Purification and characterization of, and preparation of an antibody to, transketolase from human red blood cells

Biochimica et Biophysica Acta 872 (1986) 24-32 Elsevier

24

BBA32555

Purification and characterization of, and preparation o f an antibody to, t r a n s k e t o l a s e from human red b l o o d cells Tohoru Takeuchi, Kohsuke Nishino and Yoshinori Itokawa Department of Hygiene, Faculty of Medicine, Kyoto University, Kyoto 606 (Japan) (Received November 18th, 1985) (Revised manuscript received April 2nd, 1986)

Key words: Transketolase; Anti-transketolase antibody; Thiamine pyrophosphate; (Human red blood cell)

Transketolase (sedoheptulose-7-phosphate:D-glyceraldehyde-3-phosphate glycolaldehydetransferase, EC 2.2.1.1) was purified 16000-fold from human red blood cells, using DEAE-Sephadex A-50, Sephadex O-150, FPLC on Mono P, and Sephadex 0-100. The purified enzyme migrated as a single protein band on SDS-polyacrylamide gel eiectrophoresis. The FPLC step resolved transketolase into three peaks, designated 1, II and III. From results of re-FPLC on Mono P, SDS-polyacrylamide gel electrophoresis, gel filtration, catalytic studies, amino acid analysis and immunological studies, it was concluded that I, II and III were originally the same protein, modified during storage and purification. Transketolase had a subunit (/14, 70000) and appeared to be composed of two identical subunits. 1 mol of subunit contained 0.9 mol of thiamine pyrophosphate. The pH optimumof the reaction lay within the range 7.6-8.0, and the K m values were determined to be 1.5- 10 -4 M for xylulose 5-phosphate and 4.0.10 -4 M for ribose 5-phosphate. Hg z+ and p-chioromercuribenzoate inhibited the enzyme reaction, and the inhibition of the latter disappeared upon the addition of cysteine. Thiamine and its phosphate esters did not, but cysteine (1 • 10-z M) and ethanol (10% and 1%, v / v ) did activate the enzyme reaction. Antibody prepared to II bound all forms of transketolase in the hemolysate, but inhibited the reaction only about 20%.

Introduction Transketolase (EC 2.2.1.1) is an enzyme of the pentose phosphate cycle which transfers a ketol group to an aldehyde acceptor molecule and requires thiamine pyrophosphate and divalent cations for activity [1,2]. The determination of transketolase activity in human red blood cells is used to evaluate nutritional states of thiamine and to detect some diseases such as Wernicke-Korsakoff syndrome, uremic neuropathy, and beriberi heart disease in relation to the thiamine pyrophosphate effect [3-6]. The thiamine pyrophosphate effect, which is used to detect thiamine deficiency, is said to represent the percentage of apotransketolase to

holotransketolase in samples [3], but it has not been proved. The possibilities of the existence of a congenital anomaly [7] and isoenzymes [8] of transketolase in relation to Wernicke-Korsakoff syndrome were reported, but they should be studied further using a purified enzyme and be confirmed. Although transketolase from yeast has been studied well [9-11], there are only a few reports as to purification and properties of t,ransketolase from human red blood cells [12-14] and there are no reports which describe properties of purified transketolase from human red blood cells. To clarify these problems mentioned above, it is necessary to study properties of human transketolase,

016%4838/86/$03.50 © 1986 Elsevier Science Publishers B.V. (Biomedical Division)

25 because the properties of mammalian transketolase are much different from these of yeast transketolase [15,16]. In this work, a purification method, properties, amino acid composition and preparation of antibody to transketolase from human red blood cells are presented. Materials and Methods

