Biochem. J. (2013) 450, 107–114 (Printed in Great Britain)
107
doi:10.1042/BJ20121594
Protein–protein interactions between the bilirubin-conjugating UDPglucuronosyltransferase UGT1A1 and its shorter isoform 2 regulatory
partner derived from alternative splicing
Mélanie ROULEAU, Pierre COLLIN, Judith BELLEMARE, Mario HARVEY and Chantal GUILLEMETTE1
Pharmacogenomics Laboratory, Centre Hospitalier Universitaire de Québec (CHUQ) Research Center, 2705 Boul. Laurier, Québec, Canada, G1V 4G2, and Faculty of Pharmacy, Laval
University, Pavillon Ferdinand-Vandry, Université Laval, Québec, Canada, G1V 0A6
The oligomerization of UGTs [UDP (uridine diphosphate)glucuronosyltransferases] modulates their enzyme activities.
Recent findings also indicate that glucuronidation is negatively
regulated by the formation of inactive oligomeric complexes
between UGT1A enzymes [i1 (isoform 1)] and an enzymatically
inactive alternatively spliced i2 (isoform 2). In the present
paper, we assessed whether deletion of the UGT-interacting
domains previously reported to be critical for enzyme function
might be involved in i1–i2 interactions. The bilirubin-conjugating
UGT1A1 was used as a prototype. We also explored
whether intermolecular disulfide bonds are involved in i1–
i2 interactions and the potential role of selected cysteine
residues. Co-immunoprecipitation assays showed that UGT1A1
lacking the SP (signal peptide) alone or also lacking the
transmembrane domain (absent from i2) did not self-interact,
but still interacted with i2. The deletion of other N- or Cterminal domains did not compromise i1–i2 complex formation.
Under non-reducing conditions, we also observed formation of
HMWCs (high-molecular-mass complexes) for cells overexpressing i1 and i2. The presence of UGTs in these complexes
was confirmed by MS. Mutation of individual cysteine residues
throughout UGT1A1 did not compromise i1–i1 or i1–i2 complex
formation. These findings are compatible with the hypothesis that
the interaction between i1 and i2 proteins (either transient or
stable) involves binding of more than one domain that probably
differs from those involved in i1–i1 interactions.
INTRODUCTION
an additional/shorter terminal exon (exon 5b) and leading to nine
shorter UGT1A proteins [i2 (isoform 2)] [16,17].
UGT1A i2 alternative protein isoforms lack glucuronic acid
transferase activity, but are residents of the ER despite lacking
the TMD (transmembrane domain) [17]. Functional assays with
human cells co-expressing i1 enzymes and i2 variant isoforms
have clearly demonstrated a significant decrease in cellular
glucuronidation activity for many different UGT1A substrates,
supporting a regulatory role for i2 proteins [16–21]. Furthermore,
co-immunoprecipitation experiments support protein–protein
interactions between i1 and i2 proteins in heterologous expression
systems [17–19]. Also, the cellular co-localization of UGT1A
isoforms supports molecular interaction between them. The
mechanism underlying the regulation of transferase activity by
splice forms has become even more complex with the recent
demonstration of heterodimer formation between all UGT1A
i1 and other i2 [19]. Previously reported data thus suggest that
glucuronidation is negatively regulated by interactions between
active i1 enzymes and alternatively spliced i2 regulatory proteins
through formation of enzymatically inactive complexes.
The function of UGTs as oligomeric complexes and
their regulation through protein–protein interactions has been
documented (for reviews see [11,22]). Oligomerization of two
inactive mutants was even shown to yield an active enzyme
complex [23]. Previous studies point to several potential protein
domains that may regulate oligomerization/activity or subcellular
localization (Table 1). For instance, domains of the N-terminal
The functions of most proteins depend on continuous interactions
with other proteins or molecules [1]. The study of protein
interactomes is a growing field, and numerous studies have
attempted to characterize interactomes for various biological
processes and/or for different cell types or tissues. The finding
that most human genes are subject to alternative pre-mRNA
splicing has demonstrated a key process for protein diversity and
for regulating various biological processes [2–4]. Both processes,
splicing and protein–protein interactions, reportedly affect smallmolecule biotransforming pathways such as those governed
by the superfamily of UGTs [UDP (uridine diphosphate)glucuronosyltransferases] [5–12].
Human UGTs are an important class of conjugating enzymes
that play a key role in several important metabolic processes and
exhibit broad substrate specificity. Indeed, those enzymes catalyse
the glucuronidation of a wide range of structurally unrelated
endogenous and xenobiotic chemicals, including numerous
therapeutic drugs [13,14]. UGTs are type I membrane proteins of
the ER (endoplasmic reticulum) classified in two major families,
UGT1A and UGT2 (2A and 2B) [13]. The ∼ 190-kb UGT1A locus
produces half of the enzymatically active human UGT enzymes.
Alternative usage of variable first exons encoding the substratebinding domain leads to nine active UGT1A enzymes, known as
i1 (isoform 1) [15]. In previous studies, alternative splicing at the
3 -end of this locus was revealed, as a result of the presence of
Key words: alternative splice product, dominant-negative
regulation, membrane protein, protein–protein interaction, uridine
diphosphate-glucuronosyltransferase (UGT).
Abbreviations used: ER, endoplasmic reticulum; HEK, human embryonic kidney; HMWC, high-molecular-mass complex; i1, isoform 1; i2, isoform 2;
i2SP, deletion of signal peptide within i2; LC, liquid chromatography; MRA, micro-anchoring region; MRA, deletion of MRA; MS/MS, tandem MS; SP,
signal peptide; SP, absence of the signal peptide; SP-TMD, deletion of the signal peptide and the transmembrane domain; TMD, transmembrane
domain; UDP, uridine diphosphate; UGT, UDP-glucuronosyltransferase.
