THE JOURNAL OF BIOLOGICAL CHEMISTRY Vol. 263, No. 10, Issue of April 5, pp. 4921-4925,1988 Printed in U.S.A.

Restoration of Activity to Catalytically Deficient Mutants of Ribulosebisphosphate Carboxylase/Oxygenase by Aminoethylation”

(Received for publication, November 10, 1987)

Harry B. Smith$ and Fred C. Hartman From the Protein Engineering and Molecular Mutagenesis Program of the Biology Division, Oak Ridge National Laboratory, and the University of Tennessee-Oak Ridge Graduate School of Biomedical Sciences, Oak Ridge, Tennessee 37831

Substitutions for active-site lysyl residues at posi- tions 166 and 329 in ribulosebisphosphate carboxyl- ase/oxygenase from Rhodospirillum rubrum have been shown to abolish catalytic activity. Treatment of the Cys-166 and Cys-329 mutant proteins with 2-bro- moethylamine partially restores enzyme activity, pre- sumably as a consequence of selective aminoethylation of the thiol group unique to each protein. Amino acid analyses, slow inactivation of the wild-type carboxyl- ase by bromoethylamine, and the failure of bromoeth- ylamine to restore activity to the corresponding glycyl mutant proteins support this interpretation. The ob- served facile, selective aminoethylations may reflect an active site microenvironment not dissimilar to that of the native enzyme. Catalytic constants of these novel carboxylases, which contain a sulfur atom in place of a specific lysyl ymethylene group, are significantly lower than that of the wild-type enzyme. Furthermore, the aminoethylated mutant proteins form isolable com- plexes with a transition state analogue, but with com- promised stabilities. These detrimental effects by such a modest structural change underscore the stringent requirement for lysyl side chains at positions 166 and 329. In contrast, the aminoethylated mutant proteins exhibit carboxylaseloxygenase activity ratios and K, values that are unperturbed relative to those for the native enzyme.

Site-directed mutagenesis enables systematic alteration of protein structure with absolute specificity for any preselected site, thereby overcoming two major weaknesses of traditional chemical modification. Nevertheless, interpretation of func- tional consequences of mutations requires distinguishing di- rect from indirect effects, as so vividly documented by insight- ful studies of tyrosyl-tRNA synthetase (1, 2) and staphylo- coccal nuclease (3). Ideally, the three-dimensional structures of both wild-type and mutant proteins will have been deter- mined, but, even then, calculations of quantitative structure- activity relationships may be suspect unless the mutant pro- teins can be demonstrated to conform to certain free energy equations (2).

In the absence of crystallographic data, our laboratory initiated site-directed mutagenesis studies (4-6) of ribulose-

* This research was sponsored jointly 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.

$ Predoctoral student supported by training Grant GM07438 from National Institutes of Health.

bisphosphate carboxylase (EC 4.1.1.39) from Rhodospirillum rzhrum, a homodimer of 50.5-kDa subunits (7). Our goals were to validate the earlier assignments, based on chemical procedures (8), of 2 different lysyl residues to the active site and to ascertain qualitatively the function of these residues. Subsequently, a 3-A resolution structure of the R. rubrum carboxylase was determined (9). However, this structure has not yet assisted designing and interpreting site-directed mut- agenesis experiments, because fitting of amino acid side chains to electron densities was incomplete and only the deactivated form of the enzyme, which may differ conforma- tionally from the catalytically active form, was available. Although our mutagenesis data were compatible with direct obligatory involvement of these lysines in catalysis, the pos- sibility that the amino acid replacements altered the active site conformation and thereby rendered the mutant proteins inactive could not be totally discounted.

Indications of catalytic functionality of the two active-site lysines (Lys-166 and Lys-329 in the enzyme from R. rubrum) included their strong acidities and enhanced nucleophilicities (lo), presumably reflective of an unique microenvironment. If substitutions for Lys-166 and Lys-329 did not alter the active-site conformation, other reactive side chains at these positions should also exhibit unusual properties. The cysteinyl mutant carboxylases, devoid of enzyme activity, provide a direct chemical approach to inspecting the conformational integrity of the active site. Selective aminoethylation of the relevant thiol group, dependent upon its enhanced reactivity, would generate a “lysyl-like” side chain at the active site and consequently could restore enzyme activity.’ Conversion of cysteinyl residues to aminoethylcysteinyl residues has long been recognized as a means of creating trypsin-susceptible sites in proteins to facilitate sequence studies (11, 12).

