Identification of a genetic alteration in the code for bilirubin UDP-glucuronosyltransferase in the UGT1 gene complex of a Crigler-Najjar type I patient.

Patients with Crigler-Najjar syndrome (CN) type I inherit an autosomal recessive trait for hyperbilirubinemia, which is characterized by the total absence of bilirubin UDP-glucuronosyltransferase (transferase) activity. The recent identification of two bilirubin transferase isoforms with identical carboxyl termini (Ritter, J. K., J. M. Crawford, and I. S. Owens. 1991. J. Biol. Chem. 266:1043-1047) led to the discovery of a unique locus, UGT1, which encodes a family of UDP-glucuronosyltransferase isozymes, including the two bilirubin forms (Ritter, J. K., F. Chen, Y. Y. Sheen, H. M. Tran, S. Kimura, M. T. Yeatman, and I. S. Owens. 1992. J. Biol. Chem. 267:3257-3261). The UGT1 locus features a complex of six overlapping transcriptional units encoding transferases, each of which shares the four most 3' exons (2, 3, 4, and 5) specifying the 3' half of the transferase coding regions (condons 289-533) and the entire 3' untranslated region of each mRNA. This gene model predicts that a single critical mutation in any of these four "common" exons may inactivate the entire family of encoded transferases. In agreement with this prediction, we show here that in the first CN type I individual analyzed (patient F.B.), a 13-bp deletion has occurred in exon 2. Analysis of product generated by the polymerase chain reaction and genomic DNA demonstrated that F.B. is homozygous for the defective allele (UGT1*FB), and that the consanguineous parents are both heterozygotic at this locus. The mutation is predicted to result in the synthesis of severely truncated bilirubin transferase isozymes that are lacking a highly conserved sequence in the carboxyl-terminus and the characteristic membrane (endoplasmic reticulum)-anchoring segment of the protein molecule.

Identification of two single base substitutions in the UGT1 gene locus which abolish bilirubin uridine diphosphate glucuronosyltransferase activity in vitro.

Accumulating evidence indicates that mutations in the human UGT1 gene locus abolish hepatic bilirubin UDP-glucuronosyltransferase activity and cause the subsequent accumulation of bilirubin to toxic levels in patients with Crigler-Najjar type 1 (CN-I). Genetic and biochemical criteria are required to link CN-I with mutations in UGT1. Here we present analysis of mutations at the UGT1 locus in three individuals that were clinically diagnosed with CN-I (two related and one unrelated). Each patient carries a single base substitution that alters conserved residues in the transferase enzyme molecule, serine to phenylalanine at codon 376 and glycine to glutamic acid at codon 309. Each was homozygous for the defect as demonstrated by sequencing and RFLPs. Mutant cDNAs, constructed by site-directed mutagenesis, inserted into expression vectors, and transfected into COS-1 cells, supported the synthesis of the bilirubin transferase protein but only cells transfected with the wild-type cDNA expressed bilirubin UDP-glucuronosyltransferase activity. The data provide conclusive evidence that alterations at Gly 309 and Ser 376 are the genetic basis for CN-I in these families. These results suggest that the two codons, located in conserved regions of the molecule, are part of the active site of the bilirubin enzyme.

Coding defect and a TATA box mutation at the bilirubin UDP-glucuronosyltransferase gene cause Crigler-Najjar type I disease.

