Portal de Periódicos da CAPES

Linked Rubisco Subunits Can Assemble into Functional Oligomers without Impeding Catalytic Performance

2006; Elsevier BV; Volume: 282; Issue: 6 Linguagem: Inglês

10.1074/jbc.m610479200

1083-351X

Spencer M. Whitney, Robert E. Sharwood,

Algal biology and biofuel production

Although transgenic manipulation in higher plants of the catalytic large subunit (L) of the photosynthetic CO2-fixing enzyme ribulose 1,5-bisphospahte carboxylase/oxygenase (Rubisco) is now possible, the manipulation of its cognate small subunit (S) is frustrated by the nuclear location of its multiple gene copies. To examine whether L and S can be engineered simultaneously by fusing them together, the subunits from Synechococcus PCC6301 Rubisco were tethered together by different linker sequences, producing variant fusion peptides. In Escherichia coli the variant PCC6301 LS fusions assembled into catalytically functional octameric ([LS]8) and hexadecameric ([[LS]8]2) quaternary structures that excluded the integration of co-expressed unfused S. Assembly of the LS fusions into Rubisco complexes was impaired 50–90% relative to the assembly of unlinked L and S into L8S8 enzyme. Assembly in E. coli was not emulated using tobacco SL fusions that accumulated entirely as insoluble protein. Catalytic measurements showed the CO2/O2 specificity, carboxylation rate, and Michaelis constants for CO2 and ribulose 1,5-bisphosphate for the cyanobacterial Rubisco complexes comprising fusions where the S was linked to the N terminus of L closely matched those of the wild-type L8S8 enzyme. In contrast, the substrate affinities and carboxylation rate of the Rubisco complexes comprising fusions where L was fused to the N terminus of S or a six-histidine tag was appended to the C terminus of L were compromised. Overall this work provides a framework for implementing an alternative strategy for exploring simultaneous engineering of modified, or foreign, Rubisco L and S subunits in higher plant plastids. Although transgenic manipulation in higher plants of the catalytic large subunit (L) of the photosynthetic CO2-fixing enzyme ribulose 1,5-bisphospahte carboxylase/oxygenase (Rubisco) is now possible, the manipulation of its cognate small subunit (S) is frustrated by the nuclear location of its multiple gene copies. To examine whether L and S can be engineered simultaneously by fusing them together, the subunits from Synechococcus PCC6301 Rubisco were tethered together by different linker sequences, producing variant fusion peptides. In Escherichia coli the variant PCC6301 LS fusions assembled into catalytically functional octameric ([LS]8) and hexadecameric ([[LS]8]2) quaternary structures that excluded the integration of co-expressed unfused S. Assembly of the LS fusions into Rubisco complexes was impaired 50–90% relative to the assembly of unlinked L and S into L8S8 enzyme. Assembly in E. coli was not emulated using tobacco SL fusions that accumulated entirely as insoluble protein. Catalytic measurements showed the CO2/O2 specificity, carboxylation rate, and Michaelis constants for CO2 and ribulose 1,5-bisphosphate for the cyanobacterial Rubisco complexes comprising fusions where the S was linked to the N terminus of L closely matched those of the wild-type L8S8 enzyme. In contrast, the substrate affinities and carboxylation rate of the Rubisco complexes comprising fusions where L was fused to the N terminus of S or a six-histidine tag was appended to the C terminus of L were compromised. Overall this work provides a framework for implementing an alternative strategy for exploring simultaneous engineering of modified, or foreign, Rubisco L and S subunits in higher plant plastids. The catalytic efficiency of the photosynthetic CO2-fixing enzyme d-ribulose-1,5-bisphosphate carboxylase/oxygenase (Rubisco, EC 4.1.1.39) has a central role in determining how efficiently plants use their resources of water, fertilizer nutrient, and light (1Andrews T.J. Whitney S.M. Arch. Biochem. Biophys. 