This information is current as
of June 15, 2017.
Prediction of the Immunodominant Epitope
of the Pyruvate Dehydrogenase Complex E2
in Primary Biliary Cirrhosis Using Phage
Display
Merrill J. Rowley, Marita Scealy, James C. Whisstock,
Jennifer A. Jois, Lakshmi C. Wijeyewickrema and Ian R.
Mackay
References
Subscription
Permissions
Email Alerts
This article cites 22 articles, 8 of which you can access for free at:
http://www.jimmunol.org/content/164/6/3413.full#ref-list-1
Information about subscribing to The Journal of Immunology is online at:
http://jimmunol.org/subscription
Submit copyright permission requests at:
http://www.aai.org/About/Publications/JI/copyright.html
Receive free email-alerts when new articles cite this article. Sign up at:
http://jimmunol.org/alerts
The Journal of Immunology is published twice each month by
The American Association of Immunologists, Inc.,
1451 Rockville Pike, Suite 650, Rockville, MD 20852
Copyright © 2000 by The American Association of
Immunologists All rights reserved.
Print ISSN: 0022-1767 Online ISSN: 1550-6606.
Downloaded from http://www.jimmunol.org/ by guest on June 15, 2017
J Immunol 2000; 164:3413-3419; ;
doi: 10.4049/jimmunol.164.6.3413
http://www.jimmunol.org/content/164/6/3413
Prediction of the Immunodominant Epitope of the Pyruvate
Dehydrogenase Complex E2 in Primary Biliary Cirrhosis
Using Phage Display
Merrill J. Rowley,1 Marita Scealy, James C. Whisstock, Jennifer A. Jois,
Lakshmi C. Wijeyewickrema, and Ian R. Mackay
X
-ray crystallographic studies have been used to examine
the structure of epitopes for a number of different Ags
(1). In most cases, the Abs react with conformational
epitopes formed from assembled topographic determinants made
up of residues brought into contact on the surface of the molecule
during protein folding; such regions often contain T cell epitopes
as well (2). These epitopes have an area of 650 –900 Å2, and the
affinity of Ag-Ab binding depends on the interaction of just a few
contiguous but discontinuous amino acid residues on the surface of
the Ag molecule (1). Identification of these reactive residues is
fundamental to understanding the generation of the Ab response
but conventional techniques of epitope mapping rarely define conformational epitopes more precisely than to within 100 –200 aa.
Although crystallographic studies to define critical residues for Ag
binding have been performed for immune Abs, very little is known
about the interactions of spontaneously occurring autoantibodies
with cognate autoantigens. In particular, critical residues for autoantibody binding have not been defined, although this is essential
for an understanding of the nature of the autoimmune response and
for identifying cross-reactivities that could be operative in examples of molecular mimicry.
Department of Biochemistry and Molecular Biology, Monash University, Clayton,
Victoria, Australia
Received for publication September 27, 1999. Accepted for publication January
7, 2000.
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
Address correspondence and reprint requests to Dr. M. J. Rowley, Department of
Biochemistry and Molecular Biology, Monash University, Clayton, Victoria 3800,
Australia. E-mail address: [email protected]
Copyright © 2000 by The American Association of Immunologists
Primary biliary cirrhosis (PBC)2 is an autoimmune liver disease
characterized by autoantibodies (anti-M2) reactive with components of the pyruvate dehydrogenase complex (PDC) or of the
related enzyme complexes, 2-oxoacid dehydrogenase and
branched chain oxoacid dehydrogenase (3). PDC, a multienzyme
complex (Mr 8.5 ⫻ 106), is located on the inner mitochondrial
membrane and catalyzes the conversion of pyruvate to acetyl-CoA.
The complex is composed of multiple copies of three enzymes that
act in sequence, the pyruvate dehydrogenase E1, the dihydrolipoamide acetyltransferase E2, and the flavine adenine dinucleotidedependent dihydrolipoamide dehydrogenase E3 (4). Most patients
(⬎90%) with PBC have autoantibodies to PDC-E2.
The major epitope within PDC-E2 has been mapped to the region associated with the inner lipoyl domain wherein a lysine residue binds the lipoic acid cofactor for the enzyme, and there is a
minor epitope associated with the outer lipoyl domain (5, 6). Using
recombinant polypeptides, the inner lipoyl domain epitope was
mapped to aa 91–227 (5); a short synthetic peptide representing aa
167–186 that included the inner lysine residue 173 (K173) absorbed
most reactivity with PBC sera by ELISA, albeit only at a serum
dilution of 1:80,000 (5). On the other hand, studies by Koike et al.
(7) using purified porcine PDC and PDC-E2, or a 37-kDa trypsin
fragment that represented the two lipoyl domains of PDC-E2,
showed that there was a very clear loss of autoantigenicity during
fragmentation of the PDC-E2. Moreover, although a short synthetic peptide (aa165–178) central to the inner lipoyl domain
slightly inhibited the binding of PBC sera to solid-phase PDC by
ELISA, the level of inhibition with 25 ␮g of peptide was much less
than that obtained by 0.005 ␮g of PDC. As further evidence that
2
Abbreviations used in this paper: PBC, primary biliary cirrhosis; PDC, pyruvate
dehydrogenase complex; RA, rheumatoid arthritis.
