Ribulose bisphosphate carboxylase assembly: what is the role of the large subunit binding protein?

163

TIBS13-May1988

Open Question Ribulose bisphosphate carboxylase assembly: what is the role of the large subunit binding protein? HarryRoy and Susan Cannon Ribulose 1,5-bisphosphate carboxylase/oxygenase (Rubisco), is a key enzyme in both photosynthetic carbon fixation and photorespiration. Genetic manipulation of this enzyme to increase its carboxylase activity could lead to substantial increases in net photosynthesis in several important crop plants. Unfortunately, the higher plant enzyme cannot be expressed in active form in E. coli. Several factors may account for this, the best studied of which is a binding protein that interacts with large subunits after they are synthesized but before they are assembled into holoenzyme. Various functions for this binding protein are considered.

polypeptide then crosses the chloroplast envelope and enters the stroma, losing its amino terminal transit sequence in the process 19-e~. Large subunits are cleaved and acetylated in spinach, tobacco, wheat, and muskmelon (Houtz et al., pers. commun.), it is not clear whether this occurs before or after assembly of holoenzyme. J:inally, there is strong evidence that newly synthesized large subunits associate with another cytoplasmically synthesized chloroplast protein, known as the large subunit binding protein, before they assemble into Rubisco 8"14'22-24. The binding protein is not part of the assembled holoenzyme, and it has been proposed that the binding protein may play an essential role in the assembly of Rubisco s.

The Rubisco large subunit binding for biotechnological intervention. One protein The binding protein appears to exist objective is to clone the structural genes for Rubisco in expression vectors in an equilibrium between a 720 kDa and obtain active enzyme in a conven- form (containing 12 ---60 kDa binding ient system such as E. coli. This would protein subunits) and 60 kDa to 117 permit analysis of the effects of genetic kDa forms (Fig. 1). Large subunits manipulation on the structure and associate primarily with the 720 kDa form, but they also occur as complexes function of Rubisco. The genes for the catalytic subunit of of ---117 kDa with an average sedithe enzyme, the so-called 'large' sub- mentation coefficient of 7S, based on unit, are located in chloroplast DNA 3. their behavior in sucrose gradients22. Large subunit genes from many species ATP causes net dissociation of the 720 of plants, algae and autotrophic pro- kDa form of this binding protein in karyotes have now been sequenced 4. dilute chloroplast extracts 1°-~4'22'24, Many of the genes for the less well but at high binding-protein concentraundersIIood 'small' subunits of the tions in vitro this complex is stable in enzyme, which are encoded in the nu- the presence of ATP 25'26. This is clear DNA, have also been cloned and believed to be due to the concentration dependence of aggregation, which sequenced 5. Despite these advantages, the goal of should be very pronounced with the obtaining active higher plant Rubisco doder.americ form of the binding proin E. coli has not been achieved: evi- tein. Despite the insensitivity of t.V~e dence from a variety of sources indi- dodecameric complex to ATP at high cates that the large subunits of higher protein concentration, it is believed plant Rubisco do not self-assemble that large subunits can still enter and with small subunits6'7, except in hom- leave this complex in chloroplasts in a ologous systems containing chloroplast light-dependent manner: large subcomponents 8-~4. This is curious, since units associated with the high molcyanobacterial large subunits do self- ecular weight form of the binding proassemble in E. coil ~5-~7, and yet differ tein can be mobilized or released from from the large subunits of higher plants it upon illumination of intact isolated chloroplastsg.25.26. Large subunits by only a few amino acids TM. The assembly of Rubisco in higher derived in vitro from the 720 kDa form plants is a complex process. The holo- of the binding protein by incubation of enzyme comprises eight large and eight chloroplast extracts with ATP at 4°C are indistinguishable from the native H. Roy and S. Cannonare at the Plant Science small subunits. The nuclear-coded 117 kDa complexes described above, Program, BiologyDepartment, RensselaerPoly- small subunlt is synthesized in precur- and they can be incorporated into sor form by cytosolic ribosomes. The technicInstitute, Troy,NY12180-3590, USA.