2003; 414: 159-169Crossref PubMed Scopus (107) Google Scholar, 2Zhu X.-G. Portis A.R. Long S.P. Plant Cell Environ. 2004; 27: 155-165Crossref Scopus (176) Google Scholar, 3Long S.P. Zhu X. Naidu S.L. Ort D.R. Plant Cell Environ. 2006; 29: 315-330Crossref PubMed Scopus (1106) Google Scholar). As the primary port of entry of inorganic carbon into the biosphere, it is surprising that plant Rubisco is not very efficient, particularly at limiting CO2 concentrations where the rate of catalytic turnover is less than one-thousandth that of many other plant enzymes (4Morell M.K. Paul K. Kane H.J. Andrews T.J. Aust. J. Bot. 1992; 40: 431-441Crossref Scopus (41) Google Scholar). The tendency of Rubiscos to confuse CO2 with the more abundant atmospheric gas, O2, encumbers photosynthesis in higher plants with both a requirement to invest large amounts of protein in Rubisco and also a requirement for an energy-intensive photorespiratory metabolism to recycle the oxygenated waste product. However, higher plant Rubisco is not the pinnacle of evolution as more efficient and specific forms are found in nature, particularly in a number of red algae (5Whitney S.M. Baldett P. Hudson G.S. Andrews T.J. Plant J. 2001; 26: 535-547Crossref PubMed Google Scholar). The benefits of replacing the Rubisco in C3 crops with these natural variants have been modeled and show substantial improvements that come at no additional energy or resource cost (1Andrews T.J. Whitney S.M. Arch. Biochem. Biophys. 2003; 414: 159-169Crossref PubMed Scopus (107) Google Scholar, 2Zhu X.-G. Portis A.R. Long S.P. Plant Cell Environ. 2004; 27: 155-165Crossref Scopus (176) Google Scholar, 6Tcherkez G.G.B. Farquhar G.D. Andrews T.J. Proc. Natl. Acad. Sci. U. S. A. 2006; 103: 7246-7251Crossref PubMed Scopus (559) Google Scholar). Against this background, it is no surprise efforts are continuing into understanding Rubisco's catalytic mechanism, how it has evolved, and how improvements in its efficiency might be engineered by mutating it or replacing it in plants with more efficient homologs (7Spreitzer R.J. Salvucci M.E. Annu. Rev. Plant Biol. 2002; 53: 449-475Crossref PubMed Scopus (707) Google Scholar, 8Parry M.A. Andralojc P.J. Mitchell R.A. Madgwick P.J. Keys A.J. J. Exp. Bot. 2003; 54: 1321-1333Crossref PubMed Scopus (288) Google Scholar, 9Raines C.A. Plant Cell Environ. 2006; 29: 331-339Crossref PubMed Scopus (127) Google Scholar, 10Spreitzer R.J. Peddi S.R. Satagopan S. Proc. Natl. Acad. Sci. U. S. A. 2005; 102: 17225-17230Crossref PubMed Scopus (98) Google Scholar). Engineering Rubisco in higher plants is complicated by the separate locations of the genes coding for the large (L) and small (S) 2The abbreviations used are: S, Rubisco small subunit; L, Rubisco large subunit; carboxyarabinitol-P2,2′-carboxyarabinitol-1,5-bisphosphate; carboxypentitol-P2, isomeric mixture of carboxyarabinitol-P2 and 2′-carboxyribitol-1,5-bisphosphate; CLLS, cyanobacterial (Synechococcus PCC6301) Rubisco L-linker-S fusion peptide; EPPS, 4-([2-hydroxyethyl)-1-piperazinepropanesulfonic acid; RuBP, ribulose-P2; MES, 2-[N-morpholino]ethanesulfonic acid; ribulose-P2, d-ribulose-1,5-bisphosphate; Rubisco, ribulose-P2 carboxylase/oxygenase; TLLS, tobacco Rubisco L-linker-S fusion peptide; Bis-Tris, 2-[bis(2-hydroxyethyl) amino]-2-(hydroxymethyl)propane-1,3-diol. subunits and the complex assembly mechanism that necessitates the coordinated expression, post-translational modifications, and assembly of both subunits into a hexadecamer (L8S8) within the chloroplast stroma (Fig. 1) (11Whitney S.M. Andrews T.J. Plant Cell. 2001; 13: 193-205Crossref PubMed Scopus (80) Google Scholar, 12Roy H. Andrews T.J. Leegood R.C. Sharkey T.D. von Caemmerer S. Photosynthesis: Physiology and Metabolism. Kluwer Academic Publishers, Dordrecht, The Netherlands2000: 53-83Google Scholar, 13Houtz R.L. Portis Jr., A.R. Arch. Biochem. Biophys. 