UDP-glucuronosyltransferase 1A3 (UGT1A3) is a phase II drug metabolism enzyme localized in the endoplasmic reticulum and the nuclear envelope (Radominska-Pandya et al., 1999). It catalyzes the conversion of hydrophobic endogenous compounds and xenobiotics to hydrophilic glucuronides and hence plays a major role in metabolism and detoxification (Miners and Mackenzie, 1991; Wells et al., 2004).

The UGT genes are highly polymorphic. Recently, Ehmer et al. (2004) and Iwai et al. (2004) have identified six novel single nucleotide polymorphisms (SNPs) in exon 1 of human UGT1A3 gene (Fig. 1). Of these, four cause amino acid substitutions (A17G, Q⁶R; T31C, W¹¹R; C133T, R⁴⁵W; and T140C, V⁴⁷A), and the remaining two are silent (G81A, E²⁷E and A477G, A¹⁵⁹A). The missense variants showed a significant difference in their estrone metabolism activity (Iwai et al., 2004). Considering a wide substrate spectrum of UGT1A3 (Green et al., 1998) and the fact that the activity alteration for a given enzyme could be substrate-specific, it is worth conducting a further functional evaluation of these UGT1A3 variants with other important substrates.

Flavonoids are dietary antioxidants widely existent in fruits, vegetables, cereals, dry legumes, chocolate, and beverages (Scalbert et al., 2005). Experimental studies support an important role of flavonoids in the prevention of cardiovascular diseases and cancers (Galati et al., 2000; Kris-Etherton and Keen, 2002; Vita, 2005). Since glucuronidation is an essential pathway for the elimination and biological activity of flavonoids (da Silva et al., 1998), genetic polymorphisms of UGTs are potentially of toxicological and physiological importance on the pharmacokinetics, pharmacologic effects, and toxicity of flavonoids (Wittig et al., 2001). In the present study, wild-type and variant human UGT1A3s were expressed in a Bac-to-Bac baculovirus expression system. Three common flavonoids, quercetin, luteolin, and kaempferol, were used as substrates to characterize their glucuronidating capacity. Regioselectivity in quercetin metabolism by these UGT1A3 variants was also observed. Finally, the frequency distribution of UGT1A3 polymorphisms in a Chinese Han population was reported in this work.

Materials and Methods

Materials. Polymerase chain reaction (PCR) primers and a blood genomic DNA isolation kit were obtained from Sangon (Shanghai, China). Restriction endonucleases, DNA molecular marker, and T4 ligase were obtained from MBI Fermentas (Amherst, NY). pGEM-T plasmid was obtained from Promega (Madison, WI). A MutanBest kit was obtained from Takara (Tokyo, Japan). Cellfectin reagent, pFastBac1 vector, DH10Bac-competent cells, and Grace's medium were obtained from Invitrogen (Calsbad, CA). Spodoptera frugiperda Sf9 insect cells were kind gifts from YangShengTang Pharmaceutical (Hangzhou, China). Fetal bovine serum was purchased from Hyclone (Logan, UT). Mouse anti-his antibody and peroxidase-conjugated goat anti-mouse secondary antibody were supplied by GE Healthcare Bio-Sciences (Little Chalfont, Buckinghamshire, UK) and Zhongshan Biotechnology (Guangdong, China), respectively. The polyvinylidene fluoride (PVDF) membrane was acquired from AMRESCO (Solon, OH). Uridine diphosphate glucuronic acid (UDPGA), d-saccharic acid 1,4-lactone monohydrate, Brij 58, alamethicin, and β-glucuronidase were provided by Sigma (St. Louis, MO). Quercetin, luteolin, and kaempferol were purchased from National Institute for the Control of Pharmaceutical and Biological Products (Beijing, China), and their purities were 99%.

