| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Departments of
*
Pathology and
Biophysics and Biophysical Chemistry, School of Medicine, and
Department of Molecular Microbiology and Immunology, School of Hygiene and Public Health, Johns Hopkins University, Baltimore, MD 21205
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
To localize the disease-associated (DA) epitopes of Tg, we partially degraded Tg with trypsin. The tryptic peptides were tested by Western immunoblot for their immunoreactivity with different mAbs (21). Each mAb reacted with peptides of different m.w. Some of the mAbs reacted with sequential epitopes and some reacted with conformational epitopes; after treatment with a reducing agent, the latter epitopes did not bind their corresponding mAbs. One of the mAbs (137C1) recognized a conformational epitope that was also recognized by the sera of many patients with thyroid disease but not by control normal sera (22). This mAb reacted with two peptides of Tg with molecular masses of 15 kDa and 23 kDa.
The goal of this investigation was to determine the location, the amino acid sequence, and the secondary conformation of the two peptides reactive with mAb 137C1 and sera from the patients with autoimmune thyroid disease. In a previous paper, we showed that one of the peptides (15 kDa) reactive with this mAb was located at the carboxyl-terminal end of the Tg molecule (23). In the present study, we isolated the 23-kDa peptide and subjected it to amino acid sequencing to determine its localization on the Tg molecule. The two peptides were also used in competitive inhibition assays to find out whether both peptides bear the same epitope. The probable secondary structure of each peptide was estimated by computational methods and the DA epitope was provisionally located with respect to these predictions.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Protein assay
Protein determinations were performed by the bicinchoninic acid (BCA) protein assay using BSA as the standard. A Shimadzu UV-VIS recording spectrophotometer UV-160 (Shimadzu, Kyoto, Japan) was used for spectrophotometric assays.
Murine mAbs to Tg
Preparation and characterization of the mAbs were described by Bresler et al. (19).
Human sera
Sera evaluated for reactivity to fragments of Tg were a subset
of a population previously tested by competitive inhibition ELISA using
mAb 137C1 and other mAbs (24). Sera were from individuals
with Hashimoto’s disease, Graves’ disease, thyroid carcinoma, or from
normal individuals. These sera inhibited the binding of mAb 137C1 to Tg
(total inhibition, partial inhibition, or no inhibition) as shown in
Fig. 5
.
|
Human thyroids were obtained at autopsy from normal individuals. The method for purification of Tg has been presented in a previous paper (25). The purified Tg was partially digested by incubation with trypsin for 4 h at 37°C at a ratio of 1:100 (w/w) of trypsin to Tg and used for isolation of different peptides by HPLC.
HPLC
The 4-h tryptic peptides of Tg were analyzed by HPLC using a
PolySULFOETHYL Aspartamide column (200 x 4.6 mm), from Poly LC
(Columbia, MD). The column was equilibrated with 10 mM of phosphate
buffer, pH 3 (buffer A), at a flow rate of 1 ml/min. Sample, in buffer
A + 15% acetonitrile, was applied to the column. After application of
the sample, a linear gradient from 0 to 100% of buffer B (buffer A +
0.4 M NaCl) was applied for 100 min. The column then was washed with
buffer B for 10 min and equilibrated with buffer A for 10 min.
One-milliliter fractions were collected. This column resolved the
tryptic peptides of Tg into at least seven peptide peaks (Fig. 1A). The low m.w. peptides
reactive with mAb 137C1 were present in peaks 4 and 5 as detected by
Western immunoblot using mAb 137C1 (Fig. 1B). The fractions
corresponding to each peptide peak were pooled, concentrated, and
dialyzed to remove excess salt, and used for preparative SDS-PAGE to
isolate each peptide.
