Abstract
The UGT1 complex codes for a subfamily of homologous “1A7-like” UDP-glucuronosyltransferases (UGTs), including UGT1A7 and UGT1A8. Little information is available regarding either the substrate specificities or regulation of the UGT1A7-like forms from rats. We compared the activities and tissue expression of UGT1A7 and UGT1A8, which exhibit 77% identity in their amino terminal sequence. UGT1A7 shows broad specificity, catalyzing the glucuronidation of 31 of 40 randomly selected substrates (100 μM) at rates >0.1 nmol/mg/min. UGT1A7 substrates included both planar and nonplanar compounds, mono- and polycyclic aromatics, and compounds with bulky side chain ring substitutions. UGT1A8 exhibited a narrower substrate specificity that completely overlapped with UGT1A7. UGT1A8 was most active toward the 1-OH, 4-OH, 5-OH, 6-OH, 7-OH, 10-OH, 11-OH, and 12-OH derivatives of benzo[a]pyrene. Other effective UGT1A8 substrates (>0.1 nmol/mg/min) included 9-OH-benzo[a]pyrene, 1-naphthol, 4-methylumbelliferone, 7-hydroxycoumarin, chrysin, quercetin, 4-nitrophenol, and estriol. In general, substrates preferred by UGT1A8 were polyaromatic planar structures with nonbulky substituents and a superimposable 1-naphtho ring structure. Studies of the tissue expression of the UGT1A7 and 1A8 mRNAs using RNase protection analysis suggested that each is expressed in liver and kidney of control rats. A major difference is the higher expression of UGT1A7 mRNA in intestine. These studies suggest complementary functions of the UGT1A7 and UGT1A8 forms in xenobiotic metabolism. Further studies are necessary to determine whether their relative contributions change as a function of development, hormonal status, or exposure to inducing agents.
Phase 2 enzymes play a critical role in the metabolism, detoxification, and elimination of xenobiotic chemicals from the body. UDP-glucuronosyltransferases are a family of phase 2 enzymes that catalyze the conjugation of xenobiotics with glucuronic acid, forming glucuronides. Glucuronide conjugates tend to have reduced biological activity, more restricted patterns of distribution, and enhanced rates of excretion from the body. In both humans and rodents, two families of UGTs1 are known. The UGT1A family is unique in using an exonsharing arrangement to encode a family of UGTs having identical C-terminal sequence (∼245 amino acids) (Ritter et al., 1992; Emi et al., 1995). Analysis of the various amino terminal coding exons of UGT1A reveals two distinct subclasses or clusters based on sequence similarity. One of these, the “cluster B” subgroup (Gong et al., 2001), includes UGT1A7 and the “UGT1A7-like” forms. Humans have four functional cluster B forms (UGT1A7, UGT1A8, UGT1A9, and UGT1A10) and three homologous but nonfunctional pseudogenes (1A11, 1A12, and 1A13) (Gong et al., 2001), whereas rats have three: UGT1A7, UGT1A8, and UGT1A10 (UGT1A9 is a pseudogene in rats) (Emi et al., 1995). A third cluster B exon (Emi et al., 1995), corresponding to the UGT1A9 exon, represents a pseudogene due to the presence of a frameshift mutation. The relationship of the human and rat UGT1A7-like forms based on their respective positions in the complex and the sequence homology of their encoded proteins is shown in Fig. 1. Although the substrate and tissue specificity of the human UGT1A7-like forms have been fairly well investigated (Strassburg et al., 1998, 1999; Cheng et al., 1999), few data are available in the literature for the cluster B forms from rat. This information would be useful for evaluation of the rat as a model for human drug and xenobiotic glucuronidation.
Rat UGT1A7 has been reported previously by our laboratory to catalyze glucuronidation of various phenol, diphenol, and dihydrodiol metabolites of the environmental carcinogen BaP (Grove et al., 1997; Bock et al., 1999) and the therapeutic analgesic drug acetaminophen (Kessler et al., 2002). In the latter study, recombinant expressed UGT1A8 also was found to catalyze activity toward acetaminophen but at 25-fold lower rate than UGT1A7. No other studies to our knowledge have been reported glucuronidating activities catalyzed by theUGT1A8 form. In view of the high sequence homology of the UGT1A7 and UGT1A8 forms of rat (77% identical, 83% similar), they might be predicted to share significant overlap in substrate specificity.
