UGT1A1 Eliminated Cmpds

8/3/2019 UGT1A1 Eliminated Cmpds

1/11

Regulation of the UGT1A1 Bilirubin-Conjugating

Pathway: Role of a New Splicing Event

at the UGT1A LocusEric L evesque, Hugo Girard, Kim Journault, Johanie L epine, and Chantal Guillemette

UDP-glucuronosyltransferase 1A1 (UGT1A1) is involved in a wide range of biological andpharmacological processes because of its critical role in the conjugation of a diverse array ofendogenous and exogenous compounds. We now describe a new UGT1A1 isoform, referredto as isoform 2 (UGT1A1_i2), encoded by a 1495-bp complementary DNA isolated fromhuman liver and generated by an alternative splicing event involving an additional exonfound at the 3 end of the UGT1A locus. The N-terminal portion of the 45-kd UGT1A1_i2protein is identical to UGT1A1 (55 kd, UGT1A1_i1); however, UGT1A1_i2 contains aunique 10-residue sequence instead of the 99amino acid C-terminal domain ofUGT1A1_i1. RT-PCR and Western blot analyses with a specific antibody against UGT1A1indicate that isoform 2 is differentially expressed in liver, kidney, colon, and small intestineat levels that reach or exceed, for some tissues, those of isoform 1. Western blots of differentcell fractions and immunofluorescence experiments indicate that UGT1A1_i1 andUGT1A1_i2 colocalize in microsomes. Functional enzymatic data indicate thatUGT1A1_i2, which lacks transferase activity when stably expressed alone in HEK293 cells,acts as a negative modulator of UGT1A1_i1, decreasing its activity by up to 78%. Coimmu-noprecipitation of UGT1A1_i1 and UGT1A1_i2 suggests that this repression may occur viadirect proteinprotein interactions. Conclusion: Our results indicate that this newly discov-ered alternative splicing mechanism at the UGT1A locus amplifies the structural diversity ofhuman UGT proteins and describes the identification of an additional posttranscriptionalregulatory mechanism of the glucuronidation pathway. (HEPATOLOGY 2007;45:128-138.)

Glucuronidation, catalyzed by UDP-glucurono-syltransferase (UGT) enzymes, represents a ma-

jor phase II conjugation pathway in humans andis involved in the metabolism and excretion of several

endogenous and exogenous compounds. UGTs are mem-brane glycoproteins located in the endoplasmic reticulum(ER) that transfer the glucuronic acid moiety from UDPglucuronic acid to aglycone substrates, resulting in anincreased polarity of the substrate that facilitates its excre-tion from the body through bile and urine. This meta-bolic process is involved in the elimination of bilirubin,steroids, bile acids, toxic dietary components, and severaldrugs, including morphine, irinotecan, and mycopheno-late mofetil, to name a few.1-3

The UGTgene superfamily includes four mammalianUGT families, namely UGT1, UGT2, UGT3, andUGT8.4 Over the last few years, the UGT1 and UGT2gene structures have been extensively characterized in hu-mans; together, these genes encode 16 functional pro-teins.5,6 UGT1A comprises 17 exons, spans more than198 kb, is located on chromosome 2q37, and encodes 9functional UGT1A proteins. This gene is characterizedby 13 first exons 1 shared to common exons 2 to 5, thelatter encodes the C-terminal region of UGT proteins

with the UDPglucuronic acid binding and transmem-

brane domains.6-8 Each unique exon 1, which is tran-

Abbreviations: cDNA,complementary DNA; ER, endoplasmicreticulum;UGT,UDP-glucuronosyltransferase.

From the Laboratory of Pharmacogenomics, Oncology and Molecular Endocri-nology Research Center, CHUL Research Center and Faculty of Pharmacy, LavalUniversity, Quebec, Canada.

Received August 9, 2006; accepted September 29, 2006.Supported by the Canadian Institutes of Health Research and the Canada ResearchChair Program. Eric Levesque is a recipientof a CIHR fellowship award.Hugo Girardand Johanie Lepine are recipients of a studentship award from the CIHR. ChantalGuillemette is the chairholder of the Canada Research Chair in Pharmacogenomics.

The naming of the UGT1A1 isoform 2 (UGT1A1_i2 for protein andUGT1A1_v2 for gene product) was done according to the Human Gene Nomen-clature Guidelines and was approved by the UGT nomenclature committee.

Eric Levesque and Hugo Girard contributed equally to this study.Address reprint requests to: Chantal Guillemette, Ph.D., Laboratory of Pharmacog-

enomics, CHUL Research Center, T3-48, 2705 Boulevard Laurier, Quebec, Canada,G1V 4G2. E-mail: [email protected]; fax: 418-654-2761.

Copyright 2006 by the American Association for the Study of Liver Diseases.Published online in Wiley InterScience (www.interscience.wiley.com).DOI 10.1002/hep.21464

Potential conflict of interest: Nothing to report.

128


2/11

scribed by unique promoter for tissue-specific expression,encodes the amino-terminal half of the protein that im-parts aglycone specificity.

UGT1A1 is one of the most studied UGT enzymesdue to its major role in the biliary excretion of bilirubin, atoxic breakdown product of heme metabolism. Its physi-ological role is also exemplified by its involvement in theconjugation of steroid and thyroid hormones.9,10 Geneticpolymorphisms in the promoter region ofUGT1A1, forinstance, are associated with reduced transcriptional ac-tivity and result in Gilberts syndrome (mild unconju-gated hyperbilirubinemia). In turn, more significantlesions in UGT1A1 can lead to severe forms of hyperbil-irubinemia known as Crigler-Najjar type I and II.11-13

Additional evidence supports a critical role for UGT1A1in the metabolism of many therapeutic drugs. UGT1A1 isinvolved in the metabolism of the topoisomerase I inhib-

itor irinotecan, the topoisomerase II inhibitor etoposide,and the oral contraceptive steroid 17-ethinyl estradi-ol.14-16 Therefore, any genetic and/or environmental in-fluences that alter the glucuronidation activity ofUGT1A1 may have significant physiological and phar-macological consequences.

