Rubisco Assembly: A Model System for Studying the Mechanism of

The Plant Cell, Vol. 1, 1035-1 042, November 1989 O 1989 American Society of Plant Physiologists

Rubisco Assembly: A Model System for Studying the Mechanism of Chaperonin Action

Harry Roy Plant Science Group, Biology Department and Center for Biophysics, Rensselaer Polytechnic Institute, Troy, New York 121 80

INTRODUCTION

The chloroplast enzyme Rubisco (ribulose 1,5-bisphosphate carboxylase/oxygenase; EC 4.1.1.39) catalyzes the carboxylation of ribulose bisphosphate in the Calvin cycle and the oxygenation of the same substrate in the photo- respiratory pathway. Because these two competing reac- tions determine the net photosynthetic yield of many higher plants, it is important to understand their mechanisms. This requires not only a model of Rubisco atomic structure, but also experimental systems that allow the structure to be manipulated by in vitro mutagenesis. The requirement for a structural model has been fulfilled recently (Chapman et al., 1987, 1988; Andersson et al., 1989). Unfortunately, however, in vitro mutagenesis of higher plant Rubisco has not been possible because the subunits of higher plant Rubisco, either when dissociated from each other chemi- cally in vitro (Voordouw, Van der Vies, and Boumeister, 1984) or when synthesized in Escherichia coli (Bradley, Van der Vies, and Gatenby, 1986), fail to assemble into active enzyme. Severa1 groups have attempted to overcome this problem by studying Rubisco assembly from either native subunits (Roy, Cannon, and Gilson, 1988a), prokaryotic subunits expressed in E. coli (Gatenby, 1988), or alga1 subunits (Newman and Cattolico, 1988). These studies indicate that the post-translational assembly of Rubisco holoenzyme in both prokaryotes and eukaryotes is a complex process that may require helper proteins called “chaperonins” (Gatenby, 1988; Hemmingsen et al., 1988; Roy and Cannon, 1988; Roy et al., 1988a; Ellis, Van der Vies, and Hemmingsen, 1989). Chaperonins are func- tionally analogous to other molecular chaperones such as nucleoplasmin, which is involved in nucleosome assembly, and the bovine microsomal binding protein, which partici- pates in immunoglobulin transport (Pelham, 1986; Ellis et al., 1989). The molecular chaperones, including the chaperonins, are supposed to interact transiently with other proteins, promoting assembly of those proteins into functional complexes. The molecular chaperones do not form

part of the final structure of the protein complexes whose assembly they promote. In this article I review recent experiments that persuasively support the idea that chaperonins are required for Rubisco assembly.

RUBISCO AS A MODEL FOR ASSEMBLY STUDIES

Rubisco is an excellent model system for protein assembly studies for the following reasons:

(1) The structure is known. In higher plant Rubisco, the large subunits (53 kD) are arranged in an octameric core, and the small subunits (about 14 kD) occur in layers of four on opposite sides of the molecule (Chapman et al., 1987,1988; Andersson et al., 1989). X-ray crystallographic analysis has revealed the detailed structure of tobacco Rubisco in the absence of substrate (Chapman et al., 1987). More recently, the geometry of a transition state analog and its potential ligands in the active site of spinach Rubisco has been determined (Andersson et al., 1989). These research developments have created opportunities not only for studies on the role of specific amino acids at the active site, but also for understanding the nature of the amino acid residues that participate in interactions between subunits.

(2) Rubisco is very abundant (Ellis, 1979), and is synthesized at high rates, permitting the detection of intermediates in the assembly process (Barraclough and Ellis, 1980; Roy et al., 1982).

(3) Prokaryotic Rubisco assembly can be studied in E. coli using cloned genes from cyanobacteria and auto- trophic and chemoautotrophic bacteria (Gatenby, 1988; Roy et al., 1988a).

(4) As I discuss in detail in this article, higher plant Rubisco assembly can be studied in chloroplast extracts.

Dow

nloaded from https://academ

ic.oup.com/plcell/article/1/11/1035/5970258 by guest on 25 Septem

ber 2021

1036 The Plant Cell

SYNTHESIS AND ASSEMBLY OF HIGHER PLANT RUBISCO IN CHLOROPLAST EXTRACTS WATCHING A MOLECULAR CHAPERONE AT WORK

Rubisco large subunits are encoded by a single chloroplast gene (Mclntosh, Poulsen, and Bogorad, 1980), and are synthesized in chloroplasts on free ribosomes (Gooding, Roy, and Jagendorf, 1973; Blair and Ellis, 1974). Small subunits are encoded by a small family of nuclear genes (Tobin and Silverthorne, 1985), and are made as cytosolic precursors containing an NH,-terminal “transit” sequence. The small subunit precursors are imported into the chloroplast, cleaved to mature form, and assembled into hex- adecameric holoenzyme (Chua and Schmidt, 1978; Lub- ben, Theg, and Keegstra, 1988).

