- Email: [email protected]
Analysis of the interaction between rat immunoglobulin e and rat mast cells using anti-peptide antibodies
Molecular Immunology, Vol. 24, No. 4, pp. 379-389, Printed in Great Britain.
0161-5890/87 $3.00 + 0.00 Pergamon Journals Ltd
1987
ANALYSIS OF THE INTERACTION BETWEEN RAT IMMUNOGLOBULIN E AND RAT MAST CELLS USING ANTI-PEPTIDE ANTIBODIES DAVID S. BURT,* GILLIAN Z. HASTINGS,JOHN HEALYand DENIS R. STANWORTH Rheumatology and Allergy Research Unit, Department of Immunology, University of Birmingham, Birmingham B15 2TJ, U.K. (Received
26 March 1986; accepted in revised form 26 September
1986)
Abstract-Polyclonal antisera with pre-determined specificities for a range of rat IgE epitopes were produced by immunizing rabbits with KLH-conjugates of five different synthetic peptides representing sequences 378-396, 414428, 491-503, 522-535 and 56G571 in the CH3 and CH4 domains of rat IgE. Each rabbit elicited peptide-specific antibodies which were capable of binding affinity-purified rat IgE (IR162) (titres l/lOOCrl/lO,OOO) and IgE in rat immunocytoma serum (IR162) either immobilized on microtitre-plates or in free-solution as assessed by ELISA. Heating a solution of rat IgE at 56°C for 1 hr, a treatment known to abolish the cytophilic activity of rat IgE and also induce irreversible conformational changes in the CH3 and CH4 domains, resulted in enhanced binding of the immunoglobulin to antibodies directed against IgE sequences represented by two of the synthetic peptides 414428 and 491-503, but not to the three other peptides. The five anti-peptide sera together with two previously studied antisera specific for rat IgE sequences 459472 and 542-557 were tested in functional assays designed to investigate the mode of interaction between rat IgE and its receptor on rat mast cells. Each anti-peptide serum was capable of inhibiting the binding of IgE to mast cells and furthermore, able to initiate the secretion of histamine from cells sensitized with rat IgE in an “anti-1gE”-induced manner. In view of the evidence implicating the CH3 and/or CH4 domains as the location of the mast cell receptor-site on rat IgE, we propose a model to describe the mode of interaction between IgE and its mast cell receptor.
INTRODUCTION Immunoglobulin E (IgE) plays an important role in type 1 immediate hypersensitivity reactions due to its ability to bind reversibily to high affinity receptors on mast cells and basophils (Stanworth et al., 1967; Metzger, 1983). Cross-linking of surface-bound IgE by antigen initiates a series of membrane and intracellular events terminating in the secretion of biologically active mediators. Results of limited studies suggest that IgE binds to mast cells via sites within its CH3 and/or CH4 domain (Dorrington and Bennich, 1978; Perez-Montfort and Metzger, 1982; Holowka and Baird, 1983). However, apart from one unconfirmed study (Hamburger, 1975) attempts to identify sub-domain sites within IgE which interact with its Fc receptor on mast cells have been unsuccessful. As an alternative approach to more conventional ways of tackling this problem, we have investigated the possibility of producing antibodies against selective linear sequences within IgE; which may be useful in delineating, precisely, the mast cell receptor binding site on IgE. Recently we reported that synthetic peptides corresponding to two surface accessible regions of the CH4 domain of rat IgE (His 542-Lys 557 and Tyr 459-Arg 472) were capable of eliciting
antibodies in rabbits which recognized specifically the same sequences in intact rat IgE (Burt et al., 1986). Furthermore, these anti-peptide antibodies were capable of differentiating between native rat IgE and IgE heated at 56°C for 1 hr; a treatment which induces irreversible conformational changes in the CH3 and CH4 domains of IgE (Dorrington and Bennich, 1973) and a reduction in its receptorbinding capacity (Dorrington and Bennich, 1978). These findings suggested that the heat-sensitive sequences against which the two antipeptide sera are directed may contribute to the mast cell receptor binding site on IgE. We have now extended this approach to study five additional CH4 and CH3 domain sequences of rat IgE. In this report we show that rabbit antisera produced against these peptides are also capable of binding specifically native rat IgE and that they demonstrate different capacities to distinguish between the native and heated immunoglobulin. Furthermore, these anti-peptide sera together with the two previously studied sera were shown to inhibit the binding of rat IgE to mast cells and also interact with cell-associated IgE. These results are discussed in terms of a model for the IgE-mast cell receptor interaction. MATERIALSAND METHODS
*Present and correspondence address: Division of Virology, National Institute for Medical Research, Mill Hill, London NW7 IAA, U.K.
Peptide
synthesis
Peptides were synthesied by the solid-phase pro379
DAVID S. BURT et al.
