Leslie M. Shaw, Jack W. London,1and Lorette E. Petersen. We isolated 'y-glutamyltransferase .... (Gilford, Oberlin, Oh. 44074) as de- scribed elsewhere. (15).
CLIN.CHEM.24/6, 905-915(1978)
Isolationof ‘y-Glutamyltransferasefrom Human Liver, and Comparison with the Enzyme from Human Kidney Leslie M. Shaw, Jack W. London,1 and Lorette E. Petersen
We isolated ‘y-glutamyltransferase [(y-glutamyl)-peptide:amino acid -y-glutamyltransferase, EC 2.3.2.2] from human liver and compared some of its properties with the
same enzyme prepared from human kidney. The enzymes from these two sources are very similar with respect to initial velocity kinetic constants, pH optima of the transpeptidation and autotransfer reactions, heat stability, competitive inhibitionby glutathione of the colorimetric
assay in which y-glutamyl-4-nitroanilide is substrate, stability of catalytic activity to trypsinization, and relative rates of transfer of the ‘y-glutamyl moiety from ‘y-glutamyl-4-nitroanilide and L-[glycine-2-3H] glutathione to some amino acids and small peptides. The kidney enzyme is
inhibited more by the ‘y-glutamyl acceptor substrate, glycylglycine, as reflected in a sevenfold lower value for the inhibition constant K.A. Major differences were observed in the lectin-binding properties of liver -y-glutamyltransferase compared to the kidney enzyme. Lectiri-binding property differences are retained for the trypsinized form of the liver and kidney enzymes, although the degree of precipitation was less for certain lectins as compared to the untreated enzyme. Lectin-binding properties were reversed by carbohydrates specific for each lectin. We adapted the histochemical staining technique of Rutenberg et al. [J. Histochem. Cytochem. 17, 517 (1969)] to the
detection of -y-glutamyltransferase activity in acrylamide gels; diffusion artifacts are minimized and the color produced is stable for several days. Untreated and trypsinized forms of the liver enzyme both migrated faster in acrylamide gels (as single bands) than did the corresponding forms of the kidney enzyme. AddItional Keyphrases: chromatography
enzyme kinetics enzyme activity binding of lectins differences in the same enzyme from different tissues liver disease transplant rejection y -glutamyl transfer tissue of origin of enzyme in serum
‘y-Glutamyltransferase has been isolated from kidney tissue of various species (1-3) including man (4,5). Some properties of the enzyme from this tissue have been determined, including initial velocity kinetics (6-8), molecular weight, amino
acid and carbohydrate
composition,
and subunit composition
The William Pepper Laboratory, Division of Laboratory Medicine, Department of Pathology, Hospital of the University of Pennsylvania, Philadelphia,
Pa. 19104.
Present address: Radiology Department, versity
of Pennsylvania,
Philadelphia,
Received Jan. 30, 1978; accepted
Hospital
Pa. 19104. Mar. 27, 1978.
(1, 2, 9). The enzyme has also been isolated from other mammalian tissues, but with much less information about its properties. y-Glutamyltransferase isolated from azo dyeinduced hepatoma of rats has some catalytic properties similar to those reported earlier for the enzyme purified from hog and beef kidney (10), and the catalytic properties of ‘y-glutamyltransferase isolated from human seminal fluid are similar to those previously reported for the enzyme in other human tissues, with only slight differences from the enzyme from animal kidney (11). Thus, present evidence indicates great similarities in those properties-primarily catalytic properties-of y-glutamyltransferase isolated from different mammalian tissues that have been studied. Here we compare some kinetic and lectin-binding properties of y-glutamyltransferase isolated from human liver and kidney and show that although the enzymes prepared from these two human tissue are very similar with respect to the kinetic properties investigated, there are very substantial differences in their respective lectin-binding properties. There is further interest in establishing the properties of y-glu-
tamyltransferase obtained from normal adult human liver and kidney because the activity of this enzyme is above normal in the liver and serum of patients with liver diseases (12, 13) and in the urine of kidney transplant patients during rejection episodes (14). Thus a study of the properties of the enzyme from the normal tissues will serve as an important basis for the study of the pathophysiological mechanisms responsible for the appearance of the enzyme in these biological fluids.
Materials Sephadex Pharmacia
G-200 and concanavalin A were obtained from Fine Chemicals, Piscataway, N.J. 08854; diethy-
laminoethyl-cellulose
(DE-52)
from
Whatman
Inc.,
Clifton,
N.J. 07014; DEAE-cellulose, Con A-agarose,2 glutathione, trypsin (bovine pancreas, Type III), wheat germ agglutinin, soybean agglutmnin, and asparagus pea (Lotus tetragonolobus) agglutinin from Sigma Chemical Co., St. Louis, Mo. 63178; Ricinus communis agglutinin 120, y-glutamyl-4-nitroanilide, ‘y-glutamyl-3-carboxy-4-nitroanilide, aldolase (EC 4.1.2.17, rabbit muscle), albumin (bovine serum), cytochrome c (horse heart), and ferritin from BioDynamics/bmc, Indianapolis, md. 46250; acrylamide, methylene bisacrylamide, N,N,N’,N’tetramethylenediamine, and Hydroxylapatite Bio Gel HT from Bio-Rad, Rockville Centre, N.Y. 11570; L-[glycine-23H]glutathione (240 Ci/mol) from New England Nuclear, Boston, Mass. 02118; -y-glutamyl-4-methoxy-2-naphthylamide from Vega-Fox Biochemicals, Tucson, Ariz. 85719; and
of the Uni2
abbreviations used: Con A, concanavalin tris(hydroxymethyl)aminomethane hydrochloride.
Nonstandard
tris/HCI,
CLINICAL
CHEMISTRY,
Vol. 24, No.
A; and
6, 1978 905
Fast Blue BBN from Polysciences, Inc., Warrington, Pa. 18976. All other chemicals used were of the highest analytical quality available.