Another impetus for investigating the aminoethylation of the cysteinyl mutant proteins concerns the feasibility of al- tering the carboxylase/oxygenase activity ratio of the wild- type enzyme. Ribulose-P, carboxylase’ catalyzes an oxidative degradation of ribulose-P2, which competes directly with car- boxylation and thereby limits crop yields (13, 14). The two activities are intimately associated in that they entail a com- mon reaction intermediate, the 2,3-enediol of ribulose-P, (15). The ratio of the two activities has never been altered signifi- cantly by any chemical modification or genetic manipulation. The aminoethylations reported herein offer the opportunity

As reported by Chem. Eng. News (1986) in the July 14 issue, pp. 7-14, Dr. John H. Richards at the California Institute of Technology has observed restoration of enzyme activity to the Lys-73 + Cys mutant of 8-lactamase upon arninoethylation.

The abbreviations used are: ribulose-P, carboxylase, D-ribulose 1,5-bisphosphate carboxylase/oxygenase; carboxyarabinitol-P,, 2- carboxyarabinitol 1.5-bisphosphate; carboxyribitol-P2, 2-carboxyri- bitol 1,5-bisphosphate; Bicine, N,N’-bis(2-hydroxyethyl)glycine.

4921

4922 Aminoethylation of Mutant R&dose-P. Carboxylase

to ascertain the effects on enzymatic properties of a subtle structural change at the active site, net replacement of a lysyl y-methylene of the wild-type enzyme by a sulfur atom, that could not be introduced by either chemical modification or site-directed mutagenesis alone.

A preliminary account of this report has appeared (16).

EXPERIMENTAL PROCEDURES

Materi4lls

Z-Aminoethylcysteine (a standard for amino acid analyses), 2- bromoethylamine hydrobromide, phenylmethanesulfonyl fluoride, and leupeptin were purchased from Sigma. Pancreatic DNase I was obtained from Worthington. Buffers and ninhydrin for amino acid analyses were purchased from Beckman. NaH”C03 (60 mCi/mmol) was a product of Amersham Corp. Ribulose-Pz, carboxyarabinitol-P2, and carboxyribitol-Pz were prepared following published procedures (17, 18). Other chemicals and reagents were procured at the highest level of purity readily available. Ribulose-Pz carboxylase from R. rubrum was purified as described earlier (19) and stored at -70 “C in a pH 8.0 buffer consisting of 50 mM Bicine, 66 mM NaHCO,, 10 mM MgC12, 1 mM EDTA, henceforth designated activation buffer, to which glycerol (20%, v/v) and L-mercaptoethanol (10 mM) had been added.

Methods

Purification of Mutant Proteins-The construction of the plasmids for the mutant proteins (Cys-166, Gly-166, Cys-329, Gly-329) used in the present studies and their expression in Escherichia coli JM107 have been described (5, 6). Cell paste (15 g) was suspended in 2 volumes of activation buffer (see “Materials”) containing 5 mM di- thiothreitol, 2 mM phenylmethanesulfonyl fluoride, 20 pM leupeptin, and 1 pg/ml of pancreatic DNase I. The cell suspension was passed three times through a French pressure cell (Amicon) at 4 “C, after which the cellular debris was removed by centrifugation at 100,000 X g for 1 h. The mutant carboxylase in the 31-ml supernatant (-700 mg of total protein) was isolated by the procedure used to purify the R. rubrum carhoxylase (19), except smaller columns were used. The first DEAE-cellulose column was 2.5 X 17 cm rather than 2.5 X 55 cm, and the second DEAE-cellulose column was 1.5 x 18 cm rather than 2.5 X 11.5 cm. Since the mutant carboxylases are devoid of enzyme activity, fractions were screened by polyacrylamide gel elec- trophoresis in the presence of sodium dodecyl sulfate (20). Typically, mutant protein (IO-15 mg) was obtained at 85-95% purity as judged by gel densitometry; it was stored at -70 “C in activation buffer, 20% (v/v) glycerol, 10 mM 2-mercaptoethanol.

