The UDP-glucuronosyltransferase (UGT) superfamily plays a major role in the excretion of endobiotics and xenobiotics and their metabolites (Miners and Mackenzie, 1991). Its members have widely overlapping substrate specificity and characteristic individual tissue distribution. UGT comprises two subfamilies, UGT1 and UGT2. UGT1 has a unique gene structure; there are 13 exons 1 from UGT1A1 to UGT1A13P, and exons 2 to 5 are used in common for all mRNAs expressed from the gene. Nine of these exons code for UGT1 isoforms (UGT1A1, UGT1A3, UGT1A4, UGT1A5, UGT1A6, UGT1A7, UGT1A8, UGT1A9, and UGT1A10) (Mackenzie et al., 1997), and four code for pseudo exons 1 (UGT1BP, UGT1A11P, UGT1A12P, and UGT1A13P) (Gong et al., 2001). UGT1 mRNAs are processed by differential splicing. The UGT1 gene appears to be structurally distinctive, in that the UGT1 proteins result from differential splicing of the divergent first exons and share exons 2 to 5 in common (Ritter et al., 1992).

A considerable number of UGT family polymorphisms exist. Human UGT1A1 polymorphisms have been studied particularly closely (Maruo et al., 1999), because they are the causes of hereditary unconjugated hyperbilirubinemias (Bosma et al., 1992; Aono et al., 1993; Sato et al., 1996) and are risk factors for neonatal hyperbilirubinemia and breast milk jaundice (Maruo et al., 1999, 2000). Polymorphisms of UGTs have also been found for UGT1A3 (Iwai et al., 2004), UGT1A4 (Ehmer et al., 2004; Wiener et al., 2004), UGT1A6 (Ciotti et al., 1997), UGT1A7 (Guillemette et al., 2000), UGT1A8 (Huang et al., 2002), UGT1A10 (Jinno et al., 2003), UGT2B4 (Levesque et al., 1999), UGT2B7 (Jin et al., 1993), and UGT2B15 (Levesque et al., 1997).

Human UGT1A4, whose cDNA was first cloned by Ritter et al. (1991), has been identified in the human liver and stomach (King et al., 2000). It catalyzes glucuronidation of primary, secondary, and tertiary amines, carcinogenic aromatic amines (β-naphthylamine, 4-aminobiphenyl, and benzidine), androgens, progestins, and plant steroids (hecogenin, diosgenin, and tigogenin) (Green et al., 1995; Green and Tephly, 1996; King et al., 2000). UGT1A3 and UGT1A4 catalyze the glucuronidation of important endogenous substances including amines and steroids (Mojarrabi et al., 1996; Green et al., 1998). We have earlier described UGT1A3 polymorphisms in a Japanese population (Iwai et al., 2004). UGT1A4 is 93% identical to UGT1A3, and catalyzes the glucuronidation of a number of clinically important therapeutic agents including antihistamines, tricyclic antidepressants, and antipsychotics such as clozapine (8-chloro-11-(4-methyl-1-piperazinyl)-5H-dibenzo[b,e][1,4]diazepine).

Clozapine is known to undergo a large number of biotransformation reactions in humans (Dain et al., 1997). Its metabolism is mediated by CYP1A2, CYP3A4, and CYP2D6 in the liver (Prior and Baker, 2003). When human liver microsomes were incubated with clozapine, at least 5% of the whole conjugates were the corresponding glucuronides (Breyer-Pfaff and Wachsmuth, 2001). In experiments with expressed human UGT1A4, clozapine was mainly conjugated to tertiary 5-N-glucuronides and, to a lesser extent, to a quaternary ammonium (Luo et al., 1994; Green et al., 1995; Breyer-Pfaff and Wachsmuth, 2001). Glucuronidation of clozapine by other UGT isoforms has not been studied, and its biotransformation routes have not been identified in full. The therapeutic serum concentration of clozapine as an antipsychotic drug exceeds 350 ng/ml, and in animal experiments its hepatic concentration was 20 times higher than in the blood (Gauch and Michaelis, 1971). Clozapine also has an adverse effect in agranulocytosis, which is assumed to be mediated by a chemically reactive metabolite (Pirmohamed et al., 1995; Breyer-Pfaff and Wachsmuth, 2001).

