A comparison of the kinetic properties of two different forms of microsomal UDPglucuronyltransferase.

Two forms of UDPglucuronyltransferase (EC 2.4.1.17) have been purified from microsomes of pig liver. One form is free of phospholipids and the other contains a small amount of residual phospholipids. Each form, however, is responsive to activation on addition of purified phospholipids. Comparison of kinetic properties of these enzymes, after reconstitution with identical phospholipid environments, indicate that these are unique functional forms of UDPglucuronyltransferase. The two differ by as much as 100-fold in their rates of conjugation at Vm of p-nitrophenol. Relative rates of glucuronidation of a variety of phenolic aglycones are different for the two enzymes, which suggests different reaction mechanisms. The energetic basis for binding of UDP-glucuronic acid to the active sites is different for the two forms of UDPglucuronyltransferase. Moreover, one form, but not the other, binds Mn2+, which leads to modulation of kinetic properties.

explored. The data presented in this paper indicate that the differences between forms of UDPglucuronyltransferase involve significant aspects of function other than their specificities for aglycones. Two purified forms obtained from pig liver microsomes that seem primarily to catalyze the glucuronidation of phenols have differences in activities at V, on the order of 100-fold with p-nitrophenol as aglycone, different reaction mechanisms, and different mechanisms for interacting with UDP-glucuronic acid. In addition, the function of one of these enzymes, but not the other, is affected by Mn2+.

MATERIALS AND METHODS
LPC' was prepared from egg yolk by treating purified lecithin from eggs with phospholipase Az, as in Ref. 8. This LPC was used only for purification procedures. Phospholipids used in reconstitution experiments were purchased from P-L Biochemicals, Milwaukee, WI. They were used without further purification. All other chemicals used were the best available commercial grades. All solutions were prepared with deionized, double distilled water.
Protein concentration was measured by the biuret method (9) except when glycerol was present. In this instance, protein was determined by the method of Lowry et al. (10) after precipitation of the protein with trichloroacetic acid (11). Phospholipid phosphorus was determined after digestion of the CHCL/CH30H extracts of the different forms of UDPglucuronyltransferase (12).
Purification of UDPglucumnyltransferase-The details of the purification of GT2p and of the starting material for purification of GTIP are given in Ref. 13. Partially purified GTIP (2-3 mg) was applied to a hydroxylapatite column (Bio-Rad Laboratories, Bio-Gel HT), (1.5 X 15 cm) equilibrated with 5 m~ Tris, pH 8 (0 "C), 0.1 m~ dithioerythritol, 2 m~ EDTA, 1.1% (w/v) cholate, and 10% (v/v) glycerol. The column was washed with 80 ml of 150 m~ phosphate in the same buffer. GTlp was dialyzed extensively as in Ref. 13 in order to reduce the concentration of cholate and phosphate. The enzyme solutions were frozen at -20 "C and could be stored for several months without appreciable loss of activity.
Based on activities at V,,, and after correction for detergent-induced activations, the purification procedures described in Ref. 13 yield enzyme preparations that are enriched 1350-fold for GTzP and 20-fold for GTIP, as compared with the rate of conjugation of p-nitrophenol in intact microsomes. The figure for the enrichment of GTlp versus intact microsomes has to be interpreted cautiously because GT2p, as compared with GTW, has 100-fold greater activity withp-nitrophenol (see below) and probably represents the major enzyme in microsomes for glucuronidation of this substrate. Hence, one cannot compare the specific activity of pure GTlp with that of GTlp per se i n intact microsomes. The yield of enzyme as a fraction of microsomal protein in the final preparation is less than 0.1%. Activities at V, and other relevant kinetic constants were determined as in Refs. 12 and 13, for The abbreviations used are: LPC, lysophosphatidylcholine; GT1p, the form of UDPglucuronyltransferase eluted from hydroxylapatite at 90 m~ Pi; GT=, the form of UDPglucuronyltransferase eluted from hydroxylapatite at a concentration of Pi greater than 90 m~; PC, phosphatidylcholine; DOPC, dioleoylphosphatidylcholine; the subscripta UDPGA and pNP refer to UDP-glucuronic acid and p-nitrophenol.
an enzyme with a rapid equilibrium, random order mechanism (14, 15).
