Oxygenase Large SubunitNϵ-Methyltransferase

PROTEIN EXPRESSION AND PURIFICATION ARTICLE NO.

14, 104 –112 (1998)

PT980936

Expression, Purification, and Characterization of Recombinant Ribulose-1,5-Bisphosphate Carboxylase/Oxygenase Large Subunit Ne-Methyltransferase Qi Zheng,1 Erica J. Simel, Patricia E. Klein, Malcolm T. Royer, and Robert L. Houtz2 Department of Horticulture and Landscape Architecture, Plant Physiology/Biochemistry/Molecular Biology Program, University of Kentucky, Lexington, Kentucky 40546

Received April 21, 1998, and in revised form June 9, 1998

Ribulose-1,5-bisphosphate carboxylase/oxygenase (Rubisco) large subunit (LS) Ne-methyltransferase (Rubisco LSMT, EC 2.1.1.127) catalyzes methylation of the LS of Rubisco. A pea (Pisum sativum L. cv Laxton’s Progress No. 9) Rubisco LSMT cDNA was expressed in Escherichia coli, but most of the expressed protein was found in the insoluble fraction as an inclusion body. Expression at lower temperatures increased the level of soluble Rubisco LSMT and the associated enzymatic activity. However, the soluble form of Rubisco LSMT occurred as two molecular mass forms with the lower molecular mass suggestive of N-terminal processing at Ser-37. Deletion of 108 nucleotides from the 5* end encoding the N-terminal 36 amino acids of Rubisco LSMT resulted in a 10-fold increase in solubility and activity. Further addition of a 3* nucleotide sequence coding for a hexahistidyl carboxy-terminal peptide enabled purification of the N-terminally truncated Rubisco LSMT to homogeneity. Five milligrams of pure recombinant Rubisco LSMT was obtained from a 1-liter E. coli cell culture. The apparent kinetic constants for recombinant Rubisco LSMT for spinach Rubisco and AdoMet were only slightly different from the constants determined using affinity-purified native Rubisco LSMT from pea chloroplasts. However, there was a 6- to 7-fold reduction in the kcat for Rubisco LSMT, which was apparently a consequence of catalytic inactivation due to exposure to NiSO4 during purification. The availability of larger quantities of purified Rubisco LSMT should enable 1 Present address: Tropical Research and Education Center, University of Florida/IFAS, 18905 SW 280th Street, Homestead, FL 33031. 2 To whom correspondence should be addressed. Fax: (606) 2572859.

104

studies of the structure–function relationships in Rubisco LSMT and moreover its interaction with Rubisco. © 1998 Academic Press

Ribulose-1,5-bisphosphate carboxylase/oxygenase (Rubisco)3 catalyzes the carboxylation and oxygenation of ribulose-1,5-bisphosphate, the first committed steps in the competitive metabolic pathways of photorespiration and photosynthetic CO2 fixation in higher plants (1). Previous studies have shown that the N-terminal region of the large subunit (LS) of Rubisco is essential for maximum catalytic activity, but does not directly contribute amino acid residues to the active site (2). This same region of the LS of Rubisco is posttranslationally processed by removal of Met-1 and Ser-2 followed by acetylation of Pro-3 (3,4). In a number of plant species including tobacco and pea, the LS of Rubisco is additionally modified by trimethylation of the e-amino group of Lys-14 (5). However, the LS of Rubisco from spinach and wheat and several other species contains a des(methyl)lysyl residue at position 14. Although protein methylation is a widespread and common posttranslational modification (6), the biological function of posttranslational protein methylation remains unknown. The natural existence of a des(methyl)lysyl residue in the LS of Rubisco in some but not all plant species suggests that the biological function may be speciesrelated. 3 Abbreviations used: AdoMet, S-adenosyl-L-methionine; Rubisco, ribulose-1,5-bisphosphate carboxylase/oxygenase; SDS, sodium dodecyl sulfate; LS, large subunit; PCR, polymerase chain reaction; IPTG, isopropyl-b-D-thiogalactoside; LSMT, large subunit Ne-methyltransferase.

1046-5928/98 $25.00 Copyright © 1998 by Academic Press All rights of reproduction in any form reserved.

