THE JOURNAL or Bmmcrc~r. CHEMISTBY

Vol. 252, No. 4, Issue of February 25, PP. 1336-1343, 1977 Printed in U.S.A.

Purification and Properties of Microsomal UDP- Glucuronosyltransferase from Rat Liver*

(Received for publication, May 27, 1976, and in revised form, October 21, 1976)

JEFFREY P. GORSKI AND CHARLES B. KASPER

From the McArdle Laboratory for Cancer Research, University of Wisconsin, Madison, Wisconsin 53 706

p-Nitrophenol conjugating activity associated with liver microsomal UDP-glucuronosyltransferase (EC 2.4.1.17) was purified 150- to 200-fold from cell-free homogenates. The purification scheme included solubilization with the non- ionic detergent Lubrol WX, anion exchange chromatogra- phy at pH 6.0 and 7.5, and affinity chromatography with UDP-hexanolamine Sepharose 4B. The enzyme purified as a phospholipid. protein complex and was shown to consist of a single polypeptide chain of molecular weight 59,000 by so- dium dodecyl sulfate-polyacrylamide gel electrophoresis. Amiiio acid analysis indicated approximately 531 mol of amino acids/59,000 g of enzyme and a molar ratio of nonpo- lar to polar residues of 1.08. During fractionation, the en- zyme displayed instability with such steps as gel filtration, dialysis, or ultrafiltration of dilute samples; however, upon adsorption to ion exchange resins or storage in concentrated form, the enzyme was reasonably stable.

The active lipoprotein complex showed both size and charge heterogeneity as judged by gel filtration and electro- focusing. Three forms of the enzyme resolved by isoelectric focusing had isoelectric points which averaged pH 6.68,6.56, and 6.31. Polypeptide compositions of these electrophoreti- tally distinct phospholipid . protein complexes were indistin- guishable on the basis of sodium dodecyl sulfate-polyacryl- amide gel electrophoresis, suggesting that the charge heter- ogeneity may be the result of differences in the phospholipid content of the lipoprotein complex.

UDP-glucuronosyltransferase is present in mammalian liver, kidney, skin, gastrointestinal tract (2), and lung (3). The enzyme catalyzes transfer of n-glucuronic acid from UDP-WD glucuronic acid to an acceptor compound with inversion at the C-l carbon of the sugar to yield the P-glucuronide (2). Endoge- nous acceptor substrates include bilirubin, thyroxine, and es- tradiol (2), whereas N-hydroxyacetylaminofluorene (4) and diethylstilbesterol (5) are examples of two exogenous sub- strates. UDP-glucuronosyltransferase functions in close asso- ciation with the microsomal mixed function oxidase system in that many of the hydroxylated products are substrates for glucuronidation.

* This investigation was supported by Grants CA-07175 and CA- 17300 from the National Cancer Institute. A preliminary account of this work was presented at the 1974 American Society of Biological Chemists Meeting in Minneapolis, Minnesota (1).

Liver subcellular fractionation studies of Isselbacher (6) and Beaufay et al. (7) have shown that the majority of transferase activity resides in the microsomal fraction; however, recent studies have also shown the transferase to be an integral component of the nuclear envelope.’ Efforts to purify the en- zyme from several sources have been partially successful (8- ll), but problems have been encountered in obtaining a solu- ble, stable, homogeneous preparation. Consequently, little is known about the molecular nature of the enzyme and present biochemical knowledge is based primarily on experiments per- formed with the membrane-bound form. A variety of condi- tions and agents are known to influence the glucuronidation reaction. For example, activation of microsomal transferase is obtained in vitro by sonication, alkaline pH, detergents, lim- ited phospholipase A or C digestion, storage at -20” (12-21) and exposure to organic solvents.* In many instances, the activations are not additive (14).

The question of transferase heterogeneity remains uncer- tain, but the wide chemical diversity of compounds capable of being glucuronidated does not necessarily argue for polymor- phism. As noted by Axelrod et al. (22), all of the substrates possess a nucleophilic functional group and one enzyme could catalyze a nucleophilic substitution reaction between all ac- ceptor compounds and UDP-glucuronic acid. The sequential kinetic mechanism determined for bilirubin andp-nitrophenol conjugation (20, 23) is consistent with such a mechanism. However, differences among enzymic results with various sub- strates remain to be explained in terms other than transferase polymorphism. For example, (a) phenolphthalein transferase activity is divided equally between both smooth and rough microsomes, but p-nitrophenol conjugating activity is found largely in the smooth surfaced vesicles (24); (b) the develop- mental appearance of p-nitrophenol transferase activity pre- cedes that for o-aminophenol and phenolphthalein in rat fe- tuses (25). The uncertainties of assaying a membrane-bound enzyme in crude preparations, as pointed out by Coleman (26), make this evidence for multiplicity less than definitive. In addition, earlier work (27-30) has suggested the possibility that substrate specificity may be modified by changes in the phospholipid environment of the enzyme.

Our approach to studying many of these interesting ques- tions has involved direct isolation and purification of p-nitro- phenol conjugating UDP-glucuronosyltransferase and elucida- tion of its molecular properties.

1 J. Gorski and C. B. Kasper, manuscript in preparation. ’ J. Gorski, unpublished observation.

1336

by guest on May 3, 2020

http://ww

w.jbc.org/

Dow

nloaded from

Purification of UDP-Glucuronosyltransferase 1337

MATERIALS AND METHODS

UDP-glucuronic acid, p-nitrophenol, dithiothreitol, P-mercapto- ethanol, and Tris base were purchased from Sigma Chemical Co. Organic chemicals used in the synthesis of UDP-hexanolamine were at least 95% pure and the majority were obtained from Aldrich Chemical Co., Inc., Supelco Inc., and Mann Research Laboratories Inc. were the sources of Lubrol WX and sodium deoxycholate, respec- tively, while Triton detergents were the gift of Rohm and Haas Co. Sodium dodecyl sulfate was recrystallized twice from 95% ethanol prior to use. Acrylamide and DEAE-agarose were purchased from Bio-Rad Laboratories; bisacrylamide was obtained from Eastman Kodak Co. Ampholytes were obtained from LKB Instruments and Pharmacia Fine Chemicals was the source for Sephadex and Sepha- rose 4B. All other chemicals used in this work were analytical reagent grade.

