The Journal of Immunology
Key Residues of a Major Cytochrome P4502D6 Epitope Are
Located on the Surface of the Molecule1
Yun Ma,* Mark G. Thomas,† Manabu Okamoto,* Dimitrios P. Bogdanos,* Sylvia Nagl,‡
Nanda Kerkar,§ Agnel R. Lopes,* Luigi Muratori,¶ Marco Lenzi,¶ Francesco B. Bianchi,¶
Giorgina Mieli-Vergani,§ and Diego Vergani2*
Eukaryotically expressed CYP2D6 is the universal target of liver kidney microsomal Ab type 1 (LKM1) in both type 2 autoimmune
hepatitis (AIH) and chronic hepatitis C virus (HCV) infection. In contrast, reactivity to prokaryotically expressed CYP2D6 protein
and synthetic peptides is significantly lower in HCV infection than in AIH. The aim of the present study was to characterize LKM1
reactivity against a panel of eukaryotically expressed CYP2D6 constructs in the two conditions. LKM1-positive sera obtained from
16 patients with AIH and 16 with HCV infection were used as probes to perform a complete epitope mapping of CYP2D6.
Reactivity to the full-length protein and 16 constructs thereof was determined by radioligand assay. We found that antigenicity
is confined to the portion of the molecule C-terminal of aa 193, no reactivity being detectable against the aa sequence 1–193.
Reactivity increases stepwise toward the C-terminal in both AIH and HCV, but the frequency of reactivity in the two conditions
differs significantly between aa 267–337. To further characterize this region, we introduced a five and a three amino acid swap
mutation selected from the homologous regions of CYP2C9 and HCV. This maneuver resulted in a substantial loss of LKM1
binding in both conditions, suggesting that this region contains a major epitope. Molecular modeling revealed that CYP2D6316 –327
is exposed on the surface of the protein, and may represent a key target for the autoantibody. These findings provide an initial
characterization of the antigenic constitution of the target of LKM1 in AIH and HCV infection. The Journal of Immunology,
2002, 169: 277–285.
L
iver kidney microsomal Ab type 1 (LKM1)3 was originally described as the serological hallmark of autoimmune hepatitis (AIH) type 2 (1–5), and was later reported
to be present in up to 10% of patients with chronic hepatitis C virus
(HCV) infection (6, 7). In the two conditions, LKM1 gives an
identical immunofluorescence pattern, recognizes the same antigenic target, namely the liver enzyme cytochrome P4502D6
(CYP2D6) (8, 9) and similarly blocks its in vitro catalytic activity
(10 –12). There are also differences: in AIH, LKM1 reacts equally
well with prokaryotically and eukaryotically expressed CYP2D6
(13), but only with eukaryotic CYP2D6 in HCV infection (14 –17);
in AIH, LKM1 recognizes the linear CYP2D6196 –218, CYP2D6254 –271,
and CYP2D6321–351 sequences as immunodominant epitopes but
not in HCV infection (12, 16, 17).
Epitope mapping is an important step in understanding the
mechanisms triggering autoimmunity and in providing guidance
*Institute of Hepatology, †Department of Biology, Centre for Genetic Anthropology,
and ‡Department of Biochemistry and Molecular Biology, Bloomsbury Centre for
Structural Biology, University College London, London, United Kingdom; §Institute
of Liver Studies, King’s College Hospital, London, United Kingdom; and ¶Department of Internal Medicine, Cardioangiology, and Hepatology, University of Bologna,
Bologna, Italy
Received for publication March 11, 2002. Accepted for publication April 8, 2002.
The costs of publication of this article were defrayed in part by the payment of page
charges. This article must therefore be hereby marked advertisement in accordance
with 18 U.S.C. Section 1734 solely to indicate this fact.
1
This work was supported by a Royal Society (London, U.K.) Dorothy Hodgkin
Fellowship (to Y.M.), the Children’s Liver Disease Foundation (Birmingham, U.K.;
to D.P.B. and G.M.-V., and Grant NL1720 to Y.M.), and the Children Nationwide
Medical Research Trust (London, U.K.; to G.M.-V.).
2
Address correspondence and reprint requests to Dr. Diego Vergani, Institute of
Hepatology, University College London Medical School, 69 –75 Chenies Mews, London WC1E 6HX, U.K. E-mail address: D.Vergani@ucl.ac.uk
3
Abbreviations used in this paper: LKM1, liver kidney microsomal Ab type 1; AIH,
autoimmune hepatitis; HCV, hepatitis C virus.
Copyright © 2002 by The American Association of Immunologists, Inc.
for designing immunomodulatory therapy (18, 19). Epitope mapping comprises two major approaches: linear and conformational,
the former using prokaryotically expressed proteins or synthetic
peptides as target Ags, the latter using eukaryotically expressed
proteins. To date, linear epitope mapping of CYP2D6 has been
partially performed with either a limited number of synthetic
CYP2D6 peptides or by the use of constructs prokaryotically expressed. Although Yamamoto et al. (16) identified four linear
epitopes, CYP2D6254 –271, CYP2D6321–351, CYP2D6373–389, and
CYP2D6410 – 429 recognized by LKM1 in AIH, Dalekos et al. (20)
showed that in HCV infection, anti-CYP2D6-positive sera recognize two truncated CYP2D6 proteins (aa 250 – 494 and 321– 494)
expressed in Escherichia coli.
B cell epitopes, unlike T cell epitopes, are usually conformational, i.e., highly dependent on the three-dimensional structure of
the protein (21–26). The presence of conformational epitopes on
CYP2D6 was first suggested by Duclos-Vallee et al. (10) who
were unable to reverse inhibition of the CYP2D6 enzymatic activity by preincubation of LKM1-positive sera with synthetic peptides spanning known linear epitopes. The presence of conformational epitopes was confirmed by our group and by Yamomoto et
al. (14, 15). Sera from patients with AIH reacted with linear peptides, and prokaryotic and eukaryotic full-length protein, while
those from patients with HCV infection reacted only with the eukaryotically expressed, conformationally intact CYP2D6, leading
the two groups to conclude that LKM1 from HCV-infected patients preferentially reacts with conformational epitopes (14, 15).
To perform conformational epitope mapping of CYP2D6, we
decided to use two complementary strategies: first, to express a
full-length and a series of truncated CYP2D6 proteins in a eukaryotic system and second, to characterize the identified epitope by
mutagenesis introducing two homologous regions into CYP2D6
under the conditions in which the major antigenic regions remain
0022-1767/02/$02.00
278
structurally intact (22, 23). To further characterize the epitope, we
localize the epitopes within the CYP2D6 protein by modeling locally on the homologous region of published structure of Bacillus
megaterium P450 BM-3 (27, 28).
Materials and Methods
Construction of truncated CYP2D6 fragments
A complete cDNA of human CYP2D6 (kindly provided by Prof. U. Meyer,
University of Basel, Basel, Switzerland) was subcloned into the BamHI
and KpnI sites of plasmid pSP72 in vitro transcription vector (Promega,
Southampton, U.K.) downstream to a SP6 polymerase promoter. The subcloning was performed using a rapid ligation kit according to the manufacturer’s instructions (Roche Diagnostics, Lewes, U.K.). The recombinant
plasmid DNA was amplified through transformation of competent E. coli
(DH5a; Life Technologies, Paisley, U.K.) and then extracted and purified
with a plasmid DNA Miniprep kit (Qiagen, Crawley, U.K.) and finally used
as a template to construct truncated CYP2D6 fragments.