Recent preliminary data from our laboratory using aspecific antibody against human UGT1A1 highlight thepresence of immunoreactive proteins of lower molecular

weight in several human tissues. These data suggest thatadditional forms of this enzyme may exist. Furthermore, a

search of public databases revealed a shorter UGT1Acomplementary DNA (cDNA), which supports the exis-tence of a new class of UGT proteins. We describe theisolation and the characterization of a new UGT1A1 iso-form 2 (UGT1A1_i2) that is generated by alternativesplicing of an additional exon in the common 3 region ofUGT1A. We show that UGT1A1_i2 negatively regulatesthe function of the conjugating UGT1A1 enzyme(UGT1A1_i1, isoform 1).

Materials and Methods

Materials. UDP-glucuronic acid was obtained fromSigma (St. Louis, MO), geneticin (G418) was obtainedfrom Wisent Inc. (St.-Bruno, Canada), blasticidin wasobtained from Invitrogen (Carlsbad, CA), and LipofectinReagent was obtained from Stratagene (La Jolla, CA).Protein assay reagents were obtained from Bio-Rad (Rich-mond, CA). HEK293 cells were obtained from the Amer-ican Type Culture Collection (Manassas, VA). TotalRNA from human tissues was purchased from Ambion(Austin, TX). Four of the liver tissue samples used in thisstudy were kindly provided by Dr. Ted T. Inaba from the

University of Toronto.17All other tissues and microsomes

were purchased from Tissue Transformation Technology(Edison, NJ). Superscript II reverse transcriptase wasobtained from Invitrogen (Carlsbad, CA). Estradiol(E2) and its metabolites were purchased from Steraloids(Newport, RI).

Isolation of the Human UGT1A1_v2 cDNA.UGT1A1 variant 2 (UGT1A1_v2) was amplified by RT-PCR from total RNA obtained from human liver. TotalRNA (1g) was denatured in the presence of 100 pmol ofoligo(dT) primer at 65C for 15 minutes. The reaction

was performed at 42C in 20l containing Tris-HCl (pH8.3), 2.5 mM MgCl2, 10 mM dithiothreitol, and 0.5 mMof each dNTP. The RNA was incubated at 42C for 5minutes before the addition of 200 U of Superscript II(Gibco BRL). The reaction was then incubated at 42Cfor 50 minutes and at 70C for 15 minutes and chilled onice. Specific 3 oligonucleotides were designed for intron

4 of the common region ofUGT1A upstream of the pu-tative polyadenylation signals. The UGT1A1_v2cDNA was cloned via PCR amplification as follows.The PCR reaction (50 l) contained 2 mM MgCl2, 0.2mM of each dNTP, and 0.4 M of each primer (sense#1 483 5-GAGAGAAAGCTTCGAACCTCTG-GCAGGAGCAAA-3 and antisense #1484 5-GAGA-GACTCGAGTATCCAGTGCCACCACACACA-CATTAGCACCTCAAA-3). The sense primer wasdesigned to incorporate a HindIII restriction endonu-clease cleavage site 5 of the initiation ATG codon,

whereas the antisense primer contained a Xho1 site 3

of the stop codon. Finally, 2 U of Taq polymerase wereadded, and the reaction was incubated at 94C for 30seconds followed by 40 cycles of 94C for 30 seconds,55C for 30 seconds, and 72C for 1 minute, with afinal incubation at 72C for 5 minutes. The amplifiedproduct of 1,495 bp was separated on a 1% agarose geland digested with specific restriction endonucleases asmentioned above, subcloned into vectors pcDNA3 andpcDNA6, and sequenced.

Tissue Distribution of the UGT1A1_v1 andUGT1A1_v2 mRNA. Reverse transcription was per-formed in 20 l with 1 g of total human RNA and

with pdN6 random hexamer primers according to themanufacturers protocol (Invitrogen). First, using acommon oligonucleotide located in exon 2 in UGT1A(#1470 5-GAATTTGAAGCCTACATTAATGCT-TCTGGAGAACAT-3) along with a 3 oligonucleo-tide located in exon 5b for UGT1A_v2 (#1528 5-TCACATCTGTCTTCCTGACTGC-3) or exon 5afor UGT1A_v1 (#1529 5-TCAATGGGTCTTG-GATTTGTGG-3), we performed PCR reactions us-ing cDNA samples from tissues. PCR conditions were

95C for 1 minute for denaturation, followed by 40

HEPATOLOGY, Vol. 45, No. 1, 2007 LEVESQUE ET AL. 129


3/11

cycles at 95C for 30 seconds, 55C for 30 seconds, and72C for 1 minute, with a final extension at 72C for 7minutes. PCR products were purified on Qiagen quickcolumns (Qiagen Inc., Mississauga, Ontario, Canada)and sequenced. UGT1A1_v2 mRNA was amplified

with 1 l of the reverse transcription reaction withprimers #615 5-GAGAGAGGTGACTGTCCAG-GAC-3 and #1528, and UGT1A_v1 with primers#615 and #1529 using the same PCR conditions above.

Protein Expression Analysis. The presence ofUGT1A1_i1 and UGT1A1_i2 proteins was determined

with a specific polyclonal antibody against UGT1A1(Ab1A1#518) raised against the N-terminal portion (res-idues 63-144) of UGT1A118 and with a specific commer-cial UGT1A1 antibody (BD Gentest, San Jose, CA).Immunoreactive bands were visualized using a chemilu-minescence kit (ECL; Perkin Elmer, Woodbridge, On-

tario, Canada) with exposure to Kodak XB film. Therelative levels of UGT proteins were determined by inte-grated optical density using the Bioimage program Visage110S (Genomic Solution Inc., Ann Arbor, MI).

