Abstract
The UGT1A3-1A5 genes are a highly related UDP-glucuronosyltransferase (UGT) cluster exhibiting high levels of coding and regulatory region homology. However, the ensuing proteins have both differing substrate specificities and differing expression patterns. The expression profile of each enzyme also varies considerably from one individual to the next. Differences in UGT expression have been predicted to contribute to an individual's response to pharmaceuticals and to predisposition toward cancer in the event of carcinogen exposure. Therefore, it is desirable to elucidate the mechanisms that drive the transcription of UGT genes and identify the factors responsible for their variable expression. To this end, we have isolated the UGT1A3, UGT1A4, and UGT1A5 proximal promoters and begun to investigate the regulatory elements necessary for activity in vitro. We have established that the nucleotide sequence upstream of the UGT1A5 exon 1 is an ineffective promoter, correlating with the lack of substantial expression of this UGT in human tissues. In contrast, the UGT1A3 and UGT1A4 proximal promoters are both highly active in hepatic and colonic cell lines, with maximal activity being encoded by the proximal 500 base pairs. However, the UGT1A3 and UGT1A4 promoters exhibit low activity in the human embryonic kidney cell line HEK293, unless coexpressed with hepatocyte nuclear factor (HNF) 1α. Furthermore, mutation of the consensus-like HNF1-binding site in the UGT1A3 promoter abolishes promoter function in all cell types. This study suggests an important role for HNF1α in the transcriptional regulation of the human UGT1A3 and UGT1A4 genes.
Many small lipophilic molecules require conjugation with a polar moiety before they can be efficiently excreted from the body. Catalyzed by members of the UDP-glucuronosyltransferase (UGT) enzyme superfamily, glucuronidation is a predominant example of this metabolic process, and one that usually results in a loss of bioactivity of the parent compound. Therefore, the efficacy and/or toxicity of many xenobiotics (pharmaceuticals, environmental contaminants, and dietary compounds) and the fate of numerous endogenous signaling molecules (such as steroid hormones and neurotransmitters) can be regulated through glucuronidation (Radominska-Pandya et al., 1999; Tukey and Strassburg, 2000). Differences in UGT expression have been predicted to contribute to an individual's response to pharmaceuticals and to their predisposition toward developing certain cancers. Thus, it is desirable to elucidate the mechanisms that drive the transcription of UGT genes and identify the factors responsible for their variable expression.
UGT1A3, UGT1A4, and UGT1A5 are a trio of highly related proteins encoded by the human UGT1A locus. They share greater than 90% identity in their primary amino acid sequences and more than 85% nucleotide sequence identity in their proximal promoters to 1 kilobase (Green and Tephly, 1998; Gong et al., 2001). However, despite these similarities, this gene cluster varies considerably in both substrate specificity and expression pattern. Whereas UGT1A3 and UGT1A4 mRNA transcripts have been found in many tissues including liver, biliary tissue, colon, and breast (Strassburg et al., 1997, 1998a; Chouinard et al., 2006), UGT1A5 has not been found to be expressed to any significant extent in any tissues, although highly variable (but very low) expression in liver and gastrointestinal tract has been reported recently (Finel et al., 2005). Furthermore, there are inherent differences in the expression of UGT1A3 and UGT1A4; for example, UGT1A3 is found in the stomach of some individuals, whereas UGT1A4 is not (Strassburg et al., 1998b). Because the mechanisms that determine the expression levels and/or tissue specificity of the UGT1A3-1A5 cluster are currently not well understood, we have isolated their respective proximal promoters and begun to investigate the regulatory elements necessary for their activity in vitro.
Materials and Methods
Isolation of theUGT1A3andUGT1A4Promoters. The proximal 3.3 and 3.4 kilobases of the UGT1A3 and UGT1A4 promoters, respectively, were amplified by nested PCR from NotI digested human genomic DNA. In brief, PfuTurbo (Stratagene, La Jolla, CA) was used to simultaneously amplify both promoters using the primers GTTATCATTAAATAATAATCCT and GGCAGGGGAACCTGGAGTCCT, and the resulting PCR product was used as template to specifically amplify each promoter in separate reactions. The second round PCR primers were AGCCATAAGCTTATTGGATACCAGTATTGCT and AGCCATAAGCTTCTCAGCAGAAGACACGGACA for the UGT1A3 promoter, or AGCCATGCTAGCATCTGATATCAGTAATGTG and AGCCATCTCGAGCTCAGCAGAAGCCACCGACA for UGT1A4. Both promoters were cloned into pGL3-basic (Promega, Madison, WI) and their ends were sequenced to confirm their identity.