|
The method for gel electrophoresis was described in detail previously (26). Briefly, SDS-PAGE was performed on a 5–20% gradient gel according to the method of Laemmli (27). To obtain nonreducing conditions, the reducing agent was eliminated from the loading buffer of the gel. The m.w. of peptides under nonreducing conditions are referred to as the apparent m.w. (MPap) because accurate m.w. determinations cannot be made under these conditions. Preparative SDS-PAGE were done using a 3-mm gel. The gel section, containing each peptide (compared with the lane containing the m.w. standard) was cut from the gel and the protein in the gel slice was eluted with 50 mM of phosphate buffer (pH 7.0) at 4°C for 18 h, with continuous mixing by a Labquake Rotator (Barnstead/Thermolyne, Dubuque, IA). Then the eluted protein was dialyzed against 50 mM of phosphate buffer (pH 7.0) at 4°C. The purity and the immunoreactivity of the eluted peptide was checked by SDS-PAGE and Western immunoblotting using mAb 137C1. The 23-kDa peptide was used for amino acid sequencing, and both peptides (15 kDa and 23 kDa) were used for ELISA assays. These peptides were previously reported as 15- and 20-kDa peptides (21). Using better discriminatory m.w. markers, we determined that the larger peptide has a molecular mass of 23 kDa.
ELISA
ELISA was done according to Bresler et al. (19) with these minor modifications. ELISA plates were coated with 1 µg/ml of intact Tg, or with 2.5 µg/ml of the 15-kDa peptide or the 23-kDa peptide. Then, 50 µl of diluted mAb or 50 µl of human serum (1:100) was added to the ELISA plates and incubated at room temperature for 1 h. Alkaline phosphatase-coupled anti-murine Ab or anti-human secondary Ab (Jackson ImmunoResearch, West Grove, PA) was added to the ELISA plates and incubated at room temperature for 1 h. Then the plates were washed and 100 µl of 1 mg/ml of p-nitrophenol phosphate in 10% diethanolamine (pH 9.8) was added to each well of the ELISA plates and incubated at room temperate for 1 h. At the end of the incubation time, the phosphatase reaction was stopped by the addition of 50 µl of 3 M sodium hydroxide. Then the plates were read at 405 nm in a Dynatech Microplate reader (Dynatech Laboratories, Chantilly, VA).
Amino acid sequencing
The amino acid sequence of each peptide was determined using an Applied Biosystems protein sequencer (model 492; Foster City, CA) in the protein sequencing facility of the Department of Biological Chemistry of the Johns Hopkins University. By this method, a standard Edman degradation cycle was used and the phenylthiohydantoin (PTH) amino acids were analyzed by a 140 C microgradient system and detected by a 785 A programmable absorbance PTH analyzer (Applied Biosystems).
Competitive ELISA of mAb 137C1 treated with either peptide or intact Tg
Plates were coated by addition of 50 µl of 2.5 µg/ml of tryptic peptide of Tg or 1 µg/ml of intact Tg in carbonate bicarbonate buffer (pH 9.8) to each well and incubated at 4°C for 14 h. On the next day, after being emptied, the plates were washed with PBS (pH 7.4) containing 0.05% Tween 20 (PBS-T) and the wells were blocked by adding 200 µl of 1% BSA in PBS-T to each well. Then the plates were incubated for 1 h at room temperature, washed with PBS-T, and used for competition assays. mAb 137C1 was diluted 1:2000 with PBS-T containing 1% BSA and incubated with different concentrations (0.5, 1, 1.5, 2, and 2.5 µg/ml) of either the 15-kDa peptide, the 23-kDa peptide, or intact Tg at room temperature for 4 h with continuous mixing by a Labquake Rotator. At the end of incubation, 50 µl of either peptide- or Tg-treated mAb was added to each well of ELISA plates coated with either the 15-kDa peptide, the 23-kDa peptide, or intact Tg. Plates were incubated at room temperature for 1 h. At the end of the incubation period, the plates were washed with PBS-T, and the remaining steps of the ELISA assay were done by the method described in the above section.
Prediction of secondary structure of 15- and 23-kDa peptides
The secondary structures of the 15- and 23-kDa peptides were predicted by using the program PHD (profile fed neural network systems from HeiDelberg (28, 29)). Secondary structural estimates are categorized as helix, strand, or loop; based on a cross-validation with 250 known proteins with predictions reported for each residue (0–9) and for the probability of assigning helix, strand, and loop (0 = low and 100 = high for each category).
