Purification of Yeast cw-Isopropylmalate Isomerase HIGH IONIC STRENGTH HYDROPHOBIC CHROMATOGRAPHY*

a-Isopropylmalate isomerase, the second enzyme specific for leucine biosynthesis, can be purified from extracts of yeast utilizing a chromatographic procedure that allows separation of proteins in the presence of high concentrations of (NH,)$O,. The purification procedure utilizes the stabilizing effect of glycerol and (NH,),SO, on the isomerase and their opposing effects on protein retention on valine-Sepharose and leucine-Sepharose. The method effectively separates the isomerase from fumarase, a stable internal marker protein that was co-purified in early steps. High ionic strength hydrophobic chromatography, based on differential retention as a function of the length of the hydrophobic sidearm and ionic strength, yields approximately 200-fold purified cu-isopropylmalate isomerase and may be of general utility in purifying unstable enzymes requiring high ionic strength.

a-Isopropylmalate isomerase, the second enzyme specific for leucine biosynthesis, can be purified from extracts of yeast utilizing a chromatographic procedure that allows separation of proteins in the presence of high concentrations of (NH,)$O,.
The purification procedure utilizes the stabilizing effect of glycerol and (NH,),SO, on the isomerase and their opposing effects on protein retention on valine-Sepharose and leucine-Sepharose.
The method effectively separates the isomerase from fumarase, a stable internal marker protein that was co-purified in early steps. High ionic strength hydrophobic chromatography, based on differential retention as a function of the length of the hydrophobic sidearm and ionic strength, yields approximately 200-fold purified cu-isopropylmalate isomerase and may be of general utility in purifying unstable enzymes requiring high ionic strength.

Affinity chromatography as developed by Cuatrecasas
(2) has led to many novel procedures of protein purification, in most cases based on a specific interaction between protein and a ligand that has been covalently attached to a Sepharose. Subsequent developments have demonstrated the utility of Sepharoses substituted with nonpolar ligands, whereby chromatographic separations are based on hydrophobic interactions between protein and the substituted Sepharoses (3)(4)(5)(6). The procedure has been termed hydrophobic chromatography. Recent studies have shown that interaction of protein with the substituted Sepharoses, especially Sepharoses with any one of the branched-chain amino acids attached, may be enhanced by salts of the lyotropic, or Hofmeister, series (7)(8)(9)(10)(11). The salts examined most extensively, phosphate and sulfates, are required in high concentrations for such interactions between proteins and the hydrophobic matrix. The application of high ionic strength hydrophobic chromatography on valine and leucine-Sepharoses has proved to be useful for the purification of yeast a-isopropylmalate isomerase (2-isopropylmalate dehydratase, EC 4.2.1.33), the second enzyme specific for leucine biosynthesis, for the same high ionic strength conditions (1.24  stabilizing (r-IPM2 isomerase, decreases the interaction of the enzyme with the hydrophobic matrix and promotes its release from the matrix. Thus, it has been possible to combine the antagonistic effects of the two stabilizing agents (glycerol and (NH,),SO,) and the differential retentive properties of the two Sepharose substitutents in preparing eluting gradients that resulted in a selective separation of the isomerase from an enzyme that was co-purified in the initial steps, fumarase.

EXPERIMENTAL PROCEDURE
Enzyme Assays--~Y-IPM isomerase was assayed by the method of Gross et al. (12) using dimethylcitraconate as substrate. Fumarase was assayed by a similar procedure except that the reaction mixture contained 10 rnM L-malate.  (14) and was modified as previously described (15). The trace element solution and vitamin solution were those of Lucas et al. (16) O-5". Thirty-five-milliliter portions were then disrupted by sonic oscillation with a Bronson S75 Sonifier at a power setting of 5 amperes using one 30-s treatment.