Ribulose 1,5-bisphosphate carboxylase/oxygenase (EC 4.1.1.39) (Rubisco), catalyses the CO2-fixation step in the Calvin cycle of photosynthesis, leading to the formation of two molecules of 3-phosphoglycerate for each ribulose bisphosphate molecule that reacts with CO2. The enzyme also catalyses a reaction of the same substrate with 02, leading to the production of a molecule of phosphoglycerate and a molecule of 2-phosphoglycolate. In higher plants, the latter compound is ultimately converted to glycine, which can undergo decarboxylation, leading to an actual loss of CO2 from the organism. The oxygenation of ribulose bisphosphate (photorespiration) is apparently wasteful, and organisms adapt to this reaction through anaerobiosis, CO2-concentrating systems, and evolutionary alterations of the structure of Rubisco which favor the carboxylation reaction 1. Under normal conditions of illumination in air, in well watered and fertilized soil, plant productivity is limited by CO2 fixation2: and it is believed that the carboxylase activity of Rubisco is the limiting factor. If the oxygenase reaction could be suppressed further, photosynthetic rates might increase 20-30% in many crop plants. Rubisco has therefore become a major target

~) 1988,ElsevierPublicationsCambridge 0376-5067188/$02.00

TIBS 13- May 1988

164

for the binding protein. Only if this were the case would the binding protein be in position to play a role in Rubisco assembly. Presumably the primitive interactions of the binding protein with large subunits, if any, Fig. 1. Model for the assembly of ribulose bisphosphate carboxylase. In this model, large subunits (L), synthewould not have been mandatory for sized in the chloroplast, interact with a binding protein Rubisco assembly, but would have had (BP). The molecular weights o f the complexes containing an adaptive advantage. Only later, as large subunits are indicated in kDa. A T P mediates disthe proteins co-evolved, could any sociation of the 720 kDa binding-protein complex, releasdependence of Rubisco assembly on ing large subunits. These sediment more rapidly than the binding protein grow, step by step, expected for large subunit monomers, and we estimate their mean molecular weight at 117 kDa, based on sedito its current level. mentation and gel filtration studies. The large subunits in Storage: The binding protein might these complexes may be monomers bound to one - 6 0 serve to store large subunits. This could k Da binding-protein subunit or they may be homodimers be useful to compensate for any irreguwhich subsequently interact with - 6 0 kDa bindinglarities in small subunit supplies, or protein subunits during assembly into Rubisco 13. These excessive large subunit synthesis, or complexes are represented as L *. Small subunits (S), derived from the cytoplasm, are also incorporated into Rubisco in the chloroplast. Since S can assemble into Rubisco under conditions where the > 720 kDa other reasons. This is consistent with complex is almost completely dissociated, the assembly steps are depicted as shown. There is no evidence two otherwise puzzling observations: for direct transfer o f L from the >720 kDa complex into Rubisco. (1) the high molecular weight binding protein complex releases assemblycompetent large subunits when incuRubisco upon subsequent incubation at ciple, this question may be settled in a bated in vitro with A T e 1-14'22'25'26 and room temperature 1°-~4. Large subunits number of ways, using in vitro systems, (2) the high molecular weight binding in the --117 kDa particles can also be or by cloning and expressing binding protein complex releases large subunits incorporated into Rubisco22; this incor- protein cDNAs in E. coli along with the in isolated intact chloroplasts !n the poration is stimulated by small sub- large and small subunit genes. Work on light but not in the dark 9'25. units26. Thus, the incorporation of this problem is progressing in several Modification: The binding prc;tein large subunits into Rubisco in vitro rep- laboratories, including ours. might covalently modify large subunits. resents assembly rather than mere subSolubility factor: The binding protein While this is theoretically possible, we unit exchange. The in vitro post- could function as an agent for keeping and others have found no evidence, of translational assembly of large subunits large subunits soluble. For example, phosphorylation or adenylylation of from the 720 kDa and the ---117 kDa isolated large subunits fail to renature the large subuni,~ or of the binding complexes into Rubisco can be inhi- with small subunits when resuspended protein when the 0inding-protein combited by antibody directed against the in aqueous buffers 7. Large subunits plex was exposed to radiolabeled binding protein 22. This indicates that synthesized from cloned genes in E. ATpI3, 24. the binding protein is associated with coli form precipitates or large aggreTransport: The binding protein large subunits, at least transiently, gates 6'17. However, in chloroplast might mediate transport of large subprior to assembly. This is illustrated in extracts, newly synthesized large sub- units within the chloroplast stroma Fig. 1. The high molecular weight com- units in association with the binding from the site of bynthesis to tl':e site of plex gives rise to --117 kDa complexes protein remain in solution, and can assembly. There is little evidence on (L*). These are either dimers contain- assemble into Rubisco holoenzyme this point. Hattori and Margulies2s ing large subunits, which interact with molecules in vitro 8"14'22'26. reported that substantial portions of BP transiently during assembly, or they Molecular chaperones: Recently, large-subunit synthesizing polyriboare heterodimers containing one Ellis 23 has drawn an analogy between somes are associated with thylakoids in binding-protein subunit and one large the large subunit binding protein and spinach chloroplasts. Controls indicasubunit. It should be emphasized that several animal proteins, such as nucleo- ted that this was not .due to trivial there is no evidence of direct flow of plasmin or the immunoglobulin heavy contamination of the membranes with large subunits from the (--720 kDa) chain binding protein, which appear to stromal material. Since small subunits high molecular weight binding-protein play roles as molecular chaperones, arrive in the chloroplast at the outer complex into Rubisco. preventing improper associations (see envelope, it is possible to imagine some Ref. 27 for review). The binding pro- sort of traffic, but this hypothesis may Possible binding protein requirement tein may indeed function as a chap- be very difficult to test. A number of possible functions for erone in the assembly process. The Regulation of translation: The bindthe binding protein appear to exist. evolution of such molecular chap- ing protein could interact with large Assembly: The binding protein may erones poses interesting questions. For subunits during their synthesis, either be obligatory for Rubisco assembly in example, if an amino acid substitution enhancing or slowing the rate of transhigher plants s,23. Definite evidence for occurred which rendered large subunits lation. There is no evidence for or this idea is still lacking, even ~hough it is insoluble, it is unlikely that a simultan- against this hypothesis. clear that the binding protein is a real eous compensatory mutation would participant in the reactions of large sub- have occurred, leading to the adaptaNone of the above hypotheses are units before the assembly process. The tion of the binding protein as a chap- inconsistent with any published data, binding protein may facilitate or inhibit erone. Thus, it appears necessary to and they are not mutually exclusive. assembly, or both, or neither. In prin- postulate some pre-existing function Whether or not the binding protein is