Generation ofUGT1A3andUGT1A4Promoter Deletion Constructs and Mutants. The pGL3-1A3-3.3k and pGL3-1A4-3.4k vectors were used as templates to clone the required deletion fragments of each promoter. The antisense primer sequences for all UGT1A3 and UGT1A4 fragments were AGCCATCTCGAGCTCAGCAGAAGACACGGACA and AGCCATCTCGAGCTCAGCAGAAGCCACCGACA, respectively. The sense primers annealed to bases –2541 to –2523, –1539 to –1519, –507 to –487, –200 to –184, –150 to –130, or –130 to –113 of the UGT1A3 promoter, relative to the translation start site. The three longest UGT1A4 promoter subfragments were amplified with the same sense primers as UGT1A3, giving lengths of 2610, 1574, and 506 nucleotides. For constructs containing less than 500 bases of the UGT1A4 promoter, PCR primers annealing to UGT1A4 nucleotides –200 to –184, –150 to –130, or –130 to –113 were used. To generate the mutated UGT1A3 150-bp construct, the UGT1A3 150-bp promoter was reamplified with the sense primer AGCCATGCTAGCGGCCAACGCTTCACTAGAGGA containing the desired mutations (in bold). All resulting PCR products were cloned into pGL3-basic and sequenced in full.
Isolation of theUGT1A5Promoter. The proximal 1.5 kilobases of the UGT1A5 promoter was also amplified using two sequential PfuTurbo reactions. The first round of PCR was performed on bacterial artificial chromosome 1308M2 from the human library RPCI-11 (BACPAC Resource Center, Children's Hospital Oakland Research Institute, Oakland, CA) using primers GAGGTCTTTAGACCACTTAGTC and GCTCCACACAAGACCTATGTATGAT. The resulting products were used as template to amplify the UGT1A5 1550-bp promoter using the same primers as for UGT1A4 1574 bp. The 508- and 150-bp fragments of the UGT1A5 promoter were also amplified using the corresponding UGT1A4 primers defined above. All fragments were ligated into pGL3-basic and sequenced in full.
Transient Transfection and Luciferase Reporter Assay. HepG2, Caco-2, and HEK293 cells were obtained from The American Type Culture Collection (Manassas, VA). All transfections were performed using Lipofectamine 2000 (Invitrogen, Carlsbad, CA) in 24-well plates seeded with 2 × 105 HepG2, 7.5 × 104 Caco-2, or 1 × 105 HEK293 cells the previous day. Either 0.5 μg of empty pGL3-basic or a reporter vector carrying UGT1A3, UGT1A4, or UGT1A5 promoter sequences was cotransfected with 0.25 μg of pCMX-HNF1α (Mackenzie et al., 2005) or empty pCMX vector. pRL-null (0.025 μg) was added to all transfections as an internal control for transfection efficiency. The reporter and expression plasmid concentrations were optimized to maximize promoter response without exceeding the recommended DNA/Lipofectamine ratio or inducing cytotoxicity (Gardner-Stephen and Mackenzie, 2005). After 48 h, cells were lysed in passive lysis buffer (Promega) and analyzed for firefly and Renilla luciferase activity using the Dual-Luciferase Reporter Assay System (Promega) and a TopCount luminescence and scintillation counter (Parkard, Mt. Waverly, Australia). As a minimum, all triplicate transfections were performed twice in independent experiments.
Electrophoretic Mobility Shift Assay. HNF1α electrophoretic mobility shift assay (EMSA), supershift, and competition assays were performed with Caco-2 and HepG2 nuclear extracts as described in Gardner-Stephen and Mackenzie (2005). The DNA probes used were UGT1A3-HNF1 (ATTAATGGTTAATAATTAACTAGAGG), UGT1A4-HNF1 (ATTAATGGGTAATAAGTAACTGGTGG), UGT1A3-HNF1mut (ATTAATGGCCAACGCTTCACTAGAGG), and the unrelated probe sequence FXR-consensus (GATCTCAAGAGGTCATTGACCTTTTTG). Underlined text indicates the extent of the putative HNF1-binding sites in each probe and mutations are highlighted in bold.