When protein synthesis is carried out for short periods in the presence of radioactive amino acids in isolated intact chloroplasts, Rubisco large subunits are essentially the only soluble radiolabeled product. During electrophoresis under nondenaturing conditions, the radioactive large subunits do not co-migrate with Rubisco holoenzyme. Instead, about 75% of the radioactive large subunits co-migrate with a stainable high molecular weight protein, the Rubisco subunit binding protein (subsequently defined as a “chaperonin”); the remainder show anomalous behavior during nondenaturing electrophoresis, but can be resolved by sucrose gradient centrifugation as lower molecular weight complexes (peak sedimentation coefficient about 7). After prolonged illumination under chase conditions, large subunit radioactivity is lost from both the high molecular weight and the low molecular weight complexes (Roy et al., 1982; Chaudhari and Roy, 1989), and radioactive large subunits co-migrate with Rubisco both during nondenaturing electrophoresis and sucrose velocity gradient centrifugation (Barraclough and Ellis, 1980; Roy et al., 1982). No incorporation of radioactive large subunits into Rubisco occurs under chase conditions in the dark in intact chloroplasts. However, ATP can substitute for light in promoting the assembly of large subunits into Rubisco in vitro (Bloom, Milos, and Roy, 1983).

The co-migration of radioactive large subunits with Rub- isco could be attributed to either de novo assembly of holoenzyme or an exchange reaction between the large subunits already assembled into Rubisco and the newly synthesized large subunits. Because of the presence of large amounts of endogenous Rubisco in chloroplast extracts, it is not possible to measure increased activity contributed by any newly assembled Rubisco molecules. Nevertheless, the incorporation of large subunits into Rub- isco can be stimulated by added small subunits (Roy, Chaudhari, and Cannon, 1988b). It is, therefore, very likely that genuine holoenzyme assembly is taking place in these chloroplast extracts. Nondenaturing electrophoresis can

resolve cyanobacterial Rubisco from its corresponding octameric large subunit core particles (Gatenby et al., 1988). Comparison of stained gels with fluorograms shows that the migration of radioactive higher plant large subunits is identical to that of the stained Rubisco. Soluble Rubisco large subunit core particles have not been detected in higher plants, but this could be due to a variety of factors, such as their presence below the limit of detection of fluorography, or the presence of such high amounts of radioactive Rubisco that the small amount of cores present cannot be seen. It is, therefore, unlikely that a high proportion of newly synthesized large subunits are incorporated into core particles. The vast majority of newly synthesized large subunits that co-migrate with Rubisco are, therefore, believed to be incorporated into Rubisco holoenzyme molecules in chloroplast extracts.

lncubation of dilute chloroplast extracts with ATP brings about dissociation of the high molecular weight complex of large subunit and binding protein (Bloom et al., 1983; Musgrove, Johnson, and Ellis, 1987). Low molecular weight complexes, either present from the beginning or released from the high molecular weight binding protein complexes by ATP treatment, provide large subunits for the in vitro assembly of Rubisco (Milos and Roy, 1984; Milos, Bloom, and Roy, 1985; Cannon, Wang, and Roy, 1986; Chaudhari and Roy, 1989). As diagrammed in the model for Rubisco assembly seen in Figure 1, these low molecular weight complexes could be large subunit dimers, or heterodimers containing one large subunit and one binding protein subunit. Antibody to the Rubisco subunit binding protein inhibits the incorporation of large subunits from the low molecular weight complexes into Rubisco (Cannon et al., 1986), indicating that all assembly-compe- tent large subunits are associated with the binding protein at least transiently after release from the high molecular weight binding protein complex. It has been suggested that heterodimers, consisting of one large subunit and one small subunit, are intermediates in the assembly of prokaryotic Rubisco (Gurevitz, Somerville, and Mclntosh, 1985). More recent work suggests that such heterodimers are unlikely intermediates in the assembly of cyanobacter- ia1 Rubisco in E. coli (see Gatenby, 1988, for discussion). There is no information on the occurrence Of this kind of heterodimer in higher plants.

Because assembly of large subunits into Rubisco is stimulated by illumination of chloroplasts or addition of ATP to chloroplast extracts-conditions that lead to dissociation of the high molecular weight binding protein complex-it is unlikely that the high molecular weight binding protein complex serves as a scaffold upon which assembly of Rubisco takes place. Although dissociation of the high molecular weight binding protein complex is sup- pressed at high protein concentrations comparable to those occurring in chloroplasts, it has been demonstrated that large subunits dissociate from this complex during

Dow



ber 2021

Rubisco and the Chaperonins 1037

Binding proteinprecursor

Ribosome Bindingprotein——-^ '

Large A /subunit V

/ATP flA* High molecular weight

large subunit - bindingprotein complex

Possible intermediates

Small subunitprecursor

Figure 1. Assembly of Higher Plant Rubisco within the Chloroplast.