2003; 414: 150-158Crossref PubMed Scopus (78) Google Scholar). The L subunit contains the catalytic site and the S subunits, whose precise role in the structure and function of Rubisco remains poorly understood and which are essential for catalytic viability (7Spreitzer R.J. Salvucci M.E. Annu. Rev. Plant Biol. 2002; 53: 449-475Crossref PubMed Scopus (707) Google Scholar, 8Parry M.A. Andralojc P.J. Mitchell R.A. Madgwick P.J. Keys A.J. J. Exp. Bot. 2003; 54: 1321-1333Crossref PubMed Scopus (288) Google Scholar, 14Spreitzer R.J. Arch. Biochem. Biophys. 2003; 414: 141-149Crossref PubMed Scopus (171) Google Scholar). Indeed, when stripped of S or assembled with foreign S, the catalytic properties of Rubiscos are substantially impaired. Genetically engineering Rubisco in plastids, therefore, needs to attend to both L and S. For the Rubisco L subunit gene (rbcL) located in the plastid genome (plastome), genetic manipulation by plastome transformation in tobacco is a routine but protracted process requiring typically 4–12 months to obtain mature homoplasmic transformants (15Whitney S.M. Andrews T.J. Plant Physiol. 2003; 133: 287-294Crossref PubMed Scopus (29) Google Scholar, 16Whitney S.M. von Caemmerer S. Hudson G.S. Andrews T.J. Plant Physiol. 1999; 121: 579-588Crossref PubMed Scopus (112) Google Scholar, 17Whitney S.M. Andrews T.J. Proc. Natl. Acad. Sci. U. S. A. 2001; 98: 14738-14743Crossref PubMed Scopus (110) Google Scholar, 18Kanevski I. Maliga P. Rhoades D.F. Gutteridge S. Plant Physiol. 1999; 119: 133-141Crossref PubMed Scopus (96) Google Scholar, 19Kanevski I. Maliga P. Proc. Natl. Acad. Sci. U. S. A. 1994; 91: 1969-1973Crossref PubMed Scopus (108) Google Scholar). In contrast, an appropriate means for engineering the native (or foreign) S subunit genes (RbcS) in higher plants has remained an elusive challenge due to the multiple RbcS copies in higher plant nuclei that essentially precludes them from targeted mutagenic or replacement strategies. Moreover, attempts to incorporate recombinant S into higher plants by transplastomic methods have highlighted how circumventing the assembly of cytosolic-synthesized S with plastid synthesized L into L8S8 complexes is almost immutable unless the endogenous levels of S have been substantially reduced by antisense (17Whitney S.M. Andrews T.J. Proc. Natl. Acad. Sci. U. S. A. 2001; 98: 14738-14743Crossref PubMed Scopus (110) Google Scholar, 20Dhingra A. Portis A.R. Daniell H. Proc. Natl. Acad. Sci. U. S. A. 2004; 101: 6315-6320Crossref PubMed Scopus (145) Google Scholar, 21Zhang X.H. Ewy R.G. Widholm J.M. Portis Jr., A.R. Plant Cell Physiol. 2002; 43: 1302-1313Crossref PubMed Scopus (27) Google Scholar). Resilient expression of the native S is also problematic when transplanting in foreign L subunits, as highlighted by the transplastomic replacement of the tobacco rbcL with sunflower rbcL that produced tobacco-sunflower transformants whose hybrid sunflower L8-tobacco S8 enzyme was kinetically impaired and unable to support autotrophic growth in air (18Kanevski I. Maliga P. Rhoades D.F. Gutteridge S. Plant Physiol. 1999; 119: 133-141Crossref PubMed Scopus (96) Google Scholar). Here a novel strategy for simultaneously engineering Rubisco S and L subunits was examined using the Synechococcus sp. PCC6301 enzyme that, unlike higher plant Rubiscos, is coded by a single rbcL-rbcS operon and can be functionally expressed in Escherichia coli (12Roy H. Andrews T.J. Leegood R.C. Sharkey T.D. von Caemmerer S. Photosynthesis: Physiology and Metabolism. Kluwer Academic Publishers, Dordrecht, The Netherlands2000: 53-83Google Scholar, 22Emlyn-Jones D. Woodger F.J. Price G.D. Whitney S.M. Plant Cell Physiol. 