Stable Expression of UGT1A1 Isoforms 1(UGT1A1_i1) and 2 (UGT1A1_i2). HEK293 cells

were transfected with pcDNA3.1/UGT1A1_i1-His-Myc, pcDNA6/V5-HisA-UGT1A1_i2 and pcDNA3/UGT1A1_i2 expression plasmids as describedpreviously, using geneticin (1 mg/ml) or blasticidin (10g/ml).19 Several clones expressing both UGT1A pro-

teins (clones #2 and #10) were isolated from the initialpool to obtain different expression levels of these twoproteins in the same cell and to ensure stable expressionlevel over time.

Preparation of Microsomal Fractions and Enzy-matic Assays. Membrane fractions were prepared from800 106 cells of each population as described9with cellsdisrupted using 3 10 seconds of sonication. Enzymaticassays were performed using 20 g of proteins as previ-ously reported.20 Reactions were initiated by addition ofE2 (5, 25, 75, and 200 M) or bilirubin (200 M). Allbilirubin assays were performed under minimal light con-ditions for 1 hour, or 3 hours for estradiol, at 37C. Bili-rubin and estradiol assays were stopped with 100 l ice-cold methanol (0.02% butylated hydroxytoluene) and100l ice-cold methanol, respectively. Estradiol glucuro-nide measurements were performed as described previ-ously9 and activities are expressed as pmol glucuronides/min/mg protein. Bilirubin-glucuronide analysis wasperformed via high-performance liquid chromatography(Alliance 2690, Waters, Milford, MA) using a Luna C8100 4.6-mm 3-m column (Phenomenex, Torrance,CA) and eluted at a flow rate of 0.9 ml/min. The initial

conditions were 30% A (H2O, 1 mM ammonium for-

mate) and 70% B (methanol, mM ammonium formate)followed by a linear gradient up to 90% B in 2.5 minutes.The effluent from the high-performance liquid chroma-tography system was connected directly to an API 3000triple quadrupole mass spectrometer (Sciex, Toronto,Canada) equipped with a turbo-ion spray source with asplit of 1:4 in positive mode. The mass spectrometer wasoperated in the multiple reaction monitoring modes us-ing the following conditions: spray probe temperature at500C, ionization voltage at 5,000 V, and the orifice andthe ring at 45 V and 200 V, respectively. Data were ac-quired with a dwell time of 400 milliseconds, a pause timeof 5 milliseconds, and a scan time of 1.2 seconds. Biliru-bin glucuronidation activities were calculated as area/min/mg protein and are expressed as a percentage ofactivity versus UGT1A1_i1 activity. All enzyme activities

were subsequently divided by the UGT1A1_i1 content.

Immunofluorescence. Stable HEK-293/UGT1A1_i1/UGT1A1_i2 and control HEK-293 cells (75 103) wereused for these experiments as described.9 ForUGT1A1_i1, rabbit anti-c-myc primary antibody (Sig-ma-Aldrich) was used, and for UGT1A1_i2, mouse an-ti-V5 primary antibody (Invitrogen, Burlington, Canada)

was used; both antibodies were diluted 1:200. ForUGT1A1_i1 and UGT1A1_i2 a goat anti-rabbit second-ary antibody 1:500 (Alexa Fluor 488, green and AlexaFluor 594, red, respectively) were used. The expression ofthe ER resident protein calnexin was also assessed using a

rabbit anti-calnexin primary antibody (Stressgen Biotech-nologies, Victoria, Canada) diluted 1:200. Visualizationand image acquisition was achieved using a fluorescentmicroscope with a 100 oil objective coupled with adigital camera.

Coimmunoprecipitation Experiments. Protein GSepharose 4 fast flow (GE Healthcare, Piscataway, NJ)

was centrifuged at 12,000g for 20 seconds and washedthree times with 1 ml of high-salt buffer (500 mM NaCl,1% [vol/vol] IGEPAL, 50 mM Tris [pH 7.5]). At thefinal wash, the beads were incubated for 1 hour at 4C

with rocking in 1 ml of buffer. Finally, the solution wascentrifuged at 12,000gfor 20 seconds, and a 50% slurrymix was prepared by adding an equivalent volume ofhigh-salt buffer. Both solubilized and sonicated micro-somes were used for coimmunoprecipitation experi-ments. Solubilized microsomes were prepared asdescribed previously.21 Briefly, microsomes from HEK-293transfected cells were incubated for 30 minutes at4C in 25 mM Tris-HCl (pH 7.4) containing 0.8% (w/v)sodium cholate, 0.1 mM dithiothreitol, and 20% glyc-erol. The solution was centrifuged at 105,000g for 60minutes, and the supernatant was collected for co-immu-

noprecipitation experimentation. Microsomes (50 g)

130 LEVESQUE ET AL. HEPATOLOGY, January 2007


4/11

were mixed with 1 g of specific monoclonal antibody(Invitrogen, Burlington, Ontario, Canada) in 1 ml ofhigh-salt buffer and incubated at 4C with 50 l of pro-tein G Sepharose 4 fast flow (50% slurry) for 15 hours.The beads were washed 3 times with 1 ml high-salt buffer

and finally with 1 ml 50 mM Tris (pH 7.5). Beads con-taining the immunoprecipitated proteins were resus-pended with 30 l of1 SDS-PAGE solution, heated at100C for 2 minutes, and centrifuged at 12,000gfor 20seconds. The supernatant was analyzed via SDS-PAGE.The membrane blots were probed with a specific mono-clonal antibody linked with horseradish peroxidase (In-vitrogen, Burlington, Ontario, Canada).