0022-1767/00/$02.00
Downloaded from http://www.jimmunol.org/ by guest on June 15, 2017
Primary biliary cirrhosis (PBC) is an autoimmune liver disease characterized by autoantibodies reactive with the pyruvate
dehydrogenase complex. A conformational epitope has been mapped to aa 91–227 within the inner lipoyl domain of the E2 subunit
(pyruvate dehydrogenase complex E2 (PDC-E2)). We have used phage display to further localize this epitope. A random heptapeptide library was screened using IgG from two patients with PBC, with negative selection using pooled normal IgG. Phage that
contained peptide inserts (phagotopes) selected using PBC sera differed from those selected using IgG from patients with RA or
polychondritis. Two motifs occurred only among the PBC-selected phagotopes; these were MH (13 sequences, 16 phagotopes) and
FV (FVEHTRW, FVEIYSP, FVLPWRI). The phagotopes selected were tested for reactivity with anti-PDC-E2 affinity purified
from four patients with PBC. Phagotopes that contained 1 of 15 different peptide sequences were reactive with one or more of these
four anti-PDC-E2 preparations, whereas phagotopes that contained 1of the remaining 28 sequences were negative. The peptides
(FVLPWRI, MHLNTPP, MHLTQSP) encoded by three phagotopes that were strongly reactive with all four preparations of
anti-PDC-E2 were synthesized. Each of the selected peptides, but not an irrelevant peptide, inhibited the reactivity by ELISA of
PBC serum with recombinant PDC-E2 and reduced the inhibition of the enzyme activity of PDC by a PBC serum. The peptide
sequences, along with the known NMR structure of the inner lipoyl domain of PDC-E2, allow the prediction of nonsequential
residues 131HM132 and 178FEV180 that contribute to a conformational epitope. The Journal of Immunology, 2000, 164: 3413–3419.
3414
Materials and Methods
Sera
The sera used were from four patients with PBC (PBC 1– 4) diagnosed
according to published criteria (15). All patients had high titers of Abs to
PDC-E2 and other related Ags as determined by immunoblotting on bovine
heart mitochondria (16). One patient, PBC1, also had type 1 diabetes and
her serum contained autoantibodies to glutamic acid decarboxylase. Control sera were derived from patients with rheumatoid arthritis (RA;1 subject) and relapsing polychondritis (1 subject), selected by reason of very
high serum levels of an irrelevant autoantibody to type II collagen, and
healthy laboratory staff (10 subjects). The 10 normal sera were pooled to
provide IgG for negative selection during biopanning. IgG was prepared
from the sera by affinity purification on protein A-Sepharose (Pharmacia,
Uppsala, Sweden).
Preparation of affinity-purified Abs using recombinant PDC-E2
A cDNA fragment of human PDC-E2 that represented nt 1095–2207 corresponding to aa 109 – 470 was ligated into pGEX and expressed as a GST
fusion protein (17). This fusion protein contains only the inner lipoyl domain but reacts strongly and specifically with PBC sera (17). Abs to
PDC-E2 were prepared by affinity purification on recombinant protein (16).
Phage-displayed random peptide library
The Ph.D-7 random heptapeptide library was purchased from New England
Biolabs (Beverly, MA). The library contained 2 ⫻ 1013 PFU/ml with a
complexity of 2 ⫻ 109 transformants. The displayed heptapeptides were
expressed at the amino terminus of the pIII coat protein of the filamentous
coliphage M13, and the library was used at a concentration of 8 ⫻ 1010
PFU/ml. The library was stored in TBS (pH 7.4) with 50% glycerol. Phage
was propagated in Escherichia coli strain ER2537, which was provided
with the library kit as a noncompetent glycerol stock. The DNA inserts
from individual nonselected colonies were sequenced from the library to
confirm that they contained random peptide sequences. All phages selected
contained inserts, although two of the phages had the amber stop codon
(TAG) that is expressed as glutamine in this library. All 20 aa codons were
represented in the random sequences close to their expected frequencies
based on the genetic code.
Biopanning and isolation of phage
The library was screened by biopanning using 100 ␮g of IgG mixed overnight at 4°C with 2 ␮l of the phagemid library in 500 ␮l of PBS (pH 7.3)
containing 1 mg/ml BSA (PBS-BSA). Phages that bound to the IgG were
isolated using magnetic beads coated with anti-human Ig (Chemicon International, Temecula, CA), washed 10 times in PBS-BSA, eluted from the
magnetic beads with 1 mg/ml BSA, 0.1 M HCl (pH 2.2) adjusted with
glycine, and then neutralized with 1 M Tris-HCl (pH 9). After this first
round of “positive” selection, there was a “negative” selection step in
which phages not specifically reactive with PBC sera were removed using
100 ␮g of IgG prepared from pooled normal sera. Then the phages were
amplified by infecting E. coli EWR2537 cells in the logarithmic phase of
growth and allowing them to grow at 37°C with shaking for 4.5 h. Phages
from liquid culture were obtained by clearing the supernatant twice by
centrifugation, 13,000 ⫻ g for 5 min. Phage particles were precipitated
with 4% polyethylene glycol 6000 in 0.5 M NaCl, centrifuged at 13,000 ⫻
g for 15 min, and the pellets were resuspended in 1 ml TBS. The amplified
eluate was enriched by two further rounds of positive selection, negative
selection, and amplification. Phages were purified from single colonies
after the third round of biopanning and were used for DNA sequencing, and
tested for reactivity with affinity-purified anti-PDC-E2 by capture ELISA.
Each biopanning involved three rounds of positive selection with IgG
from one patient. In all, there were performed four separate biopannings
using IgG from two patients with PBC: one with RA and one with relapsing polychondritis.