2006; 47: 1630-1640Crossref PubMed Scopus (55) Google Scholar, 23Goloubinoff P. Gatenby A.A. Lorimer G.H. Nature. 1989; 337: 44-47Crossref PubMed Scopus (528) Google Scholar, 24Kettleborough C.A. Parry M.A.J. Keys A.J. Phillips A.L. J. Exp. Bot. 1990; 41: 1287-1292Crossref Google Scholar, 25Gatenby A.A. Plant Mol. Biol. 1992; 19: 677-687Crossref PubMed Scopus (29) Google Scholar). Using different peptide linkers, the PCC6301 L and S were tethered together to produce an array of SL and LS fusion peptides. Presented here are results that show the subunit fusions can correctly fold and assemble into functional Rubisco oligomers in E. coli and that unlinked S is excluded from assembly. Moreover the kinetic properties of the Rubisco oligomers were evaluated and found that the catalytic prowess of most fusion-Rubiscos closely mimicked the wild-type PCC6301 L8S8 enzyme. Materials—Unlabeled and carboxyl-14C-labeled carboxypentitol-P2 was synthesized as described (26Pierce, J., Tolbert, N. E., and Barker, R. (1980) 19, 934-942Google Scholar). Ribulose-P2 was synthesized and purified as described (27Kane H.J. Wilkin J.M. Portis A.R. Andrews T.J. Plant Physiol. 1998; 117: 1059-1069Crossref PubMed Scopus (83) Google Scholar). All cloned DNA sequences were fully sequenced using BigDye terminator sequencing (Applied Biosystems) on an ABI 3730 sequencer (Biomolecular Resource Facility, The John Curtin School of Medical Research, Australian National University) following the manufacturer's protocol (Applied Biosystems). Amplification of Linker Genes—A gene coding for the linker-60 peptide was assembled by splice overlap extension using successive PCR reactions with Herculase Enhanced DNA polymerase (Stratagene). The initial PCR reaction contained 2 μm concentrations each of primers Linker1, Linker2, Linker3, and Linker4 that overlapped with adjacent primers by 20–23 nucleotides (supplemental Fig. 1). The PCR products were diluted 100-fold in a subsequent PCR reaction that amplified the 193-bp full-length linker-60 gene using primers 5′EcoRVlinker and 3′Nde60mer (Fig. 2A). A 208-bp gene coding for a tetrahistidine tagged linker-60 peptide (60H4) was assembled using primer Linker2His-4 instead of Linker2 in the initial PCR reaction (supplemental Fig. 1). Both linker-60 genes were cloned into pGEM-T Easy (Promega) to give plasmids plinker60TVE and plinker60H4TVE. Genes coding for shorter linker peptides of 20 and 40 amino acids were amplified from plinker60TVE using the primer pairs 5′EcoRVlinker/3′Nde20mer and 5′EcoRVfusion/3′Nde40mer and cloned into pGEM-T Easy to give plasmids plinker20TVE and plinker40TVE. Cloning Wild-type PCC6301 Rubisco Genes—The Synechococcus PCC6301 Rubisco rbcL-S operon was amplified from pSH1 (28Andrews T.J. J. Biol. Chem. 1988; 263: 12213-12219Abstract Full Text PDF PubMed Google Scholar) using primers 5′NdeCrbcL (5′-CATATGCCCAAGACGCAATCTGCCGCAG-3′, the NdeI site is underlined, and the rbcL initiator codon is in bold) and 3′SacISynS (5′-TGAGCTCTTAGTATCGGCCGGGACGATGAACGAT-3′, the SacI site is underlined, and the complement of rbcS terminator codon is in bold) and the 1851-bp NdeI-SacI product was cloned into pET30Xa/LIC (pET30, Novagen) to give plasmid pETCLS (Fig. 2B). Cloning PCC6301 S-linker-L (CSLL) Rubisco Fusion Genes— The Synechococcus PCC6301 rbcL and rbcS genes were amplified separately from pSH1 using the primer pairs 5′NdeCrbcL/3′SynrbcL (5′-TTAGAGCTTGTCCATCGTTTCAAATTCGAA-3′, the complement of rbcL terminator