21, and d24 in plastids labeled for 10 min (Fig. 2). The decrease in d23 gen- eration suggests that chlorophyll binding (and other cofactors) to D l is more efficient in the plants illuminated for 16-72 h. An increase in cofactor binding efficiency may be required to allow D l accumulation in mature chloroplasts. The rate of synthesis and accumulation of chlorophyll and chlorophyll proteins is high for at least 12 h after 4.5-day-old dark-grown barley plants are illuminated (Klein and Mullet, 1987). During this phase, high rates of accumulation of thylakoid membrane proteins are important in order to establish a functional photosynthetic ap- paratus, Between 16 and 36 h of illumination, chlorophyll stops accumulating, and the rate of synthesis of many chloroplast proteins declines (Klein and Mullet, 1987). However, the rate of synthesis of D l remains high in mature chloroplasts because these proteins are damaged as a function of photochemistry and must be degraded and resynthesized in order to maintain PSI1 activity. During the first 12 h of chloroplast development, chlorophyll and other cofactors needed for D l accumulation are readily available because of the rapid synthesis of the photo- synthetic apparatus. In contrast, in mature chloroplasts, chlo- rophyll and other cofactors needed for D l synthesis and accu- mulation may become limiting. Chlorophyll biosynthesis rates are low, and cofactor availability may depend to a large extent on the recycling of cofactors following protein turnover. There- fore, in mature chloroplasts, mechanisms that improve the ef- ficiency of coupling between protein synthesis and cofactor binding may be very important.

We hypothesize that ribosome pausing during the synthesis of D l in mature chloroplasts improves the efficiency of D l synthesis by providing additional time for nascent chains to bind cofactors such as chlorophyll prior to polypeptide release from the ribosomes. The importance of the paused D l transla- tion intermediates is demonstrated in the barley mutant uir- 115. Translation of D l is normal in uir-115 plants illuminated for 1 h. However, the D l translation intermediates d15-dl8, d21, and d24 are not observed later in chloroplast development,

Synthesis of Reaction Center Protein D l 17923

and accumulation of D l in mature chloroplasts is severely re- stricted, eventually leading to seedling death (Gamble and Mullet, 1989h2

REFERENCES

Adir, N., Shochat, S., and Ohad, I. (1990) J. B i d . Chem. 266, 12563-12568 Allen, J. P., Feher, G., Yeates, T. O., Komiya, H., and Rees, D. C. (1987) P m . Natl.

Barber, J., and Anderson, B. (1992) %rids Biochem. Sci. 17,61-66 Callahan, F. E., Wergin, W. P., Nelson, N., Edelman, M., and Mattoo, A. K. (1989)

Christopher et al. (1994) Plant Physiol. 104, 1119-1129 Deisenhofer, J., and Michel, H. (1989) EMBO J. 8, 2149-2170 Deisenhofer, J., Epp, O., Miki, K., Huber, R., and Michel, H. (1985) Nature 318,

Diner, B. A,, Ries, D. F., Coben, B. N., and Metz, J. G. (1988) J. Bid. Chem. 263,

Eichacker, L., SOU, J., Lauterbach, P., Rudiger, W., Klein, R. R., and Mullet, J. E.

Gamble, P. E., and Mullet, J. E. (1989) J. Biol. Chem. 264, 7236-7243 Feher, G., Allen, J. P., Okamura, M. Y., and Rees, C. (1989) Nature 339, 111-116

Greenberg, B. M., Gaba, V., Mattoo, A. K, and Edelman, M. (1987) EMBO J. 6,

Hartz, D., McPheeters, D. S., Traut, R., and Gold, L. (1988) Methods Enzymol. 164,

Acad. Sci. U. S. A. 84, 6162-6166

Plant Physiol. 91,629-635

618-624

8972-8980

(1990) J. Biol. Chem. 265,13566-13571

2865-2869

419-425

* Kim, J., Klein, P. G., and Mullet, J. E. (1994) Plant Mol. Biol., in press.

Herrin, D. L., Battey, J. F., Greer, K., and Schmidt, G. W. (1992) J. Biol. Chem. 267,

Kim, J. K., and Hollingsworth, M. J. (1992) Anal. Biochern. 206,183-188 Kim, J., Klein, P. G., and Mullet, J. E. (1991) J. Biol. Chem. 266, 14931-14938 Kim, J., Eichacker, L. A,, Riidiger, W., and Mullet, J. E. (1994) Plant Physiol. 104,

Klein, R. R., and Mullet, J. E. (1986) J. Biol. Chem. 261, 1113%11145 Klein, R. R., and Mullet, J. E. (1987) J. Biol. Chem. 262,43414348 Klein, R. R., Gamble, P. E., and Mullet, J. E. (1988a) Plant Physiol. 88,1246-1256 Klein, R. R., Mason, H. S., and Mullet, J. E. (1988b) J. Cell Biol. 106, 289-301 Lipp, J., Dobberstein, B., and Haeuptle, M.-T. (1987) J. Biol. Chern. 262, 1680-

Marder, J. B., Goloubinoff, P., and Edelman, M. (1984) J. Biol. Chem. 269,3900-

Mattoo, A. K, and Edelman, M. (1987) Proc. Natl. Acad. Sci. U. 5. A. 84,1497-

Mattoo, A. K, Marder, J. B., and Edelman, M. (1989) Cell 56,241-246 Michel, H., Hunt, D. F., Shabanowitz, J., and Bennett, J. (1988) J. Biol. Chem. 263,

Mullet, J. E., Klein, R. R., and Grossman, A. R. (1986) Eur. J . Biochem. 185,

Mullet, J. E., Klein, P. G., and Klein, R. R. (1990) Proc. Natl. Acad. Sci. U. S. A. 87,

Sayer, R. T., Anderson, B., and Bogorad, L. (1986) Cell 47,601-608

Trebst, A. (1986) Z. Naturforsch. 41, 240-245 Shipton, C. A,, and Barber, J. (1991) Proc. Natl. Acad. Sci. U. 5. A. 88,6691-6695

Wettern, M., and Galling, G. (1985) Planta 166,474-482 Yeates, T. O., Komiya, H., Rees, D. C., Allen, J. P., and Feher, G. (1987) Proc. Natl.

8260-8269

907-916

1684

3908

1501

1123-1130

331438

4038.4042

Acad. Sci. U. 5. A. 84, 6438-6442