At the regulation level, UGT1A7 mRNA has been reported to be expressed in liver and many extrahepatic tissues, including kidney, intestine, ovary, spleen, and lung (Emi et al., 1995; Grove et al., 1997). In the liver and to a lesser extent the intestine and kidney, UGT1A7 is induced by exposure to polycyclic aromatic hydrocarbon-type inducing agents such as 3-methylcholanthrene (Emi et al., 1995; Metz and Ritter, 1998; Grams et al., 2000) and β-naphthoflavone (Kobayashi et al., 1998) and also by exposure to the dithiole thione oltipraz (Grove et al., 1997; Metz and Ritter, 1998). Conflicting results have been reported for tissue expression of the UGT1A8 mRNA in rats. Some laboratories have reported lack of detectable UGT1A8 mRNA expression in liver (Emi et al., 1995; Grams et al., 2000), whereas a third study reported expression of UGT1A8 mRNA in liver and kidney with lower levels in gastrointestinal tract (Shelby et al., 2003). The finding that UGT1A8 is expressed in rat hepatoma H4IIE cells provides indirect support for expression of this form in rat liver (Leung and Ho, 2002).
In the current study, our major objective was to compare the substrate specificities of the UGT1A7 and UGT1A8 forms from rat. Membranes from human embryonic kidney (HEK) 293 cells expressing recombinant UGT1A7 and UGT1A8 were used. In addition, the tissue specificity of UGT1A7 and UGT1A8 was studied using an alternative strategy based on RNase protection analysis (RPA).
Materials and Methods
Substrates and Reagents. 4-Hydroxyestradiol and 3,5-dinitrocatechol were from MP Biomedicals (Irvine, CA). All other chemicals were purchased from Sigma-Aldrich (St. Louis, MO). Uridine diphosphate-[14C]glucuronic acid (326 mCi/mmol) was obtained from MP Biomedicals.
Generation of Polyclonal Antisera to Rat UGT1 Common Region and Western Analysis of UGT1A7 and UGT1A8 Expressed in HEK Cells. A fusion protein coding for common region residues 360 to 455 of rat UGT1A7/UGT1A8 was expressed in Escherichia coli strain SG13008 harboring plasmid pQE30-r1CR. The fusion protein was purified and injected into female B6C3F1 mice to generate polyclonal antisera as described by our laboratory previously (Ritter et al., 1999). The specificity of the antibody was demonstrated by Western immunoblot analysis of a panel of rat UGT1A- and UGT2B-expressing cells. Protein (25 μg) from control or stable UGT-expressing HEK cells was subjected to electrophoresis through 7.5% SDS-polyacrylamide gels, transferred to nitrocellulose membranes, and probed as described previously (Ritter et al., 1999). The primary antisera were used at a dilution of 1:2000. The secondary antibody was horseradish peroxidase-conjugated sheep anti-mouse IgG antibody (Amersham Biosciences Inc., Piscataway, NJ) at a 1:20,000 dilution. Signal was detected using the ECL Western blotting detection system (Amersham Biosciences Inc.).
Sources of Recombinant Rat UGT1A7 and UGT1A8. Rat UGT1A7 cDNA (clone p7-5) was cloned from an oltipraz-induced male Sprague-Dawley rat liver cDNA library as described previously (Kessler et al., 2002). Rat UGT1A8 cDNA was obtained by a PCR-based strategy, in which the amino terminal coding exon and carboxyl terminal coding sequence were amplified in separate PCR reactions and joined in a third reaction, taking advantage of a 20-base overlap between the two fragments. The amino terminal coding sequence (876 base pairs) was amplified using a Sprague-Dawley rat UGT1A8 genomic clone (RGC-3) as template, and 5′-ACC CCA TCG ATA CCA TGG CTC CTT CAG GCT-3′ and 5′-CTT CAA ATT CCT TGG AAA GTG GCT TCC-3′ as the forward and reverse primers. The carboxyl-terminal coding fragment (∼1.4 kilobase pairs) was amplified using p7-5 UGT1A cDNA as template and 5′-ACT TTC CAA GGA ATT TGA AGC CTA TGT-3′ and the T7 primer (5′-GTA ATA CGA CTC ACT ATA GGG C-3′) as the forward and reverse primers. The proof-reading Pfu DNA polymerase was used for all steps in cloning of rat UGT1A8 cDNA. The fragments were joined by denaturing and annealing 1 ng of each fragment at 52°C and extending the product at 68°C. The final product (∼2.2 kilobase pairs) was amplified using 5′-ACC CCA TCG ATA CCA TGG CTC CTT CAG GCT-3′ and the T7 sequence as primers and subsequently cloned into pBluescript SK+ (designated p8-1 clone). The sequence of the UGT1A8 cDNA was confirmed using an ABI PRISM model 377 automated fluorescence sequencer in the Nucleic Acids Synthesis and Analysis core of the Massey Cancer Center at Virginia Commonwealth University.