Rats were obtained from Sprague-Dawley Farm, Madison, Wis. Preparation of Gradient-purified Microsomes -Induction of rats

with 0.12 g/kg of phenobarbital was carried out as previously re- ported (31), except that 150- to 175-g rats were used throughout this work. Rats were maintained on a 12 h light-12 h dark schedule (321, with injections given 3.5 h into the light period. Food and water were available at will except that animals were fasted 24 h prior to killing, and all operations following liver excision were at O-4”. Fresh liver was homogenized in 0.25 M sucrose, 0.05 M TKM” buffer. After filtration through four layers of cheesecloth, the homogenate was centrifuged at 20,000 x g,,, for 15 min. The lipid plaque was re- moved by aspiration, and the supernatant was centrifuged at 105,000 x g,,, for 90 min. The crude microsomal pellet was sonicated in 10% (w/v) potassium citrate, 0.05 M TKM buffer, and the density was adjusted to greater than 1.20 g/ml with 2.3 M sucrose/TKM buffer containing 10% (w/v) potassium citrate. This suspension was over- laid in a stepwise manner with sucrose/TKM/citrate solutions hav- ing densities of 1.20, 1.18, and 1.16 g/ml. The discontinuous gradient was developed by centrifugation for 16 h at 105,000 x g,,,; the membrane localized in the 1.16 and upper half of the 1.18 g/ml layers, and was isolated by lateral puncture. After dilution with water, membrane was pelleted by centrifuging at 105,000 x g,,, for 2 h; the pellet was sonicated in 20 rnM Trislacetate buffer, pH 7.5, and again sedimented at 105,000 x g,,, for 2 h. This membrane pellet, sonicated in 20 mM Tris/acetate (pH 7.5) containing 0.2 mM EDTA, was distributed into tubes at a concentration of approximately 30 mg/ml of protein. The contents of the tubes were flushed with nitro- gen gas, tightly stoppered, and frozen at -20”. The yield of micro- somal membrane was approximately 11 mg/g wet liver.

Total protein was determined with the Lowry method (331 using ovalbumin as standard.

Purification of UDP-glucuronosyltransferase -All of the following operations were performed at O-4” and 0.02% sodium azide was present in all buffers.

Microsomal membrane, 245 to 265 mg of protein, was dispersed by sonication in 50 mM Tris/acetate buffer, pH 7.5, containing 1 M urea to give a protein concentration of 10 mg/ml. The resulting suspension was centrifuged at 326,000 x g,,, for 60 min, and the pellet fraction containing approximately 91% of the transferase activity, was solu- bilized in 20 ml of 20 mM Tris/acetate, pH 7.5, containing 4% (w/v) Lubrol WX and 0.5 mM dithiothreitol. After dilution B-fold with 20 mM Tris/acetate (pH 6.0) containing 0.5 mM dithiothreitol and mak- ing a final adjustment to pH 6.0, the solubilized membrane was applied to a chromatographic column (1.9 cm diameter) containing 50 ml of settled DEAE-agarose equilibrated with 20 rnM Tris/acetate (pH 6.0) containing 0.5 mM dithiothreitol. The column was subse- quently washed with 2 bed volumes of buffer and the flow-through volume containing the transferase activity was collected. The pH was then adjusted to 7.5 and solid magnesium chloride was added to a final concentration of 5 mM. The sample was applied directly to a bed (3.8 x 14 cm) of UDP-hexanolamine Sepharose 4B equilibrated with 20 rnM Tris/acetate buffer, pH 7.5, containing 5 rnM magnesium chloride and 0.5 mM dithiothreitol. Unbound proteins were removed from the affinity column by washing with 350 ml of this buffer. With a pressure head of 40 cm, the column was eluted in two steps: (a) 350 ml of 20 rnM Tris/acetate, pH 7.5, containing 25 mM EDTA and 0.5 rnM dithiothreitol, was filtered through the affinity column followed

n The abbreviations used are: TKM, 50 rnM Tris/HCl buffer, pH 7.5, containing 25 rnM potassium chloride, and 5 mM magnesium chloride; SDS, sodium dodecyl sulfate; UDP-hexanolamine, P’-(6- amino-1-hexyll-P2-(5’-uridine)-pyrophosphate.

by (b) 350 ml of 20 mM Tris/acetate, pH 7.5, containing 5 mM UDP- glucuronic acid, 0.5 mM dithiothreitol, and 0.05% Lubrol WX (w/v). Fractions of 11 to 12 ml were collected both in Steps a and 6, and the contents of all fractions having absorbance at 280 nm were pooled into two respective fractions. The pooled UDP-glucuronic acid wash containing the majority of the enzymic activity was then dialyzed for 11 h in Nojax dialysis tubing against 7 liters of 10 mM Trislacetate, pH 7.5, containing 0.5 rnM dithiothreitol and 0.05% (w/v) Lubrol WX (Buffer A). After dialysis, the volume measured 240 ml, and the concentration of UDP-glucuronic acid was between 2.5 and 3 mM. Moist DEAE-agarose gel (12 ml) equilibrated with Buffer A was added and the slurry was mixed gently for 1 h, after which time it was centrifuged for 5 min at 280 x g,,, The supernatant containing the active enzyme was removed and the pellet was resuspended in Buffer A, transferred to a chromatographic column, washed with 2 bed volumes of Buffer A, and the washes combined with the superna- tant fraction. In order to remove UDP-glucuronic acid from the preparation, the purified transferase was applied to a column (6 x 42 cm) of Sephadex G-25 (coarse) equilibrated with Buffer A. After gel filtration, the enzyme was concentrated by adsorption to a column (1.9 x 6 cm) of DEAE-agarose equilibrated with Buffer A and eluted with a small volume of Buffer A containing 0.2 M sodium chloride. Purified enzyme, at a concentration of at least 0.25 mg/ml, was stored at 4” in a nitrogen atmosphere.

UDP-glucuronic acid may be reclaimed after separation from the enzyme. The appropriate Sephadex G-25 fractions were pooled and then concentrated to dryness by rotary evaporation. Lubrol WX was removed from the residue by several extractions with chloroform at room temperature. UDP-glucuronic acid was then desalted on a small column of Sephadex G-10 at 4”.