380
cedure (Merrifield, 1963) as described previously (Burt et al., 1986). Peptides were characterized by quantitative amino acid analysis (following hydrolysis in 6 M HCl for 24 hr at llOC), their homogeneity being checked by reverse-phase HPLC and TLC. Fractions giving the highest peptide yield and amino acid composition consistent with that expected from their sequences were used in subsequent studies. Peptides were stored in a desiccator at 4°C over P,Os. Antisera production
Purified synthetic peptides were coupled to KLH as previously described (Burt et al., 1986). Essentially, 5 mg peptide in 0.1 M phosphate buffer (pH 7) was coupled to 10 mg KLH using 1 ml 12 mM glutaraldehyde. Female New Zealand White rabbits (Buxted Rabbit Co., Sussex, U.K.) were immunized S.C.with 500 pg peptide-conjugate in Freund’s complete adjuvant. Injections of 500 pg conjugate were repeated on day 44 and 74. Animals were bled 30 days after the primary immunization and 14 days after each subsequent immunization. Serum was separated and stored at -20°C. Polyclonal rabbit serum specific for the Fc region of rat IgE was prepared as described by Batchelor and Stanworth, 1980. Purification of IgE
Rat IgE (IR162) (serum, a gift from Dr H. Bazin, University of Louvain, Brussels, Belgium) and human IgE (WT) (plasma by courtesy of Dr C. West, Walton Hospital, Liverpool, U.K.) were purified by ammonium sulphate precipitation, gel filtration and affinity chromatography as previously described (Burt et al., 1986). The resultant purified rat immunocytoma (IR162) IgE (prIgE)i showed positive reactivity with rabbit antisera against rat IgE(Fc) and rat light chain, but failed to react with antisera raised against: whole rat serum proteins, rat albumin, rat IgM(Fc) and rat IgM(Fc) and rat IgG(Fc). Purified human (WT) IgE (phIgE) reacted positively with goat anti-human IgE(Fc) and a sheep anti-human IgE(Fc), but not with mouse monoclonal antibodies specific for human IgG and IgM. Rat IgE was radiolabelled with I-125 using chloramine-T (Hunter, 1970) and isolated by chromatography on a sephadex-G25 column. The resultant material had a sp. act. of approx. 3.7 x 106cpm/pg protein.
was incubated at 56°C in a water bath for 1 hr, whilst the other was maintained at 20°C for the same timeperiod. Unused prIgE (56°C) was maintained at 4°C and used within 34 days. Rat IgE anti-ovalbumin serum
Hooded Lister rats (Bantin and Kingman Ltd, Hull, UK) were immunized i.p. with 10 pg ovalbumin mixed with AI( (4.63 mg) in a total vol of 0.6 ml. Animals were bled 14 days later and the serum isolated and pooled. The ability of the serum to passively sensitize rat mast cells was determined as outlined below. Antisera characterization
The abilities of the rabbit antisera to bind peptide and whole IgE, and the specificities of these reactions were determined by indirect and solution-phase inhibition enzyme-linked immunosorbent assays (ELISA). The assay conditions used have been described fully elsewhere (Burt et al., 1986). Essentially they involve the addition of serial dilutions of test antiserum or pre-immune serum to microtitre plate wells pre-coated with either free-peptide (2.5pM), prIgE or phIgE (1 ,ug/ml) or whole rat immunocytoma serum (IR162) (1 in 1000 dilution). Peptidereactive rabbit antibodies were identified by adding a 1 in 1000 dilution of goat anti-rabbit IgGhorseradish peroxidase conjugate followed by pphenylendiaminehydrogen peroxide substrate. The coloured reaction was measured at 492nm (O.D. 492) in an automated plate reader. Inhibition ELISA was performed as above, except that the rabbit anti-peptide sera were first preincubated at 4°C overnight with various concns of either free-peptide or prIgE or rMyelE prior to addition to the antigen-coated microtitre-plates. Binding of anti-peptide sera to I-125 rat IgE
Preparation of heated rat IgE [prZgE (56”C)]
Various dilutions of anti-peptide sera or preimmune serum were incubated with approx. 0.08 pg (300,000 cpm) I-125 rat IgE in PBS in a final vol of 30 ~1 for 2 hr at room temp. Fifty microlitres of 10% v/v Staphylococcus aureus Protein A bacterial adsorbent (Miles Laboratories, Slough, U.K.) was then added for 15 min at room temp followed by the addition of 1 ml PBS and centrifugation at 750g for 10 min. The radioactivity in each sedimented residue was counted after washing twice in PBS by repeated resuspension and centrifugation.
Immediately prior to use, a solution of prIgE (1 mg/ml) was divided into two equal aliquots. One
Isolation and purification of mast cells
TAbbreviations: prIgE,
Rat serosal mast cells were obtained by washing the peritoneal and pleural cavities of male Wistar rats (Bantin and Kingman, Hull, U.K.), with 25 ,uM Tris-buffered medium containing 123 mM NaCl, 2.7 mA4 KCl, 1 mM MgCl,, 5.6 mM glucose and 0.1% gelatin, pH 7.2, as described elsewhere (Burt and Stanworth, 1983). For histamine release studies
affinity-purified rat IgE; prIgE (56”Q affinity-purified rat IgE heated at 56°C for 1 hr; phIgE, affinity-purified human IgE; rMye1 E, serum from rats secreting IgE immunocytoma protein IR162; KLH, keyhole limpet hemocyanin.