Results

Identification of a New cDNA Encoding UGT1A1Isoform 2 (UGT1A1_i2). A cDNA that lacks the por-tion encoded by the terminal exon 5 at the UGT1A locus(GenBank AF297093) was recently isolated from humankidney (GenBank BC053576). The deduced amino acidsequence of the protein predicts a shorter C-terminal por-tion compared with other UGT1As. Based on this obser-vation, we screened the common region of UGT1A foradditional putative exons and identified an intronic ac-ceptor site followed by a 28-bp open reading frame asso-ciated with several downstream polyadenylation signals(GenBank AF297093). We subsequently designed spe-

cific oligonucleotides that recognize these regions to iso-

late potential UGT1A1 cDNAs via RT-PCR. In humanliver, a 1,495-bp UGT1A1 cDNA was isolated with anopen reading frame of 1,335 bp and a 3-untranslatedregion of 133 bp (GenBank DQ364247). The sequenceof a new shorter UGT1A1 cDNA, UGT1A1 variant 2

(UGT1A1_v2), corresponded to exons 1-4 flanked with apreviouslyunidentified 28-bp 3 sequence followedbya stopcodon (TGA) (31 bp total) (Fig. 1B, right panel). The se-quence from exon 2 to the 3 UTR was identical to thecDNA sequence BC053576. Comparison of the sequence

with the full-length UGT1A1 cDNA (UGT1A1_v1herein)revealed that UGT1A1_v2 lacked 298 bp (encoding 99amino acids) of thecommon exon 5. This portion isreplacedby a shorter sequence of 31 bp encoded by a new exon lo-cated in intron 4 (GenBank AF297093) (Fig. 1A). Thus, theexon 5 originally described by Ritter et al.7 was renamed

exon 5a, and the new exonlocated inintron 4 was referredtoas exon 5b (Fig. 1A).

The UGT1A1_v2 cDNA is predicted to encode a45-kd protein. Its deduced primary structure (Fig. 1B)reveals an ER-targeting signal peptide and complete bind-ing sites for UDP-GlcA and aglycone substrates; however,it also reveals a lack of a characteristic hydrophobic trans-membrane domain (residues 445-530) encoded by exon5a. The 10amino acid sequence (RKKQQSGRQM)encoded by the newly discovered exon 5b contains a typ-ical dilysine motif KKXX, presumably for ER retention,

as well as four positively charged residues that may pro-

Fig. 1. Identification of an additional exon in the common region ofUGT1A1 and isolation of the corresponding UGT1A1_v2 cDNA. (A) Schematic

representation of the UGT1A1 gene structure on chromosome 2q37. UGT1A1 exons 1-5a and b are presented with the novel exon 5b located 1091

bp downstream of exon 4. The two alternatively spliced UGT1A1 mRNAs are named UGT1A1_v1 and UGT1A1_v2. (B) Schematic representation of

the protein primary structures. Blue boxes correspond to the shared exons 1-4, whereas exons 5b and 5a are indicated in purple and yellow,

respectively.



5/11

mote the interaction of the protein with negativelycharged ER membranes.

Wide Expression Profile of UGT1A1_v2 in HumanTissues. Tissue distribution of UGT1A1 isoform 2 wasstudied via RT-PCR and Western blotting. Using a com-mon oligonucleotide located in exon 2 of UGT1A(#1470; Fig. 2A) associated with a 3 oligonucleotide lo-cated in the new exon 5b (#1528) or an oligonucleotide inexon 5a (#1529) (Fig. 2A), it was demonstrated that bothUGT1A amplicons are expressed in all the human tissuestested (Fig. 2B). We then used a specific UGT1A1 primerfor exon 1 (#615; Fig. 2A) combined with each of the 2primers for exons 5b and 5a, which yielded a 886-bpUGT1A1_v1 amplicon in the liver, colon, and small in-testine, in agreement with previous reports.22 On the

other hand, the 619-bp amplicon specific to the

UGT1A1_v2 transcript was expressed in the same tissuesas UGT1A1_v1 with the exception of the kidney, whichshowed only the short form for UGT1A1_v2. These datasuggest that both UGT1A1 spliced mRNAs are not al-

ways coexpressed in the same tissues.To correlate gene expression data with protein expres-

sion, we used a UGT1A1-specific antibody18 in additionto the commercially available anti-UGT1A1. We de-tected at least 2 forms of UGT1A1 of different molecular

weights (Fig. 3). Human liver microsomes expressed the55-kD UGT1A1_i1 and a protein of approximately 45kD, consistent with the predicted molecular weight ofUGT1A1_i2. In agreement with RT-PCR results, thekidney expressed only the 45-kD UGT1A1_i2 protein.

To evaluate the relative expression of both proteins in

individual liver samples and explore their subcellular lo-

Fig. 2. Distribution of UGT1A1_v1 and UGT1A1_v2 transcripts in human tissues. (A) Sense and antisense oligonucleotides used in RT-PCR

experiments are indicated with specific arrows, and the expected amplicons are illustrated. (B) RT-PCR was performed with the UGT1A-specific sense

#1470 (exon 2), antisense #1528 (exon 5b, UGT1A_v2 specific primer), and #1529 primers (exon 5a, UGT1A_v1 specific primer) in human samples

from five tissues. (C) UGT1A1-specific RT-PCR was performed with primers #615 (located in exon 1) and with the same antisense primers #1528

(UGT1A1_v2) and #1529 (UGT1A1_v1). All amplifications were sequenced and repeated at least three times for each tissue using independent

reverse transcription reactions from commercially purchased RNAs.



6/11


7/11

48-kD UGT1A1_i2 because a His-V5tag sequence (3

kD) was introduced in frame with the UGT1A1_i2 se-quence for further characterization experiments.UGT1A1_i1 is expressed significantly higher in micro-somes compared with homogenates (Fig. 5B). In contrast,UGT1A1_i2 is expressed more highly in homogenatescompared with microsomes, suggesting possible extrami-crosomal localization of the shorter UGT1A1_i2 proteinas inferred with human tissues. In UGT1A1_i1/_i2 cellmicrosomes, UGT1A1_i2 expression was 2-fold higherthan UGT1A1_i1.