Below, using clozapine as the substrate in gene expression experiments, we analyze the incidence and type of UGT1A4 polymorphisms in a Japanese population and study their activity.

Materials and Methods

Materials. [¹⁴C]UDP-glucuronic acid (15.5 GBq/mmol) was obtained from PerkinElmer Life and Analytical Sciences (Boston, MA), and UDP-glucuronic acid and clozapine, trans-androsterone (5α-androstan-3β-ol-17-one), imipramine (10,11-dihydro-N,N-dimethyl-5H-dibenz[b,f]azepine-5-propanamine hydrochloride), cyproheptadine (4-(5H-dibenzo[a,d]cyclohepten-5-ylidene)-1-methylpiperidine hydrochloride), and tigogenin [(25R)-5α-spirostan-3β-ol) were obtained from Sigma-Aldrich (St. Louis, MO). A human liver cDNA library, Mutant Km Kit, and pkF18 vector were obtained from Takara (Kyoto, Japan), and pCR2.1 vector, pCR3.1 vector, and TA-cloning kit were obtained from Invitrogen (Carlsbad, CA). The thin-layer chromatography (TLC) plastic sheet 5748 used below was obtained from Merck (Darmstadt, Germany), the GenePORTER transfection reagent from Gene Therapy Systems Inc. (San Diego, CA), the polyvinylidene fluoride membrane from Bio-Rad (Hercules, CA), and the ECL Plus Western blotting detection reagents from Amersham Biosciences UK, Ltd. (Little Chalfont, Buckinghamshire, UK).

Sequence Analysis of UGT1A4. Genomic DNA was isolated from the leukocytes of 100 healthy Japanese volunteers, with informed consent. The study was approved by the ethics committee of Shiga University of Medical Science. Exon 1 of UGT1A4, along with part of an intron, was amplified by PCR with primer pair 5′-TTTGTCTTCCAATTACATGC-3′/5′-AGATATGGAAGCACTTGTAAG-3′. The forward primer was located upstream of the initiation codon, and the reverse primer was in the intron. The 1075-base pair product was amplified by PCR under the following conditions: initial denaturing for 5 min at 94°C, followed by 1 min at 94°C, 1 min at 62°C, and 2 min at 72°C for 30 cycles with a MiniCycler (MJ Research, Watertown, MA). A final reaction for 8 min at 72°C ensured complete extension of the PCR products. The amplified DNA fragment sequences were determined directly with a dRhodamine Terminator FS Ready Reaction Kit and PRISM 310 (PerkinElmer Life and Analytical Sciences). Sequencing primers were as follows: 5′-CCTTGAGTGTAGCCCAGCGT-3′, 5′-AGGCGGTGGTCCTCACCCCA-3′, 5′-TGCTGTGTTTTTTTGGAGGT-3′, 5′-ATCTTGTCAGCTATGCATCC-3′, and 5′-CTCCAGGTTCCCCTGCCGCG-3′.

Analysis of Allelic Polymorphisms. We subcloned the PCR products (40 ng) to pCR 2.1 vectors (50 ng) using a TA-cloning kit. Transformation by the ligated products took place using Competent Cells JM109 (Toyobo Engineering, Osaka, Japan).