codon is in bold) and 5′NcoSynS (5′-CCATGGGCATGAAAACTCTGCCCAAAGAG-3′, the NcoI site is underlined, and rbcS initiator codon is in bold)/3′EcoRVSynS (5′-TGATATCGGCCGGGACGATGAACGATG-3′, the EcoRV site is underlined), respectively. The 1422- and 337-bp DNA fragments were cloned into pGEM-T Easy to give plasmids pCLTVE(NdeI) and pCSTVE(NcoI), respectively. The rbcL gene was excised from pCLTVE(NdeI) with NdeI and NotI, and the 1438-bp fragment was cloned into pET30 to give pETCL. The rbcS gene from pCSTVE(NcoI) was cloned in-frame with different length linker peptides by cloning the 332-bp NcoI-EcoRV fragment into plasmids plinker20TVE, plinker40TVE, plinker60TVE, and plinker60H4TVE to give plasmids pCS20TVE, pCS40TVE, pCS60TVE, and pCS60H4TVE, respectively. The 513-bp NcoI-NdeI rbcS60 gene from pCS60TVE was cloned into pET28a+ (Novagen) to give pETCS60, and then the 552-bp XbaI-NdeI fragment was cloned into pETCL, giving plasmid pETCS60L (Fig. 2B). The 393-bp and 453-bp NcoI-NdeI fragments from pCS20TVE and pCS40TVE were cloned into pETCS60L to give plasmids pETCS20L and pETCS40L, respectively (Fig. 2B). Plasmid pETCS60LS that contains a fused and a non-fused copy of rbcS was assembled by replacing the 1078-bp SphI fragment in pETCLS with the 1589-bp fragment from pETCS60L. The 513-bp NcoI-NdeI rbcS60 fragment was then replaced with the 528-bp NcoI-NdeI rbcS60H4 fragment from pCS60H4TVE to give plasmid pETCS60H4LS (Fig. 2B). Sequence coding for a C-terminal His6 tag was cloned 3′ to rbcL by amplifying the gene with primers 5′NdeCrbcL and 3′XhoSynL (5′-CTCGAGCTTGTCCATCGTTGCAAATTCGAAC-3′, the XhoI site is underlined) and cloning the 1237-bp KpnI-XhoI fragment into pETCS40L to give pETCS40LH6 (Fig. 2B). Cloning the PCC6301 CL40S Rubisco Gene—The rbcL gene from pSH1 was amplified using primers Nae5′SynL (5′-GCCGGCGTGAAGGACTACAAAC-3′, the NaeI site is underlined) and 3′EcoRVSynL (5′-GATATCCCTTGTCCATCGTTTCGAATTCGAACTTG-3′, the EcoRV site is underlined). The 1386-bp DNA product was cloned into pGEM-T Easy to give plasmid pCLEVTVE into which the 121-bp EcoRV-NdeI linker-40 gene from plinker40TVE was cloned to give plasmid pCL40TVE. The rbcS gene was amplified from pSH1 using primers 5′NdeSynS (5′-TCATATGAGCATGAAAACTCTGCCCAAAGA-3′, the NdeI site is underlined, and the rbcS initiator codon is in bold) and Syn3SacI, and the 347-bp product was cloned into pGEM-T Easy to give plasmid pCSTVE(NS). The 344-bp NdeI-SacI rbcS fragment was then cloned in-frame 3′ to the fused rbcL40 gene in pCL40TVE to give plasmid pCL40S from which the 1757-bp KpnI-SacI fragment was used to replace the corresponding 1665-bp fragment in pETCLS to give pETCL40S (Fig. 2B). Cloning the Tobacco Rubisco TSLL Fusions—Sequence coding 83 nucleotides of the tobacco ribosomal RNA (Prrn) promoter, 63 nucleotides of the T7g10 5′-untranslated region, and a codon modified rbcS gene (cmrbcS) was assembled by splice overlap extension using overlapping primers by successive PCR reactions as described above (supplemental Fig. 1). The translated product from cmrbcS matched that coded by a native Nicotiana tabacum (tobacco) Rubisco RbcS (11Whitney S.M. Andrews T.J. Plant Cell. 2001; 13: 193-205Crossref PubMed Scopus (80) Google Scholar). The initial PCR reaction used primers ModSun1, ModSun2, ModTob3, ModTob4r, ModTob5, ModTob6r, ModTob7, and ModTob8rH6 that overlapped by 15 nucleotides with adjoining primers (supplemental Fig. 1). The second PCR reaction amplified a 523-bp Prrn-T7g10 5′-untranslated region-cmrbcS sequence using primers 5′HindPrrn and 3′EcoRVmNtS and was cloned into pGEM-T Easy to give plasmid pPcmSTVE. To clone the linker-40 and -60 genes in-frame to the 3′ end of cmrbcS, the gene was amplified from pPcmSTVE using primers 5′NcoImNtS and 3′EcoRVmNtS, and the 368-bp NcoI-EcoRV cmrbcS gene cloned into plinker40TVE and plinker60TVE to give pTS40 and pTS60, respectively. The 489-bp cmrbcS40 and 549-bp cmrbcS60 NcoI-NdeI fragments were then cloned into pET28a(+) to give plasmids pETTS40 and pETTS60. A tobacco rbcL was cloned in-frame to the 3′ end of the cmrbcS-linker genes by amplifying a copy of rbcL from pLEV3 (15Whitney S.M. Andrews T.J. Plant Physiol. 