Methods Enzyme assay. Unless otherwise indicated, all assays of enzyme activity were conducted at 30 #{176}C, with use of a Stasar III spectrophotometer (Gilford, Oberlin, Oh. 44074) as described elsewhere (15). The final concentrations of substrates and buffer for the standard assay were: 4 mmol/liter 7-glutamyl-3-carboxy-4-nitroanilide, 100 mmol/liter glycyiglycine, 100 mmol/liter tris(hydroxymethyl)aminomethane HC1, pH 8.25. As described in detail by Shaw et al. (15) y-glutamyltransferase activities determined by assays containing 4 mmol/liter y-glutamyl-4-nitroanilide and 40 mmol/liter glycylglycine are comparable to the standard assay used in this work. The standard assay was used in all steps of the isolation of liver and kidney ‘y-glutamyltransferase. We measured the relative rates of y-glutamyl transfer to some amino acids and small peptides, catalyzed by the hepatic and renal enzyme, with either ‘y-glutamyl-4-nitroanilide or L-[glycine-2-3H]glutathione as y-glutamyl donor substrates. The reaction rates with the latter substrate were determined by the radioassay method of Zelazo and Orlowski (2), in which the rate of formation of L-[glycine-2-3Hjcysteinylglycine was measured. In this assay procedure L-[glycine-2-3H]glutathione (80 000 counts/mm) and each acceptor substrate were present at concentrations of 5 and 20 mmol/liter, respectively, in 100 mmol/liter tris/HC1,2 pH 8.7, containing 10 mmol of dithiothreitol and 10 mmol of MgCl2 per liter. After incubating each reaction mixture for 15 mm at 30 #{176}C, we terminated the reactions by adding 1.5 ml of picric acid solution (10 g/liter). These mixtures were then loaded on top of small Dowex-1 acetate columns (0.5 X 3 cm) and each column was washed with 2 ml of distilled water. Glutathione and picric acid are retained on these columns, but cysteinylglycine is recovered quantitatively in the effluent. One-milliliter portions of each effluent were mixed with Bray’s solution (16) and counted in a Tn Carb liquid scintillation counter (Packard Instruments Co., Inc., Downers Grove, Ill. 60515). The control used for this experiment was the total reaction mixture, including glycylglycine as acceptor substrate but no enzyme, taken through the entire incubation and assay procedures. Lectin binding. We studied the selective precipitation of the isolated hepatic and renal enzymes by various soluble plant lectins by incubating a constant amount of the respective enzyme at room temperature (25 #{176}C) for 2 h in 25 mmol/ liter tris/HC1, pH 7.5, Triton X-100 (2.5 g/liter), 0.5 mmol/liter NaCl, 0.05 molJliter MgC12 with various amounts of each leetin (0-1 g per liter of total reaction mixture). After this incubation the solutions were centrifuged at 5000 X g for 30 mm and the supernates assayed for y-glutamyltransferase activity by the assay procedure described. The controls for these studies were the respective enzyme mixed with all of the constituents of the reaction mixture, but no lectins, and taken through the entire incubation and centrifugation steps. Acrylamide gel electrophoresis. We electrophoresed enzyme in 61/2% acrylamide according to the method of Davis (17). The gels were stained for y-glutamyltransferase activity by incubation for 1 h at room temperature in a substrate mixture containing, per liter, 0.4 mmol of y-glutamyl-4methoxy-2-naphthylamide, 3.8 mmol of glycylglycine, and 0.33 g of Fast Blue BBN in 0.1 mol/liter tris/HC1, pH 7.4, and 0.14 mol/liter NaC1. Before -y-glutamyl-4-methoxy-2-naphthylamide was added to the rest of the reaction mixture it was dissolved in dimethyl sulfoxide and an equal volume of 1 mol/liter NaOH was added to it. Thus, to prepare 20 ml of total reaction mixture, 25 mg of y-glutamyl-4-methoxy-2-
906
CLINICALCHEMISTRY,Vol. 24, No. 6, 1978
naphthylamide is first added to 0.5 ml of dimethyl sulfoxide and dispersed with a glass rod until dissolved; 0.5 ml of 1 mol/liter NaOH is then added and this solution is mixed with 19 ml of a solution of the rest of the substrate mixture in 0.15 mol/liter NaC1. After being incubated with the substrate mixture, the gels were removed from the substrate solution and placed in 0.15 mol/liter NaCl for 30 mm. Then they were placed in a 0.1 mol/liter solution of CuSO4 for 30 mm. The gels were stored in 0.15 mol/liter NaCl. The color produced by this procedure, a deep rose-red, is stable for several days, after which it gradually fades. This procedure is based on the histochemical method for y-glutamyltransferase developed by Rutenberg et al. (18). It has the advantage over other methods of minimal diffusion of the naphthylamine product before diazotization, thus allowing formation of discrete bands. Duplicate gels were stained for carbohydrate by the method of Zacharius et al. (19) and for protein with Coomassie Brilliant Blue R-250. Enzyme purification. (a) Human Liver -y-Glutamyltransferase. Step 1: We obtained normal liver tissue at autopsy less than 10 h postmortem and froze it at -30 #{176}C immediately. The tissue was stored at this temperature until used for an enzyme preparation (about two weeks). For a typical preparation, 100 g of frozen liver was thawed and homogenized at 4 #{176}C in a Waring Blendor with three volumes of a solution of 80 mmol/liter MgC12 and 750 imol/liter NaOH. The homogenization was continued for 3.5 mm, during which time it was necessary to interrupt homogenization several times and place the homogenizer vessel in ice to prevent heat denaturation of the enzyme. We never allowed the temperature to exceed 15 #{176}C during the homogenization procedure. The homogenate was incubated at 37 #{176}C for 2 h and then centrifuged at 10 000 X g for 30 mm. We performed all subsequent steps at 4 #{176}C. Step 2: Deoxycholate Extraction. To the pellets from the 10 000 X g centrifugation was added 150 ml of deoxycholate (10 g/liter) in 0.1 mol/liter Tris/HC1 buffer, pH 8.0. This mixture was stirred overnight and centrifuged at 15000 X g for 45 mm. The supernatant fluid was dialyzed for 36 h against three successive 18-liter changes of deoxycholate (1 g/liter). Step 3: Batchwise DEAE-Cellulose Chromatography. The dialyzate was mixed with DEAE cellulose (bed volume of 100 ml) that had been equilibrated with phosphate buffer (10 mmol/liter, pH 7.0), and stirred for 2 b. This mixture was placed on a B#{252}chner funnel. The DEAE-cellulose cake, to
which the enzyme was bound, was rinsed, suction filtration, with 500 ml of equilibration
with the aid of buffer followed