Construction of Expression Vector. UGT1A4 DNA isolated from a human liver cDNA library was amplified by PCR with the primer pair 5′-TTTGTCTTCCAATTACATGC-3′/5′-AGATATGGAAGCACTTGTAAG-3′. The cDNA was inserted into a pCR3.1 expression vector using a TA-cloning kit. Mutation was induced by site-directed mutagenesis, using a Mutan Km Kit according to the manufacturer`s instructions. The constructed cDNA was excised from the pCR3.1 vector by two restriction enzymes (PstI and BamHI) and was ligated into a pkF18 vector (Takara) for mutagenesis. To introduce the mutations, we used the following primer for the T to G transversion at nucleotide 142: 5′-AGCTCCCGCACGGCCTCCCGC-3′, where the mutation is indicated by underlining. After the substitution had been introduced into the pkF18 vector, the converted cDNA was cut out and re-ligated into the pCR3.1 vector. We checked the cDNA by sequencing.

Expression of UGT1A4 in COS-7 Cells. Twenty-four hours before transfection, 6 × 10⁵ COS-7 cells suspended in Dulbecco's modified Eagle's medium with 10% fetal bovine serum were seeded onto 100-mm-diameter culture plates. For transfection, 5 ml of Dulbecco's modified Eagle's medium containing 50 μl of GenePORTER and 3.0 μg of DNA was poured onto the cells. Four hours later, 5 ml of Dulbecco's modified Eagle's medium containing 20% fetal bovine serum was added to each plate. Two models were generated: wild-type UGT1A4 and L48V UGT1A4, and nontransfected cells were used as the controls. After 48 h, the transfected cells were harvested and stored at –80°C before use. Protein content was measured with a BCA Protein Assay Kit (Pierce Chemical, Rockford, IL).

Preparation of Antibody and Measurement of Expressed Protein. For preparation of a polyclonal antibody of UGT1A4, we chose a segment such that the amino acid sequence was unique among UGT1A4 isoforms and the mutation site (L48V) was excluded. We synthesized a 10-amino acid segment (EHLLKRYSRS: residues 104–113) and generated its polyclonal antibody by immunizing rabbits. We purified the antibody by affinity chromatography of 18 ml of Thiol Sepharose 4B (Amersham Biosciences UK, Ltd.) coupled with 6 mg of the synthesized peptide. The column was equilibrated using Dulbecco's phosphate-buffered saline (Nacalai Tesque, Kyoto, Japan), and the antibody was eluted by 0.2 M glycin-HCl buffer, pH 2.5. The final titer of the purified antibody was 1:51,200.

Western Blotting. Cell homogenates underwent sodium dodecyl sulfate-polyacrylamide gel electrophoresis. The protein was transferred to a polyvinylidene fluoride membrane and visualized with ECL Plus Western blotting detection reagents. The membrane was incubated for 1 h in blocking solution, 1 h in a solution of anti-UGT1A4 antibody (1:500), and 1 h in a solution of anti-rabbit antibody (1:15,000). The detection solution was added and the membrane was exposed to film for 5 min. The relative amounts of UGT1A4 expressed at the protein band peaks were measured with an Image Master-CL (Amersham Biosciences AB, Uppsala, Sweden). We blotted the wild-type and L48V samples as a pair on the same membrane, and confirmed that the measurements were in the linear range by assaying three different concentrations of an identical cell homogenate at the same time. The relative amount of L48V UGT1A4 was always within about 0.9 to 1.1 times that of wild-type UGT1A4.