1
To whom correspondence should be addressed (email [email protected]).
c The Authors Journal compilation c 2013 Biochemical Society
108
Table 1
M. Rouleau and others
UGT protein domains and their implication in oligomerization
Literature
UGT protein (a.a.)
The present paper
Oligomerization
Localization
Activity
Reference(s)
Native protein
SP (1–26)
MRA (152–180)
TMD (490–510)
+
NA
−
+
ER
−
NA
+
+
−
−
−
[11,12,22]
[25–27]
[24,27,28]
[23,27,31,45,47]
N-terminus (24–276)
−
+
−
[23]
C-terminus (291–534)
+
+
−
[23,27,28,48]
UGT1A1 (a.a.)
Native protein
SP (1–20)
MRA (152–180)
TMD (490–510)
SP and TMD (1–20 and 490–510)
N-term (1–150)
180–250
250–375
375 – 490 (or 375–434 for i2)
L519X
K530X
Western blot expression Interaction
+
+
++
+
+
+
+
i1-i1
i1-i2
+
−
+
+
−
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
i1-i2
i2-i2
−
+
+
−
+
+
+
+ , presence of; − , absence of; , deletion; a.a., amino acids; NA, not available.
portion of UGT, such as the SP and a hydrophobic region
[called the MRA (micro-anchoring region); residues 152–180 in
UGT1A1], are critical for enzymatic activity and localization
to the ER membrane [24–27]. The MRA is also likely to be
involved in UGT oligomerization [28]. Recent protein homology
modelling identified a putative dimerization domain for UGT2B7
located between residues 180 and 200 [29]. The C-terminal
portion has also been studied. Experiments showed that the Cterminal tail (residues 510–534) and TMD (residues 490–510)
are not essential for hetero-oligomerization of UGT [23,28],
although deletion of one or the other significantly affects enzyme
activity [23,27,30,31]. Although protein–protein interactions are
typically non-covalent and thus reversible [32,33], some studies
also suggest that UGT oligomerization could occur via covalent
interactions. For instance, the presence of a HMWC (highmolecular-weight complex) under non-reducing conditions in
SDS/PAGE suggests the possibility of disulfide bonds between
UGTs [23,28,34].
On the basis of these observations, we assessed whether
deletion of UGT domains reported previously to be critical
for enzyme function, oligomerization and cellular localization
might mediate protein–protein interactions between i1 and i2. We
then explored whether relatively stronger intermolecular disulfide
bonds are involved in these interactions and the potential role of
specific cysteine residues using site-directed mutagenesis. The
bilirubin-conjugating UGT1A1 was used as a prototype. We
provide direct experimental evidence that the domain(s) critical
for interaction between i1–i2 probably differs from that involved
in i1–i1 interactions. The formation of large molecular complexes
was further demonstrated, but the identity of the cysteine residues
responsible for those interactions remains unknown.
EXPERIMENTAL
Preparation of microsomes and SDS/PAGE
Human cells were disrupted by sonication (3× 10 s on ice with a
10 s pause in between each sonication; 550 Sonic Dismembrator,
Fisher Scientific) and centrifuged twice at 12 200 g for 20 min to
remove the nuclear components cell debris. The ER membrane
fraction was then collected by sedimentation of the supernatant
at 105 000 g for 1 h. The protein concentration was determined
using the Bradford method. The microsomal protein extracts were
diluted in 4× non-reducing buffer [6.85 % (v/v) SDS, 233.3 mM
c The Authors Journal compilation c 2013 Biochemical Society
Tris/HCl (pH 6.8), 24 % (v/v) glycerol, 0.008 % Bromophenol
Blue and resolved by SDS/PAGE (10 % gel)]. Under reducing
conditions, samples were heated in the 4× buffer, which was
completed with 0.85 % final concentration of 2-mercaptoethanol
as a reducing agent. Protein levels in HEK (human embryonic
kidney)-293 cells overexpressing either i1–Myc–His alone or
i1–Myc–His coexpressed with i2–V5–His were determined by
Western blotting using an antibody against either myc (Cedarlane,
1:1500 dilution) or v5 (Life Technologies, 1:5000 dilution).
LC (liquid chromatography)-MS/MS (tandem MS) analysis
SDS/PAGE lanes corresponding to non-reducing conditions of
UGT1A1-overexpressing HEK-293 cells were cut into six gel
slices per lane. In-gel protein digestion was performed on a
MassPrepTM liquid handling station (Waters) according to the
manufacturer’s instructions and using sequencing-grade modified
trypsin (Promega). Cell extracts containing peptides were
dried using a SpeedVac® (Thermo Scientific) and then separated
by online reversed-phase nanoscale capillary LC and analysed by
electrospray MS/MS as described in [35]. Briefly, an LTQ (linear
trap quadrupole) MS (Thermo Electron) equipped with a nanoelectrospray ion source (Thermo Electron) was used. Peptides
were separated with a PicoFrit® column BioBasic C18, 10 cm
× 0.075 mm internal diameter (New Objective), with a linear
gradient from 2 to 50 % solvent B (acetonitrile and 0.1 % formic
acid) over 30 min at 200 nl/min. Each full-scan mass spectrum
(400–2000 m/z) was followed by collision-induced dissociation
of the seven most intense ions. All MS/MS samples were analysed
using Mascot (Matrix Science, version 2.3) with the human
UniRef 100 protein database (Homo sapiens: Taxon 9606, 100683
entries) assuming trypsin as the digesting enzyme. Two missed
cleavages were allowed. Scaffold (version 03.00.02; Proteome
Software) was used to validate MS/MS-based peptide and protein
identifications, which were accepted if they could be established at
>95.0 % probability as specified by the Peptide Prophet algorithm
and contained at least two identified peptides [36].