Mutations at the bilirubin UDP-glucuronosyltransferase (transferase) gene in a severely hyperbilirubinemic Crigler-Najjar (CN) type I individual was compared with that in a moderately hyperbilirubinemic CN II individual. The CN-I (CF) patient in this study sustained a TATA box insertional mutation which was paired with a coding defect at the second allele, unlike all coding defects previously seen in CN-I patients. The sequence of the mutant TATA box, [A(TA)8A], also seen in the CN-II patient, was compared with that at the wild-type box, [A(TA)7A]. Transcriptional activity with [A(TA)8A] was 10-15% that with the wild-type box when present in the -1.7 kb upstream regulatory region (URR) of the bilirubin transferase UGT1A1 gene which was fused to the chloramphenicol acetyl transferase reporter gene, pCAT 1.7H, and transfected into HepG2 cells. Also, a construct with a TA deletion, [A(TA)6A], was prepared and used as a control; transcriptional activity was 65% normal. The coding region defect, R336W, seen in CF (CN-I) was placed in the bilirubin transferase UGT1A1 [HUG-Br1] cDNA, and its corresponding protein was designated UGT1A1*32. The UGT1A1*32 protein supported 0-10% normal bilirubin glucuronidation when expressed in COS-1 cells. The I294T coding defect seen at the second allele in SM (CN-II) generated the UGT1A1*33 mutant protein which supported 40-55% normal activity with a normal Km (2.5 microM) for bilirubin. The hyperbilirubinemia seen in SM decreased in response to phenobarbital treatment, unlike that seen in CF. Parents of the patients were carriers of the respective mutations uncovered in the offspring. The TATA box mutation paired with a deleterious missense mutation is, therefore, completely repressive in the CN-I patient, and is responsible for a lethal genotype/phenotype; but when homozygous, i.e. paired with itself, as previously reported in the literature, it is far less repressive and generates the mild Gilbert's phenotype.

The novel bilirubin/phenol UDP-glucuronosyltransferase UGT1 gene locus: implications for multiple nonhemolytic familial hyperbilirubinemia phenotypes.

At least three types of congenital nonhemolytic unconjugated hyperbilirubinemias, including the rare Crigler-Najjar (CN) diseases (Types I or II) and Gilbert's syndrome (affecting 6% of the population) are associated with either absent or reduced hepatic UDP-glucuronosyltransferase (transferase) activity towards the potentially toxic endogenous acceptor, bilirubin. Here, we review the biochemical studies associated with these deficiencies. Accumulated evidence from studies with an animal model of CN Type I syndrome, the Gunn strain of hyperbilirubinemic rats, suggested that multiple isozymes are absent. These confounding observations have been clarified by a flurry of reports which have revealed the molecular basis for the complex disease phenotype in the Gunn rat and by the isolation and description of a novel human gene complex, UGT1, which encodes multiple and independently-regulated transferase isozymes that contain identical carboxyl terminal regions (246 amino acids). Finally, we discuss the implications of the gene organization and genetic defects determined for four different CN Type I individuals as a basis for a model which explains the inheritance pattern and genotypes of other familial unconjugated hyperbilirubinemias.

Genetic polymorphism in the human UGT1A6 (planar phenol) UDP-glucuronosyltransferase: pharmacological implications.

Two missense mutations were uncovered in the UGT1A6 (HLUG P1) cDNA which codes for a human phenol-metabolizing UDP-glucuronosyltransferase. The mutant and a wild-type UGT1A6 cDNAs were isolated from a custom synthesized human liver lambda Zap cDNA library. Both an A to G transition at nucleotide 541 (T181 A) and an A to C transversion at nucleotide 552 (R184S) occurred in exon 1 of the UGT1A6 (UGT1F) gene at the UGT1 locus. The two mutations on a single allele created a heterozygous genotype. Newly created BsmI and BsoFI sites at the T181 A and R184S locations, respectively, were confirmed by endonuclease treatment of PCR-generated DNA using the donor-liver genomic DNA as template. Screens with endonuclease treatment showed that 33/98 DNA samples were heterozygous with both mutations on one allele. One other individual also carried the R184S mutation on the second allele. Wild-type UGT1A6 generated a broad plateau of activity from pH 5.0 to pH 8.0 with certain experimental phenols, while activity was 1.3-2.5-fold higher at pH 6.4 than at pH 7.2 for others. UGT1A6*2 (181 A+ and 184S+) metabolized 4-nitrophenol, 4-tert-butylphenol, 3-ethylphenol/4-ethylphenol, 4-hydroxycoumarin, butylated hydroxy anisole and butylated hydroxy toluene, with the pH 6.4 preference, at only 27-75% of the rate of the wild-type isozyme whereas 1-naphthol, 3-iodophenol, 7-hydroxycoumarin, and 7-hydroxy-4-methylcoumarin were metabolized at essentially the normal level. Furthermore, UGT1A6*2 metabolized 3-O-methyl-dopa and methyl salicylate at 41-74% of that of the wild-type, and a series of beta-blockers at 28-69% of the normal level. This evidence suggests that the UGT1A6 enzyme activity is affected by different amino acids depending upon the substrate selection.