Amirwethylution of Mutant Proteins-Thirty min prior to treat- ment, the protein stock was diluted with activation buffer at 25 ‘C to 0.4 mg/ml (if only kinetics of modification were to be examined) or 7 mg/ml (if the modified protein was to be characterized). An aliquot of freshly prepared 2 M 2-bromoethylamine (pH 8.0) was then added to the protein solution to give a final reagent concentration of 100 mM. Periodically, aliquots of the reaction mixture were assayed for carboxylase activity by the spectrophotometric method (19). The direct “CO&ixation assay (see below) could not be used to monitor the reaction because of interference by bromoethylamine. After max- imal restoration of activity had been attained, the bulk of the reaction mixture was dialyzed against activation buffer, 10 mM 2-mercapto- ethanol for subsequent investigation of the properties of the amino- ethylated protein. A small portion of the reaction mixture was di- alyzed against 10 mM NaCl for the purpose of amino acid analyses.

Protein and Enzyme Assays-The dye-binding method of Bradford (21), with the required reagent obtained from Bio-Rad, was used to determine protein concentration; the R. rubrum carboxylase served as the standard.

After thorough dialysis of the aminoethylated mutant proteins to remove bromoethylamine and by-products, they were examined for both carboxylase and oxygenase activities. Carboxylase activity was determined by the modified (4) “C02-fixation procedure (22). Assay mixtures (final volume = 250 pcl; pH = 8.0) contained 24 mM NaH”C03 (16 mCi/mmol), 32 mM Bicine, 6.5 mM MgCl,, 0.05 mM EDTA, 0.4 mM ribulose-P,, 250 pg of bovine serum albumin, and 0.4- 35 wg of the aminoethylated mutant protein. Incorporation of “CO2 into 3-phosphoglycerate was reliably monitored during 0.5-6 min. Carboxylase activity of the unmodified mutant proteins was deter- mined with increased levels of protein (65 rg) and increased reaction

times (lo-60 min). Oxygenase activity was determined with an oxygen electrode (Hansatech, Norfolk, Great Britain) as described by Lori- mer et al. (22). Assay mixtures (final volume = 250 ~1; pH = 8.0) contained 32 mM Bicine, 6.5 mM MgCl,, 0.05 mM EDTA, 0.4 mM ribulose-PZ, and 50-165 c(g of the aminoethylated mutant protein.

Compiexation of the Aminoethylated Mutant Proteins with Carbox- yarabinitol-Pp-To a solution (3.5 ml) of the protein (0.4 mg/ml) in activation buffer was added an b-fold molar excess of [‘“Clcarboxyar- abinitol-P2 (6.5 x 10’ cpm/pmol). Forty-five min after mixing, an aliquot (0.5 ml) of the solution was subjected to gel filtration on a 1 X 51-cm column of Sephadex G-50 (fine). Fractions were monitored for protein (AWnm ) and for radioactivity so that the stoichiometry of carboxyarabinitol-Pz binding could be calculated. This value estab- lished maximal binding at time “zero.” Unlabeled carboxyarabinitol- PZ (20 times the amount of ‘“C-labeled material) was then added to the remaining original protein solution (3.0 ml). The rate of exchange of labeled for unlabeled ligand was then determined by periodically subjecting 0.5-ml aliquots to gel filtration as just described.

IsoeZectric Focusing-Gels and solutions for isoelectric focusing were prepared according to the method of O’Farrell (12), excepting omission of Nonidet P-40. Focusing on vertical slabs of 7% polyacryl- amide (9.5 X 6.5 x 0.05 cm) required 2100 V-h.

Amino Acid Analyses-Proteins were hydrolyzed at 100 “C for 21 h in evacuated sealed tubes with 6 N HCl, 0.01 M 2-mercaptoethanol. Hydrolysates were dried on a Speed Vat Concentrator (Savant) and subjected to chromatography on a Beckman 121 M amino acid ana- lyzer using Beckman’s “3-hour-single-column system,” consisting of isocratic elution by three different buffers (pH 3.25, 4.25, and 6.4, respectively) in succession. To achieve complete separation of S- aminoethylcysteine from lysine, elution of the column with the second buffer (pH 4.25) was omitted entirely. The analyzer was hard wired to a PDP data acquisition system. Computer programs used in data collection and analysis were provided by S. S. Stevens and J. T. Holderman of the Oak Ridge National Laboratory.

RESULTS

Reactions of Mutant Proteins with 2-Bromoethylamine- Time-dependent restorations of carboxylase activity to both the Cys-166 and Cys-329 mutant proteins upon incubation with 2-bromoethylamine are shown in Fig. 1. The observed plateaus did not reflect the true upper limits of activities, because analogous treatment of wild-type enzyme resulted in -30% inactivation during the 19-h reaction period. The res- torations were absolutely dependent on the presence of the alkylating agent and required the thiol group introduced by mutagenesis, because identical treatment of either the Gly-

u k! m .