Chloroplast ribosomes synthesize the Rubisco large subunit. The Rubisco subunit binding protein, which is related to the £. coli groELprotein (a "chaperonin"), is synthesized as a precursor in the cytoplasm. After processing in the Chloroplast, it forms a high molecularweight complex with the large subunit. This interaction may occur during or very shortly after synthesis of the large subunit. ATP producedby photophosphorylation, and perhaps other ATP-dependent factors [such as a hypothetical groES-like protein (not shown)], lead to therelease of large subunits as dimeric complexes with binding protein subunits or with other large subunits. The large subunits associatewith each other and with small subunits (also imported from the cytosol) in an uncharacterized series of steps leading to the formation ofhexadecameric Rubisco holoenzyme molecules. The binding protein itself is not incorporated into Rubisco holoenzyme. The figure showsheterodimers of large subunit and binding protein, large subunit dimers, and an octameric large subunit core particle. There is tentativeevidence for all these complexes, either in higher plants or in E. coli, but they remain uncharacterized in chloroplasts. Red = large subunit;blue = small subunit, small subunit precursor; black = binding protein, binding protein precursor; tan = ribosome.

illumination of intact chloroplasts (Roy, Hubbs, and Can-non, 1988c). Moreover, the relative concentration of thelow molecular weight large subunit complexes can beinfluenced by controlling the osmotic conditions underwhich Chloroplast protein synthesis takes place (Chaudhariand Roy, 1989). Thus, it appears likely that, despite therelative stability of the high molecular weight complexes,the low molecular weight large subunit complexes areavailable for Rubisco assembly in the intact Chloroplast.

Neither the binding protein nor the large subunits arephosphorylated or adenylylated by ATP in Chloroplast ex-tracts (Hemmingsen and Ellis, 1986; P. Milos and H. Roy,unpublished data). However, other types of post-transla-tional processing have not been ruled out, and it is known

that amino termini of large subunits are modified in manyspecies (Mulligan, Houtz, and Tolbert, 1988).

The Rubisco subunit binding protein has two types ofsubunits, « (61 kD), and 0 (60 kD), in roughly equivalentamounts (Hemmingsen and Ellis, 1986; Musgrove et al.,1987). These are synthesized from polyA-containing RNAas polypeptides of higher molecular weight, which presum-ably are transported from the cytosol into the Chloroplast(Hemmingsen and Ellis, 1986). Using antibodies to the peaRubisco subunit binding protein and immunoblotting,cross-reactive material of about 60 kD has been found inE. coli, Chromatium vinosum, several other photosyntheticbacteria, maize bundle sheath chloroplasts, spinach,wheat, castor bean, sycamore, and tobacco (Hemmingsen

Dow



ber 2021

1038 The Plant Cell

and Ellis, 1986; Macherel et al., 1986; Hemmingsen et al., 1988; -Torres-Ruiz and McFadden, 1988). In the chromo- phytic alga Olisthodiscus luteus, several polypeptides cross-react with the pea binding protein antibody (Newman and Cattolico, 1988). Apparently, proteins related to the binding protein are extremely widespread, if not ubiqui- tous. Most importantly, as expected of a molecular chaperone, there is no evidence that the binding protein itself is ever incorporated into Rubisco.

RUBISCO SUBUNIT BlNDlNG PROTEIN IS RELATED TO THE E. coli groEL PROTEIN, A CHAPERONIN

It has been demonstrated recently that the higher plant Rubisco subunit binding protein is 46% identical to an E. coli protein encoded by the groEL gene (Hemmingsen et al., 1988). Moreover, electron micrographs of the pea binding protein (S. Cannon and H. Roy, unpublished data; Pushkin et al., 1982) strongly resemble micrographs of the groEL protein, which exhibits a sevenfold rotational sym- metry and is believed to contain 14 identical subunits of about 60 kD (Hendrix, 1979; Hohn et al., 1979). Defects in groEL function have pleiotropic effects on RNA, DNA, and protein synthesis in E. coli (Wada and Itikawa, 1986), and deletion of groEL sequences is lethal at all temperatures (Fayet, Ziegelhoffer, and Georgopoulos, 1989). groEL function is also needed for post-translational assembly of bacteriophages T4, T5, and X (Hohn et al., 1979; Hendrix, 1979; Hemmingsen et al., 1988). The groEL gene is adjacent to a gene encoding the 10.3-kD groES protein. The groEL and groES proteins are synthesized at elevated rates during heat stress in E. coli. Proteins that cross-react with groEL protein antibodies have been found in Myco- bacterium spp. and Pseudomonas aeruginosa (Shinnick, Vodkin, and Williams, 1988), Caulobacter (Gomes et al., 1986), and Legionella pneumophila (Lema et al., 1988). These proteins apparently correspond to a previously iden- tified bacterial common antigen (Shinnick et al., 1988). The sequences of the Tetrahymena mitochondrial 58-kD heat shock protein and the yeast mitochondrial 60-kD heat shock protein are also highly similar to E. coli groEL protein (McMullin and Hallberg, 1988; Reading, Hallberg, and Myers, 1989). Based on the widespread occurrence of the

ROLE OF CHAPERONINS IN THE ASSEMBLY OF SOME OLlGOMERlC PROTEINS IN YEAST MITOCHONDRIA AND IN THE ASSEMBLY OF RUBISCO INE. coli