Thirteen UDPglucuronosyltransferase genes are encoded at the human UGT1 gene complex locus.

The original novel UGT1 complex locus previously shown to encode six different UDP-glucuronosyltransferase (transferase) genes has been extended and demonstrated to specify a total of 13 isoforms. The genes are designated UGT1A1 through UGT1A13p with four pseudo ones. UGT1A2p and UGT1A11p through UGT1A13p have either nucleotide deletions or flawed TATA boxes and are therefore pseudo. In the 5' region of the locus, the 13 unique exons 1 are arranged in a tandem array with each having its own proximal TATA box element and, in turn, are linked to four common exons to allow for the independent transcriptional initiation to generate overlapping primary transcripts. Only the lead exon in the nine viable primary transcripts is predicted to undergo splicing to the four common exons generating mRNAs with identical 3' ends and transferase isozymes with an identical carboxyl terminus. The unique amino terminus specifies acceptor-substrate selection, and the common carboxyl terminus apparently specifies the interaction with the common donor substrate, UDP-glucuronic acid. In the extended region, the viable TATA boxes are either A(A)TgA(AA)T or AT14AT; in the original locus the element for UGT1A1 is A(TA)7A and TAATT/CAA(A) for all of the other genes. UGT1A1 specifies the critically important bilirubin transferase isoform. The relationships of the exons 1 to each other are as follows: UGT1A2p through UGT1A5 comprises a cluster A that is 87-92% identical, and UGT1A7 through UGT1A13p comprises a cluster B that is 67-91% identical. For the two not included in a cluster, UGT1A1 is more identical to cluster A at 60-63%, whereas UGT1A6 is identical by between 48% and 56% to all other unique exons. The locus was expanded from 95 kb to 218 kb. Extensive probing of clones beyond 218 kb with coding nucleotides for a highly conserved amino acid sequence present in all transferases was unable to detect other exons 1. The mRNAs are differentially expressed in hepatic and extrahepatic tissues. This locus is indeed novel, indicating the least usage of exon sequences in specifying different transferase isozymes that have an expansive substrate range.

The UDP glycosyltransferase gene superfamily: recommended nomenclature update based on evolutionary divergence.

This review represents an update of the nomenclature system for the UDP glucuronosyltransferase gene superfamily, which is based on divergent evolution. Since the previous review in 1991, sequences of many related UDP glycosyltransferases from lower organisms have appeared in the database, which expand our database considerably. At latest count, in animals, yeast, plants and bacteria there are 110 distinct cDNAs/genes whose protein products all contain a characteristic 'signature sequence' and, thus, are regarded as members of the same superfamily. Comparison of a relatedness tree of proteins leads to the definition of 33 families. It should be emphasized that at least six cloned UDP-GlcNAc N-acetylglucosaminyltransferases are not sufficiently homologous to be included as members of this superfamily and may represent an example of convergent evolution. For naming each gene, it is recommended that the root symbol UGT for human (Ugt for mouse and Drosophila), denoting 'UDP glycosyltransferase,' be followed by an Arabic number representing the family, a letter designating the subfamily, and an Arabic numeral denoting the individual gene within the family or subfamily, e.g. 'human UGT2B4' and 'mouse Ugt2b5'. We recommend the name 'UDP glycosyltransferase' because many of the proteins do not preferentially use UDP glucuronic acid, or their nucleotide sugar preference is unknown. Whereas the gene is italicized, the corresponding cDNA, transcript, protein and enzyme activity should be written with upper-case letters and without italics, e.g. 'human or mouse UGT1A1.' The UGT1 gene (spanning > 500 kb) contains at least 12 promoters/first exons, which can be spliced and joined with common exons 2 through 5, leading to different N-terminal halves but identical C-terminal halves of the gene products; in this scheme each first exon is regarded as a distinct gene (e.g. UGT1A1, UGT1A2, ... UGT1A12). When an orthologous gene between species cannot be identified with certainty, as occurs in the UGT2B subfamily, sequential naming of the genes is being carried out chronologically as they become characterized. We suggest that the Human Gene Nomenclature Guidelines (http://www.gene.acl.ac.uk/nomenclature/guidelines.html++ +) be used for all species other than the mouse and Drosophila. Thirty published human UGT1A1 mutant alleles responsible for clinical hyperbilirubinemias are listed herein, and given numbers following an asterisk (e.g. UGT1A1*30) consistent with the Human Gene Nomenclature Guidelines. It is anticipated that this UGT gene nomenclature system will require updating on a regular basis.