D-Ribose 5-phosphate and thiamine pyrophosphate were purchased from Boehringer; thiamine triphosphate was from Sankyo Pharmaceutical Co. D-Xylulose 5-phosphate and D-sedoheptulose 7phosphate were obtained from Sigma. DEAE-Sephadex A-50, Sephadex G-150, Mono P, Sephadex G-100, Pharmalyte 8-10.5, and Polybuffer 96 were from Pharmacia. LiChrosorbNH 2 (the particle size was 5/~m) was from Merck, ultrafiltration membrane PM-10 from Amieon. Bio-Rad protein assay kit and Affi-Gel Protein A MAPS kit were from Bio-Rad. All other reagents were reagent grade. Outdated human blood was donated by the Department of Gynecology, Kyoto University Hospital. Transketolase activity was determined by the rate of sedoheptulose 7-phosphate formation. The reaction mixture contained 2.0 mM ribose 5-phosphate, 1.0 mM xylulose 5-phosphate in 75 mM diethanolamine-HC1 buffer (pH 7.9) in a final volume of 0.24 ml. Incubations were for 10 rain at 37°C. The enzyme reaction was stopped by addition of 0.26 ml 10% (w/v) trichloroacetic acid or 0.6 ml 7.5% (w/v) trichloroacetic acid if precipitates were expected. When precipitates were detected, centrifugations were done at 1600 × g for 10 min. Sedoheptulose 7-phosphate in 0.5 ml of the reaction mixture was determined by the method of Takeuchi et al. [17]. Thiamine pyrophosphate was identified by the method of Ishii et al. [18] using a LiChrosorbNH 2 column (4 x 150 mm). The quantity of thiamine pyrophosphate was determined by the method of Fujiwara and Matsui [19]. Protein was determined by the method of Lowry et al. [20] and with a Bio-Rad protein assay kit. Polyacrylamide gel electrophoresis was performed in the presence of SDS on 10% polyacrylamide slab gels according to the procedure of

Weber and Osborn [21]. The molecular weight of the native enzyme was determined by gel filtration on Sephadex G-150 which was previously equilibrated and eluted with 100 mM NaC1 in 50 mM Tris-HC1 buffer (pH 8.0). Ferritin (450000), aldolase (160000) and bovine serum albumin (67 000) served as marker proteins. Determination of the molecular weight of the denatured transketolase was performed according to the method of Weber and Osborn [21], with phosphorylase b (94000), bovine serum albumin (67000), ovalbumin (43000), carbonic anhydrase (30000) and trypsin inhibitor (20100) as standard proteins. Proteins were dialyzed exhaustively against distilled water and dried in a desiccator, then hydrolyzed at l l 0 ° C for 22 h in 0.5 ml of 6 M HC1 in vacuo, Samples were evaporated to dryness, redissolved in the analysis buffer, and assayed on a K-101AS amino acid analyser from Kyowa Seimitsu Co. LTD. Tokyo. 250/~g of fraction II in 3 ml of 50 mM Tris-HC1 buffer (pH 8.0) were mixed with 3 ml of complete Freund's adjuvant, and injected intradermally at 40 sites in the back of the rabbit, subcutaneously at four sites in the food-pads, and intramuscularly at four sites in the thighs. Further injections of 250 #g of fraction II, prepared and injected as above, were given 4 weeks later. The blood was obtained 1 week after the last injection, and the antiserum was divided into small volumes (1 ml) and stored at - 90°C. At the time of experiments, 1 ml of the antiserum was applied to 1 ml of Affi-Gel Protein A (MAPS kit) according to BioRad Bulletin 1152. The buffer of purified IgG was changed to 20 mM Hepes-NaOH (pH 7.9). The protein concentration of IgG solution was adjusted to 1.9 mg/ml with the same buffer (anti-transketolase IgG). As a control study, non-immunized rabbit serum was processed in the same way as the antiserum mentioned above (control IgG). Immunological studies. Ouchterlony's procedure [22] was carried out. To clarify whether anti-transketolase IgG bound all forms of transketolase in the hemolysate, the following experiment was done. 150 pl of hemolysate and 100 pl of 20 mM Hepes-NaOH buffer (pH 7.9) were incubated with 50/~1 of anti-transketolase IgG, with an IgG control, or with 50 pl of 20 mM Hepes-NaOH (pH