Recent experiments have provided convincing evidence of a requirement for one of the chaperonins in the assembly of several oligomeric proteins whose subunits are imported into mitochondria. Human ornithine transcarbamylase, cloned into yeast nuclear DNA, is imported from the cytosol into yeast mitochondria; the assembly of this trimeric enzyme in mitochondria has recently been shown to be dependent on the yeast mitochondrial chaperonin. In similar experiments, the assembly of cytosolically synthesized subunits of cytochrome b2, oligomeric yeast mitochondrial F1-ATPase, and the mitochondrial Rieske Fe/S protein have been shown to depend upon the mitochondrial chaperonin (Cheng et al., 1989). The chaperonin is thought to act shortly after the import of these proteins, but is not believed to catalyze removal of their transit peptides. Other types of post-translational processing by the chaperonin were not ruled out.

More interestingly for plant researchers, Goloubinoff, Gatenby, and Lorimer (1 989) have found that the formation of active cyanobacterial Rubisco in genetically engineered E. coli is dependent on the presence of functional E. coli groES and groEL genes. They also found that the cloned Rhodospirillum rubrum Rubisco, which consists of a dimer of large subunits and lacks small subunits, also requires both of the groE genes for synthesis of active holoenzyme in E. coli. Rubisco large subunits are synthesized in the presence of mutant groEL and groES genes, but are not assembled into active holoenzyme molecules unless nor- mal groEL and groES genes are present. It is not clear whether the assembling Rubisco subunits are covalently modified by the groEL or groES proteins. Nevertheless, the data support the idea that both of these proteins are required for Rubisco assembly in E. coli, and suggest that there may be a requirement for both groEL-like and groES- like activity for assembly of Rubisco in cyanobacteria. Because there is considerable evidence that chloroplasts are descended from endosymbiotic cyanobacteria, the apparent involvement of the binding protein in the assembly of Rubisco in higher plants appears to represent a conserved mechanism.

approximately 60-kD groEL-like proteins, and the evidence that some of these proteins are involved in the post- translational assembly of macromolecular complexes, Hemmingsen et al. (1988) proposed that the groEL class of proteins be called “chaperonins.” These are not postu- lated to work by covalently modifying proteins. Instead, they are supposed to prevent premature association of incompletely folded proteins. In fact, little is known about their mechanism of action.

THE M E ~ ~ A N I S M OF CHAPERONIN ACT~ON 1s NOT U N D E R S T O ~ ~

Roy and Cannon (1988) suggest that the Rubisco large subunit binding protein might interact with large subunit nascent chains. Evidence for association of groEL protein

Dow



ber 2021


with nascent chains has come from studies of in vitro synthesized p-lactamase and chloramphenicol acetyltrans- ferase (Bochkareva, Lissin, and Girshovich, 1988). These authors employed modified initiator tRNA molecules that had been covalently linked to a photoactivatable cross- linker and an E. coli extract that had been depleted of groEL protein. They added groEL protein to the extracts after the initiation of in vitro protein synthesis in the pres- ente of the modified tRNA, and then illuminated the sam- ple. This led to cross-linking of the in vitro initiated chains to groEL protein. When re-addition of groEL protein was delayed for various intervals af?er the initiation of protein synthesis, photochemical cross-linking of the nascent chains to groEL protein gradually declined. The authors inferred that the groEL protein interacts transiently with newly synthesized proteins. Based on results of experiments on the assembly properties of large subunits synthesized in chloroplasts under alternative osmotic conditions, Chaudhari and Roy (1989) suggest that the pea binding protein interacts with nascent Rubisco large subunits in intact chloroplasts. Further studies are needed to confirm the proposed interactions between nascent chains and chaperonins.

It is of interest to know whether chaperonins function while in their characteristic high molecular weight form, or in the dissociated state. Cheng et al. (1 989) propose that the mitochondrial chaperonin might act as a workbench upon which folding and assembly of polypeptide chains can occur. Consistent with this hypothesis, intact groEL protein particles have been seen attached to bacteriophage proheads (Hohn et al., 1’979). In contrast to this, it is clear that assembly of Rubisco can take place when the high molecular weight complex containing the Rubisco subunit binding protein has been dissociated in the pres- ente of ATP (Milos and Roy, 1984; Cannon et al., 1986). In this case, the binding protein interacts with the large subunits in low molecular weight complexes prior to assembly (Cannon et al., 1986). Thus, the reaction mechanisms of chaperonins may differ, depending on the organism. It is also of interest to know at what stage of the assembly process the chaperonin functions. For example, it is possible, even likely, that plant and bacterial chaperonins interact with Rubisco at several stages of the assembly process (Cannon et al., 1986; Goloubinoff et al., 1989).