Construction of UGT1A3 Variants. All five reported UGT1A3 enzymes were constructed in the present study: UGT1A3.1, UGT1A3.2, UGT1A3.3, UGT1A3.4, and UGT1A3.5. A UGT1A3*1 gene had been obtained in our laboratory previously, with an AflII site and a XhoI site at two sides of the coding sequence and a his-tag ahead of the stop codon. In this research, using pGEM-T-UGT1A3*1 as the template (Chen et al., 2005a), UGT1A3 alleles were generated by site-directed mutagenesis following the instructions of the MutanBest kit. Oligonucleotide primers for introducing nucleotide transversions are shown in Table 1. To ensure that no mistake happened during the mutagenesis, all PCR products were verified by DNA sequencing.

Expression of UGT1A3s. Expression of UGT1A3s was performed in an insect Bac-to-Bac system as we described previously (Chen et al., 2005a). In brief, the UGT1A3 gene was excised from the pGEM-T vector, subcloned into the pFastBac1 vector, and transformed into Escherichia coli DH10Bac. PCR with UGT1A3-specific primer and M13 primer was used to confirm the transposition of UGT1A3 into the bacmid. The minipreparation of bacmid-UGT1A3 was transfected into Sf9 cells. After 72 h, the first passage of baculovirus was harvested from supernatant and titrated by viral plaque assay. Infection conditions were optimized for subsequent amplification or expression. Infected cells were collected by centrifugation and disrupted by sonication. Cell homogenate was quantified by GeneQuant (GE Healthcare BioSciences) at 280 nm and stored at –80°C until further analysis. Transfection with an untransposed bacmid was performed as a negative control. All five UGT1A3 enzymes were expressed in the same condition.

Western Blot. To ascertain the relative expression level of each UGT1A3 enzyme, a semiquantitative Western blot method was established using the UGT1A3.1 preparation as a standard. The cell homogenate of each UGT1A3 in the same loading amount was separated by a 12% SDS-polyacrylamide gel, transferred onto a PVDF membrane, and probed with the mouse anti-his antibody (1:3000). A peroxidase-conjugated goat anti-mouse antibody (1: 5000) was used as the secondary antibody. Immunocomplexes were exposed to 3,3′-diaminobenzidine detection solution for 10 min. The relative expression of UGT1A3s was measured with a universal hood image system (Bio-Rad Laboratories, Segrate, Italy) using Quantity One 1-D analysis software. To optimize the loading amount, the linear range of Western quantification was determined by assaying UGT1A3.1 homogenates at different concentrations.

Glucuronidation Assay. Some preliminary experiments were performed to determine the optimal incubation condition. The effects of pH value (7.0, 7.4, 7.8, 8.0, 8.2, and 8.6), β-glucuronidase inhibition reagent (d-saccharic acid 1,4-lactone monohydrate), and de-latency reagents (alamethicin and Brij 58) on the glucuronidation assay were evaluated. A typical incubation mixture (100 μl of total volume) contained 100 mM Tris-HCl buffer, pH 8.0, 10 mM MgCl₂, 2 mM UDPGA [the UDPGA apparent K_m is 200 μM∼300 μM with the various UGT isoforms as previously determined (Green et al., 1998)], 15 μg/mg pro alamethicin, 5 mM d-saccharic acid 1,4-lactone monohydrate, 0.5 mg/ml cell homogenate, and 0.1 mM flavonoids. The reaction was initiated by adding UDPGA, incubated at 37°C, and terminated with 290 μl of methanol. After adding an internal standard (morin for quercetin, quercetin for kaempferol, and kaempferol for luteolin), the mixture was centrifuged to remove precipitated protein. Twenty-five microliters of supernatant were injected into an LC-10A HPLC system (Shimadzu, Kyoto, Japan), equipped with two LC-10AD pumps, a UV detector, and a Diamonsil C18 column (5-μm particle size, 250 mm × 4.6 mm). Data acquisition and integration were performed using an HS2000 chromatography workstation. The mobile phase was used at a flow rate of 1 ml/min, with 52%, 52%, or 54% methanol (volume percentage) in 0.02 M phosphoric acid (pH 2.0) for quercetin, luteolin, and kaempferol, respectively. Peaks were detected at 368 nm for quercetin, 350 nm for luteolin, and 365 nm for kaempferol. Glucuronides were identified with an Agilent 1100 liquid chromatography-mass spectrometry system (Agilent Technologies, Palo Alto, CA) equipped with an electrospray ionization source and operated in a negative ion mode. The enzymatic kinetic assay of each UGT1A3 enzyme was conducted under conditions linear with time and protein concentration. Kinetic parameters were calculated from three independent expressions according to the Lineweaver-Burk equation. (For detailed concentrations of each tested flavonoid, see Tables 3, 4, 5.)