26

7.9) at room temperature for I h. After addition of 200 pl of saturated ammonium sulfate solution, the reaction mixture was incubated at room temperature for 1 h and centrifuged at 12 000 x g for 20 min. Supernatants were dialyzed against 50 mM Tris-HC1 buffer (pH 8.0). Precipitates were washed with 40~ saturated ammonium sulfate solution twice, then dissolved with 450/~l of distilled water and dialyzed against 50 mM Tris-HCl buffer (pH 8.0). The volumes of the dialyzates were measured and the transketolase activity of the dialyzates was determined by the following method. 150 pl of the dialyzate was incubated with 20 pl of 24 mM ribose 5-phosphate/12 mM xylulose 5-phosphate and incubated at 37°C for 30 rain. After addition of 300 /~l of 7.5~ (w/v) trichloroacetic acid, the reaction mixture was centrifuged at 1600 x g for 10 rain. Sedoheptulose 7-phosphate in 0.2 ml of the supernatant was determined as mentioned above. Results and Discussion

Purification of transketolase Results of purification procedures are shown in Table I. All operations except for FPLC (that was done at room temperature) were done at 5°C. Step I: Preparation of hemolysate. Outdated red blood cells (100 ml) were washed three times with 0.15 M NaC1, then frozen at -90°C. After thawing, 100 ml of distilled water were added. The hemolysate was dialyzed against 10 mM Tris-HCl buffer, pH 8.3, (buffer A) for 36 h. Dialyzed hemolysate was centrifuged at 40000 x g for 60 min.

~

0.05

2.0

FRACTION NUMBER Fig. 1. DE.AE-Sephadex A-50 chromatography. About 200 mJ of dialyzed hemolysate (step 1) were applied to a DEAE-Sephadex A-50 column (6.5 x 60 cm) previously equilibrated with 10 mM Tris-HCl buffer, pH 8.3 (buffer A). After washing the column with 2 1 of buffer A, we ehited with a linear gradient of equal volumes (1.6 I) of buffer A and buffer A containing. 1.0 M NaCI. Transketolase activity was resolved into three peaks as indicated (a, b, and c). The yield of total enzyme activity ( a + b + c ) was 83.2% in this case. The arrow indicates the starting point of the gradient. - - , protein concentration; • • , enzyme activity.

Step 2: DEAE-Sephadex A-50 column chromatography. The supernatant from Step 1 was applied to a DEAE-Sephadex A-50 column (6.5 x 60 cm) previously equilibrated with buffer A. After the column had been washed with 2 1 of buffer A, a linear gradient elution with equal volumes (1.6 1) of buffer A and buffer A containing 1.0 M NaC1 was started. As shown in Fig. 1, the enzyme activity resolved into three peaks (a, b and c). Fractions of peak b were collected and were con-

TABLE I PURIFICATION OF TRANSKETOLASE FROM HUMAN RED BLOOD CELLS This table is the summation of ten purification cycles. Protein was determined by a Bio-Rad protein assay kit. The enzyme activity was determined as described under Materials and Methods. Step

Total protein (mg)

Total activity (nmol/min)

Specific activity (nmol/min per mg)

Yield (~)

Purification (fold)

Hemolysate DEAE-Sephadex A-50 (peak b) Sephadex (3-150 FPLC on Mono P and Sephadex G-150 (I + II + Ill)

830000 500 5.6

142000 26 700 9560

0.17 53.4 1710

100 18.8 6.7

1 314 10000

2.2

5970

4.2

16000

2700

27

0,01

i

N Z W

30"40

5O ~ "