In preparation for amino acid analysis, the enzyme was precipi- tated by the addition of acetone to a final concentration of 70% (v/v) at 0”. After centrifugation, the protein pellet was washed with cold ethanol (95% v/v), transferred to a thick-walled glass hydrolysis tube, and taken to dryness. The protein was hydrolyzed as previ- ously described (34) and amino acid analysis was carried out accord- ing to Spackman et al. (351, except as modified for single column analysis by Beckman Instruments, Inc. (Publication A-TB-059Al. On-line quantitation was provided by an Autolab AA computing integrator.

Kinetic Assay of Glucuronosyltransferase -The degree of conjuga- tion was based on the disappearance of p-nitrophenol absorption in the visible region (61. Assays were carried out in a cuvette contain- ing 6 rnM UDP-glucuronic acid, 0.6 mxlp-nitrophenol, 20 mM magne- sium chloride, 50 rnrvr Tris/acetate, pH 7.5, and enzyme in a final volume of 1.0 ml. After a 2.5-min temperature pre-equilibration at 24”, the reaction was followed at 440 nm with a Cary model 15 recording spectrophotometer; blanks were composed of enzyme and magnesium chloride in Tris buffer. The rate of glucuronide forma- tion was calculated directly from the absorbance change.

Single Time Point Assay for Glucuronosyltransferase - Samples were assayed for enzymic conjugation at 24” in a 0.5-ml volume containing 6 rnM UDP-glucuronic acid, 0.6 mM p-nitrophenol, and 20 mM magnesium chloride in 50 rnM Trislacetate buffer, pH 7.5. Blanks were of the same composition except that UDP-glucuronic acid was omitted. Reactions were stopped by addition of an equal volume of cold 95% ethanol before they had reached 25 to 30% completion and precipitated protein was removed by low speed cen- trifugation. Supernatants were appropriately diluted with 0.1 M

EDTA, pH 9.6, and the difference in absorbance at 400 nm between blanks and reaction mixtures noted. A molar extinction coefficient for p-nitrophenol of 18,200 was used to calculate the amount of phenol conjugated.

Preparation of UDP-Hexanolamine Sepharose 4B - UDP-hexano- lamine was synthesized according to the published procedure of Barker et al. (361. This ligand was coupled to the Sepharose 4B support with the cyanogen bromide method of activating cross- linked dextrans (371; 300 mg of cyanogen bromide/ml of settled Sepharose was used. Activated Sepharose was added as a moist gel to an equal volume of 25 mM UDP-hexanolamine in 0.1 M sodium carbonate buffer, pH 10.0. After the coupling reaction had proceeded for 18 h at 4”, free ligand was separated from that covalently bound by washing the Sepharose with 60 volumes of 0.1 M sodium carbonate buffer, pH 9.0, saturated with chloroform. Unreacted UDP-hexano- lamine was reclaimed after this wash had been desalted, using Sephadex G-10. Quantitation of covalently attached ligand was by direct phosphorus analysis following acid digestion; aliquots of the same suspension were also dried and weighed. The concentration of

by guest on May 3, 2020

http://ww

w.jbc.org/

Dow

nloaded from

1338 Purification of UDP-Glucuronosyltranwferase

bound ligand was calculated by using 31.2 mg dry weight/ml of wet, settled Sepharose 4B. Unless otherwise noted, affinity chromatogra- phy was carried out with resin which contained 2.0 rn~ UDP-hexan- olamine.

Isoelectric Focusing-Isoelectric focusing was carried out accord- ing to the method of Vesterberg (381, employing a llO-ml column. Operating temperature was maintained at 5”; 1% (w/v) ampholyte was routinely used. The sample, 2 to 4 mg of protein in 20% (w/v) sucrose and 1% (w/v) ampholvte, was introduced into the gradient at the appropriate point. Electrophoretic runs were of 3-to 4 days duration at a constant voltage of 500 V. Column fractions of approxi- mately 0.8 ml were collected and assayed for transferase activity at room temperature by the single time point assay procedure.

Isoelectric focusing was also carried out in polyacrylamide gels (39), 10 x 0.5 cm in size. The sample in 10% (w/v) sucrose was layered beneath 200 ~1 of 5% (w/v) sucrose and 2.5% (w/v) ampholyte to protect it from alkaline denaturation by the upper electrode solution (40). Gels were equilibrated at 4” prior to and during the 12-h electrophoretic run. Determination of transferase activity in gels was as follows: total length of eel was noted, the nel was sliced - I - manually or with a Gilson Aliquogel slicer, and the pieces were placed into 0.1 M Tris/acetate buffer (pH 7.5) which contained 3 mM CDP-glucuronic acid, 0.6 rnM p-nitroihenol, 5 rnM magnesium chlo- ride, and 0.02% sodium azide. Assays were developed at room tem- perature for 24 to 48 h. The pH gradient was determined with extracts prepared by incubating 0.5-cm sections of gel in 0.6 ml of carbon dioxide-free water.

Other Methods -Phosphorus was determined with the Bartlett method (41). Polyacrylamide disc gel electrophoretic studies used the standard SDS svstem (34) and stacking SDS svstem (42). Proteins were visualized” in gels by staining &h 0.09% (w/v) Coomassie brilliant blue in water/ethanol/acetic acid (45/45/10). The method for calculating molecular weights of proteins after SDS-gel electropho- resis was that of Fairbanks et al. (43). The standard proteins used were lysozyme, chymotrypsinogen, pepsin, ovalbumin, bovine serum albumin, and human transferrin. Molecular weight values are the average of multiple determinations.

RESULTS

Gradient-purified liver microsomal membrane from pheno- barbital-treated rats was used throughout this work as a source of UDP-glucuronosyltransferase. Two factors recom- mend use of this preparation. First, microsomal membrane prepared by the sucrose/citrate discontinuous density gradient procedure has an average specific activity of 56 nmol/min/mg which is approximately 48% higher than non-gradient-purified preparations (44). Secondly, phenobarbital treatment en- hanced microsomal enzyme activity another 37% and, impor- tantly, also increased the yield of membrane per g of liver about 4-fold (45, 46). Transferase activity is completely stable for up to 3.5 months when the membrane is stored at -20.

