trol, Riviera Beach, FL) was used as both an analytical and prepara- tive procedure. Peptide mixtures (1-15 nmol) were injected onto a. 0.46 X 25-cm Lichrosorb ...
Vol. 263, No. 1, Issue of January 5, pp. 123-129,1988 Printed in U.S.A.
THEJOURNALOF BIOLOGICAL CHEMISTRY
Characterization of the Regulatory Thioredoxin Site of Phosphoribulokinase* (Received for publication, July 20, 1987)
Michael A. Porter$, Claude D. Stringer, and FredC. Hartman From the Biology Division, Oak Ridge National Laboratory, and the University of Tennessee-Oak Ridge Graduate School of Biomedical Sciences, Oak Ridge, Tennessee 37830
Phosphoribulokinase is light-regulated via thioredoxin by reversible oxidationlreduction of sulfhydryl/ disulfide groups. To identify the cysteinyl residues that are involved in regulation, the S-carboxymethyl labeling patterns of the fully reduced (active) andoxidized (inactive) formsof the enzyme were compared. Tryptic digests of the reduced, [ 14C]carboxymethylatedenzyme contained four labeled peptides, all of which were purified and sequenced by Edman degradation. If the enzyme was oxidized by 5,5’-dithiobis-(2-nitrobenzoic acid) prior to carboxymethylation and tryptic digestion, only two labeled peptides were observed, thereby revealing the identity of the regulatory cysteines as Cys-16 and Cys-55. The former waspreviously implicated as part of the nucleotide-binding domain of the active site (Porter, M. A., and Hartman, F. C. (1986) Biochemistry 25, 7314-7318),a conclusion reinforced by the present observation that the sequence around the Cys-16 is similar to a consensus sequence of ATP-binding sites from a number of proteins of diverse phylogenetic origin (Higgins, C. F., Hiles, I. D., Salmond, G. P. C., Gill, D. R., Downie, J. A., Evans, I. J., Holland, I. B., Gray, L., Buckel, S. D., Bell, A. W., and Hermondson, M. (1986) Nature 323, 448-450). The regulatory disulfide of phosphoribulokinase was found to be intrasubunit based on the stoichiometry of the oxidation and the failure to resolve oxidized and reduced enzyme by gel filtration under dissociation conditions.
Although many enzymes function only in thereduced form, formost organismsthere is little compelling evidence to a significant suggest thatoxidation/reductionrepresents mode of regulation in vivo (1, 2). A notable exception is a group of photosynthetic enzymes, where catalytic activity is clearly linked to the light/dark cycle by oxidation/reduction (3). These enzymes include fructose bisphosphatase, NADPglyceraldehyde-phosphate dehydrogenase, sedoheptulose bis-
phosphatase, NADP-malate dehydrogenase, NADP-glucose6-phosphate dehydrogenase, and phosphoribulokinase. With the exception of glucose phosphate dehydrogenase, all are activated via reduction by thioredoxin (3). This mechanism of regulation is absent in anaerobic prokaryotes, but appears to function in all aerobic photosynthetic organisms (3, 4). Although considerable information has accumulateddescribing the physiological process of thioredoxin-dependent activation, the relationship between activation and catalysis a t the molecular level hasnot beendescribed. Conceivably, oxidationcouldlead to inactivation indirectly by inducing conformational changes ordirectly by altering the properties of the activesite. An essential cysteinyl residue of phosphoribulokinase has recently been identified near the amino terminus(5-7). This cysteine(Cys-16) is selectively alkylated by BAEP,’ with substantial protection against inactivation afforded by ATP or ADP (5, 6), andselectively sulfonylated by the ATP analogue 5’-p-fluorosulfonylbenzoyladenosine(7). These observations led to the conclusion that Cys-16 is located near the ATP-binding domain of the active site. ATP also retarded the rate of oxidative deactivation of the kinase by molecular oxygen (8), dehydroascorbate (5), oroxidized glutathione (9), suggestingapossibleregulatory role for Cys-16 (5).The kinetics of theseprotective effects as well assimilar p H dependencies of oxidation and alkylationimplied that a single sulfhydryl was the site of both modifications (5). These parallels prompted experiments toverify directly that Cys-16 of phosphoribulokinase is involved in regulation of enzyme activity through oxidation/reduction and to identify the other regulatory cysteinyl residue with which Cys-16 is paired to form a disulfide. EXPERIMENTAL PROCEDURES~
Materials-Phosphoribulokinase was purified from spinach leaves as previously described (8), excepting two modifications of the final affinity chromatography step. The protein to be applied to the reactive-red column was dialyzed overnight against 30 mM Bicine-KOH (pH 6.5), 10 mM DTT, which had been thoroughly flushed with N,. After washing the loaded column with dialysis buffer, contaminating
* This research was sponsored jointly by the Science and Education Administration of the United States Department of Agriculture under Grant 84-CRCR-1-1520from the Competitive Research Grants Office and by the Office of Health and Environmental Research, United States Department of Energy, under Contract DE-AC05-840R21400 with the Martin Marietta Energy Systems, Inc. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. $ Postdoctoral Investigator supported by the United States Department of Agriculture through Subcontract 3322 from the Biology Division of OakRidge National Laboratory to the University of Tennessee.