Regioselectivity of UGT1A3s to Quercetin. Regioselectivity of UGT1A3 variants to quercetin was measured using enzymatic kinetic parameters according to our previous work (Chen et al., 2005a). In brief, each monoglucuronide was separated and collected from the HPLC effluent and quantified by determination of the amount of quercetin released after hydrolysis. The identities of the monoglucuronides were determined with UV spectra in methanol and some diagnostic reagents (sodium methoxide solution and unmelted sodium acetate solution).

DNA Isolation. One hundred twenty-five unrelated healthy Chinese Han adults were included in this study. All these recruitments were from the clinical laboratory in the Second Affiliated Hospital of Zhejiang University. The research protocol was approved by the Ethics Committee of Zhejiang University. An aliquot of 2.5 ml of venous blood was drawn from each volunteer into a tube containing 0.3 ml of sodium citrate and stored at –80°C before DNA isolation. Genomic DNA was extracted with a UNIQ-10 column blood genomic DNA isolation kit, following the manufacturer's specifications, and quantified by measuring absorbance at 260 nm and 280 nm with a GeneQuant spectrophotometer (GE Healthcare Bio-Sciences). Typical yields of DNA were 15 ng/μl∼25 ng/μl, and A₂₆₀/A₂₈₀ ratios were 1.7∼1.8. Samples were stored in the Tris-EDTA buffer (pH 8.0) at –20°C until further analysis.

Distributions ofUGT1A3Alleles and SNPs inUGT1A3in a Chinese Han Population. Two-step allele-specific amplification (ASA) was developed for mutation detections. At the first step, genomic DNA was amplified for 20 cycles with two UGT1A3-specific primers: AAGAAAGCAAATGTAGCAGGC/TACCTTATTTCCCACCCACTTC. Ten PCR products were chosen randomly for sequencing and as positive controls for the second step. The second amplification was performed following an allele specific procedure. For each SNP in UGT1A3, two parallel reactions were performed. A pair of allele-specific primers (Table 2) was designed and added into two parallel PCRs, respectively. PCR was performed in 20 μl of reaction mixture containing 1.75 mM MgCl₂, 0.2 mM dNTPs, 0.3 μM each primer (Table 2), 0.1 μl of product of the first step, 0.5 U of TaqDNA polymerase. After an initial denaturation at 94°C for 2 min, the DNA was amplified by 25 cycles of 20 s at 94°C and 60 s at an optimized temperature (Table 2). All PCR processes were performed on a Mastercycler gradient (Eppendorf, Hamburg, Germany). Special attention was paid to keep the same load and reaction conditions for two parallel PCRs. After PCR, amplification products were electrophoresed on a 1.5% agarose gel containing ethidium bromide. The images were scanned and analyzed with a universal hood gel image system (Bio-Rad Laboratories). Predicted lengths of fragments are also shown in Table 2. For the second step, 10 positive controls (sequenced samples) and 2 negative controls (reactions containing no template) were performed. The result of analysis of each example was agreed to by two individual operations.

