PURIFICATION AND PROPERTIES OF YEAST TRANSALDOLASE*

Sedoheptulose-7-phosphate, originally identified as a product of photosynthesis in algae and higher plants (3), has been shown to arise by the action of transketolase on pentose phosphate (4, 5). This reaction occurs in a wide variety of animal (5, 6), plant (5, 7), and microbial cells (8-lo), and evidence has been obtained that the heptulose ester is an intermediate in the formation of hexose monophosphate from pentose phosphate (7, 11, 12, 13). With the isolation of sedoheptulose-7-phosphate from transketolase reaction mixtures (5), it became possible to study directly the conversion of this ester to hexose monophosphate. In an earlier publication from this laboratory (14) it was reported that this reaction requires a source of triose phosphate, and evidence was presented for the accompanying mechanism. The enzyme has been named transaldolase, since it catalyzes the

Sedoheptulose-7-phosphate, originally identified as a product of photosynthesis in algae and higher plants ( 3), has been shown to arise by the action of transketolase on pentose phosphate (4,5). This reaction occurs in a wide variety of animal (5,6), plant (5,7), and microbial cells (8-lo), and evidence has been obtained that the heptulose ester is an intermediate in the formation of hexose monophosphate from pentose phosphate (7,11,12,13).
With the isolation of sedoheptulose-7-phosphate from transketolase reaction mixtures (5), it became possible to study directly the conversion of this ester to hexose monophosphate. In an earlier publication from this laboratory (14) it was reported that this reaction requires a source of triose phosphate, and evidence was presented for the accompanying mechanism. The enzyme has been named transaldolase, since it catalyzes the transfer of aldol linkages and no reaction occurs in the absence of the acceptor. In the present paper a procedure for the purification of transaldolase from yeast is described, together with evidence for the formation YEAST TRAXSALDOLASE of fructose-6-phosphate and tetrose phosphate and for the mechanism of the reaction.

Methods1
Xedoheptulose-7-phosphate-This was prepared by the action of spinach transketolase on R-5-P, as previously described (5). The product used for most of this work was 70 to 80 per cent pure.
Hezose &phosphate-Commercial HDP was purified by ion exchange chromatography on Dowex 1 formate columns and precipitated as the barium salt. The product was 80 per cent pure, uncorrected for moisture content, and was free of other phosphate esters.
Glyceraldehyde-S-phosphate-This was prepared from HDP according to a suggestion of Dr. T. Biicher of the University of Marburg. HDP was treated with aldolase in the presence of a large excess of acetaldehyde. Under these conditions, DHAP reacted to form methyltetrose-l-phosphate (15) and G-3-P accumulated.
The product was isolated by ion exchange chromatography.
The reaction mixture was put on a Dowex 1 formate column, 9.6 sq. cm. X 15 cm., and eluted by gradient elution (16), with 2 liters of water in the mixing bottle and a solution containing 0.25 M ammonium formate and 0.1 M formic acid in the reservoir.
Fractions of 20 ml. were collected and aliquots analyzed with DPN and glyceraldehyde-3-phosphate dehydrogenase.
Fractions 69 to 77, containing 181 pmoles of G-3-P, were treated with 1.0 ml. of 1.0 M calcium chloride and neutralized with 0.60 ml. of saturated NaOH.
The calcium salt was precipitated with 4 volumes of ethanol, washed with 80 per cent ethanol, and dried in vacua. The product, 57 mg., contained 163 Imoles of G-3-P, which accounted for 78 per cent of the total phosphate present.
Other Xubstrates-Barium R-5-P was a commercial preparation. Uniformly labeled HDP was obtained from the Nuclear Instrument and Chemical Corporation. S-7-P, labeled with Cl* in positions 1 and 3, was made from R-5-P labeled in position 1; the preparation was carried out as previously described (5).
Coenzymes and Enzymes-Commercial TPN and DPN were purified by ion exchange chromatography (17,18). Aldolase was prepared according to Taylor et al. (19). Hexosephosphate isomerase was prepared as previously described (5). Zwischenferment was isolated from brewers' yeast by the method of Kornberg (20) and phosphogluconic dehydrogenase by the method of Horecker and Smyrniotis (21). Glyceraldehyde-3-phosphate dehydrogenase was prepared from rabbit muscle according to Cori,Slein,and Cori (22). Acid phosphatase was purified from potato by the method of Kornberg.2 Brewers' yeast was kindly furnished by the Chr. Heurich Brewing Company of Washington, D. C. Spinach leaf extracts were prepared by homogenizing 60 gm. of fresh leaves with 360 ml. of water in a Waring blendor for 3 minutes and filtering the homogenate with Schleicher and Schuell No. 588 filter paper. The extracts of rat liver acetone powder were fractionated with ammonium sulfate as described previously (13). Analytical Methods-Absorption measurements were made with a Beckman DU spectrophotometer.