Site-directed mutagenesis
Site-directed mutagenesis was carried out to replace specific
cysteine residues of i1 cDNA with tyrosine (Table 2). The
UGT1A1 cDNA was subjected to oligonucleotide-mediated
mutagenesis with turbo Pfu DNA polymerase (DNA polymerase
Protein–protein interactions between UGT1A1 spliced isoforms
Table 2
Point mutations of selected cysteine residues for analysis of disulfide bonds involved in UGT oligomerization
Literature
Cysteine
18
127
186
223
280
383
509
510
517
109
The present paper
Oligomerization
NA
+
−
−
NA
NA
NA
NA
NA
Enzyme activity
+
−
−
−
−
−
+
+
+
Reference(s)
[46]
[28]
[34,46]
[28,46]
[46]
[46]
[46]
[46]
[46]
Western blot expression
++
++
+++
Interaction
i1–i1
i1–i2
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+ , presence of; − , absence of; NA, not available
from Pyrococcus furiosus; Agilent, Mississauga). Oligonucleotides were obtained from Integrated DNA Technologies
(Coralville). Human UGT1A1 i1 was used as a template DNA
for PCR, which was cloned into pcDNA3.1A-Myc-His (Life
Technologies). The PCR was performed on pcDNA3.1A-MycHis for i1 at 94 ◦ C for 30 s, followed by 21 cycles of 94 ◦ C
for 30 s, 55 ◦ C for 60 s and 68 ◦ C for 16 min. Site-directed
mutagenesis was followed by digestion with DpnI to eliminate
parental DNA (Roche Applied Science). Digestion products were
then transformed into Escherichia coli XL1 Blue cells. Growing
colonies were assessed for the presence of the correct point
mutation, which was confirmed by direct sequencing.
Generation of UGT1A1-mutated proteins
Mutated cDNAs for i1–Myc–His and i2–V5–His were generated
with overlap extension PCR as described in [37]. This method
relies on two oligonucleotides that overlap the deletion to be made.
Human UGT1A1 i1 and UGT1A1 i2 were used as a template DNA
for the PCR. UGT1A1 i1 was cloned into pcDNA3.1A-MycHis (Life Technologies), whereas UGT1A1 i2 was cloned into
pcDNA6A-V5-His (Life Technologies). PCRs were performed
on UGT1A1 pcDNA3.1A-Myc-His for i1 or on pcDNA6A-V5His for i2 cDNA as follows: 95 ◦ C for 5 min, 30 cycles of 95 ◦ C
for 30 s, 55 ◦ C for 30 s, 72 ◦ C for 2 min and a final extension at
72 ◦ C for 7 min. All constructs were verified by direct sequencing.
Co-immunoprecipitation assays using proteins from transiently
transfected cells
HEK-293 cells were transfected for transient expression using
the Neon® electroporation kit (Life Technologies). Briefly, 7.5×
106 cells were mixed with DNA (10 μg of i1 plasmid, 10 μg
of i2 or 5 μg each of i1 and i2) and then electroporated at
1300 V for 10 ms (three pulses). After 48 h, the cells were washed
with PBS, harvested and lysed for 45 min on ice with 1 ml
of lysis buffer [0.05 M Tris/HCl (pH 7.4), 0.15 M NaCl, 0.3 %
deoxycholic acid, 1 % (w/v) Igepal CA-630 (Sigma) and 1 mM
EDTA]. Cell lysates were then homogenized by pipetting up
and down through fine needles (18G followed by 20G) for ten
to twenty times on ice. Lysates were centrifuged for 15 min at
13 000 g, and the supernatants were collected. Cell-free lysates
were mixed with Protein G–Sepharose® 4 Fast Flow (50 % slurry;
GE Healthcare) and stirred for 30 min at 4 ◦ C to block nonspecific binding. After centrifugation (13 000 g for 1 min), 1 mg
of supernatant protein was added to 1 μg of specific polyclonal
anti-Myc (Sigma) in 1 ml of lysis buffer and incubated at 4 ◦ C
with 50 μl of beads (50 % slurry) for 15 h. The beads were washed
three times with 1 ml of lysis buffer and finally with 1 ml of 50 mM
Tris/HCl (pH 7.5). The beads containing the immunoprecipitated
proteins were resuspended with 30 μl of SDS/PAGE loading
buffer, heated at 100 ◦ C for 5 min and then centrifuged at 12 000 g
for 20 s. The supernatant was subjected to SDS/PAGE followed by
Western blotting. Blotted membranes were probed with a specific
monoclonal anti-V5 antibody linked with horseradish peroxidase
(Life Technologies), as specified in Figures 2 and 4B.
Subcellular localization of UGT1A1-mutated protein by differential
centrifugation
HEK-293 cells were plated in 100-mm dishes (5× 106 cells/dish)
and co-transfected for transient expression with 10 μg pUGT1A
i1 using Lipofectine (Life Technologies). After 48 h, the cells
were washed with PBS, harvested, centrifuged for 5 min at 888 g,
resuspended in 500 μl of microsome buffer [2.62 mM KH2 PO4 ,
1.38 mM K2 HPO4 , 20 % glycerol (pH 7.0) and 0.5 mM DTT
(dithiothreitol) freshly added] and frozen. The next day, the
cytosolic and microsomal fractions were collected as follows:
the cells were sonicated (5× 10 s on ice with a 10-s pause
in between each sonication using 550 Sonic Dismembrator),
centrifuged once at 12 200 g for 15 min and then the resulting
supernatant was further centrifuged at 107 400 g for 1 h. The
supernatant corresponded to the cytosolic fraction and the pellet to
the microsomal fraction. Each fraction (80 μg of protein each) was
resolved by SDS/PAGE followed by Western blotting, with the
membrane blots probed using monoclonal anti-Myc tag antibodies
(Cedarlane) followed by a secondary anti-mouse IgG antibodies
linked with horseradish peroxidase (GE Healthcare).