I The abbreviations used are: BAEP, (bromoacety1)ethanolamine phosphate; DTNB, 5,5’-dithiobis-(2-nitrobenzoic acid); TNB, 5-thio2-nitrobenzoate; Bicine, N,N‘-bis(2-hydroxyethyl)glycine;Ru5P, Dribulose 5-phosphate; HPLC, high performance liquid chromatography; DEAE, diethylaminoethyl; DTT, dithiothreitol; Hepes, N-2hydroxyethylpiperazine-N’-2-ethanesulfonic acid; PTH, phenylthiohydantoin; PRK, phosphoribulokinase, in Miniprint. * Portions of this paper (including part of “Experimental Procedures,” part of “Results,” and Tables 11-VIII) are presented in miniprint at theend of this paper. Miniprint is easily read with the aid of a standard magnifying glass. Full size photocopies are included in the microfilm edition of the Journal that is available from Waverly Press.
123
124
Regulatory Thioredorin Site
aldolase was eluted by the inclusion of 0.1 mM fructose bisphosphate in the pH 6.8 phosphate wash. Prior to the elution of phosphoribulokinase with the ATP-containing buffer, the column was washed with pH 6.8 phosphate buffer that lacked fructose bisphosphate. Spinach thioredoxin-f was partially purified through the G-75 gel filtration step of a previously published procedure (10) with some modifications: buffer A was replaced by 50mM Bicine-KOH (pH 8.0), 1 mM EDTA, 10 mM 2-mercaptoethanol; the acetone precipitation, overnight dialysis, and buffer exchange steps were omitted; and the gel filtration buffer was not supplemented with high salt. The pooled fractions containing thioredoxin-f were concentrated by dialysis against the Bicine buffer that was saturated with ammonium sulfate followed by centrifugation to collect the precipitated protein, which was then dialyzed against the Bicine buffer that contained 20% (v/v) glycerol. The concentratedpreparation of thioredoxin (30% pure based on denaturing polyacrylamide gel electrophoresis) was stored at -80 “C until use. The following materials were obtained from the indicated vendors: assay reagents for phosphoribulokinase, Sigma; high purity HPLC solvents, Burdick & Jackson; iodoacetic acid, Aldrich; [l-”C]iodoacetic acid, Du Pont-New England Nuclear; sequencing reagents, Applied Biosystems; trifluoroacetic acid, Pierce Chemical Co.; DTNB, Behring Diagnostics; DEAE-cellulose, Whatman; Ultrogel AcA34, LKB; and DTT, Bethesda Research Laboratories. Protein Concentration-The concentration of purified phosphoribulokinase was determined from the absorbancy (280 nm) of the solution and an Et:,,, of7.2, a value determined experimentally by amino acid analysis of kinase solutions of known absorbancies (8). Activation and Assay of Phosphoribulokinase-The modified (8) enzyme-coupled assay of Racker (11) was used to monitor phosphoribulokinase activity. One unit of enzyme activity causes an absorbancy change (340 nm) of6.22 min” in the 1-ml assay solution. Activation of phosphoribulokinase by DTT or thioredoxin was conducted at 25 “C by dilution of the oxidized enzyme into 45 mM BicineKOH (pH 8.0), 0.9 mM EDTA, 10% (v/v) glycerol, containing the desired concentration of thiol. Oxidation of Phosphoribulokime-Prior to oxidation of purified kinase, exogenous thiol was removed by rapid gel filtration on Sephadex G-25 with 45 mM Bicine-KOH (pH 8.0), 0.9 mM EDTA, 10% (v/v) glycerol as the equilibration buffer. The resulting kinase solution wasoxidized with 1 molof DTNB/mol of subunit at room temperature for 1 min. Activity was typically only 1.0% of the initial level after this treatment. TNB, the reaction by-product was removed by dialysis against the equilibration buffer. Stoichiometry of oxidation of the kinase was determined by monitoring TNB formation at 412 nm (12). Two-pl aliquots of0.5 mM DTNB were added at 4-5-min intervals to 0.5 ml of phosphoribulokinase (0.34 mg/ml) in equilibration buffer. After each addition of reagent, the absorbance at 412 nm was measured, and a 10-pl aliquot of the reaction mixture was removed and stored onice for later enzyme assays. Gel Filtration of Denatured Phosphoribulokime-A column of Ultrogel AcA 34 (1 X 105 cm) was equilibrated at room temperature with a solution of 6 M urea, 50 mM Hepes-KOH (pH 6.81,50 mM KC1, and 1mM EDTA that was saturated with Nz.Samples of oxidized or reduced kinase were brought to 6 M urea in a volume of 1 ml and applied to the column which was then eluted with thiol-free, NPsaturated equilibration buffer at 4 ml/h. Effluent was monitored for absorbance at 280 nm. As a control, the column was re-equilibrated with the same buffer containing 10 mM DTT, anda sample of reduced phosphoribulokinase was chromatographed as above. Carboxymethylation of Phosphoribulokinase-For radioactive