Solubilization Studies - Because the instability of glucuron- osyltransferase has been a major obstruction to its characteri- zation, our approach to this problem was to use the mildest available methods for solubilization and fractionation. Conse- quently, the effects of varying concentrations of several deter- gents (Triton X-100, Triton X-114, Triton X-165, Lubrol WX, and deoxycholate) and urea on transferase activity were exam- ined and assays were conducted at both 37” and 23”. For the nonionic detergents, enzymic activity was generally higher at 23” than at 37”; this was particularly notable in the case of Lubrol WX. The greatest stimulation of activity (208% of control) was obtained with 0.04% deoxycholate at a detergent to protein ratio of 2.4, although above this level activity gradu- ally decreased until complete inhibition resulted at 0.14%. Urea progressively inhibited enzyme activity above 0.8 M;

however, only a marginal temperature effect was noted. A major point of importance is that gradient-purified microsomal membrane used in these studies does not exhibit the marked Lubrol activation of UDP-glucuronosyltransferase reported for

less purified microsomal fractions (47). Starting with a micro- somal preparation having a specific activity of 1.75 (nanomoles of p-nitrophenoliminimg) Howland et al. (47) found a lo-fold increase in activity upon treatment with Lubrol. Even with this striking increase, however, the specific activity was only one-third that obtained for citrate gradient-purified mem- brane. The results indicate that Lubrol activation was most likely not the result of a direct detergent-enzyme interaction but instead a modification of the environment of the enzyme in the membrane.

Neither urea nor detergents provided a selective extraction of enzymic activity, and consequently no significant purifica- tion was achieved by this approach. The method of solubiliza- tion did, however, dramatically influence transferase stabil- ity. For example, the half-life of the enzyme was 13, 48, and 1 to 2 days in the case of Lubrol WX, 1 M urea, and deoxycho- late, respectively. Even though transferase solubilized by 1 M

urea displayed a significantly longer half-life than that ob- served with Lubrol WX (48 days versus 13 days), urea released less than 50% of the total activity into the supernatant (105,000 x g for 60 min), in contrast to 87% with Lubrol WX. On the basis of this latter distinction and the fact that the enzyme exists as a larger aggregate after urea solubilization

than after detergent treatment, Lubrol WX was selected as the solubilizing agent of choice. Results with Triton X-100 were similar to those with Lubrol WX, but since the latter has a reduced absorption at 280 nm, it allowed for more sensitive ultraviolet monitoring of protein during fractionation in the presence of detergent.

Purification of Solubilized UDP-Glucuronosyltransferase- p-Nitrophenol conjugating activity was purified 25-fold using the five-step procedure outlined in Table I. Gradient-purified microsomal membrane represents an enzyme enrichment of 6- to g-fold over liver homogenate and provides an excellent

source of enzyme for further purification. Extraction with 1 M

urea (48) followed by centrifugation for 60 min at 326,000 x g,,, releases predominantly loosely bound membrane proteins not associated with transferase activity. Solubilization of the extracted residue with Lubrol WX yielded a heterogeneous population of different size classes of active phospho- lipid. protein complexes (Fig. 1). Ion exchange chromatogra- phy on DEAE-agarose of the Lubrol WX-solubilized mem- brane matrix followed by affinity chromatography with UDP- hexanolamine Sepharose 4B accounted for almost half of the total increase in specific activity. In the latter step 70% of the

applied transferase units are bound to UDP-hexanolamine Sepharose 4B, while only 13% of the total protein is retained. Initially, direct elution with 5 mM UDP-glucuronic acid was attempted in order to selectively release glucuronidating ac- tivity. Although enzyme was eluted under these conditions, the process was nonspecific, as several other proteins presum- ably bound by ionic or hydrophobic association to the chro- matographic support were also removed. The protocol finally adopted was a two-step procedure in which the affinity column was first washed with 25 mM EDTA to remove the bulk of the contaminating polypeptides and then with 5 mM cofactor (Fig. 2). One disadvantage to this sequential elution scheme is the fact that close to one-third of the bound transferase activity eluted with the first wash. It is interesting to note that the amount of activity removed in the EDTA wash is approxi- mately equivalent to the increase in units bound in the pres- ence of magnesium. Thus, it appears that two different modes of binding occur or that two populations of transferase mole- cules exist. Phospholipid co-eluted with protein in both the

by guest on May 3, 2020

http://ww

w.jbc.org/

Dow

nloaded from

Purification of UDP-Glucuronosyltransferase 1339

EDTA and UDP-glucuronic acid washes (Fig. 2). The phospho- rus/protein ratios of pooled Fractions 50 to 75 (EDTA wash) and 77 to 95 (UDP-glucuronic acid wash) were 45 and 79% that of the initial membrane, or 14.3 and 25.2 pg of phosphoruslmg of protein, respectively. It is possible to substitute 0.6 M so- dium chloride or sodium sulfate for the nucleotide sugar and obtain similar results. An examination of the polypeptide composition of the EDTA and UDP-glucuronic acid affinity column washes by SDS-polyacrylamide gel electrophoresis yielded the results presented in Fig. 3 (left side), Gels A and B. A common M, = 59,000 protein is seen in both samples, but the EDTA wash also contained a major component of M, = 52,600 as well as some proteins of lower electrophoretic mobil- ity. It should be noted that the specific activity of the enzyme eluted with EDTA was approximately one-half that for the enzyme eluted subsequently with UDP-glucuronic acid. This

TABLE I

Purification of UDP-glucuronosyltransferase from rat liver microsomes

Methods Transferase Apparent unit.9 Protein pWifk3-

tion

1. Phenobarbital-induced, gradient-purified micro- somal membrane

2. 1 M urea extraction Supernatant Pellet

3. DEAE-agarose flow- through, pH 6.0

4. Affinity chromatography on UDP-hexanolamine Sepha- rose 4B a. 0.025 M EDTA wash of

affinity column b. 0.005 M UDP-glucuronic

acid wash of affinity col- umn

5. DEAE-agarose flow throuah. DH 7.5c

16,220

2,800 51 20,910 183 1.2 15,330 a7 1.86

3,540 W8OP 7,250

13,440l

4,830 19301

m&T

248

-fold

6.3

5.2 14.7

2.0 25.0

D Expressed as nanomoles of p-nitrophenol conjugated/min. L Numbers in brackets are uncorrected activity values. c Dialysis for 11 h prior to this ion exchange step resulted in a 59%

loss of enzyme units This loss may be circumvented by maintaining the protein concentration in the range of 2 to 4 mg/ml.

increase in specific activity is consistent with the near total absence of the M, = 52,600 protein in the nucleotide sugar wash.