FRACTION

70

80

applied to a Sephadex G-150 column (2.5 x 90 cm) previously eqnifibrated with 50 mM Tris-HC1 buffer, pH 8.0 (buffer B). Elution was done with buffer B as shown in Fig. 2. Only one peak containing transketolase activity was obtained and was concentrated as described above. Step 4: FPLC on Mono P. The buffer of the fraction from step 3 was changed to 25 mM diethanolamine-HCl buffer (pH 9.5). The solution was applied to Mono P which had been equilibrated with the same buffer. Elution was done with a buffer (pH 8.1) containing 1.0 ml of Pharmalyte 8-10.5 and 5.2 ml of Polybuffer 96 in

Fig. 2. Sephadex G-150 chromatography. Concentrated peak b of step 2 was applied to a Sephadex G-150 column (2.5x90 cm) previously equilibrated with 50 mM Tris-HC1 buffer, pH 8.0, and eluted with the same buffer. - - , protein concentration; • • , enzyme activity.

centrated to an appropriate volume using PM-10 under nitrogen. Step 3: Sephadex G-150 column chromatography. The fraction concentrated from peak b was

',\

90

i° 30

"

40 TIME

rain

50

Fig. 3. FPLC on Mono P. The fraction of step 3 was applied to Mono P previously equilibrated with 25 mM diethanolamineHC1 buffer, pH 9.5, and eluted with a buffer (pH 8.1) containing 1.0 ml of Pharmalyte 8-10.5, and 5.2 ml of Polybuffer 96 in 100 ml solution. Transketolase was resolved into three peaks as indicated (I, II and III). The yield of total enzyme activity was almost 100%. The bracket ( u ) indicates the fractions which were taken for re-FPLC on Mono P (see Fig. 8). - - , protein concentration; . . . . . . , pH gradient; • @, enzyme activity.

Fig. 4. SDS-polyacrylamide gel electrophoresis. Purified enzyme preparations of step 5 (I, II and Ill) were appfied and electrophoresed according to the method of Weber and Osborn [21]. The gel was stained with Coomassie brilliant blue R 250.

28

100 ml solution (buffer C). As shown in Fig. 3, the enzyme resolved into 3 peaks (I, II, and III). The dotted line indicates the pH of the elutant. All three peaks were eluted at pH 8.6. On FPLC on Mono P, proteins are usually eluted at the same pH as their isoelectric points.

A

Step 5: Sephadex G-IO0 column chromatography. I, II and III from Step 4 were applied separately to a Sephadex G-100 column (2.5 × 90 cm) previously equilibrated with buffer B in order to eliminate buffer C. All three peaks were eluted at the same elution volume. Purity of the enzyme preparation. All three enzyme preparations (I, II and III) showed a single protein band and had the same R v value on SDS-polyacrylamide gel electrophoresis (Fig. 4). Specific actwity. Specific activities of I, II and III were 3.9, 3.9 and 3.2 /~mol of sedoheptulose 7-phosphate/min per mg protein, respectively. (Protein was determined by the method of Lowry et al. [20] with bovine serum albumin as standard protein.) Molecular weight. The molecular weight of transketolase was estimated to be 70000 by SDSpolyacrylamide gel electrophoresis (Fig. 5A) and 104000 by gel filtration (Fig. 5B). Heinrich et al. [12] reported that the molecular weight of partially purified transketolase was 104000 by gel filtration and 136000 by ultracentrifugation, and that the discrepancy was due either to dissociation into subunits or to an interaction of transketolase with the gel filtration media. We concluded that the molecular weight of a subunit was 70 000 and the native transketolase was composed of two possibly identical subunits. Coenzyme. The purified enzyme contained only thiamine pyrophosphate. 1 mol of the subunit contained 0.9 mol of thiamine pyrophosphate (protein was determined by the method of Lowry et al. [20]). Additional thiamine pyrophosphate ( 1 - 1 0 -3 M) with or without Mg 2+ did not enhance the enzyme activity. Thiamine, thiamine monophosphate and thiamine triphosphate at the same concentration, with or without Mg 2+, had also no effects on the enzyme activity. Isoelectric point. The isoelectric point was estimated to be p H 8.6 at 25°C from the result of FPLC on Mono P. Catalytic properties. Transketolase was most ac-