Such a treatment released virtually all of the isomerase yet only a fraction of the fumarase and protein that could be obtained with a longer treatment and in effect constituted a purification of severalfold (Fig. 1). These results agree with the findings of Ryan et al. (20) who found that the isomerase was released much more readily than cu-IPM synthase, the first enzyme specific for leucine biosynthesis, and the mitochondrial markers cytochrome oxidase and citrate synthase. Since S288c(u yeast was more resistant to breakage, four 30-s sonic oscillation treatments were used. (NH,),SO, fractionations were performed by the slow addition of solid (NH,),SO, with stirring. The suspensions were allowed to stand for 20 min after which they were centrifuged in a Sorvall RC-2 centrifuge for a period of 10 min at 27,000 x g. The 0 to 50% saturation (NH,),SO, fraction thus obtained was discarded and the 50 to 65% fraction resuspended in 2.0 ml of the column equilibration buffer desired. This protein preparation was loaded onto a valine-or leucine-Sepharose column. Protein resolution was enhanced by applying the protein gently onto the surface of the substituted agarose from a pipette tip extended beneath the column buffer. Five-milliliter fractions were collected and assayed immediately.
Chromatography 4317 matography of a 50 to 65% (NH,),SO, fraction resuspended in protein and both enzymes were completely retained (Fig. 5). the column equilibration buffer resulted in a retardation of the Prolonged washing with the original buffer failed to release protein and a further retardation of the two enzyme activities, protein or enzyme activity. A shift to a potassium phosphate isomerase and fumarase, which eluted together. A turbid 30% glycerol buffer at Fraction 88, however, released the fraction containing a small amount of protein was excluded.
ret,ained protein, isomerase, and fumarase. Thus, glycerol and The presence of 2 mM leucine in the elution buffer did not alter (NH,),SO, exerted opposing effects on protein retention and this profile. The absence of a leucine effect was expected since produced an intermediate effect when both were present. the presence of 10 mM of any of the branched-chain amino Chromatography on unsubstituted Sepharose 4B under condiacids, separately or in concert, in the reaction mixture did not tions identical with those in Fig. 5 produced no retardation of influence the activity of the isomerase or fumarase.
either the isomerase or the fumarase. Chromatography on Valine-Sepharose in Presence of Glycerol and (NH,),SO,-On valine-Sepharose (Fig. 3), employing the same glycerol-(NH,),SO, conditions, the isomerase and fumarase were only slightly separated from the bulk of the protein. The length of the hydrophobic group of the amino acid substitutent seemed influential in producing this result.
Chromatography on Leucine-Sepharose in Presence of Glycerol-The effect of glycerol and (NH,),SO, on the chromatographic behavior of isomerase and fumarase was clearly demonstrated by the elution profiles produced when only one of these agents was present. Leucine-Sepharose equilibrated with a potassium phosphate buffer 30% (v/v) in glycerol resulted in the protein, isomerase, and fumarase being eluted together (Fig. 4). Glycerol proved ineffective in allowing retention on leucine-Sepharose.
The trailing minor peak of isomerase activity is presumed to have been caused by residual (NH,)$O, in the 50 to 65% protein fraction resuspended in the phosphate-glycerol buffer, (NH,),SO, which thereby produced the effect depicted in Fig. 2.
Ionic Strength Dependence of Interaction of Protein with Leucine-Sepharose-Interaction of the isomerase and fumarase with leucine-Sepharose was clearly ionic strength-dependent. The capacity, defined as recoverable enzyme units after a phosphate-glycerol buffer shift, increased markedly when ionic strength was increased to about 3.5 (Fig. 6). In contrast to the behavior of the isomerase, the fumarase interacted with leucine-Sepharose at a lower ionic strength. Furthermore, the stability of the isomerase on the columns and in the eluate increased markedly with increased ionic strength.