A total of 16 CYP2D6 constructs are shown in Fig. 1. They are divided
into three groups (A–C) and were obtained using different methods. Group
A (N-terminal region) and group C (C-terminal region) fragments were
constructed by digestion of the template with different combinations of
restriction endonucleases, which enabled maintenance of the reading
frame. Group B fragments were built by removing nuclear bases at the 3⬘
end of the full cDNA using Bal31 nuclease. The expected sizes of the
constructs were first examined by Agarose MS (molecular screening agarose; Roche Diagnostics) gel electrophoresis and then their sequences were
confirmed by nucleotide sequencing analysis (dye terminator) on an
ABI377 sequencer (PerkinElmer, Foster City, CA).
N-terminal region fragments CYP2D61–79 (A1), CYP2D61–190 (A2),
and CYP2D61–580 (A3) were obtained by partial digestion with restriction
EPITOPE MAPPING OF CYTOCHROME P4502D6
endonuclease enzyme NarI, which has three digestion sites on the CYP2D6
molecule at positions 79, 190, and 580. Partial digestion was applied so that
the enzyme would cut only on one or two of the three digestion sites in
each reaction. This task was achieved by using different quantities of enzyme and reaction times. NarI digestion products were further digested
with ClaI, which is located downstream of the KpnI site within vector
pSP72 and creates a NarI compatible end with that digested by NarI. The
aim of this step was to delete CYP2D6 sequence regions 79 –1567, 1901567, and 580-1567 and then to re-ligate the upstream and downstream
regions of the DNA template. Re-ligation was also performed with a rapid
ligation kit (Roche Diagnostics).
The second group of truncated fragments (B1 to B9) was constructed
using full-length CYP2D61–1567 as a template that remained continuous
with vector at the 5⬘ end and with reducing lengths at the 3⬘ end by removing nuclear bases using Bal31 nuclease according to the manufacturer’s instructions (Amersham Pharmacia Biotech, Little Chalfont, U.K.).
The 3⬘ end of CYP2D6 was then rendered blunt by Klenow reaction according to the manufacturer’s instructions (Promega). A series of truncated
fragments was finally released by 5⬘ BamHI digestion and subcloned into
BamHI and EcoRV sites of a new pSP72 vector.
C-terminal region fragments CYP2D61–79/580–927 (C1), CYP2D61–79/580–977
(C2), CYP2D61–79/580-1174 (C3), and CYP2D61–79/580-1567 (C4) were constructed through two steps. First, C4 was constructed by NarI partial digestion to delete CYP2D79 –580 and then followed by re-ligation of the
upstream and downstream regions of the template DNA. Second, C1 to C3
were constructed with the use of C4 as a template. Following initial digestion with EcoRV (location within vector pSP72), C4 was digested by
BstEII (partial digestion), BspEI (complete digestion), and Bsu36I (partial
digestion) respectively to build up fragments CYP2D6580 –927 (C1),
CYP2D6580 –977 (C2), and CYP2D6580 –1174 (C3).
FIGURE 1. Strategy for epitope mapping of CYP2D6 and summary of reactivity. The first dotted line (u) represents 1567 bp, the full-length CYP2D6
cDNA. Four restriction enzymes, NarI, BstEII, BspEI, and Bsu3, were used to construct truncated fragments, and their digestion sites on CYP2D6 are
shown. The second dotted line (u) represents the full-length CYP2D6 (497 aa) protein. The subsequent lines represent three groups of truncated CYP2D6
proteins. Their lengths are indicated by amino acid numbers. A1 to A3 (t) were created by NarI partial digestion, C1-C4 (solid lines) were created by
different combinations of restriction enzyme digestion, such as C4 by NarI at 79 and 580 sites and re-ligation; C1-C3 were created using C4 as a template
and then were digested by BstEII, BspEI, and Bsu3, respectively. B1-B9 (o) were created by removing bases at the 3⬘ end with nuclease Bal31 digestion.
The frequency of reactivity against each construct is shown in brackets followed by the number of sera reacted; a– c, frequency of reactivity in AIH
compared with that in HCV infection; p ⫽ 0.03, 0.05, and 0.06, respectively.
The Journal of Immunology
279
Production of chimeric constructs and in vitro mutagenesis
procedures
active measurement in a MicroBeta counter (EG & G, Milton Keynes,
U.K.) to evaluate the efficiency of the reaction.
Based on preliminary results, CYP2D6316 –327 was selected as a potential
immunodominant epitope, and its sequence was replaced by homologous
regions from cytochrome P4502C9 (CYP2C9, GenBank accession no.
P11712) and HCV genome polyprotein 1A (GenBank accession no.
P26664) to generate chimeric constructs CYP2D6/2C9 and CYP2D6/
HCV, respectively (Fig. 2). The homologous regions between CYP2D6,
CYP2C9, and HCV proteins were identified by using the protein database
local homology search tool BLAST2 program (http://www.ncbi.nlm.nih.
gov/blast/b12seq/b12.html). rCYP2D61–1567 (full-length) and CYP2D61–
1275 (B8) in the pSP72 vector (Promega) were used as templates. A twostep strategy was applied: the first step was to produce chimeric fragments
by PCR and the second was to subclone them into a new pSP72 vector.
Each 50-␮l PCR consisted of 35 cycles of denaturation for 30 s at 94°C,
annealing for 30 s at 55°C, and extension at 72°C for 1 min in an automated
thermal cycler (Roche Diagnostics) with the use of Super TAQ polymerase
(HT Biotechnology, Cambridge, U.K.) and a TaqStart Ab (Clontech Laboratories, Basingstoke, U.K.). A 5 U/␮l solution of Taq and a 7-␮mol
solution of TaqStart Ab were mixed in a ratio of 2:1 before setting up the
PCR. The reaction was initiated by a 5-min incubation at 94°C and was
completed with a 30-min extension at 72°C.
Two pairs of internal oligonucleotide primers that overlapped by 10 –11
nucleotide bases and were engineered with endonuclease restriction sites
were used to introduce five amino acid substitutions of CYP2C9 (alanine,
leucine, leucine, lysine, and glutamic acid with corresponding CYP2D6
glycine (317), methionine (321), isoleucine (322), leucine (323), and aspartic acid (326)) and three amino acid substitutions of HCV protein (proline, leucine, and leucine of HCV protein with corresponding CYP2D6
317, 321 and 322). Combining the use of two flanking primers spanning the
5⬘ and the 3⬘ ends of multiple cloning site of the vector pSP72, the two
overlapping fragments (for each construct) carrying the mutations were
amplified and joined by AflII (CYP2C9 mutation) or AatII (HCV mutation)
cleavage and ligation. Following BamHI and EcoRV cleavage at the 5⬘ and
3⬘ ends, respectively, the final chimeric products were subcloned into corresponding sites of vector pSP72.
The rDNA was amplified by transforming competent E. coli (DH5a;
Life Technologies), which was cultured on agar plates using ampicillin
resistance for selection of potential clones. The correct chimeric constructs
were first checked by restriction enzyme digestion and finally confirmed by
DNA sequencing analysis using an ABI377 sequencer (PerkinElmer).
The primers used were (restriction enzyme sites underlined): CYP2C9
mutating primers, an AflII sense primer: 5⬘-TGC TTA AGC ATC CGG
AGG TGC AGC GCC GTG T-3⬘, and an AflII anti-sense primer: 5⬘-TGC
TTA AGC AAG AGC AGG AGG GCC CAG GCC-3⬘; HCV mutating
primers, an AatII sense primer: 5⬘-CGG ACG TCC AGC GCC GTG TCC
AAC AG-3⬘, and an AatII antisense primer 5⬘-TGG ACG TCC GGA TGT
AGA AGC AAG AGC AGG AGG GGC CAG G-3⬘; flanking primers, a
sense primer: 5⬘-GAA CTC GAG CAG CTG AAG CTT GCA TG-3⬘, and
an antisense primer: 5⬘-CGA CTC ACT ATA GGG AGA CCG GCA
GAT-3⬘.