labeling of cysteinyl residues of reduced phosphoribulokinase, protein solutions (containing 10mM DTT) were brought to 30 mM iodoacetic acid in 1.2 M NaHCOd; 10 acid (from a stock of 1 M [14C]i~doacetic mM iodoacetic acid was used to alkylate oxidized phosphoribulokinase, which lacked exogenous DTT. Incubation was carried out on ice in the dark for 10 min. Samples of phosphoribulokinase, which were initially in the reduced (active) form, were quenched by the addition of a 2-fold molar excess of DTT and dialyzed. Samples of phosphoribulokinase, which were initially in the oxidized (inactive) form, were not quenched with thiol butplaced directly into the dialysis chamber. If total alkylation of protein sulfhydryls was desired, an equal volume of 8 M guanidine HC1,lOO mM phosphate, 3mM EDTA (pH 8.0) was added immediately after the iodoacetic acid. All alkylated samples were freed of reagent and denaturant by exhaustive dialysis against 50 mM ammonium bicarbonate (pH 8.0). Tryptic Digestion-The carboxymethylated protein, after thorough
of Phosphoribulokinase dialysis, was digested overnight with L-1-tosylamido-2-phenylethyl chloromethyl ketone-treated trypsin equivalent to 2% of the kinase weight. Digestion was conducted a t 37 “C and terminated by freezing (5). HPLC of Peptides-Reverse phase HPLC (Laboratory Data Control, Riviera Beach, FL) was used as both an analytical and preparative procedure. Peptide mixtures (1-15 nmol) were injected onto a 0.46 X 25-cm Lichrosorb 5RP8 column (HPLC Technology, Palos Verdes Estates, CA), whichwas eluted with a 15 to 75% (v/v) acetonitrile gradient in 0.1% (v/v) trifluoroacetic acid (13). The total gradient was developedin 25 min at a flow rate of 1ml/min. Effluent was monitored by absorbance (215 nm) and by radioactivity. Peptide Purification-The cysteinyl residues of phosphoribulokinase (9 mg) were carboxymethylated with [’4C]iodoaceticacid in the presence of DTT and guanidine as described above.As expected based on prior amino acid analyses (8), four labeled peptides (designated I, 11, 111, and IV) were revealed by HPLC of a tryptic digest (Fig. 1B).The peptides were partially purified by gel filtration of the total digest on a 1.5 X 230-cm Sephadex G-25 column equilibrated with 4 M urea, 50 mM ammonium bicarbonate (pH 7.8), 2 mM 2mercaptoethanol (Fig. 2). Fractions were pooled based on radioactivity and qualitative analysis on the HPLC. Peptides I1 and IV were both located in the second radioactive peak (Fig. 2), while peptide I was associated with the third radioactive peak (Fig. 2). The pools containing these three peptides were dialyzed against 100 mM acetic acid and lyophilized. The dry residues wereredissolved in 50 mM ammonium bicarbonate and preparatively fractionated by HPLC, thereby yielding peptides I, 11, and IV in pure form. The first radioactive peak from gel filtration (Fig. 2), which contained peptide 111, was applied directly to a 0.5 X 36-cm column of DEAE-cellulose equilibrated with 4 M urea, 50 mM ammonium bicarbonate (adjusted to pH 7.0 with acetic acid), 2 mM 2-mercaptoethanol. After initial washing of the column with 10 ml of equilibration buffer, the peptides were eluted with a 50 to 400 mM gradient of ammonium bicarbonate (pH 7.0) in 4 M urea, 2 mM 2-mercaptoethanol (Fig. 3). Based on radioactivity and analytical HPLC analyses, fractions from the two major peaks (ZZIA and ZIZ B ) were pooled, dialyzed exhaustively (in Spectrapor dialysis tubing with a molecular weight cutoff of 1000, Spectrum Medical, Los Angeles,CA) against 100 mM acetic acid, and lyophilized. After dissolution of the residual samples in 6 M urea, 50 mM ammonium bicarbonate (pH 7.8), both variants of peptide 111 (see “Results”) were purified to apparenthomogeneity by preparative HPLC. RESULTS
Inactivation of Phosphoribulokinase by DTNB-As previously reported (6), phosphoribulokinase is very sensitive to I1
1“-
f 200
s’
lp
il
..e,.-.. ”.._. ,r.. I.-.
i,\ ...... A;,.
-...”...... .....”_? ~
~
I
A.
..,
IE
i
FIG. 1. Radioactivity profiles of tryptic digests of “C-carboxymethylated phosphoribulokinasefractionated by reverse phase HPLC. Conditions of carboxymethylation were as follows: A, fully reduced phosphoribulokinase in absence of denaturant; B, fully reduced phosphoribulokinase in presence of denaturant; C, oxidized phosphoribulokinase in absence of denaturant; D, fully reducedphosphoribulokinase in presence of 1mM ATP andabsence of denaturant; E, oxidized phosphoribulokinase in presence of denaturant; F, phosphoribulokinase, previously alkylated with BAEP, in the presence of denaturant. Additional details areprovided in the text.