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Most of the mAbs produced against Tg, including mAb 137C1, reacted with multiple tryptic peptides of Tg with different m.w. (21). mAb 137C1 reacted with two small peptides with molecular masses of 15 kDa and 23 kDa. In a previous paper, we showed that one of the peptides (15 kDa) was located at the carboxyl-terminal end of Tg, starting with amino acid 2657 (23). Our first goal was to isolate the 23-kDa peptide by HPLC followed by preparative SDS-PAGE.
To establish the purity of the peptides isolated by preparative
SDS-PAGE, the eluted peptides were analyzed by gel electrophoresis.
Then the gel was either developed with a protein stain (silver nitrate)
or tested by Western immunoblot for the immunoreactivity of the
peptides with mAb 137C1. Fig. 2A shows the gel pattern of
the two peptides when they were stained by silver nitrate, and Fig. 2B shows the Western immunoblot pattern. These results
showed that although mAb 137C1 reacted with both of the bands, the
silver nitrate stain of eluted proteins showed that one peptide was not
contaminated with the other.
|
The 23-kDa peptide isolated by SDS-PAGE was subjected to amino
acid sequencing. Fig. 3 shows that this
peptide is located further from the carboxyl-terminal end of the Tg
molecule than the 15-kDa peptide, starting with amino acid 2089. The
23-kDa peptide did not overlap with the 15-kDa peptide, and the two
peptides showed no homologous sequences. Thus, these two peptides are
located in different parts of the Tg molecule, and they do not share a
sequential epitope.
|
To study the specificity of the epitopes presented by each
peptide, we next tested the immunoreactivity of the 15-kDa peptide by
ELISA, using mAb 137C1 and other mAbs produced against Tg
(19). As shown in Fig. 4A, this peptide reacted
strongly with mAb 137C1. Minor immunoreactivity was also seen with mAbs
42C3 and 154C6. These findings accord with our previous observation
that these three mAbs bind neighboring or overlapping determinants
(19).
|
B). Immunoreactivity of the 15- and 23-kDa peptides with sera of patients with autoimmune thyroid disease
Because mAb 137C1 reacted with the same peptides as the sera from
patients with thyroid diseases (23), we tested the
immunoreactivity of the peptides reactive with mAb 137C1 with selected
sera from patients with thyroiditis, Graves’ disease, and thyroid
carcinoma, as well as with those from normal individuals. As shown in
Fig. 5A, the 15-kDa peptide
reacted strongly with sera from three patients with Hashimoto’s
thyroiditis, weakly with sera of patients with Graves’ disease, and
not at all with sera from patients with thyroid carcinoma. In a
previous investigation, the three sera from thyroiditis patients
strongly inhibited the binding of mAb 137C1 to Tg in a competitive
ELISA (24). Two patient’s sera failed to recognize the
15-kDa peptide, even though they inhibited the binding of mAb 137C1 to
Tg (24).
We next tested the reactivity of the 23-kDa peptide with the sera from
patients with the thyroid diseases and with those from normal
individuals. As shown in Fig. 5B, this peptide reacted
strongly with sera from the same three patients with Hashimoto’s
thyroiditis, which inhibited the binding of mAb 137C1 to Tg in a
competitive ELISA (24). Some of the Hashimoto’s disease
sera did not completely inhibit the binding of mAb to Tg in the
competitive assays (24), and these sera did not inhibit
the binding of mAb 137C1 to Tg. In a previous publication
(22), we showed that the 23-kDa peptide did not react with
some sera from patients with Hashimoto’s disease or with sera from
most of the patients with Graves’ disease or thyroid carcinoma in
Western immunoblot assays.
Competitive inhibition of 15-kDa peptide, 23-kDa peptide, and intact Tg for binding of mAb 137C1
Because the 15-kDa and 23-kDa peptides are located in different
parts of the Tg molecule and have no homologous amino acid sequences,
we needed to determine whether the two peptides bear the same epitope.