Radioimmunoprecipitation assay for detection of Abs to
CYP2D6 protein
Inhibition of reactivity to 35S-labeled CYP2D6 was investigated using four
competitors: 1) full-length CYP2D6 expressed in a eukaryotic system
(Gentest, Woburn, MA); 2) truncated CYP2D61–375 (construct B8); 3)
CYP2D61–375/CYP2C9310 –320 mutant; and 4) CYP2D61–375/HCV794 – 801
mutant.
Two sera, one from a patient with AIH (No. 3, Table I) and the other
from a patient with HCV infection (No. 1, Table II), were selected for the
inhibition studies, because they reacted with, and were inhibited by, all the
constructs used in the inhibition assay.
Fifty-microliter aliquots of the two sera diluted at 1/200 were incubated
at room temperature for 4 h with one of the following: 1) 3.0 ␮g of
rCYP2D6 protein (⬃0.03 ␮g purified CYP2D6; Gentest) or 2) 2 ␮l (⬃0.06
␮g) of individual truncated/mutant CYP2D6 protein. The quantities of the
Ags used in the inhibition assays and serum dilution were determined from
preliminary experiments. Ab binding was then conducted under identical
conditions to those described above. Percent inhibition was calculated as
100 ⫻ (1 ⫺ (inhibited cpm/uninhibited cpm)).
In vitro-coupled transcription and translation of CYP2D6
Three-dimensional modeling of CYP2D6
Full-length, truncated, or mutant rCYP2D6 DNA were used as templates to
express protein eukaryotically by an in vitro transcription/translation system using a TNT-coupled reticulocyte lysate kit (Promega) as described
previously (14). In brief, 1.0 ␮g of plasmid DNA was incubated at 30°C for
2 h in a 50-␮l mixture containing 25 ␮l rabbit reticulocyte, 1 ␮l RNA
polymerase (40 U/␮l), 1 ␮l amino acid mixture (minus methionine) (1
mM), RNasin inhibitor (20 U/␮l), and [35S]methionine (10 Ci/ml) (Amersham Pharmacia Biotech). The translation product was then evaluated by
incubating 1 ␮l of the reaction mixture with 1 M NaOH/2% H2O2, precipitated with 25% trichloroacetic acid and was then harvested onto a
Whatman GF/A glass fiber filter (Waterman, Maidstone, U.K.) for radio-
The Swiss-Model, Swiss-Pdb Viewer (29) (both from http://www.
expasy.ch/swissmod/SWISS-MODEL.html), and RasMol 2.6 (http://www.
umass.edu/microbio/rasmol/) programs were used for investigating and analyzing the derived CYP2D6 protein structure.
Sera from patients were diluted 1/50 and incubated in immunoprecipitation
buffer (20 mM Tris/150 mM NaCl, 0.15% Tween, 0.1% aprotinin, 10 mM
benzamidine, 0.1% BSA) at 4°C overnight with a 10,000 cpm aliquot of
recombinant protein. The Ab-bound protein was separated from free Ag by
the addition of protein A-Sepharose (Pharmacia Biotech, St. Albans, U.K.)
and incubated for 1 h at 4°C. The total immunoprecipitation mixture was
then transferred into 96-well filtration plates with 0.45-mm filters (multiscreen durapore HVPP membranes; Millipore, Bedford, MA). The immunocomplex bound Sepharose beads were washed 10 times with washing
buffer (20 mM Tris/150 mM NaCl, 0.1% Tween 20, 0.1% BSA) and then
fixed to dried multiwell filters by MeltiLex (Millipore). The radioactivity of
the precipitated protein was counted on multiwell filters in a MicroBeta
count (EG & G).
All test sera were assessed in duplicate. A value was considered positive
when the cpm was ⬎1050, this representing 3 SD above the mean in 57
normal controls against the full-length protein. Cut-off point for truncated
proteins was calculated transferring into each assay 20 normal control
samples.
cpm due to reactivity against wild type minus cut-off cpm was assigned
a value of 100%. Reduction of reactivity to the mutants (cpm minus cutoff) was expressed as a percentage of the value against wild type.
The intraassay coefficient of variation, calculated on 30 occasions,
ranged between 4 and 6% against the full-length CYP2D6 protein. The
interassay coefficient of variation, calculated on 80 occasions, ranged between 9 and 16%. To control for intra and interassay variation, one strong
positive and five negative sera were transferred in each assay, and were
tested in quadruplicate.
Inhibition studies
LKM1
LKM1 was investigated by indirect immunofluorescence on 5-␮m cryostat
sections of rat livers, kidneys, and stomachs at the initial dilution of 1/10
in PBS (30). Positive sera were double-diluted to extinction.
Viral tests
Abs to HCV were detected by second-generation ELISA (ELISA II; Chiron, Emeryville, CA) and confirmed by RIBA II (United Biomedical,
Hauppage, NY). HCV RNA was detected by nested PCR within the 5⬘
highly conservative noncoding region (Amplicor; Hoffmann La Roche,
Basel, Switzerland).
Patients and controls
FIGURE 2. Homologous regions between CYP2D6316–327, CYP2C9309–320,
and HCV794 – 801. Shaded boxes correspond to CYP2D6/2C9 and CYP2D6/
HCV point mutants. Vertical bars indicate boundaries for swap mutants.
Highly conservative amino acid differences are shown in bold.