Reconstitution of GTP and GT2-Enzyme was mixed with the indicated lipid in a ratio of 1:lO (w/w) when enzyme was reconstituted with LPC and in a ratio of 1:6 when enzyme was reconstituted with PC. These ratios of protein to lipid yield maximal activity. Mixtures of enzyme and lipid were kept at 0 "C under a stream of argon until aliquots were removed for assay. The activities of the enzyme-lipid mixtures were stable for at least 8 h, which was the longest interval tested. Further details of reconstitution are given in Refs. 12 and 13. Dispersion of Phospholipids-LPC was dispersed in water. PC was microdispersed in 10 m~ Tris, pH 7.5 (30 "C), 1-2 ml, by ultrasonication with a microtip (Heat Systems-Ultrasonication, Plainview, NY, model W185F) for 10 min. The sonicated mixture was centrifuged at 10,OOO X g in order to remove multilamellar structures.

RESULTS
Comparison of activities at V,,, of GTlp and GTZ-GTlp, when assayed with p-nitrophenol as aglycone, has no measurable activity in the absence of added phospholipids. This is true even though the preparation of GTlp contains a small amount of residual phospholipid phosphorus, which we were unable to remove by washing lyophilized enzyme with cold dry acetone or CHC13/CH30H. The reasons for the strong apparent avidity of GTlp for a small amount of phospholipid are not clear. No attempts were made to characterize the residual phospholipid.
The conditions for assay of GTp, in the absence of added phospholipids, would have enabled detection of rates of glucuronidation on the order of 10"' mol/min/mg of enzyme protein. Hence, the activity of GTlp in the absence of added phospholipids is exceedingly low, if present at all. By contrast, the delipidated form of GTzp has a small activity. The activities of both forms of UDP glucuronyltransferase are stimulated by phospholipids; but, for a given lipid, the activity of GTzp is far larger than that for GTIP, when activity was measured with p-nitrophenol as aglycone (Table I). On the other hand, to the extent that the relationship was examined, the length and unsaturation of the acyl chain of LPC had similar quantitative effects on the activities at V,,, for GTP and GTzp. Activities at V,,, were not determined for GTlp in lipid environments other than those in Table I. Activities were measured, however, at a single set of substrate concentrations for GTlp reconstituted with a broader range of LPC than listed in Table I. These data were in agreement with the idea that longer chain length and unsaturation of the acyl chain of LPC are associated with greater activation of GTlp as compared with LPC containing shorter and saturated acyl chains (data not shown). Also, as for GTw, LPC was a better activator of GTlp than was PC. Despite these similarities, interactions between GTlp and a given lipid must have functional consequences that differ from those due to interactions between TABLE I Comparison of actiuities a t V, of GT,, and GTzp as a bnction of the phospholipid environment of each enzyme GTlp and GTzp were purifed and assayed at 30 "C as under "Materials and Methods." When enzyme was reconstituted with lipids, a 10-fold excess (w/w) of LPC or a 6-fold excess of DOPC was added to pure enzyme. These amounts of lipid gave maximal activation in all instances. V,,, at 30 "C was determined as in the text. Activities are micromoles ofp-nitrophenol glucuronidated/min/mg of enzyme protein. GTzp and the same lipid. For example, the ratio of activities at V,,, for GTP to activity at V,,, for GTzp was different as the lipid environments of the reconstituted enzymes were vaned.