RESULTS
Effect of deleted domains on the i1–i2 interaction
We first assessed whether the deletion of UGT1A-interacting
domains reported previously to be critical in certain cases for
enzyme function and those involved in subcellular localization
might be relevant to protein–protein interactions between i1
and i2 proteins (summarized in Table 1). UGT1A1 was used
as a prototype because homo-oligomerization has been well
documented [28,38,39]. We generated ten mutant UGT1A1
proteins (Table 1 and Figure 1). The engineered i1–Myc
mutants were each co-transfected with either i1–V5 or i2–V5.
c The Authors Journal compilation c 2013 Biochemical Society
110
Figure 1
M. Rouleau and others
UGT protein domains and conserved cysteine residues
Schematic representation of UGT1A1 i1 protein in the ER membrane, showing the domains and localization of the cysteine residues. Most of the polypeptide resides in the ER lumen, except for the
C-terminal tail and the transmembrane domain. The variable region confers substrate specificity and the conserved region is responsible for binding the co-substrate UDP-glucuronic acid. Each
cysteine residue is represented by (䉱), except for Cys18 (), which shows the N-terminal peptide being cleaved to produce the mature protein. Abbreviations of several domains that were deleted
are also indicated.
Expression of mutant proteins was confirmed by Western blotting
(results not shown) for all constructs. We initially performed
co-immunoprecipitation control experiments to visualize the
interaction between i1–i1, i1–i2 and i2–i2 (Figure 2A) [18,19].
Following transient expression, i1–Myc and its associated
proteins were specifically pulled out from whole-cell extracts
by immunoprecipitation using anti-Myc antibodies (Figure 2B).
The immunoprecipitates were then assessed by SDS/PAGE and
Western blotting, and an interaction between proteins was probed
with anti-V5 antibodies.
The i1–i1 interaction was not detected in the absence of the
SP (SP) alone or with additional deletion of the TMD (SPTMD), although the interaction of i1SP-TMD with i2 was
detected. Deletion of the i1 TMD alone (i1TMD) did not
preclude the i1–i1 or i1–i2 interaction. Deletion of additional
i1 domains, such as the MRA (MRA), the first 150 residues
(1–150), the end of exon 1 (180–250), the conserved region or
signature sequence (250–375 and 375–490) or the C-terminus
(L519X or K530X), did not prevent the interaction of either
i1–i1 or i1–i2. On the basis of these observations, we also tested
deletions within i2, i.e. the SP (i2SP) and its first 150 residues
(i21–150), each of which abolished the interaction with i1, but
not with intact i2 (Figure 2C). Part of the signature sequence and
conserved C-terminal region of i2 (i2375–444) was also deleted,
and this did not prevent the interaction with i1 or intact i2.
As a missing domain of UGT1A1 could lead to its aberrant
localization and thus erroneous interpretation of the coimmunoprecipitation results, subcellular fractionation was
c The Authors Journal compilation c 2013 Biochemical Society
performed to confirm that each mutant protein indeed localized
appropriately to the ER membrane to allow oligomerization with
UGTs (Figure 3). The i1SP mutant integrated into the ER
membrane, but to a lesser extent compared with native UGT1A1
i1; we estimated that one half of the total i1SP was in the
cytosolic fraction, thus still leaving a substantial proportion of
the protein available for interaction with native i1 in the ER.
This translocation from the ER membrane to the cytosol was
also observed for i1SP-TMD. Deletion of other i1 domains,
namely 1–150, TMD, L519X and K530X, did not affect the
subcellular localization.
Disulfide bond analysis by SDS/PAGE under non-reducing
conditions
A cellular heterologous expression system of UGT1A1 in HEK293 cells was generated. Using antibodies directed against
the tagged proteins of each UGT1A protein isoforms and in the
absence of a strong reducing agent such as 2-mercaptoethanol, we
observed monomeric UGT1As and also formation of the HMWC
(Figure 4A, left-hand panel). Those monoclonal antibodies, antiMyc for i1 and anti-V5 for i2, were used to demonstrate the
presence of i1 and i2 as monomers and as part of the HMWC.
The HMWCs were absent from the parental HEK-293 cell line
and were not observed in the presence of 2-mercaptoethanol, a
reducing agent (Figure 4A, right-hand panel). Furthermore, the
presence of UGT1A in these HMWCs, as determined by MS,
was confirmed (results not shown); it is noteworthy that it was
Protein–protein interactions between UGT1A1 spliced isoforms
Figure 2
111
Protein–protein interactions between UGT1A1 i1 and i2 mutants
HEK-293 cells were transfected for transient expression. An equal amount of cell lysates was used for immunoprecipitation of protein complexes with a polyclonal anti-Myc antibody and the presence
of the interacting partner was revealed by a monoclonal antibody directed against V5 tag protein. (A) Controls for the co-immunoprecipitation (IP) experiments with UGT1A1 full-length protein.
Lanes with odds numbers are immunoprecipitated with a polyclonal anti-Myc antibody, whereas the lanes with even numbers were left without immunoprecipitation antibody. Lanes 1–4, negative
control; expression of either UGT1A1 i1–Myc or UGT1A1 i2–V5 alone. Lanes 5–10, positive control; lanes 5 and 6, control for UGT1A1 i1–Myc + UGT1A1 i2–V5 interaction; lanes 7 and 8, control
for UGT1A1 i1–Myc + UGT1A1 i1–V5 interaction; lanes 9 and 10, control for UGT1A1 i2–Myc + UGT1A1 i2–V5 interaction. (B) Co-immunoprecipitation experiments with various UGT1A1
i1-myc mutant proteins with subsequent visualization of interacting partner UGT1A1 i1–V5 or UGT1A1 i2–V5 in the immunoprecipitation. A band corresponding to i1 is observed at the expected
molecular mass (indicated on the right-hand side by the arrows) except for SP and SPTMD; in the latter case an unsure band is detected (B, lane 16). A band corresponding to i2 is observed
in each corresponding wells. (C) Co-immunoprecipitation experiments of either UGT1A1 i2–Myc or UGT1A1 i1–Myc and with subsequent visualization of UGT1A1 i2–V5 mutated proteins. A band
is detected at the expected molecular mass except for the interaction of i1–i2SP and also i1–i21–150. For (B) and (C), the control lane (ctl) correspond to UGT1A1 i2–V5 full-length protein.