by 200 ml of the same buffer containing deoxycholate (1 g/ liter). The enzyme was then eluted from the DEAE-cellulose cake with 500 ml of the phosphate buffer with added NaCl (1 mol/liter) and Triton X-100 (5 g/liter). This eluate was collected in six 80-ml fractions. Fractions 1-4 contained most of the activity, and these were combined and dialyzed against three changes of 18 liters of tris/HC1 (50 mmol/liter, pH 7.5, in Triton X-100, 5 g/liter) for 36 h. Step 4: Hydroxylapatite Chromatography. The dialyzate was loaded on a 5.0 X 11.5 cm column of hydroxylapatite, which had been equilibrated with buffer of the same composition as that used for dialysis in Step 3. The column was then eluted with 900 ml of equilibration buffer. Under these conditions most of the enzyme was adsorbed onto the apatite, with only a small proportion (12%) passing through the column at this point in the procedure. The adsorbed -y-glutamyltransferase was eluted from the column with phosphate buffer (0.1 mol/liter, pH 7.5, in Triton X-100, 5 g/liter). The flow rate throughout chromatography with hydroxylapatite was 150 ml/h. The fractions containing enzyme activity (Figure 1) eluted from the column with the buffer were combined and dialyzed against three changes of 18 liters of tris/
0
,,oI/h,.r
phOpP
against three 18-liter changes of tris/HC1 (50 mmolfliter, pH 7.5, in Triton X-100, 5 g/liter). (b) Human kidney --glutamyltransferase: We obtained normal kidney tissue at autopsy less than 10 h postmortem and froze it at -30 #{176}C without delay. The tissue was stored at
b,ff*,
1800 -
1400 >
this temperature until used for an enzyme preparation (about three weeks). The first two steps in the preparation of ‘y-glutamyltransferase from kidney cortex are the same as for the hepatic enzyme. The third is also batchwise chromatography with DEAE-cellulose, but in contrast to this step for the hepatic enzyme, the DEAE-cellulose was equilibrated in the
10CC
L
600 200 0
50
00
50
Fraction
Number
200
250
Fig.1.Chromatography on hydroxylapatite The dlalysate of the fractions containing the liver enzyme obtained from the batch DEAE-cellulose step was subjected to adsorption chromatography on hydroxylapatite as described under Methods. Six-milliliter fractions were collected and enzyme activity was determined in them by the -y-glutamyltransferase assay procedure described under Methods
HC1 (50 mmol/liter, pH 7.5, in Triton X-100, 5 g/liter). The small portion of the enzyme that was not adsorbed by this column of hydroxylapatite was not further studied. StepS: Affinity Chromatography with Insolubilized Concanavalin A. The dialyzate from Step 4 was loaded on a 1.5 X 25 cm column of concanavalin A-agarose, which had been equilibrated with 50 mmol/liter tris/HC1, pH 8.0, containing 0.5 mol/liter NaCl, Triton X-100 (5 g/liter), 1 mmol/liter CaCl2, and 1 mmol/liter MgCl2. After loading the enzyme solution onto the column, it was rinsed with 180 ml of equilibration buffer. The enzyme was eluted from the column as a broad peak with equilibration buffer containing methyl-aD-mannopyranoside (0.2 mol/liter). The flow rate throughout the entire chromatographic procedure was 20 ml/h. Some 22%
of the enzyme did not bind to the column and was recovered in the fractions preceding elution of the main enzyme fraction. The total recovery of hepatic enzyme with the Con A affinity column was 85-100% for three separate preparations. The portion of the enzyme that did not bind to this column was not studied further. At this stage of purification the overall enrichment of enzyme activity was 426-fold, with a yield of 16%. Table 1 summarizes the steps in a typical isolation procedure of y-glutamyltransferase from 100 g of liver. All protein determinations reported in Table 1 and throughout the re-
mainder of this work were done by the method of Lowry et al. (20). The hepatic enzyme, purified through Step 5 as outlined above,
was used
in all studies
described
below
after
dialysis
presence of deoxycholate (1 g/liter). With the renal, but not the hepatic enzyme, this is necessary; otherwise, highly variable and relatively low yields of enzyme activity were obtained in this step. The renal enzyme was eluted from the DEAE-
cellulose cake with 0.01 mol/liter phosphate buffer, containing NaCl (0.5 mol/liter) and Triton X-100 (5 The eluted enzyme was dialyzed against three 18-liter of 10 mmol/liter phosphate buffer, pH 8.5, in Triton
pH 6.0, g/liter). changes
X-100
(5 g/liter). Step 4: DE-52 Chromatography. DE-52 was prepared as described in Whatman circular no. 1L2 (Whatman, Inc., Clifton, N.J. 07014). The equilibration buffer was 10 mmol/ liter phosphate, pH 8.5, containing Triton X-100 (1 g/liter). After loading the dialyzed enzyme from Step 3 onto the DE 52 column (2.5 X 15 cm) and washing it with 500 ml of equilibration buffer, the enzyme was eluted by means of a linear gradient established between 150 ml of phosphate buffer (10 mmolf liter, pH 7.0, in Triton X-100, 5 g/liter) and 150 ml of 10 mmol/liter phosphate buffer, pH 6.0, containing 0.2 mol of NaCl per liter, in Triton X-100 (5 g/liter). Figure 2 depicts
a typical chromatogram.
The flow rate throughout
this step
was 80 ml/h. In addition to the primary peak (fractions 29-50) a small secondary peak (fractions 60-65) contained 5% of the total enzyme activity. The latter was not further studied. A summary of the partial purification of the y-glutamyltransferase from 50 g of kidney cortex up to this point is also given in Table 1. We obtained a 46-fold enrichment of enzyme activity, with a 36% yield. The renal enzyme purified through Step 4 was used in all studies described below after dialysis against three 18-liter changes of tris/HC1 (50 mmol/liter, pH 7.5, in Triton X-100, 5 g/liter). In contrast to the hepatic enzyme the human renal enzyme did not bind very effectively to concanavalin A-agarose. With use of the same procedure
outlined for the hepatic enzyme most of the renal enzyme (an average affinity
of 80% for several preparations) did not bind to the column. Thus, an affinity-chromatography step with
Table 1. PurificatIon of 7-Glutamyltransferase Protein
Homogenate Deoxycholate extract DEAE-cellulose batch Hydroxyapatite
Volume, ml
Total mg
387 144 327 476
From human liver 22 252 1273 572 163
ActIvity Concn., gil
Speclflc kU!g
Total
U
57.5 8.84 1.75 0.34
606 573 398 235
0.33
95
0.027 0.45 0.70 1.44
Yield, %
Purlf IcatIon, -fold
-
-
95 66 39
16.7 25.9 53.3
16
426
chromatogr.
Concanavalin A-agarose
25
8.23
11.5
chromatogr.
From human kidney cortex Homogenate
300
5,580
3,290
0.59
-
Deoxycholate extract DEAE cellulose batch
168 255
988 387
5.88 1.52
2,800 3,520
2.83 9.10
85 107
4.9 15.4
33
44.6
1.35
1,200
36
46
DE-52 chromatogr.
18.6
26.9
CLINICAL CHEMISTRY,
-
Vol. 24, No. 6, 1978
907
confidence limits were calculated for the kinetic constants, to indicate how well the constants were defined. Trypsinization of -y-glutamyltrans [erase. We treated both the hepatic and the renal -y-glutamyltransferase, purified from D
.