125
Regulatory Thioredoxin Siteof Phosphoribulokinase
2
4
8
6
DTNB ADDED
($4)
FIG. 4. Stoichiometry of the reaction of phosphoribulokinase (8 PM) with DTNB. 0
, 100
120
140
160
mL
FIG. 2. Gel filtration of tryptic digest of 14C-carboxymethylated phosphoribulokinase. The column was equilibrated and eluted with 4 M urea, 50 mM ammonium bicarbonate, 2 mM 2mercaptoethanol (pH 7.8). Additional details are providedin the text.
50
\
J
.
9
t
z 8
n w
5
>
90
0 V W
a >
t
2 Iu
U
mL
FIG. 3. Ion exchange chromatography on DEAE-cellulose of peptide 111 (the first peak from the gel filtration profile shown in Fig.2). The column was preparedand eluted as described under “Experimental Procedures.”
99
5
DTNB.Titration of phosphoribulokinase with DTNB revealed a direct correlation between reagent conncentration and degree of inactivation; total loss of kinase activity was observed at a DTNB concentrationapproximating that of the phosphoribulokinase subunit (Fig. 4). For each molar equivalent of DTNB added, 2 mol eq of TNB were released during the course of inactivation. This ratio was consistent with the loss of kinase activity concomitant with the formation of a single disulfide bond per subunit. Spectroscopic examination of the inactivated enzyme, subjected to dialysis to remove TNB, confirmed that the reagent moiety was not incorporated. The inactivation of phosphoribulokinase by DTNB was fully reversed by treatment with DTT. Partially purified spinach thioredoxin-f substantially enhanced the DTT-dependent rate of activation (Fig. 5).
10 15 MINUTES
20
.,
FIG. 5. Reactivation of DTNB-oxidized phosphoribulokinase (1.6 QM) by DTT and thioredoxin. Activation conditions are described in the text. Activators were as follows: A, 20 p~ thioredoxin; 0, 5 mM D m , A, 5 mM DTT + 4.5 p~ thioredoxin; 0, 5 mM D m + 11 p~ thioredoxin; 40 mM DTT.
To determine whether the disulfide formed by the action of DTNB was intrasubunit or intersubunit, oxidized and reduced phosphoribulokinase were denatured and fractionatedon U1trogel AcA-34 in the presence of 6 M urea (see “Experimental Procedures”). Both samples coeluted as a single, symmetrical peak (datanot shown). Inclusion of 10 mM DTT in the column equilibration buffer did not alter the elution position of the reduced enzyme (data not shown). These results showed that the oxidized enzyme behaved as a monomer, thereby illus-
126
Regulatory Thioredoxin Site of Phosphoribulokinase
trating that the oxidation process entailed the formation of an intrasubunit disulfide rather than an intersubunit or intermolecular linkage. Carboxymethylation of Fully Reduced Phosphoribulokinase-Under denaturing conditions in the presenceof DTT, alkylation of phosphoribulokinase with ["C]iodoacetic acid revealed four sites of reaction (Fig. lB), consistent with the reportednumber of cysteinylresidues (8). Based on acid hydrolysis and amino acid analyses, all of the radioactivity was associated with carboxymethylcysteine (data not shown). Peptide I1 was absent from digests of phosphoribulokinase that was alkylated with BAEP prior to carboxymethylation (Fig. 1F ). This observation identified the cysteinyl residue of peptide I1 as Cys-16, previously assigned to the active site by selective modification with BAEP (5, 6) and 5'-p-fluorosulfonylbenzoyladenosine (7). Carboxymethylation under nondenaturing conditions revealed that the cysteinyl residues of peptides I and I11 were essentially inaccessible to iodoacetic acid (Fig. l.4) and could thus be classified as buried. Under these same conditions, ATP protected the cysteinyl residue of peptide I1 (Cys-16) against carboxymethylation (Fig. lo). Identification of the Regulatory Sulfhydryl Groups of Phosphoribulokinase-Two approaches were used toidentify those cysteines of phosphoribulokinase that are involved in redox regulation. One approach entailed a comparison of carboxymethylation patterns, under denaturing conditions, of fully reduced enzyme (Fig. 1B) andenzyme that had been oxidized with DTNB (Fig. 1E). In the latter case, peptides I1 and I11 were absent from tryptic digests, thereby implicating their respective cysteinyl residues as sites of oxidation. The only cysteinyl residue of the oxidized enzyme that was accessible to iodoacetic acid under nondenaturing conditions was that of peptide IV (Fig. IC). A secondapproach to identifying the oxidant-sensitive cysteines of phosphoribulokinase was based on the relatively large size of peptide 111, which resulted