e

RECOMBINANT RUBISCO LARGE SUBUNIT N -METHYLTRANSFERASE

It has been speculated that the functional significance of methylated Rubisco may be related to stability through a reduction in the proteolytic susceptibility of the LS of Rubisco. To facilitate in vitro functional studies, a relatively large amount of pure protein is needed. An affinity purification procedure for pea Rubisco large subunit Ne-methyltransferase (LSMT, EC 2.1.1.127) has been described and characterized (7). Native pea Rubisco LSMT was purified approximately 7000-fold from isolated pea chloroplasts in one step using immobilized spinach Rubisco as a ligand. However, the yield of Rubisco LSMT obtained by this method was poor, typically 2.8%. With cloning of the pea Rubisco LSMT cDNA, direct in vitro analysis of substrate specificity, catalytic mechanisms, and functional significance of methylation of Lys-14 in the LS of Rubisco should be possible. A full-length cDNA for Rubisco LSMT from pea has been cloned utilizing the polymerase chain reaction and conventional bacteriophage library screening (8). The 1802-bp cDNA of pea Rubisco LSMT encodes a 489-amino-acid polypeptide with a predicted molecular mass of 55 kDa. A deduced N-terminal amino acid sequence with features common to other chloroplast transit peptides was identified, but the exact processing site could not be determined. The expression of heterologous proteins in E. coli has become a standard technique in molecular biology. The overexpression of recombinant proteins is desirable in the study of structure–function relationships, as well as protein–protein interactions. It is particularly useful for proteins with a natural low abundance and/or those that necessitate sophisticated purification protocols. The availability of a full-length Rubisco LSMT cDNA enabled us to engineer a form of Rubisco LSMT which when produced in E. coli resulted in a high level of expression. Using the pET expression system to introduce a metal-binding peptide tag to the recombinant protein also provided a means for convenient purification. The objectives of the present study were to develop and optimize the expression of soluble recombinant pea Rubisco LSMT in E. coli, develop an affinity purification scheme for bacterially expressed Rubisco LSMT, and subsequently characterize the recombinant Rubisco LSMT for its substrate specificity and kinetic parameters. MATERIALS AND METHODS

Materials. E. coli strain DH5a was purchased from Gibco BRL (Gaithersburg, MD) as competent cells and used for cloning purposes. BL21(DE3)pLysS cells were purchased from Novagen, Inc. (Milwaukee, WI) also as competent cells which were used for protein expression. Plasmid pET23d was purchased from Novagen, Inc. (Milwaukee, WI), and used for expression of target proteins.

105

The Bluescript KS(1)–LSMT is a construct containing a full-length cDNA of pea Rubisco LSMT which was described previously by Klein and Houtz (8). All enzymes used for molecular cloning were purchased from Gibco BRL. Plasmid preparations and PCR products were obtained using commercially available supplies (Promega, Madison, WI). Coomassie Protein Assay Reagent was from Pierce Chemical Co. (Rockford, IL). Immobilon-P transfer membranes were purchased from Millipore Corp. (Bedford, MA). His z Bind Resin was purchased from Novagen Inc., [methyl-3H]AdoMet (70 – 80 Ci/mmol) was from DuPont NEN (Boston, MA). Unlabeled AdoMet from Sigma Chemical Co. (St. Louis, MO) was purified by chromatography on Whatman CM-52 before use according to the method of Chirpich (9). Goat antirabbit IgG alkaline phosphatase (AP) conjugate, AP color development reagents, p-nitro blue tetrazolium chloride, and 5-bromo-4-chloro 3-indolyl were purchased from Bio-Rad Laboratories (Hercules, CA). Oligonucleotide primers were synthesized by the Macromolecular Structure and Analysis Facility at the University of Kentucky using an Applied Biosystems 380B DNA Synthesizer (Foster City, CA). The antibody to pea Rubisco LSMT was elicited in rabbits using a denatured inclusion body of recombinant pea Rubisco LSMT by Charles River PharmServices (Southbridge, MA) and partially purified using Protein G–Sepharose. Spinach (Spinacia oleracea) seed was from Stokes Seeds Inc. (Buffalo, NY). Pea (Pisum sativum) seed was from Asgrow Seed Co. (Kalamazoo, MI). Recombinant VU-1 calmodulin was provided by Dr. Daniel Roberts, University of Tennessee. Rubisco was purified from mature pea and spinach leaf tissue according to the method of McCurry et al. (10). All other chemicals were from Sigma Chemical Co. Construction of pea Rubisco LSMT cDNA in pET23d. An internal NcoI site in the coding region of the pea Rubisco LSMT cDNA was removed using a sitedirected oligonucleotide mutagenesis kit from Amersham (Arlington Heights, IL). A 730-bp SstI–BamHI fragment of the Rubisco LSMT full-length cDNA containing the internal NcoI site was subcloned in an M13 phage vector. The mutant strand was synthesized and ligated according to the Amersham manual. The nonmutant strand was then nicked with NcoI, and digested with exonuclease III. The gapped DNA was then repolymerized and ligated. The mutant Rubisco LSMT fragment was cloned back into the full-length Rubisco LSMT cDNA. The open reading frame for pea Rubisco LSMT was cloned as a NcoI–XhoI fragment into the bacterial expression vector pET-23d (pET23d–LSMT). Using mutated Bluescript KS(1)–LSMT as a DNA template, 108 nucleotides were removed from the 59

106

ZHENG ET AL.