>
I
#{176}
20
30 40 50 Fraction Number
Fig.2. Chromatography on DE-52 The dialysate of the fractions containing kidney -y-glutamyltransf erase obtained from the batch DEAE-cellulose step was chromatographed on DE-52 as described under Methods. Fractions of 4 ml were collected and enzyme activity was determined on these by the assay as described under Methods; protein was determined in each fraction by the method of Lowry et al. (20)
use of Con A-agarose
was not used to further
kidney -y-glutamyltransferase. Analysis of kinetic data. 15),
we developed
equation
Results
purify human
Enzyme
studies of the -y-glutamyltnansferase reaction (8, and applied the following initial velocity
of the mechanism
to the reactions
glutamyltransferase above:
Guided
by our previous
catalyzed
prepared
by liver and kidney -y-
from these tissues
the respective tissues as outlined above, with trypsin (EC 3.4.21.4) and determined the catalytic activity and lectinbinding properties of the resulting form of the enzyme. To 3.0 mg of enzyme preparation (liver or kidney) in 6 ml of tris/HC1 buffer (50 mmol/liter, pH 7.5, in Triton X-100, 5 g/liter, and containing 50 mmol of NaCl and 10 mmol of CaC12 per liter) at 37 #{176}C, we added 0.5 mg of trypsin every half hour for 6 h. After incubating for a total of 8 hat 37 #{176}C, this reaction mixture was brought to room temperature for an additional 14 h. When used for lectin-binding studies, the enzyme was further purified by passage through a 0.75 X 16.5 cm column of DE 52 that had been equilibrated with 50 mmol/liter tnis/HC1, pH 7.5. In this step the trypsin-treated renal enzyme passed through the column unimpeded; in contrast, the trypsintreated liver enzyme was bound to the column and could be eluted with equilibration buffer containing 0.4 mol of NaCl per liter.
as described
As summarized in Table 1, we achieved a 426-fold purification and 16% yield of -y-glutamyltransferase from human liver. When the purified material was electrophoresed in 61/2% acrylamide gels and stained for y-glutamyltransferase activity,
Vmt([D]EA] + + aK
KmA
K m A \[D]+KD[A]+(1+ l)A/
m
/
substrate
concentration.
We determined kinetic constants enzymes by a nonlinear regression (Digital
(equation
Corp.,
base for calculating tamyltransferase
for the hepatic and renal fit of the initial velocity
1) to the data, using a PDP-10 computer
Equipment
Marlboro,
Mass. 01752). The data consisted of y-glumeasurements over a y-glutamyl
the kinetic constants activity
donor substrate
concentration range of 0.2-6.0 mmol/liter (0.25-40 mmoVliter for the more soluble 3-carboxy derivative) and glycyiglycine concentration range of 2-150 mmol/liter. For the hepatic
enzyme,
we made
over these concentration
39 activity
measurements
ranges in each of two separate
ex-
periments with y-glutamyl-4-nitroanilide as donor substrate and 42 activity determinations over the above ranges in each
of two separate nitroanilide
experiments
as donor
with -y-glutamyl-3-carboxy-4-
substrate.
For the renal
made 36 activity
determinations
experiments
‘y-glutamyl-4-nitroanilide
with
measurements
as the
concentration
we
in each of three separate
in each of two experiments
3-carboxy-4-nitroanilide
above substrate
enzyme,
donor
and 36 activity
with y-glutamylsubstrate,
over
the
ranges.
In deriving the kinetic constants by the nonlinear regression fit of initial velocity equation 1 to the kinetic data, we used the method of Fletcher and Powell (22) to minimize the deviation between calculated and observed enzyme velocities. From the Hessian Matrix (23) of this minimization procedure, 95% 908
CLINICAL CHEMISTRY,
Vol. 24, No. 6, 1978
K m 1) “rn Ii DA) A trn
where v is the initial velocity, KmA is the acceptor (glycylglycine) Michaelis constant, Km’ is the donor (-y-glutamyl-4nitroanilide or y-glutamyl-3-carboxy-4-nitroanilide) Michaelis constant, KmDA in the donor-acting-as-acceptor (autotransfer) Michaelis constant, a is the ratio of autotransfer to transfer turnover numbers, Vmt is the -y-glutamyl transfer maximum velocity, K1A is the acceptor competitive inhibition constant, [D] is the donor substrate concentration, and LA] is
equation
1D12) (1)
(
the acceptor
Purification
+ K
A
K D D2+__IA]2
Krn
a slow-moving single band was detected. When chromatographed on a column of Sephadex G-200, the purified enzyme eluted in the void volume, with no apparent further purification. The hepatic enzyme preparation was not homogeneous, however, because in gels stained for protein several small faster-moving bands were visible in addition to the major band which corresponded to the slow-moving band in the gel that was stained for enzyme activity. Only one band stained for carbohydrate, and it corresponded to the slowmoving band of enzyme activity. y-Glutamyltransferase was obtained in a partly purified form from human kidney cortex as outlined in Table 1. A 46-fold purification was obtained, with a yield of 36%. When this preparation was electrophoresed in 61/2% acrylamide gels and stained for y-glutamyltransferase activity, a slow-moving single band was observed. The electrophoretic mobility of the renal enzyme was even slower than that of the hepatic enzyme. The renal enzyme preparation was not homogeneous; several protein bands in addition to that of the enzyme band being visible. Only one carbohydrate-containing band was seen, and this corresponded to the band with enzyme activity.
Trypsin
Treatment
No loss of catalyticactivity was detected for either enzyme after incubating the -y-glutamyltransferase preparations (prepared through Step 5 for the liver enzyme and Step 4 for the kidney enzyme) with trypsin for 8 hat 37 #{176}C as described in the Methods section. The trypsin-treated hepatic enzyme eluted significantly later on a Sephadex G-200 column than did the untreated enzyme and had an approximate relative molecular mass of 120 000 (Figures 3 and 4). Untreated enzyme eluted in the void volume of this column as a single peak. Of the total trypsinized hepatic enzyme activity, 76% was present in the fraction of low relative molecular mass (100-132
6O >
.40
a
E 2O
“0
40
80120 ISO Elution Volume (ml)
200
Fig.3. Chromatography on Sephadex G-200 Liver -y-glutamyltransferase, 3.4 U in a total volume of 2 ml, treated with trypsin for 8 h as described under Methods, was chromatographed on a 2.5 X 58 cm column of Sephadex G-200. Before chromatography, the column had been equilibrated with buffer consisting of 50 mmol/liter NaCI and Triton X-100 (1 g/llter). The flow rate was maintained at 12 mI/h, the same buffer used for column equilibration serving as elution buffer