Activity Assays. Crude membrane preparations from clonal HEK cells stably expressing recombinant rat UGT1A7 and UGT1A8 were prepared as described previously (Guillemette et al., 2000). Negative control membranes were prepared from untransfected HEK cells. UGT assays (50-μl total volume) contained 50 mM Tris-HCl, pH 7.5, 5 mM MgCl2, 25 μg of crude membrane protein, 1.5 mM UDP-glucuronic acid containing 0.1 μCi of UDP-[14C]glucuronic acid, and 100 μM substrate. Concentrated stocks of most substrates were prepared in dimethyl sulfoxide to achieve a final concentration of dimethyl sulfoxide of 0.1% in the assay. Bisphenol A and buprenorphine were dissolved in ethanol and methanol, respectively, to give final alcohol concentrations of 0.1% in the assay. Reactions were incubated at 37°C for 60 min and analyzed for formation of glucuronide conjugates using silica thin-layer chromatography as described previously (Bansal and Gessner, 1980; Guillemette et al., 2000). For some substrates (BaP metabolites, epicatechin, and bisphenol A), an alternative mobile phase (Nemoto and Gelboin, 1976) gave improved chromatography of radioactive glucuronides.
Analysis of UGT1A7 and UGT1A8 mRNAs in Rat Tissues. Total RNA was isolated from liver, kidney, and intestine of adult male Sprague-Dawley rats as described previously (Chomczynski and Sacchi, 1987) and assayed for UGT1A7 and 1A8 mRNAs by RPA. Specific probes for UGT1A7 and UGT1A8 were synthesized from plasmids pr1A7-223 and pr1A8-262, respectively, containing bases 48 to 261 of UGT1A7 and 194 to 455 of UGT1A8. Plasmids containing the rat cyclophilin and glyceraldehyde 3-phosphate dehydrogenase control RNAs were obtained from Ambion (Austin, TX). Probe synthesis, hybridization, and analysis were carried out essentially as described for analysis of the rat UGT1A6 mRNA (Auyeung et al., 2001).
Results
Generation of Antisera to Rat UGT1 Common Region. For the study of UGT1A expression and regulation in rats, an antisera was developed that specifically recognizes UGT1A-derived UGTs. Figure 2A shows the sequence of the bacterially expressed r1CR fusion protein that was used as antigen (UGT1A common region residues 360-455). Testing of the resulting antisera revealed strong reactivity with UGT1A family members with no detectable reactivity toward UGT2B forms, even after long exposure times (Fig. 2B). Reprobing of the blot with a nonspecific polyclonal antibody 1168 raised against purified rat UGT (Mackenzie et al., 1984) demonstrated the presence of UGT2B isozymes in the membranes from UGT2B-expressing cells. The apparent high specificity is consistent with the low overall similarity (59-64%) of UGT1A residues 360 to 455 to UGT2B isoforms, as illustrated for UGT2B3 (Fig. 2A). We have observed that this antisera shows improved specificity for UGT1A isoforms in the analysis of complex tissue homogenates from rats.