The last step of the procedure was passage through a DEAE- agarose column, pH 7.5, in the presence of approximately 3 mM UDP-glucuronic acid. This step yielded a l&fold increase in specific activity and was necessary to remove traces of contaminating proteins. The substrate was then removed by Sephadex G-25 chromatography, and the purified transferase concentrated by adsorption to DEAE-agarose, pH 7.5, and elution in a small volume with 0.2 M sodium chloride as eluent.

Glucuronosyltransferase purified by the procedure in Table I exhibits only one major band after SDS-polyacrylamide gel electrophoresis (Fig. 5, left side, Gel C) with a calculated molecular weight of 59,000. The amino acid composition of UDP-glucuronosyltransferase is presented in Table II and its (amino acid) content is compared to that of intact microsomal membrane. An interesting feature is the rather close resem- blance of the amino acid composition of the purified enzyme when expressed in mole per cent with that of the membrane matrix from which it was derived. On the basis of the subunit molecular weight of the transferase, there are approximately 531 amino acid residues, of which 48 mol % are polar residues and 52 mol % nonpolar. The fact that methionine is present at the level of 14 residues suggests that cyanogen bromide degra- dation would be a fruitful means of further studying the primary structure of the enzyme, since only 15 fragments would be expected.

Throughout the purification procedure, enzymic activity ex- hibited marked instability when manipulations were carried out at dilute protein concentrations (approximately 30 pg/ml). For example, only 40% of the initial activity was recovered after dialysis for 11 h prior to Step 5 and similar losses were experienced during gel filtration and membrane ultrafiltra- tion. The enzyme after concentration by adsorption to and desorption from DEAE-agarose was stable for at least 2 days at 4”.

The degree of enzyme purification by affmity chromatogra- phy was based on the number of transferase units bound to the column and on the relative distribution and recovery of activ- ity in the UDP-glucuronic acid and EDTA washes after elu- tion. That is to say, for the experiment cited, of the 15,330 units applied to the affinity column, 30% of the total enzymic activity was not adsorbed. We have assumed that the transfer-

FRACTION NUMBER

FIG. 1. Fractionation of Lubrol WX- solubilized microsomal membrane on Sepharose 4B. Twelve milliliters of a 0.8% protein solution was applied to a column (5 x 89 cm) of Sepharose 4B equilibrated with 50 mM Tris/HCl, pH 7.5, containing 0.1 mM EDTA and 0.5 mM dithiothreitol and fractionated by upflow gel filtration at 4”. A flow rate of 42 ml/h was maintained and frac- tions of 13 ml were collected. Phospho- rus content and transferase activity are expressed as micrograms/ml and as the change in absorbance at 400 nm divided by the corrected 280 nm absorb- ance of the respective fraction.

by guest on May 3, 2020

http://ww

w.jbc.org/

Dow

nloaded from

1340 Purification of UDP-Glucuronosyltransferase

FRACTION NUMBER

FIG. 2. Affinity chromatography of UDP-glucuronosyltransferase on UDP-hexanolamine Sepharose 4B. Lubrol WX-solubilized micro- somes from Step 3, Table I, were applied to a column (3.8 x 14 cm) of UDP-hexanolamine Sepharose. Conditions for elution are described under “Materials and Methods”; the arrows indicate the points at which designated changes in composition of the column buffer were made. Fractions of 11.5 ml were collected during adsorption and elution of the affinity column, and a pressure head of 20 cm was maintained throughout. UDPGA , UDP-glucuronic acid.

A B C A B c

163,000 -

86,500~

59,000-

52,600-

dye front-

FIG. 3. Left panel, SDS-polyacrylamide gel electrophoresis of UDP-glucuronosyltransferase at various stages of purification. Gels A and B are representative of the polypeptides eluted with EDTA and UDP-glucuronic acid, respectively, from the affmity column as illustrated in Fig. 2. Gel C is the pattern of UDP-glucuronosyltrans- ferase at the final stage of purification. Right panel, comparison of polypeptide compositions of the three transferase forms isolated after isoelectric focusing. Fractions of the transferase activity peaks at pH 6.69 (Gel A), pH 6.56 (Gel B), and pH 6.35 (Gel 0, shown in the lower half of Fig. 5, were subjected to sodium dodecyl sulfate-poly- acrylamide gel electrophoresis with the standard gel system, and the molecular weight values calculated as described under “Materials and Methods.”

ase retained by the column is distributed between Washes a and b in proportion to the activity recovered in each wash (figures in brackets, Table I). On the average, 47% of the bound enzyme units were recovered in the two-step elution process. However, recovery of protein in the initial drop- through (Fractions 3 to 30, Fig. 2) and the two washes was close to quantitative (i.e, these three fractions account for essentially all of the protein applied to the column). Failure to recover all of the bound transferase activity is most likely due to instability of the purified transferase in dilute solutions, as already mentioned, or to a change in one or more of its kinetic parameters, or both. The recovery of transferase units in Steps 4 and 5 (Table I) has been corrected (unbracketed values) for these losses. Several control experiments justify our making

TABLE II Amino acid composition of purified UDP-glucuronosyltransferase

Amino acid” Transferase Microsomal membrane*

Lysine 6.45 Histidine 2.51 Arginine 4.69 Aspartic acid 9.10

Threonine 4.81 Serine 6.67 Glutamic acid 9.99 Proline 5.58 Glycine 8.00 Alanine 6.21 Half-cystine 0.51 Valine 6.76 Methionine 2.59 Isoleucine 5.28 Leucine 11.53 Tyrosine 3.45 Phenvlalanine 5.81

mol 96 mollmol en. zymec

34.2 13.6 24.8 48.2 25.5 35.3 52.9 29.5 42.4 32.9

2.7 35.8 13.7 28.0 61.1 18.3 30.8

mOl%

6.46 2.18 4.79 9.60 5.39 6.48

10.70 6.17 7.06 7.34

6.62 2.53 4.91

11.10

3.52 5.26

o Cysteiue and tryptophan were not determined. b Data taken from Ref. 48. Half-cystine content was not deter-

mined. c Calculated on the basis of a M,, = 59,000 polypeptide chain.

this correction. (a) Greater than 94% of the activity applied to an underivatized Sepharose 4B column was recovered in the effluent fraction, suggesting that the act of passage through the column did not lead to loss of activity, and (b) an average 85% of the bound activity was recovered from affinity columns loaded to near-saturation. That is to say, it is possible to significantly increase the total number of transferase units bound to the affinity column by a factor of 3 or 4 over that reported in Table I by simply applying proportionately more of the solubilized enzyme. Elution by the conventional two-step procedure resulted in the retrieval of 85% of the total units retained by the column in comparison to the 47% recovery just described. Although at first inspection this protocol would appear to markedly improve the efficiency of the affinity chro- matographic step, a severe limitation was that the binding of extraneous protein increased disproportionately to the reten- tion of glucuronosyltransferase. Consequently, the degree of purification achieved with saturated affinity columns was sub- stantially reduced and SDS-gel electrophoresis of eluted frac- tions demonstrated a complex polypeptide composition.