We next assessed the enzymatic function ofUGT1A1_i2 expressed alone or coexpressed withUGT1A1_i1 using the UGT1A1 substrates bilirubin andestradiol. UGT1A1_i2 had no detectable transferase ac-tivity for either substrate (Fig. 6 A, B). Next, we used thestably cotransfected HEK293-UGT1A1 _i1_i2 cells totest if UGT1A1_i2 alters UGT1A1_i1 activity. In thepresence of UGT1A1_i2, UGT1A1_i2 transferase activ-ity for bilirubin and estradiol was significantly lower com-pared with UGT1A1_i1 microsomes from both stable celllines (Fig. 6). The inhibition of UGT1A1_i1 activity byUGT1A1_i2 for both bilirubin and estradiol varied from21%-78% and was significant for all substrate concentra-

tions tested. These differences in velocity cannot be ex-

plained by differences in microsomal UGT1A1_i1concentration because the data were normalized to theamount of UGT1A1_i1 (Fig. 6C).

Subcellular Localization of UGT1A1_i1 andUGT1A1_i2. To further explore the subcellular localiza-tion of UGT1A1_i2, a series of immunofluorescenceanalyses were performed with the 2 antibodies againstUGT1A1 (Ab#518) and anti-UGT1A (Ab#RC-71) us-ing HEK-293 cells stably expressing UGT1A1_i1 orUGT1A1_i2. Both proteins were expressed in the ER andperinuclear structure (data not shown). Colocalizationanalyses were then performed with anti-Myc and anti-V5for UGT1A1_i1 and UGT1A1_i2 visualization, respec-

Fig. 5. Western blot analysis of HEK-293 stable cell lines expressing

UGT1A1_i1, UGT1A1_i2, or both. pcDNA3-UGT1A1_i2 and pCDNA6/

HisA/V5-UGT1A1_i2 vectors were transfected into HEK-293 and HEK-

293-UGT1A1_i1/HisMyc cells, respectively. Homogenates (h) or

microsomal () proteins (20 g for each) obtained from stable cell lines

were separated on 10% SDS-PAGE gel for protein quantification. The

relative abundance (relative OD units) of each protein is shown in the

lower panel.

Fig. 6. Glucuronidation rates of (A) bilirubin and (B) estradiol (E2)

catalyzed by microsomes derived from UGT1A1_i1, UGT1A1_i2, and

UGT1A1_i1/UGT1A1_i2 HEK-293 stable cell lines. Microsomes were

incubated for 60 minutes (bilirubin) or 3 hours (estradiol) at 37C with

200 M bilirubin and various concentrations of E2 (5, 25, 75, and 200

M). Transferase activity values were normalized to the amount of

UGT1A1_i1. All data represent means SD of three independent

experiments performed in triplicate. P value 0.05.



8/11

tively, in cells expressing both proteins. Subsequent colo-calization analyses confirmed the colocalization ofUGT1A1_i1 and UGT1A1_i2 in the ER as well as in theperinuclear structure (Fig. 7).

Interaction Between UGT1A1_i1 and UGT1A1_i2.Based on the inhibitory effect of UGT1A1_i2 onUGT1A1_i1-mediated glucuronidating activity and theirpartial subcellular localization, we tested if UGT1A1_i1and _i2 interact by using coimmunoprecipitation assays

with microsomes from stable cell lines expressing bothisoforms. Western blotting showed a 48-kD protein cor-responding to UGT1A1_i2-His/V5, indicating thatUGT1A1_i2 coprecipitates with UGT1A1_i1 (Fig. 8,lane 5). Coimmunoprecipitation experiments withUGT1A1_i1 or UGT1A1_i2 revealed that anti-Myc didnot cross-react with UGT1A1_i2 (Fig. 8, lanes 3 and 4).These observations suggest that UGT1A1_i1 and

UGT1A1_i2 may interact directly.

Discussion

We report the isolation and characterization of a newUGT1A1 protein, UGT1A1_i2, generated by alternativesplicing of an additional exon in the common 3 region ofthe gene UGT1A. UGT1A1_i2 comprises 444 residues,

Fig. 7. Immunofluorescence localization of UGT1A1_i2 and

UGT1A1_i1. HEK-293 cells were stably transfected with UGT1A1_i1/_i2

isoforms, and the cells were incubated with antibodies specific for

UGT1A1_i1 or UGT1A1_i2 followed by fluorescently labeled secondary

antibodies. Nuclei are stained with DAPI (blue). (A) Subcellular localiza-

tion of UGT1A1_i1 (green). (B) Subcellular localization of UGT1A1_i2

(red). (C) Colocalization of both UGT1A proteins (yellow-orange). (D)

Localization of calnexin, an ER resident protein (green).

Fig. 8. Coimmunoprecipitation of UGT1A1_i1 and UGT1A1_i2. Microsomal fractions (50 g) obtained from UGT1A1_i1-His/Myc, UGT1A1_i2-

His/V5, and cell lines expressing both proteins were immunoprecipitated with () or without () 1 g of anti-Myc. The protein complexes were

visualized via western blotting with monoclonal anti-V5. Lanes 1 and 2: UGT1A1_i1-His/Myc. Lanes 3 and 4: UGT1A1_i2-His/V5. Lanes 5 and 6:

UGT1A1_i1-His/Myc/UGT1A1_i2-His/V5. Lanes 7 and 8: positive control immunoprecipitations with anti-His (1 g); lane 7, visualization of

UGT1A1_i2-His/V5; lane 8, visualization of UGT1A1_i2-His/V5. Lane 9: UGT1A1_i2-His/V5 (5 g) was loaded on the gel as positive control for the

western blot. Negative controls showed that neither UGT1A1_i1 nor UGT1A1_i2 precipitated in the absence of anti-Myc (lanes 2, 4, and 6).