Each transfection and co-immunoprecipitation experiment was done at least twice as independent experiments. The vertical grey lanes indicates that there was a rearrangement or a grouping of the
Western blot images. = deletion.
not possible to discriminate between the highly similar i1 or i2
proteins by MS. In light of this observation and given that disulfide
bonds are formed via oxidation of the thiol moiety of the cysteine
residue, several cysteine residues present in UGT1A1 i1 were further mutated to tyrosine (Figure 1). Co-immunoprecipitation experiments revealed that these selected point mutations in i1 (Cys18 ,
Cys127 , Cys186 , Cys223 , Cys280 , Cys383 , Cys509 , Cys510 and Cys517 ) did
not prevent interaction with i1 or with i2 (Figure 4B and Table 2).
DISCUSSION
Identification of molecular interactions is an essential step towards
a better understanding of cellular processes. Homo- and heterooligomerization of UGTs is well documented from independent
studies [39–43], and growing evidence suggests that these
protein–protein interactions are one of the mechanisms by which
i2 regulates cellular glucuronidation activity. We demonstrated
previously that the splice variant UGT1A1 i2, which lacks the
transmembrane helix encoded by exon 5a, interacts strongly with
the full-length native UGT1A1 i1 [17–19]. However, the identity
of the domain(s) mediating interactions between UGTs and their
regulatory i2 partners is still unresolved.
UGT1A1 i1 and i2 proteins have highly similar C-terminal
portions, although i2 lacks the TMD encoded by exon 5a,
suggesting that i2 does not require this membrane anchor for
its interaction with i1. Similarly, deletion of the i1 TMD does
not prevent i1–i2 interaction. However, we observed that deletion
of N-terminal domains of UGT1A1 i1, such as the SP alone or
with the TMD, precluded the i1–i1 interaction, but not that of
i1–i2. The presence of the SP and the first 150 residues of i2
appear to be critical for the i1–i2 interaction, but not that of i2–i2.
Conversely, the i1–i1 complex still formed even on deletion of
the first 150 residues. These observations indicate that different
domains are involved in native i1–i1 interactions compared with
i1–i2 interactions. The SP of i1 enzymes is cleaved on insertion
c The Authors Journal compilation c 2013 Biochemical Society
112
Figure 3
M. Rouleau and others
Subcellular localization of mutated Myc-tagged UGT1A1 mutants
HEK-293 cells transiently expressing UGT1A1 i1–Myc-mutant proteins were used to prepare extracts and submitted to differential centrifugation. Subcellular fractions (80 μg of protein) were
subjected to SDS/PAGE with subsequent Western blotting with an anti-Myc antibody. A signal is observed in the microsomal fraction for i1 full-length protein and 1–150, TMD, L519X and
K530X. For SP and SPTMD, a signal is also observed in the cytosolic fraction. = deletion.
Figure 4
Formation of HMWC under non-reducing conditions
(A) Microsomal proteins from UGT1A1-overexpressing HEK-293 cells were resolved with SDS/PAGE under non-reducing or reducing conditions (2-mercaptoethanol). Detection of tagged UGT1A1
i1–Myc (upper panels) and UGT1A1 i2–V5 (lower panels) proteins under reducing (right-hand panel) or non-reducing (left-hand panel) conditions with either anti-Myc antibody for i1 (Cedarlane,
1:1500 dilution) or anti-V5 for i2 (Life technologies, 1:5000 dilution). (B) Cysteine mutation and co-immunoprecipitation analysis. HEK-293 cells were co-transfected with UGT1A1 i1–Myc cysteine
mutants and either UGT1A1 i1–V5 or UGT1A1 i2–V5. Cell lysates were immunoprecipitated with rabbit polyclonal anti-Myc antibody, resolved with SDS/PAGE and Western blotting and visualized
with a mouse monoclonal anti-V5 antibody. Lanes 1 and 2 are positive co-immunoprecipitation control for either the interaction of i1 with i1, or the interaction of i1 with i2. The vertical grey lane
indicates that there was a rearrangement or a grouping of the Western blot images.
c The Authors Journal compilation c 2013 Biochemical Society
Protein–protein interactions between UGT1A1 spliced isoforms
into the ER membrane, and thus the fact that its deletion affects
the i1–i2 interaction is intriguing because it suggests a structural
role for this domain in the i1–i2 interaction. In fact, our data
support that only a fraction of the observed effect is attributable to
mislocalization of mutated proteins in the cell. Thus it is possible
that deletion of the i1 SP may impair its self-association owing
to misfolding of the mutated i1 in the ER membrane, but the
topology of mutated proteins was not assessed to confirm whether
they were either located on the luminal or cytosolic side of the ER
membrane. Differences between the domains involved in homoor hetero-oligomerization could be plausible because the topology
of i2, which lacks a TMD in the ER membrane, may differ from
that of i1. We suggested previously that localization/retention of
i2 in the ER may be conferred, at least in part, by the dilysine motif
KKXX encoded by the new exon 5b [27,30]. Finally, it was not
possible to infer protein–protein domains based on structural data
owing to the extremely limited availability of three-dimensional
structures for UGT [44].