Rat immunoglobulin E and rat mast cell interaction analysis
these samples were pooled, washed by repeated centrifugation at 400 g for 5 min and the mixed preparation of cells consisting of approx. 3-5% mast cells used directly. For I-125 rat IgE-binding studies mast cells were purified through 65% Percoll (Pharmacia). Pooled cells from one to two rats were resuspended in 1 ml T&buffered medium and layered onto 3 ml 65% Percoll. After centrifugation at 490 g for 15 min the supernatant was aspirated and the pellet comprising approximately 8&95% mast cells was washed three times in medium. Histamine release studies Binding of anti-peptide antibodies to cell-associated ZgE. The ability of antibodies in the antipeptide sera
to bind receptor-occupied rat IgE was assessed by their capacity to cross-link IgE on mast cells and subsequently mediate the secretion of histamine in an “anti-1gE”-induced manner (Humphrey et al., 1963). Pooled, washed mixed serosal cells from four rats were pre-incubated with 0.1 ml rat IgE immunocytoma serum (IR162) in a final vol of 1 ml for 1 hr at 37°C in order to saturate the mast cell IgEreceptors. Cells were washed and aliquots containing 5 x lo4 mast cells in Tris-medium supplemented with 1 mM CaCl, and 20 pg/ml phosphatidylserine (PS) (Sigma, London, U.K.), were challenged for 30 min at 37°C in a total incubation vol of 1 ml with various dilutions of anti-peptide sera or pre-immune serum previously decomplemented by heating at 56°C for 30 min. In some experiments cells were incubated with a mixture of 10e3 M 3-deazaadenosine (Southern Research Laboratories Inc., Birmingham, Alabama)/ 10d4 M L-homocysteine thiolactone (Sigma, London, U.K.), known inhibitors of IgE-mediated mast cell histamine secretion, for 5 min at 37°C prior to addition of anti-peptide serum. After the incubations, cells were sedimented by centrifugation at 400g for 5 min and each supernatant separated from its corresponding cell pellet. Percentage of the total histamine released into the supernatant was determined for each sample using an automated fluorimetric assay (Evans et al., 1973) and expressed as % histamine release histamine in supernatant x 100 = histamine in cell pellet ’ + histamine in supernatant Znhibition of passive sensitization. Various dilutions
of anti-peptide sera or pre-immune serum (50 ~1) were pre-incubated with 25 ~1 of a 1 in 2 dilution of rat IgE anti-ovalbumin serum. After an overnight incubation at 4”C, 50~1 of washed rat serosal cells comprising approx. 5 x lo4 mast cells were added and the tubes incubated at 37°C for a further 2 hr. Cells were then washed three times by repeated addition of cell medium and centrifugation at 400 g for 5 min. To
381
the cell pellet was added 0.5 ml Tris medium containing 20pg/ml PS and 1 mM CaCl, with, or without, 5 pg/ml ovalbumin for 15 min at 37°C. The cells were then processed for the measurement of histamine release as described above. Inhibition of binding of I-125 rat IgE to mast cells.
To approx. 0.08 pg of I-125 rat IgE in T&buffered medium were added various dilutions of anti-peptide sera or pre-immune serum in a final vol of 200 ~1. After a 2 hr incubation at room temp 5 x lo5 purified rat mast cells (100 ~1) were introduced for a further 2 hr at 37°C. Cell-associated I-125 rat IgE was separated from unbound labelled immunoglobulin according to the method of Kulczycki et al., 1974, by layering 75 ~1 of each incubation mixture onto 250 ~1 feotal calf serum in microcentrifuge tubes and sedimenting the cells by centrifugation for 30 set in a high-speed microcentrifuge. Supernatants were aspirated and the radioactivity in each pellet counted. Results were expressed as a percentage of the cpm I-125 rat IgE bound in the presence of the same dilution of pre-immune serum after subtraction of non-specific binding determined in the presence of 50 pg prIgE. RESULTS
Selection of rat ZgE heavy chain sequences for synthesis
Amino acid sequences predicted to be located on the surface regions of the intact rat IgE Fc-dimer (and therefore possibly accessible for interaction with cellular receptors) were selected from the primary structure of the rat IgE heavy chain deduced by Hellman et al., (1982). The criteria of peptide selection have been outlined elsewhere (Burt et al., 1986). The schematic structure of the IgE CH3 and CH4 domains and the tentative location of the peptides chosen for solid-phase synthesis (Table l), are shown in Fig. 1. In native rat IgE the sequences represented by peptides 378-396 and 560-571 possess putative glycosylation sites at Asn 381 and Asn 561 (Hellman et al., 1982). One of these sites (Asn 381) has also been identified in human (ND) IgE (Max et al., 1982), whilst Asn 561 is located in the additional CH4 C-terminal decapeptide present in rat, but absent in human IgE (Hellman et al., 1982). Amino acid analysis of the prepared peptides showed their compositions to be in good agreement with expected values (Table 2). Anti-peptide responses
Rabbits were immunized in duplicate with each KLH-peptide conjugate. All animals produced antibodies which recognized their own immunizing peptide giving titres of between 1 in 12,500 and 1 in 312,500 as assessed by indirect ELISA. The peptide specificity profiles of the most potent antisera from each pair of animals, taken after the second booster immunization, are shown in Fig. 2. These sera were
382
DAVID S. BURT et Table
1. Rat IgE primary
sequences
al.
selected for synthesis
Amino acid sequence” 378
396
,NH Glu Ser Glu Glu Asn Ile Thr Val Thr Trp Val Arg Glu Arg Lys Lys Ser Ile Gly NH, 414 428 ,NH Tyr Ser Ile Leu Pro Val Asp Ala Lys Asp Trp Ile Glu Gly Glu Gly NH, 491 503 ,NH Tyr Leu Gin Asp Ser Lys Leu Ile Pro Lys Ser Gin His Ser NH, 522 535 ,NH Tyr Arg Leu Glu Val Thr Lys Ala Leu Trp Thr Gin Thr Lys Gin NH, 571 560 ,NH Gly Asn Thr Ser Leu Arg Pro Ser Gin Ala Ser Met NH, “Amino acids are numbered according to their position in the primary sequence of IgE (IR2) heavy chain (Hellman et al., 1982). Peptides 491-503, 522-535 and 416428 were synthesized with additional N-terminal tyrosines not present in native rat IgE for future radio-iodination purposes. All peptides were amidated at their C-terminus to avoid a C-terminal negative charge not normally present in the intact IgE molecule.
used, in all subsequent experiments. Antisera produced against peptides 414-428, 491-503, 522-535 and 560-571 exhibited a strict specificity for their own immunizing sequence. Surprisingly, antibodies raised against peptide 378-396 consistently recognized solid-phase peptide 560-571 as well as their own immunizing peptide [Fig. 2(A) and (D)]. This crossreactivity was observed in sera from two different rabbits and for every bleed including the pre-boost bleeds. This result cannot be explained on the basis of sequence homology (Fig. 1) and furthermore, the lack of recognition of peptide 378-396 by anti560-571 antibodies suggests that these antibodies are not recognizing common sequences or conformations.