2003; 133: 287-294Crossref PubMed Scopus (29) Google Scholar) using primers 5′NdeNtL (5′-TCATATGTCACCACAAACAGAGACTAA-3′, the NdeI site is underlined, and the rbcL initiator codon is in bold) and Trb-cLHind (5′-CAAGCTTTTACTTATCCAAAACGTCCAC-3′, the HindIII site is underlined, and the complement of rbcL terminator codon is in bold) and cloning the 1436-bp NdeI-Hin-dIII rbcL fragment into pETTS40 and pETTS60 to give plasmids pETTS40L and pETTS60L, respectively (Fig. 2B). Growth and Expression of Rubisco in E. coli—The Rubisco genes were cloned downstream of the T7 promoter in pET28a(+) and pET30 and transformed into BL21(DE2) cells (Promega). Rubisco expression was auto-induced by growing the cells at 22 °C in Luria-Bertani medium containing 30 μg·ml-1 kanamycin supplemented with additional carbon sources (0.5% (v/v) glycerol, 2.8 mm glucose, 5.6 mm α-lactose), nutrients (1 mm MgSO4, 25 mm (NH4)2SO4, 40 mm KH2PO4, 50 mm Na2HPO4) and trace metals (50 μm FeCl3, 20 μm CaCl2, 2 μm Na2SeO3, 2 μm H3BO3, 10 μm MnCl2, 10 μm ZnSO4, 2 μm CoCl2, 2 μm CuCl2, 2 μm NiCl2, 2 μm Na2NoO4) (29Studier F.W. Protein Expression Purif. 2005; 41: 207-234Crossref PubMed Scopus (4313) Google Scholar). The cultures (100 ml) were shaken at 200 oscillations per minute for 120 h with final cell densities measured by absorbance at 600 nm between 6.6 and 9.2. The cells were harvested by centrifugation (6000 × g, 10 min, 4 °C), frozen in liquid nitrogen, and stored at -70 °C. Protein Extraction—E. coli cells pellets were suspended in ice-cold extraction buffer (100 mm EPPS-NaOH. pH 8, 1 mm EDTA, 20 mm MgCl2, 2 mm dithiothreitol, 0.043% (w/v) protease inhibitor mixture for bacterial cells (Sigma)) and lysed by passage through a French pressure cell (140 megapascals). An aliquot of lysate sample was mixed with an equal volume of SDS buffer (125 mm Tris-Cl, pH 6.8, 4% (w/v) SDS, 20% (v/v) glycerol, 150 mm 2-mercaptoethanol, 0.01% (w/v) bromphenol blue), and the remainder was centrifuged at 38,000 × g for 15 min at 4 °C. Aliquots of the supernatant were either assayed for protein content using the dye binding Pierce Coomassie Plus kit, treated with a equal volume of SDS buffer (soluble protein sample for SDS-PAGE analysis), and diluted 2-fold with Native PAGE buffer (30% (v/v) glycerol, 0.001% (w/v) bromphenol blue) or incubated with 25 mm NaHCO3 at 25 °C for 30 min before measuring Rubisco content ([2-14C]carboxyarabinitol-P2 binding, see below) and carboxylase activity under substrate ribulose-P2-limiting and -saturating conditions (see below). PAGE and Immunoblot Analyses—Proteins were separated by SDS-PAGE and non-denaturing PAGE using 4–12% NuPAGE Bis-Tris and 4–12% Tris-glycine gels (Invitrogen), respectively. The Bis-Tris gels were buffered with MES and electrophoresed at 200 V according to the supplier's instructions. Tris-glycine gels were run at 4 °C in 60 mm Tris, 191 mm glycine buffer at 60 volts for 18 h. Protein bands were visualized using Gelcode Blue reagent (Pierce) or blotted onto nitrocellulose (Hybond C, APBiotech) using an Xcell transfer cell (Novex) according to the manufacturer's specifications. Immunoblot analyses were performed as described previously (5Whitney S.M. Baldett P. Hudson G.S. Andrews T.J. Plant J. 2001; 26: 535-547Crossref PubMed Google Scholar) using polyclonal antiserum raised in rabbits to pure spinach Rubisco, tobacco