Enzyme that was not adsorbed to the leucine-Sepharose still interacted with the matrix and was retarded. The retardation, expressed in terms of VJV,, was also ionic strength dependent (Fig. 7). Fumarase again was retarded at a lower ionic strength than was the isomerase. The enzymes could be separated on leucine-Sepharose only at higher salt concentrations as de- 4318 pitted in the inset of Fig. 7. The affinity of the hydrophobic matrix for fumarase was generally stronger than its affinity for isomerase. Gradient Elution on Leucine-Sepharose-When a 50 to 65% (NH,),SO, fraction was adsorbed on leucine-Sepharose from the phosphate-l.24 M (NH,),SO, buffer described, elution with a buffer containing a gradient of increasing glycerol and decreasing (NH,),SO, concentration reproduced the effects demonstrated in earlier experiments (Fig. 8). Proteins were distributed in response to the gradient. The fumarase again interacted more tightly with the matrix.
Chromatography on Valine-Sepharose in Presence of (NH,),SO,-On valine-Sepharose equilibrated with a potassium phosphate-l.24 M (NH,),SO, buffer, some protein was retarded while some protein was completely retained and could be released only upon a phosphate-glycerol buffer shift at Fraction 210 (Fig. 9). The isomerase was retarded on valine-Sepharose under these conditions and could be separated from the bulk of the protein. A fraction of the fumarase was not adsorbed, the result of a capacity effect. The fumarase that was eluted could be rechromatographed and adsorbed. All of the fumarase applied was adsorbed on more highly substituted valine-Sepharose, one containing 16 pmol of valine/g of Sepharose, dry weight (Fig. 10).

Enzyme Purification
Purification procedures that retained a-IPM isomerase in a favorable environment of high ionic strength yet allowed chromatographic separations were devised. An early procedure involved chromatography on leucine-Sepharose as illustrated in Fig. 2. Fractions with maximal activity were concentrated by ultrafiltration with an Amicon XM50 membrane and the concentrate dialyzed to remove glycerol in order to allow (NH,),SO, fractionation. Because fumarase was precipitated at a slightly higher (NH,),SO, concentration, several precipitations yielded electrophoretically pure isomerase, but owing to the dialysis step, at a purification of only 13-fold. The more so-so0 so-830 efficient procedure based on differential retention as a function of sidearm length and ionic strength yielded 200-fold purified isomerase (Table I).
Chromatography on Valine-Sepharose-The 50 to 65% (NH,),SO, fraction obtained as described above was applied onto a valine-Sepharose column (2 x 21.5 cm) equilibrated with 0.05 M potassium phosphate, pH 6.8, 1.24 M (NH,),SO,. Elution with this buffer resolved the isomerase from excluded and adsorbed proteins (Fig. 11). A shift to a buffer of low ionic strength again released adsorbed proteins. Elution with the buffer of decreasing (NH,),SO, and increasing glycerol concentration resulted in an early release of the isomerase (Fig. 12).
The leucine-Sepharose fractions containing the major amount of isomerase activity were pooled and (NH,),SO, was added to approximately 70% saturation. After standing, the suspension was centrifuged in order to eliminate the glycerol remaining in the supernatant fluid. The 70% (NH,),SO, precipitate was washed with 0.05 M potassium phosphate, pH 6.8, 62% saturated with (NH,),SO,, centrifuged, and the  (v/v) glycerol and stored at -20". In these fractionation procedures, 1 hour of equilibration was allowed before the precipitates were removed by centrifugation.
Variability in Fold Purification-The specific activity, fold purification, and yield varied and depended on the amount of time allowed for each step. The range of fold purification was 70 to 200 for bakers' yeast ol-IPM isomerase. The range of the specific activity for the final preparation was 2.3 to 6.2. Thus, the indication is that active and inactive enzyme were not separated during the purification procedure. The enzyme from yeast strain S288ccu could be purified to an extent falling in the lower portion of this range, suggesting that the enzyme is somewhat less stable in this strain.