Recognition of rat IgE The ability of the various anti-peptide sera to bind affinity purified rat IgE (prIgE) immobilized on microtitre-plate wells is illustrated in Fig. 3(A). Compared with pre-immune serum, each anti-peptide serum demonstrated positive reactivity with prIgE. Minimum ELISA titres for the reactions were in the range 1 in 1000-l in 10,000, several orders of magnitude lower than those observed for reactivity with homologous peptide (Fig. 2). Antibodies in serum produced against the C-terminal sequence 560-571 exhibited the highest titre (> 1 in 10,000) and the greatest absolute optical density reading. The species specificity of these interactions are demonstrated in
CH 3
CH 4
Fig. 1. Schematic representation of rat heavy-chain showing the CH3 and CH4 domains, adapted from the computer-graphic derived representation of human IgG CH2 and CH3 domains (Dr B. Sutton, Molecular Biophysics, University of Oxford), based on the Deisenhofer (1981) X-ray crystallographic co-ordinates. The presumed location of the synthetic peptides are shown together with their sequence number according to their position in rat IgE (IR2) heavy-chain as deduced by Hellman et al. (1982).
Rat imrnunoglobulin E and rat mast cell interaction analysis
383
Fig. 3(B), which shows that antibodies in sera raised against peptide 414-428, 491-503 and 522-535 were able to recognize affinity purified human IgE (phIgE). However, these anti-~ptide antisera reacted more strongly with rat IgE. Tbis cross-reactivity can be explained on the basis of varying degrees of homology between rat and human IgE in sequences represented by these peptides (Hellman et al., 1982); antibodies against peptide 414-428, which reflects greater than 60% sequence homology recognized phIgE to the greatest extent [Fig. 3(B)]. When tested for their capacity to bind IgE in rat immunocytoma serum (IRl62), the anti-peptide sera demonstrated identical patterns of reactivity as those for their interaction with prIgE (data not shown). Results from inhibition ELISA studies showed that the IgEreactive fraction of each antiserum resided in the peptide-reactive population of antibodies; since, IgEcontaining immunocytoma serum inhibited the binding of anti-peptide antibodies to their homologous peptides, or alternatively, pre-incubation with homologous peptide (but not other peptides) inhibited the binding of anti-peptide antibody to immobilized rat IgE (data not shown). Binding of anti-peptide sera to rat IgE heated at 56°C
The interactions between each anti-peptide serum and solid-phase prIgE were inhibited to various extents in a dose-de~ndent manner by prIgE (Fig. 4). These findings indicate that in each of the antisera there is a population of anti-peptide antibodies which are capable of binding rat IgE in free-solution. Heated (56°C for 1 hr) rat IgE, previously demonstrated to show reduced binding to antibodies specific for the Fc region of rat IgE (Burt et al., 1986), proved to be considerably more effective compared to prIgE at inhibiting the binding of anti-414428 and anti-491-503 sera to solid-phase prIgE [Fig. 4(B) and (C)l. PrIgE (10 pg/ml) inhibited the anti-414-428-prIgE interaction by 10.5%, whilst the same concn of prIgE (56°C) gave 30% inhibition [Fig. 4(B)]. For the ~ti-491-5O~IgE interaction [Fig. 4(C)], prIgE was only inhibitory at concns greater than SOpg/ml; whilst, prIgE (56°C) showed significant inhibition (2@47%) over a much lower concn range of O.l-lOpg/ml. As was previously shown (Burt et al., 1986), prIgE (56°C) was also more effective than prIgE at inhibiting the binding of antibodies raised against peptide 459472 to solidphase prIgE [Fig. 4(D)]. Conversely, antisera directed against sequences 378-396, 522-535 and 560-571 were unable to distinguish between the two forms of IgE, as shown in Fig. 4(A), (E) and (F), where the inhibition curves produced by prIgE and prIgE (56°C) were superimposable. Inhibition of binding of ZgE to rat mast cells
Z-125 rat ZgE. Antibodies in a polyclonal rabbit serum specific for rat IgE(Fc) bound greater than
DAVID S. BURT et al.
384
f
. A&-378-396 A II 414-428 cl 1' 491-503 0 II 560-571 .
0”
II 522-535
CD)
(El
2
I
0
103
104
105
106
103
104
I05
106
102
IO'
104
105
Fig. 2. Binding of anti-peptide antibodies to various peptides. Serial dilutions of each antiserum were added to microtitre plates coated with 2.5 nM free peptide as described in the Materials and Methods section. Peptide coating: (A) 378-396; (B) 414428; (C) 491-503; (D) 560-571; and (E) 522-535. Each point is the mean of duplicate determinations. 3-
(A) 0
ti
: 0"
1
.
Ix\ \
‘L_ 0
’ IO
\
I 50
-x
9
1
2-
I ---x-_-x_ 100
2%L_*_~, , 500
1000
5000
~0000
Dilution-’
Figure 3. Binding of anti-peptide antibodies to affinity-purified IgE. Serials dilutions of pre-immune serum (X) or sera raised against peptides 378-396 (A); 414428 (A); 491-503 (0); 5222535 (0) or 56&571 (0) were added to microtitre plates coated with (A) 1 pg/ml prIgE or (B) 1 pg/ml phIgE. Each point is the mean of duplicate determinations.
Rat immunoglobulin
E and rat mast cell interaction
r 0
01
051
5
IO
analysis
385
(E)
50100
Fig. 4. Binding of anti-peptide sera to rat IgE heated at 56°C for 1 hr. Antisera raised against sequence (A) 378-396 (1 in 100); (B) 414428 (1 in 200); (C) 491-503 (1 in 100); (D) 459472 (1 in 1000); (E) 522-535 (1 in 200); and (F) 56tL571 (1 in 1000) were pre-incubated with various concns of either prIgE (--) or prIgE (56°C) (---) prior to addition to microtitre-plates coated with 1 ng/ml prIgE. Reactions were processed as described in Materials and Methods. Each point is the mean & SEM for three determinations.