Each peptide or intact Tg was incubated with mAb 137C1 for 4 h at
room temperature, and the peptide- or intact Tg-treated mAb was tested
by ELISA for binding to the same peptide, the other peptide, or the
intact Tg. Fig. 6 shows the effects of
competitive inhibition by the 15-kDa and 23-kDa peptides and by intact
Tg on binding of mAb 137C1 to the 15-kDa peptide, the 23-kDa peptide,
and intact Tg, respectively. These results show that each peptide not
only inhibited binding of mAb 137C1 to the same peptide, but it also
inhibited the binding of this mAb to the other peptide and intact
Tg.
|
To look for possible conformational determinants of the two
peptides which reacted with mAb 137C1, the peptides were submitted to
analysis by the PHD program. The resulting predictions showed mixed
class (helix, strand, and loop) for each peptide. The individual
residue probabilities for helix, strand, and loop for each peptide were
graphically plotted with KaleidaGraph MacII (version 2.1.3, Abelbeck
Software, Reading, PA). Fig. 7A shows an expansion of the
first 85 residues of 23-kDa peptide to delineate the contrasting helix,
strand, and loop regions, with the initial portion of the sequence
showing a predominant helix (residues 15–29) surrounded by loop and
strand regions. The remaining secondary structure of this peptide (Fig. 7B) shows predominantly mixed strand and loop regions. In
contrast, the 15-kDa peptide (Fig. 7C) is composed primarily
of helix and loop regions that alternate along the primary sequence.
The probability of strand region existing in this peptide is very low.
Thus, the loops and helices of each peptide probably play the major
role in the conformational epitope of mAb 137C1, with very little
contribution of strand regions for epitope recognition by this
mAb.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
As shown previously, the 15-kDa fragment reactive with mAb 137C1 is
located at the carboxyl-terminal end of the Tg molecule, from amino
acid 2657 to the end of the molecule at amino acid 2748
(23). In the present study, when we sequenced the 23-kDa
peptide (amino acids 2089–2316), we found that it was not located at
the carboxyl terminal end of the Tg and that the two peptides do not
share any sequence homology of three or more amino acids (Fig. 3).
Thus, the 15-kDa peptide is not derived from the 23-kDa peptide.
A striking finding of our study, therefore, is that two apparently unrelated fragments of the Tg molecule bind the same mAb. Furthermore, because of its susceptibility to denaturation, we concluded that the epitope recognized by mAb 137C1 is conformational (21).
To find out what determines the conformation of the epitope present on
each peptide, we subjected the amino acid sequences of these peptides
to the PHD program for prediction of their secondary structures. The
results showed an overwhelming proportion of loop regions within each
sequence (Fig. 7). The 23-kDa peptide is composed of 55.7% loop,
34.6% strand, and 9.6% helix. The 15-kDa peptide showed an even
higher percentage of loop (71.7%), with 28.3% helix and virtually no
ß-sheet prediction. Outside of other mammalian Tg precursor protein
sequences found in the database SWISSPROT, multiple sequence alignments
resulted in low (43% or less) alignment predictions. The secondary
structures of the 15-kDa and 23-kDa peptides are not sufficiently
similar to any other known amino acid sequence published in SWISSPROT.
Consequently, although we can estimate the secondary structures of the
15-kDa and 23-kDa peptides, it is not possible to ascertain their
probable tertiary structures due to the lack of a homologous protein
whose structure have been previously been determined by experimental
methods such as nuclear magnetic resonance or crystallography. However,
we can say that loop regions, which are predominant in the structures
of both peptides, are most likely to be involved in a conformational
epitope recognized by mAb 137C1. Further work with successively shorter
fragments of both peptides will clarify the nature of this
conformational epitope.
Because there is no sequence identity between the two peptides, it is possible that the epitope is formed by discontinuous segments contributed by different regions of each peptide. An alternative, nonexclusive possibility is that the epitopes on the two fragments are only structurally similar based on shape and charge of the amino acid residues. Both possibilities have precedent in other systems (30, 31, 32, 33), as will be discussed below, although they have not been reported for thyroid-specific autoimmunity.
Evidence in support of a discontinuous epitope was initially provided by Geysen et al. (30). According to these investigators, the minimal requirement for significant binding of a peptide to an Ab is that three amino acids within the sequence of the peptide have both the correct identity and position. With viral proteins, they showed that a neutralizing Ab recognized a discontinuous epitope consisting of residues at three different positions in the same peptide chain. Thus, the fragments need to share only a few critical amino acid residues and not a large sequential segment. More recent reports extended these findings to include autoantibodies in systemic autoimmunities, such as rheumatoid factor or Ro, in which groups of investigators found multiple epitopes with different primary sequences reactive to single autoantibodies (32, 33).