Electrophoresis of Enzyme-Polyacrylamide disc gel electrophoresis of 10 pg of purified bakers' yeast wIPM isomerase yielded a single band (Fig. 13A). The necessary dialysis of the protein to remove stabilizing salt destroyed most of the enzyme activity. However, using a simple gel slicing procedure, remaining traces of activity were correlated with the protein band. A diffuse band was commonly observed (Fig. 13B) that 4320 appeared as two bands with higher protein concentration (Fig.  13C). Purified enzyme from yeast strain S288c(u consistently yielded two bands (Fig. 13C). Sodium dodecyl sulfate disc gel electrophoresis of 10 pg of purified S288ccu cu-IPM isomerase yielded one band (Fig. 14E). DISCUSSION The purification procedure for cu-IPM isomerase described here has utilized conditions that preserve its activity and effect its interaction with a hydrophobic matrix, high ionic strength. The sulfate ion employed allows both stabilization and hydrophobic interaction.
The effect of sulfate and other lyotropic series salts on the behavior of proteins in solution has long been studied and is believed to be mediated by the structure of water and therefore entropy effects (21,22), or by a more direct effect of these salts on the protein itself (23)(24)(25). However, the effect of such salts on the intermolecular and intramolecular interactions of proteins is not completely understood. High ionic strength hydrophobic chromatography, besides being useful as a preparative tool, may allow the study of many purified proteins that require high ionic strength.
Many enzymes have been shown to respond to or require high salt concentrations.
Multiple RNA polymerases have been found to be released from sea urchin embryos, rat liver nuclei (26), and yeast (27) (28). Yeast phosphofructokinase has been chromatographed on Sepharose 6B in the presence of a buffer 30% saturated in (NH,),SO,, a salt that was an effective stabilizer and protease inhibitor (29). Salmonella typhimurium phosphoribosyladenosine triphosphate synthetase has been shown to be stabilized by high (NH,),SO, concentration (30). The quarternary structure of the enzyme appears dependent on ionic strength and pH. For these and many other enzymes with similar properties, one would expect that high ionic strength hydrophobic chromatography could be a useful tool. However, for proteins not dependent on high ionic strength, the valine-Sepharose procedure alone could be used in conjunction with (NH,),SO, fractionation in the quick partial purification of (NH,),SO, fractions (Fig. 12). Considering its application to the purification of (Y-IPM isomerase and its potential for other enzymes, high ionic strength hydrophobic chromatography employing (NH,) $0, represents a useful procedure for the separation of biological materials.
The possibility of an effect of salt on the quarternary structure of the isomerase must be considered. Whereas a single band or clear doublet were observed upon disc gel electrophoresis of purified bakers' yeast isomerase, the enzyme from yeast strain S288c(u produced two bands consistently. Kohlhaw et al. (31) reported the appearance of two bands in the case of purified S. typhimurium (Y-IPM synthase and later reported this to be a result of an association-dissociation equilibrium (32). However, in the case of the isomerase, multiple aggregation states cannot as yet be distinguished from conformational states. Especially interesting is the effect of salts on the structure of succinate-ubiquinone reductase, which can be dissociated by chaotropic salts and reconstituted by removal of these salts or addition of anti-chaotropic ions such as S0,2-, HP0,2-, or Fm (33). The integrity of the complex is believed to be maintained by hydrophobic associations, interactions strengthened by anti-chaotropic salts in aqueous media. The effect of anti-chaotropic salts on the stability of LU-IPM isomerase from Salmonella typhimurium and yeast has been studied and appears general.' Sodium dodecyl sulfate-treated yeast isomerase appears as one band upon disc gel electrophoresis.
The presence of one subunit species or two very similar subunits cannot as yet be distinguished.
However, based on genetic evidence in S. typhimurium, the isomerase is believed to be coded for by two genes, 1euC and 1euD (12,34). Since infrequent unlinked suppressor mutations, termed supQ (35,36), of a total 1euD deletion are possible, it is conceivable that the role of the 1euD gene is one other than catalytic function in uitro. If so, the genetics of the isomerase in S. typhimurium would be similar to that in yeast where many genes have been demonstrated to be involved in determining isomerase activity (15). If one type of subunit is present, the role of these other genes becomes an interesting problem for further investigation.