(60 kDa) (53 kDa) BP + L -.

( > 720 kDa) = L1BP12

L* (117 kDa)

S BP L8S8

TIBSI3-May1988

7 Voordouw, G., van der Vies, S. M. and Boumeister, P. P. (I984) Fur. J. Biochem. 141,313-318 8 Barraclough, R. aad Ellis, R. J. (1980) Biochin,. Biophys. Ao'a 608,19-31 9 Roy, H., Bloom, M., Milos,P. and Monroe, M. (1982)J. CellBiol. 94, 20-27 10 Bloom, M., Milos, P. and Roy, H. (1983) References Proc. Natl Acad. Sci. USA 80,1013-1017 10gren, w. L. (1984) Annu. Rev. Plant. 11 Milos,P. and Roy, H. (1984)J. CellBiochen.. Physiol. 35,415--442 24,153-162 2 Hardy, R. W. F., Havelka, U. D. and Quebedeaux, B. (1978) in Photosynthetic 12 Mflos, P., Bloom, M. and Roy, H. (1985) Plant Mol. Biol. Rep. 3, 33-42 Carbon Assimilation, (Siegelman,H. W. and 13 Milos, P. (1985) PhD Thesis. Rensselaer Hind, G., eds), pp. 165-178,PlenumPress Polyt¢chnicInstitute, Troy, NewYork 3 Mclntosh, L., Poulsen, C. and Bogorad, L. 14 Milos, P. and Roy, H. (1985) in Molecular (1980)N ~ r e 288,556--560 Biolt~gy of the Photosynthetic Apparatus, 4 Miziorko, H. and Lorimer, G. H. (1983) (Steinback, K. E., Arntzen, C., Bogorad, L. Annu. Re¢. Biochen.. 52,507-535 and Bonitz, S., eds), pp. 349-354, Cold 5 Cashmore, A. R. (1983)in Genetic EngineerSpringHarbor Press ing of Plants, (Kosuge, T., Meredith, C. P. and Hollaender,A., eds), pp. 29-38, Plenum 15 Gurevitz, M., Somerville,C. R. and Mclntosh, L. (1985)Proc. Natl Acad. Sci. USA 82, Press 6546--6550 6 Bradley, D., van der Vies, S. and Gatenby, A. A. (1986) Philos. Trans. R. Soc. London 16 Tabita, F. R. and Small, C. L. (1985) Proc. Natl Acad. Sci. USA 82, 6100-6103 Set. B 313,447--458

required for assembly of Rubisco, its functions should be studied, for they appear likely to tell us something about the regulation of the biogenesis of this key photosynthetic enzyme.