DNA sequencing
Single-stranded DNA was prepared by phenol extraction as described by
Sambrook et al. (18) and sequenced using the Sequenase version 2.0 T7
DNA polymerase sequencing kit (United States Biochemical, Cleveland,
OH). The M13 (-28 gIII) sequencing primer supplied in the library kit was
used and ␣-35S-dATP Redivue (Amersham International, Little Chalfont,
U.K.) was used as label. Sequencing was conducted using dGTP, and ambiguous sequences were resequenced using the nucleotide analogue dITP
to minimize secondary structure formation.
Capture ELISA
Phages were tested for reactivity with Abs to PDC-E2 by a capture ELISA
in which 96-well microtiter plates (Nunc, Roskilde, Denmark) were coated
with 100 ␮l/well of affinity-purified anti-PDC-E2 in sterile PBS (pH 7.3)
overnight at 4°C in a humidified chamber. The plates were blocked for 2 h
at room temperature with 200 ␮l/well of 1% skimmed milk powder in
0.05% Tween 20 in PBS and washed three times with 0.005% Tween 20
in TBS (pH 7.4). Duplicate samples of 100 ␮l of purified phage diluted
1:20 in sterile PBS were added to the wells and held overnight at 4°C. The
plates were blocked and washed as before, and 100 ␮l of a 1:2000 dilution
of an Ab to M13 raised in sheep (Pharmacia Biotech, Uppsala, Sweden) in
1% skimmed milk powder in 0.05% Tween 20 in PBS was added to each
well. IgG was detected using HRP-conjugated anti-sheep/goat Ig (Silenus,
Hawthorn, Australia) with 0.5 mg/ml 2,2⬘-azino-bis-(3-ethylbenzathiazoline-6-sulfonic acid) (Boehringer Mannheim, Mannheim, Germany) in 0.03
M citric acid, 0.04 M Na2HPO4, and 0.003% H2O2 (pH 4) as substrate. A
result greater than the mean ⫹ 2 SDs of the A415 readings obtained for eight
different preparations of phage without inserts on the same plate was considered positive.
Peptide ELISA
Peptides representing the sequences in the three most reactive phagotopes
in the capture ELISA (FVLPWRI, MHLNTPP, and MHLTQSP) and an
irrelevant control peptide CIAPKRHNSAC were purchased from Chiron
Technologies (Melbourne, Australia). The peptides were synthesized with
an N-terminal free amine and a C-terminal free acid and were of 90 –95%
purity as assessed by HPLC and mass spectrometry. The reactivity of PBC
and control sera with the peptides was tested by ELISA. Microtiter plates
were coated with 20 ␮g/ml of peptide in 50 mM carbonate buffer (pH 9.6)
or PBS (pH 7.4) and held overnight at 4°C. Plates were washed, blocked
with 3% BSA in PBS, and reacted with doubling dilutions from 1:200 of
sera from PBC patients or controls overnight at 4°C. Bound Ab was detected using HRP-conjugated anti-human IgG (Silenus) with 0.5 mg/ml
2,2⬘-azino-bis(3-ethylbenzthiazoline-6-sulfonic acid) as substrate, as described above.
Peptide inhibition of reactivity of PBC sera with rPDC-E2
The ability of the phage peptides to inhibit the reactivity of a PBC serum
with rPDC-E2 was tested in a direct ELISA. Microtiter plates were coated
with 1 ␮g/ml of PDC-E2 in PBS (pH 7.4) overnight at 4°C. Serum from
Downloaded from http://www.jimmunol.org/ by guest on June 15, 2017
the reactive epitope is not precisely defined by a short peptide
central to the inner lipoyl domain, immunoreactivity of PDC-E2
may not require the lipoic acid cofactor, since enzymatic delipolyation or relipoylation did not change the reactivity of purified
PDC or PDC-E2 with PBC sera (7), although a peptide 167–184
was more reactive after lipoylation (8). Thus, although the potent
reactivity of PBC sera by immunoblotting on recombinant
PDC-E2 is suggestive of a sequential epitope, it is more likely that
there is an undefined conformational epitope residing within the
peptide sequence 91–227 of PDC-E2.
A new approach to the identification of nonlinear epitopes is Ab
probing of phage-displayed peptide libraries which can reveal conformational epitopes and also linear mimotopes that mimic the
shape of conformational epitopes (9). Conformational epitopes
have been identified using mAbs (10 –12) and also using polyclonal Abs in immune sera from both animals and humans (13,
14). However, there have been few definitive studies on polyclonal
autoantibodies in human sera using phage display libraries. The
very high levels of Abs to PDC-E2 in PBC, and the ready availability of recombinant PDC-E2, indicate that phage display technology should be applicable to identification of mimotopes of
PDC-E2 that are reactive with autoantibodies. We describe the
isolation of phage that contain peptide inserts (phagotopes) using
individual sera from two PBC patients and the identification of
particular phagotopes that react with affinity-purified PDC-E2
from four patients. This has enabled a prediction of the major
conformational epitope for PDC-E2 in PBC based on comparison
of the reactive sequences derived by phage library screening with
the known NMR structure of the inner lipoyl domain.