Differences in UDP-glucuronosyltransferase activities in congenic inbred rats homozygous and heterozygous for the jaundice locus.

The genic transfer of the jaundice locus (jj) from the Gunn rat into the inbred RHA/++ rat produced congenic inbred homozygous RHA/jj rats which lacked detectable bilirubin UDP-glucuronosyltransferase activity. Congenic inbred RHA/j+ rats contained half the activity for bilirubin of the RHA/++ strain. Constitutive activities for glucuronidation of sixteen substrates of twenty-one tested were inherited additively. Approximately seven groups were discernible based on the defect in activity for these substrates in the RHA/jj strain. Activity for 1-hydroxybenzo[a]pyrene was, after that for bilirubin, the most severely reduced (188-fold), while no differences in the glucuronidation of three androgens and of the 6-hydroxy-, 10-hydroxy-, and 11-hydroxybenzo[a]pyrenes were observed. The conjugation of other substrates was affected to an intermediate extent. Most of the twenty-one glucuronidating activities were induced by phenobarbital in the RHA/jj strain as well as in the RHA/++ and RHA/j+ strains. Activities for 9-hydroxybenzo[a]pyrene and for the 2-hydroxy- and 4-hydroxybiphenyls were induced such that the defect was overcome, and the RHA/jj had the same level of activity as the RHA/++ strain. Cytochrome p-450 content and cytochrome c reductase and aminopyrine demethylase activities were unaffected in the congenic strains. Cytochrome p-450 content and cytochrome c reductase activity were induced approximately 2.5- and 2.0-fold, respectively, by phenobarbital while aminopyrine demethylase activity was induced about 30% in each strain. The congenic inbred rats should provide a stable and reproducible genetic model for studying defective UDP-glucuronosyltransferase specified by the jaundice (jj) locus.

Separation of different UDP glucuronosyltransferase activities according to charge heterogeneity by chromatofocusing using mouse liver microsomes. Three major types of aglycones.

Hepatic UDP glucuronosyltransferase (EC 2.4.1.17) (GT) enzymes in control, phenobarbital- and 3-methylcholanthrene-induced microsomes from C57BL/6N mice have been fractionated according to charge heterogeneity on a chromatofocusing system using a pH 9.5 to 6 gradient. Transferase activities for eleven different substrates were determined on column fractions. Activities toward 3-hydroxybenzo[a]pyrene, phenolphthalein and estrone (type 1 substrates) were enhanced by both effector compounds and always eluted primarily at pH 8.5. In control and phenobarbital-induced microsomes, activities toward testosterone, 4-hydroxybiphenyl, morphine, naphthol and 9-hydroxybenzo[a]pyrene (type 2 substrates) eluted primarily at about pH 6.7. Activities toward p-nitrophenol, 4-methylumbelliferone and 2-hydroxybiphenyl (type 3 substrates) in control and phenobarbital-induced microsomes exhibited two peaks which eluted at pH 8.5 and 6.7. 3-Methylcholanthrene treatment increased almost exclusively activities which eluted at pH 8.5 for each of the three types of substrates. The pH value of elution corresponds to the approximate isoelectric point of the eluted protein. Immunoabsorption studies with an antibody preparation raised against a purified low pI form confirmed that a 51,000-dalton transferase form, GTM1, eluted primarily at pH 6.7 and that a 54,000-dalton form, GTM2, eluted at pH 8.5. A mathematical treatment of the ratios of activity after 3-methylcholanthrene treatment to that after phenobarbital treatment versus pH produced six patterns of activity. A minimum of two enzymes at the low pH region and one enzyme at the high pH region, all with broad-substrate specificity, could account for these patterns.