70% of the total cpm I-125 rat IgE added at 1 in 60 dilution (Table 3). In the same assay, compared with pre-immune serum, anti-sera directed against sequences 378-396,522-535 and 560-571 bound considerable amounts of radiolabelled immunoglobulin. Anti-peptide antibodies in sera specific for rat IgE sequences 414-428, 459472, 491-503 and 542-535 bound low, but significant levels of radioiodinated rat IgE. Pre-incubation of I-125 rat IgE with various dilutions of anti-peptide sera directed against sequences 378-396, 522-535 and 56&571 resulted in the inhibition of binding of radiolabelled prIgE to mast cells (Fig. 5). Antibodies speciffc for the CH4, C-terminal sequence, 560-571 showed the greatest activity, reducing IgE-binding to 68.5% of the control value at a dilution of 1 in 20. Not surprisingly antibodies directed against sequences 414-428, 459472, 491-503 and 542-557 in rat IgE, which
Table 3. Bindine
bound low levels of I-125 rat IgE (Table 3), had no effect on the binding of I-125 rat IgE to mast cells. Passive sensitization. Since four of the seven antipeptide sera failed to bind appreciable amounts of I-125 rat IgE, the ant&era were tested for their capacity to interfere with the ability of native IgE, in a rat anti-ovalbumin serum, to bind to rat mast cells. As Fig. 6 shows pre-treatment of rat anti-ovalbumin serum with various dilutions of each of the antipeptide sera prior to addition of rat mast cells resulted in a dose-dependent inhibition of passivesensitization as assessed by the subsequent reduced ability of the cells to release histamine in response to ovalbumin. Antibodies in the same dilution of preimmune serum had no significant effect on passive sensitization compared with cells sensitized in the presence of buffer alone, indicating that the observed inhibition with the anti-peptide sera was not due to
of anti-DeDtide “‘I-rat
Antiserum specificity Non-immune 378-396 414428 459472 491-503 522-535 542-557 560-571 Rat IgE (Fc)
l/2 1942 f 21.7 69,408 f 3057 2052 + 2800 + 2147k 73,058 f 4119 k 117,224 i -
32.1 108 116 517 287.5 6672
antibodies
to ‘2SI-rat IaE
IgE - Bound f SEM Serum dilution
l/6 1664 f 26.9 17,926 + 1774 2327 2792 2596 21,478 3555 90,733 193,118
+ f + + + + +
90.4 201.5 429 2392 122.4 3831 12,651
l/60 -
235,511 f 10,715
Antipeptide sera or non-immune serum at final dilutions indicated were incubated with 324,000 cpm I-125 rat IgE in a total vol of 30~1 for 2hr at 20°C. I-125 IgE bound to rabbit antibodies was precipitated with 50 ~1 10% v/v Staphylococcus aureus Protein A as described in Materials and Methods. Values represent the mean + SEM of triplicate determinations.
DAVID S. BURT et
386
al.
immunoglobulins in the rabbit sera competing with IgE antibodies for IgE(Fc) receptors on the mast cells. These results suggest that antibodies directed against the surface sequences of the CH3 and CH4 domains of rat IgE can inhibit the binding of IgE to its mast cell receptor. Interaction between anti-peptide receptor-occupied IgE
0-0
AntIserum
dilutbon-’
Fig. 5. Inhibition of binding of I-125 rat IgE to mast cells. Approximately 300,OOOcpm I-125 rat IgE were incubated with various dilutions of anti-peptide sera for 2 hr at 20°C. Purified rat mast cells (5 x 10s) were added for a further 2 hr at 37°C and cell-associated I-125 IgE determined as described in Materials and Methods. Results are expressed as the percentage of CPM I-125 IgE bound in presence of the same dilution of pre-immune serum (13,494 f 469.3) after subtraction of non-specific binding (901.7 f 12.1). Each point is the mean + SEM for triplicate determinations. Anti-459472 (x ), anti-542-557 (B). Other symbols as in Fig. 2.
0
I I
I
2 AntIserum
I
4 dilution-’
‘+----A
Fig. 6. Inhibition of passive sensitization of mast cells. Twenty-five microlitres of a 1 in 2 dilution of rat antiovalbumin serum was preincubated with 50~1 of various dilutions of anti-peptide serum overnight at 4°C. Mixed rat serosal cells (50 ~1) were added for a further 2 hr at 37°C. Cells were challenged with 0.5 pg/ml ovalbumin and processed for histamine release as already described. Results are expressed as percentage histamine released compared to that observed for cells sensitized in the presence of the same concn of pre-immune serum. Spontaneous histamine release = 6.05 f 1.1%. Histamine release for cells sensitized in presence of buffer alone = 2 1.7 f 0.1%. Each point is the mean + SEM for triplicate determinations. Symbols as in Figs 2 and 5.