end of the full-length cDNA using PCR with a forward primer of 59-CTTCCATGGCAAAATCAGTAGCCTCTG-39 and a reverse primer of 59-CACCACCACCACCACCACTCGAGTG-39. PCR amplification was carried out in a 100-ml reaction mixture containing a 200 mM concentration of each dNTP, a 100 mM concentration of each primer, 2 ng template DNA, 2 mM MgCl2, 10 ml reaction buffer (200 mM Tris-HCl, pH 8.4, 500 mM KCl), and 3 ml of Taq polymerase. PCR were performed with a Robocycler Gradient 40 (Stratagene, La Jolla, CA) for 30 cycles (30 s of denaturation at 94°C, 30 s of annealing at 50°C, and 90 s of polymerization at 72°C). The PCR products were purified and after digestion with NcoI and XhoI, the DNA fragments were cloned into the E. coli expression vector pET23d (pET23d–DLSMT). Ser-37 was replaced with Met as the start codon for translation of truncated Rubisco LSMT. To add a carboxy-terminal His-tag to the recombinant Rubisco LSMT, a 39 AvrII–XhoI fragment containing the original stop codon was deleted. The construct was blunt-end ligated (pET23d–DLSMT(H)). DNA cloning followed the procedure of Sambrook et al. (11). LB plates with carbenicillin at 50 mg/ml were used for the selection of transformants containing the pET23d vector. Colonies were selected based on plasmid DNA restriction enzyme analysis. Sequences in the ligation region were confirmed by DNA sequencing. Cell induction and cell lysate preparation. For expression, the plasmids pET23d–LSMT, pET23d–DLSMT and pET23d–DLSMT(H) were transformed into E. coli BL21(DE3)pLysS cells. In addition to carbenicillin, chloramphenicol at 34 mg/ml was also added to LB plates for the pLysS plasmid. A single colony of BL21(DE3)pLysS cells harboring pET-LSMT from a freshly streaked plate, or a few microliters from a glycerol stock, was inoculated into LB liquid medium containing the appropriate antibiotics and incubated with shaking at 37°C until the OD600 reached 0.6–1.0. Cultures were transferred to a 15°C incubator, transferred to room temperature (25°C), or remained at 37°C and were induced with 0.4 mM IPTG for 3.5 h. Cells were harvested, washed with sterilized water, and resuspended in standard buffer (50 mM Tris– HCl, pH 8.2, 5 mM MgCl2, 1 mM EDTA), at 1/50 of the original culture volume or 1/25 of the original culture volume for affinity purification of Rubisco LSMT. After cell lysis by freeze–thawing and treatment with DNase (100 mg/ml) for 30 min, the soluble and insoluble fractions were separated by centrifugation at 14,000g for 5 min in a benchtop microcentrifuge or for large-scale preparations in a JA-20 Beckman rotor. Affinity purification of Rubisco LSMT. Recombinant Rubisco LSMT produced from the construct containing a full-length pea Rubisco LSMT cDNA or mature Rubisco LSMT from pea chloroplasts was affinity

purified with spinach Rubisco immobilized on PVDF membranes according to the procedure described by Wang et al. (7). The procedure for affinity purification of His-tagged recombinant proteins was adapted from the pET System Manual (12) with slight modifications for optimization of the purification of recombinant Rubisco LSMT. Two milliliters of His z Bind resin was charged with 10 ml of 50 mM NiSO4 and washed with 10 ml binding buffer (5 mM imidazole, 500 mM NaCl, 20 mM Tris–HCl, pH 7.9). The soluble cell lysate was loaded on the resin column at a flow rate of 10 ml/h. After being washed with 10 bed volumes of binding buffer, the resin was washed with 10 ml washing buffer (40 mM imidazole, 500 mM NaCl, 20 mM Tris–HCl, pH 7.9). Bound Rubisco LSMT was eluted with 5 ml of elution buffer (200 mM imidazole, 500 mM NaCl, 20 mM Tris–HCl, pH 7.9). The eluted Rubisco LSMT was concentrated to 1–2 mg/ml by centrifugal ultrafiltration using a Centricon-30 (Amicon Inc., Beverly, MA) and dialyzed against standard buffer in a microdialyzer (System 500, Pierce Chemical Co.) for 2 h. Protein concentration was determined by the method of Bradford (13) using Rubisco as a standard. Purified Rubisco LSMT was stored in 100-ml aliquots at 220°C. Rubisco LSMT activity determination. Determination of Rubisco LSMT activity was modified slightly from the previously described protocol (14). Assay mixtures contained 100 mM Hepes–KOH (pH 8.0), 2 mM MgCl2, 28 mM [methyl-3H]AdoMet (36 Ci/ mmol), and 200 –300 mg purified spinach Rubisco in a final volume of 20 ml. The enzyme source for Rubisco LSMT was 2 ml of soluble cell lysate prepared as previously described, 1–2 m l of purified Rubisco LSMT, or other sources as described accordingly. Reactions were initiated by addition of AdoMet and incubated at 30°C for 2–10 min. The reactions were terminated by addition of 500 ml of 10% TCA. After centrifugation at 14,000g for 5 min, the protein pellets were dissolved in 150 ml of 0.1 N NaOH and reprecipitated with 500 ml of 10% TCA. The protein pellets were dissolved in 50 ml of 88% formic acid followed by the addition of 50 ml of water. Incorporated radioactivity was determined with a liquid scintillation analyzer (Packard 2000, Packard Instrument Co., Downers Grove, IL) after addition of 1.3 ml scintillation cocktail (Bio-Safe II, Research Products International Corp., Mount Prospect, IL). Rubisco LSMT activity is expressed as picomoles of CH3 groups transferred from [methyl-3H]AdoMet to spinach Rubisco per minute per milligram of protein or per unit sample as indicated. Protein electrophoresis, blotting, and phosphorimaging. Cell lysates or debris suspensions were prepared by adding an equal volume of SDS–PAGE sample prep-