0.8
K0
rome Ca Albumin
0.6
0.4 Aldolose...’
trypein digestion)
“‘Clver6GTGGT(after
0.2
Iver GGT I ixIO
5
10
50a10’
Molecular Weight Fig.4. Standards (rabbit muscle adolase, bovine serum albumin, horse heart cytochrome c, and ferritin), trypsin-treated and trypsin-untreated liver y-glutamyltransferase, and Blue Dextran
were chromatographed on a 2.5 X 58 cm G-200 as described under Figure 3
column
The liver and kidney enzymes were isolated trwot4i Steps 5 and 4, respectively, and chromatographed on DE-52 columns as described under Methods. 0.03 U of trypsin-treated liver y-glutamyItransferase and 0.028 U of trypsin-treated kidney -y-giutamyltransferase, each in a volume of 0.01 ml, were electrophoresed in 61/2% gels prepared in 0.3 mol/liter trls/HCI, pH 9.0, by the method of Davis (1?). The gels in this figure were run at the same time and stained for y-glutamyltransferase activity as described under Methods. The gel containing the liver enzyme is labeled with an L, that containing the kidney enzyme is labeled K. Arrows in the photograph indicate the position of the small quantity (,
10
C 6.8
U
N
7.2
-
7.6
8.4
8D
-
88
pH
110
KIDNEY
w 90
70
50
30 jtrsf lo
I/[y- Glutamyl-4-
nitrooni lide] (mmol / I )
68
7.2
Fig. 7. Glutathione inhibition of the formation of 4-nitroaniline of initial velocity (1/v) is plotted vs. the reciprocal of y-glutamyl-4-nitroanlllde concentrations (liter/mmol) at the indicated glutathione concentrations and 20 mmol/llter of glycylglycine The reciprocal
ceptors. Table 3 summarizes the rates of -glutamyl transfer to some amino acids and small peptides, obtained for the isolated hepatic and renal enzymes, with the model substrate ‘y-glutamyl-4-nitroanilide and the natural substrate glutathione (labeled with 3H in the glycine moiety) as the ‘y-glutamyl donor substrates. According to the data in Table 3 the relative y-glutamyl transfer rates for the hepatic and renal enzymes were comparable in the sense that glycylglycine gave the highest rate, followed by L-cystine (this amino acid was not used in the glutathione assays because the dithiothreitol present in these to maintain glutathione in the reduced state would reduce cystine to cysteine), L-glutamine, and L-me-
thionine.
The relationship
between
transfer to glycylglycine,
the rate of y-glutamyl
as well as the rate of the autotransfer reaction (y-glutamyl transfer to the y-gIutamyl donor substrate) to pH was determined for both the hepatic and renal enzymes. These data are illustrated in Figure 8. The pH curves for the two enzymes were very similar, with an optimum pH of 8.1 for the transpeptidation reaction and 8.6 for the autotransfer reactions. Inactivation
Two IUB units (U) of the renal or hepatic enzyme in 2 ml of tris/HCI (0.5 mmolfliter, pH 7.5, in Triton X-100, 5 g/liter), was incubated in a water bath at 58 #{176}C in the absence and in
the presence of 20 mmol of glutathione of these mixtures
‘
8.4
8.8
Fig. 8. The rate of the -y-glutamyltransfer and autotransfer reactions vs. pH At the indicated pH values the final concentration of tris/HCI buffer was 150 mmol/liter. For the -y-glutamyl transfer reaction the final concentration of glycyiglycine was 20 mmol/liter and for both the -gIutamyl transfer and the autotransfer reactions the final concentration of y-glutamyl-3-carboxy-4-nitroanilide was 4 mmol/liter. Each reaction mixture contained 5 mU of enzyme in the case of liver y-glutamyltransferase. 5.6 mU In the case of kidney y-glutamyltransferase. Incubation temperature, 37 #{176}c
the hepatic and the renal ‘y-glutamyltransferase steadily decreased with time in the mixtures lacking glutathione: a 64% decrease in activity after 100 mm for the renal enzyme and a 76% decrease in activity for the hepatic enzyme over the same time interval. In the presence of glutathione, however, we saw no loss of activity for either enzyme throughout the 100-mm incubation period.
Studies
-
from y-glutamyl-3-carboxy-4-nitroanilide
Thermal
80 pH
Lectin-Binding
pH Optima
76
were tested
up to a total incubation
for enzyme
per liter, and portions activity
every 20 mm
time of 100 mm. The activity of both
We indicated in the description of the isolation procedures for ‘y-glutamyltransferase from liver and kidney that the hepatic enzyme binds to concanavalin A-agarose and is eluted from this affinity column with buffer containing a carbohydrate known to specifically inhibit the binding of Con A to
polysaccharide-containing
macromolecules,
namely,
methyl-a-D-mannopyranoside. On the other hand, only a very small proportion of the human kidney enzyme binds to the Con A affinity column under the same experimental conditions. We have studied the interaction of Con A with the hepatic and renal enzymes in another way-namely, by incubating the enzyme with various concentrations of the soluble form of the lectin. As illustrated in Figure 9 (a and b) the hepatic enzyme was precipitated out of solution in increasing amounts as the concentration of Con A was increased. This precipitation is largely prevented by including 0.2 mol/liter methyl-a-D-mannopyranoside in each incubation mixture CLINICAL CHEMISTRY,
Vol. 24, No. 6, 1978
911
b.
a.
4> 4-
0 >‘
8)
E
>
>.‘
0
N C
ILl
8)
E >‘.
N
C
w
[Lectin]
(g/l)
[Lectin}
(g/l)
[Lectin]
(g/l)
d.
C.
800
600
>..
4-
>
4)
E 200
0 [Lectin]
(g/l)
Fig. 9. Lectin binding studies with liver and kidney y-glutamyltransferases 7.Glutamyltransferase actIvities were determined in the stemate after centrifugatlon of reaction mixtures containing the indicated lectins and y-glutamyltransSerase: liver enzyme. (b) untreated kidney enzyme, (C) trypsinized liver enzyme, and (COtrypsinized kidney enzyme. Symbols and the corresponding lectlns: #{149}, concanavalin A; 0, concanavalin A + 0.2 mol of methyl-a-o-mannopyranoslde per liter; Y, Ricinus communisagglutinin 120; V, Ricinus communisagglutinln 120 + 0.2 mol of lactose per lIter; 0, wheat germ agglutinln: , Lotus fetragonolobus agglutinin; 0, soybean agglutinin. All other details of this procedure are as described under Methods.