EPITOPE MAPPING IN PBC USING PHAGE DISPLAY
The Journal of Immunology
3415
Table I. Peptide sequences from phagotopes isolated by biopanning with two PBC sera (PBC1 and PBC2),
sera from patients with RA and relapsing polychondritis (RP), and isolated by biopanning with anti-human
IgG beads alone
PBC1
AGFNVPW
AMDPPYS
APKNSSV
FPKNSTV
FVEIYSP
HFHRPYH
HFHRPTQ
HFHRPPT (3)
a
AGFNVPW (2)a
DRWSSLF
DYKNSSV
FPKNSTV
FVEHTRW
FVLPWRI
GWGQLFQ
HFHRPPT
HFHRPYS
HHKYWHR
HSSHRLS
HYHRPPE (2)
IHKNSSV
KNWDTPP
KNWSTPF
MHLNTPP
MHLTQSP
MHLTTNI
MHRPNLE
MHSPAAT
MHSPLPI (2)
MHTPNSR
PRSRLTP
QSILPFP
SAPWPPT
RA
ALPPPYM
ESMPALS
GTNTAQK
KWSFTPP
IPNSTHT
LHSDARD
LNTTLKV
LPYKLTQ
MVPELTE
PGPSTTS (⫻2)
RP
ALPPPYM
KNWTTPL
KPWKHLK
KSTPQFT
PGPSTTS (⫻3)
PPSKMET
PRSRLTP
VTKWYSR
YPMSWRG
Beads
AMDPPTA
FPKNSTV
GSLLTLH
HHNYNMR
HPNTNLY
HTSTDKP
ITMGPNN
KNWDTPH
KSWFTPP
LEKNSTV
NHWTMNL
SHHHKPR
VSPTLHL
VSRLQQP
WVPLAST
YANNNHN
YPLKPAR
YTTIAPV
Numbers in parentheses are the number of times each sequence was isolated.
PBC2 at a final dilution of 1:100,000 was mixed with dilutions of peptide
inhibitors, preincubated for 1 h at room temperature, transferred to the
rPDC-E2-coated plate, and left overnight at 4°C before development as
described above. The serum dilution (1:100,000) was chosen to give an
A450 in the ELISA of ⬃0.5 in the absence of inhibitor. The peptides were
tested at 1, 2, 5, and 25 ␮g/ml, and rPDC-E2 was used as a positive control
at 10 ␮g/ml. Each dilution was tested in quadruplicate and compared with
the results of 16 wells that contained no inhibitor.
The effect of peptides on the ability of PBC sera to inhibit
enzyme activity of PDC
The enzyme inhibitory activity of PBC sera was measured in a microtiter
plate assay as described previously (19). To test the ability of the peptides
to interfere with the strong enzyme inhibitory function of anti-PDC-E2, 2
␮l of a highly inhibitory serum diluted 1:20 in PBS was preincubated for
1 h at room temperature in a total volume of 60 ␮l with 4 ␮g of rPDC-E2,
50 ␮g of peptide, or the equivalent volume of peptide diluent. The enzyme
reaction was initiated by the addition of 50 ␮l of enzyme mixture containing PDC from porcine heart (Sigma, St. Louis, MO) in 50 mM phosphate
buffer (pH 8.0) containing 2% BSA and 10 mM DTT, and 100 ␮l of a
reaction mixture containing the essential cofactors for the reaction, cocarboxylase, CoA, and the NAD analogue thio-NAD. Absorbance was read
immediately at 405 nm and then at minute intervals for 10 min (19). The
rate of change of absorbance was calculated, and the percent inhibition was
obtained using the standard reaction rate in wells to which PBS (pH 7.4)
was added instead of serum. The serum PBC2 was tested at a dilution of
1:20, which represented a final dilution in the assay of 1:4000; at this dilution,
the serum totally inhibited the activity of the enzyme. The counterinhibitory
effect of each peptide and each diluent was tested in quadruplicate.
Results
Phagotopes isolated by biopanning with polyclonal sera
The peptide inserts in phagotopes isolated after three rounds of
biopanning using IgG from each of the two PBC sera were sequenced (Table I). Sequences were obtained for 27 phagotopes
derived from PBC1 serum and 28 derived from PBC2 serum. In
each case, six phagotopes did not have DNA inserts. The sequences obtained were compared with those of inserts in
phagotopes obtained by biopanning with serum from the patient
with RA (11 sequences) and the patient with relapsing polychondritis (11 sequences), and from the biopanning itself in the absence
of IgG, using magnetic beads coated with anti-human IgG (18
sequences) and unselected sequences from the library (20 sequences) (Table I).
The sequences derived with the PBC sera showed distinct similarities
and most differed clearly from sequences derived from all other sources.
This was reflected by the differing frequency of each amino acid in the
isolated peptides, when a comparison was made between the frequencies
of particular amino acid in peptides from nonselected phagotopes, peptides derived from PBC sera, peptides from the control sera, and peptides
derived after biopanning with beads only. Using beads only, tryptophan
(W), asparagine (N), and tyrosine (Y) were each increased 5–6-fold and
histidine (H) and lysine (K) more than 2-fold. Although the frequency of
these particular amino acids tended to be increased in peptides selected
when Abs were used for biopanning, there were, in addition, selective
increases in frequency of particular amino acids according to the source
of the Ab. For PBC sera, several motifs were distinguishable. Of the 43
different sequences isolated, 8 sequences that were derived from 11
phagotopes (isolated from biopannings with the two PBC sera) contained
the motif H-aromatic-HRPxx, and 13 sequences derived from 16
phagotopes began with MH. There were two sequences, including one
isolated with IgG from each PBC serum, that began with FVE
(FVEHTRW, FVEIYSP) and one other sequence was FVLPWRI. Of
another two sequences (AGFNVPW, HHKYWHR) that were isolated
using each PBC serum, HHKYWHR was considered to be a “nuisance”
peptide since it resembled a sequence HHKYWxR previously identified
from biopanning with beads coated with anti-mouse Ig but no selecting
Ab (20).
Sequences from phagotopes positive by capture ELISA
All phagotopes derived with IgG from either PBC1 or PBC2 were
tested by capture ELISA for reactivity with preparations of affinity-purified anti-PDC-E2 from four different subjects with PBC.