Comparison of Mechanism of Catalysis by GTIp and GTP-Recent studies of the function of GT2p indicate that the relative rates of glucuronidation at V, of substituted phenols, catalyzed by enzyme reconstituted with oleoyl-LPC, depend only on the pK, of the phenolic aglycone (13). These data are compatible with the conclusion that glucuronidation catalyzed by GT2p reconstituted with oleoyl LPC proceeds via a mechanism in which the rate-determining step is nucleophilic attack of phenolate ion on the bond between UDP and C-1 of glucuronic acid (12, 13). The large differences between GTIP and GTPP (for enzymes reconstituted with oleoyl-LPC) in rates of glucuronidation of p-nitrophenol, could be due to a kinetic mechanism for GTlp in which the rate-determining step differs from that observed for catalysis by GT2p. Another possibility is that the active sites of GTlp and GT2p have differential effects on the nucleophilicity of the phenolate form of phenolic aglycones. These possibilities were investigated by determining for GTlp the rate of glucuronidation at V, with a series of para-substituted phenols. The data are shown in Fig. 1 in the form of log V,,, uersus the Hammett function for para-substituted phenols. Rates of conjugation catalyzed by GTlp reconstituted with oleoyl-LPC vary with the acidity of the phenolic aglycones, but the slope of the plot is 0.58. The theoretical slope for this plot is 2.23 (16). Hence, the rate of glucuronidation of phenols catalyzed by GTlp as compared with GTzP is dependent only partially on the concentration of phenolate ion. There are several possible explanations for the differences, but these cannot be evaluated by the present data, nor is it possible to draw conclusions about the nucleophilicity of phenols within the active sites of GTlp and GTzp. It seems clear, nevertheless, that the mechanism of the conjugation reaction catalyzed by GTIP (reconstituted with oleoyl-LPC) is different from the mechanism of the reaction catalyzed by GTzp (reconstituted with oleoyl-LPC). Another important aspect of the data in Fig. 1 is that the rates at which GTlp catalyzes the glucuronidation ofp-aminophenol are 10-fold greater than the rate at which G T~P catalyzes the conjugation of this aglycone (13). Therefore, whereas GTPP uersus GTlp seems to be especially suited for catalyzing the glucuronidation of acidic phenols, GTIP may be a more efficient enzyme for catalyzing the glucuronidation of weakly acidic phenols, at least under the condition that each enzyme is reconstituted with oleoyl-LPC. We also determined whether GTlp has high catalytic specificity for functional groups other than phenols. This does not seem to be the case. GTlp has lower specific activity for conjugation of "COO-and "SH as compared with rates of conjugation of p-nitrophenol; and it does not catalyze formation of Nand C-glucuronides.  Table 11. It has not been confirmed by kinetic isotope effects (17) that these enzymes have truly rapid equilibrium mechanisms, but patterns of product inhibition and initial rate data are consistent with a rapid equilibrium, random order kinetic mechanism (14). We think it is reasonable, therefore, to use the kinetically determined binding constants in Table I1 as reflections of the affinity of GTlp and GTzp for ligands. The data show a small difference in the value of KUD~GA for GTlp uersus GTPP. The values of other K terms depend on the phospholipid environment of each form of UDPglucuronyltransferase. Of interest are the differential effects of variable phospholipid environments on the ratios KuDpGA/RUDpGA (or K p~p / R p~p ) for GTlp uersus GTPP which reflect, most likely, ligand-induced changes in the affinity of the enzyme for other ligands. For example, for GT,p reconstituted with myristoyl-LPC, the data in Table  I1 suggest, based on the presumed kinetic mechanism, that UDP-glucuronic acid binds to free GTlp with less avidity as compared with the affinity of the complex GTlp .p-nitrophenol for UDP-glucuronic acid. The ligand-induced alteration in affinity of free enzyme for the second ligand depends on the lipid environment since it is different for GTlp in each of the lipid environments tested. It is always different for GTlp, however, as compared with the same effect in GT2p.