A total of 32 LKM1-positive patients were studied. Sixteen patients had
classical LKM1-positive AIH diagnosed according to international criteria
(31). Thirteen were female with a median age of 10 years, ranging from 1
to 20 years. The median LKM1 titer was 1/640, ranging from 1/20 to
⬍0.02
⬍0.01
⬍0.005
⬍0.01
⬍0.02
0.01
0.02
NS
NS
0.06
⬍0.05
NS
NS
p Valueb
4,328
⫺
⫺
⫺
2,166
2,218
2,276
2,390
2,447
2,634
2,679
4,238
4,161
2,545
2,699
3,288
4,242
1
10,240
4,422
⫺
⫺
⫺
2,126
2,140
2,307
2,231
2,591
2,554
2,884
3,268
3,525
2,747
2,618
3,302
4,660
2
10,240
3,580
⫺
⫺
⫺
2,681
2,741
2,680
2,698
2,432
2,471
2,824
3,299
3,361
2,928
3,281
3,232
3,542
3
1,280
3,741
⫺
⫺
⫺
2,102
2,128
2,306
2,341
2,274
2,543
2,430
3,028
3,165
2,126
2,114
2,280
3,910
4
640
2,245
⫺
⫺
⫺
⫺
1,195
1,156
1,123
1,394
1,421
1,607
1,629
1,955
1,700
1,473
1,908
1,883
5
40
4,032
⫺
⫺
⫺
⫺
2,125
2,148
2,392
2,760
2,681
2,846
2,941
3,490
⫺
2,141
3,768
3,550
6
640
b
4,910
⫺
⫺
⫺
⫺
⫺
⫺
2,362
2,849
2,781
3,042
3,188
3,720
3,345
3,122
3,601
4,201
7
1,280
3,418
⫺
⫺
⫺
⫺
⫺
⫺
2,128
2,492
2,646
2,880
2,886
2,269
3,205
2,982
3,396
3,458
8
1,280
4,292
⫺
⫺
⫺
⫺
⫺
⫺
2,199
2,341
2,406
2,704
2,764
3,385
2,386
2,838
3,706
3,840
9
1,280
3,699
⫺
⫺
⫺
⫺
⫺
⫺
2,106
2,207
2,532
2,501
2,620
2,890
2,906
3,341
3,566
3,476
10
640
1,471
⫺
⫺
⫺
⫺
⫺
1,089
⫺
1,170
1,122
1,261
1,143
1,278
⫺
⫺
1,398
1,246
11
160
2,068
2,302
2,398
2,377
2,364
2,453
2,543
2,680
2,174
2,325
2,617
2,676
2,743
Mean cpm
0.08
0.008
0.03
0.01
NS
NS
NS
NS
0.004
NS
NS
NS
p Value
2,814
⫺
⫺
⫺
2,134
⫺
⫺
2,304
2,440
2,505
2,456
3,027
3,340
2,102
2,386
2,262
2,710
1
640
3,110
⫺
⫺
⫺
⫺
2,292
2,368
2,461
2,429
2,457
2,701
2,834
2,947
1,560
1,598
2,460
2,792
2
320
3,100
⫺
⫺
⫺
⫺
2,104
2,112
2,418
2,385
2,743
2,708
2,986
3,011
2,417
2,428
2,548
2,562
3
320
3,482
⫺
⫺
⫺
⫺
⫺
2,161
2,827
2,474
2,512
2,882
3,014
3,365
2,310
3,031
2,982
3,578
4
320
2,935
⫺
⫺
⫺
⫺
⫺
⫺
2,180
2,444
2,760
2,784
2,961
3,060
2,282
2,964
3,228
3,292
5
640
3,275
⫺
⫺
⫺
⫺
⫺
2,161
⫺
2,225
2,320
2,369
2,707
2,831
2,438
2,406
2,966
3,155
6
320
3,320
⫺
⫺
⫺
⫺
2,389
2,432
2,322
⫺
⫺
⫺
⫺
2,741
⫺
⫺
2,619
3,035
7
320
3,067
⫺
⫺
⫺
⫺
2,384
2,580
⫺
⫺
⫺
⫺
⫺
2,886
⫺
⫺
3,206
3,158
8
640
2,777
⫺
⫺
⫺
⫺
⫺
⫺
2,274
⫺
⫺
2,204
3,094
3,252
2,106
2,436
2,961
2,892
9
640
2,448
⫺
⫺
⫺
⫺
⫺
⫺
⫺
2,243
⫺
2,425
⫺
⫺
⫺
⫺
⫺
2,431
10
320
Patient No./LKM1 Titer a
1,806
⫺
⫺
⫺
⫺
⫺
⫺
⫺
⫺
1,253
1,550
1,556
1,502
⫺
1,353
1,323
1,846
11
320
12
320
1,593
⫺
⫺
⫺
⫺
⫺
⫺
⫺
⫺
1,056
1,116
1,235
1,262
1,256
1,165
1,313
1,458
12
160
2,245
⫺
⫺
⫺
⫺
⫺
⫺
⫺
⫺
⫺
⫺
2,476
2,691
⫺
⫺
2,227
2,446
LKM1 titers are shown as reciprocal values. Reactivity to CYP2D6 constructs is shown in cpm, the cut-off point is 1,050; ⫺, cpm below the cut-off point.
A value of p comparing mean cpm of truncated constructs with that of full-length protein.
1–497
1–26
1–63
1–193
1–225
1–265
1–267
1–317
1–324
1–327
1–337
1–375
1–417
1–26/193–309
1–26/193–325
1–26/193–391
1–26/193–497
Wild type
Wild type
A1
A2
A3
B1
B2
B3
B4
B5
B6
B7
B8
B9
C1
C2
C3
C4
a
CYP2D6 Ag
Constructs
Table II. Reactivity to full-length and truncated CYP2D6 proteins in patients with HCV infection
b
2,269
2,091
1,883
2,197
2,269
2,152
2,215
2,487
2,501
2,404
2,345
2,675
2,957
3,110
Mean cpm
Patient No./LKM1 Titer a
LKM1 titers are shown as reciprocal values. Reactivity to CYP2D6 constructs is shown in cpm, the cut-off point is 1,050; ⫺, cpm below the cut-off point.
A value of p comparing mean cpm of truncated constructs with that of full-length protein.
1–497
1–26
1–63
1–193
1–225
1–265
1–267
1–317
1–324
1–327
1–337
1–375
1–417
1–26/193–309
1–26/193–325
1–26/193–391
1–26/193–497
Wild type
A1
A2
A3
B1
B2
B3
B4
B5
B6
B7
B8
B9
C1
C2
C3
C4
a
CYP2D6 Ag
Constructs
Table I. Reactivity to full-length and truncated CYP2D6 proteins in patients with autoimmune hepatitis
2,937
⫺
⫺
⫺
⫺
⫺
⫺
⫺
⫺
⫺
⫺
1,928
2,254
⫺
⫺
⫺
3,140
13
640
2,005
⫺
⫺
⫺
⫺
⫺
⫺
⫺
⫺
1,134
1,138
1,331
1,317
1,298
1,430
1,543
2,402
13
160
2,278
⫺
⫺
⫺
⫺
⫺
⫺
⫺
⫺
⫺
⫺
1,391
⫺
⫺
⫺
⫺
1,806
14
320
2,232
⫺
⫺
⫺
⫺
⫺
1,106
⫺
⫺
⫺
1,091
⫺
1,354
⫺
1,279
1,146
1,937
14
160
1,908
⫺
⫺
⫺
⫺
1,169
⫺
⫺
⫺
⫺
⫺
⫺
1,179
⫺
⫺
⫺
1,790
15
80
1,817
⫺
⫺
⫺
⫺
⫺
⫺
⫺
⫺
⫺
⫺
1,249
1,607
⫺
⫺
⫺
1,589
15
20
2,379
⫺
⫺
⫺
⫺
⫺
⫺
⫺
⫺
⫺
⫺
⫺
2,456
⫺
⫺
⫺
2,178
16
320
1,970
⫺
⫺
⫺
⫺
⫺
⫺
⫺
⫺
⫺
⫺
⫺
1,280
⫺
⫺
⫺
1,921
16
640
16/16 (100)
0/16
0/16
0/16
1/16 (6)
5/16 (31)
6/16 (38)
7/16 (44)
7/16 (44)
7/16 (44)
9/16 (56)
11/16 (69)
14/16 (88)
7/16 (44)
8/16 (50)
11/16 (69)
16/16 (100)
Frequency of Reactivity (%)
16/16 (100)
0/16
0/16
0/16
4/16 (25)
6/16 (38)
8/16 (50)
10/16 (63)
11/16 (69)
13/16 (81)
14/16 (88)
14/16 (88)
16/16 (100)
11/16 (69)
13/16 (81)
14/16 (88)
16/16 (100)
Frequency of Reactivity (%)
280
EPITOPE MAPPING OF CYTOCHROME P4502D6
The Journal of Immunology
1/10,240. Ab to HCV and HCV RNA were negative in all of them. Of the
16 patients, 12 were investigated at diagnosis before starting immunosuppressive treatment, while 4 were on immunosuppression at the time of
investigation consisting of prednisolone (0.5–2 mg/kg/day), 3 with and 1
without azathioprine (1–2 mg/kg/day). The other 16 patients had chronic
HCV infection. Eleven were female with a median age of 48 years, ranging
from 19 to 70 years. The median LKM1 titer was 1/320, ranging from 1/80
to 1/640. All were HCV RNA-positive. Four of the patients had received
IFN treatment. The remaining 12 patients were untreated. This project was
approved by the Ethics Committees of King’s College Hospital (London,
U.K.) and University of Bologna (Bologna, Italy). The project was also
approved by the Safety Advisory Unit of University College London (GM
Project No. 663).