Downloaded from http://www.jimmunol.org/ by guest on June 15, 2017
HHKYWHR (2)
HWHRPSE
HWHRPTN
IISGKDS
KQDPPSL
LGRNSTV
MHETTKA
MHIPTPD
MHLNTAS
MHNPSMT (3)
MHSPSTQ
MHTPAIG
SLTGPRE
WHRPNIH
PBC2
3416
EPITOPE MAPPING IN PBC USING PHAGE DISPLAY
Table II. Reactivity of phagotopes by ELISA after capture with affinity-purified Abs to PDC-E2 from four different sera
OD
Sequence
AGFNVPW
FVEHTRW
FVEIYSP
FVLPWRI
GWGQLFQ
HYHRPPE
KNWDTPP
MHLNTPP
MHLTQSP
MHLTTNI
MHNPSMT
MHSPAAT
MHSPLPI
MHTPAIG
SAPWPPT
No insert
Isolated by
PBC1
PBC2
PBC3
PBC4
PBC1, PBC 2
PBC2
PBC1
PBC2
PBC2
PBC2
PBC2
PBC2
PBC2
PBC2
PBC1
PBC2
PBC2
PBC1
PBC2
0.819
0.765
0.805
0.846
0.905
00.695
0.735
0.960
1.034
1.189
0.827
0.740
0.776
0.673
0.754
0.605 ⫾ 0.046a
0.782
0.671
0.634
1.141
0.724
0.658
0.502
1.465
1.599
0.540
0.780
0.696
0.839
0.694
0.608
0.373 ⫾ 0.139
0.323
0.708
0.607
1.232
0.776
0.720
0.393
1.326
1.457
NTc
NT
0.596
0.598
NT
0.608
0.367 ⫾ 0.137
0.873
0.818
0.740
1.500
0.950
0.911
0.569
1.737
1.942
0.587
0.949
0.832
0.979
0.683
0.725
0.447 ⫾ 0.206
Mean ⫾ SD for eight different samples tested in duplicate.
The mean OD in the capture ELISA is shown in bold for all values ⬎2 SD above the mean of the controls. All other phagotopes were negative.
c
NT, not tested.
a
Phagotopes that contained 1of 15 different peptide sequences were
reactive with 1or more of the four affinity-purified preparations of
anti-PDC-E2, whereas phagotopes that contained 28 different sequences were negative with all 4 Ab preparations (Table II). Of the
phagotopes that were positive, most reacted with more than one of
the preparations of anti-PDC-E2, and four phagotopes with sequences FVLPWRI, MHLNTPP, MHLTQSP, and GWGQLFP
were reactive with all four of the antisera. The reactive phagotopes
included 7of 13 in which the peptide included the motif MH, all of
the three with the motif FV, and that containing the peptide
AGFNVPW that was isolated from both biopannings.
Reactivity of synthetic peptides with PBC sera
PDC-E2 or the peptides MHLTQSP, MHLNTPP, and FVLPWRI
were 0.003 ⫾ 0.00013, 0.0026 ⫾ 0.00010, 0.0021 ⫾ 0.00013, and
0.0014 ⫾ 0.0002, respectively, compared with 0.0005 ⫾ 0.00014
with the control peptide, which represented percent inhibition of
73, 80, 84, 89, and 96%. The addition of rPDC-E2 led to the
greatest restoration of reactivity, although the molar concentration
of the rPDC-E2 was ⬃1000 times lower than that of any of the
peptides.
Discussion
In the present study, we have used Ab screening of a phage-displayed library of random heptapeptides expressed in the pIII coat
The peptides FVLPWRI, MHLNTPP, and MHLTQSP that were
displayed on the three phagotopes that were strongly reactive with
all preparations of anti-PDC-E2 (OD ⬎ 1) and an irrelevant control peptide CIAPKRHNSAC were synthesized and tested for reactivity with PBC sera. None of the four peptides reacted with any
of the PBC or control sera by direct ELISA under the conditions
tested. However, each of the three selected peptides, but not the
control peptide, inhibited the reactivity of the PBC sera with
rPDC-E2 by ELISA. The most inhibitory peptide was FVLPWRI,
which gave complete inhibition (100 ⫾ 2%, mean ⫾ SD) at 25
␮g/ml, as did 10 ␮g/ml of rPDC-E2 as the positive control (99 ⫾
2%), whereas the negative control peptide CIAPKRHNSAC gave
5 ⫾ 3% inhibition at the highest concentration. The two similar
peptides, MHLTQSP and MHLNTPP, were also inhibitory at 25
␮g/ml, although to a lesser extent, 70 ⫾ 3% and 30 ⫾ 11%.
Effect of peptides on enzyme inhibition
Fig. 1 shows the PDC enzyme activity under varying conditions,
based on changes with time in readings at A405 representing production of thio-NADH. The reaction rate in the absence of any Ab
(inset) was 0.013 ⫾ 0.0006 (mean ⫾ SD) per min and 0.00016 ⫾
0.00046 in the presence of serum from PBC2, which represented
99% inhibition of the enzyme activity. Addition of an irrelevant
peptide had no significant effect on the rate of reaction, whereas
the addition of rPDC-E2 and each of the phage-selected peptides
MHLTQSP, MHLNTPP, and FVLPWRI was counterinhibitory in
significantly restoring the rate of reaction over the 10-min period
of observation. Thus, the rates of reaction after addition of
FIGURE 1. Effect of peptides on the inhibition of PDC enzyme activity
by PBC serum. The results show changes with time in absorbance representing the production of thio-NADH in the presence of Ab alone (G), or
Ab preincubated with 4 ␮g of r-PDC-E2 (B), or 50 ␮g of the peptide
MHLTQSP (C), peptide MHLNTPP (D), peptide FVLPWRI (E), or control
peptide CIAPKRHNSAC (F). Results are the means ⫾ SD of four replicates. The natural uninhibited enzyme activity in the absence of Ab (A,
inset, 0.13 absorbance) was totally (99%) inhibited by Ab (anti-PDC-E2)
whereas r-PDC-E2 or peptide MHLTQSP, MHLNTPP, or FVLPWRI was
significantly counterinhibitory; the control peptide had no effect.