Results and Discussion
Relatively little is currently known about the transcriptional regulation of the UGT1A3-1A5 genetic cluster. Accordingly, we have initiated a detailed analysis of the UGT1A3-1A5 promoters and hereby describe a number of fundamental functional similarities and differences between these highly related sequences.
Basal Activities of theUGT1A3,UGT1A4, andUGT1A5Proximal Promoters. In both liver and colon-derived cell lines, the UGT1A3 and UGT1A4 130-bp promoters had minimal activity, exhibiting less than 3-fold increases over basal reporter gene expression by the promoter-less pGL3-basic vector. However, inclusion of an additional 20 nucleotides of either promoter substantially increased luciferase expression in both cell types (Fig. 1, A and B). Further increases in promoter activity could be obtained in either cell line by inclusion of up to 500 base pairs of the UGT1A3 or UGT1A4 promoters, with the exception of the UGT1A3 promoter in Caco-2 cells, which showed greatest activity at a length of 200 bp (Fig. 1B). The largest increases obtained in promoter activity for UGT1A3 were 53-fold in HepG2 and 40-fold in Caco-2 cells, whereas the UGT1A4 promoter had maximal activities of 21 and 15 times that of pGL3-basic in the same cell lines (Fig. 1, A and B). Increasing promoter length beyond 500 nucleotides resulted in reduced promoter activity for both UGT genes, although this phenomenon was more marked in HepG2 cells than in Caco-2. It was also found that, for all promoter lengths, the UGT1A3 gene had greater activity in vitro than UGT1A4, regardless of host cell type. Much of the increased activity of the UGT1A3 promoter over that of UGT1A4 was found to require nucleotides –150 to –200, although the transcription factors mediating this effect remain unidentified.
Comparison of the UGT1A5 promoter with the regulatory regions of UGT1A3 and UGT1A4 revealed that the former had the least activity. This was true for all three promoter lengths tested, in both HepG2 and Caco-2 cells (Fig. 2, A and B). The UGT1A5 promoter also differed from UGT1A3 and UGT1A4 in that the shortest fragment tested, 150 bp, was the most active. Increasing the promoter length to 500 nucleotides decreased promoter function in both HepG2 and Caco-2 cells, a result in direct opposition to that obtained for UGT1A3 and UGT1A4 (Fig. 1, A and B). UGT1A5 expression has not been detected at a substantial level in any human tissue to date, and it has been suggested that this may be due to lack of a functional promoter (Tukey and Strassburg, 2001). These results support this hypothesis and suggest that although the UGT1A5 core promoter is sufficient for assembly of a preinitiation complex, one or more crucial regulatory elements between –150 and –500 bp are missing, and/or that the UGT1A5 promoter contains negative regulatory sequences not present in UGT1A3 and UGT1A4.
HNF1 Is Required for Basal Activity of theUGT1A3andUGT1A4Proximal Promoters. For both UGT1A3 and UGT1A4, it was found that nucleotides –130 to –150 were important for basal activity in HepG2 and Caco2 cells. This region of the UGT1A4 promoter has previously been predicted to contain an HNF1-binding site (Tronche et al., 1997), based on the high identity (11 of 13 nucleotides) of this region with the HNF1-binding site consensus sequence GTTAATNATTAAC. Furthermore, the equivalent region of the UGT1A3 promoter contains a 100% match to the HNF1-binding site consensus. To date, however, no experimental evidence has been presented to ascertain whether these sites are functional in the context of their promoters. To investigate the influence of HNF1α on the activity of the UGT1A3 and UGT1A4 promoters, constructs containing 500 bp or less of each regulatory region were cotransfected with HNF1α into cells known to express HNF1 factors (HepG2 and Caco-2 cells) or a cell line devoid of HNF1 factors, namely HEK293 (Bernard et al., 1999; Gardner-Stephen and Mackenzie, 2005). In HEK293 cells, it was found that UGT1A3 and UGT1A4 promoters of sufficient length to include the putative HNF1-binding site were highly responsive to heterologous expression of HNF1α. Reporter gene expression under the control of the UGT1A3 or UGT1A4 promoters could be increased up to 22-fold over basal levels (Fig. 2C). In contrast, vectors containing only the most proximal 130 bp of the UGT1A3 or UGT1A4 promoters were completely unresponsive to the presence of HNF1α.