in its elution at the solvent front during gel filtration (the first radioactive peak illustrated in Fig. 2) and resolution from the other labeled peptides. If the deduced pairing of peptides I1 and I11 in the oxidized enzyme was correct, the disulfide-linked fragment (composed of peptides I1 and 111) in tryptic digests should emerge at the front duringgel filtration with the net effect of moving peptide I1 forward inthechromatogram.Thus, DTNB-oxidized phosphoribulokinasewas carboxymethylated with unlabeled iodoacetic acid in the presence of guanidine (without added DTT), and the subsequent trypticdigest was fractionated by gel filtration on Sephadex G-25 (in the absence of DTT).Thebreakthrough peak (basedonnm) was pooled, dialyzed against 100 mM acetic acid, and lyophilized. The peptides were redissolved ih 4 M guanidine, 50 mM sodium phosphate (pH 8.0), 1.5 mM EDTA, 5 mM DTT (to reduce the putativedisulfide) and thenwere treated with["C] iodoacetic acid. Afterdialysis against50 mM ammonium bicarbonate (pH 8.0), the peptides were subjected to reverse phase HPLC. The radioactivity profile showedthe appearance of peaks at the previously established positions for peptides I1 and I11 (Fig. 6B). As shown in Fig. 6A, the corresponding breakthrough peak fromgel filtration of reduced, 14C-carboxymethylated kinase revealed peptide I11 as the only radioactive component. Purification and Characterization of the Carboxymethylcysteinyl Peptides-Theonly notable problem encountered in the purification of these peptides reflected the limited solubility of peptide 111. In the absence of urea, the recoveries of peptide I11 from gel filtration and ion exchange column chromatography were only 36 and lo%, respectively. Inclusion of
W
I-
2
20-
I
cc W a v)
I-
,......r..12.r......*............................-' 600-
B
i\'I
L,*..*...-
m
z
3
0
u
200-
.-.."....../%"E. .... 21 5
18(0
....
...."..
b..-
15 MINUTES
24
FIG. 6. Radioactivity profiles from reverse phase HPLC of the breakthrough peaks from Sephadex G-25 chromatography of tryptic digests of phosphoribulokinase (see Fig. 2). A, phosphoribulokinase was reduced and denatured during carboxymethylation. B, phosphoribulokinase was oxidized with DTNB before denaturation and carboxymethylation with ['2C]iodoaceticacid. The breakthrough peak from gel filtration of the digest was then reduced, carboxymethylatedwith ['4C]iodoaceticacid, and subjectedto HPLC. Additional details are provided in the text.
4 M urea in thebuffers used during thepurification of peptide I11 increasedthe recoveries to amoreacceptable 70-80% range. Peptide I11 consistently appeared as a broad peak or poorly resolved doublet during reverse phase HPLC(Fig. 1B). During ion exchange chromatography of peptide I11 on DEAE-cellulose, two majorpeaks and two minor peaks were resolved (Fig. 3). Analysis of these peaks by HPLC showed that all four eluted innearly the same positions. The twomajor peaks were purified further by HPLC andsequenced. The sequence of the peptide (111 A) derived from the first major peak from the DEAE column shown is below. The peptide (I11B) isolated from the second peak differedonly in lacking an arginyl residue at the amino terminus, thereby revealing adjacent sites in phosphoribulokinase for trypsin cleavage and explaining the origin of the two closely related peptides that included the same cysteinyl residue. Automated Edman degradation indicated that all of the labeled peptides were at least 90% pure and provided the following sequences (see Miniprint). I. Lys-Leu-Thr-Cys-Ser-Tyr-Pro-Gly-Ile-Lys 11. Ser-Gln-Gln-Gln-Thr-Ile-Val-Ile-Gly-Leu-A~a-A~a-AspSer-Gly-Cys-Gly-Lys
I11 A. Arg-Leu-Thr-Ser-Val-Phe-Gly-Gly-Ala-Ala-Glu-Pro-ProLys-Gly-Gly-Asn-Pro-Asp-Ser-Asn-Thr-Leu-Ile-SSer-AspThr-Thr-Thr-Val-Ile-Cys-Leu-Asp-Asp-Phe-His-Ser-LeuAsp-Arg IV. Phe-Phe-Asn-Pro-Val-Tyr-Leu-Phe-Asp-Glu-Gly-Ser-Thr.
Ile-Ser-Trp-Ile-Pro-Cys-Gly-Arg
In each case, the observed sequence was in reasonable agreement with the amino acid composition (see Miniprint). As expected, based on the results of successive alkylation of phosphoribulokinase by BAEP andiodoacetic acid,peptide I1 (which represents one of the regulatory sulfhydryls) was the NH2-terminal fragment that included the active-site cysteinyl residue at position 16 ( 5 , 6). Edman degradation of intact phosphoribulokinase(see Miniprint) provided sequence data that spans peptides I1 and 111, thereby placing the second regulatory sulfhydryl at position 55.