Statistics
␹2 and one-tailed Fischer’s exact tests were used to compare frequencies of
reactivity to different constructs in patient groups. The normality of variable distributions was tested using the Kolmogorov-Smirnov goodnessof-fit test. Differences in anti-CYP2D6 levels between patient groups, and
between Ab levels to full-length and truncated proteins were analyzed using the Student’s t test. A value of p ⬍ 0.05 was considered significant.
281
AIH and 7 of 16 (44%) patients with HCV infection. Reactivity to
C2 was observed in 81 and 50%, respectively ( p ⫽ 0.06), which
was comparable to that against B5, the B construct (without deletion) of similar length. Reactivity to C3, containing the four previously defined linear epitopes, 196 –218, 254 –271, 321–351, and
373–389, was seen in 14 of 16 (88%) in AIH and 11 of 16 (69%)
in HCV infection, which was similar to that against the comparable
construct (B8). The frequency of reactivity to construct C4, which
contains the full 193– 497 sequence, was 100% in both AIH and
HCV infection groups.
In patients with AIH, mean cpm levels against the full-length
protein were significantly higher than those against truncated proteins (B1-B7, C1, and C2 shown in Table I). In patients with
LKM1/HCV infection, mean cpm levels against full-length protein
were also higher than those against truncated proteins (B2-B5 and
C1, Table II). There were no significant differences in Ab levels
between patient groups.
Reactivity to CYP2D6 chimeric mutants
Results
Reactivity to full-length and truncated CYP2D6 proteins
Frequency of reactivity to full-length and truncated CYP2D6 proteins in sera from patients with AIH and HCV infection is shown
in Fig. 1. Individual readings are shown in Tables I and II.
A 100% frequency of reactivity to the full-length CYP2D6 protein was observed in the sera from both LKM1-positive AIH and
HCV infection, while no reactivity was observed against the amino-terminal constructs A1, A2, and A3 in the sera from both AIH
and HCV infection.
Group B constructs were introduced to define, first, the residues
critical for Ab binding and second the regions of differential reactivity between AIH and HCV infection. A total of nine truncated
CYP2D6 proteins were expressed, covering the region from starting codon to aa 417. The differences between their lengths ranged
from 2 (CYP2D61–265 and CYP2D61–267) to 50 (CYP2D61–267
and CYP2D61–317) aa.
Reactivity to group B constructs showed a stepwise increase
among the patients with AIH and HCV infection in terms of frequency (Fig. 1) and Ab levels (cpm; mean cpm shown in Tables I
and II). Reactivity to B1 was observed in 4 of 16 (25%) patients
with AIH, significantly more frequent than against A1, A2, and A3
( p ⫽ 0.03 for all), and in 1 of 16 (6%) patients with HCV infection. Reactivity to B2 was observed in 38% of patients with AIH
and in 31% with HCV infection, respectively. Frequency of reactivity to all group B constructs (B1 to B9) was higher in AIH than
in HCV infection. Frequency of reactivity to CYP2D61–327 (B6)
and to CYP2D61–337 (B7) was significantly higher in AIH than in
HCV infection (13 of 16, 81% vs 7 of 16, 44%, p ⫽ 0.03, and 14
of 16, 88% vs 9 of 16, 56%, p ⫽ 0.05, respectively). When the
length of CYP2D6 protein increased from 317 to 327 aa, the frequency of reactivity increased from 63 to 81% in AIH but remained the same (44%, p ⫽ 0.03) in HCV infection.
Differences in only two amino acids could remarkably alter frequency of reactivity which increased from 38 to 50% in patients
with AIH and from 31 to 38% in patients with HCV infection
when B2 (CYP2D61–265) and B3 (CYP2D61–267) were used as
targets.
Four truncated C-terminal constructs, C1-C4, lacking the amino-terminal region 27–192 and increasing in length toward the C
terminus, were introduced to investigate further the role of region
27 to 192 on Ab binding. Frequency and levels of Abs increased
gradually from C1-C4 (see Tables I and II). Reactivity to C1, containing major linear epitopes of CYP2D6, 196 –218 and 254 –271,
respectively (12, 16), was observed in 11 of 16 (69%) patients with
Based on the results obtained by epitope mapping showing that
CYP2D6316 –327 was able to differentiate between AIH and HCV
infection, this sequence was selected for further analysis by introducing site-specific mutations into it. Within this region,
CYP2D6317–327 shares 81% (9/11) homology with CYP2C9, including six identical and three conservative amino acids (Fig. 2)
and CYP2D6316 –323 shares 88% (7/8) homology including five
identical and two conservative amino acids, with polyprotein 1A of
HCV (Fig. 2), a viral sequence suggested to act as the molecular
mimic capable of inducing tolerance breakdown against CYP2D6
(32). CYP2D6/2C9 and CYP2D6/HCV mutations were introduced
into construct B8 (CYP2D61–375) and full-length CYP2D6 thus
generating four mutants: 1) CYP2D61–375/CYP2C9310 –320; 2)
CYP2D61–375/HCV794 – 801; 3) CYP2D61– 497/CYP2C9310 –320;
and 4) CYP2D61– 497/HCV794 – 801.
Reduction of reactivity to all four mutants was observed in both
groups of patients. Frequency of sera showing a significant reduction in binding (cpm), i.e., ⱖ 30% when compared with that
against the unmodified polypeptides (B8 and wild-type CYP2D6),
was analyzed. Reduction in binding to CYP2D61–375/CYP2C9310 –320,
CYP2D61–375/HCV794 – 801, and CYP2D61– 497/CYP2C9310 –320
(mutants I, II, and III) was seen in 5 of 14 (36%), 10 of 14 (71%),
and 11 of 16 (69%) patients with AIH, and in 5 of 11 (45%), 7 of
11 (64%), and 9 of 16 (56%) patients with HCV infection. Although the reduction in reactivity was similar in AIH and HCV
patients against mutants I, II, and III, that against mutant IV was
significantly greater in AIH (12 of 16, 75%) than in HCV patients
(6 of 16, 38%, p ⫽ 0.03).
Inhibition studies
Results of inhibition are shown in Fig. 3. Four proteins were used
as target Ags and also as inhibitors. They were: 1) full-length
CYP2D6 protein; 2) construct B8 (CYP2D61–375); 3) mutant I
(CYP2D61–375/CYP2C9310 –320); and 4) mutant II (CYP2D61–375/
HCV794 – 801). Preincubation of the AIH serum with the full-length
protein resulted in 68, 82, 91, and 95% inhibition against full protein, construct B8, and mutants I and II, respectively (Fig. 3A).
Preincubation with construct B8 led to inhibition against itself and
against mutant I, but neither against the full-length protein nor
against mutant II. Preincubation with mutant I led to a high-level
inhibition against itself and mutant II and a low-level inhibition
against construct B8. Similarly, mutant II could reduce Ab binding
against itself and also against mutant I (Fig. 3A). When the serum
from a patient with HCV infection was preincubated with the four
competitors, a similar pattern of inhibition was observed (Fig. 3B),
282
EPITOPE MAPPING OF CYTOCHROME P4502D6
epitope displays a hydrophobic patch that is situated between an
N-terminal aromatic residue (W316) and a C-terminal histidine
(H324). The negatively charged D326 is expected to play an important role in epitope recognition through electrostatic interaction.