Occurrence of Transaldolase
In Liver-With the ammonium sulfate fraction prepared from extracts of rat liver acetone powder, S-7-P is converted to G-6-P only if triose phosphate is present (Fig. 1). Either R-5-P or HDP will serve as a source of triose phosphate; the reaction is most rapid when a mixture of HDP and crystalline muscle aldolase is added. With the crude transaldolase preparation the effect of HDP is catalytic, since nearly 3 moles of G-6-P were formed per equivalent of triose phosphate added (Table I). This suggests that triose phosphate can be regenerated from the tetrose ester; a mechanism for this conversion is proposed in the succeeding paper.
In Spinach-Extracts of spinach leaves will form G-6-P from S-7-P at a rapid rate (Fig. 2), whereas with R-5-P as the substrate G-6-P appears only after an appreciable lag period.
An effect of HDP cannot be demonstrated with these extracts, since they contain a very active fructose diphosphatase which forms F-6-P (and therefore G-6-P) from HDP. Presumably such unfractionated extracts contain endogenous sources of triose phosphate.
In Microorganisms-Dried brewers' yeast is an excellent source of transaldolase (14), and the enzyme has been purified from this source. It has also been detected in extracts of Escherichia coli (11). In the experiments with HDP 0.05 mg. of crystalline aldolase was added.
The total volume was 1.5 ml. G-6-P formation was calculated from the density change at 340 rnp resulting from the reduction of TPN. The reaction mixture contained excess isomerase and Zwischenferment, but even under these conditions the rate of TPN reduction was not proportional to the amount of transaldolase added, and the true rate was therefore assumed to be given by the broken line in Fig. 3. Thus for an observed rate of 0.012 density unit per minute, the corrected rate was estimated to be 0.016. In order to reduce the size of this correction, the observed rate was usually maintained between 0.010 and 0.015 density units per minute; however, larger corrections proved to be sufficiently precise for purification purposes.
A preparation of 6-phosphogluconic dehydrogenase from yeast was employed as a source of Zwischenferment (21)  cause of the high activity of Zwischenferment in this preparation and its greater stability in the frozen state. A unit of transaldolase is defined as the amount which causes a change in optical density of 1.0 per minute, corrected for deviation from proportionality, Specific activity is defined as the number of units per mg. of protein.
Extraction-Brewers' yeast was washed with 10 volumes of cold water, allowed to settle overnight in the cold room, pressed, and dried on wire screens in a stream of air. The dried yeast (240 gm.) was suspended in 720 ml. of 0.1 M sodium bicarbonate and allowed to autolyze at 25" for 7.5 hours. The mixture was treated with 4.0 liters of cold water and centrifuged (International, size 2). The residue was discarded and the supernatant solution stored at 2" overnight (Table II,  YEAST TRANSALDOLASE Acetone Fractionation-2.15 liters of autolysate were brought to pH 4.8 with 12.0 ml. of 5 N acetic acid and treated with 1.10 liters of acetone, previously chilled to -12". The acetone was added slowly (over approximately 10 minutes), and during this time the mixture was stirred mechanically and cooled to -5" in a freezing bath. After 10 minutes longer at this temperature, the precipitate was removed by a brief centrifugation at 2" (International, size 2, 2 minutes). The supernatant solution was cooled to -8" and treated with 455 ml. of cold acetone, added during 6 minutes.
After 30 minutes at -8" the precipitate was collected by centrifugation, dissolved in 120 ml. of water, and adjusted to pH 7.0 with 0.27 ml. of 5 N ammonium hydroxide.
The procedure was repeated with the remaining autolysate and the solutions combined (acetone fraction, 283 ml.).
Calcium Phosphate Adsorption I-The acetone fraction was diluted to 650 ml. with water, to bring the protein concentration to 8.0 mg. per ml., and treated with 283 ml. of calcium phosphate gel (4.0 gm. dry weight). The gel was collected by centrifugation and the enzyme eluted with two 141 ml. portions of 0.15 M pyrophosphate buffer, pH 7.6 (calcium phosphate eluate, 335 ml.).
Ammonium Sulfate Fractionation-The calcium phosphate eluate was treated with 100 gm. of ammonium sulfate and centrifuged and the precipitate discarded.
Two more fractions were collected by the addition of 21