Regulatory Thioredoxin Siteof Phosphoribulokinase
127
10 5 dized (Fig. 1E) and reduced(Fig. 1B) phosphoribulokinase H~N-Ser-Gln-Gln-Gln-Thr-Ile-Val-Ile-Gly-Leu-Ala-Ala-Asp-Serclearly indicates that peptides I1 (bearing Cys-16) and 111 15 20 25 (bearing Cys-55) are paired to form the regulatory disulfide. Gly-Cys-Gly-Lys-Ser-Thr-Phe-Met-Arg-Arg-Leu-Thr-Ser-ValConfirmation is provided by partial purification of the disul30 35 40 Phe-Gly-Gly-Ala-Ala-Glu-Pro-Pro-Lys-Gly-Gly-Asn-Pro-Aspfide peptide (fromoxidized phosphoribulokinase), followed by 45 50 55 I1 and 111 Ser-Asn-Thr-Leu-Ile-Ser-Asp-Thr-Thr-Thr-Val-Ile-Cys-Leu-its reduction and alkylation to regenerate peptides (Fig. 6). These experiments represent the most direct docu60 Asp-Asp-Phe-His-Ser-Leu-Asp-Arg mentation to date that the deactivated form of a thioredoxinregulated enzyme is a disulfide. The alternative possibilities Thus, the oxidation of phosphoribulokinase by DTNB genof sulfhydryloxidation to the level of sulfenic or sulfinic acids erated a disulfide between Cys-16 and Cys-55. The positional can thus be discounted,a t least for phosphoribulokinase. assignments for the cysteinyl residues of peptides I and IV The identification of both regulatory sulfhydryls of phoswere not establisheddue to lack of additional sequence inforphoribulokinase and the determinationof the encompassing mation. primary structure are the first such datafor a thioredoxin-fregulated enzymeand will provide a basis for comparison with DISCUSSION analogous structural features of similarly regulated proteins Three major conclusions can be supportedby these exper- as they become available. Limited sequence data from chloiments: (i) oxidative inactivation of phosphoribulokinase re- roplast fructose-1,6-bisphosphatasehas appeared, which acsults from the formation of a single, intrasubunit disulfide counts for only 2 of the enzyme’s 6 cysteinyl residues, but bond; (ii) the regulatory disulfide of phosphoribulokinase is homology with phosphoribulokinase is not evident (19). Parformed between Cys-16 and Cys-55; and (iii) Cys-16, which tial characterization, without sequence information, of the was previously identified as anessential active-siteresidue of regulatory site of NADP-dependentmalate dehydrogenase (which is activated by thioredoxin-m) has been described in phosphoribulokinase (5, 7), isalso involved inthethiola preliminary report (ZO), in which Cys-10 and Cys-15 are dependent regulation of enzyme activity. Inactivation of phosphoribulokinase with the potential in identified as the relevant groups. Hence, the 38-amino acid vivo oxidant, dehydroascorbate, had implicated Cys-16 as one loop of phosphoribulokinaseformed by the disulfide between of proteins that of the regulatory cysteines (5).However, large molar excesses Cys-16 and Cys-55 is not an invariant feature of dehydroascorbate were required to achieve complete inac- are regulated by thioredoxin. Among all protein disulfides of tivation,andpreliminarycharacterization of the oxidized known connectivity, 49% involve half-cystines separated by protein indicated a lack of specificity. In contrast, phospho- less than 24 residues, a value considerably greater than preribulokinase is totally inactivated by stoichiometric levels of dicted by probability. In contrast, pairing of cysteines that DTNB with the highly specificinvolvement of only 2 cysteinyl are separated by 35-40 residues (as in phosphoribulokinase) residues. Although clearly not a physiologial oxidant, DTNB accounts for only 7% of observeddisulfides,in agreement probably mimics the in uiuo oxidation process with respect to with random distribution (21). Since the a-caTbons of Cys-16 and Cys-55 must approach product formation, as thioredoxin (in the presence of DTT to maintain the reduced form) reactivates the oxidized enzyme within -4.5-7.5 A of each other in order to form a disulfide former is exposed and the latter (Fig. 5). Furthermore, there seems little reason topresuppose (21), it is surprising that the that the cysteinyl residues most susceptible to oxidation in buried (as determined by accessibility to iodoacetate). This observation suggests that formation of the mixed disulfide vitro are not the same as those susceptible in uiuo. Collectively, datafromthreedisparateexperimentsare between phosphoribulokinase and thioredoxin, the