Rubisco, or Synechococcus PCC6301 Rubisco. Immunoreactive proteins were visualized using AttoPhos reagent (Promega) with a Vistra FluorImager. Rubisco Purification—Rubiscos without histidine tags were purified from E. coli cells by size exclusion chromatography or by ultracentrifugation through sucrose density gradients. Cells that had been stored at -70 °C were suspended in ice-cold extraction buffer, lysed by a French press, and centrifuged as described above. The supernatant was chromatographed through a Superdex 200HR 10/30 column equilibrated with column buffer (50 mm EPPS-NaOH, 100 mm NaCl, pH 8) using an ÁKTA explorer system (APBiotech). Alternatively, a saturated ammonium sulfate solution (pH 7) was slowly added to the supernatant to a final concentration of 20% (w/v) on ice, and the extract was centrifuged at 30,000 × g for 15 min at 2 °C. The supernatant was collected, and Rubisco was precipitated by adding ammonium sulfate to 50% (w/v) and pelleted by centrifugation. The precipitate was dissolved in a 0.7-ml gradient buffer (25 mm EPPS-NaOH, pH 8, 1 mm EDTA) and centrifuged at 31,000 rpm for 26 h at 4 °C through an exponential density gradient (mixing volume, 12.15 ml; gradient volume, 15 ml) of 7–28.8% (w/v) sucrose in gradient buffer using an SW31Ti rotor (Beckman). Fractions collected from the Superdex 200 chromatography (0.3 ml) and the sucrose gradients (1 ml) were assayed for substrate-saturated RuBP carboxylase activity (see below) and an aliquot mixed with an equal volume of SDS or Native PAGE buffer for PAGE analysis. The three fractions with the highest Rubisco activities for each sucrose gradient were pooled and dialyzed for 16 h at 4 °C against 2 liters of gradient buffer and then for 90 min against 0.5 liters of gradient buffer containing 20% (v/v) glycerol. The dialyzed samples were frozen in liquid nitrogen and stored at -70 °C. Rubiscos with histidine tags were purified using Ni2+-nitrilotriacetic acid-agarose (Qiagen). E. coli cells were suspended in ice-cold affinity extraction buffer (50 mm Tris-Cl, pH 8, 0.3 m NaCl, 10 mm imidazole, 0.043% (w/v) bacterial protease inhibitor mixture), lysed by a French press, and centrifuged (as above). The soluble protein was chromatographed through Ni2+-nitrilotriacetic acid-agarose and washed with 20 bed volumes of affinity extraction buffer. Bound protein was eluted in 1.5–2 ml of elution buffer (50 mm Tris-Cl, pH 8, 0.3 m NaCl, 200 mm imidazole) and immediately dialyzed with successive changes of gradient buffer and stored as described above. Kinetic Measurements—The Michaelis constant for ribulose-P2 (KmRuBP) and catalytic turnover rate (Vmaxc) were measured in cell free soluble E. coli protein extracts. After lysis and centrifugation (see above), the E. coli-soluble protein extract was preincubated with 25 mm NaHCO3 for 20 min at 25 °C to activate the Rubisco. Carboxylase activities were measured using NaH14CO3 assays (28Andrews T.J. J. Biol. Chem. 1988; 263: 12213-12219Abstract Full Text PDF PubMed Google Scholar, 30Paul K. Morell M.K. Andrews T.J. Biochemistry. 1991; 30: 10019-10026Crossref PubMed Scopus (28) Google Scholar) containing different amounts of substrate ribulose-P2 (0–2 mm). Assays were buffered with 100 mm EPPS-NaOH, pH 8, containing 20 mm MgCl2 and were performed in duplicate with unbuffered ribulose-P2 added to initiate catalysis. Catalytic turnover rate was calculated by dividing the substrate-saturated carboxylase activity by the concentration of Rubisco active sites measured by the stoichiometric binding of the tight binding inhibitor [2-14C]carboxyarabinitol-P2 as described previously (17Whitney S.M. Andrews T.J. Proc. Natl. Acad. Sci. U. S. A. 2001; 98: 14738-14743Crossref PubMed Scopus (110) Google Scholar, 31Ruuska S. Andrews T.J. Badger M.R. Hudson G.S. Laisk A. Price G.D. von Caemmerer S. Aust. J. Plant Physiol. 