Deletion of other conserved domains of UGT1A1 i1, such as the
MRA, reportedly substantially reduces its homo-oligomerization
[28]. Conversely, and according to our observations, the MRA
is probably not critical for i1–i2 interaction. The fact that i2
is enzymatically inactive suggests that the amino acid sequence
encoded by exon 5a (residues 435–530) is required, functionally
or structurally, for UGT1A1 transferase activity. However,
deletion of the C-terminal region (TMD, L519X and K530X) does
not affect the interacting potential between i1 and i2. Consistent
with these observations, the TMD and the cytoplasmic ‘tail’
are not essential for functional oligomerization in the case of
UGT1A9 [45]. However, the use of only one approach is a
limitation of the present study, as the co-immunoprecipitation
of mutated protein could lead to the misfolding of proteins. The
results thus need to be confirmed with another approach. Another
limitation is the fact that the co-immunoprecipitation results
provide a qualitative outcome since the level of expression of
mutated proteins, as well as their localization to the ER membrane,
could be different from the full-length protein. The present paper
remains the first to study the interaction between splice variant
proteins derived for the UGT1A locus in humans.
The present study also evaluated the involvement of
disulfide bonds between UGT1A isoforms on homo- and
hetero-oligomerization, i.e. via SDS/PAGE under non-reducing
conditions. Under these conditions, we could assess the presence
of monomers and the formation of UGT oligomers. Our data are
consistent with formation of oligomers with complexes migrating
at 130–250 kDa. The presence of UGT in those complexes was
further confirmed following tryptic digestion of bands extracted
from gels and sequencing by MS. The data are thus consistent with
previous reports indicating that disulfide bonds are indeed formed
between UGTs [23,28,34]. UGT protein–protein interactions have
also been studied under milder conditions using native-PAGE,
which concluded that UGTs indeed oligomerize, probably in
part via disulfide bonds [39,43]. As cysteine residues are widely
conserved among UGT1A subfamily members [46], the study
of their potential role in UGT protein–protein interactions is
relevant. A cysteine-focused point mutation analysis did not
compromise the interactions between UGT1A isoforms (i1–i1
or i1–i2). However, this analysis cannot rule out the possibility
that a combination of cysteine residues may be responsible for
the formation of HMWCs under non-reducing conditions. A
study by Olson et al. [34] indicated that Cys183 is involved
in homodimerization of UGT1A9, although mutation of this
codon to tyrosine (Figure 4B) or glycine (results not shown) in
UGT1A1 i1 did not prevent its homo- or hetero-oligomerization
with i2.
113
In conclusion, we report that i1–i2 interactions (either stable or
transient) probably involve association between more than one domain that likely differs from those involved in i1–i1 interactions.
Further investigations are needed for a fuller understanding of the
interactions between i1 enzymes and their regulatory i2 partners.
AUTHOR CONTRIBUTION
Mélanie Rouleau and Pierre Collin performed the experiments and analysis. Chantal
Guillemette and Judith Bellemare were involved in the design of the study. Mélanie
Rouleau and Chantal Guillemette wrote the paper. Mario Harvey was involved in data
interpretation. Chantal Guillemette managed and financed the project.
FUNDING
The present work was supported by The National Sciences and Engineering Research
Council of Canada (NSERC) and the Canada Research Chair Program [grant number
CG086976]. M.R. is recipient of a Frederick Banting and Charles Best Canada Graduate
Scholarship award (CIHR) and of a doctoral scholarship award from Fonds de recherche
en santé du Québec (FRSQ). P.C. is recipient of a studentship award from FRSQ, the Fonds
de l’Enseignement et de la Recherche of the Faculty of Pharmacy of Laval University,
and received partial support from the Oncology, Molecular Endocrinology and Human
Genomics Research Center. C.G. holds the Canada Research Chair in Pharmacogenomics
(Tier II).
REFERENCES
1 Eisenberg, D., Marcotte, E. M., Xenarios, I. and Yeates, T. O. (2000) Protein function in the
post-genomic era. Nature 405, 823–826
2 Schwerk, C. and Schulze-Osthoff, K. (2005) Regulation of apoptosis by alternative
pre-mRNA splicing. Mol. Cell 19, 1–13
3 Wei, J., Zaika, E. and Zaika, A. (2012) p53 Family: role of protein isoforms in human
cancer. J. Nucleic Acids 2012, 687359
4 Graveley, B. R. (2001) Alternative splicing: increasing diversity in the proteomic world.
Trends Genet. 17, 100–107
5 Fremont, J. J., Wang, R. W. and King, C. D. (2005) Coimmunoprecipitation of
UDP-glucuronosyltransferase isoforms and cytochrome P450 3A4. Mol. Pharmacol. 67,
260–262
6 Ishii, Y., Iwanaga, M., Nishimura, Y., Takeda, S., Ikushiro, S., Nagata, K., Yamazoe, Y.,
Mackenzie, P. I. and Yamada, H. (2007) Protein–protein interactions between rat hepatic
cytochromes P450 (P450s) and UDP-glucuronosyltransferases (UGTs): evidence for the
functionally active UGT in P450–UGT complex. Drug Metab. Pharmacokinet. 22,
367–376
7 Takeda, S., Ishii, Y., Iwanaga, M., Nurrochmad, A., Ito, Y., Mackenzie, P. I., Nagata, K.,
Yamazoe, Y., Oguri, K. and Yamada, H. (2009) Interaction of cytochrome P450 3A4 and
UDP-glucuronosyltransferase 2B7: evidence for protein–protein association and possible
involvement of CYP3A4 J-helix in the interaction. Mol. Pharmacol. 75, 956–964
8 Akizawa, E., Koiwai, K., Hayano, T., Maezawa, S., Matsushita, T. and Koiwai, O. (2008)
Direct binding of ligandin to uridine 5 -diphosphate glucuronosyltransferase 1A1.