0.2

7cavfla 200 400 600 800 1000 1200

TIME (min)

FIG. 1. Carboxylase activity during treatment of wild-type ribulose-P, carboxylase (m, the Cys-329 mutant (A), the Cys- 166 mutant (A), the Gly-329 mutant (0), and the Gly-166 mutant (0) with 2-bromoethylamine. Untreated wild-type en- zyme (Cl) served as the control. See “Methods” for experimental details.

Aminoethylation of Mutant Ribulose-P2 Carboxylase 4923

166 or Gly-329 mutant protein did not lead to any detectable activity.

Extents of Aminoethylations-Isoelectric focusing under de- naturing conditions provided direct visualization of the pro- gression of the reactions (Fig. 2). As anticipated on the basis of its partial inactivation, the treated wild-type carboxylase showed a product with a higher PI than that of the untreated sample (lanes 3 and 4 in both panels). This alkylation product represented -35% of the total protein as determined by densitometric scanning of the gels, in good agreement with the 30% loss of activity. The increase in PI is attributed to conversion of a neutral (or negative) thiol to a basic amino- ethylthioether.

Consistent with Lys + Cys substitutions, both of the un- treated cysteinyl mutants have lower PI values than that of wild-type enzyme (lanes 1 and 4 in both panels). After alkyl- ation, the preparations of both the Cys-329 (Fig. 2A, lane 2) and the Cys-166 (Fig. 2B, lane 2) mutant proteins contain two

I 2 3 4 5 6 7

1 2 3 4 5 6 7

FIG. 2. Isoelectric focusing (denaturing conditions) of ami- noethylated Cys-329 (panel A ) and Cys-166 (panell?) mutant proteins. In each panel, lane 1 is the untreated mutant protein; lanes 2 and 3 are treated mutant and wild-type proteins, respectively; lane 4 is the untreated wild-type enzyme; lanes 5 6 , and 7 are, respectively, mixtures of untreated mutant with treated native proteins, untreated mutant with treated mutant proteins, and treated mutant with treated native proteins.

new species with higher PI values. One of these coincides with wild-type enzyme and presumably accounts for the partial restoration of carboxylase activity as a consequence of ami- noethylation of the thiol group unique to each mutant protein. The other new species, presumably catalytically inactive, co- incides with the reaction product formed by alkylation of wild-type enzyme and no doubt reflects two sites of modifi- cation.

If the specific activities of the treated mutant proteins shown in Fig. 1 are adjusted based on the fractional concen- trations of those species that exhibit the wild-type PI (0.2 for the Cys-166 mutant and 0.4 for the Cys-329 mutant), the specific activity of the Cys-166 mutant becomes 1.2 units/mg and that of the Cys-329 mutant becomes 3.6 units/mg com- pared to 6.0 units/mg for wild-type enzyme.

Direct proof that aminoethylation occurred a t cysteinyl residues was provided by amino acid analyses. The treated wild-type enzyme contained 0.3-0.7 residue of aminoethyl- cysteine/molecule, whereas the analogously treated mutant proteins contained 0.7-1.2 residues of aminoethylcysteine/ molecule. Other alkylated amino acids were not observed.

Enzymatic Characteristics of the Aminoethylated Mutant Proteins-Oxygenase activity was restored to the cysteinyl mutant proteins during aminoethylation to the same relative extent as the carboxylase activity; hence, evidence of altered carboxylase/oxygenase ratios with either of these novel en- zymes is lacking. K, and Ki values for the enzymically restored proteins did not appear to differ significantly from those of wild-type enzyme determined in side-by-side experiments: K, (ribulose-P2) = 6-13 p ~ ; K, (CO,) = 350-500 p ~ ; Ki (carbox- yribitol-P2) = 2-3 p ~ . In contrast, the kcat values (corrected as above) for the aminoethylated Cys-166 and Cys-329 mutant proteins were only 20 and 60% of the kat for wild-type enzyme, respectively.