9/11

with the first 434 residues identical to UGT1A1_i1; theconserved C-terminal domain encoded by exon 5a ofUGT1A1_i1, however, is replaced by a 10-residue se-quence encoded by an additional exon (5b) located withinintron 4. Despite these structural changes, UGT1A1_i2retains the consensus substrate and cosubstrate bindingdomainsencoded by exons 1-4that are characteristicof UGT proteins. UGT1A1_i2 resides mainly in the ER,and in human tissues it is widely expressed, although theexpression level varies between individuals. Functionalenzymatic data strongly imply that UGT1A1_i2, whichlacks transferase activity when expressed alone, acts as anegative modulator of the conjugating activity ofUGT1A1_i1 potentially through direct proteinproteininteractions.

We have shown that the UGT1A locus encodes oneadditional exon and that the common region of theUGT1A locus comprises 5 common exons (exons 2-5aand 5b) instead of 4 (exons 2-5), as described.6,7We fur-ther demonstrated that exons 5a and 5b are alternativelyspliced in human tissues, generating 2 different UGT1A1mRNAs, and that the levels of these transcripts vary sig-nificantly in different tissues and among individuals.These data suggest that other transcriptional regulatorymechanisms are involved in the expression of a shorterform of UGT1A1namely UGT1A1_i2in varioushuman tissues, in addition to those mechanisms that dic-tate tissue-specific expression for UGT1A family mem-

bers.25 In support of these results, we observed thatUGT1A1_i1 and UGT1A1_i2 mRNAs are differentiallyexpressed in human hepatic and extrahepatic tissues,namely the liver, colon, and small intestine (Fig. 2).

At physiological levels, UGT1A1 represents a criticalenzyme involved in the inactivation of bilirubin as well asseveral steroid hormone molecules and therapeutic drugs.The constitutive expression ofUGT1A1 is under the influ-ence of a common polymorphic repeat sequence (TA5-8)located in its regulatory region. The presence of the TA re-peat is now recognized as one of the key factors contributing

to interindividual and racial heterogeneity in UGT1A1 pro-tein levels and bilirubin-conjugating capacity in hu-mans.13,26 The most frequent allele, UGT1A1*28 (TA7), isassociated with a 30%-50% reduction in transcriptional ac-tivityandis linkedto mild hyperbilirubinemia, susceptibilityto cancer, and drug-induced toxicity.26-29 The reduction inUGT1A1_i1 conjugating activity, apparently induced bythe presence of the UGT1A1_i2 in vitro, suggests that theexpression of the isoform 2 is an additional determinant ofvariable glucuronidation capacity. We propose that the ex-tent of the role of this posttranscriptional regulatory mecha-

nism is dictated by the UGT1A1_i2 expression level.

Alternative splicing-induced variant proteins that exert re-pressive activity have been reported for many proteins.30,31

The tissue expression pattern for UGT1A1_i2 suggeststhat the functional consequences of its putative negativeregulatory function predominate in extrahepatic tissues.In the liver, the UGT1A1_i2 level in microsomes appearsto be minimal (10%) compared with UGT1A1_i1; thisproportionality probably is needed to preserve the high levelof glucuronidation activity that maintain homeostasis of keyfactors such as bilirubin. In contrast, UGT1A1_v2 seems tobe enriched in the kidney and in the proximal and distalsegments of the small intestine (jejunum and ileum). Thepresence of high levels of the UGT1A1_v2 form in the kid-ney remains undefined and deserves further investigation.Interestingly, the UGT1A1_v1 content decreases along thedistal gastrointestinal tract, whereas that of UGT1A1_v2 in-creases; this trend would be expected to effect a progressive

decline in transferaseactivity.Therefore, theinhibitoryeffectof UGT1A1_i2 on UGT1A1_i1 represents an additionalposttranscriptionalmechanism involved in thefinetuningofendogenous and exogenous metabolism in humans and ispredicted to be greater in extrahepatic tissues. It is proposedthat specific regulatory mechanisms may exist for the pro-cessing of both the 5 and 3 end regions ofUGT1A.

We found that UGT1A1_i2 has no detectable UGTactivity toward common UGT1A1_i1 substrates, sug-gesting that the amino acid sequence encoded by exon 5a(residues 435-530) is required for UGT1A1 transferase

activity. This requirement could reflect either a functionalor structural role. Several studies demonstrated that trun-cation of specific UGT domains, including the C-termi-nal domain, impairs UGT enzymatic activity but does notprevent binding to the ER.32,33 These studies also supportour observed ER localization of UGT1A1_i2 despite thelack of the transmembrane domain; we note, however,that a portion of the protein did not localize within theER (Fig. 7). The subcellular localization to the ER may beconferred, at least in part, by the dilysine motif KKXX,encoded by the new exon 5b, that can serve as an ERretention/retrieval signal.34,35 Another study also sug-gested that Asp446, encoded by exon 5 of rat UGT1A6, isessential for proper secondary structure, overall folding,and activity of the enzyme.36 Because the entire exon 5a ismissing in UGT1A1_i2, this isoform may fold differentlythan UGT1A1_i1, with a consequent loss of transferaseactivity.

Based on our findings, we hypothesize that UGTdimerization is one of the mechanisms involved in thereduced UGT1A1 glucuronidation activity caused byUGT1A1_i2. In support of this notion, several reportshave described the homodimerization of UGT1A1 en-

zymes.23,24,37 Futhermore, Ghosh et al.20 reported that



10/11

the interactions between UGT1A1 proteins are not abol-ished by partial deletion of the C-terminal domain andsuggested that homodimerization of UGT1A1 may ex-plain some of the dominant-negative effects of mutatedUGT1A1 proteins. Homodimerization of UGT1A1 pro-teins is also supported by the clinical description of aCrigler-Najjar patient with only one nonsense mutationin the coding region of the gene, inferring that the dys-functional protein had a dominant negative effect withsevere clinical consequences.38 Our results are in agree-ment with the concept of a dominant-negative effectthrough direct proteinprotein interactions betweenUGT1A1_i1 and UGT1A1_i2, as previously reported byGhosh et al.24 This hypothetical mechanism is plausiblebecause UGT1A1_i1 and UGT1A1_i2 both localize inmicrosomes (Fig. 4) where they potentially could interact.However, other mechanisms might also be responsible for

the decrease in UGT1A1_i1 transferase activity inducedby the presence of UGT1A1_i2. These two proteinscould form homodimers, which could prohibit properfolding of the catalytic domain of UGT1A1_i1. This, inturn, could accelerate the degradation of the protein,thereby altering the net transferase activity of the complextoward various substrates. In addition, UGT1A1_i2 maybind specific substrates and/or cofactors that are requiredfor UGT1A1_i1 activity.