Cloning of UGT1A8 cDNA and Expression in HEK Cells. Since we were unable to obtain a copy of the UGT1A8 cDNA by standard cDNA library screening methods, a PCR strategy was devised as described under Materials and Methods. Sequencing of the final full coding product (pSK+ clone p8-1) revealed a cluster of differences spanning codons 76 to 81 compared with the previously reported UGT1A8 genomic DNA sequence (GenBank accession no. D38063). The sequence of clone 8-1 encodes Thr76 Tyr77 Ser78 Val79 Ser80 Tyr81 and seems authentic, based on sequence confirmation of an independent UGT1A8 genomic clone (clone RGC-7). Our finding that UGT1A7 contains five of the six residues at the corresponding position further supports this conclusion.
Given the high sequence homology of the UGT1A7 and UGT1A8 forms (Fig. 1), it was of interest to compare their glucuronidating specificities. The cDNAs were stably expressed in HEK cells, and resulting membrane fractions were analyzed by Western immunoblotting with the anti-rat UGT1360-455 antisera as shown in Fig. 2C. UGT1A7 and UGT1A8 exhibit similar electrophoretic mobilities. Densitometric analysis indicated a 4.3-fold higher UGT1A7 expression level relative to UGT1A8. The results for subsequent activity analyses accounted for this difference in expression level by multiplying the observed UGT1A8 activity by a normalization factor of 4.3.
The UGT1A7 and UGT1A8 cell membranes were tested for activity toward a group of 40 substrates, including BaP phenols and diols and other phenols and carboxylic acids. Metabolism by untransfected HEK cell membranes was assessed as a negative control. These data are presented in Table 1. Two of the most salient findings are 1) the broad activity of the UGT1A7 isoform toward most of the substrates tested; and 2) a completely overlapping but much narrower specificity of the UGT1A8 form, particularly for planar aromatic phenols containing multiring systems. UGT1A7 exhibited >0.1 nmol/mg/min glucuronidating activity toward 31 of the 40 substrates tested. Substrates for UGT1A7 include the simpler aromatic phenols, 1-naphthol and p-nitrophenol, which are commonly used as general markers for UGT activity (Fig. 3). In contrast, UGT1A8 was active toward only 17 substrates at the >0.1 nmol/mg/min level. The most effective UGT1A8 substrates were BaP phenols (10-OH, 5-OH, 12-OH-, 7-OH, 11-OH, 1-OH, 4-OH, and 6-OH benzo[a]pyrene) and 1-naphthol. The narrower specificity of the UGT1A8 form compared with UGT1A7 is exemplified by its more selective glucuronidation of 1-naphthol versus 4-nitrophenol (Fig. 3). Like UGT1A7, UGT1A8 exhibited significant activity toward the coumarins (4-methylumbelliferone and 7-OH coumarin) and the flavonoid chrysin. But in contrast to UGT1A7, UGT1A8 had little or no activity toward any of the estrogens (estriol, 4-OH estradiol, estrone, 17β-estradiol, and 17β-ethynylestradiol) or phenols with bulky ring substituents (acetaminophen, propofol, phenol red, bisphenol A, and 4-OH biphenyl). All UGT1A8 substrates identified in this study were found to also be efficient substrates of UGT1A7.
Relative Expression of UGT1A7 and UGT1A8 mRNAs in Liver, Intestine, and Kidney. Tissue levels of the UGT1A7 and UGT1A8 mRNA were studied in liver, kidney, and intestine from adult male rat using the RPA technique. UGT1A8 mRNA was detected using an antisense probe spanning the XcmI and BamHI sites (bases 194-455), which share 83% identity with the UGT1A7 sequence. A band corresponding in size to the predicted length of the full-protected UGT1A8 probe (262 bases) was visible in the liver and kidney samples using the r1A8 probe (Fig. 4). The 1A8 band was slightly more abundant in liver. Intestinal UGT1A8 was below the limit of detection. A similar analysis performed with the specific antisense probe for UGT1A7 (bases 48-261) provided contrasting results. In contrast, UGT1A7 expression was most highly abundant in intestine with less detected in liver and kidney (Fig. 4, 223-base product). The results with the UGT1A7 probe are supported indirectly by data from the UGT1A8 analysis (Fig. 4, bottom). The sequence of the UGT1A8 probe contains a 104- to 108-base region with high identity to UGT1A7 (data not shown). Bands that seem to correspond to this region of 1A8 probe cross-hybridization to the UGT1A7 mRNA are visible in the 104- to 108-base region of the ladder. The results support abundant expression of UGT1A7 mRNA in intestine with lower levels in liver and kidney. Data for two control probes demonstrated that our analysis involved equivalent amounts of mRNAs for the three tissues (316-base product for glyceraldehydes 3-phosphate dehydrogenase and 105-base product for cyclophilin mRNA).