putative consistent with the conclusion that oxidative inactivation of intermediate in the regulatory reaction, involves Cys-16 and phosphoribulokinase reflects the formation of a single intra- induces a conformational change in phosphoribulokinase. In subunit disulfide. First, the stoichiometry of the reaction is the direction of reduction (activation), this conformational that required for disulfidebond formation,i.e. 1mol of DTNB change couldplaceCys-55 in amoreburied environment; is consumed permole of subunit inactivated with concomitant whereas inthedirection of oxidation(deactivation),this generation of 2 mol of T N B (Fig. 4), and TNB is not incor- residue could become more exposed. porated into the protein. Second, after oxidation of phosphoBased on previous chemical modification studies, Cys-16 ribulokinaseby DTNB, only2 of the 4cysteinylresidues was assigned to the nucleotide-binding domain of the active feund in the reduced form of the enzyme can be carboxy- site (5, 7) and was also suggested as one of the regulatory methylated (Fig. l , B and E ) . Third, the oxidized enzyme residues (5). The latter postulate is confirmed by characterization of the disulfide peptide as reported herein, and the chromatographs as amonomer during gel filtrationunder dissociation conditions that maintain the integrity of disulfide former conclusion is reinforced by the recognition that the linkages. Hence, thedisulfide of oxidized phosphoribulokinase primary structure in the vicinity of Cys-16 is similar to a must be intrasubunit, a fact that could not be deduced from consensus sequence thathasbeen observed amongATPthe stoichiometry of oxidation or the pattern of carboxyme- binding proteinsof diverse function andorigin (22,23) (Table thylation. I). Although a similar sequence is observed in phosphoribuThe propensity of phosphoribulokinase to undergo intra- lokinase from Rhodobacter spheroides (24), this bacterial ensubunit oxidation is especially noteworthy because aggregate zyme lacks a cysteine at position 16 and, significantly, is not forms of catalyticallydeficientphosphoribulokinasehave regulatedby thiols (4, 25). As Cys-16 is not common to been detected by gel filtration of whole cell extracts or par- phosphoribulokinase from all sources, this residue is probably tially purified preparations (14-18). In retrospect, one might not essential t o catalytic activity. However, a facilitative role envision thatthese aggregates represent complexation of cannot be excluded, because the catalytic efficiency (k,,,/K,) phosphoribulokinase with other proteinsvia disulfides rather of the higher plant enzyme is approximately 100-fold greater than self-association. Our experiments do not offer any inthan that of the bacterial enzyme (8, 25, 26). Another notesightsconcerningthe origin of such complexes northeir worthy observation derived from the comparative sequence physiological significance, if they are indeed formed in uivo. information in Table I is that Lys-18 of spinach phosphoriComparison of the patterns of carboxymethylation of oxi- bulokinase corresponds toLys-21 of adenylate kinase, a well-
Regulatory Thioredoxin Site of Phosphoribulokinase
128
TABLE I Sequence similarities among phosphoribulokinuse and diverse ATP-binding proteins Sequences of phosphoribulokinase that are found in the other proteins illustrated are boxed. The abbreviations used are: PRK, phosphoribulokinase; Mal k, maltose transport protein; Pst B, phosphate transport protein; Opp D, oligopeptide transport protein; His P, histidine transport protein. Sequence Protein Source Position PRK Spinach 11-23 PRK. spheroides Rh. 11-23 Mal k b
E . coli
P a t Bb E . coli OPP Db E . coli H i s Pb imur S . typh Rabbit AdenylatekinaseC Myos i nc Rabbit Sequence data from Ref. 24. * Sequence data from Ref. 23. e Sequence data from Ref. 22.
15-27 41-53
i urn
53-65 38-50
14-26 179-1 90