In conclusion, the evolution of alternative splicingmechanisms at the UGT1A locus has served to amplify

the structural diversity of these biologically and pharma-cologically important proteins beyond the basic patternprovided by the distinct gene classes. It is expected thatthe alternative splicing mechanism occurring in the com-mon region of the gene generates several other novelUGT1A_i2 proteins expressed in various tissues. Ourpreliminary findings support this statement (unpublisheddata). Thus, UGT1A expression may involve highly reg-ulated and complex transcriptional and splicing mecha-nisms at the extremities of this locus, which may explainthe observed diversity among UGT proteins expressed

from this gene.Acknowledgment: We thank Dr. Ted T. Inaba from

the University of Toronto for kindly providing liver tissuefrom 4 individuals. We also thank Drs. Alain Belangerand Olivier Barbier for critical reading of the manuscriptand Olivier Bernard for helping with DNA preparations.

References

1. Wells PG, Mackenzie PI, Chowdhury JR, Guillemette C, Gregory PA,

Ishii Y, et al. Glucuronidation and the UDP-glucuronosyltransferases in

health and disease. Drug Metab Dispos 2004;32:281-290.

2. King CD, Rios GR, Green MD, Tephly TR. UDP-glucuronosyltrans-

ferases. Curr Drug Metab 2000;1:143-161.

3. Bullingham RE, Nicholls AJ, Kamm BR. Clinical pharmacokinetics of

mycophenolate mofetil. Clin Pharmacokinet 1998;34:429-455.

4. Mackenzie PI, Walter Bock K, Burchell B, Guillemette C, Ikushiro S,

Iyanagi T, et al. Nomenclature update for the mammalian UDP glycosyl-

transferase (UGT) gene superfamily. Pharmacogenet Genomics 2005;15:

677-685.

5. Riedy M, Wang JY, Miller AP, Buckler A, Hall J, Guida M. Genomic

organization of the UGT2b gene cluster on human chromosome 4q13.Pharmacogenetics 2000;10:251-260.

6. Gong QH, Cho JW, Huang T, Potter C, Gholami N, Basu NK, et al.

Thirteen UDPglucuronosyltransferase genes are encoded at the human

UGT1 gene complex locus. Pharmacogenetics 2001;11:357-368.

7. Ritter JK, Chen F, Sheen YY, Tran HM, Kimura S, Yeatman MT, et al. A

novel complex locus UGT1 encodes human bilirubin, phenol, and other

UDP-glucuronosyltransferase isozymes with identical carboxyl termini.

J Biol Chem 1992;267:3257-3261.

8. Strassburg CP, Oldhafer K, Manns MP, Tukey RH. Differential expres-

sion of the UGT1A locus in human liver, biliary, and gastric tissue: iden-

tification of UGT1A7 and UGT1A10 transcripts in extrahepatic tissue.

Mol Pharmacol 1997;52:212-220.

9. Lepine J, Bernard O, Plante M, Tetu B, Pelletier G, Labrie F, et al. Spec-

ificity and regioselectivity of the conjugation of estradiol, estrone, and their

catecholestrogen and methoxyestrogen metabolites by human uridinediphospho-glucuronosyltransferases expressed in endometrium. J Clin En-

docrinol Metab 2004;89:5222-5232.

10. Findlay KA, Kaptein E, Visser TJ, Burchell B. Characterization of the

uridine diphosphate-glucuronosyltransferase-catalyzing thyroid hormone

glucuronidation in man. J Clin Endocrinol Metab 2000;85:2879-2883.

11. Bosma PJ, Chowdhury JR, Huang TJ, Lahiri P, Elferink RP, Van Es HH,

et al. Mechanisms of inherited deficiencies of multiple UDP-glucurono-

syltransferase isoforms in two patients with Crigler-Najjar syndrome, type

I. FASEB J 1992;6:2859-2863.

12. Bosma PJ, Goldhoorn B, Oude Elferink RP, Sinaasappel M, Oostra BA,

Jansen PL. A mutation in bilirubin uridine 5-diphosphate-glucuronosyl-

transferase isoform 1 causing Crigler-Najjar syndrome type II. Gastroen-

terology 1993;105:216-220.

13. Bosma PJ, Chowdhury JR, Bakker C, Gantla S, de Boer A, Oostra BA, et

al. The genetic basis of the reduced expression of bilirubin UDP-glucu-

ronosyltransferase1 in Gilberts syndrome.N Engl J Med1995;333:1171-

1175.

14. Watanabe Y, Nakajima M, Ohashi N, Kume T, Yokoi T. Glucuronidation

of etoposide in human liver microsomes is specifically catalyzed by UDP-

glucuronosyltransferase 1A1. Drug Metab Dispos 2003;31:589-595.

15. Gupta E, Lestingi TM, Mick R, Ramirez J, Vokes EE, Ratain MJ. Meta-

bolic fate of irinotecan in humans: correlation of glucuronidation with

diarrhea. Cancer Res 1994;54:3723-3725.

16. Ebner T, Remmel RP, Burchell B. Human bilirubin UDP-glucuronosyl-

transferase catalyzes the glucuronidation of ethinylestradiol. Mol Pharma-

col 1993;43:649-654.