(a) untreated
912
CLINICALCHEMISTRY.Vol. 24, No. 6, 1978
(Figure 9a). In contrast, only 7% of the renal enzyme is precipitated out of solution by the highest concentration of Con A used, 1 g/liter. We also tested the precipitability of the isolated hepatic and renal enzymes by several other soluble plant lectins. A comparison the graphical data in Figure 9 (a and b) shows other noteworthy differences: (a) The significant precipitation of the kidney enzyme by the Lotus tetragonolobus (asparagus pea) lectin (80%, 1 g/liter) compared to precipitation of only 6.6% of the liver enzyme; the precipitation of the renal enzyme is prevented by including L-fucose (0.2 mol/liter) in the incubation mixture. (b) Precipitation of 73% of the hepatic enzyme, but only 6% of the renal enzyme, by wheatgerm agglutinin. (c) Both enzymes are precipitated by Ricinus communis agglutmnmn 120, but the hepatic enzyme to a greater extent (88% for liver vs. 51% for kidney). The precipitation of the two enzymes by this lectin at a concentration of 1 g/liter is prevented by the inclusion of 0.2 mol of lactose per liter of incubation mixture. In another experiment we tested the precipitability of the small proportion of the total of the renal y-glutamyltransferase that did bind to the Con A-agarose column. This fraction of the total renal -y-glutamyltransferase (20% of the total activity applied to the column) was selectively eluted from the column with the same elution buffer used for the hepatic enzyme containing 0.2 mol/liter methyl-a-D-mannopyranoside. The major difference between this fraction of the renal enzyme and the total renal enzyme preparation used in the above lectin-binding experiments was the fact that at a concentration of 1 g/liter of concanavalin A, 53% of this fraction of the renal enzyme was precipitated and at the same concentration of Lotus tetragonolobus only 19% of the activity was precipitated. The precipitability with the other lectins was found to be comparable to that of the total renal-enzyme preparation. We also studied the interaction of the soluble form of lectins with the enzyme preparations from liver and kidney after incubation with trypsin and chromatography on DE-52 of a portion of each of the same hepatic and renal y-glutamyltransferase preparations used for the lectin-binding studies summarized in Figure 9, a and b. After incubating the trypsinized forms of these enzymes with the soluble plant lectins under the same conditions used for the nontrypsinized form of the enzymes, a much finer precipitate was produced in each case than the more copious precipitate formed with the nontrypsinized forms. At a concentration of 1 g/liter of each respective lectin we observed the following decreases of y-glutamyltransferase activity (Figure 9 c and d): (a) 39% of the hepatic enzyme, but less than 1% of the renal enzyme, by concanavalin A; (b) 57% of the renal enzyme, as compared to less than 1% of the hepatic enzyme, by Lotus tetragonolobus agglutinin; (c) 52% of the liver vs. 8.6% of the kidney enzyme by wheat-germ agglutinin; and (d) 83% of the hepatic enzyme, as compared to 27% of the renal enzyme, by Ricinus communis agglutinin 120.
Discussion We have shown that the initial velocity kinetic constants for -y-glutamyltransferase from human liver and kidney are very similar except for the large difference in the degree of inhibition by the substrate glycylglycine. The inhibition constant, K’0’, for the renal enzyme was sevenfold lower than the value of this constant for the hepatic enzyme with ‘y-glutamyl-4-nitroanilide or its 3-carboxy analog as donor substrate. Furthermore, the relative rates of y-glutamyl transfer to a variety of amino acid, dipeptide, and tripeptide acceptors, the glutathione competitive inhibition constants, and the pH
optima were
of the transpeptidation determined
and
found
and autotransfer to be quite
similar.
reactions We have
shown that both enzyme preparations
were completely
pro-
tected from heat inactivation by including glutathione in the incubation mixtures and that neither enzyme lost catalytic activity after incubation with trypsin for 8 h at 37 #{176}C. For all of these studies, we used the hepatic and renal enzymes prepared by extraction of the particulate fraction from the respective tissue with deoxycholate, followed by several chromatographic steps in the presence of the non-ionic detergent Triton X-100. In previous studies in which the same experimental design, initial velocity equation, and data-reduction procedure were used as described in this work, we obtained the kinetic constants for ‘y-glutamyltransferase from human serum and hog kidney (8, 15). With y-glutamyl-4-nitroanilide as the donor substrate, the values for the human serum enzyme kinetic constants Km’0’, KmD, Km’, and u were 12.5, 1.83, 7.98 mmol/liter and 0.178, respectively, which compare very well with the values obtained for these constants for both the liver and the kidney enzymes. We observed only slight inhibition of the human serum enzyme by glycylglycine (15), which is comparable to our results in the present study for the hepatic enzyme and contrasts with the finding of significant inhibition of the renal enzyme. This finding is compatible with the view that ‘y-glutamyltransferase in human serum originates from
liver but not kidney, even though
human
kidney has a sig-
nificantly greater (about 10-fold, L. M. Shaw, unpublished observation) content of this enzyme as compared to liver (24). However, further detailed studies of the structural properties of ‘y-glutamyltransferase isolated from liver, bile, serum and other tissues are needed in order to firmly establish the origin of the enzyme in normal and pathological human sera. Kidney is considered to be the tissue of origin of the y-glutamyltransferase present in the urine of normal individuals (25). The constants Km’0’ and KmDA for the human hepatic and renal enzymes are rather different from those of the hog kidney enzyme E(Km’0’ and KmD0’ values for the latter enzyme were 24.9 and 2.08 mmol/liter, respectively, with y-glutamyl-4-nitroanilide as the y-glutamyl donor substrate (8)]. Thus, the hog-kidney enzyme exhibited a lesser affinity for glycyiglycine and a greater capability to catalyze the auto!Lransfer reaction with y-glutamyl-4-nitroanilide as the ‘yglutamyl donor substrate than the enzyme from either human
liver or kidney. Tate and Ross (26), in a publication that came to our attention while this manuscript was in preparation, found that human kidney ‘y-glutamyltransferase showed a greater preference for L-glutamine as a -glutamyI acceptor relative to other amino acids as compared to the rat kidney enzyme. These investigators also noted that this greater preference for L-glutamine as compared to the rat kidney enzyme was also shown by the enzyme from human peripheral blood lymphocytes, several lymphoblastic lines, and human seminal fluid (27, 11). In these studies, L-cystine was not included as an acceptor. Our results for human liver and kidney 7-glutamyltransferase are comparable to these findings, in that we observed both enzymes to have a greater preference for Lglutamine as acceptor as compared to the other amino acids tested (excluding L-cystine from the comparison). Thus our
data support
the conclusion
that the enzyme from a variety
of human tissues shows a greater preference for L-glutamine than does the rat-kidney enzyme. In other studies of the ratkidney enzyme in which L-cystine was included, this amino acid was the best acceptor substrate (28,29). Our finding that L-cystine is the best of the amino acids we used as acceptor substrates is in agreement with these results. We have determined that y-glutamyltransferase from human liver has a relative molecular mass greater than 200 000. The high-molecular-weight form of the enzyme