o.'2

Rf

o!6

50C . B 40C 3O(

I-

S0

10

I

200 ELUTIONVOLUME(ml)

I

3O0

Fig. 5. Molecular weight of the enzyme. A, estimation of the molecular weight by SDS-polyacrylamidegel electrophoresis. Enzyme preparations of step 5 were applied to a 10% polyacrylamide gel containing 0.1% SDS. The standard proteins were; e, phosphorylase b (Mr 94000); O, bovine serum albumin (67000); ×, ovaibumin (43000); II, carbonic anhydrase (30000); A, trypsin inhibitor (20100). B, estimation of the molecularweight by gel filtration. Enzymepreparations were applied to a SephadexG-150 column (2.5X90 cm) previously equilibrated with 100 mM NaCI in 50 mM Tris-HC1 buffer (pH 8.0), and eluted with the same buffer. The standard proteins were; n, ferritin (Mr 450000); z~, aldolase (160000); v, bovine serum albumin (67000).

tive in the p H range 7.6-8.0 with 0.1 M diethanolamine-HC1 buffer (Fig. 6). The K m values for xylulose 5-phosphate and ribose 5-phosphate were estimated to be 1.5.10 -4 M and 4 . 0 . 1 0 -4 M, respectively. The maximum velocity was calculated to be 4.2-10 2 m o l / m i n per tool enzyme. Stability. Purified enzyme (60-100 ~tg/ml) was stable at - 9 0 ° C for several weeks. Freezing and

29 TABLE II EFFECTS OF METAL IONS AND INHIBITORS ON TRANSKETOLASE

~1~0 -

-

The reaction mixture contained 24 #1 of the material, 20 VI of the enzyme solution (70/tg/ml), 156/~1 of 0.1 M diethanolamine-HCl (pH 8.0) and 40 Itl of 12 mM ribose 5-phosphate/6 mM xylulose 5-phosphate solution. The pH values of the reaction mixtures lay within the range 7.8-8.1. After incubation at 37°C for 10 min, the enzyme reaction was stopped by addition of 0.26 ml of 10% (w/v) trichloroacetic acid. Each experiment had its control run. The composition of the control was the same as the reaction mixture described above except that 24 ~1 of the corresponding material were added to the control after the enzyme reaction was stopped. Material

pH Fig. 6. Enzyme activity as a function of pH. Enzyme activity was determined with the following buffers, e, 0.1 M citrateNaOH, O, 0.1 M Hepes-NaOH; x , 0.1 M diethanolamineHCI. The pH was measured at 37°C in the final assay mixture.

thawing had no effects on the enzyme activity. Incubation at 55°C for 5 min reduced 50~ of the enzyme activity, but in the presence of albumin (0.2 mg/ml) reduced only 3~ of activity. With regard to the above properties ( M r, coenzyme, pI, catalytic properties and stability), the results were identical for I, II and Ill. Effect of metal ions and inhibitors. (These results were obtained from preparation II for lack of preparations I and III.) As shown in Table II, Hg 2+ (1.10 -4 M) and p-chloromercuribenzoate (5.10 -4 M) inhibited the reaction almost completely. By the addition of cysteine, the enzyme activity was recovered from the inhibition caused by the latter. Cysteine (1.10 -2 M) and ethanol (10~ and 1%, v/v) enhanced the enzyme activity. In the present experiment, phosphate (1 • 10-2 M) did not inhibit the enzyme activity. But partially purified transketolase from human red blood cells is reported to be inhibited by the same concentration of phosphate [13]. We supposed that the discrepancy derived from the difference in assay systems for transketolase activity. The authors in Ref. 13 used a coupled enzyme system containing