Using small subunit precursor molecules, radiolabeled to a high specific activity and supplied to intact chloroplasts in vitro, it has been discovered recently that small subunits can associate with Rubisco large subunit binding protein (Gatenby et al., 1988; Ellis and Van der Vies, 1988). However, the kinetics of the association are not suggestive of a precursor-product relationship with Rubisco holoenzyme (Gatenby et al., 1988). Moreover, there is evidence that the proportion of small subunits that are associated with the binding protein in vivo is below levels of routine

detection by fluorography (Roy et al., 1982). The groES protein of E. coli is known to interact with

the groEL protein in the presence of ATP, but not in the absence of ATP (Chandrasekar et al., 1986; Fayet et al., 1989). Although the data of Goloubinoff et al. (1989) suggest that a groES-like protein should be present in chloroplasts, there is no evidence that such a protein is actually present in plants. The groES-like protein would be expected to turn up as an ATP-dependent factor required in higher plant Rubisco assembly (see legend to Figure 1).

What biochemical properties are shared by the many proteins that interact with chaperonins? It seems unlikely that all these proteins share a common amino acid sequence, although this remains to be shown. They could have sets of related sequences or sets of amino acids with similar chemical properties. On the other hand, incompletely folded polypeptide chains might have some structural regularity despite their partially denatured condition. For example, they might possess elements of secondary structure but lack tertiary structure. Perhaps all forms of chaperonins can recognize such sets of related arnino acids or other structural regularities; it seems more likely that variations in sequence or aggregation state may tailor the chaperonins to the systems in which they function. This might explain, for example, why higher plant Rubisco cannot assemble in E. coli, whereas cyanobacterial Rub- isco can (Gurevitz et al., 1985; Bradley et al., 1986; Van der Vies, Bradley, and Gatenby, 1986; Ellis et al., 1989).

Recent experiments with other molecular chaperones (the bovine microsomal binding protein and the cytosolic uncoating ATPase for clathrin-coated vesicles) have shown that several different short peptides can interact stably with these proteins (Flynn, Chappell, and Rothman, 1989). The interactions are stable to 1 M NaCI, but can be disrupted by low concentrations of ATP. Peptide-dependent ATP hydrolysis catalyzed by these chaperones was measured to calculate a K, for each peptide-chaperone complex. The K,,, values range over 3 orders of magnitude, but generally occur between 12 pM and 990 pM. Binding constants in this range would be consistent with the re- quirements for transient binding of newly made proteins in both the endoplasmic reticulum and cytosol. Based on the structural diversity of the peptides that can bind, Flynn et al. (1 989) propose that steric accessibility is the main factor governing the ability to associate with a molecular chaperone. Apparently, if a peptide is folded into a tertiary structure, the chaperone will not associate with it (Flynn et al., 1989). This suggests that the peptide binding site of the chaperone has to be able effectively to “surround” a peptide in order to bind stably. It remains to be seen whether these speculations are correct; however, it is interesting in this context that antibody to Rubisco large subunits cannot precipitate large subunits that are associated with the high molecular weight binding protein complex (Barraclough and Ellis, 1980).

Dow



ber 2021

1 O40 The Plant Cell

CONCLUSION the chloroplast, thus circumventing the nucleocytoplasmic pathway of expression. Mutations of Rubisco in higher plants are rare (Somerville, 1984), but there are Rubisco- deficient mutants in primrose (Winter and Herrmann, 1988) and tobacco (Avni et al., 1989), so genetic studies of the binding protein in higher plants may be possible. In this case, however, nuclear-based transformation mechanisms would have to be used (White, 1989), and one would have to rely on the nucleocytoplasmic expression of the binding protein genes.

Rubisco has served for many years as an important model system for studying chloroplast biogenesis. Major contributions to the literature regarding chloroplast gene expression, uptake and processing of proteins, and regulation of photosynthesis began with studies on Rubisco. Because of both the apparent role of chaperonin in Rubisco assembly and the occurrence of other types of molecular chaperones in animal systems, it seems likely that many processes in plants will turn out to be directly dependem on the action of molecular chaperones. Severa1 of the molecular chaperones described so far are either heat shock proteins themselves, or are related to them by homology. Because of this, I suggest that some of the plant heat shock proteins or their cognates are also molecular chaperones.

In summary, the Rubisco subunit binding protein is closely related to the E. coli groEL protein and a mitochondrial protein, both of which have been shown to function as molecular chaperones. The Rubisco subunit binding protein has been implicated independently in the assembly of higher plant Rubisco. The Rubisco subunit binding protein in higher plants is a particularly favorable model for studying the biochemical and biophysical mechanisms of molecular chaperone action.

The principle of self-assembly of proteins already has been modified by the discovery of routine post-translational processing and by the discovery of extensive and complex protein transport mechanisms. To this list of modifications it now appears that we should add a third set of processes mediated by molecular chaperones (Ellis et al., 1989). The evidence for this is persuasive for the assembly of cyanobacterial Rubisco in E. coli (Goloubinoff et al., 1989), and for the assembly of severa1 other oligomeric proteins in yeast mitochondria (Cheng et al., 1989). However, the mechanisms by which the chaperonins assist the assembly processes in these in vivo systems are not understood.