e

RECOMBINANT RUBISCO LARGE SUBUNIT N -METHYLTRANSFERASE

aration buffer (50 mM Tris–HCl, pH 6.8, 10 mM DTT, 2% SDS, 0.01% bromophenol blue, and 10% glycerol) and heating at 100°C for 2 min. Samples (10 –15 ml) were loaded on 15% acrylamide gels for electrophoresis. After electrophoresis, proteins were detected by staining with Coomassie brilliant blue R-250 and destaining with methanol:acetic acid (40:10). The gels were mounted with BioDesign Gel Wrap Membrane (BioDesign Inc., Carmel, New York) according to the manufacturer’s instructions. For Western blotting, proteins were electrophoretically transferred to PVDF membranes using a MilliBlot-SDE Transfer System (Millipore Corp.). The membranes were incubated with blocking buffer (1% BSA in TTBS buffer containing 100 mM Tris–HCl, pH 7.5, 500 mM NaCl, 0.05% Tween 20) at room temperature for 1 h, followed by incubation with Rubisco LSMT antibody (1:100,000) in blocking buffer for 2 h. After three 5-min washes with TTBS buffer, the membranes were incubated with goat anti-rabbit IgG alkaline phosphatase conjugate (Bio-Rad) in blocking buffer for 1 h. Following a second TTBS wash, proteins were visualized by addition of substrate solution containing AP color development reagents NBT (300 mg/ ml) and BCIP (150 mg/ml) in substrate buffer (100 mM NaHCO2, 1 mM MgCl2, pH 9.5). Development was stopped by rinsing with H2O. For phosphorimaging, proteins were electrophoretically transferred to PVDF membranes as described above. The membranes were washed with methanol and then H2O. The air dried membranes were exposed for at least 24 h to a phosphorimager screen. Proteins methylated with [methyl-3H]AdoMet were detected using a PhosphorImager Model 425 (Molecular Dynamics, Sunnyvale, CA). Determination of recombinant Rubisco LSMT enzyme specificity. To determine the specificity of recombinant Rubisco LSMT, potential substrates for Lys methylation, such as histone type III-s from calf thymus, horse heart cytochrome c, and recombinant VU-1 calmodulin (15), as well as methylated pea Rubisco and BSA, were used as substitutes for spinach Rubisco in enzyme assays. Additional assays were performed and terminated by direct addition of SDS–PAGE sample preparation buffer. The samples were processed for electrophoresis on 15% SDS–PAGE gels, followed by electrophoretic transfer to PVDF membranes for phosphorimaging. Edman degradative protein sequencing. Proteins were resolved on SDS–PAGE gels and electrophoretically transferred to Immobilon-P membranes as previously described. After a brief staining with Coomassie brilliant blue R-250 and destaining with 40% methanol/10% acetic acid, the membrane containing the protein of interest was excised with a razor blade and

107

FIG. 1. The effect of induction temperature on expression of pea Rubisco LSMT in E. coli. E. coli cells were cultured at 37°C until the OD600 reached 0.6 –1.0 and were then transferred to 15, 25, or 37°C. IPTG (0.4 mM) was added to the cell cultures and incubation continued for 3.5 h. Cells were harvested, resuspended, and treated with DNase as described under Materials and Methods. Recombinant Rubisco LSMT was resolved on a 15% SDS–PAGE gel, electrophoretically transferred to a PVDF membrane, and detected with Rubisco LSMT antibody. Cell lysates were processed for determination of Rubisco LSMT activity as described under Materials and Methods. MW, low-range molecular mass markers. Lanes 1–3, insoluble fractions from E. coli lysates induced at 15, 25, and 37°C, respectively; lanes 4 –5, soluble fractions from E. coli lysates induced at 15, 25, and 37°C, respectively. ND, not detectable.

submitted for direct amino acid sequencing. Sequencing was performed at the Macromolecular Structure Analysis Facility at the University of Kentucky, using an Applied Biosystems 477A automated sequencer. RESULTS

Apparent processing of recombinant pea Rubisco LSMT in E. coli. When plasmid pET23d–LSMT containing the full-length Rubisco LSMT cDNA was used to express Rubisco LSMT in E. coli, the majority of expressed protein was in the insoluble fraction as determined by SDS–PAGE gels (Fig. 1). Induction of E. coli at a lower temperature (25°C versus 37°C) increased the enzymatic activity in the soluble fraction of E. coli cell lysates. Western blots of cell lysates induced at different temperatures indicated that the amount of soluble Rubisco LSMT as well as enzymatic activity increased with a lower induction temperature, whereas the amount of insoluble Rubisco LSMT decreased (Fig. 1). Affinity purification utilizing PVDF-immobilized spinach Rubisco as described previously (7) revealed that recombinant pea Rubisco LSMT expressed in E. coli was present in two slightly different molecular mass forms (Fig. 2). The lower molecular mass form of recombinant Rubisco LSMT was similar in mass to mature Rubisco LSMT isolated from pea chloroplasts. When the two recombinant forms of Rubisco LSMT were

108

ZHENG ET AL.