Ethanol
Fractionation-The calcium phosphate eluate was dialyzed overnight against flowing 0.05 M acetate, pH 7.4, and the dialyzed solution (3.2 ml.) was treated with 0.08 ml. of 2 N acetic acid to bring the pH to about 4.5. The solution was cooled in a freezing bath, and 50 per cent (volume per volume) ethanol was added slowly, with mechanical stirring. After 10 minutes at -8" the solution was centrifuged at -10" and the precipitate dissolved in 1.0 ml. of glycylglycine buffer, pH 7.4. The supernatant solution was treated with 0.7 ml. of absolute ethanol, cooled in a freezing bath to 15", and the precipitate centrifuged and dissolved as before. A third fraction was collected by the addition of 0.8 ml. of absolute ethanol in the same manner. This fraction usually contained most of the activity (ethanol fraction, 1.1 ml.).

Properties of Enzyme
Stability-The final preparation is moderately stable when stored at -16"; after 1 week 80 to 90 per cent of the original activity remains. This product still contains traces of triosephosphate isomerase, although this activity represents less than 0.05 per cent of that present in the original autolysate. For most purposes the enzyme can be used after the acid fractionation step, since this preparation is free of transketolase and hexosephosphate isomerase. At this stage of purification the enzyme can be stored at -16" for several weeks with little loss in activity.
Effect of pH and Substrate Concentration-Under the conditions of the assay the rate changes only slightly with variation in pH from 7.3 to 8.1. Below pH 7.3 the rate falls off rapidly, and at pH 7.0 is only 30 per cent of the maximum.
The effect of substrate concentration is illustrated in Fig. 4. The affinity constants (K,), calculated according to Lineweaver and Burk (25), were 1.8 X 1W4 and 2.2 X lo-* mole per liter for S-7-P and G-3-P, respectively. In view of the complicated assay method, these values must be regarded as only approximate.
Substrate Specificity-Transaldolase exhibits a high degree of specificity. S-7-P was not replaced by free sedoheptulose or ribulosed-phosphate. Free glyceraldehyde, glycolaldehyde, or glycolaldehyde phosphate did not serve as acceptor in place of G-3-P.
Absorption Spectrum-The absorption spectrum of the purified enzyme preparation showed it to be relatively free of nucleic acid or nucleotide components, with a ratio of absorption at 280 to 260 rnp of 1.77 (26). No evidence for a coenzyme or metal cofactor has been obtained; the fact that the enzyme can be precipitated from ammonium sulfate solution at pH 2.5 and dialyzed without loss of activity suggests that no readily dissociable cofactors are present.