11. Racker, E. (1957)Arch. Biochem. Biophys. 69,300-310 12. Ellman, G. L. (1959)Arch. Biochem. Biophys. 82, 70-77 13. Mahonev. W. C.. and Hermondson, M. A. (1980)J. Biol. Chem. 255, ii199-1i203 14. Wara-AswaDati. 0.. Kemble. R. J.. and Bradbeer. J. W. (1980) Plant Ph&ol.’ (Bethesda)66,34139 15. Ruffer-Turner, M. E., and Bradbeer, J. W. (1984)in Advances in Photosynthesis Research (Sybesma, C., ed) Vol. 111, pp. 597600, Martinus Nijhoff/Dr. W. Junk Publishers, The Hague, The Netherlands 16. Wolosiuk, R. A,, and Buchanan, B.B. (1978)Arch. Biochem. Biophys. 189,97-101 17. Surek, B., Heilbronn, A., Austen, A., and Latzko, E. (1985)Planta (Berl.)165, 507-512 18. Lazaro, J. J., Sutton, C. W., Nicholson, S., and Powls, R. (1986) REFERENCES Eur. J. Biochem. 156,423-429 19. Harrsch, P. B., Kim, Y., Fox, J. L., and Marcus, F. (1985) 1. Ziegler, D. M. (1985)Annu. Rev. Biochem. 54,305-329 Biochem. Biophys. Res. Commun. 133,520-526 2. Gilbert, H. F. (1984)Methods Enzymol. 107,330-351 20. DeCottignies, P., Schmitter, J., Miginiac-Maslow, M., Le3. Buchanan, B. B. (1980)Annu. Rev. Plant Physiol. 31,341-374 Marechal, P., Jacquot, J., and Gadal, P. (1987)Plant Physiol. Johnson, T. C., 4. Crawford, N.A., Sutton, C.W.,Yee,B.C., (Bethesda)83, S890 Carlson, D.C., and Buchanan, B.B. (1984)Arch. Microbiol. 21. Thornton, J. M. (1981)J. Mol. Biol. 151,261-287 139,124-129 5. Porter, M. A., and Hartman,F. C. (1986)Biochemistry 25,7314- 22. Walker, J. E., Saraste, M., Runswick, M. J., and Gay, N. J. (1982) EMBO J. 1,945-951 7318 6. Omnaas, J.,Porter, M.A., andHartman, F. C. (1985)Arch. 23. Higgins, C. F., Hiles, I. D., Salmond, G. P. C., Gill, D. R., Downie, J. A., Evans, I. J., Holland, I. B., Gray, L., Buckel, S. D., Bell, Biochem. Biophys. 236,646-653 A. W., and Hermondson, M. (1986)Nature 323, 448-450 7. Krieger, T. J., Mende-Mueller, L., and Miziorko, H. M. (1987) 24. Hallenbeck, P., and Kaplan, S. (1987)J. Bacteriol. 169, 3669Biochim. Biophys. Acta 915,112-119 3678 8. Porter, M. A., Milanez, S., Stringer, C. D., and Hartman, F. C. 25. Gibson, J. L., and Tabita, F. R. (1987)J. Bacteriol. 169, 3685(1986)Arch. Biochem. Biophys. 245,14-23 3690 9. Ashton, A. R. (1983)in Thioredoxins, Structure and Function, proceedings of the conference held June 21-24,1981,Berkeley, 26. Rippel, S.,and Bowien, B. (1984)Arch. Microbiol. 139,207-212 CA, pp. 245-250,Centre Nationalde la Recherche Scientifique, 27. Tagaya, M., Yagami, T., and Fukui, T. (1987)J. Biol. Chem. 262,82574261 Paris 28. Fry, D.C., Kuby, S. A., and Mildvan, A. S. (1986)Proc. Natl. 10. Schurmann, P., Maeda, K., and Tsugita, A. (1981) Eur. J. Acad. Sci. U. S. A . 83,907-911 Biochem. 116.37-45
documented active-site residue (27). Crystallographic modeling places Lys-21 near the @- or y-phosphatemoiety of bound ATP (28). Irrespective of the precise function of Cys-16 and Lys-18 of spinach phosphoribulokinase, the important conclusion to be drawn from our studies is that modulation of catalytic activity by thioredoxin entails a structural change at the active site. The inactivation that accompanies modification of Cys-16, by either proteinreagents or by intrasubunit disulfide bond formation, is readily rationalized by steric or conformational perturbation of the binding domain for nucleotide.
129
16 2
3 5 6 7
14aS
17
1440
18
1142
19
1161
21 22 23 2u 25 26 27 28 29 30
375 441 626
20
1082 884 755 890 888
9 10 11
12 13
863 642
14
15
5%
4i5 410 151 130 185
me See Val
1 20 1 s
Phe
1Ub
T a b l e "111:
C o m p o r i t i o n r of C a r h a i y m e f h y l c y 3 t s l n y lP e p t i d e s
mol
residues
mol
residues
Wr(om1 ASP
1.64
0.8
111
1.53
mr
1.78
0.9
S W
2.01
1.0
(11 (11
1.45 1.52.90 4.93
0.8 0.9 0.8
OI PRKs
111 1
A m
3385
2
Leu Thrc
21
7104
3
a
5 6 7 8 9 10 11 I2 13
14 15 16 17 18 20 19
Sere Val
Phe
CIY GIY Ala
A h
GI" Pro Pro Lln 011 011 A m
Pro Asp
sere
A""
22
mrc
1860
23
Leu
1268
Ile
1230
24
4293 6813 5512
Ul8U 5032 U891 3069 34007 3198 3464 3672 3592 3100 2572 1651
25 26 27 28 29 30
Sere AIP
mrc m e
me Val
5ao
457
1.72
GI" PF.3
1.94
G1Y
2.38
l.Ob111 1 . 2 (11
Ala Val
.."
5.07 3.20
2.6
2.7 I21 1.7
-01
(11 (11 (1)
0.80 5.52 3.10 ( 2 ) 3.62.85 (31 1.20.95 2.92.32 (31 4.13.25 12) 1.56 2.0
Il*
TYr
1.53 1.81 1.81
0.8 0.9 0.9
(11 11)
806 5?9 734 494 490 "05
457
1.6
(11
6.9 4.6
(7) (5) IUI
1.39 9.84 6.44 4.99 1.85 4.15 5.67 2.86
111
(31 (41
2.1
1.0 7.1 11.6 3.6 1.3 3.0
4.0
mol
iosiduel
1.UU
1.0
11)
2.1
(21
3.03 1.39 2.77 7.21.77 2 . 9 3121 (212.23.15
1.0 1.9
(11 (2)
(11 2.0
0.6
(11
1.U7
1.8
121
2.71
2.0
1.35
0.9
(1)
(21
1.34 3.16
1.7 4.0
(21
111
(41
2.46 5.67
1.8 4.1
2.55 1.35
1.8 0.9
(21 (11
3
1 n
,.> ,, \
0.8
(11
(1)
(21
I v
l.U 0.9
HiS
3.14
1.0
0s
residues
1.16
€ha LYl h P Ir8
mol
2.74 1.67
t."
Le"
rsaidurs
1.9
1.0:
4.292.01.68 2.79121 2.1 0.T9 1.0 (1) 0.80 1.0!
(11
(11
1.91
1.0
1.12
1.00
..>, 20
1.17
.." 3.0
(3)