Hepatol. Res. 38, 402–409
9 Taura, K. I., Yamada, H., Hagino, Y., Ishii, Y., Mori, M. A. and Oguri, K. (2000) Interaction
between cytochrome P450 and other drug-metabolizing enzymes: evidence for an
association of CYP1A1 with microsomal epoxide hydrolase and UDPglucuronosyltransferase. Biochem. Biophys. Res. Commun. 273, 1048–1052
10 Udomsinprasert, R., Bogoyevitch, M. A. and Ketterman, A. J. (2004) Reciprocal regulation
of glutathione S-transferase spliceforms and the Drosophila c-Jun N-terminal kinase
pathway components. Biochem. J. 383, 483–490
11 Ishii, Y., Takeda, S. and Yamada, H. (2010) Modulation of UDP-glucuronosyltransferase
activity by protein–protein association. Drug Metab. Rev. 42, 145–158
12 Ishii, Y., Takeda, S., Yamada, H. and Oguri, K. (2005) Functional protein–protein
interaction of drug metabolizing enzymes. Front. Biosci. 10, 887–895
13 Meech, R. and Mackenzie, P. I. (1997) Structure and function of uridine diphosphate
glucuronosyltransferases. Clin. Exp. Pharmacol. Physiol. 24, 907–915
14 Fisher, M. B., Paine, M. F., Strelevitz, T. J. and Wrighton, S. A. (2001) The role of hepatic
and extrahepatic UDP-glucuronosyltransferases in human drug metabolism. Drug Metab.
Rev. 33, 273–297
15 Gong, Q. H., Cho, J. W., Huang, T., Potter, C., Gholami, N., Basu, N. K., Kubota, S.,
Carvalho, S., Pennington, M. W., Owens, I. S. and Popescu, N. C. (2001) Thirteen
UDPglucuronosyltransferase genes are encoded at the human UGT1 gene complex locus.
Pharmacogenetics 11, 357–368
c The Authors Journal compilation c 2013 Biochemical Society
114
M. Rouleau and others
16 Girard, H., Levesque, E., Bellemare, J., Journault, K., Caillier, B. and Guillemette, C.
(2007) Genetic diversity at the UGT1 locus is amplified by a novel 3 alternative splicing
mechanism leading to nine additional UGT1A proteins that act as regulators of
glucuronidation activity. Pharmacogenet. Genomics 17, 1077–1089
17 Levesque, E., Girard, H., Journault, K., Lepine, J. and Guillemette, C. (2007) Regulation of
the UGT1A1 bilirubin-conjugating pathway: role of a new splicing event at the UGT1A
locus. Hepatology 45, 128–138
18 Bellemare, J., Rouleau, M., Harvey, M. and Guillemette, C. (2010) Modulation
of the human glucuronosyltransferase UGT1A pathway by splice isoform
polypeptides is mediated through protein–protein interactions. J. Biol. Chem. 285,
3600–3607
19 Bellemare, J., Rouleau, M., Girard, H., Harvey, M. and Guillemette, C. (2010)
Alternatively spliced products of the UGT1A gene interact with the enzymatically active
proteins to inhibit glucuronosyltransferase activity in vitro . Drug Metab. Dispos. 38,
1785–1789
20 Bellemare, J., Rouleau, M., Harvey, M., Tetu, B. and Guillemette, C. (2010)
Alternative-splicing forms of the major phase II conjugating UGT1A gene negatively
regulate glucuronidation in human carcinoma cell lines. Pharmacogenomics J. 10,
431–441
21 Jones, N. R., Sun, D., Freeman, W. M. and Lazarus, P. (2012) Quantification of Hepatic
UDP glucuronosyltransferase 1A splice variant expression and correlation of UDP
glucuronosyltransferase 1A1 variant expression with glucuronidation activity. J.
Pharmacol. Exp. Ther. 342, 720–729
22 Finel, M. and Kurkela, M. (2008) The UDP-glucuronosyltransferases as oligomeric
enzymes. Curr. Drug Metab. 9, 70–76
23 Meech, R. and Mackenzie, P. I. (1997) UDP-glucuronosyltransferase, the role of the amino
terminus in dimerization. J. Biol. Chem. 272, 26913–26917
24 Ciotti, M., Cho, J. W., George, J. and Owens, I. S. (1998) Required buried α-helical
structure in the bilirubin UDP-glucuronosyltransferase, UGT1A1, contains a
nonreplaceable phenylalanine. Biochemistry 37, 11018–11025
25 Seppen, J., Steenken, E., Lindhout, D., Bosma, P. J. and Elferink, R. P. (1996) A mutation
which disrupts the hydrophobic core of the signal peptide of bilirubin
UDP-glucuronosyltransferase, an endoplasmic reticulum membrane protein, causes
Crigler–Najjar type II. FEBS Lett. 390, 294–298
26 Ohnishi, A. and Emi, Y. (2003) Rapid proteasomal degradation of translocation-deficient
UDP-glucuronosyltransferase 1A1 proteins in patients with Crigler–Najjar type II.