Assay of UGT1A4 Activity. UGT1A4 activity was assayed as described previously, with a minor modification (Iwai et al., 2004). The glucuronidation process of the substrate was assayed with [¹⁴C]UDP-glucuronic acid. The incubation mixture contained 100 to 150 μg of cell homogenate, 0 to 500 μM clozapine, 500 μM UDP-glucuronic acid, 9.25 kBq [¹⁴C]UDP-glucuronic acid (5.97 μM), dimethyl sulfoxide (1%), 10 mM MgCl₂, 50 mM Tris-HCl buffer (pH 8.4), and 8.5 mM saccharolactone in a final volume of 100 μl. Incubation took place at 37°C for 30 min. The clozapine-glucuronide that resulted was isolated by thin-layer chromatography (TLC) on TLC plastic sheet 5748. The TLC plates were developed in a solvent system containing n-butanol, acetone, water, and ammonium hydroxide (35:35:20:10 v/v), and were scanned on an Instant Imager (PerkinElmer Life and Analytical Sciences). Major and minor glucuronide metabolites of clozapine were detected, and their combined density was taken to represent the reaction product. Under the same conditions, we measured the UGT1A4 activities of the wild-type and L48V polymorphisms, using as substrates trans-androsterone (0–500 μM), a tricyclic antidepressant, imipramine (0–500 μM), an antihistamine, cyproheptadine (0–500 μM), and tigogenin (0–100 μM).

Data Analysis. We calculated kinetic parameters using Prism 3.0 software (GraphPad Software Inc., San Diego, CA) and nonlinear regression on the Michaelis-Menten equation. All data displayed are the means of three separate experiments. The enzyme efficiencies were subjected to a t test for pairwise comparison.

Results

Identification of UGT1A4 Mutations and Their Frequencies. We identified three nucleotide substitutions in exon 1 of UGT1A4 in our Japanese test population. One was a T to G transversion at nucleotide position 142 (142T>G), which predicted a substitution of valine for leucine at codon 48 of the enzyme (L48V). The other two were novel, silent transitions (448T>C: L150L, 804G>A: P268P). We identified a novel substitution mutation in the intron (867 + 43C>T) and found no substitutions in exons 2 to 5. The incidence of the substitutions identified ranged from 0.155 to 0.165 (Table 1), indicating that they represented polymorphisms in the sample population. Indeed, analysis of subcloned exon 1 revealed the presence of four alleles, with frequencies ranging from 0.01 to 0.825 (Fig. 1). Two of these (L48V-L150L-P268P-867 + 43C>T and P268P) were novel. We detected five patterns of allelic combinations (Fig. 2).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 1.

Amino acid substitutions found in UGT1A4 in the Japanese population sample, and their frequencies.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 2.

Amino acid substitution combinations found in UGT1A4 in the Japanese population sample, and their frequencies.

Expression and Identification of UGT1A4. In a Western blotting analysis, 55-kDa protein bands (Fig. 3) (Miners and Mackenzie, 1991) were detected in all the expression models but not in the mock transfection; the molecular weight of the band was as reported previously.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 3.

Identification of expressed UGT1A4 protein by Western blotting. A, mock transfection; B, wild type; C, L48V.

Activities of UGT1A4 Isoforms. Clozapine-glucuronide production was proportional to the amount of cell homogenate at up to 1000 μg. At 100 μg, the reaction product was linear with time up to 90 min. The reaction velocity increased steeply with pH and then reached a plateau at pH 8.0–9.0. Lineweaver-Burk plots showed the apparent K_m of the reaction to be 285 μM at three clozapine concentrations (25 μM, 50 μM, and 200 μM), and 66 μM at three UDP-glucuronic acid concentrations (125 μM, 250 μM, and 500 μM).

At about 200 μM clozapine, glucuronidation by wild-type UGT1A4 reached its maximum velocity and then declined, indicating substrate inhibition (Fig. 4). Table 2 shows the kinetic parameters of the expressed wild-type and L48V enzymes. The difference in V_max between these enzymes was not statistically significant, and the K_m of L48V was half that of the wild-type enzyme, implying that the relative efficiency (V_max/K_m) of the homozygous model of L48V was twice that of the wild type.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 4.

Velocity of reaction of wild-type UGT1A4 with clozapine. The reaction mixture contained 100 μg of cell homogenate and 500 μM UDP-glucuronic acid, and was incubated for 30 min at pH 8.4. Note inhibition at >200 μM clozapine.