CLINICALCHEMISTRY,Vol. 24,
No. 6, 1978
913
probably results from aggregation. The tendency of this enzyme to form high-molecular-weight aggregates has been observed by others (5, 9). We produced a low-molecularweight form of liver y-glutamyltransferase (Mr ‘-‘ 120 000) by digestion with trypsin. Richter (4) studied the enzyme from human kidney and also observed that trypsinization of the aggregated form produced a low-molecular-weight non-aggregating form of ‘y-glutamyltransferase. The tendency of ‘y-glutamyltransferase to aggregate is probably due to the presence of a hydrophobic peptide in the enzyme molecule, which is removed by proteolytic digestion (9). Such a hydrophobic region is also a property of another microsomal enzyme, namely, cytochrome b5 isolated from rabbit liver (32), proteolytic digestion of which releases an extremely hydrophobic peptide chain of 40 amino acid residues. Perhaps the hydrophobic regions of these enzymes form a tight complex with inner membrane constituents, thus anchoring them to this cell structure. ‘y-Glutamyltransferase obtained from rat kidney (1) and human kidney (26) in purified form by procedures that included proteolytic enzyme treatment was shown to contain hexose, hexosamine, and sialic acid. In the present study of the liver and kidney -y-glutamyltransferases (both trypsintreated and untreated forms), carbohydrate staining of replicate acrylamide gels indicated the presence of carbohydrates. Further definition of the glycoprotein nature of human liver and kidney ‘y-glutamyltransferases was shown by the study of the inhibition of precipitation with lectins by carbohydrates specific for each lectin. The lectin-binding properties of human liver and kidney y-glutamyltransferase suggest that there are significant structural differences in the carbohydrate moieties of these enzymes. That the lectin-binding properties of human liver y-glutamyltransferase are notably different from those of the enzyme from human kidney is based on several experimental results. (a) Approximately 80% of the hepatic enzyme binds to insolubilized Con A and is subsequently eluted with buffer containing methyl-a-D-mannopyranoside. The latter is known to bind to the carbohydrate binding sites of the lectin and in that way inhibit the binding of Con A to mannosyl residues of polysaccharide macromolecules (30). By comparison, under the same experimental conditions, only a small fraction (20%) of human kidney ‘y-glutarnyltransferase was bound to the Con A affinity column. (b) Paralleling the above finding in a is the fact that 89% of the hepatic enzyme was precipitated out of solution by soluble Con A, as compared with precipitation of
bohydrate specific for those lectins, present in either membrane glycoprotein contaminants co-purified with the enzyme or a glycopeptide that is a part of the native enzyme molecule. An alternative explanation for these differences is that a conformational change occurs in the enzyme molecule after trypsinization, which produces a change in the accessibility of some of the carbohydrate moieties to the lectins. The possibility that a glycopeptide was removed by trypsinization is consistent with the findings of Hughey and Curthoys (9), who demonstrated by comparison of molecular weights of rat kidney y-glutamyltransferase prepared by two different methods-one utilizing Triton X-100 throughout all steps but no proteolytic enzyme treatment and the other using proteolytic digestion with papain but no detergent-that a peptide was eliminated from the native enzyme by proteolytic treatment. y-Glutamyltransferase from rat kidney binds very effectively to insolubilized Con A and is eluted with buffer containing methyl-a-D-mannopyranoside (1, 21). Using experimental conditions identical to those for the preparation and Con-A affinity chromatography of liver ‘y-glutamyltransferase, we confirmed this finding, because 97% of the rat kidney enzyme was bound to the column and subsequently eluted with buffer containing methyl-a-D-mannopyranoside. Thus there appears to be a species difference with respect to ability to bind to Con A. This conclusion is supported by the recent report of KOttgen and Lindinger (31), who observed that crude extracts of -y-glutaniyltransferase obtained from the kidneys of several animal species other than rat-namely, dog, pig, and cow-did not bind to Con A affinity columns (only 4 to 6% was bound). In the same study these investigators found that the enzyme extracted from the livers of a variety of species, including man, did bind to Con A affinity columns (between 94 and 97% was bound) and was subsequently eluted from the columns with buffer containing methyl-a-D-mannopyranos-
only 7% of the renal enzyme.
Con A. Furthermore,
The precipitation
of liver -y-
glutamyltransferase was prevented by inclusion of methyla-D-mannopyranoside in the incubation mixture. (c) We have shown that there are very substantial differences in the precipitability of the hepatic and renal enzymes by a variety of other soluble plant lectins. (d) Finally, these differences in lectin-binding properties are retained for the trypsinized form of liver and kidney -y-glutamyltransferase, although the degree of precipitation is less for certain lectins as compared to the non-trypsin treated form of the respective enzyme. For example, Lotus tetragonolobus lectin precipitated 80% of the non-trypsin treated form of kidney ‘y-glutamyltransferase as compared to 57% of the trypsinized form; 73% of the untreated form of the hepatic enzyme was precipitated by wheat-germ agglutinin, as compared to 52% of the trypsinized form; 89% of the intact hepatic enzyme was precipitated by Con A compared to 39% of the trypsinized form; and the percent activity precipitated by Ricinus communis agglutmnin 120 remained virtually the same for the trypsinized form of liver ‘y-glutamyltransferase (83%, compared to 88% for the nontrypsinized form the liver enzyme). The differences observed in the degree of precipitation with some of the lectins may reflect the removal by trypsin of car-
914
CLINICALCHEMISTRY,Vol. 24, No. 6, 1978
ide. Twenty-two per cent of human liver 7-glutamyltransferase did not bind to insolubilized Con A. Although we have not characterized this fraction, Kottgen et al. (33) have described a fetal form of the enzyme from fetal rat liver that does not bind to insolubilized concanavalin A. In these studies it was further shown that after treatment of y-glutamyltransferase from fetal rat liver with neuraminidase the enzyme did bind to insolubilized concanavalin A. The investigators interpret this to mean that the fetal form of the enzyme contained terminal sialic acid residues that sterically hindered binding to
it was shown that the small fraction of
adult liver -y-glutamyltransferase (approximately 5% of the total enzyme activity) which did not bind to the Con A column did bind after neuraminidase treatment. This suggests that a small fraction of the total enzyme from adult liver is the fetal form. Perhaps the fraction of human adult liver ‘y-glutamyltransferase that does not bind to concanavalin A is the fetal form. This question will be the subject of future experimentation in our laboratory.
We thank Drs. David T. Rowlands, Jr. and Pascal J. Viola for their encouragement of this work and Miss Mary Anne Dillon for her excellent assistance in preparing the manuscript.