17. Sumida A, Kinoshita K, Fukuda T, Matsuda H, Yamamoto I, Inaba T, et

al. Relationship between mRNA levels quantified by reverse transcription-

competitive PCR and metabolic activity of CYP3A4 and CYP2E1 in hu-man liver. Biochem Biophys Res Commun 1999;262:499-503.

18. Duguay Y, McGrath M, Lepine J, Gagne JF, Hankinson SE, Colditz GA,

et al. The functional UGT1A1 promoter polymorphism decreases endo-

metrial cancer risk. Cancer Res 2004;64:1202-1207.

19. Guillemette C, Ritter JK, Auyeung DJ, Kessler FK, Housman DE. Struc-

tural heterogeneity at the UDP-glucuronosyltransferase 1 locus:functional

consequences of three novel missense mutations in the human UGT1A7

gene. Pharmacogenetics 2000;10:629-644.

20. Girard H, Villeneuve L, Court MH, Fortier LC, Caron P, Hao Q, et al.

The novel UGT1A9 intronic I399polymorphism appears as a predictorof

7-ethyl-10-hydroxycamptothecin glucuronidation levels in the liver. Drug

Metab Dispos 2006;34:1220-1228.

21. Taura KI, Yamada H, Hagino Y, Ishii Y, Mori MA, Oguri K. Interaction

between cytochrome P450 and other drug-metabolizing enzymes: evi-

dence for an association of CYP1A1 with microsomal epoxide hydrolase



11/11

and UDP-glucuronosyltransferase. Biochem Biophys Res Commun 2000;

273:1048-1052.

22. Tukey RH, Strassburg CP. Genetic multiplicity of the human UDP-glu-

curonosyltransferases and regulation in the gastrointestinal tract. Mol

Pharmacol 2001;59:405-414.

23. Meech R, Mackenzie PI. UDP-glucuronosyltransferase, the role of the

amino terminus in dimerization. J Biol Chem 1997;272:26913-26917.

24. Ghosh SS, Sappal BS, Kalpana GV, Lee SW, Chowdhury JR, Chowdhury

NR. Homodimerization of human bilirubin-uridine-diphosphoglucur-

onate glucuronosyltransferase-1 (UGT1A1) and its functional implica-

tions. J Biol Chem 2001;276:42108-42115.

25. Tukey RH, Strassburg CP. Human UDP-glucuronosyltransferases: me-

tabolism, expression, and disease. Annu Rev Pharmacol Toxicol 2000;40:

581-616.

26. Beutler E, Gelbart T, DeminaA. Racialvariabilityin the UDP-glucurono-

syltransferase 1 (UGT1A1) promoter: a balanced polymorphism for regu-

lation of bilirubin metabolism? Proc Natl Acad Sci U S A 1998;95:8170-

8174.

27. Guillemette C. Pharmacogenomics of human UDP-glucuronosyltrans-

ferase enzymes. Pharmacogenomics J 2003;3:136-158.

28. Kadakol A, Ghosh SS,Sappal BS,SharmaG, Chowdhury JR, Chowdhury

NR. Genetic lesions of bilirubin uridine-diphosphoglucuronate glucu-

ronosyltransferase (UGT1A1) causing Crigler-Najjar and Gilbert syn-dromes: correlation of genotype to phenotype. Hum Mutat 2000;16:297-

306.

29. Marsh S, McLeod HL. Pharmacogenetics of irinotecan toxicity. Pharma-

cogenomics 2004;5:835-843.

30. Sabatino L, Casamassimi A, Peluso G, Barone MV, Capaccio D, Migliore

C, et al. A novel peroxisomeproliferator-activated receptor gammaisoform

with dominant negative activity generated by alternative splicing. J Biol

Chem 2005;280:26517-26525.

31. McElvaine AT, Mayo KE. A dominant-negative human growth hormone-

releasing hormone (GHRH) receptor splice variant inhibits GHRH bind-

ing. Endocrinology 2006;147:1884-1894.

32. Meech R, Mackenzie PI. Determinants of UDP glucuronosyltransferase

membrane association and residency in the endoplasmic reticulum. Arch

Biochem Biophys 1998;356:77-85.

33. OuzzineM, Magdalou J, Burchell B, Fournel-GigleuxS. An internal signal

sequence mediates the targeting and retention of the human UDP-glucu-

ronosyltransferase 1A6 to the endoplasmic reticulum. J Biol Chem 1999;

274:31401-31409.

34. OuzzineM, Barre L, NetterP, Magdalou J, Fournel-Gigleux S. Role of the

carboxyl terminal stoptransfer sequence of UGT1A6membrane protein in

ER targeting and translocation of upstream lumenal domain. FEBS Lett

2006;580:1953-1958.

35. Barre L, Magdalou J, Netter P, Fournel-Gigleux S, Ouzzine M. The stop

transfer sequence of the human UDP-glucuronosyltransferase 1A deter-

mines localization to the endoplasmic reticulum by both static retention

and retrieval mechanisms. FEBS J 2005;272:1063-1071.

36. Iwano H, Yokota H, Ohgiya S, Yuasa A. The significance of amino acid

residue Asp446 for enzymatic stability of rat UDP-glucuronosyltransferase

UGT1A6. Arch Biochem Biophys 1999;363:116-120.37. Ikushiro S, Emi Y, Iyanagi T. Protein-protein interactions between UDP-

glucuronosyltransferase isozymes in rat hepatic microsomes. Biochemistry

1997;36:7154-7161.

38. Koiwai O, Aono S, Adachi Y, Kamisako T, Yasui Y, Nishizawa M, et al.

Crigler-Najjar syndrome type II is inherited both as a dominant and as a

recessive trait. Hum Mol Genet 1996;5:645-647.


UGT1A1 Eliminated Cmpds

Documents

Transcript of UGT1A1 Eliminated Cmpds