The amino acid composition of CYP2C9310 –320 is very similar
with the exception of a bulky positively charged residue (K316)
just C-terminal of the hydrophobic patch. This L323K swap (conservative amino acid replacement) may substantially affect Ab
binding, because the insertion of a large positively charged residue
(K) may interfere with surface complementarity or electrostatic
interactions. The structure of HCV794 – 801 is similar to its corresponding region of CYP2D6316 –323, because the hydrophobic patch still
remains after introducing M321L and I322L mutations (conservative
amino acid replacement). N-terminal amino acid replacement was
performed by switching glycine 317 to proline 796.
Discussion
FIGURE 3. Inhibition with full-length CYP2D6 protein, CYP2D61–375,
CYP2D61–375/CYP2C9310 –320, and CYP2D61–375/HCV794 – 801 mutants.
The two sera, one from a patient with AIH and the other from a patient with
HCV infection, were preincubated with the inhibitors shown in the x-axis:
full-length CYP2D6 (f); CYP2D61–375 (construct B8) (p); CYP2C9 mutant (mutant I) (t), and HCV mutant (mutant II) (䡺). Percent inhibition is
shown on the y-axis (specified under the bars). A, The percent inhibition
after preincubation of the serum from a patient with AIH; B, The serum
from a patient with HCV infection.
the only difference being that construct B8 inhibited Ab binding
against mutant I to a lesser extent than that seen in AIH serum
(35% in HCV serum vs 72% in AIH serum).
Three-dimensional modeling of CYP2D6316 –327 and its mutants
CYP2D6316 –327, CYP2D6/CYP2C9310 –320, and CYP2D6/
HCV794 – 801 mutants were modeled locally on the homologous
region in B. megaterium cytochrome P450 BM-3 (27, 28). Their
three-dimensional structures are shown in Fig. 4. The sequence
CYP2D6316 –327 is located on the surface of the protein. This
Using eukaryotically expressed full-length protein and constructs
thereof, we show that CYP2D6 is divided into two main regions
with respect to antigenicity. The N-terminal region comprising
amino acid 1–193 appears to be devoid of antigenicity, while the
C-terminal 193– 497 accounts for all LKM1 reactivity. This is true
whether the LKM1 positive sera used as probes are derived from
patients with AIH or patients with HCV infection. Moreover, we
show that the sequence 316 –327, which is part of a larger region
capable of differentiating LKM1 reactivity in AIH and HCV infection, is exposed on the surface of the molecule.
The choice of using eukaryotically expressed full-length protein
and constructs was dictated by the fact that some 70% of HCV/
LKM1-positive sera do not react with the prokaryotically expressed
molecule (13), indicating that there must be posttranslational modifications of critical importance for recognition of the enzyme by HCV/
LKM1-positive sera. In the present study, we were able to confirm
that the totality of the LKM1-positive sera from either condition recognize the eukaryotically expressed cytochrome.
The use of a set of three N-terminal constructs A1, A2, and A3
enabled us to show that none of the LKM1-positive sera was reactive with the region comprising aa 1–193. Reactivity increased
stepwise from 0 to 100% using constructs increasing in length
toward the C terminus. It could be argued that the region 1–193,
though not harboring identifiable B cell epitopes, may influence
the antigenicity of the molecule, contributing to its tertiary structure. Experiments using the C constructs with the aa 27–192 deletion, do not support this hypothesis. All LKM1-positive sera reacted with construct C4, which contains the entire sequence
C-terminal to aa 193, but is devoid of the N-terminal region aa
27–192. This indicates that the removal of 159 aa from the Nterminal 1–192 sequence has no direct or indirect effect on the
antigenicity of the whole molecule expressed eukaryotically.
Constructs of differing lengths, with or without the 27–192 deletion, were used to characterize further areas of antigenicity. The
finding that constructs B6 (1–327) and C3 (1–26, 193–391) are
recognized by four-fifths of the LKM1-positive sera from patients
with AIH was expected because both contain the sequence 254 –
271, the major linear epitope as defined by synthetic peptides (16).
Of interest is the finding that one-quarter of the patients with AIH
recognize the construct B1 (1–225), which does not contain any of
the reported linear epitopes; this indicates that 33 aa C-terminal to
position 193 impart antigenicity to the molecule. In keeping with
this finding, our group has recently shown that the sequence comprising aa 193–212, expressed as a synthetic peptide, harbors a
linear epitope which is recognized by all sera of patients with AIH
and by 50% of those with HCV infection. The lower frequency of
reactivity against B1, as observed in this study, compared with that
The Journal of Immunology
283
FIGURE 4. Homology models of epitopes regions.
CYP2D6316 –327, CYP2D6/CYP2C9310 –320, and CYP2D6/
HCV794 – 801 mutants were modeled locally on the homologous region in the B. megaterium cytochrome P450 BM-3
using SWISS-MODEL (PDB code 1FAH.pdb). Epitope
residues in CYP2D6 and its CYP2C9 and HCV mutants are
displayed in stick mode and colored according to physicochemical properties or individual amino acid identities: orange, aromatic; brown, small; yellow, hydrophobic; cyan,
histidine; green, proline; red, negatively charged; blue,
positively charged. Amino acids in the CYP2D6 sequence
are shown in italics.
using synthetic peptide 193–212 could be due to the fact that a
partial unfolding of the molecule is required for this linear epitope
to be recognized.
The sequence containing the first 265 aa (B2) is recognized by
a similar proportion of sera from patients with AIH (38%) and
patients with HCV infection (31%). The frequency of reactivity
increases steadily with the C-terminal extension of the molecule in
AIH, but not in HCV infection. Thus 80% of sera from patients
with AIH react with the construct B6 (1–327) compared with only
44% of those from patients with HCV infection. This significant
difference is also detectable using a sequence 10 aa longer (B7) or
one containing the 27–192 deletion (C2).
This lower reactivity against region 316 –337 in HCV/LKM1 as
compared with AIH/LKM1 sera, is an intriguing finding. An octameric sequence WGLLLMIL (CYP2D6316 –323) contained in this
region shares 7 aa identities or conserved substitutions with the
octamer WPLLLLLL (HCV794 – 801) of the HCV polyprotein and
was suggested to be a major trigger of HCV-induced anti-CYP2D6
autoimmunity. Our findings would appear to question this notion.
However, it is possible for this sequence to be involved in Ag
recognition as a CD4 cross-reactive epitope in chronic HCV infection. When we replaced this wild-type sequence within
CYP2D6 with the homologous viral sequence, sera from both AIH
and HCV infection lost reactivity significantly. Studies of the
284
structure of the HCV region (HCV794 – 801) corresponding to that
of CYP2D6 (CYP2D6316 –323) show that they are similar toward
the C terminus, the two amino acids swapped being similar in size.
The change from glycine 317 to proline 796 may have major consequence, however. Although both glycine and proline are small
amino acids, proline may be introducing a “kink” in the epitope
helix, leading to a changed geometry between the aromatic W
(316) and the hydrophobic patch. In the middle of a helix, the
proline ring pushes away the preceding turn of the helix by ⬃1 Å
on that side, producing a bend of ⬃30° in the helix-axis and also
breaking the next hydrogen bond. These changes would readily
explain the lost reactivity we observed.