Hg 2+ Cu 2+ Zn 2+ Mg 2+ Ca 2+ p-Chloromercuribenzoate Iodoacctamide N-Ethylmalcimide EDTA Phosphate Cysteine Ethanol Acetaldehyde

Final concentration

Ratio

(M)

(v/v%)

(%)

10.0 1.0 0.01

0 68.1 66.2 107.2 110.1 2.9 35.9 83.3 84.9 90.8 51.2 100.9 154.1 178.2 123.8 104.5

1.10 -4 1" 10- 3 1.10- 3 1.10- 2 1.10 -2 5.10- 4 1.10- 4 1-10- 3 1.10- 3 1.10- 2 1.10-1 1.10 -2 1-10- 2

triose phosphate isomerase and a-glycerophosphate dehydrogenase. As the latter enzyme was inhibited by phosphate and sugar phosphates [23], we considered that a simple assay system such as we used in this experiment was better able to determine the effects of materials on the enzyme reaction. Transketolase from yeast shows different responses to these sulfhydryl reagents and phosphate from human. Amino acid analysis. The amino acid compositions of I, II, and III are shown in Table III. I, II and III had almost the same composition. Human transketolase had a different amino acid composition from yeast and pig [24]. Immunological studies. On Ouchterlony's diffusion method, only one precipitation line was detected between antiserum and each hemolysate,

30

Fig. 7. Ouchterlony's diffusion method. The agar was stained with Coomassie Brilliant Blue R 250. Precipitation lines were detected between antiserum and hemolysate, a, b, c, I, II, and III. Each line fused completely. On the contrary no precipitation lines were detected between normal rabbit serum and enzyme preparations. The enzyme preparations of a, b, and c were obtained from DEAE Sephadex A 50 column chromatography, and I, II, and III were from FPLC on Mono P.

peak a, b, c, I, II and III, and precipitation lines fused completely (Fig. 7). No precipitation line formed in the presence of normal rabbit serum. Anti-transketolase I g G to II bound all forms of transketolase in the hemolysate but inhibited the enzyme reaction only about 20% (Table IV). These forms of the enzyme could not be distinguished with the preparation of anti-transketolase I g G which was used. Polymorphism of transketolase. Transketolases from yeast [9-11,24], rat [24,25] and pig [15,16,24] seem to have only one form, but h u m a n transketolase seems to have m a n y forms [7,8]. At step 2, transketolase activity was resolved into three peaks (a, b and c). In this study, only peak b of step 2 was taken for purification to homogeneity, because the other peaks had some contaminating proteins after the last step of the procedures. When peaks a and c were applied to Mono P after the same procedures taken to purify peak b, peak

a yielded the same elution profile as peak b, while the transketolase activity of peak c was recovered in the same position of peak III. Peaks a, b, and c were eluted with the same elution volume on gel filtration and had the same R F value on SDSpolyacrylamide gel electrophoresis as I, II and III. The elution profile of re-FPLC on Mono P of fractions marked by the horizontal bracket in Fig. 3 revealed that I and II could be converted to III (Fig. 8). However, re-FPLC on Mono P of III, showed that III did not change. I, II and I I I had the s a m e M r on gel filtration, the s a m e R F value on SDS-polyacrylamide gel electrophoresis, the same number of thiamine pyrophosphates, the same p H optimum and K m values, the same amino acid composition and could not be distinguished with the antibody prepared to II. We concluded that I, II and III were originally the same protein and were modified during storage and the purification process. It was supposed from

31 TABLE III

r-

AMINO ACID COMPOSITION OF TRANSKETOLASE Each enzyme preparation was hydrolyzed at 110°C for 22 h in 0.5 ml of 6 M HC1 in vacuo, n.d., not detected under these hydrolytic conditions, n.c., detected but not calculated. Amino acid

Trp Lys His Arg Asp Thr Ser Glu Pro Gly Ala Cys Val Met lie Leu Tyr Phe

/I

Amino acid composition (%)