Discussion
The rat is one of the most extensively used and investigated laboratory animal models in pharmacological and toxicological studies of xenobiotic metabolism. Since phase 2 enzymes such as the UGTs are presumed to be critical determinants of the overall kinetics of chemicals and/or their metabolites, it is imperative to understand the basis of glucuronidation in rats, to facilitate interpretation of pharmacological and toxicological data obtained using this species.
A main objective of this study was to characterize and compare the substrate-glucuronidating activities of the rat UGT1A7 and UGT1A8 isozymes. In agreement with the relatively high overall amino acid identity of UGT1A7 and UGT1A8 (77% identity), the two isoforms were found to exhibit overlapping substrate specificities. However, distinct differences were apparent in the types of compounds glucuronidated by these two isozymes.
Rat UGT1A7 enzyme has previously been shown to glucuronidate BaP phenols and diols (Grove et al., 1997), quinol metabolites of polycyclic aromatic hydrocarbons (Bock et al., 1999), and the analgesic drug acetaminophen (Kessler et al., 2002). The more extensive analysis performed in this study suggests that rat UGT1A7 is an unusually active and efficient UGT isoform with very broad and high catalytic activity toward a wide spectrum of xenobiotic compounds. As documented previously (Grove et al., 1997), UGT1A7 was highly active toward virtually every BaP monophenols and dihydrodiols tested. This finding alone suggests a more open or flexible substrate binding site for the UGT1A7 enzyme, capable of accommodating substrates in orientation(s) necessary for transfer of glucuronic acid to eligible functional groups. This suggestion is further supported by our finding that UGT1A7 is very active toward many bulky, nonplanar type substrates, such as bisphenol A and the estrogenic steroids such as 17β-ethynylestradiol and 17β-estradiol. Our result with bisphenol A contrasts with previous negative data reported for UGT1A7-catalyzed glucuronidation of this substrate (Yokota et al., 1999). In the latter study, the expression levels of UGT1A7 in the AH22 yeast cell model were not ascertained and may have been low, as suggested by low rates of 1-naphthol glucuronidation. UGT1A7 was also found in our study to be highly active toward the green tea flavonoid epicatechin, the glucuronidation of which has been reported to be species-specific. Rat but not human liver microsomes (Vaidyanathan and Walle, 2002) showed the capacity to catalyze epicatechin glucuronidation. Further studies with other rat UGTs are needed to determine whether epicatechin is a selective rat UGT1A7 substrate, but high selectivity of this form in epicatechin catalysis is supported by 1) high inducibility of rat liver microsomal epicatechin UGT by Arochlor 1254 (since UGT1A7 gene expression is under control of the arylhydrocarbon receptor) (Metz and Ritter, 1998), and 2) the established absence of UGT1A7 expression in human liver (Strassburg et al., 1997). The characteristically broad substrate specificity of UGT1A7 is also reflected in its high activity toward compounds containing single aromatic rings, for example, 4-nitrophenol, dinitrocatechol, and propofol.
With the exception of our report of UGT1A8 activity having low but detectable glucuronidating activity toward acetaminophen (Kessler et al., 2002), no other compounds to our knowledge have been tested as substrates for rat UGT1A8. Our data demonstrate significant UGT1A8 activity toward a number of compounds, but its specificity is considerably more restricted than that of UGT1A7. UGT1A8 was active toward several BaP phenol metabolites, particularly the 1-OH, 4-OH, 5-OH, 6-OH, 7-OH, 10-OH, 11-OH, and 12-OH phenol derivatives. It is noteworthy that each of these eight high activity BaP phenol substrates contains a superimposable 1-naphthol structure, whereas three of the four BaP metabolites found to be inefficient substrates for UGT1A8 lack this structural characteristic (Fig. 5). 1-Naphthol itself was a fairly effective substrate for UGT1A8. Although 2-naphthol was not tested, our data predict that this compound would be a less efficient UGT1A8 substrate. One exception to this apparent structural rule is 3-hydroxybenzo[a]pyrene, which possesses the 1-naphtho phenolic ring structural motif but was not found to be an efficient 1A8 substrate. This suggests that other structural aspects are important as well. Overall, factors that seem to favor high UGT1A8 activity include a high degree of planarity of the substrate, presence of multiple rather than single aromatic ring systems, and ring substituents that lie in the plane of the ring system. Nonplanar compounds such as the phenol red, bisphenol A, 4-hydroxybiphenyl, and the various estrogen derivatives were found to be inefficient substrates as were compounds with a single aromatic ring and/or bulky side chains (4-nitrophenol, propofol, acetaminophen, and dinitrocatechol).