FIG. 2. Western blot of precursor and processed forms of pea Rubisco LSMT. E. coli cells were cultured and induced with IPTG at 25°C. Soluble cell lysates were prepared as described under Materials and Methods. Rubisco LSMT was affinity-purified using PVDFimmobilized spinach Rubisco as previously described. The mature form of Rubisco LSMT was purified from intact pea chloroplasts using the same method. Purified Rubisco LSMT was resolved on a 15% SDS–PAGE gel, electrophoretically transferred to the PVDF membrane, and detected with Rubisco LSMT antibody. Lane 1, processed mature pea Rubisco LSMT; lane 2, recombinant pea Rubisco LSMT. The relative location of low-range protein molecular mass markers is indicated to the left. The N-terminal amino acids of precursor (upper band) and processed forms (lower band) are indicated to the right.

submitted for N-terminal Edman degradative sequencing, the precursor form could not be sequenced, perhaps because of an N-terminal formyl-methionyl residue, while the lower molecular mass form started at Ser-37 (Table 1). This result suggested that the full-length pea Rubisco LSMT was processed to remove the N-terminal 36 amino acids when expressed in E. coli cells and that processing increased the solubility of Rubisco LSMT. The similarity in molecular mass between bacterially N-terminally processed Rubisco LSMT and mature Rubisco LSMT from pea chloroplasts suggests that Ser-37 may represent the site of chloroplast processing for removal of the transit peptide. An analysis of the N-terminal region (60 residues) of pea Rubisco LSMT for the presence and location of a signal peptide cleavage site was negative (16; http://genome.cbs.dtu.dk/ services/SignalP/). N-terminal truncation increases the solubility of recombinant Rubisco LSMT. The plasmid pET23d– DLSMT containing a truncated Rubisco LSMT cDNA

with 108 59 nucleotides removed was used to express pea Rubisco LSMT in the E. coli strain BL21(DE3)pLysS. When cells harboring this construct were induced with IPTG, a protein with a molecular mass expected for pea Rubisco LSMT without the N-terminal 36 amino acids was detected (Fig. 3). However, unlike the full-length pea Rubisco LSMT, the majority of truncated Rubisco LSMT was found in the soluble fraction. Furthermore, cell lysates from pET23d–DLSMT had a Rubisco LSMT activity of 40,299 pmol z min21 z mg protein21, but lysates from pET23d–LSMT had an activity of 4776 pmol z min21 z mg protein21. The tremendous increase in enzyme activity in cell lysates from pET23d–DLSMT compared to the cell lysates from pET23d–LSMT was attributable to an increase in soluble Rubisco LSMT in the cell lysate. Expression and purification of truncated Rubisco LSMT with a C-terminal His Tag. The plasmid pET23d–DLSMT(H) containing the insert corresponding to the pea Rubisco LSMT sequence with both the N-terminus and the C-terminus (to accommodate the His tag) deleted was used to express a high level of soluble Rubisco LSMT. The expressed protein existed in the soluble lysate as a large percentage of the total Rubisco LSMT protein (;90%). Rubisco LSMT enzymatic activity was apparently not affected by the deletion of four amino acids and addition of six histidyl residues at the C terminus. Under optimal conditions, a 100-ml culture produced more than 1 mg Rubisco LSMT in the soluble fraction as estimated on stained SDS–PAGE gels. Approximately 90% of the Nterminally truncated and carboxy-His-tagged Rubisco LSMT bound to a His z Bind resin with 73% recovered using the modified purification protocol. Thus, approximately 500 – 600 mg of pure Rubisco LSMT could be TABLE 1 N-Terminal Sequence of the E. coli Processed Form of Recombinant Pea Rubisco LSMTa

Cycle No.

Amino acid and yield (pmol)

1 2 3 4 5 6

S (20) A (28) K (20) S (20) V (39) A (34)

a Recombinant Rubisco LSMT was affinity-purified from soluble cell lysates of E. coli harboring pET23d–LSMT using PVDF-immobilized spinach Rubisco as previously described (7). Purified Rubisco LSMT was resolved on a 15% SDS–PAGE gel and electrophoretically transferred to a PVDF membrane, and the membrane regions containing precursor and processed forms of pea Rubisco LSMT were excised with a razor blade and submitted for Edman degradative sequencing. The precursor form of recombinant pea Rubisco LSMT could not be sequenced.

e

RECOMBINANT RUBISCO LARGE SUBUNIT N -METHYLTRANSFERASE

FIG. 3. Expression of full-length Rubisco LSMT and N-terminally truncated Rubisco LSMT in E. coli. Recombinant full-length Rubisco LSMT and N-terminally truncated Rubisco LSMT were expressed under E. coli strain BL21(DE3)pLysS under identical conditions. Cells were cultured and induced with IPTG at 25°C. Cell lysates were prepared as described under Materials and Methods. Five micrograms of total protein from soluble cell lysates was used for enzyme assays. MW, low-range molecular mass markers. Lane 1, soluble protein from pET23d–DLSMT; lane 2, soluble protein from pET23d–LSMT; lane 3, insoluble protein from pET23d–DLSMT; lane 4, insoluble protein from pET23d–LSMT. Arrows indicate positions of full-length and N-terminally truncated forms of Rubisco LSMT.