Data Analysis. SNP frequencies were calculated by gene counting. Student's t test was used for statistical comparisons. A value of p < 0.05 was considered to be statistically significant. All tests in the present study were two-sided.

Results

Expression of UGT1A3s. In the present study, five UGT1A3 sequences were obtained by mutagenesis and confirmed by sequencing. Their amino acid sequences correspond to those of all known UGT1A3s in the population: UGT1A3.1, UGT1A3.2, UGT1A3.3, UGT1A3.4, and UGT1A3.5. To investigate the catalytic activity of the UGT1A3 variants in vitro, we established a Bac-to-Bac insect expression system, which has been used for functional characterizations of many UGT alleles (Jinno et al., 2003) and applied in our laboratory for expression of UGT1A3, UGT1A4, UGT1A6, UGT1A9, and UGT2B7 (Qian et al., 2004; Zheng et al., 2004; Chen et al., 2005a). The optimized expression conditions were a multiplicity of infection value of 5 and an infection time of 72 h. Typical concentrations of cell homogenates were 15 mg/ml∼18 mg/ml. Expressions were finally confirmed by Western blot. We used anti-his as the primary antibody, so that the immunoreactivity of UGT1A3s would not change with mutations. The linear range of Western quantification was 0.2 mg∼0.5 mg of cell homogenate. We blotted one sample of each UGT1A3 on the same membrane. Three individual membranes were measured and confirmed. According to DNAStar software (DNASTAR Inc., Madison, WI), the molecular mass of UGT1A3 was 60 kDa. Protein bands of approximately 60 kDa were detected in all expression models except the negative control transfected with blank bacmid. As shown in Fig. 2, expression levels of UGT1A3 variants were 0.782 to 1.186 times as many as that of UGT1A3.1. These data were used for the normalization of glucuronidation activity.

Glucuronidation of UGT1A3s. At pH 8.0, the glucuronidation catalyzed by recombinant enzymes reached its maximal velocity, and then showed a plateau. Alamethicin showed a positive effect on the glucuronidation at the concentration of 12 μg/mg pro∼30 μg/mg pro, whereas Brij 58 had no detectable effect on it. In addition, d-saccharic acid 1,4-lactone monohydrate was used to inhibit the β-glucuronidase activity in the incubation, although little β-glucuronidase activity was detected in the Sf9 preparation, which is different from the liver microsome (O'Leary et al., 2001). The glucuronidation catalyzed by UGT1A3.1 produced more than one monoglucuronide for each tested flavonoid. All products were confirmed by liquid chromatography-mass spectrometry. Glucuronidation was proportional to the concentration of cell homogenate up to 1.2 mg/ml. At 0.5 mg/ml, the reaction was linear with time up to 90 min. The apparent K_m, V_max, and catalytic efficiencies (V_max/K_m) shown in Tables 3,4,5 represent a composite of all glucuronides formed using quercetin, luteolin, or kaempferol as substrate. Glucuronidation activities of 100, 11, 23, 464, and 77% to quercetin were observed for UGT1A3.1, UGT1A3.2, UGT1A3.3, UGT1A3.4, and UGT1A3.5, respectively. It suggests statistically significant differences among five UGT1A3s (p < 0.05). The UGT1A3.4 also improved the function in glucuronidation of luteolin and kaempferol, with a relative efficiency (V_max/K_m) approximately 4 times that of UGT1A3.1, whereas UGT1A3.2 and UGT1A3.3 led to a considerable decrease of glucuronidation, to less than 20%. It explained the reason why only one glucuronide was identified in our previous activity assay of quercetin with UGT1A3.2 (Chen et al., 2005b). Compared with UGT1A3.2 and UGT1A3.3, UGT1A3.5 showed a relatively mild decrease in glucuronidation efficiency for all tested flavonoids.