Evidence in support of a shared, structurally related epitope was described by Limpaiboon et al. (31). In findings similar to ours, they reported that a mAb recognized multiple bands of a Plasmodium falciparum extract in immunoblots. Like us, these investigators were initially uncertain whether the smaller peptide was part of the larger one or if the bands represented different portions of the same Ag. They used a mAb to screen a genomic expression library in which seven clones were identified. Interestingly, after the clones were sequenced, similar sequences were not found among the clones, suggesting that the same structural epitope was formed by different amino acids (31). We as yet do not know whether we are dealing with a discontinuous epitope that has the same sequences that bind to 137C1 or with a structural epitope. Regardless, our finding may have significant ramifications with regard to the progression of the autoimmune response to disease.
The identification of these epitopes may have an important diagnostic value. Because Tg carries epitopes that bind sera of normal individuals (20), using the whole Tg in the diagnosis assays of autoimmune thyroiditis leads to many "false positive" reactions. The peptides recognized by mAb 137C1 can be used as a more specific probe for detection of autoimmune thyroid disease in humans. However, many but not all patients with thyroiditis develop Abs to the epitope recognized by mAb 137C1, implicating other epitopes in thyroiditis (20, 24). Our findings, then, can be used to initiate further investigation along these lines to develop a "mixture" of specific fragments to increase the specificity of Ab testing for the diagnosis of autoimmune thyroiditis.
The long-range goal of our research is to understand why many individuals develop benign autoimmunity characterized only by production of Tg-specific autoantibodies, whereas others progress to pathogenic autoimmunity as defined by the signs and symptoms of clinical autoimmune thyroid disease. We have previously reported that both patients and normal individuals recognize conserved portions of the Tg molecule, but only patients develop Abs to novel, species-specific restricted determinants (20). There is also evidence that the B cell is an important APC in autoimmune thyroiditis (34). Thus, our findings have potential biological significance relating Ag presentation by the B cell to epitope spread and the pathogenesis of autoimmune thyroiditis.
Although we have identified two peptides on different parts of the molecule that are recognized by mAb 137C1, other, higher m.w. fragments were also recognized by mAb 137C1 (21). These higher m.w. fragments could contain the same epitopes as the two smaller fragments or they may represent similar conformational epitopes. Only one specificity of B cell receptor would be necessary to bind any of the epitopes; however, each of these epitopes would most probably have different flanking residues. After binding to the B cell, flanking residues from each site may be internalized as bystander residues. When these Ags are processed and presented by the MHC class II molecules, the precise specificity of the peptide fragments presented could be substantially different from one another because of the differing flanking residues.
Mamula and Janeway (35) proposed a model of epitope diversification whereby Ag-specific B cells, as APC, allow heightened presentation of cryptic determinants at levels sufficient to stimulate CD4 T cells. T cells of differing specificities would be recruited and T cell responses thus amplified may favor progression of a benign autoimmune response to pathogenic autoimmunity. Even subtle changes in the flanking residues are known to produce profound changes in the immunogenicity of specific epitopes (36). The proposed mechanism relates to differences in the relative affinities of B cell Ag receptors (36). A recent study by Vijayakrishnan et al. (37) reports that the on-rate of Ag binding to surface receptors of primed B cells can influence the proportion of Th1 or Th2 cells, thereby affecting the pathogenic potential of the autoimmune response. Thus, the peptide fragments presented by the B cells may direct the T helper cell in determining the ultimate outcome of the autoimmune response. Consequently, delineating the fine specificities of the autoantibodies can aid in identifying individuals who may be at greatest risk of disease. These individuals may be candidates for early intervention for prevention of disease, before progression of benign autoimmunity has escalated to that of an irreversible pathological process.
|
Footnotes |
|---|
2 Address correspondence and reprint requests to Dr. C. Lynne Burek, Department of Pathology, School of Medicine, Ross Research Building, Room 648, Johns Hopkins University, 720 Rutland Avenue, Baltimore, MD 21205. E-mail address:
3 Abbreviations used in this paper: Tg, thyroglobulin; DA, disease associated.
Received for publication February 22, 1999. Accepted for publication September 9, 1999.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