1998; 25: 859-870Google Scholar). After preincubation with 25 mm NaHCO3, duplicate aliquots of extract were incubated for up to 30 min with 13–39 μm [2-14C]carboxypentitol-P2 (an isomeric mixture of [2-14C]carboxyarabinitol-P2 and [2-14C]carboxyribitol-P2) at 25 °C, and the amount of Rubisco-bound [14C]carboxyarabinitol-P2 was recovered by gel filtration (17Whitney S.M. Andrews T.J. Proc. Natl. Acad. Sci. U. S. A. 2001; 98: 14738-14743Crossref PubMed Scopus (110) Google Scholar). Purified Rubisco preparations were used to measure the Michaelis constants for CO2 (Km(CO)2) and the CO2/O2 specificity (Sc/o at pH 8.3 (32Kane H.J. Viil J. Entsch B. Paul K. Morell M.K. Andrews T.J. Aust. J. Plant Physiol. 1994; 21: 449-461Crossref Scopus (103) Google Scholar)). Km(CO)2 was measured by 14CO2 fixation at 25 °C, pH 8, according to Andrews (28Andrews T.J. J. Biol. Chem. 1988; 263: 12213-12219Abstract Full Text PDF PubMed Google Scholar) in nitrogen-sparged septum-capped scintillation vials. The assays were initiated by adding purified enzyme (that had been preincubated for 20 min in buffer containing 20 mm MgCl2 and 25 mm NaHCO3) into N2-equilibrated assay buffer (100 mm EPPS-NaOH, 20 mm MgCl2, 0.6 mm ribulose-P2, 0.1 mg·ml-1 carbonic anhydrase) containing varying concentrations of NaH14CO3. The assays were stopped after 2 min with 0.5 volumes of 25% (v/v) formic acid and dried at 80 °C, and then the residue was dissolved in 0.5 ml of water before adding 2 volumes of scintillant (UltimaGold, Packard Bioscience) for scintillation counting. Engineering Linker Peptides of Different Lengths—Genes coding for linker peptides of different lengths for fusing the large (L) and small (S) subunits of Rubisco from Synechococcus PCC6301 (cyanobacteria) and N. tabacum (tobacco) were synthesized by splice overlap extension using PCR (supplemental Fig. 1). The linker sequences were designed based on flexible Ser-Gly-Gly motif repeats with a basic amino acid every 10th residue that had been previously used to assemble E. coli chaperonin complexes from GroES-GroEL fusions (33Farr G.W. Fenton W.A. Chaudhuri T.K. Clare D.K. Saibil H.R. Horwich A.L. EMBO J. 2003; 22: 3220-3230Crossref PubMed Scopus (62) Google Scholar). Examination of available x-ray structures for different L8S8 Rubisco hexadecamers indicated considerable variability in the spatial separation between the N and C termini of the adjoining L and S subunits (Fig. 1). Because there were numerous permutations for linking the termini of a subunit to its cognate partners, four different sized linkers were tested that comprised 20, 40, 60, and 65 (60H4) amino acids corresponding to flexible peptides of ∼ 65, 130, 185, and 200 Å. Restriction sites at the 3′ (EcoRV) and 5′ (NdeI) ends were incorporated into the coding sequence of each linker gene to facilitate the in-frame cloning of L and S genes (Fig. 2A). An internal His4 tag comprising the sequence His-Ser-His-His-His-His was engineered into the 60H4 fusion peptide to facilitate purification by affinity chromatography. Subunit Organization of Rubisco Fusions—Because the N and C termini of each L and S are positioned on the surface of L8S8 Rubiscos (Fig. 1), the viability of linking the subunits in either an L-S or S-L arrangement was examined. The genes coding for Synechococcus PCC6301 Rubisco L and S are coded by a rbcL-rbcS operon and can assemble into a functional L8S8 enzyme when expressed in E. coli (28Andrews T.J. J. Biol. Chem. 1988; 263: 12213-12219Abstract Full Text PDF PubMed Google Scholar). Fusion peptides containing alternately arranged L and S were assembled by cloning different linker peptides between the codons for the C-terminal Tyr-