Experiments performed using another swap mutant, where the
CYP2D6 sequence was replaced by its homologue from CYP2C9,
the target of the autoantibody LKM2 (33), also showed a decreased
reactivity by LKM1-positive sera from both patients with AIH or
HCV infection. This confirms that the region modified in the swap
mutants is part of a major antigenic determinant. Despite the fact
that the three locally modeled structures of this region are similar,
all containing a hydrophobic patch between an N-terminal aromatic residue (W316) and a C-terminal histidine, the two swap
mutants not only react less with LKM1-positive sera, but are also
less efficient competitors in the inhibition studies. Though the two
swap mutants differ between themselves, the CYP2C9 mutant having a C-terminal amino acid mutation L323K and the HCV mutant
an N-terminal switch between glycine 317 and proline 796, they
are more similar to each other than to the wild protein and behave
similarly in inhibition studies. Taken together, the combinations of
all these mutations may rearrange the epitope helix and greatly
alter the properties of the epitope region, particular in terms of
surface shape and charge, and consequently result in the loss of Ab
binding.
Relevant to the present findings is our previous observation that
the region spanning 316 –337 aa, which we show in the present
study to be capable of differentiating LKM1 reactivity in AIH and
HCV infection, is the focus of cross-reactive autoimmunity (34).
In that report, we showed that CYP2D6321–339 shares extensive
homology with carboxypeptidase H (35) 33–51 and 21-hydroxylase (an autoantigen in Addison’s disease; Ref. 36) 307–325 and
that the serum of a patient suffering from AIH type 2, insulindependent diabetes, and Addison’s disease contained autoantibodies reacting with the CYP2D6 sequence and cross-reacting with
the homologous regions of carboxypeptidase H and 21-hydroxylase (34), a finding outlining the key role of the region in the
generation of autoimmunity.
The aa sequence 316 –327, which is part of the major antigenic
determinant described in this study, is exposed on the surface of
the protein. This, together with the recent demonstration that the
CYP2D6 molecule is expressed on the hepatocyte membrane and
is targeted by LKM1 (37, 38), suggests that this sequence may be
directly involved in autoantibody mediated liver cell damage, akin
to the scenario in myasthenia gravis (39), autoimmune thyroiditis
(40), autoimmune thrombocytopenia purpura (41), hemolytic anemia (42), and neonatal systemic lupus erythematosus (43).
The use of sera from two distinct diseases, AIH and chronic
HCV infection, sharing the same autoantibody has allowed us to
provide an initial characterization of the antigenic constitution of
the target. Further studies should aim at establishing human mAbs
to CYP2D6 from both clinical conditions to obtain information on
the fine antigenic specificity. Moreover, in view of the fact that
LKM1 belongs to the IgG class implying a T cell-directed class
switch, the role of CD4 cells in the generation of the autoimmune
attack should be investigated.
EPITOPE MAPPING OF CYTOCHROME P4502D6
Acknowledgments
We thank Dr. Kaushik Choudhuri for reviewing the manuscript and giving
critical comments.
References
1. Alvarez, F., O. Bernard, J. C. Homberg, and G. Kreibich. 1985. Anti-liver-kidney
microsome antibody recognizes a 50,000 molecular weight protein of the endoplasmic reticulum. J. Exp. Med. 161:1231.
2. Rizzetto, M., F. B. Bianchi, and D. Doniach. 1974. Characterization of the microsomal antigen related to a subclass of active chronic hepatitis. Immunology
26:589.
3. Maggiore, G., O. Bernard, J. C. Homberg, M. Hadchouel, F. Alvarez,
P. Hadchouel, M. Odievre, and D. Alagille. 1986. Liver disease associated with
anti-liver-kidney microsome antibody in children. J. Pediatr. 108:399.
4. Homberg, J. C., N. Abuaf, O. Bernard, S. Islam, F. Alvarez, S. H. Khalil,
R. Poupon, F. Darnis, V. G. Levy, P. Grippon, et al. 1987. Chronic active hepatitis associated with antiliver/kidney microsome antibody type 1: a second type
of “autoimmune” hepatitis. Hepatology 7:1333.
5. Gueguen, M., A. M. Yamamoto, O. Bernard, and F. Alvarez. 1989. Anti-liverkidney microsome antibody type 1 recognizes human cytochrome P450 db1.
Biochim. Biophys. Acta 159:542.
6. Lenzi, M., G. Ballardini, M. Fusconi, F. Cassani, L. Selleri, U. Volta, D. Zauli,
and F. B. Bianchi. 1990. Type 2 autoimmune hepatitis and hepatitis C virus
infection. Lancet 335:258.
7. McFarlane, I. G., H. M. Smith, P. J. Johnson, G. P. Bray, D. Vergani, and
R. Williams. 1990. Hepatitis C virus antibodies in chronic active hepatitis: pathogenetic factor or false-positive result? Lancet 335:754.
8. Zanger, U. M., H. P. Hauri, J. Loeper, J. C. Homberg, and U. A. Meyer. 1988.
Antibodies against human cytochrome P-450db1 in autoimmune hepatitis type II.
Proc. Natl. Acad. Sci. USA 85:8256.
9. Manns, M. P., E. F. Johnson, K. J. Griffin, E. M. Tan, and K. F. Sullivan. 1989.
Major antigen of liver kidney microsomal autoantibodies in idiopathic autoimmune hepatitis is cytochrome P450db1. J. Clin. Invest. 83:1066.
10. Duclos-Vallee, J. C., O. Hajoui, A. M. Yamamoto, E. Jacz-Aigrain, and
F. Alvarez. 1995. Conformational epitopes on CYP2D6 are recognized by liver/
kidney microsomal antibodies. Gastroenterology 108:470.
11. Parez, N., D. Herzog, E. Jacqz-Aigrain, J. C. Homberg, and F. Alvarez. 1996.
Study of the B cell response to cytochrome P450IID6 in sera from chronic hepatitis C patients. Clin. Exp. Immunol. 106:336.
12. Klein, R., U. M. Zanger, T. Berg, U. Hopf, and P. A. Berg. 1999. Overlapping but
distinct specificities of anti-liver-kidney microsome antibodies in autoimmune
hepatitis type II and hepatitis C revealed by recombinant native CYP2D6 and
novel peptide epitopes. Clin. Exp. Immunol. 118:290.
13. Ma, Y., M. Peakman, A. Lobo-Yeo, L. Wen, M. Lenzi, J. Gäken, F. Farzaneh,
G. Mieli-Vergani, F. B. Bianchi, and D. Vergani. 1994. Differences in immune
recognition of cytochrome P4502D6 by liver kidney microsomal (LKM) antibody
in autoimmune hepatitis and chronic hepatitis C virus infection. Clin. Exp. Immunol. 97:94.
14. Ma, Y., G. Gregorio, J. Gäken, L. Muratori, F. B. Bianchi, G. Mieli-Vergani, and
D. Vergani. 1997. Establishment of a novel radioligand assay using eukaryotically expressed cytochrome P4502D6 for the measurement of liver kidney microsomal type 1 antibody in patients with autoimmune hepatitis and hepatitis C
virus infection. J. Hepatol. 26:1396.
15. Yamamoto, A. M., C. Johanet, J. C. Duclos-Vallee, F. A. Bustarret, F. Alvarez,
J. C. Homberg, and J. F. Bach. 1997. A new approach to cytochrome CYP2D6
antibody detection in autoimmune hepatitis type-2 (AIH-2) and chronic hepatitis
C virus (HCV) infection: a sensitive and quantitative radioligand assay. Clin.
Exp. Immunol. 108:396.
16. Yamamoto, A. M., D. Cresteil, O. Boniface, F. F. Clerc, and F. Alvarez. 1993.
Identification and analysis of cytochrome P450IID6 antigenic sites recognized by
anti-liver-kidney microsome type-1 antibodies (LKM1). Eur. J. Immunol. 23:
1105.