Isoelectric Characterization of Affinity Column-purified Transferase - Isoelectric focusing of UDP-glucuronosyltrans- ferase purified as described in Table I was carried out in polyacrylamide gels of varying concentrations (Fig. 4). Two peaks of transferase activity centered at pH 6.85 and pH 6.27, respectively, were resolved in the 3% polyacrylamide gel sys- tem. A similar separation was obtained in the 4% gel with the two electrophoretic forms being localized at pH 6.93 and 6.33; however, one important difference was the appearance of a small percentage of active enzyme at the gel surface. As the gel concentration was raised to 5%, an increased amount of activity was found at the gel surface, and the level of transfer- ase with the lower isoelectric point was significantly reduced. In summary, two peaks of transferase activity were generally observed after polyacrylamide gel isoelectrofocusing. The ab- solute isoelectric points in the various gel systems fluctuated within a relatively narrow range and was most likely the

by guest on May 3, 2020

http://ww

w.jbc.org/

Dow

nloaded from

Purification of UDP-Glucuronosyltransferase 1341

result of subtle changes in phospholipid composition of the lipoprotein complex.

Isoelectric focusing of partially purified transferase (equiva- lent to the EDTA wash from the affinity column) was also carried out in sucrose gradients. Two representative experi- ments are shown in Fig. 5. The purpose of these experiments was to determine the effectiveness of preparative isoelectric focusing as a possible purification step as well as to measure isoelectric points in a non-gel system. In the experiment illus- trated in the upper half of Fig. 5, conjugating activity was distributed over a rather wide pH range with principally two

(j/- 3XGEL 1’

h- 4

DISTANCE, km)

FIG. 4. Polyacrylamide gel isoelectrofocusing of affinity column- purified glucuronosyltransferase. Enzyme was focused in gels for 12 h at 4” and the activity in gel slices was measured as described under “Materials and Methods.” ‘I’ransferase activity is expressed as the difference in p-nitrophenol absorbance at 400 nm between assay mixtures and reaction blanks; pH and activity values are illustrated as that present at the center of the gel slice.

6

5

4

3

FRACTION NUMBER

FIG. 5. Column isoelectrofocusing of affinity column-purified glu- curonosyltransferase. Lubrol WX-solubilized enzyme was electrofo- cused for 72 h at 5”. Conditions of these procedures are described under “Materials and Methods.” Transferase activity is expressed as the difference in p-nitrophenol absorbance at 400 nm between assay mixtures and reaction blanks. Numbers above the peaks of activity refer to their respective p1 values. The profiles shown in the upper and lower halves of the figure represent the results from two similar experiments.

peaks centered at pH 6.66 and 6.27. This broad distribution within the gradient could be due to the fact that glucuronosyl- transferase tends to form insoluble complexes which slowly settle during focusing. When a different enzyme preparation was subjected to similar electrophoretic conditions, the profile in the lower half of Fig. 5 was obtained. In this case, three distinct peaks of activity were resolved with isoelectric points of 6.68, 6.56, and 6.35. The polypeptide compositions of these three active fractions were examined by polyacrylamide gel electrophoresis and were found to be essentially identical (Fig. 3, right side). Each was composed of a prominent doublet at M, = 59,000 and 52,600 along with minor bands at M, = 86,500 and 163,000. The gel patterns clearly show that isoelectric focusing does not separate the M, = 52,600 protein from glucu- ronosyltransferase and, furthermore, does not detectably alter the relative proportions of the two proteins in the three frac- tions exhibiting activity. This is a somewhat surprising obser- vation considering the high resolving power of the focusing technique and suggests that both of these proteins may be part of the same phospholipid ‘protein complex in the microsomal membrane.

Factors Affecting Affinity Chromatography- Since there are few examples of affinity chromatography of solubilized membrane lipoprotein complexes, a brief description of condi- tions affecting glucuronosyltransferase binding to and elution from UDP-hexanolamine Sepharose 4B will be presented. Best results were obtained when a partially purified transferase preparation rather than the total detergent-solubilized mem- brane was applied to the affinity resin. This is most likely the result of reducing the extent of nonspecific binding as well as the elimination of degradative enzymes capable of removing the UDP moiety from the support.

Magnesium chloride, which has previously been shown to increase the rate of p-nitrophenol conjugation (21, enhanced binding approximately 2-fold when incorporated into the col- umn buffer at the level of 5 mM. It should be emphasized, however, that close to 42% of the enzymic activity was bound in the absence of added magnesium ion. A second important factor was the degree of substitution of the Sepharose. The binding of transferase was found to be proportional to the UDP-hexanolamine concentration of the gel bed. Hence maxi- mum binding was obtained at high ligand concentrations and in the presence of magnesium ion. Under these conditions, significant amounts of non-enzyme protein were also bound to the affinity column some of which co-eluted with glucuronosyl- transferase. It was for this reason that multiply loaded col- umns were not used. If, however, affinity chromatography was carried out in the presence of 0.2 M NaCl, the amount of enzyme bound was reduced to approximately one-third the normal level, but less than 1% of the applied protein was adsorbed. This permitted a 16- to 20-fold purification of the transferase, although at a very low yield. SDS-polyacrylamide gel electrophoresis of this preparation revealed a major protein component at M, = 59,000, in agreement with the previously established subunit size of glucuronosyltransferase (Fig. 3, left side, Gel C). In this preparation, the M, = 52,600 polypeptide that co-purifies with the transferase is present at a markedly reduced level.