References 1. Tate, S. S., and Meister, A., Identity of maleate-stimulated glutaminase with -y-glutamyltranspeptidase in rat kidney. J. Biol. Chem. 250, 4619 (1975). 2. Zelazo, P., and Orlowski, M., -y-Glutamyl transpeptidase of sheep kidney cortex. Isolation, catalytic properties and dissociation into two polypeptide chains. Eur. J. Biochem. 61, 147 (1975). 3. Orlowski, M., and Meister, A., Isolation of ‘y-glutamyl transpeptidase from hog kidney. J. Biol. Chem. 240, 338 (1965).
4. Richter, R., Some properties of 7-glutamyl transpeptidase from human kidney. Arch. Immunol. Ther. Exp. 17,476 (1969). 5. Miller, S. P., Awasthi, Y. C., and Srivastava, S. K., Studies of human kidney y-glutamyl transpeptidase. Purification and structural, kinetic and immunological properties. J. Biol. Chem. 251, 2271 (1976). 6. Tate, S. S., and Meister, A., Interaction of ‘y-glutamyl transpeptidase with amino acids, dipeptides, and derivatives and analogs of glutathione. J. Biol. Chem. 249, 7593 (1974). 7. Elce, J. S., and Broxmeyer, B., -y-Glutamytransferase of rat kidney. Simultaneous assay of the hydrolysis and transfer reactions with [glutamate-14C]glutathione. Biochem. J. 153, 223 (1976). 8. London, J. W., Shaw, L. M., Fetterolf, D., and Garfinkel, D., Determination of the mechanism and kinetic constants for hog kidney -y-glutamyltransferase. Biochem. J. 157, 609 (1976). 9. Hughey, R. P., and Curthoys, N. P., Comparison of the size and physical properties of y-glutamy1transferase purified from rat kidney following solubilization with papain or with Triton X-100. J. Biol. Chem. 251, 7863 (1976). 10. Taniguchi, N., Purification and some properties of y-glutamy1 transpeptidase from azo dye-induced hepatoma. J. Biochem. 75,473 (1974).
11. Krishnaswamy, P. R., Tate, S. S., and Meister, A., -y-Glutamyl transpeptidase of human seminal fluid. Life Sci. 20, 681 (1977). 12. Schmidt, E., and Schmidt, F. W., y-Glutamyl transpeptidase. Dtsch. Med. Wochenschr. 98, 1 (1973). 13. Schmidt, F. W., Rationale for the use of enzyme determinations in the diagnosis of liver disease. In Evaluation of Liver Function: A Multifaceted Approach to Clinical Diagnosis, L. Demers and L. M. Shaw, Eds., Urban and Schwarzenberg, Baltimore, Md. In press. 14. Ward, J. P., Urinary gamma-glutamyl transpeptidase, an indicator of renal ischaemic injury and homograft rejection. Br. J. Urol. 47, 765 (1976). 15. Shaw, L. M., London, J. W., Fetterolf, D., and Garfinkel, D., -yGlutamyltransferase: Kinetic properties and assay conditions when 7-glutamyl-4-nitroanihde and its 3-carboxy derivative are used as donor substrates. Clin. Chem. 23, 79 (1977). 16. Bray, G. A., A simple efficient liquid scintillator for counting aqueous solutions in a liquid scintillation counter. Anal. Biochem. 1, 279 (1960). 17. Davis, B. J., Disc electrophoresis-Il. Method and application to human serum proteins. Ann. N.Y. Acad. Sci. 121,404 (1964). 18. Rutenberg, A. M., Kim, H., Fischbein, J. W., et al., Histochemical and ultrastructural demonstration of -y-glutamyl transpeptidase activity. J. Histochem. Cytochem. 17, 517 (1969).
19. Zacharius, R., Zell, T., Morrison, J., and Woodlock, J., Glycoprotein staining following electrophoresis on acrylamide gels. Anal. Biochem. 30, 148 (1969). 20. Lowry, 0. H., Rosebrough, N. J., Farr, A. L., and Randall, R. J., Protein measurement with the Folin phenol reagent. J. Biol. Chem. 193, 265 (1951). 21. Takahashi, S., Pollack, J., and Seifter, S., Purification of ‘y-glutamyltransferase of rat kidney by affinity chromatography using concanavalin A conjugated with Sepharose 4B. Biochim. Biophys. Acta 371,71(1974). 22. Fletcher, R., and Powell, M. J. D., A rapid descent method for minimization. Comput. J. 6, 163 (1963). 23. Davidon, W. C., Variable metric method for minimization, A.E.C. Res. and Dev. Report, ANL-5990 (rev.), U.S. Dept. of Commerce, Washington, D.C., 1959. 24. Rosalki, S. B., Gamma glutamyl transpeptidase. Adv. Clin. Chem. 17,53 (1975). 25. Szasz, G., y-Glutamyl transpeptidase activity in urine. J. Clin. Chem. Clin. Biochem. 8, 1(1970). 26. Tate, S. S., and Ross, M. E., Human kidney y-glutamyl transpeptidase. Catalytic properties, subunit structure and localization of the -y-glutamyl binding site on the light subunit. J. Biol. Chem. 252, 6042 (1977).
27. Novogrodsky, A., Tate, S. S., and Meister, A., -y-Glutamyl transpeptidase, a lymphoid cell-surface marker: Relationship to blastogenesis, differentiation, and neoplasia. Proc. Natl. Acad. Sd.
USA
73, 2414 (1976).
28. Thompson, G. A., and Meister, A., Utilization of L-cystine by the -y-glutamyl transpeptidase--y-glutamyl cyclotransferase pathway. Proc. Natl. Acad. Sci. USA 72, 1985 (1975). 29. Thompson, G. A., and Meister, A., Interrelationships between the binding sites for amino acids, dipeptides, and y-glutamyl transpeptidase. J. Biol. Chem. 252,6792 (1977). 30. Sharon, N., and Lis, H., Uses of lectins for the study of membranes. Methods Membrane Biol. 3, 147 (1975). 31. Kottgen, E., and Lindinger, G., Detection of multiple molecular forms of -y-glutamyltransferase by concanavalin A affinity chromatography. Z. Physiol. Chern. 357, 1439 (1976). 32. Spatz, L., and Strittmatter, P., A form of cytochrome b5 that contains an additional hydrophobic sequence of 40 amino acid residues. Proc. NatI. Acad. Sci. USA 68, 1042 (1971). 33. Kottgen, E., Reutter, W., and Gerok, W., Two different gammaglutamyltransferases during development of liver and small intestine: A fetal (sialo-) and an adult (asialo-) glycoprotein. Biochem. Biophys. Res. Commun. 72,61(1976).
CLINICALCHEMISTRY,Vol. 24, No. 6, 1978 915