antibodies
and
The capacity of antibodies in each of the antipeptide sera to recognize sites on receptor-occupied IgE was assessed by their ability to cross-link receptor-associated IgE and initiate the secretion of histamine (Table 4). Compared with pre-immune serum, each of the anti-peptide sera were capable of releasing significant amounts of histamine from IgE-sensitized mast cells in a dose-related manner. Anti-56CL571 serum was by far the most effective, initiating beween 3&60% histamine release between 1 in 2 and 1 in 16 dilution. Antisera comprising antibodies specific for peptides 378-396,459472 and 542-557 induced the release of more than 10% of the total histamine content of the cells at the lowest dilutions whilst antipeptide antibodies directed against sequences 414428 and 491-503 induced the release of low (2% above controls), but significant levels of histamine. Interestingly, these results demonstrate that the same antipeptide antibodies which impair the binding of rat IgE to mast cells, presumably by masking the receptor-site on IgE can also interact with receptor-occupied rat IgE. It has been shown that 3-deazaadenosine (in the presence of L-homocysteine thiolactone) and phosphatidylserine (PS) inhibit and enhance respectively the secretion of histamine from mast cells mediated by the cross-linking of membrane-associated IgE; but not mediated by other non-IgE-dependent mechanisms (Morita et al., 1981; Goth et al., 1971). In this present study, histamine secretion induced by a 1 in 100 dilution of a rabbit anti IgE(Fc) serum (76.7%) was reduced to 56.7 and 28.7% in the presence of 100 PM 3-deazaadenosinei 10 mML-homocysteine thiolactone and absence of PS, respectively (Fig. 7). In the same assay, histamine secretion induced by the most active anti-peptide sera was likewise suppressed in the presence of 3-deazaadenosine, and apart from the anti 459-472 mediated reaction, enhanced by PS. These results suggest that the anti-peptide sera and the rabbit anti-rat IgE(Fc) serum mediate the secretion of histamine from mast cells by similar mechanisms, involving the cross-linking of membranebound rat IgE. DISCUSSION The results of this present study extend and confirm our previous findings, which demonstrated that antibodies produced against peptides corresponding to various regions of the heavy chain of rat IgE are capable of binding to the intact, native IgE molecule
387
Rat immunoglobulin E and rat mast cell interaction analysis Table 4. Anti-peptide
sera-induced
secretion
of histamine
from rat mast cells
% Histamine release Serum dilution Serum specificity
112
114
118
l/l6
2.2 k 0.06 8.5 + 0.3
2.2 * 0.03 5.6 k 0.3
1.4+0.1 3.6 f 0.2
Pre-immune 378-396
3.1 * 0.3 11.1 f0.6
414-428 459-472 491-503 522-535
5.1 f 0.1 11.6kO.7 4.9 f 0.3 20.3 + 0.9
3.8 8.6 3.8 17.0
542-557 560-571 Rat IgE (Fc)
13.0 * 0.5 60.1 kO.6
7.9 * 0.3 60.1 + 0.08
Sensitized rat containing described “P < 0.05. NS alone was
k * + *
0.2” 0.03 0.07” 0.2
2.9 4.8 2.8 12.7
f f f f
0.3(NS) 0.2 0.2(NS) 0.5
6.9 + 0.3 51.8 k 0.9 -
3.4 4.3 2.0 9.0
+ + f f
l/100
l/l000
-
0.10 0.4 O.l(NS) 0.2
5.0 f 0.4 37.2 f 0.6
74.4 k 0.8
58.8 f 0.6
serosal cells were incubated at 37°C for 30min with various dilutions of anti-peptide sera or pre-immune in Tris-buffer 1 mM CaCI, and 20 pg/ml PS in a total vol of I ml. Samples were processed for the determination of histamine release as in Materials and Methods. Values represent the mean + SEM of triplicate determinations. = not significant. P values were between 0.001 and 0.005 for all others. Percentage histamine released in presence of buffer 2.9 + 0.3%.
(Burt et al., 1986). In common with other reports (Neurath et al., 1982; Burt et al., 1986), the antipeptide sera recognized their homologous peptides to a greater extent than the intact IgE protein. The minimum ELISA titres for anti-peptide antibody binding to homologous peptide and rat IgE were 1 in 62,500 to 1 in 312,500 and 1 in 1000 to 1 in 10,000, respectively (Figs 2 and 3). As discussed elsewhere (Sachs et al., 1972) such differences in affinities may be expected since sequences represented by synthetic peptides would in the native protein have structures which are more ordered than those of the peptides themselves in free-solution. Interestingly, antibodies raised against sequence 560-571 demonstrated the greatest ability to bind rat IgE as expressed in terms of both intensity of interaction and minimum ELISA titre of the antiserum. In native IgE this sequence would probably possess the greatest conformational freedom compared with other nonterminal sequences.
Heating IgE at 56°C markedly reduces its cytophilicity (Stanworth and Kuhns, 1965; Dorrington and Bennich, 1973). When each anti-peptide antiserum was tested for its capacity to bind native rat IgE and IgE heated at 56°C for 1 hr in a solutionphase inhibition assay, antibodies directed against peptides 414428 and 491-503 demonstrated enhanced binding to heated rat IgE as shown by a shift in the IgE-inhibition curve to the left (Fig. 4). These data support the previous findings of Dorrington and Ben&h, 1978, suggesting that irreversible conformational changes occur within areas of both the CH3 and CH4 domains of IgE as a result of heating at 56°C. Within these domains, such alterations in conformation appear to be localized to sequences 414428 and 491-503 according to findings from this study, and sequences 45!9472 and 542-557 from our previous work (Burt et al., 1986). Conversely, sequences represented by peptides 378-396, 522-535 and 560-571 appear to remain unaffected by heating 1
0
Control 3-DZA/homo.th~o
0
Minus
PS
3
0
0 NRS
378-396
459-472
AntIserum
522-535
speclf uty
542-557
)f
I RatIgE(Fc)
Fig. 7. Effect of 3-deamadenosine (3-DZA) and phosphatidylserine (PS) on anti-peptide-induced secretion of histamine. Rat serosal cells were either challenged with a 1 in 2 dilution of each anti-peptide serum or a 1 in 500 dilution of a rabbit anti-rat IgE(Fc) serum in the presence (open bars) or absence (shaded bars) of 20yg/ml PS; or after a pre-incubation with 10-j M/3-DZA (in the presence of lo-’ ML-homocysteine thiolactone) and challenged in the presence of 20 ug/ml PS (dark bars). Conditions as described in Table 3. Results are expressed as mean percentage histamine released &-SEM for three determinations.