DISCUSSION

UDP-glucuronosyltransferase was purified 200-fold from liver homogenate of phenobarbital-induced animals. Previous purification procedures have reported increases in specific ac- tivity of 32-fold from guinea pig liver homogenate (9), 30-fold

by guest on May 3, 2020

http://ww

w.jbc.org/

Dow

nloaded from

1342 Purification of UDP-Glucuronosyltransferase

from rabbit microsomes (8), and loo-fold from rat microsomal membrane (11). Included in these latter estimates, however, were 2- and &fold activations brought about by phospholipase A digestion, and sequential trypsin and digitonin treatments, respectively. Hence, a significant portion of the stated in- crease in specific activity was the result of enzyme activation rather than purification. The degree of purification reported for our affinity chromatographic procedure does not take into account an enzyme activation resulting from exposure to 1 M

urea; when this stimulation is considered an increase in spe- cific activity of 300-fold is calculated. Furthermore, with those methods employing digestion with crude phospholipase A (8) and combined digitonin and trypsin treatments (111, the solu- bilized enzyme may represent a partial degradation product of the native protein, as has been noted in the case of cytochrome b, (49), TPNH-cytochrome c reductase (50), and DPNH-cyto- chrome bj reductase (51). Detergent solubilization offers the distinct advantage of eliminating the need for lytic enzymes to release the transferase from the membrane. Gel filtration of the solubilized enzyme indicates the active form exists as a phospholipid.protein complex with a molecular weight in the range of 200,000. Purification by ion exchange and affinity chromatography yielded a lipoprotein complex composed of a polypeptide chain of M, = 59,000 and phospholipid. The precise stoichiometry of the complex was not established but based on phosphorus and protein analyses close to 47% on a weight basis was phospholipid. Thus, a complex with an approximate molecular weight of 200,000 would contain two polypeptide chains. It is not known whether this represents the minimum or maximum size for an enzymically active complex. In this context, Gregory and Strickland (52) noted that bilirubin con- jugating activity from digitonin-solubilized microsomes eluted from Sepahdex G-200 in the M, = 150,000 range, as did the majority of the membrane protein. Preparative isoelectric fo- cusing clearly resolved three uniquely charged species of en- zyme. Since the polypeptide composition of these three active forms was invariant, the most likely explanation for these results is that each active lipoprotein complex had slightly different phospholipid compositions. Complete nonenzymic re- moval of phospholipid resulted in total loss of activity, while reconstitution experiments involving the addition of either microsomal lipids or purified synthetic phospholipids restored activity.’ It should also be noted that the existence of charge differences among distinct transferase proteins possessing the same molecular weight would also be consistent with the observed isoelectric heterogeneity.

The co-purification of a M, = 52,600 protein with the p- nitrophenol-conjugating enzyme suggests that these two pro- teins may be closely associated in the native membrane and that upon solubilization of the transferase with Lubrol WX the resulting enzymically active phospholipid . protein complex contains both proteins. Isoelectric focusing fails to separate these proteins; however, the 25 mM EDTA wash of the affinity column does preferentially elute the M, = 52,600 species. This separation is also accomplished by performing the affinity chromatographic step in the presence of 0.2 M NaCl.

Enzyme instability was observed at various stages of purifi- cation, particularly where protein concentrations fell below approximately 1 to 2 mg/ml. As a result, increases in specific activity were not observed that were commensurate with the increase in physicochemical homogeneity. A similar problem was encountered by Dahl and Hokin (53) during the purifica- tion of adenosine triphosphatase. These authors make the point that enzyme purity is best evaluated by physical meth-

ods rather than in terms of specific activity. Both the transfer- ase and phosphatase exhibit an absolute lipid dependence which may be one of the major contributing factors to the problem of activity loss during purification. Factors regulating transferase activity in the microsomal membrane as well as the effect of a variety of substances on biological activity have recently been reviewed by Zakim and Vessey (54).

Our studies do not solve the question of transferase polymor- phism; however, they do provide a method for the preparation of highly purified hepatic UDP-glucuronosyltransferse which is the first step in an unambiguous approach to this question. It is interesting to conjecture that the M, = 52,600 species that remains associated with the active phospholipid protein com- plex could be a transferase with a specificity for an acceptor other than p-nitrophenol or could act in conjunction with the M, = 59,000 polypeptide to modulate specificity. Studies are in progress to explore this point. Arguments for and against the heterogeneity of microsomal UDP-glucuronosyltransferase have been presented in detail (55) and definitive data unequiv- ocally supporting one or the other position are not yet availa- ble. The availability of a highly purified transferase, however, does provide the opportunity to do the in depth kinetic analysis and specificity survey required to answer this question of enzyme polymorphism.

Acknowledgments-We wish to thank John Sheehan and Marilynn Sikes for their expert technical assistance.

Note Added in Proof-Partially purified preparations of UDP-glucuronosyl transferase from rabbit liver microsomes are comprised primarily of two major polypeptides with molec- ular weights of 55,000 and 60,000 and are essentially free from cytochrome P-45O.4

REFERENCES 1. 2.

3. 4. 5.

6. 7.

8.

9.

10.

11.

12. 13. 14.

15.

16.

17.

18. 19. 20.

21.

Gorski, J. P., and Kasper, C. B. (1974) Fed. Proc. 33, 1300 Dutton, G. J. (1966) in Glucuronic Acid, Free and Combined

(Dutton, G. J., ed) pp. 186-299, Academic Press, New York Aitio, A. (1973) Xenobiotica 3, 13-22 Gabaldon, M., and Lacomba, T. (1972)Eur. J. Cancer 8,275-279 Miller, J. A., and Miller, E. C. (1969) in The Jerusalem Sympo-

sia on Quantum Chemistry and Biochemistry (Bergman, E. D., and Pullman, B., eds) Vol. 1, pp. 237-261, Israel Academy of Sciences and Humanities, Jerusalem

Isselbacher. K. J. (1956) Recent Proe. Harm. Res. 12. 134-146 Beaufay, H., Amar-Costesec, A., ?‘hines-Sempoux: D., Wibo,

M., Robbi, M., and Berthst, J. (1974). J. Cell Biol. 61, 213-231 Isselbacher, K. J., Chrabas, M. F., and Quinn, R. C. (1962) J.