The current study also examined the tissue specificity of expression of the mRNA for UGT1A8 in comparison with UGT1A7. Using the RNase protection technique, our data suggest that UGT1A7 and UGT1A8 are expressed at comparable levels in liver and kidney of control rats with slight differences apparent in their relative abundances. In the intestine, however, UGT1A7 mRNA was found to be highly expressed relative to UGT1A8, which was not detectable in intestine using this technique. Several previous studies have addressed the question of the relative tissue specificities and relative levels of expression of the UGT1A7-like subfamily mRNAs in rats. Our data most closely match those of Shelby et al. (2003), supporting the utility of the branched chain DNA methodology used in that study. Two additional studies, using a semiquantitative reverse transcription-PCR approach to address UGT1A mRNA expression in various tissues from rat (Emi et al., 1995; Grams et al., 2000), failed to detect UGT1A8 expression in any tissue. A sequence error affecting the sequence for 1A8 codons 76 to 81 (as described under Results) was recognized in our study and would be predicted to have prevented efficient annealing and amplification of the UGT1A8 mRNA in the earlier studies.
Although UGT1A8 mRNA in intestine was not detectable by RPA in our experiment and seemed to be absent from intestine in the study by Shelby et al. (2003), it is noteworthy that results from reverse transcription-PCR analysis in our laboratory (data not shown) indicated that the UGT1A8 mRNA is in fact expressed in intestine at an ∼5-fold lower level. In the analysis, differential expression of UGT1A8 in liver and kidney was evident, in agreement with our results obtained using RPA.
Our data raise questions about the relative abundance and contribution of UGT1A7 and UGT1A8 to xenobiotic glucuronidation in the three tissues under specific conditions, for example, as a function of development or exposure to different inducing agents. UGT1A7 expression is increased after exposure to arylhydrocarbon receptor agonist in both liver and intestine (Kobayashi et al., 1998; Metz and Ritter, 1998). A recent study using H4IIE rat hepatoma cells suggests that UGT1A8 may be regulated through this mechanism as well (Leung and Ho, 2002). The effects of other inducing agents on expression of these two forms in liver, kidney, and intestine remain largely uncharacterized and will require further study.
In conclusion, the current study provides initial results from characterization and comparison of the substrate specificities and expression patterns of rat UGT1A7 and UGT1A8. UGT1A7 exhibited very broad xenobiotic-glucuronidating specificity, including many nonplanar compounds with bulky side groups. UGT1A8 exhibited a much more restricted pattern of substrate specificity, including primarily planar substrates with nonbulky ring substituents. The relative roles that these UGTs play in glucuronidation of aromatic phenol and other substrates will depend on their relative levels of expression which in turn will depend on exposure status. In livers of uninduced rats, UGT1A8 and UGT1A7 each seem to be expressed at low levels.
Acknowledgments
We acknowledge Dr. Ida S. Owens (National Institutes of Health, Bethesda, MD) for the kind gift of 1168 antibody.
Footnotes
-
This work was supported by Grant R01ES07762 from the National Institute of Environmental Health Sciences.
-
doi:10.1124/dmd.104.001321.
-
ABBREVIATIONS: UGT, UDP-glucuronosyltransferase; BaP, benzo[a]pyrene; HEK, human embryonic kidney; RPA, RNase protection analysis; PCR, polymerase chain reaction.
- Received July 8, 2004.
- Accepted October 20, 2004.
- The American Society for Pharmacology and Experimental Therapeutics