obtained from a 100-ml cell culture (Fig. 4). The increased solubility and relatively large amount of Rubisco LSMT in the cell lysate considerably enhanced the binding of His-tagged Rubisco LSMT to the His z Bind resin. However, the binding of Rubisco LSMT was still not tight, and a large part of the bound Rubisco LSMT was washed off with 60 mM imidazole buffer. To increase the recovery, the concentration of imidazole in the wash buffer was reduced to 40 mM. The remaining Rubisco LSMT was eluted with 200 mM imidazole. Substrate specificity of purified recombinant Rubisco LSMT. To determine whether the modified recombinant Rubisco LSMT had suffered changes in protein substrate specificity, BSA, histone type III-2 from calf thymus, horse heart cytochrome c, recombinant VU-1 calmodulin, and pea and spinach Rubisco were tested as methylation substrates. Among these protein substrates, only des(methyl) spinach Rubisco was capable of accepting methyl-3H groups from [methyl-3H]AdoMet (Fig. 5). Other protein methyltransferase III substrates had only minor levels of methyl-3H group incorporation, and this observation was corroborated by activity assays. A phosphorimage of the methylated protein substrates confirmed that methyl-3H groups

109

incorporated into spinach Rubisco by recombinant pea Rubisco LSMT were located exclusively in the LS based on the molecular mass of the radiolabeled protein band. Thus, under the conditions in these studies, methylation of spinach Rubisco by recombinant pea Rubisco LSMT with sequence alterations in the C- and N-termini, was identical to the specificity of native pea Rubisco LSMT isolated from intact pea chloroplasts. Kinetic analysis of recombinant pea Rubisco LSMT. The kinetic parameters of recombinant pea Rubisco LSMT were determined. Figure 6 shows the Michaelis–Menten kinetic curves indicating a Km for Rubisco of 8.30 mM, a Km for AdoMet of 8.27 mM, and a Vmax between 42.5 and 60.2 nmol z min21 z mg protein21. Native pea Rubisco LSMT purified by affinity chromatography using PVDF-membrane-immobilized spinach Rubisco had a Km for spinach Rubisco of 2.7 mM, a Km for AdoMet of 13.7 mM, and a Vmax between 272 and 378 nmol z min21 z mg protein21 (data not shown). While there were only slight differences in the Km values between recombinant and native Rubisco LSMT, the changes in Vmax were large and may be indicative of the presence of denatured or catalytically inactive enzyme. This could be due to the slight modifications of recombinant Rubisco LSMT required for affinity purification using the His- tag and/or the con-

FIG. 4. Affinity purification of N-terminally truncated Rubisco LSMT with a carboxy-terminal His- tag. Cell lysates were prepared from 100-ml cultures which had been induced for 3.5 h at room temperature with IPTG (0.4 mM). The cell lysates were loaded onto a His z Bind resin column with a 2-ml bed volume and after washing with 40 mM imidazole, Rubisco LSMT was eluted with an elution buffer containing 200 mM imidazole. MW, low-range protein molecular mass markers. Lane 1, cell lysate from pET23d–DLSMT(H); lane 2, cell lysate after passage over the His z Bind resin column; lane 3, 40 mM imidazole buffer wash; lane 4, 200 mM imidazole buffer elution. (Bottom) The respective Rubisco LSMT activities and yields for each of the fractions shown in lanes 1– 4. Arrow indicates the position of Rubisco LSMT.

110

ZHENG ET AL.

changes in the measured affinities for spinach Rubisco and AdoMet by recombinant Rubisco LSMT would also support this deduction. DISCUSSION

These results demonstrate that after induction with IPTG at 37°C a high level of expression of full-length pea Rubisco LSMT could be achieved in E. coli. However, the majority of Rubisco LSMT was insoluble, and little enzyme activity could be detected. Induction with IPTG at lower temperatures resulted in substantial increases in the level of soluble Rubisco LSMT in crude cell lysates and a decrease in the level of insoluble Rubisco LSMT. Because of the low level of soluble Rubisco LSMT in cell lysates, reconstructing the Rubisco LSMT cDNA into a pET28b vector with a fused His tag at the N-terminus would not have allowed efficient affinity purification of recombinant Rubisco LSMT from the cell lysates. The full-length precursor form of pea Rubisco LSMT includes a chloroplast-specific N-terminal transit sequence of unknown length (8). We observed that some of the soluble precursor form of Rubisco LSMT expressed in E. coli was processed to produce a lower

FIG. 5. Protein substrate specificity of recombinant pea Rubisco LSMT. Approximate 50 mg of each test protein was used as a substrate for methylation in Rubisco LSMT activity assays. After incubation at 30°C for 10 min, one set of assays was processed for liquid scintillation analysis, and another set of assays was processed by electrophoresis on 15% SDS–PAGE gels and electrophoretic transfer to PVDF membranes for phosphorimage analysis. (A) 15% SDS– PAGE gel stained with Coomassie brilliant blue R-250. (B) Phosphorimage of methyl-3H incorporation into test protein substrates. MW, low-range molecular mass markers. Lane 1, BSA; lane 2, histone type III-S; lane 3, horse heart cytochrome c; lane 4, recombinant VU-1 calmodulin; lane 5, spinach Rubisco; lane 6, pea Rubisco.