Based on in vitro studies of Rubisco assembly in chloroplast extracts, we do understand in broad outline the interaction of large subunits with the binding protein during assembly. However, we still do not have definite experimental evidence that the binding protein is obligatory for higher plant Rubisco assembly. To obtain this evidence may require cloning and expression in E. coli of genes for both of the higher plant binding protein subunits and a higher plant version of the groES protein (if it exists), in addition to higher plant large and small subunit genes. The Success of a similar approach to the problem of cyanobac- teria1 Rubisco assembly in E. coli is encouraging in this context (Goloubinoff et al., 1989). An alternative strategy is to develop from chloroplast extracts a fully in vitro protein synthesis and assembly system that can be manipulated to delete the binding protein. This approach requires careful control of in vitro protein synthesis conditions and the development of techniques that permit protein synthesis to continue despite manipulation.

Are genetic approaches to test the role of the binding protein feasible in algae or higher plants? It is not clear whether the copy number of the binding protein genes is low enough to permit easy recovery of missense mutations. Deletion of the binding protein genes is likely to be lethal, if E. coli behavior is any guide. Even missense mutations might be pleiotropic, interfering not only with Rubisco assembly but also with the assembly of other macromolecular complexes in the chloroplast. However, assuming these potential problems can be overcome, there is some hope for a genetic approach: It is possible to obtain temperature-sensitive conditional mutants deficient in photosynthesis in Chlamydomonas. Some of these mutants are deficient in Rubisco (e.g., Spreitzer, 1988). If the binding protein exists in Chlamydomonas, new Rubisco- deficient mutants might emerge in which the binding protein is specifically affected. Because it is possible to trans- form chloroplast DNA with tungsten microprojectiles fired at the cells (Boynton et al., 1988), Chlamydomonas might then become an especially useful organism for studying the binding protein. This technique would allow the introduction and expression of altered binding protein genes in

ACKNOWLEDGMENT

I thank Joseph Mascarenhas for a critical reading of this manuscript.

Received June 7, 1989; revised September 18, 1989.

REFERENCES

Andersson, I., Knight, S., Schneider, G., Lindqvist, Y., Lund- qvist, T., Branden, C.4, and Lorimer, G.H. (1989). Crystal structure of the active site of ribulose-bisphosphate carboxylase. Nature 337, 229-234.

Avni, A., Edelman, M., Rachailovich, I., Aviv, D., and Fluhr, R.

Dow



ber 2021


(1989). A point mutation in the gene for the large subunit of ribulose bisphosphate carboxylase/oxygenase affects holoenzyme assembly in Nicotiana tabacum. EMBO J. 8,1915-1918.

Barraclough, R., and Ellis, R.J. (1980). Assembly of newly synthesized large subunits into ribulose bisphosphate carboxylase in isolated pea chloroplasts. Biochim. Biophys. Acta 608,

Blair, G.E., and Ellis, R.J. (1974). Protein synthesis in chloroplasts. I. Light driven synthesis of the large subunit of fraction I protein by isolated pea chloroplasts. Biochim. Biophys. Acta

Bloom, M., Milos, P., and Roy, H. (1983). Light dependent assembly of ribulose-l,5-bisphosphate carboxylase. Proc. Natl. Acad. Sci. USA 80,1013-1017.

Bochkareva, E.S., Lissin, N.M., and Girshovich, AS. (1988). Transient association of newly synthesized unfolded proteins with the heat-shock groEL protein. Nature 336, 254-257.

Boynton, J.E., Gillham, N.W., Harris, E.H., Hosler, J.P., John- son, A.M., Jones, AR., Randolph-Anderson, B.L., Robertson, D., Klein, T.M., Shark, K.B., and Sanford, J.C. (1988). Chlo- roplast transformation in Chlamydomonas with high velocity microprojectiles. Science 240, 1534-1 538.

Bradley, D., Van der Vies, S., and Gatenby, A.A. (1 986). Expres- sion of cyanobacterial and higher-plant ribulose-l,5-bisphosphate carboxylase genes in Escherichia coli. Philos. Trans. R. SOC. Lond. B 313,447-458.

Cannon, S., Wang, P., and Roy, H. (1986). lnhibition of ribulose bisphosphate carboxylase assembly by antibody to a binding protein. J. Cell Biol. 103, 1327-1335.

Chandrasekar, G.N., Tilly, K., Woolford, C., Hendrix, R., and Georgopoulos, C. (1 986). Purification and properties of the groES morphogenetic protein of Escherichia coli. J. Biol. Chem.

Chapman, M.S., Suh, S.W., Cascio, D., Smith, W.W., and Ei- senberg, D.S. (1 987). Sliding-layer conformational change lim- ited by the quaternary structure of plant Rubisco. Nature 329,

Chapman, M.S., Suh, S.W., Curmi, P.M.G., Cascio, D., Smith, W.W., and Eisenberg, D.S. (1988). Tertiary structure of plant Rubisco: Domains and their contacts. Science 241, 71-74.

Chaudhari, P., and Roy, H. (1989). The delayed osmotic effect on in vitro assembly of Rubisco: Relationship to large subunit- binding protein complex dissociation. Plant Physiol. 89, 1366- 1371.