In HepG2 or Caco-2 cells, cotransfections of the UGT1A3 or UGT1A4 promoters with HNF1α resulted in little or no additional response (Fig. 2, A and B). Since HepG2 and Caco-2 cells express HNF1 factors (Kuo et al., 1990; Rey-Campos et al., 1991), we postulated that the endogenous levels of HNF1α and/or HNF1β in these cells were sufficient to support expression of the reporter gene from the UGT1A3 and UGT1A4 promoters in vitro. Therefore, we mutated the putative HNF1-binding site in the UGT1A3 150-bp promoter to abolish any binding of HNF1α. The functional result of this mutation was a loss of basal activity of the UGT1A3 150-bp promoter in both HepG2 and Caco-2 cells, and prevention of HNF1α responsiveness in HEK293 cells. In all cases, the mutated UGT1A3-150bp promoter construct behaved in the same manner as the UGT1A3 130-bp promoter that contains no recognized HNF1α-binding site (Fig. 2). In support of the above evidence that the UGT1A3 and UGT1A4 promoter HNF1 sites are functional, we were able demonstrate that HNF1 factors from Caco-2 and HepG2 nuclear extracts can bind these sequences in EMSA (Fig. 3). Furthermore, the mutation used to abolish HNF1α responsiveness of the UGT1A3 150-bp promoter also prevented binding of HNF1 factors to this region in vitro, whereas neither the mutated nor the unrelated (FXR) probes could interfere with HNF1 binding when added in 500-fold excess (Fig. 3).
TheUGT1A3andUGT1A4Proximal Promoters Differ in Their HNF1 Responses. During the course of this study, two notable differences between the UGT1A3 and UGT1A4 HNF1 responses were observed. First, whereas none of the UGT1A3 promoter constructs exhibited any activity in HEK293 cells in the absence of HNF1α, UGT1A4 promoters of 150 bp or longer could support a small degree of basal activity. This activity, which was 2- to 3-fold greater than the empty vector control, is presumably HNF1-independent (Fig. 2C). The second observation was that whereas UGT1A3 promoter activity could not be increased in HepG2 cells by overexpression of HNF1α, UGT1A4 promoter activity was increased up to 2.3-fold by excess HNF1α for promoter fragments ≥ 150 bp (Fig. 2A). One possible explanation is that the perfect UGT1A3 HNF1-binding element is fully occupied at physiological HNF1 concentrations, whereas the slightly flawed site of the UGT1A4 promoter is less efficient at competing with the multitude of genomic sites for limited HNF1. Therefore, addition of excess HNF1α into the system can only increase the occupancy rate of the UGT1A4 HNF1-binding site. There is also likely a cell type-specific component to this second difference between the promoters, as it was only observed in cells of hepatic origin (Fig. 2).
Summary. In summary, we have shown that whereas the nucleotide sequence upstream of the UGT1A5 first exon is ineffective as a promoter, the UGT1A3 and UGT1A4 proximal promoters are transcriptionally active in vitro. We have also shown that HNF1α can interact functionally with both the UGT1A3 and UGT1A4 promoters, and that the HNF1 site of the UGT1A3 gene is essential for transcriptional activity in vitro. In this regard, the UGT1A3 promoter appears to be most similar to UGT1A1 (Bernard et al., 1999), and distinct from that of the human UGT1A8-1A10 gene cluster (Gregory et al., 2003). Since HNF1 factors are present in cell types where UGT1A3 and UGT1A4 are expressed (Kuo et al., 1990; Rey-Campos et al., 1991), it would be reasonable to expect that HNF1α is important in UGT1A3 and UGT1A4 transcription in vivo.
Footnotes
-
This work was supported by grants from the Cancer Council of South Australia and the National Health and Medical Research Council of Australia. P.I.M. is a National Health and Medical Research Council Senior Principal Research Fellow.
-
Article, publication date, and citation information can be found at http://dmd.aspetjournals.org.
-
doi:10.1124/dmd.106.012203.
-
ABBREVIATIONS: UGT, UDP-glucuronosyltransferase; PCR, polymerase chain reaction; bp, base pair(s); HNF, hepatocyte nuclear factor; EMSA, electrophoretic mobility shift assay; FXR, farnesoid X receptor.
- Received August 2, 2006.
- Accepted October 17, 2006.
- The American Society for Pharmacology and Experimental Therapeutics