sequences of exposing recombinant pea Rubisco LSMT to nickel during the immobilization phase of purification. Experiments with purified native pea Rubisco LSMT showed that up to 100 mM imidazole and 100 mM NiSO4 in the assay mixture did not affect enzymatic activity (data not shown). However, when purified native pea Rubisco LSMT was treated with 10 mM NiSO4 for 2.5 h followed by exhaustive dialysis, complete and irreversible inactivation of Rubisco LSMT activity was observed. Thus, we speculate that the interaction of carboxy-His-tagged Rubisco LSMT with Ni ions during affinity purification probably results in a similar type of inactivation. The relative small

FIG. 6. Kinetic analysis of recombinant pea Rubisco LSMT. Rubisco LSMT activity was determined as previously described (100 mM Hepes–KOH, pH 8.0, 2 mM MgCl2, 50 mM [methyl-3H]AdoMet, 30°C, 10 min) with various concentrations of spinach Rubisco from 0.25 to 24.00 mM (A) and [methyl-3H]AdoMet from 0.15 to 38.40 mM (B). The data were graphed, analyzed, and plotted using Fig. P for Windows from Biosoft (Cambridge, UK).

e

RECOMBINANT RUBISCO LARGE SUBUNIT N -METHYLTRANSFERASE

molecular mass form. This form of Rubisco LSMT had a molecular mass similar to that of native Rubisco LSMT isolated from pea chloroplasts and had an N-terminal sequence starting at Ser-37. It has been reported that leader peptidase from E. coli is able to process the precursor form of pea cytochrome f synthesized in vitro (17). Cleavage by leader peptidase generated the same mature sequence as in pea chloroplasts. Utilizing these observations, plasmid pET23d–DLSMT was constructed with the 108 59 nucleotides removed from the full-length Rubisco LSMT cDNA. The N-terminally truncated form of recombinant Rubisco LSMT had substantially greater solubility in the E. coli cell lysates and was accompanied by increased enzymatic activity. van Wuytswinkel et al. (18) recently reported the influence of an N-terminal deletion on protein solubility. Recombinant pea-seed ferritin expressed from constructs with the transit peptide deleted accumulated in the soluble fraction while the precursor form of pea-seed ferritin was insoluble. Moreover, deletion of the transit peptide was necessary to express correctly assembled recombinant plant ferritin in E. coli. The transit peptide of Rubisco LSMT responsible for chloroplast targeting apparently had no effect on enzyme activity since the methylation of des(methyl) Rubisco could be demonstrated using fulllength recombinant Rubisco LSMT. However, deletion of the transit peptide was necessary to maximize expression of soluble pea Rubisco LSMT in E. coli. Fusion of a His tag to the carboxy terminus of Nterminally truncated Rubisco LSMT required the removal of the original stop codon. Since there was a convenient restriction site (AvaII) near the stop codon, deletion of an AvaII–XhoI fragment removed the stop codon and a C-terminal hexapeptide. Binding between the His-tagged recombinant Rubisco LSMT and the His z Bind resin was not exceptionally tight, since 60 mM imidazole removed more than 50% of the bound Rubisco LSMT. As a consequence of this, the imidazole concentration in the wash buffer was reduced to 40 mM, even though this resulted in some loss of the bound Rubisco LSMT. Experiments were conducted to ensure that recombinant Rubisco LSMT retained proper substrate specificity and kinetic constants. Enzyme assays were conducted using BSA, histone type III-s from calf thymus, horse heart cytochrome c, recombinant VU-1 calmodulin, and pea and spinach Rubisco as substrates. Histone type III-s from calf thymus, horse heart cytochrome c, and recombinant VU-1 calmodulin are known protein substrates for other protein methyltransferase III enzymes, and native pea Rubisco is known to stoichiometrically contain a trimethyllysyl residue at position 14 in the LS (5). The results from these experiments demonstrated that none of these

111

proteins served as a substrate for recombinant Rubisco LSMT except for des(methyl) spinach Rubisco. It has been reported that limited tryptic proteolysis of spinach Rubisco results in the ordered release of two adjacent N-terminal peptides from the large subunit identified as Pro-3 to Lys-8 and Ala-9 to Lys-14 (3,4). Trimethylation at Lys-14 of the LS of Rubisco results in a much slower release of the penultimate Nterminal peptide Ala-9 to Lys-18 (2). When tryptic proteolysis was applied to spinach Rubisco following in vitro methylation by recombinant pea Rubisco LSMT, a distinctly slower release of radiolabeled tryptic peptides was observed (data not shown). This observation, as well as the data in Fig. 5, demonstrates that the methylation of spinach Rubisco catalyzed by recombinant Rubisco LSMT occurred exclusively at the Nterminus of the large subunit. The apparent substrate affinity constants for recombinant Rubisco LSMT were only slightly different from the constants determined using affinity-purified native Rubisco LSMT from pea chloroplasts. However, the Vmax values were 6- to 7-fold less. This may be speculated to be partially due to the affinity purification process which uses nickel as the binding ligand and imidazole as the elution agent. Experiments conducted to determine the effect of nickel and imidazole on native Rubisco LSMT activity revealed that up to 100 mM imidazole and 100 mM NiSO4 in activity assays showed no obvious effect on purified native pea Rubisco LSMT. However, exposure to 10 mM NiSO4 for 2.5 h completely inactivated Rubisco LSMT and the enzyme remained inactive even after extensive dialysis. Comparison of recombinant Rubisco LSMT activity before and after affinity purification indicated a 30 –50% reduction of enzymatic activity following elution from the His z Bind resin. No obvious differences in enzymatic activity of His-tagged Rubisco LSMT and Rubisco LSMT without a His tag in the absence of affinity purification were observed. Still, we cannot rule out the possibility that some posttranslational modification might occur in plant chloroplasts that is absent in bacterial cells and that might account for the differences in kinetic parameters. However, given the small difference in the Km values for Rubisco and AdoMet between native and recombinant Rubisco LSMT, the reduced Vmax values are likely due to the presence of irreversibly inactivated Rubisco LSMT and not kinetically compromised enzyme. Successful heterologous expression and purification of recombinant pea Rubisco LSMT in E. coli and the availability of larger quantities of purified Rubisco LSMT will enable further studies of the structure– function relationships of Rubisco LSMT and its interaction with Rubisco.