Biol. Chem. 237, 3033-3036 Mowat, A. P., and Arias, I. M. (1970) Biochim. Biophys. Acta

212, 65-78 Puukka, R., and Laaksonen, M. (1974) Znt. J. Biochem. 5, 507-

513 Puukka, R., Laitinen, M., Vainio, H., and Hannine, 0. (1975)

Int. J. Biochem. 6, 267-270 Winsnes, A. (1969) Biochim. Biophys. Acta 191, 279-291 Mulder, G. J. (1970) Biochem. J. 117, 319-324 Vessey, D. A., and Zakim, D. (1971) J. Biol. Chem. 246, 4649-

4656 Attwood, D., Graham, A. B., and Wood, G. C. (1971)Biochem. J.

123, 875-882 Hannine, O., and Puukka, R. (1971) Chem:Biol. Interact. 3,282-

284 Graham, A. B., and Wood, G. C. (1972) i3iochim. Biophys. Acta

276, 392-398 Winsnes, A. (1972) Biochim. Biophys. Acta 289, 88-96 Puukka, R., and Hanninen, 0. (1972)Int. J. Biochem. 3,357-362 Vessey, D. A., and Zakim, D. (1972) J. Biol. Chem. 247, 3023-

3028 Graham, A. B., and Wood, G. C. (1973) Biochim. Biophys. Acta

311, 45-50

4 R. Sato, personal communication.

by guest on May 3, 2020

http://ww

w.jbc.org/

Dow

nloaded from

Purification of UDP-Glucuronosyltransferase 1343

22. Axelrod, J., Inscoe, J. K., and Tomkins, G. M. (1958) J. Biol. 38. Vesterbere. 0. (1971) Methods Enzvmol. 22. 389-411 Chem. 232, 835-841

23. Potrepka, R. F., and Spratt, J. L. (1972) Eur. J. Biochem. 29, 39. Righetti, P:, and D&dale, J. W. (i971) Biochim. Biophys. Acta

236, 17-28 433-439

24. Gram, T. E., Hansen, A. R., and Fouts, J. R. (1968) Biochem. J. 40. Mauer, H. R. (1971) Disc Electrophoresis and Related Techniques

106, 587-591 of Polyacrylamide Gel Electrophoresis, 2nd Ed, p. 135, Walter de Gruvter. Berlin

25. Dutton, G. J. (1966) Biochem. Pharmacol. 15, 947-951 26. Coleman, R. (1973) Biochim. Biophys. Actu 300, l-30 27. Graham, A. B., and Wood, G. C. (1969) Biochem. Biophys. Res

Commun. 37, 567-575

41. Bartlett, “G. R. (1959) J. Biol. Chem. 234, 466-468 42. Laemmli. U. (19701 Nature 227. 680-685 43. Fairbanks, G. 11971) Biochemistry 10, 2606-2617 44. Bock, K. W.. Frohline. W.. Remmer. H.. and Rexer, B. (1973)

28. Rao, G. S., Rao, M. L., and Breuer, H. (1972) J. Steroid Biochem. 3, l-10

29. Zakim, D., Goldenberg, J., and Vessey, D. A. (1973) Eur. J. Biochem. 38, 59-63

30. Eleter, S., Zakim, D., and Vessey, D. A. (1973) J. Mol. Biol. 78, 351-362

31. Khandwala, A. S., and Kasper, C. B. (1973) Biochem. Biophys. Res. Commun. 54, 1241-1246

32. Potter, V. R., Baril, E. F., Watanabe, M., and Whittle, E. D. (1968) Fed. Proc. 27, 1238-1245

33. Lowry, 0. H., Rosebrough, N. J., Farr, A. L., and Randall, R. J. (1951) J. Biol. Chem. 193, 265-275

34. Bornens, M., and Kasper, C. B. (1973) J. Biol. Chem. 248, 571- 579

35. Spackman, D. H., Stein, W. H., and Moore, S. (1958) Anal. Chem. 30, 1190-1206

36. Barker, R., Olsen, K. W., Shaper, J. H., and Hill, R. L. (1972) J. Biol. Chem. 247, 7135-7147

37. Porath, J., Axen, R., and Ernback, S. (1967) Nature 215, 1491- 1492

Biochim. Biophys. &a 327, 46-56 45. Goldblatt, P. J. (1972) Sub-cell. Biochem. 1. 147-169 46. Henderson, P. T. (1971) Life Sci. 10, Part II, 655-660 47. Howland, R. D., Burkhalter, A., Trevor, A. J., Hegeman, S.,

and Shirchi, D. Y. (1971) Biochem. J. 125, 991-997 48. Blackburn, G. R., Bornens, M., and Kasper, C. B. (1976)

Biochim. Biophys. Acta 436, 387-398 49. Spats, L., and Strittmatter, R. (1971)Proc. Natl. Acad. Sci. U. S.

A. 68, 1042-1046 50. Welton, A. F., Pederson, T. C., Buege, J. A., and Aust, S. D.

(1973) Biochem. Biophys. Res. Commun. 54, 161-167 51. Spatz, L., and Strittmatter. P. (19731 J. Biol. Chem. 248,793-799 52. Gregory, D. H., II, and Strickland, R. D. (1973) Biochim. Bio-

phys. Acta 327, 36-45 53. Dahl, J. L., and Hokin, L. E. (1974) Anna. Reu. Biochem. 43,

327-356 54. Zakim, D., and Vessey, D. A. (1974) Biochem. Sot. Trans. 2,

1165-1167 55. Mulder, G. J. (1974) Biochem. Sot. Trans. 2, 1172-1176

by guest on May 3, 2020

http://ww

w.jbc.org/

Dow

nloaded from

J P Gorski and C B Kasperliver.

Purification and properties of microsomal UDP-glucuronosyltransferase from rat

1977, 252:1336-1343.J. Biol. Chem. 

  http://www.jbc.org/content/252/4/1336Access the most updated version of this article at

 Alerts:

  When a correction for this article is posted• 

When this article is cited• 

to choose from all of JBC's e-mail alertsClick here

  http://www.jbc.org/content/252/4/1336.full.html#ref-list-1

This article cites 0 references, 0 of which can be accessed free at

by guest on May 3, 2020

http://ww

w.jbc.org/

Dow

nloaded from