17. Yamamoto, A. M., D. Cresteil, J. C. Homberg, and F. Alvarez. 1993. Characterization of anti-liver-kidney microsome antibody (anti-LKM1) from hepatitis C
virus-positive and -negative sera. Gastroenterology 104:1762.
18. Riemekasten, G., J. Marell, G. Trebeljahr, R. Klein, G. Hausdorf, T. Haupl,
J. Schneider-Mergener, G. R. Burmester, and F. Hiepe. 1998. A novel epitope on
the C-terminus of SmD1 is recognized by the majority of sera from patients with
systemic lupus erythematosus. J. Clin. Invest. 102:754.
19. Im, S. H., D. Barchan, S. Fuchs, and M. C. Souroujon. 1999. Suppression of
ongoing experimental myasthenia by oral treatment with an acetylcholine receptor recombinant fragment. J. Clin. Invest. 104:1723.
20. Dalekos, G. N., H. Wedemeyer, P. Obermayer-Straub, A. Kayser, A. Barut,
H. Frank, and M. P. Manns. 1999. Epitope mapping of cytochrome P4502D6
autoantigen in patients with chronic hepatitis C during ␣-interferon treatment.
J. Hepatol. 30:366.
21. Laver, W. G., G. M. Air, R. G. Webster, and S. J. Smith-Gill. 1990. Epitopes on
protein antigens: misconceptions and realities. Cell 61:553.
22. Schwartz, H. L., J. M. Chandonia, S. F. Kash, J. Kanaani, E. Tunnell,
A. Domingo, F. E. Cohen, J. P. Banga, A. M. Madec, W. Richter, and
S. Baekkeskov. 1999. High-resolution autoreactive epitope mapping and structural modeling of the 65 kDa form of human glutamic acid decarboxylase. J. Mol.
Biol. 287:983.
23. Welin Henriksson, E., M. Wahren-Herlenius, I. Lundberg, E. Mellquist, and
I. Pettersson. 1999. Key residues revealed in a major conformational epitope of
the U1–70K protein. Proc. Natl. Acad. Sci. USA 96:14487.
The Journal of Immunology
24. Hassfeld, W., G. Steiner, A. Studnicka-Benke, K. Skriner, W. Graninger,
I. Fischer, and J. S. Smolen. 1995. Autoimmune response to the spliceosome: an
immunologic link between rheumatoid arthritis, mixed connective tissue disease,
and systemic lupus erythematosus. Arthritis Rheum. 38:777.
25. Deshmukh, U. S., J. E. Lewis, F. Gaskin, P. K. Dhakephalkar, C. C. Kannapell,
S. T. Waters, and S. M. Fu. 2000. Ro60 peptides induce antibodies to similar
epitopes shared among lupus-related autoantigens. J. Immunol. 164:6655.
26. Mackay, I. R., S. Whittingham, S. Fida, M. Myers, N. Ikuno, M. E. Gershwin,
and M. J. Rowley. 2000. The peculiar autoimmunity of primary biliary cirrhosis.
Immunol. Rev. 174:226.
27. Ravichandran, K. G., S. S. Boddupalli, C. A. Hasermann, J. A. Peterson, and
J. Deisenhofer. 1993. Crystal structure of hemoprotein domain of P450BM-3, a
prototype for microsomal P450’s. Science 261:731.
28. de Groot, M. J., M. J. Ackland, V. A. Horne, A. A. Alex, and B. C. Jones. 1999.
Novel approach to predicting P450-mediated drug metabolism: development of a
combined protein and pharmacophore model for CYP2D6. J. Med. Chem. 42:
1515.
29. Guex, N., and M. C. Peitsch. 1997. SWISS-MODEL and the Swiss-PdbViewer:
an environment for comparative protein modeling. Electrophoresis 18:2714.
30. Smith, M. G., R. Williams, G. Walker, M. Rizzetto, and D. Doniach. 1974.
Hepatic disorders associated with liver-kidney microsomal antibodies. Br.
Med. J. 2:80.
31. Alvarez, F., P. A. Berg, F. B. Bianchi, L. Bianchi, A. K. Burroughs,
E. L. Cancado, R. W. Chapman, W. G. Cooksley, A. J. Czaja, V. J. Desmet, et
al. 1999. International autoimmune hepatitis group report: review of criteria for
diagnosis of autoimmune hepatitis. J. Hepatol. 31:929.
32. Manns, M. P., K. J. Griffin, K. F. Sullivan, and E. F. Johnson. 1991. LKM-1
autoantibodies recognize a short linear sequence in P450IID6, a cytochrome
P-450 monooxygenase. J. Clin. Invest. 88:1370.
33. Lecoeur, S., C. Andre, and P. H. Beaune. 1996. Tienilic acid-induced autoimmune hepatitis: anti-liver and-kidney microsomal type 2 autoantibodies recognize
a three-site conformational epitope on cytochrome P4502C9. Mol. Pharmacol.
50:326.
285
34. Choudhuri, K., G. V. Gregorio, G. Mieli-Vergani, and D. Vergani. 1998. Immunological cross-reactivity to multiple autoantigens in patients with liver kidney
microsomal type 1 autoimmune hepatitis. Hepatology 28:1177.
35. Castano, L., E. Russo, L. Zhou, M. A. Lipes, and G. S. Eisenbarth. 1991. Identification and cloning of a granule autoantigen (carboxypeptidase-H) associated
with type I diabetes. J. Clin. Endocrinol. Metab. 73:1197.
36. Wedlock, N., T. Asawa, A. Baumann-Antczak, B. R. Smith, and J. Furmaniak.
1993. Autoimmune Addison’s disease: analysis of autoantibody binding sites on
human steroid 21-hydroxylase. FEBS Lett. 332:123.
37. Loeper, J., B. Louerat-Oriou, C. Duport, and D. Pompon. 1998. Yeast expressed
cytochrome P450 2D6 (CYP2D6) exposed on the external face of plasma membrane is functionally competent. Mol. Pharmacol. 54:8.
38. Muratori, L., M. Parola, A. Ripalti, G. Robino, P. Muratori, G. Bellomo,
R. Carini, M. Lenzi, M. P. Landini, E. Albano, and F. B. Bianchi. 2000. Liver/
kidney microsomal antibody type 1 targets CYP2D6 on hepatocyte plasma membrane. Gut 46:553.
39. Karachunski, P. I., N. S. Ostlie, D. K. Okita, and B. M. Conti-Fine. 1997. Prevention of experimental myasthenia gravis by nasal administration of synthetic
acetylcholine receptor T epitope sequences. J. Clin. Invest. 100:3027.
40. Estienne, V., C. Duthoit, L. Vinet, J. M. Durand-Gorde, P. Carayon, and J. Ruf.
1998. A conformational B-cell epitope on the C-terminal end of the extracellular
part of human thyroid peroxidase. J. Biol. Chem. 273:8056.
41. Elson, C. J., and R. N. Barker. 2000. Helper T cells in antibody-mediated, organspecific autoimmunity. Curr. Opin. Immunol. 12:664.
42. Shen, C. R., D. C. Wraith, and C. J. Elson. 1999. Splenic but not thymic autoreactive T cells from New Zealand Black mice respond to a dominant erythrocyte
band 3 peptide. Immunology 96:595.
43. Eftekhari, P., L. Salle, F. Lezoualc’h, J. Mialet, M. Gastineau, J. P. Briand,
D. A. Isenberg, G. J. Fournie, J. Argibay, R. Fischmeister, et al. 2000. Anti-SSA/
Ro52 autoantibodies blocking the cardiac 5-HT4 serotoninergic receptor could
explain neonatal lupus congenital heart block. Eur. J. Immunol. 30:2782.