Reaction Mechanism
With C14-Labeled Triose Phosphate-The requirement for triose phosphate with liver preparations is indicated in Fig. 1; evidence for the same requirement with partially purified yeast preparations has been reported earlier (14). The rale of triose phosphate in the reaction was elucidated with The assay system was as described in Fig. 3, except that in the experiments with G-3-P, HDP and aldolase were omitted; the concentration of S-7-P was 3.5 X 10-B M.
For the lower curve, HDP (1.8 X 1OW M) was present.
uniformly labeled HDP-Cl4 as the source of triose phosphate (Table III). The G-6-P formed, isolated as the crystalline barium salt, had one-half the specific activity of the HDP added, indicating that 3 of the 6 carbon atoms were labeled. This was confirmed by degradation (13) of the glucose after enzymatic dephosphorylation, when it was found that nearly 90 per cent of the isotope was located in carbon atoms 4, 5, and 6; carbon atoms 1, 2, and 3, which contained little radioactivity,3 must have been derived from S-7-P.
With S-7-P-l ,S-C14-On the basis of the experiment described in the preceding section, it was apparent that in the reaction catalyzed by trans- 3 The radioactivity in C-2 and C-3 may not be significant, since a large correction was required due to the addition of carrier ethanol and the actual count obtained was only a little above background.
YEAST TRANSALDOLASE aldolase a dihydroxyacetone group was transferred from S-7-P to triose phosphate, presumably G-3-P, to form F-6-P.
Free DHA does not equilibrate with the group being transferred, since no dilution of isotope oc- curred when the reaction was carried out with S-7-P-l ,3-Cl4 in the presence of an excess of non-radioactive DHA (Table IV).

Products of Reaction
Fructose-6-phosphate-It has previously been shown (14) that without the addition of the hexosephosphate isomerase preparation little or no G-6-P is formed from S-7-P on incubation with transaldolase.
The forma- tion of F-6-P as a major product under these conditions is illustrated in Fig. 5. The enzyme preparation used for this experiment was relatively impure and contained small amounts of glucose phosphate isomerase which slowly converted the F-6-P formed to the equilibrium mixture containing about 70 per cent of G-6-P.
Tetrose Phosphate-Evidence for the formation of a tetrose ester has been obtained by paper chromatography (Fig. 6). This experiment was carried out with a crude transaldolase preparation, and, as a result, a small amount of S-7-P was converted to pentose phosphate by transketolase. A strong spot corresponding to erythrose was present which, under ultraviolet illumination, had the bright yellow fluorescence characteristic YEAST TRANSALDOLASE of tetroses. However, attempts to isolate tetrose phosphate from the reaction mixture have been unsuccessful.
Several secondary reactions involving this product have been observed; these will be discussed in the succeeding paper (29 Equilibrium Measurements-Direct evidence for the reversibility of the transaldolase reaction is not yet available, but the effect of changes in concentration of reactants on the amount of fructose-6-phosphate is consistent with the formation of an equilibrium mixture (Table V). Because triosephosphate isomerase is present, the over-all equilibrium (K') favors S-7-P; when a correction is made for the formation of DHAP, the equilibrium constant (K) for the transaldolase reaction itself is estimated to be nearly unity.

DISCUSSION
The reaction catalyzed by transaldolase represents a new type of sugar transformation in which a dihydroxyacetone group is transferred from 1 molecule to another.
In contrast to transketolase, which catalyzes a similar transfer of '(active glycolaldehyde," there is no evidence for a prosthetic group to bind the active group, and the specificity with respect to acceptor molecules is much more limited.
To date, only S-7-P and F-6-P have been found to serve as DHA donors.
In view of the biological occurrence of mannoheptulose (31), it is of interest to speculate on the formation of compounds having 3,4-cis linkages in this reaction in addition to those which, like S-7-P and F-6-P, have trans linkages.
The identification of F-6-P as the primary reaction product accounts for the earlier observation of Glock (32) that this is the first, hexose ester to arise from pentose phosphate with rat, liver preparations.
The isolation of the tetrose ester formed from S-7-P and G-3-P remains to be accomplished.
Evidence for the identity of this compound with n-erythrose-4phosphate is given in the succeeding publication (29).
The r81e of transaldolase in the formation of hexose monophosphate from pentose phosphate has been considered in previous communications (13,33,34). This reaction may also be of importance in the generation of the COZ acceptor in photosynthesis (35).