Compkxation of Aminoethylated Mutant Proteins with Car- boxyarabinitol-P2-The activated form (a Mg2”stabilized car- bamate) of ribulose-P2 carboxylase tenaciously binds the tran- sition state analogue carboxyarabinitol-P2 to form a stable quaternary complex, which can be isolated by gel filtration (24). Although the Cys-166 and Cys-329 mutant proteins did not form a stable complex (5, 6), this binding functionality was exhibited by the aminoethylated samples (Fig. 3, data shown only for the treated Cys-329 protein). The binding

VOLUME (mL)

FIG. 3. Complexation of untreated Cys-329 (A), aminoeth- ylated Cys-329 (O), and wild-type (0) proteins with the tran- sition state analogue [“C]carboxyarabinitol-P2. Each protein was incubated with excess radioactive ligand and then subjected to gel filtration as described under “Methods.” The first peak of radio- activity coincided with protein; the excess ligand appeated in the salt region.

4924 Aminoethylation of Mutant Ribulose-P2 Carboxylase - 1.0-. L A

0 12 24 36 48 TIME ( h )

FIG. 4. Exchange of [‘4C]carboxyarabinitol-P2 from qua- ternary complexes of aminoethylated Cys-329 (.) and wild- type (0) proteins. The binding data for the mutant protein are uncorrected for incomplete aminoethylation of the preparation used. See “Methods” for experimental details.

stoichiometry for both proteins, corrected for incomplete ami- noethylation (see above), was 1 mol of [14C]carboxyarabinitol- P2/mol of subunit. Despite formation of the complex, its stability was considerably less than that of the corresponding complex prepared with wild-type enzyme. The exchange rate of free unlabeled carboxyarabinitol-P2 with the protein-bound labeled ligand was characterized by a half-time of 4.5 h with the aminoethylated Cys-329 mutant protein compared to 21 h with the wild-type enzyme (Fig. 4). A similarly shortened half-time was observed with the aminoethylated Cys-166 mu- tant protein (data not shown).

DISCUSSION

Although we have not identified directly the site(s) of aminoethylation of the mutant proteins at the peptide level, considerable indirect evidence implicates the cysteinyl residue which replaced Lys-166 or Lys-329 of the native enzyme. Prima facie, restoration of enzyme activity to two different cysteinyl mutants is most logically explained by chemical “reversal” of the mutations. Consistent with this envisioned chemistry, the change in PI of the native carboxylase due to removal of either active-site lysyl side chain by mutagenesis is counterbalanced precisely by subsequent aminoethylation. If the restoration of function were due to modification at some other site, bromoethylamine should stimulate the native en- zyme and activate other catalytically deficient mutant car- boxylases; neither situation prevails. Lastly, amino acid anal- yses confirm alkylation of cysteinyl residues and demonstrate a greater extent of modification of the cysteinyl mutant proteins than of the native enzyme.

Isoelectric focusing also shows that the alkylations by bro- moethylamine are highly selective; among the 5 cysteinyl residues/subunit of the native enzyme (7), only one appears to undergo modification. With each of the cysteinyl mutant proteins, the newly introduced thiol provides a second site for modification. We believe that the preferential reactivities of Cys-166 and Cys-329 toward the weakly electrophilic bro- moethylamine are reflective of the unusual microenvironment at the active site that endows Lys-166 and Lys-329 of the native enzyme with hyperreactivity (10). If this conclusion is valid, it reinforces previous arguments that the deleterious catalytic consequences of single amino acid replacements do not reflect conformational changes. Otherwise, any confor- mational perturbations that had occurred are reversed by the aminoethylations. Our earlier assertions (5, 6) that Lys-166

and Lys-329 are involved directly in catalysis (Lys-166 as the base that enolizes ribulose-P2) are likewise strengthened by the findings that the functionally restored carboxylases ex- hibit reduced catalytic constants but normal Michaelis and inhibition constants compared to the native enzyme.

The weakened interactions of the aminoethylated proteins with the transition state analogue despite their normal Mi- chaelis and inhibition constants are not incongruous. Inter- action of ribulose-P2 carboxylase with carboxyarabinitol-P2 is a two-step process in which the initial, reversible complexa- tion is followed by a conformational change that renders the ligands of the complex exchange-resistant (18). Thus, reduced stability of the quaternary complex (enzyme. CO,. Mg’. an- alogue) as a consequence of even minor structural change of an amino acid side chain proximal to a reaction intermediate or transition state during normal catalytic turnover is not totally unexpected. However, the detrimental effects on both kc, and the stability of the quaternary complex by mere replacement of a lysyl y-methylene with a sulfur atom do emphasize the exacting requirements for lysyl side chains at positions 166 and 329.