Genetic defects at the UGT1 locus associated with Crigler-Najjar type I disease, including a prenatal diagnosis.

Characterization of the UGT1 gene complex locus encoding both multiple bilirubin and phenol UDP-glucuronosyltransferases (transferases) has been critical in identifying mutations in the bilirubin isoforms. This study utilizes this information to identify the bases of deficient bilirubin UDP-glucuronosyltransferase activity encoded by the UGT1A gene for the major bilirubin isozyme, HUG-Br1, in 3 Crigler-Najjar type I individuals and the genotype of an at-risk unborn sibling of one patient. A homozygous and heterozygous two-base mutation (CCC to CGT) created the HUG-Br1P387R mutant of the major bilirubin transferase in 2 different Crigler-Najjar type I patients, B.G. and G.D., respectively. Both parents of B.G. and his unborn sibling, J.G., were determined to be carriers of the P387R mutation. G.D. also contains the CAA to TAA nonsense mutation (G1n357st). Y.A. has a homozygous CT deletion in codons 40/41. The HUG-Br1P387R mutant protein was totally inactive at the major pH optimum (6.4), but retained 26% normal activity at the minor pH optimum (7.6), which was 5.4% of the combined activities measured at the two pH values.

UDPGT cDNA expression and UDPGT1 in human liver.

Following expression of UDPGTh1 and UDPGTh2 in Cos-1 cells, each isoform metabolized three types of dihydroxy- or trihydroxy-substituted ring structures, including the 3,4-catechol estrogen (4-hydroxyestrone), estriol and 17-epiestriol, and hyodeoxycholic acid (HDCA), but the UDPGTh2 isozyme was 100-fold more efficient than UDPGTh1. UDPGTh1 and UDPGTh2 are 86% identical overall (76 differences out of 528 amino acids), including 55 differences in the first 300 amino acids of the amino terminus, a domain which confers isoform substrate specificity. The data indicate a high level of conservation in the amino terminus is not required for the preservation of substrate specificity. Analysis of glucuronidation activity encoded by UDPGTh1/UDPGTh2 chimeric cDNAs constructed at their common restriction sites, Sac I (codon 279), Nco I (codon 385), and Hha I (codon 469), showed that nine amino acids between residues 385 and 469 are important for catalytic efficiency, suggesting that this region represents a domain which is critical for catalysis but distinct from that responsible for aglycon selection. Screening of leukocyte DNA cosmid library with human UDPGT-Br1 (1-470 bps) or UDPGT-Br2 (1-450 bps) resulted in three overlapping clones, which were isolated and mapped by endonucleases. Construction of subclones and DNA sequencing, Southern blot analysis revealed that a cluster of 4 exons (132, 88, 220, 1032 bps in one clone) encodes the entire region of 3' identity shared between human UDPGT-phenol, human UDPGT-Br1 and human UDPGT-Br2. A similar strategy but using probes which correspond to the unique regions of human UDPGT-Br1 and human UDPGT-Br2 showed that the exon 1 of UGT1A and UGT1D encodes the unique region of human UDPGT-Br1, and human UDPGT-Br2 and is located 5.6 and 49 Kb, respectively, upstream of the 4 common exons.

Cloning and expression of human liver UDP-glucuronosyltransferase cDNA, UDPGTh2.