388
DAVID S. BURT et al.
under the same conditions. Hence, heating rat IgE at 56°C induces localized structural changes in the CH3 and CH4 domains of IgE rather than a generalized change in conformation. If as we predict, the peptides used in these studies represent surface accessible regions of the CH3 and CH4 domains, our findings suggest that one or more of the IgE-Fc sequences represented by amino acids 414428, 459472, 491-503 and 542-557 may be part of the heatsensitive cytophilic site on rat IgE. Indeed recent experiments have shown that synthetic peptides corresponding to these four rat IgE sequences, but not others, are able to inhibit the binding of I-125 rat IgE to rat mast cells (Burt and Stanworth, submitted). Antibodies in sera specific for each of the seven IgE sequences studied were able to block, to various extents the binding of IgE to mast cells (Figs 5 and 6). The capacity of antibodies directed against any part of the exposed surface of the CH3 and CH4 domains of rat IgE to inhibit IgE receptor binding suggests that this region of rat IgE is involved in the receptor interaction. This proposal is consistent with data from other laboratories based on the localization of heat-sensitive domains on IgE (Dorrington and Bennich, 1978); the relative enzymatic susceptibility of IgE in solution and bound to its receptor (Perez-Montfort and Metzger, 1982); and the mapping of distances between various regions on receptor-occupied IgE and the mast cell membrane surface (Holowka and Baird, 1983). Furthermore, a mouse monoclonal anti-rat IgE(Fc) recently shown to recognize an epitope in the CH2 domain (Holowka et al., 1985) was unable to inhibit the binding of rat IgE to rat basophilic leukemic cells (Conrad et al., 1983), arguing against the involvement of the CH2 domain in receptor-binding. However: whereas the anti-peptide antisera were able to differentiate non-heat-sensitive between heat-sensitive and sequences (Burt et al., 1986; Fig. 4 in this paper), they were unable to distinguish between sites which may be involved in receptor-binding and other nonreceptor-interacting sites. This result may be explained in terms of stearic effects associated with the binding of anti-peptide antibodies to sequences adjacent to, but not corresponding to the mast cell Similar observations have been receptor-site. reported by Baniyash and Eshhar (1984) using a variety of mouse monoclonal antibodies specific for non-identified conformation determinants within the Fc region of rat IgE. The possibility of stearic interference of the IgE-receptor sites by antibodies in this present study could be tested by using isolated F(ab); and Fab’ fragments of the intact antipeptide antibodies. However, other investigators have found that both intact IgE(Fc)-reactive monoclonal antibodies and their Fab’ fragments are equally effective at inhibiting the binding of IgE to its receptor (Baniyash and Eshhar, 1984). An alternative explanation for the generalised inhibitory effect observed may be that binding of anti-peptide antibodies at
distant sites may induce conformational changes at the receptor-site. This possibility remains to be investigated. Each peptide-specific serum used in this present study was also capable of interacting with receptoroccupied IgE as assessed by their ability to induce the secretion of histamine from IgE-sensitized mast cells (Table 4). The rank order of effectiveness of the anti-peptide sera in stimulating secretion was similar to their efficacies in binding free rat IgE (Burt et al., 1986; Fig. 3(A) in this paper), supporting the proposal that the anti-peptide antibodies in these sera are cross-linking surface IgE on the mast cell membrane. More importantly, these results suggest that surface regions of the CH3 and CH4 domains are exposed on receptor-occupied rat IgE. It could be argued that sequences within these two domains of rat IgE which are not involved in receptor-binding would be accessible to specific antibodies. In such cases, the ability of the same anti-peptide antibodies to also inhibit the binding of rat IgE to mast cells could be explained in terms of the stearic or conformational effects already alluded to. A similar dual activity for monoclonal anti-rat IgE(Fc) antibodies has also been reported (Baniyash and Eshhar, 1984). However, the ability of antibodies directed against sequences, based on our studies to be putative receptor-interacting sites, to interact with receptor-occupied rat IgE has important implications in relation to the orientation of the IgE heavy-chain when bound to its specific receptor on mast cells. Our finding that antibodies directed against most of the CH3-CH4 surface, including those specific for the terminal sequence of the last domain (56&571) can bind to receptor-occupied IgE argues against the existence of a “pocket” receptor for rat IgE, where the CH3-CH4 domains are masked and the IgE molecule sits perpendicular to the membrane surface. That the Fc region of receptor-occupied rat IgE is accessible to proteolytic enzymes (Perez-Montfort and Metzger, 1982) and monoclonal antibodies (Baniyash and Eshhar, 1984) supports this view. In common with other immunoglobulin classes, the structure of rat IgE is such that any sequences comprising the receptor-site (which we predict to involve regions 414428, 459-472, 491-503 and 542-557) would exist in duplicate with corresponding sites on each heavy-chain rotated 180” with respect to one another (Metzger, 1983). Furthermore, the antipeptide antibodies described in this present study would recognize linear sequences which are likewise duplicated on each identical heavy-chain, unlike conventional antibodies which normally recognize conformational determinants which may comprise amino acid residues contributed by both heavychains. This latter property of the anti-peptide antibodies is crucial to the model that we now propose for the IgE-receptor interaction. On the basis of our own data together with that of other groups, we suggest that only one of the identical pair of se-
Rat immunoglobulin
E and rat mast cell interaction
quences postulated to comprise receptor-site residues actually interacts with the receptor, thereby leaving the symmetry-related set accessible and available for combination with its specific anti-peptide antibody. This may occur in one of two ways; either, the receptor contact residues are provided by both heavychains, or alternatively, all receptor contact residues may be derived from a single heavy-chain. In conclusion, irrespective of which of these arrangements is correct, the proposed model predicts a lateral interaction between the receptor and rat IgE along its two-fold axis of symmetry. This idea is consistent with results from Holowka and Baird (1983), and Holowka et al. (1985) which imply that the terminal end of the heavy-chain of rat IgE is closely associated with the mast cell membrane, whilst the remainder of the molecule is not fully extended from the membrane, but rather, disposed at an angle to the surface with the CH2 domains approx. 35-52A away from the membrane. This proposal necessitates that the IgE heavy-chain has some degree of flexibility at the CH2-CH3 interface; a prediction supported by at least one study (Slattery et al., 1985). A similar type of model has recently been suggested for human IgGl bound to its receptor on monocytes (Woof et al., 1986). Acknowledgements-The authors would like to thank Dr John Fox, Macromolecular Service Laboratory, Department of Chemistry, University of Birmingham for assistance with peptide synthesis and Janet Hadley for typing this manuscript. This work was supported by the Wellcome Trust.