It is interesting that regioselectivities of UGT1A3s to hydroxyl groups in quercetin were also significantly different (Table 6). UGT1A3.3 showed an obvious preference to 3′-glucuronide. It catalyzed 3′-glucuronidation with more than 5 times the efficiency of 7-glucuronidation. UGT1A3.2, UGT1A3.4, and UGT1A3.5 produced much less 4′-glucuronide and 3′-glucuronide than 7-glucuronide and 3-glucuronide. Especially for the latter two, almost no 4′-glucuronide and 3′-glucuronide were detected after incubation.

Distributions ofUGT1A3Alleles and SNPs inUGT1A3in a Chinese Han Population. The 125 genomic DNA samples from whole blood were screened by ASA to examine the distribution frequency of UGT1A3 alleles and SNPs in UGT1A3 in a Chinese Han population (Tables 7 and 8). Sequences of 10 positive controls were identical with the results of sequencing. The prevalence of T31C, G81A, and T140C mutants was 26.8%, 26.8%, and 10.4% of donors, respectively, percentages that are significantly different from those reported for German-Caucasians (p < 0.05) (Ehmer et al., 2004). German-Caucasians have much higher distribution frequencies (65% to T31C, 65% to G81A, and 58% to T140C) than the Chinese Han population has, whereas the other three SNPs (A17G, C133T, and A477G) reported in Japanese (Iwai et al., 2004) but not in German-Caucasians were also detected in the Chinese Han population. Significant distribution differences of UGT1A3 alleles were observed between the German-Caucasian and Chinese Han populations except UGT1A3*3a. These findings reveal that the types and incidence of UGT1A3 alleles and SNPs in UGT1A3 have significant race differences. Such a difference is relatively weak between Japanese and the Chinese Han population, with C133T having a significant difference (p < 0.05) in the SNP distribution and UGT1A3*4 showing a statistically significant frequency difference between the Chinese Han population and Japanese. In the previous report (Iwai et al., 2004), the 477 substitution is always found to be together with two other SNPs (T31C and G81A) in the Japanese population, whereas some subjects without the 477 substitution also show 31 and 81 mutations in a Chinese Han population. This result leads to the significant distribution difference of UGT1A3*2a and UGT1A3*3a between the Chinese Han population and Japanese.

Discussion

Genetic polymorphisms have been reported for several functional human UGT genes (Maruo et al., 2005), including UGT1A1 (Bosma, 1995; Maruo et al., 1999), UGT1A3 (Ehmer et al., 2004; Iwai et al., 2004), UGT1A4 (Ehmer 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). These genetic variants may cause different phenotypes and play an important role in drug efficacy and xenobiotic toxicity, as well as in hormonal regulation and pathogenesis. For example, mutant UGT1A1 alleles lead to inherited diseases of bilirubin metabolism (Ritter et al., 1992; Bosma et al., 1994; Labrune et al., 2002). UGT1A6, UGT1A7, UGT1A9, UGT2B7, and UGT2B15 polymorphisms have been proposed for toxicological and clinical significance (Bhasker et al., 2000; Vogel et al., 2001; Miners et al., 2002; Carlini et al., 2005; Chung et al., 2005). Novel UGT1A10 polymorphisms are associated with orolaryngeal carcinoma risk (Elahi et al., 2003). Recently, in German-Caucasians and in a Japanese population, six UGT1A3 SNPs have been identified (Ehmer et al., 2004; Iwai et al., 2004). Some UGT1A3 variants show significantly different activities in estrone metabolism. However, their activities toward other substrates have not been determined. Since flavonoids are common constituents of foods of plant origin and display positive effects on many important diseases, such as cardiovascular diseases and cancers (Galati et al., 2000; Kris-Etherton and Keen, 2002; Vita, 2005), three widely existing flavonoids were used in this study as the substrates for glucuronidation to increase understanding of UGT1A3 variants.