112

ZHENG ET AL.

ACKNOWLEDGMENTS This work is Kentucky Agricultural Experiment Station Article No. 98-11-58 and was supported by U.S. Department of Energy Grant DE-FG05-92ER20075 to R.L.H.

REFERENCES 1. Andrews, T. J., and Lorimer, G. H. (1987) in “The Biochemistry of Plants” (Stumpf, P. K., and Conn, E. E., Eds.), Vol. 10, pp. 131–218, Academic Press, New York. 2. Houtz, R. L., and Mulligan, R. M. (1991) Protection of trypticsensitive sites in the large subunit of ribulose bisphosphate carboxylase/oxygenase by catalysis. Plant Physiol. 96, 335–339. 3. Mulligan, R. M., Houtz, R. L., and Tolbert, N. E. (1988) Reactionintermediate 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, 1513–1517. 4. Houtz, R. L., Stults, J. T., Mulligan, R. M., and Tolbert, N. E. (1989) Post-translational modifications in the large subunit of ribulose bisphosphate carboxylase/oxygenase. Proc. Natl. Acad. Sci. USA 86, 1855–1859. 5. Houtz, R. L., Poneleit, L., Jones, S. B., Royer, M., and Stults, J. T. (1992) Posttranslational modifications in the amino-terminal region of the large subunit of ribulose-1,5-bisphosphate carboxylase/oxygenase from several plant species. Plant Physiol. 98, 1170 –1174. 6. Paik, W. K., and Kim, S. (1990) “Protein Methylation,” CRC Press, Boca Raton, FL. 7. Wang, P., Royer, M., and Houtz, R. L. (1995) Affinity purification of ribulose-1,5-bisphosphate carboxylase/oxygenase large subunit {N-methyltransferase. Protein Expression Purif. 6, 528 – 536. 8. Klein, R. R., and Houtz, R. L. (1995) Cloning and developmental expression of pea ribulose-1,5-bisphosphate carboxylase/oxygen-

9.

10.

11.

12. 13.

14.

15.

16.

17.

18.

ase large subunit N-methyltransferase. Plant Mol. Biol. 27, 249 – 261. Chirpich, T. P. (1968) “Lysine-2,3-amminomutase: Purification and Properties.” Ph.D. Thesis, University of California, Berkeley. McCurry, S. D., Gee, R., and Tolbert, N. E. (1982) Ribulose-1,5bisphosphate carboxylase/oxygenase from spinach, tomato or tobacco leaves. Methods Enzymol. 90, 515–521. Sambrook, J., Fritsch, E. F., and Maniatis, T. (1989) “Molecular Cloning: A Laboratory Manual,” 2nd ed., Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY. Novagen, Inc. (1995) “pET System Manual,” 5th ed., Novagen, Inc., Madison, WI. Bradford, M. M. (1976) A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Anal. Biochem. 72, 248 –255. Houtz, R. L., Royer, M., and Salvucci, M. E. (1991) Partial purification and characterization of ribulose-1,5-bisphosphate carboxylase/oxygenase large subunit sN-methyltransferase. Plant Physiol. 97, 913–920. Roberts, D. M., Crea, R., Malecha, M., Alvarado-Urbina, G., Chiarello, R. H., and Watterson, D. M. (1985) Chemical synthesis and expression of a calmodulin gene designed for site-specific mutagenesis. Biochemistry 24, 5090 –5098. Nielsen, H., Engelbrecht, J., Brunak, S., and von Heijne, G. (1997) Identification of prokaryotic and eukaryotic signal peptides and prediction of their cleavage sites. Protein Eng. 10, 1– 6. Anderson, C. M., and Gray, J. (1991) Cleavage of the precursor of pea chloroplast cytochrome f by leader peptidase from Escherichia coli. FEBS Lett. 280, 383–386. van Wuytswinkel, O., Savino, G., and Briat, J.-F. (1995) Purification and characterization of recombinant pea-seed ferritins expressed in Escherichia coli: Influence of N-terminus deletions on protein solubility and core formation in vitro. Biochem. J. 305, 253–261.