Recent elegant studies by Pierce et al. (15) utilizing keto- carboxyarabinitol-P,, the authentic reaction intermediate in the carboxylase pathway, provided considerable insight into the requirements for altering the carboxylase/oxygenase ratio. They showed that the enzyme catalyzes the forward partition- ing of the intermediate to product (3-phosphoglycerate) but not its backward partitioning as measured by decarboxylation. Hence, once the six-carbon intermediate is formed, catalysis is irrevocably committed to the carboxylase pathway. These new findings, in conjunction with the recognition (25, 26) of mutual competition between CO, and O2 for the reactive enediol of ribulose-P2, prompted the deduction that the only means for modulating substrate specificity is through differ- ential changes in the reactivities of the gaseous substrates toward the enediol. We rationalized that such changes might be induced by slight manipulation of the steric and polar characteristics of amino acid side chains in close contact with the enediol of ribulose-P2. Although our present observations of an unaltered activity ratio exhibited by either aminoeth- ylated mutant protein do not lend credence to such rational- ization, success through other designed structural changes is not precluded.

REFERENCES

1. Fersht, A. R., Leatherbarrow, R. J., and Wells, T. N. C. (1987)

2. Lowe, D. M., Winter, G., and Fersht, A. R. (1987) Biochemistry

3. Hibler, D. W., Stolowich, N. J., Reynolds, M. A., Gerlt, J. A., Wilde, J. A,, and Bolton, P. H. (1987) Biochemistry 26,6278- 6286

4. Niyogi, S. K., Foote, R. S., Mural, R. J., Larimer, F. W., Mitra, S., Soper, T. S., Machanoff, R., and Hartman, F. C. (1986) J.

5. Hartman, F. C., Soper, T. S., Niyogi, S. K., Mural, R. J., Foote, R. S., Mitra, S., Lee, E. H., Machanoff, R., and Larimer, F. W. (1987) J . Biol. Chem. 262,3496-3501

6. Soper, T. S., Mural, R. J., Larimer, F. W., Lee, E. H., Machanoff, R., and Hartman, F. C. (1988) Protein Engineering, in press

7. Hartman, F. C., Stringer, C. D., and Lee, E. H. (1984) Arch. Biochem. Biophys. 232, 280-295

8. Hartman, F. C., Stringer, C. D., Milanez, S., and Lee, E. H. (1986) Philos. Trans. R. SOC. Lond. B Biol. Sci. 313, 379-395

9. Schneider, G., Lindqvist, Y., Branden, C.-I., and Lorimer, G.

10. Hartman, F. C., Milanez, S., and Lee, E. H. (1985) J. BWZ. Chem.

11. Lindley, H. (1956) Nature 178,647-648 12. Cole, R. D. (1967) Methods Enzymol. 11,315-317

Biochemistry 26,6030-6038

26,6038-6043

Biol. Chem. 261,10087-10092

(1986) EMBO J. 5,3409-3415

260.13968-13975

Aminoethylation of Mutant Ribulose-P2 Carboxylase 4925

13. Tolbert, N. E. (1980) in The Biochemistry of Plants (Davies, D. D., ed) Vol. 2, pp. 487-523, Academic Press, New York

14. Miziorko, H. M., and Lorimer, G. H. (1983) Annu. Rev. Biochem.

15. Pierce, J., Andrews, T. J., and Lorimer, G. H. (1986) J. Biol.

16. Smith, H. B., and Hartman, F. C. (1987) Fed. Proc. 46, 1976 17. Horecker, B. L., Hurwitz, J., and Weissbach, A. (1958) Biochem.

18. Pierce, J., Tolbert, N. E., and Barker, R. (1980) Biochemistry 19,

19. Schloss, J. V., Phares, E. F., Long, M. V., Norton, I. L., Stringer,

52,507-535

Chem. 261,10248-10256

Prep. 6,83-90

934-942

C. D., and Hartman, F. C. (1982) Methods Enzymol. 90, 522- 528

20. Laemmli, U. K. (1970) Nature 227,680-685 21. Bradford, M. M. (1976) Anal. Biochem. 72, 248-254 22. Lorimer, G. H., Badger, M. R., and Andrews, T. J. (1977) Anal.

23. O’Farrell, P. H. (1975) J. Biol. Chem. 250,4007-4021 24. Miziorko, H. M. (1979) J. Biol. Chem. 254,270-272 25. Pierce, J., Lorimer, G. H., and Reddy, G. S. (1986) Biochemistry

26. Van Dyk, D. E., and Schloss, J. V. (1986) Biochemistry 25,5145-

Biochem. 78,66-75

25,1636-1644

5156