120-~

E OO3

I

II

III

n.d. 7.6 3.0 4.8 9.7 5.5 7.1 10.9 2.5 8.0 10.8 n.c. 6.2 n.c. 7.7 7.8 2.7 5.3

n.d. 7.5 3.0 4.8 9.9 5.6 6.9 11.2 2.4 7.9 11.2 n.c. 6.4 n.c. 7.5 8.2 2.5 5.1

n.d. 7.2 3.0 5.2 9.3 5.8 6.5 11.1 2.6 7.4 11.5 n,c. 6.4 n.c. 7.8 8.4 3.1 4.7

c

!oo2

~0

3b

40

w

TIME rain Fig. 8. Re-FPLC on Mono P of the marked fractions in Fig. 3. Conditions were the same as Fig. 3. - - , protein concentration; • • , enzyme activity.

be no transketolase variants other than the a, b and c peaks. But the mechanism with which human transketolase had resolved into three peaks was not known. Unlike the report of Nixon et al. [8], this paper shows that transketolase in human red blood cells was originally a single protein. We considered that there were at least four reasons for this discrepancy. (1) Modification during storage and purification. Our results led us to conclude that the polymorphism of transketolase was due to the modification of transketolase during storage and

the results of specific activity and re-FPLC on Mono P that III was partially denatured form of I and II. With the results from FPLC on Mono P, gel filtration, SDS-polyacrylamide gel electrophoresis, and immunological studies, we also concluded that I, II and III corresponded to the a, b and c peaks of step 2, respectively. Judging from the high yield of transketolase activity on DEAESephadex A-50 (a + b + c, 87%), there seemed to

TABLE IV PRECIPITATION OF TRANSKETOLASE IN HEMOLYSATE BY ANTI-TRANSKETOLASE IgG Hemolysate was incubated with anti-transketolase IgG, with control IgG , or with 20 mM Hepes-NaOH buffer (pH 7.9) at room temperature for 1 h. Then saturated ammonium sulfate solution was added to make 40% saturated solution. After dialysis of supernatants and dissolved precipitates against 50 mM Tris-HCI buffer (pH 8.0), the enzyme activities were determined. Incubation

Supernatant

Total

Recovery

with

activity (nmol/ml per rain)

volume (pl)

activity (nmol/ml per min)

Dissolved precipitate volume (VI)

activity (nmol/min)

(%)

Anti-transketolase IgG Control IgO 20 mM Hepes buffer

0 54.2 61.9

610 780 710

79.5 4.1 5.0

470 470 420

37.4 44.2 46.0

81.3 96.1 100

32

purification. Nixon et al. [8] seemed to consider some forms of the modified transketolase as isoenzymes. (2) The difference in assay system for detection of transketolase activity. It seemed that they [8] used a less specific assay system than ours. We supposed that there might be some materials, other than transketolase, in the hemolysate which reduced phenazine methosulphate in the presence of xylulose 5-phosphate, ribose 5-phosphate, NAD, nitroblue tetrazolium, glyceraldehyde-3-phosphate dehydrogenase, MgCl2, sodium arsenate, and contaminants in these reagents. (3) The possibility of complex formation between transketolase, transaldolase and glyceraldehyde-3-phosphate dehydrogenase. Human transketolase in the hemolysate might form enzyme complexes with transaldolase and glyceraldehyde-3-phosphate dehydrogenase like transketolase from Candida utilis [26]. These enzyme complexes might have different isoelectric points and be separated by isoelectric focussing. Nixon et al. [8] might have determined these enzyme complexes as transketolase variants. (4)The difference in race. This seemed less likely, but might play some role on the discrepancy, for there exists a different susceptibility to WernickeKorsakoff syndrome between the white and yellow races. This is the first report of a purification method, the properties, amino acid composition and preparation of antibody to transketolase from human red blood cells. Human transketolase was originally a single protein and is easily modified during storage and purification. An antibody prepared to II bound all forms of transketolase in the hemolysate. Using purified enzyme and anti-transketolase IgG, we may be able to determine the enzyme protein concentration in the hemolysate and to clarify the relationship between thiamine pyrophosphate concentration, enzyme activity and enzyme protein concentration. It will be a great help to study thiamine pyrophosphate effects and genetic mutations of the enzyme assumed to cause Wemicke-Korsakoff syndrome.