Downloaded from http://www.jimmunol.org/ by guest on June 15, 2017
b
The Journal of Immunology
3417
protein of bacteriophage to derive peptide mimotopes of PDC-E2.
Our strategy was to use polyclonal IgG from patients with PBC in
two separate sequences of three rounds of positive selection using
an individual IgG and negative selection using IgG from pooled
normal sera. This allowed the development of panels of “individual specific” phagotopes containing insert sequences that could be
compared with sequences obtained using sera from other diseases.
We tested the phagotopes for reactivity with four preparations of
affinity-purified anti-PDC-E2 from PBC sera and so identified a set
of reactive phagotopes with sequences that represented mimotopes
of the immunodominant conformational epitope of PDC-E2. Since
similar sequences were derived by biopanning with IgG from two
different patients with PBC, and the derived phagotopes reacted
unequivocally with affinity-purified anti-PDC-E2 from four patients with PBC, we suggest that there is a single major conformational and discontinuous epitope within PDC-E2 that is recognized by most PBC sera.
Not all of the phagotopes derived by biopanning with PBC sera
were reactive with anti-PDC-E2. This is understandable, given that
polyclonal IgG includes a myriad of Abs so that, even after extensive rounds of negative selection, the phagotopes isolated will
include many that react with other Abs that are either specific or
nonspecific to the disease. The difficulty in discriminating between
sequences that are specific or nonspecific for any given disease (or
for Abs that accompany that disease) has greatly hampered the use
of phage-displayed technology with polyclonal sera. In the present
study, disease-specific phagotopes were identified using a capture
ELISA in which plates were coated with affinity-purified antiPDC-E2. However, for other autoantibodies, affinity-purified Abs
may not be readily available. Such limitations have been offset by
our utilization of an alignment algorithm PILEUP (21) to group
peptide sequences into clusters according to their physicochemical
relatedness (20). Using PILEUP, it was found that PBC-derived
phagotopes that contained sequences that were reactive by ELISA
with affinity-purified anti-PDC-E2 aggregated into disease-specific
clusters, whereas other disease-specific clusters were unreactive
(20). It is likely that some of the PBC-specific sequences derived
from the phage library will represent mimotopes for autoantibodies
other than anti-PDC-E2, particularly mimotopes of conformational
epitopes of the related molecules 2-oxoglutarate dehydrogenase
complex-E2 and branched chain oxoacid dehydrogenase-E2, or
various PBC-associated nuclear autoantigens. In addition, some of
the nondisease-specific phagotopes sequences that we derived
would include either nuisance peptides intrinsic to the biopanning
Downloaded from http://www.jimmunol.org/ by guest on June 15, 2017
FIGURE 2. a, A side view of the inner lipoyl domain of PDC-E2 and the
proposed contact residues for the conformational epitope of PDC-E2 based on
the published NMR structure (22). The
residues H132M133 juxtaposed with F178
and V180 are shown in red, with amino
acids surface exposed in the same plane
in the vicinity of the single histidine
(H132) shown as space-filled models.
The region represented by the linear
peptide 167–183 surrounding the lipoyl
lysine K173 is shown in yellow. b, Alignment of the inner and outer lipoyl domains of human PDC-E2 performed using Clustal W multiple sequence
alignment (25). Numbers shown refer to
the amino acid sequence of the inner lipoyl domain. Colors are as shown in the
structure in a. ⴱ, K173.
3418
least two separate regions of the linear sequence, H132 and M133,
and F178 and V180, and that each of these residues are critical for
Ab binding. Moreover, this is the region at which enzyme inhibitory Abs bind, since each of the peptides tested markedly reduced
the enzyme inhibitory effect of a PBC serum.
Phage display has not previously been used to identify the PDC-E2
epitope using polyclonal Abs; a random dodecapeptide phage-displayed library was screened with a mouse mAb C355.1 to the inner
lipoyl domain of PDC-E2 (23). In that study, among 36 phagotopes
selected, 3 common amino acid motifs were identified: SYP, TYVS,
and VRH. It is notable that these amino acids occur within our putative epitope and that the actual sequence SYP occurs in the same
epitope region (SYPPHM) of the inner but not the outer lipoyl domain
of PDC-E2. The mAb C355.1 reacted not only with PDC-E2, but also
with the luminal surface of bile duct epithelium only from patients
with PDC, and not from normal subjects or patients with other liver
diseases (23). This reactivity was similar to that of human combinatorial Abs specific to the inner lipoyl domain of PDC-E2, and the
peptides identified by phage display using C355.1 inhibited immunohistological staining of human combinatorial Abs to PBC biliary epithelial cells (24). The colocalization of the epitope isolated with
C355.1 and the epitope defined using polyclonal anti-PDC-E2 in the
present study suggest that this epitope is relevant to the pathogenesis
of PBC.
In conclusion, we have, for the first time, amalgamated findings
from Ab screening of a phage library with the known structure of
an immunodominant region of an autoantigenic molecule to locate
precisely a conformational epitope, in this case the epitope for
anti-PDC-E2, which is the major autoantibody in PBC. These findings should give new impetus to the search for mimicking sequences among Ags of microbial provenance that could be implicated in the pathogenesis of PBC and could stimulate similar
studies with other well-defined autoantigen-Ab systems.