The human liver cDNA clone UDPGTh2, encoding a liver UDP-glucuronosyltransferase (UDPGT) was isolated from a lambda gt 11 cDNA library by hybridization to mouse transferase cDNA clone, UDPGTm1. UDPGTh2 encoded a 529 amino acid protein with an amino terminus membrane-insertion signal peptide and a carboxyl terminus membrane-spanning region. There were three potential asparagine-linked glycosylation sites at residues 67, 68, and 315. In order to obtain UDPGTh2 protein encoded from cloned human liver UDP-glucuronosyltransferase cDNA, the clone was inserted into the pSVL vector (pUDPGTh2) and expressed in COS 1 cells. The presence of a transferase with Mr approximately 52,000 in transfected cells cultured in the presence of [(35)S]methionine was shown by immunocomplexed products with goat antimouse transferase IgG and protein A-Sepharose and analysis by sodium dodecyl sulfate-polyacrylamide gel electrophoresis and autoradiography. The expressed UDPGT was a glycoprotein as indicated by electrophoretic mobility shift in Mr approximately 3,000-4,000 when expressed in the presence of tunicamycin. The extent of glycosylation was difficult to assess, although one could assume that glycosyl structures incorporated at the level of endoplasmic reticulum were always the core oligosaccharides. Thus, it is likely that at least two moieties inserted can account for the shift of Mr approximately 3,000-4,000. This study demonstrates the cDNA and deduced amino acid sequence of human liver UDP-glucuronosyltransferase cDNA, UDPGTh2.

Comparison of glucuronidating activity of two human cDNAs, UDPGTh1 and UDPGTh2.

Two human liver UDP-glucuronosyltransferase cDNA clones, HLUG25 and UDPGTh2 were previously shown to encode isozymes active in the glucuronidation of hyodeoxycholic acid (HDCA) and certain estrogen derivatives (e.g., estriol and 3,4-catechol estrogens), respectively. In this study we have found that the UDPGTh-2-encoded isoform (UDPGTh2) and HLUG25-encoded isoform (UDPGTh1) have parallel aglycone specificities. When expressed in COS 1 cells, each isoform metabolized three types of dihydroxy- or trihydroxy-substituted ring structures, including the 3,4-catechol estrogen (4-hydroxyestrone), estriol, 17-epiestriol, and HDCA, but the UDPGTh2 isozyme was 100-fold more efficient than UDPGTh1. UDPGTh1 and UDPGTh2 were 86% identical overall (76 differences out of 528 amino acids), including 55 differences in the first 300 amino acids of the amino terminus, a domain which conferred the substrate specificity. The data indicated that a high level of conservation in the amino terminus was not required for the preservation of substrate selectivity. Analysis of glucuronidation activity encoded by UDPGTh1/UDPGTh2 chimeric cDNA constructed at their common restriction sites,Sac 1 (codon 297),Nco 1 (codon 385), andHha 1 (codon 469), showed that nine amino acids between residues 385 and 469 were important for catalytic efficiency, suggesting that this region represented a domain which was critical for the catalysis but distinct from that responsible for aglycone selection. These data indicate, that UDPGTh2 is a primary isoform responsible for the detoxification of the bile salt intermediate as well as the active estrogen intermediates.

Expression of human liver 3, 4-catechol estrogens UDP-Glucuronosyltransferase cDNA in COS 1 cells.

The human cDNA clone UDPGTh2, encoding a liver UDP-glucuronosyltransferase (UDPGT), was isolated from a lambdagt 11 cDNA library by hybridization to mouse transferase cDNA clone, UDPGTm1. The two clones had 74% nucleotide sequence identities in the coding region UDPGTh2 encoded a 529 amino acid protein with an amino terminus membrane-insertion signal peptide and a carboxyl terminus membrane-spanning region. In order to establish substrate specificity, the clone was inserted into the pSVL vector (pUDPGTh2) and expressed in COS 1 cells. Sixty potential substrates were tested using cells transfected with pUDPGTh2. The order of relative substrate activity, was as follows: 4-hydroxyestrone > estriol >2-hydroxyestriol > 4-hydroxyestradiol > 6alpha-hydroxyestradiol > 5alpha-androstane-3alpha, 11beta, 17beta-triol=5beta-androstane-3alpha, 11beta, 17beta-triol. There were only trace amounts of glucuronidation of 2-hydroxyestradiol and 2-hydroxyestrone, and in contrast to other cloned transferase, no gulcuronidation of either the primary estrogens and androgens (estrone, 17beta-estradiol/testosterone, androsterone) or any of the exogenous substrates tested was detected. A lineweaver-Burk plot of the effect of 4-hydroxyestrone concentration on the velocity of glucuronidation showed an apparent Km of 13 muM. The unique specificity of this transferase might play an important role in regulating the level and activity of these potent and active estrogen metabolites.