Table 3 shows kinetic parameters determined for UGT1A4 for trans-androsterone substrate, comparing that type, for which the glucuronidation activity decreased in L48V, with the wild type (Ehmer et al., 2004); comparisons are also presented for three other substrates, imipramine, cyproheptadine, and tigogenin. The L48V had only 30 to 60% of the V_max value of the wild type for all four substrates. Apparent K_m values of L48V for trans-androsterone, imipramine and cyproheptadine were also considerably smaller than for the wild type, but that of tigogenin was not changed. The L48V/wild ratios of efficiencies for trans-androsterone, imipramine, cyproheptadine, and tigogenin were, respectively, 10.4, 1.8, 1.6, and 0.6.

Discussion

Glucuronidation is a major excretion pathway for diverse endobiotics and xenobiotics catalyzed by the UGT family of enzymes. A considerable number of UGT polymorphisms have been identified, some of which are associated with disease or adverse drug effects. Polymorphisms of UGT1A1 are known to cause Gilbert syndrome, a mild hereditary unconjugated hyperbilirubinemia and breast milk jaundice (Maruo et al., 2000), and may contribute to breast cancer (Adegoke et al., 2004). Polymorphisms of UGT1A4, UGT1A6, UGT1A7, and UGT1A8 may also be risk factors for carcinogenesis (Gagne et al., 2002; Strassburg et al., 2002; Ehmer et al., 2004; Wiener et al., 2004). In our previous report we detected six novel UGT1A3 SNPs and considered their possible contribution to differences between individuals in drug metabolism and susceptibility to side effects (Iwai et al., 2004). UGT1A4 has a fairly similar amino acid sequence to UGT1A3 and overlapping substrate specificity for several amines, suggesting the presence of its polymorphism.

In the experiments reported here, we analyzed UGT1A4 in our Japanese test population. We identified two novel and one known single nucleotide substitution in exon 1, and one novel substitution in intron 1 (Table 1). All incidences were greater than 0.01, indicating that the substitutions were polymorphisms. Recently, P24T and L48V polymorphisms have been detected in a German population (Ehmer et al., 2004). We did not detect P24T in our sample population, but L48V is a common polymorphism in Japanese as well as European populations. The present analysis shows that polymorphism of UGT1A4 differs between races.

Linkage analysis of the alleles coded by the four UGT1A4 SNPs revealed that their combinations generate three polymorphisms (Fig. 2), of which two (L48V-L150L-P268P-867 + 43C>T and P268P) are novel. The amino acid sequence of the enzymes expressed by L48V-L150L-P268P-867 + 43C>T and L48V were the same, but the former had an additional intronic polymorphism (867 + 43C>T). Thus, based on amino acid sequences, there were only two enzymes in the sample population: wild-type and L48V. The location of 867 + 43C>T in the first intron was not in the typical consensus sequence of splicing, and the nature of its contribution in the splicing of UGT1A4 mRNA is not clear.

In the present study, L48V was twice as efficient as the wild type in glucuronidating the antipsychotic drug clozapine (Table 2). We also confirmed, by in vitro expression of the wild-type enzyme, that the glucuronidation of clozapine shows substrate inhibition (Fig. 4). Ehmer et al. (2004) reported that the L48V activity for trans-androsterone was lower than that of the wild type. Although our data also showed a reduced V_max of L48V for trans-androsterone (Table 3), the efficiency of L48V was 10 times higher than that of the wild type. Likewise, the efficiencies of L48V for imipramine and cyproheptadine were 1.6 to 1.8 times higher than that of the wild type, but for tigogenin, there was reduction by a factor of 0.6. These results indicate that L48V accelerates or slows the glucuronidation of UGT1A4, depending on substrates.

In conclusion, our results show a polymorphism of UGT1A4 (L48V) in a Japanese population. An understanding of the effects of the mutation on the glucuronidation of substrate drugs would help clinicians who prescribe these drugs to calculate the appropriate dosage and avoid adverse effects.