Acknowledgments
We thank Dr. Janet Davies and Dr. Mark Myers for critically reviewing the
manuscript and helpful discussions.
References
1. 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.
2. Hannum, C. H., M. Utlee, L. A. Matis, R. H. Schwartz, and E. Margoliash. 1982.
The major B and T cell determinant on pigeon cytochrome c in B10.a mice. Adv.
Exp. Med. Biol. 150:37.
3. Gershwin, M. E., and I. R. Mackay. 1991. Primary biliary cirrhosis: paradigm or
paradox for autoimmunity. Gastroenterology 100:822.
4. Yeaman, S. J. 1989. The 2-oxo acid dehydrogenase complexes: recent advances.
Biochem. J. 257:625.
5. Van de Water, J., M. E. Gershwin, P. Leung, A. Ansari, and R. L. Coppel. 1988.
The autoepitope of the 74 Kd mitochondrial autoantigen of primary biliary cirrhosis corresponds to the functional suite of dihydrolipoamide acetyltransferase.
J. Exp. Med. 167:1791.
6. Fussey, S. M. P., S. T. Ali, J. R. Guest, O. F. W. James, M. F. Bassendine, and
S. J. Yeaman. 1990. Reactivity of primary biliary cirrhosis sera with Escherichia
coli dihydrolipoamide acetyltransferase (E2p): characterization of the main immunogenic region. Proc. Natl. Acad. Sci. USA 87:3987.
7. Koike, K., H. Ishibashi, and M. Koike. 1998. Immunoreactivity of porcine heart
dihydrolipoamide acetyl- and succinyl-transferases (PDC-E2, OGDC-E2) with
primary biliary cirrhosis sera: characterization of the autoantigenic region and
effects of enzymatic delipoylation and relipoylation. Hepatology 27:1467.
8. Tuaillon, N., C. Andre, J.-P. Briand, E. Penner, and S. Muller. 1992. A lipoyl
synthetic octadecapeptide of dihydrolipoamide acetyltransferase specifically recognized by anti-M2 antibodies in primary biliary cirrhosis. J. Immunol. 148:445.
9. Scott, J. K., and G. P. Smith. 1990. Searching for peptide ligands with an epitope
library. Science 249:386.
10. Luzzago, A., F. Felici, A. Tramontano, A. Pessi, and R. Cortese. 1993. Mimicking of discontinuous epitopes by phage-displayed peptides. I. Epitope mapping of
human H ferritin using a phage library of constrained peptides. Gene 128:51.
11. Felici, F., A. Luzzago, A. Folgori, and R. Cortese. 1993. Mimicking of discontinuous epitopes by phage-displayed peptides. II. Selection of clones recognized
by a protective monoclonal antibody against the Bordella pertussis toxin from
phage peptide libraries. Gene 128:21.
Downloaded from http://www.jimmunol.org/ by guest on June 15, 2017
process itself or mimotopes of bacterial or viral Ags for which Abs
were absent from the particular pool of IgG used for negative
selection. Examples include HHKYWHR that may react with
beads used for biopanning in the absence of any selecting Ab, and
sequences with the motif H-aromatic-HRPxx, as judged by the
study of Davies et al. (20); in that study, there was a branch of the
guide tree that contained a cluster of sequences with the motif
HFHRPxx derived from sera of patients with type I diabetes as
well as PBC sera.
Our interpretation of the epitope structure for anti-PDC-E2 has
been facilitated by the derivation of an NMR structure for aa 131–233
of human PDC-E2 that encompasses the major immunodominant
epitope(s) (22) (Fig. 2a). Hitherto epitope mapping with PBC sera has
linked reactivity to the highly conserved amino acids surrounding the
lipoyl-lysine K173, particularly linear peptides 167–183 (AEIETD
KATIGFEVQEGYL) (5, 8) or 165–178 (LLAEIETDKATIGF) (7).
However, the lipoyl-lysine K173 may not contribute to the epitope as
judged by the marked scarcity of lysines among the selected
phagotopes. Moreover, lipoylation or delipoylation of PDC-E2 does
not affect its antigenicity (7), although this may not be the case for Abs
reactive with PDC-E2 from E. coli (6). Finally, although the amino
acids of the previously identified “linear” epitope (5, 7, 8) are predominantly surface exposed, most are actually not on the face of the
molecule that contains the lipoyl-lysine K173 as judged by the structural model (Fig. 2a). Accordingly there are good grounds for redefining the major immunodominant region of PDC-E2.
We note that there are two particular paired residues, FV and
MH, that featured among the phagotope sequences with frequent
reactivity with affinity-purified anti-PDC-E2, and that the inner
lipoyl domain contains only one histidine within the linear sequence P131 HMQVLL137. The NMR structure of the lipoyl domain shows that both this single histidine and the adjacent methionine are surface exposed and, moreover, these are in juxtaposition
to phenyalanine (F178) and valine (V180) which are also surface
exposed. These four amino acids are too far apart in the structure
to form part of an epitope that includes the lipoyl-lysine (K173)
(Fig. 2a). However, F178 and V180 are part of the previously described linear epitope comprising aa 167–183 (5). The amino acids
surrounding the region of juxtaposition of M133, H132, F178, and
V180 within the ⬃20 Å diameter of a putative Ab epitope include
proline (P), leucine (L), glutamine (Q), glutamate (E), tyrosine (Y),
and serine (S), each of which was well represented among the
reactive phagotopes. Accordingly, this region could readily be proposed as the major immunodominant epitope, although it excludes
the lipoyl-lysine K173. As further evidence in support of this proposed epitope structure, we can cite sequence differences between
the inner and outer lipoyl domains, noting the infrequent reactivity
of the outer lipoyl domain with PBC sera. Thus, the PHMQ of the
inner lipoyl domain is represented by PHQK in the outer domain,
and the sequence GFEV in the inner domain is represented by
GFES in the outer domain (Fig. 2b). The characteristically potent
inhibition of PDC enzyme activity by Abs to the PDC-E2 epitope
would be explained according to our epitope structure by steric
interference with the swinging movement of the lipoyl domain,
since the epitope is close to the flexible regions that link lipoyl
domains and the E3-binding domain.