Although statistically significant differences in catalytic activities of UGT1A3s to flavonoids were observed in the present study, results are not the same as observations reported for estrone glucuronidation (Iwai et al., 2004). The UGT1A3 variants clearly showed different activity alteration, depending on substrates. UGT1A3.5 exhibited a mild decrease in both flavonoids and estrone. UGT1A3.3 had a mild increase in estrone glucuronidation but a considerable decrease in flavonoid glucuronidation. A distinct activity difference was observed in UGT1A3.2 and UGT1A3.4. The former, having a 369% activity in estrone metabolism, showed a decreased activity in the metabolism of flavonoids to approximately 10%. The latter showed approximately 4 times glucuronidation efficiency to flavonoids of UGT1A3.1, with a decreased activity to 70% to estrone. These results indicate that amino acid residues at positions 45 and 47 are in a key region of UGT1A3 protein. Polymorphism research on UGT1A4, which shares more than 93% similarity in DNA sequence with UGT1A3 (Green et al., 1998), also indicated such an assumption. An L⁴⁸V UGT1A4 mutant completely lost dihydrotestosterone glucuronidation activity ((Ehmer et al., 2004) and gave twice the efficiency for clozapine as the wild-type UGT1A4. Its efficiencies for some important clinical drugs, trans- androsterone, imipramine, cyproheptadine, and tigogenin, also changed (Mori et al., 2005). Attention should be paid to mutations in this region in the structural and functional research of UGT1A3.

It is of interest to observe a regioselectivity variation in quercetin metabolism by the UGT1A3 variants. It has been reported that different quercetin glucuronides showed different free radical scavenging, antioxidant, antithrombotic, and anticarcinogenic activities (Moon et al., 2001; Silva et al., 2002). The quercetin 7-glucuronide and 3-glucuronide were the main active forms exerting the physiological activity in vivo (Morand et al., 1998). In the previous study, we have described that UGT1A3.1 played its minor role in quercetin physiological activities by preferring low-activity 3′- and 4′-glucuronides to high-activity 7-glucuronide and 3-glucuronide (Chen et al., 2005a). In the present work, our data showed that three variants (UGT1A3.2, UGT1A3.4, and UGT1A3.5) had much higher glucuronidation efficiency to the 7- and 3-hydroxyl groups of quercetin than to the other hydroxyls. Among these variants, the UGT1A3.4 may alter susceptibility to quercetin exposure, because UGT1A3.4 had 464% of the total glucuronidation efficiency of UGT1A3.1 and preferably produced high-activity 7-glucuronide, with efficiencies more than 10 times that to 3′- and 4′-hydroxyls.

Finally, we determined the frequency distributions of UGT1A3 alleles and SNPs in UGT1A3 in a Chinese Han population. A significant race difference in distributions was noted. UGT1A3*2, leading to decreased flavonoid glucuronidation, has a lower frequency in Chinese Han than in German-Caucasian, whereas UGT1A3*4, a key mutant in producing more high-activity quercetin glucuronides, is distributed more widely in the Chinese Han population than in German-Caucasians, but significantly less than in Japanese. Chinese Han may have a different susceptibility to flavonoid exposure from the Caucasian and Japanese and, hence, may affect human susceptibility to cancer and cardiovascular diseases, according to the beneficial effects of flavonoids to the prevention of these diseases (Galati et al., 2000; Kris-Etherton and Keen, 2002; Vita, 2005).

The present study demonstrated a significant difference in catalytic activities of the UGT1A3s to quercetin, luteolin, and kaempferol and, for the first time, determined the frequencies of UGT1A3 alleles and several UGT1A3 SNPs in Chinese Han subjects. The UGT1A3.4 showed a 464% increase in total glucuronidation efficiency and a clear preference to producing quercetin glucuronides with high biological activity. The statistically different distribution frequency of this variant allele in a Chinese Han population indicates that Chinese Han may have a different susceptibility to flavonoid exposure from the Caucasian and Japanese. The clinical relevance of these genetic variants of UGT1A3 remains to be assessed.