Acknowledgements We are grateful to Dr. Rikimaru Hayashi, The Institute of Food Science, Kyoto University, for

the amino acid analysis of the enzyme, and to Dr. Yutaka Kojima, Department of Environmental Medicine, Graduate School of Environmental Science, Hokkaido University, for his helpful comments on the manuscript.

References 1 Horex:ker, B.L. and Smyrniotis, P.Z. (1953) J. Am. Chem. Soc. 75, 1009-1010 2 Racker, E., De la Haba, G. and Leder, I.G. (1953) J. Am. Chem. Soc. 75, 1010-1011 3 Brin, M. (1966) Methods Enzymol. 9, 506-514 4 Dreyfus, P.M. (1962) N. Engl. J. Med. 267, 596-598 5 Lonergan, E.T., Semar, M., Sterzei, R.B., Treser, G. Needle, M.A., Voyles, L. and Lange, K. (1971) N. Engl. J. Med. 284, 1399-1403 6 Akbarian, M. and Dreyfus, P.M. (1968) J. Am. Med. Assoc. 203, 77-80 7 Blass, J.P. and Gibson, G.E. (1977) N. Engl. J. Med. 297, 1367-1370 8 Nixon, P.F., Kaczmarek, M.J., Tare, J., Kerr, R.A. and Price, J. (1984) Eur. J. Clin. Invest. 14, 278-281 9 Datta, A.G. and Racker, E. (1961) J. Biol. Chem. 236, 617-628 10 Egan, R.M. and Sable, H.Z. (1981) J. Biol. Chem. 256, 4877-4883 11 Shreve, D.S., HoUoway, M.P., Haggerty, J.C. and Sable, H.Z. (1983) J. Biol. Chem. 258, 12405-12408 12 Heinrich, P.C. and Wiss, O. (1971) Helv. Chim. Acta 54, 2658-2668 13 Warnock, L.G. and Prudhomme, C.R. (1982) Biochem. Biophys. Res. Commun. 106, 719-723 14 Schellenberg, G.D., Wilson, N.M., Copeland, B.R. and Furlong, C.E. (1982) Methods Enzymol. 90, 223-228 15 Simpson, F.J. (1960) Can. J. Biochem. Physiol. 38, 115-124 16 Tomita, I., Saitou, S. and Ishikawa, M. (1979) J. Nutr. Sci. Vitaminol. 25, 175-184 17 Takeuchi, T., Nisluno, K. and Itokawa, Y. (1984) Clin. Chem. 30, 658-661 18 Ishii, K., Sanemori, H. and Kawasaki, T. (1979) Anal. Biochem. 97, 191-195 19 Fujiwara, M. and Matsui, K. (1953) Anal. Chem. 25, 810-812 20 Lowry, O.H., Rosebrough, N.J., Farr, A.L. and Randall, R.J. (1951) J. Biol. Chem. 193, 265-275 21 Weber, K. and Osbom, M. (1969) J. Biol. Chem. 244, 4406-4412 22 Ouchterlony, O. (1958) Prog. Allergy. 5, 1-78 23 Lee, Y.P., Jonathan, E.C. and Lee, A.S.-T. (1973) J. Biol. Chem. 248, 5610-5615 24 Kochetov, G.A. (1982) Methods Enzymol. 90, 209-223 25 Paoletti, F. (1983) Arch. Biochem. Biophys. 222, 489-496 26 Wood, T., Muzariri, C.C. and Malaba, L. (1985) Int. J. Biochem. 17, 1109-1115