Alternatively, given that phagotopes were selected using polyclonal IgG, the peptides selected could represent more than one
epitope rather than regions of the immunodominant epitope for
anti-PDC-E2. However, the observation that the peptide
MHLTQSP was 70% inhibitory for the reactivity of a PBC serum
with rPDC-E2 by ELISA, and that the peptide FVLPWRI was
totally inhibitory for the reaction with the same serum suggest that
the epitope is a complex topographic determinant formed from at
EPITOPE MAPPING IN PBC USING PHAGE DISPLAY
The Journal of Immunology
12. Cook, A. D., J. M. Davies, M. A. Myers, I. R. Mackay, and M. J. Rowley. 1998.
Mimotopes identified by phage display for the monoclonal antibody CII-C1 to
type II collagen. J. Autoimmun. 11:205.
13. Osman A. A., H. Uhlig, B. Thamm, J. Schneider-Mergener, and T. Mothes. 1998.
Use of the phage display technique for detection of epitopes recognized by polyclonal rabbit gliadin antibodies. FEBS Lett. 433:103.
14. Folgori A, R. Tafi, A. Meola, F. Felici, G. Galfré, R. Cortese, P. Monaci, and
A. Nicosia. 1994. A general strategy to identify mimotopes of pathological antigens using only random peptide libraries and human sera. EMBO J. 13:2236.
15. Kaplan, M. M. 1996. Primary biliary cirrhosis. N. Engl J. Med. 316:521.
16. Rowley, M. J., T. Maeda, I. R. Mackay, B. E. Loveland, G. L. McMullen,
G. Tribbick, and C. C. A. Bernard. 1992. Differing epitope selection of experimentally induced and natural antibodies to a disease-specific autoantigen, the E2
subunit of pyruvate dehydrogenase complex (PDC-E2). Int. Immunol. 4:1245.
17. Coppel, R. L., L. J. McNeilage, C. D. Suhr, J. Van de Water, T. W. Spithill,
S. Whittingham, and M. E. Gershwin. 1988. Primary structure of the human M2
mitochondrial autoantigen of primary biliary cirrhosis: dihydrolipoamide acetyltransferase. Proc. Natl. Acad. Sci. USA 85:7317.
18. Sambrook, J., E. F. Fritsch, and T. Maniatis. 1989. Molecular Cloning: A Laboratory Manual, 2nd Ed. Cold Spring Harbor Lab. Press, New York.
19. Teoh, K.-L., M. J. Rowley, H. Zafirakis, E. R. Dickson, R. H. Wiesner,
M. E. Gershwin, and I. R. Mackay. 1984. Enzyme inhibitory autoantibodies to
pyruvate dehydrogenase complex in primary biliary cirrhosis: applications of a
semiautomated assay. Hepatology 20:1220.
3419
20. Davies, J. M., M. Scealy, Y.-P. Cai, J. Whisstock, I. R. Mackay, and
M. J. Rowley. 1999. Multiple alignment and sorting of peptides derived from
phage displayed random peptide libraries with polyclonal sera allows discrimination of relevant phagotopes. Mol. Immunol. 36:659.
21. Sneath, P. H. A., and R. R. Sokal. 1973. Unweighted pair-group method using
arithmetic averages. In Numerical Taxonomy: The Principles and Practice of
Numerical Classification. P. H. A. Sneath, ed. Freeman, New York, p. 230.
22. Howard, M. J., C. Fuller, R. W. Broadhurst, R. N. Perham, J. G. Tang, J. Quinn,
A. G. Diamond, and S. J. Yeaman. 1998. Three-dimensional structure of the
major autoantigen in primary biliary cirrhosis. Gastroenterology 115:139.
23. Cha, S., P. S. C. Leung, J. Van de Water, K. Tsuneyama, R. E. Joplin,
A. A. Ansari, Y. Nakanuma, P. L. Schatz, S. Cwirla, L. E. Fabris, et al. 1996.
Random phage mimotopes recognized by monoclonal antibodies against the
pyruvate dehydrogenase complex-E2 (PDC-E2) Proc. Natl. Acad. Sci. USA 93:
10949.
24. Leung, P. S. C., S. Cha, R. E. Joplin, C. Galperin, J. Van de Water, A. Ansari,
R. L. Coppel, P. J. Schatz, S. Cwirla, L. E. Fabris, et al. 1996. Inhibition of
PDC-E2 human combinatorial autoantibodies by peptide mimotopes.
J. Autoimmun. 9:785.
25. Thompson, J. D., D. G. Higgins, and T. J. Gibson. 1994. Clustal W: improving
the sensitivity of progressive multiple sequence alignment through sequence
weighting, positions—specific gap penalties and weight matrix choice. Nucleic
Acids Res. 22:4673. http://www.infobiogen.Fr/docs/ClustalW/clustalw.html
Downloaded from http://www.jimmunol.org/ by guest on June 15, 2017