Although the differences between GTlp and GTZp in their affinities for UDP-glucuronic acid appear to be small, the data in Table 1 1 1 indicate that there are major differences in the manner in which each of these enzymes (reconstituted with oleoyl-LPC) interacts with UDP-glucuronic acid. GTip has

KIUDP. Mn2+
KIglUC-nic ecld high affinity for UDP and poor affinity for glucuronic acid. Also, the binding of UDP-glucuronic acid to GTIP is as strong as the binding of UDP. This suggests that interactions between UDP and GTlp provide nearly all the binding energy for interactions between enzyme and UDP-glucuronic acid. By contrast, GTPP has an appreciable affinity for binding for glucuronic acid. GTzP, as compared with GTIP, also has a higher affinity for UDP. As pointed out previously, however, the binding of UDP-glucuronic acid to GTzp must be associated with a considerable loss of inherent binding energy from interactions between UDP and GTZP and glucuronic acid and GTPP (12). This loss of inherent binding energy does not appear to occur when UDP-glucuronic acid binds to GT1p, as evidenced by a close correspondence of K U D P G A and KIuDp for GTP (Table 111). Differential Effects of Mn2+ on the Properties of GT,p and G T z p T h e activity of UDPglucuronyltransferase in intact microsomes is stimulated by divalent cations (18). Divalent cations, however, have no effect on the activity of GTZP (12). We observed, however, that Mn2+ and other divalent cations activated GTlp (Table IV). The kinetic mechanism of this activation was enhancement of activity at V,,, and of affinity of enzyme for substrates. The data in Table IV are for GTlp reconstituted with oleoyl-LPC; but the effects of Mn2' on GTlp were independent of the type of lipid used to reconstitute delipidated GTlp. Since Mn2+ enhances the affinity of GTlp for UDP-glucuronic acid, we determined the effects on Mn2+ on affinity of GTlp for the UDP and glucuronic acid regions of this substrate. Addition of Mn2+ (1.5 mM) to GTlp enhanced the affinity of GTlp for each of these moieties (Table IV). This concentration of Mn2+ gave maximal enhancement of activity. At the concentrations of Mn2+ and UDP present in the assays, virtually all the UDP existed as a UDP-Mn2+ chelate (19). It is not possible with these data to decide whether GTlp has a higher affinity for UDP-Mn2+ uersus UDP, or whether binding of Mn2+ to GTlp enhances affinity for UDP by a mechanism  Table V, UDP-Mn" does not bind to GTPP.
The data, therefore, show that there must be substantial differences between GTlp and GT,p in the regions of these enzymes that interact with UDP.

DISCUSSION
We have purified from pig liver two forms of UDPglucuronyltransferase, for which phenols appear to be the best substrates. These two enzymes, however, appear to have different rate-limiting catalytic steps, widely variable catalytic activities with the same substrates, and a different energetic basis for the binding of UDP-glucuronic acid to the active site of each enzyme; and one, but not the other, is responsive to activation by divalent cations. These two forms of UDPglucuronyltransferase, therefore, would appear to contain significant differences in their primary structures; and, in fact, the only homology of function between them is that they catalyze the glucuronidation of the Same substrates. Because of the similarity of substrates metabolized in vitro by GTlp and GTzp and the low specific activity of the former uersus the latter, the physiological sigdicance of the GTlp is not completely clear. It is of interest, however, that whereas GTZP as compared with GTlp has higher activity with relatively strongly acidic phenols, GTlp versus GTzp has significantly higher activity with relatively weakly acidic phenols, such as p-aminophenol. Possibly, the best substrate for GTlp has not been established by the above experiments. It is possible, too, that its physiologic lipid environment has an effect on the activity of GTlp that cannot be discerned from these experiments. These uncertainties do not affect, however, the conclusions about the marked differences in the kinetic properties of GTIP versus GTPP.
There is reason to believe that the uniqueness of different forms of UDPglucuronyltransferase cannot be established firmly by comparing the relative rates of glucuronidation of multiple aglycones catalyzed by different functions of microsomal proteins. The basis for this idea is the demonstration that the relative rates of glucuronidation at V, of phenolic aglycones catalyzed by a single pure form of UDPglucuron-yltransferase can depend on the chemical composition of the phospholipid used to reconstitute the activity of delipidated enzyme (7). For example, the ratio of activity at V,,, for conjugation of p-nitrophenol and 1-naphthol is 80 for pure, delipidated GTZP reconstituted with oleoyl-LPC, but 3 when this Same form of the enzyme is reconstituted with palmitoyl-LPC (13). Determination of the uniqueness of the catalytic function of a form of UDPglucuronyltransferase hence appears to require careful control of the enzyme's lipid environment. This is not a trivial problem in working with proteins like UDPglucuronyltransferase, which are integral components to membranes. These enzymes can form mixed micelles with detergents and phospholipids during purification so that separation of the enzymes of interest from contaminating proteins could depend on the properties of mixed micelles of variable composition and not only on the properties of the enzyme. Determinations of the uniqueness of function of different fractions of presumed pure UDPglucuronyltransferme depend at this time, therefore, on characterization of these enzymes after their delipidation, followed by reconstitution with phospholipids of known composition.