Cheng, M.Y., Hartl, F.-U., Martin, J., Pollock, R.A., Kalousek, F., Neupert, W., Hallberg, E.M., Hallberg, R.L., and Horwich, A.L. (1 989). Mitochondrial heat-shock protein hsp6O is essential for assembly of proteins into yeast mitochondria. Nature 337,

Chua, N.-H., and Schmidt, G.W. (1 978). Post-translational import into intact chloroplasts of a precursor to the small subunit of ribulose-1 ,Bbisphosphate carboxylase. Proc. Natl. Acad. Sci.

Ellis, R.J. (1979). The most abundant protein on earth. Trends

Ellis, R.J., and Van der Vies, S.M. (1988). The Rubisco subunit

Ellis, R.J., Van der Vies, S.M., and Hemmingsen, S.M. (1989).

19-31.

319,223-224.

261,1241 4-1 241 9.

354-356.

620-625.

USA 75, 6110-6114.

Biochem. Sci. 4, 241-244.

binding protein. Photosynth. Res. 16, 101-1 15.

The molecular chaperone concept. Biochem. SOC. Symp. 55,

Fayet, O., Ziegelhoffer, T., and Georgopoulos, C. (1989). The groES and groEL heat shock gene products of Escherichia coli are essential for bacterial growth at all temperatures. J. Bacte- riol. 171, 1379-1385.

Flynn, G.C., Chappell, T.G., and Rothman, J.E. (1989). Peptide binding and release by proteins implicated as catalysts of protein assembly. Science 245, 385-390.

Gatenby, A.A. (1 988). Synthesis and assembly of bacterial and higher plant Rubisco subunits in Escherichia coli. Photosynth. Res. 17, 145-157.

Gatenby, A.A., Lubben, T.H., Ahlquist, P., and Keegstra, K. (1 988). lmported large subJnits of ribulose bisphosphate carboxylase/oxygenase, but not imported p-ATP synthase subunits are assembled into holoenzyme in isolated chloroplasts. EMBO J. 7, 1307-1314.

Goloubinoff, P., Gatenby, A.A., and Lorimer, G.H. (1989). GroE heat shock proteins promote assembly of foreign prokaryotic ribulose bisphosphate carboxylase oligomers in Escherichia coli. Nature 337, 44-47.

Gomes, S.L., Juliani, M.H., Maia, J.C.C., and Silva, A.M. (1986). Heat shock protein synthesis during development in Caulobac- ter crescentus. J. Bacteriol. 168,923-930.

Gooding, L.R., Roy, H., and Jagendorf, A.T. (1973). Immunolog- ical identification of nascent subunits of wheat ribulose bisphosphate carboxylase on ribosomes of both chloroplast and cytoplasmic origin. Arch. Biochem. Biophys. 159, 324-335.

Gurevitz, M., Somerville, C.R., and Mclntosh, L. (1985). Pathway of assembly of ribulose bisphosphate carboxylase/oxygenase from Anabaena 7120 expressed in Escherichia coli. Proc. Natl. Acad. Sci. USA 82,6546-6550.

Hemmingsen, S.M., and Ellis, R.J. (1986). Purification and properties of Rubisco LSU binding protein. Plant Physiol. 80, 269- 276.

Hemmingsen, S.M., Woolford, C., Van der Vies, S.M., Tilly, K., Dennis, D.T., Georgopoulos, C.P., Hendrix, R.W., and Ellis, R.J. (1 988). Homologous plant and bacterial proteins chaperone oligomeric protein assembly. Nature 333, 330-334.

Hendrix, R.W. (1979). Purification and properties of groE, a host protein involved in bacteriophage assembly. J. MOI. Biol. 129,

Hohn, T., Hohn, B., Engel, A., Wurtz, M., and Smith, P.R. (1979). lsolation and characterization of the host protein groE involved in bacteriophage lambda assembly. J. MOI. Biol. 129, 359-373.

Lema, M.W., Brown, A., Butler, C.A., and Hoffman, P.S. (1988). Heat shock response in Legionella pneumophila. Can. J. Micro- biol. 34, 1 148-1 153.

Lubben, T.H., Theg, S.M., and Keegstra, K. (1988). Transport of proteins into chloroplasts. Photosynth. Res. 17, 173-194.

Macherel, D., Kobayashi, H., Valle, E., and Akazawa, T. (1986). Expression of amyloplast DNA in suspension-cultured cells of sycamore (Acer pseudoplatanus L.). FEBS Lett. 201,315-320.

Mclntosh, L., Poulsen, C., and Bogorad, L. (1980). Chloroplast gene sequence for the large subunit of ribulose bisphosphate carboxylase of maize. Nature 288, 556-560.

McMullin, T.W., and Hallberg, R.L. (1 988). A highly evolutionarily

145-1 53.

375-392.

Dow



ber 2021

1042 The Plant Cell

conserved mitochondrial protein is structurally related to the protein encoded by the fscherichia coli groEL gene. MOI. Cell Biol. 8, 371 -380.

Milos, P., and Roy, H. (1 984). ATP-released large subunits participate in the assembly of ribulose bisphosphate carboxylase. J. Cell. Biochem. 24, 153-1 62.