Biochem. Biophys. Res. Commun. 310, 735–741
27 Ouzzine, M., Magdalou, J., Burchell, B. and Fournel-Gigleux, S. (1999) An internal signal
sequence mediates the targeting and retention of the human UDP-glucuronosyltransferase
1A6 to the endoplasmic reticulum. J. Biol. Chem. 274, 31401–31409
28 Ghosh, S. S., Sappal, B. S., Kalpana, G. V., Lee, S. W., Chowdury, J. R. and Chowdury,
N. R. (2001) Homodimerization of human bilirubin-uridine-diphosphoglucuronate
glucuronosyltransferase-1 (UGT1A1) and its functional implications. J. Biol. Chem. 276,
42108–42115
29 Lewis, B. C., Mackenzie, P. I. and Miners, J. O. (2011) Homodimerization of
UDP-glucuronosyltransferase 2B7 (UGT2B7) and identification of a putative dimerization
domain by protein homology modeling. Biochem. Pharmacol. 82, 2016–2023
30 Ouzzine, M., Barre, L., Netter, P., Magdalou, J. and Fournel-Gigleux, S. (2006) Role of the
carboxyl terminal stop transfer sequence of UGT1A6 membrane protein in ER targeting
and translocation of upstream lumenal domain. FEBS Lett. 580, 1953–1958
31 Meech, R., Yogalingam, G. and Mackenzie, P. I. (1996) Mutational analysis of the
carboxy-terminal region of UDP-glucuronosyltransferase 2B1. DNA Cell Biol. 15,
489–494
Received 16 October 2012/12 November 2012; accepted 14 November 2012
Published as BJ Immediate Publication 14 November 2012, doi:10.1042/BJ20121594
c The Authors Journal compilation c 2013 Biochemical Society
32 Janin, J., Bahadur, R. P. and Chakrabarti, P. (2008) Protein–protein interaction and
quaternary structure. Q. Rev. Biophys. 41, 133–180
33 Brinda, K. V. and Vishveshwara, S. (2005) Oligomeric protein structure networks: insights
into protein–protein interactions. BMC Bioinformatics 6, 296–311
34 Olson, K. C., Dellinger, R. W., Zhong, Q., Sun, D., Amin, S., Spratt, T. E. and Lazarus, P.
(2009) Functional Characterization of low-prevalence missense polymorphisms in the
UDP-glucuronosyltransferase 1A9 gene. Drug Metab. Dispos. 37, 1999–2007
35 Lachapelle, S., Gagne, J. P., Garand, C., Desbiens, M., Coulombe, Y., Bohr, V. A., Hendzel,
M. J., Masson, J. Y., Poirier, G. G. and Lebel, M. (2011) Proteome-wide identification of
WRN-interacting proteins in untreated and nuclease-treated samples. J. Proteome Res.
10, 1216–1227
36 Nesvizhskii, A. I., Keller, A., Kolker, E. and Aebersold, R. (2003) A statistical model for
identifying proteins by tandem mass spectrometry. Anal. Chem. 75, 4646–4658
37 Lee, J., Shin, M. K., Ryu, D. K., Kim, S. and Ryu, W. S. (2010) Insertion and deletion
mutagenesis by overlap extension PCR. Methods Mol. Biol. 634, 137–146
38 Operana, T. N. and Tukey, R. H. (2007) Oligomerization of the UDP-glucuronosyltransferase 1A proteins: homo- and heterodimerization analysis by fluorescence
resonance energy transfer and co-immunoprecipitation. J. Biol. Chem. 282,
4821–4829
39 Fujiwara, R., Nakajima, M., Yamanaka, H., Katoh, M. and Yokoi, T. (2007) Interactions
between human UGT1A1, UGT1A4, and UGT1A6 affect their enzymatic activities. Drug
Metab. Dispos. 35, 1781–1787
40 Kurkela, M., Patana, A. S., Mackenzie, P. I., Court, M. H., Tate, C. G., Hirvonen, J.,
Goldman, A. and Finel, M. (2007) Interactions with other human UDP-glucuronosyltransferases attenuate the consequences of the Y485D mutation on the activity and
substrate affinity of UGT1A6. Pharmacogenet. Genomics 17, 115–126
41 Fujiwara, R., Nakajima, M., Yamanaka, H., Nakamura, A., Katoh, M., Ikushiro, S., Sakaki,
T. and Yokoi, T. (2007) Effects of coexpression of UGT1A9 on enzymatic activities of
human UGT1A isoforms. Drug Metab. Dispos. 35, 747–757
42 Nakajima, M., Yamanaka, H., Fujiwara, R., Katoh, M. and Yokoi, T. (2007) Stereoselective
glucuronidation of 5-(4 -hydroxyphenyl)-5-phenylhydantoin by human
UDP-glucuronosyltransferase (UGT) 1A1, UGT1A9, and UGT2B15: effects of UGT–UGT
interactions. Drug Metab. Dispos. 35, 1679–1686
43 Fujiwara, R., Nakajima, M., Oda, S., Yamanaka, H., Ikushiro, S.-I., Sakaki, T. and Yokoi, T.
(2010) Interactions between human UDP-glucuronosyltransferase (UGT) 2B7 and UGT1A
enzymes. J. Pharm. Sci. 99, 442–454
44 Miley, M. J., Zielinska, A. K., Keenan, J. E., Bratton, S. M., Radominska-Pandya, A. and
Redinbo, M. R. (2007) Crystal structure of the cofactor-binding domain of the human
phase II drug-metabolism enzyme UDP-glucuronosyltransferase 2B7. J. Mol. Biol. 369,
498–511
45 Kurkela, M., Hirvonen, J., Kostiainen, R. and Finel, M. (2004) The interactions between
the N-terminal and C-terminal domains of the human UDP-glucuronosyltransferases are
partly isoform-specific and may involve both monomers. Biochem. Pharmacol. 68,
2443–2450
46 Ghosh, S. S., Lu, Y., Lee, S. W., Wang, X., Guha, C., Roy-Chowdhury, J. and
Roy-Chowdhury, N. (2005) Role of cysteine residues in the function of human UDP
glucuronosyltransferase isoform 1A1 (UGT1A1). Biochem. J. 392, 685–692
47 Kurkela, M., Morsky, S., Hirvonen, J., Kostiainen, R. and Finel, M. (2004) An active and
water-soluble truncation mutant of the human UDP-glucuronosyltransferase 1A9. Mol.
Pharmacol. 65, 826–831
48 Koiwai, O., Aono, S., Adachi, Y., Kamisako, T., Yasui, Y., Nishizawa, M. and Sato, H.
(1996) Crigler–Najjar syndrome type II is inherited both as a dominant and as a recessive
trait. Hum. Mol. Genet. 5, 645–647