165 17 Van der Vies, S. M., Bradley, D. and Gatenby, A. A. (1986) EMBO J. 5, 24392444 18 Curtis, S. E. and Haselkorn, R. (1983) Proc. NatlAcad. Sci. USA 80,1835-1839 19 Cline, K., Werner-Washburne,M., Lubben, T. H. and Keegstra, K. (1985)J. Biol. Chem. 26O,3691-3696 20 Robinson,C. and Ellis, R. J. (1984) Fur. J. Biochen.. 142,337-342 21 Mishkind,M. L., Wessler,S. R. and Schmidt, G. W. (1985)J. CellBiol. 100,226--234 22 Cannon, S., Wang, P. and Roy, H. (1986) J. Cell Biol. 103,1327-1335 23 Ellis,R. J. (1987)Nature 328,378-379 24 Hemmingsen,S. M. and Ellis, R. J. (1986) Plant Physiol. 80,269-276 25 Roy, H., Hubbs, A. and Cannon, S. (1988) Plant Physiol. 86, 50-53 26 Roy, H., Chaudhari, P. and Cannon, S. (1988) Plant PhysioL 86, 44--49 27 Pelham,H. R. B. (1986)Ce1146,959-961 28 Hattori, T. and Margulies, M. M. (1986) Arch. Biochem. Biophys. 244,630--640

Talking Point How finicky is mitochondrial protein import? Nikolaus Pfanner, RupertP~dler and Walter Neupert Recently it was reported that artificial targeting signals or signals specific f o r organelles other than mitochondria could direct proteins into mitochondria. Here we discuss findings which suggest that specific steps o f mitochondrial protein import can be bypassed. Non-specific targeting signals appear to use this bypass pathway. Such import occurs at very low rates under physiological conditions and therefore does not affect the uniqueness o f mitochondrial protein composition.

cluding: (1) binding to receptor proteins on the mitochondrial surface (which may require a cytosolic cofactor); (2) subsequent insertion into proteinaceous sites in the outer membrane; and (3) transport into or across the inner mitochondrial membrane via contact sites between both membranes (requiring the electrical potential, A¥, across the inner membrane). These findings suggested that the recognition of a targeting signal (usually contained in the presequence) by a receptor protein served as the control step for selective import of mitochondrial proteins. Recently, however, it has been reported that sequences that are not speIV. Pfanner, R. Pfaller and W. Neupert are at the cific for mitochondrial protein import institut fiir Physiologische Chemie der Universitiit could also target "passenger' proteins M~nchen, Goethestrasse 33, D-8000 Miinchen 2, to mitochondria. These included FRG. N. Pfanneris presently at the Department of Molecular Biology, ,~rinceton University, a chloroplast 'transit' sequence 6, sequences selected out of the E. coli gePrinceton, NJ 08544, USA.

Several features of the targeting of nuclear encoded proteins to mitochondria and the translocation of proteins across the mitochondrial membranes have been unravelled during recent years. Many precursor proteins carry aminoterminal peptide extensions (presequences) which contain mitochondrial targeting information. This was shown by fusing such presequences to non-mitochondrial 'passenger' proteins and transporting these chimaeric proteins into mitochondria in vivo and in vitro ~. Recent studies have resolved the translocation across the membranes into several distinct steps 2-5, in-

nome 7, a region of the cytosolic protein dihydrofolate reductase s, a mitochondrial gene product 9, and artificial presequences ~°. Furthermore, changes of individual amino acid residues at the amino terminus of the mature protein part allowed import of a precursor protein from which the presequence had been removed ~. These results suggest that mitochondrial protein import is much less selective than previously thought. It is a generally accepted view, however, that all subcellular compartments contain a unique set of proteins. The question thus arises as to how mitochondria are able to maintain the specificity of their protein composition. It migl~t be proposed that misrouted prote;as are degraded when they arrive in the wrong compartment. In such a case the degradation system would have to possess a high specificity to distinguish between misrouted proteins and correctly targeted and imported ones. Furthermore, it would be rather uneconomical for the cell to control intracellular sorting predominantly by selective degradation. How then can this conflict be resolved7 The following observation may provide the answer. Import of proteins targeted by nonmitochondrial targeting signals seems to occur at low rates when compared to the import of physiological precursor proteins. In the studies mentioned above, it was already shown that the import rates of proteins carrying nonphysiological targeting sequences were ~) 1988.Elsevier PublicationsCambridge 0376-5[167/88/$02.(~1