Milos, P., Bloom, M., and Roy, H. (1985). Methods for studying the assembly of ribulose bisphosphate carboxylase. Plant MOI. Biol. Rep. 3, 33-42.

Mulligan, R.M., Houtz, R.L., and Tolbert, N.E. (1988). Reaction- intermediate analogue binding by ribulose bisphosphate carboxylase/oxygenase causes specific changes in proteolytic sensitivity: The amino terminal residue of the large subunit is acetylated proline. Proc. Natl. Acad. Sci. USA 85, 151 3-1 51 7.

Musgrove, J.E., Johnson, R.A., and Ellis, R.J. (1987). Dissocia- tion of the ribulosebisphosphate-carboxylase large-subunit binding protein into dissimilar subunits. Eur. J. Biochem. 163,

Newman, S., and Cattolico, R.A. (1988). Synthesis of active Olisthodiscus luteus ribulose-bisphosphate carboxylase in Escherichia coli. Plant MOI. Biol. 11, 821-831.

Pelham, H.R.B. (1 986). Speculations on the functions of the major heat shock and glucose-regulated proteins. Cell 46, 959-961.

Pushkin, A.V., Tsuprun, V.L., Solovjeva, N.A., Shubin, V.V., Estigneeva, Z.G., and Kretovich, W.L. (1 982). High molecular weight pea leaf protein similar to the groE protein of fscherichia coli. Biochim. Biophys. Acta 704, 379-384.

Reading, D.S., Hallberg, R.L., and Myers, A.M. (1989). Charac- terization of the yeast hsp60 gene coding for a mitochondrial assembly factor. Nature 337, 655-658.

Roy, H., and Cannon, S. (1988). Ribulose bisphosphate carboxylase assembly: What is the role of the large subunit binding protein? Trends Biochem. Sci. 13, 163-164.

Roy, H., Cannon, S., and Gilson, M. (1988a). Assembly of Rub- isco from native subunits. Biochim. Biophys. Acta 957, 323- 334.

Roy, H., Chaudhari, P., and Cannon, S. (1988b). lncorporation of large subunits into ribulose bisphosphate carboxylase in chloroplast extracts: lnfluence of added small subunits and of conditions during synthesis. Plarit Physiol. 86, 44-49.

Roy, H., Hubbs, A., and Cannon, S. (1988~). Stability and dis-

529-534.

sociation of the large subunit Rubisco binding protein in vitro and in organello. Plant Physiol. 86, 50-53.

Roy, H., Bloom, M., Milos, P., and Monroe, M. (1982). Studies on the assembly of large subunits of ribulose bisphosphate carboxylase in isolated pea chloroplasts. J. Cell Biol. 94,

Shinnick, T.M., Vodkin, M.H., and Williams, J.C. (1988). The Mycobacterium tuberculosis 65 kilodalton antigen is a heat shock protein which corresponds to common antigen and to the Escherichia coli groEL protein. Infect. Immun. 56,446-451.

Somerville, C.R. (1 984). The analysis of photosynthetic carbon dioxide fixation and photorespiration by mutant selection. Oxf. Surv. Plant MOI. Cell Biol. 4, 103-131.

Spreitzer, R.J. (1 988). Temperature-sensitive, photosynthesis- deficient mutants of Chlamydomonas reinhardtii. Plant Physiol.

Tobin, E.M., and Silverthorne, J. (1985). Light regulation of gene expression in higher plants. Annu. Rev. Plant Physiol. 36, 569- 593.

Torres-Ruiz, J.A., and McFadden, B.A. (1988). A homolog of ribulose bisphosphate carboxylase-oxygenase binding protein in Chromatium vinosum. Arch. Biochem. Biophys. 261, 196- 204.

Van der Vies, S.M., Bradley, D., and Gatenby, A.A. (1986). Assembly of cyanobacterial and higher plant ribulose bisphosphate carboxylase subunits into functional homologous and heterologous enzyme molecules in fscherichia coli. EMBO J.

Voordouw, G., Van der Vies, S.M., and Boumeister, P.P. (1984). Disassociation of ribulose-l,5-bisphosphate carboxylase/oxygenase from spinach by urea. Eur. J. Biochem. 141, 31 3-31 8.

Wada, M., and Itikawa, H. (1986). Participation of fscherichia coli K-12 groE gene products in the synthesis of cellular DNA and RNA. J. Bacteriol. 157, 694-696.

White, F.F. (1 989). Vectors for gene transfer in higher plants. In Plant Biotechnology, S.D. Kung and C.J. Arntzen, eds (Boston: Butterworths) pp. 3-34.

Winter, P., and Herrmann, R.G. (1988). A 5-base-pair deletion in the gene for the large subunit causes the lesion in the ribulose bisphosphate carboxylase/oxygenase-deficient plastome mutant sigma of Oenothera hookeri. Bot. Acta 101, 68-75.

20-27.

86,773-777.

5,2439-2444.

Dow



ber 2021

Documents

Rubisco Assembly: A Model System for Studying the Mechanism of