REFERENCES
Baniyash M. and Eshhar Z. (1984) Inhibition of IgE binding to mast cells and basophils by monoclonal antibodies to murine IgE. Eur. J. Immun. 14, 799-807. Batchelor K. W. and Stanworth D. R. (1980) A comparison of the histamine-releasing properties of rat pleural and peritoneal mast cells. Immunology 41, 271-278. Burt D. S., Hastings G. Z. and Stanworth D. R. (1986) Use of synthetic peptides in the production and characterization of antibodies directed against predetermined specificities in rat immunoglobulin E. Molec. Immun. 23, 181-191. Burt D. S. and Stanworth D. R. (1983) The effect of ribose and purine modified adenosine analogues on the secretion of histamine from rat mast cells induced by ionophore A23 187. Biochem. Pharmac. 32, 2729-2732. Conrad D. H., Studer E., Gervasoni J. and Mohanakumar T. (1983) Properties of two monoclonal antibodies directed against the Fc and Fab’ regions of rat IgE. Int. Archs Allergy appl. Immun. 70, 352-360. Deisenhofer J. (198 I) Crystallographic refinement and atomic models of a human Fc fragment and its complex with fragment B of protein A from Staphylococcus aureus at 2.9 and 2.8 A resolution. Biochemistry 20, 2361-2370. Dorrington K. J. and Bennich H. H. (1973) Thermally induced structural changes in immunoglobulin E. J. biol. Chem. 248, 8378-8384. Dorrington K. J. and Bennich H. H. (1978) Structurefunction relationships in human immunoglobulin E. Immun. Rev. 41, 3-25.
analysis
389
Evans D. I’., Lewis J. A. and Thompson D. S. (1973) An automated fluorimetric assay for the rapid determination of histamine in biological fluids. Life Sci. 12, 327-336. Goth A,, Adams H. R. and Knoohuizen M. (1971) Phosphatidylserine: selective enhancer of histamine release. Science U.S.A. 173, 103&1035. Hamburger R. N. (1975) Peptide inhibition of PrausnitzKustner reaction. Science 189, 389-390. Hellman L., Pettersson U., Engstrom A., Karlsson T. and Bennich H. H. (1982). Structure and evolution of the heavy-chain from rat immunoglobulin E. Nucl. Acid Res. 10, 6041-6049. Holowka D. and Baird B. (1983) Structural studies on the membrane-bound immunoglobulin E receptor complex. 2. Mapping of distances between sites on IgE and the membrane surface. Biochemistry 22, 3475-3484. Holowka D., Conrad D. H. and Baird B. (1985) Structural of membrane-bound immunoglobulin Emapping receptor complexes: use of monoclonal anti-IgE antibodies to probe the conformation of receptor-bound IgE. Biochemistry 24, 626&6267. Humphrey J. H., Austen K. F. and Rapp H. J. (1963) In vitro studies of reversed anaphylaxis with rat cells. Immunology 6, 226245. Hunter R. (1970) Standardization of the Chloramine-T method of protein iodination. Proc. Sot. exp. Eiol. Med. 133, 989-992. Kulczycki A., Jr, Isersky C. and Metzger H. (1974) The interaction between IgE and rat basophilic leukemia cells. I. Evidence for specific binding of IgE. J. exp. Med. 139, 600-616. Max E. E., Battey J., Ney R., Kirsch I. R. and Leder P. (1982) Duplication and deletion in the human immunoglobulin E genes. Cell 29, 691-699. Merrifield R. E. (1963) Solid phase peptide synthesis. 1. The synthesis of a tetrapeptide. J. Am. them. Sot. 85, 2149-2154. Metzger H. (1983) The receptor on mast cells and related cells with high affinity for IgE. In Contemporary Topics in Molecular Immunology (Edited by Inman F. P. and Kindt T. J.), Vol. 9, pp. 115-145. Plenum Press, New York. Morita Y., Chiang P. K. and Siraganian P. (1981) Effects of inhibitors of transmethylation on histamine release from human basophils. Biochem. Pharmac. 30, 785-791. Neurath A. R., Kent S. B. H. and Strick N. (1982) Specificity of antibodies elicited by a synthetic peptide having a sequence in common with a fragment of a virus protein, the hepatitis B surface antigen. Proc. Natn. Acad. Sci. U.S.A. 79, 7871-7875. Perez-Montfort R. and Metzger H. (1982) Proteolysis of soluble IgE-receptor complexes: localization of sites on IgE which interact with the Fc receptor. Molec. Immun. 19, 113-125. Sachs D. H., Schechter A. N., Eastlake A. and Anifinsen B. (1972) An immunologic approach to the conformational equilibria of peptides. Proc. Nain. Acad. Sci. USA. 69, 3790-3794. Slattery J., Holowka D. and Baird 8. (1985) Segmental flexibility of receptor-bound immunoglobulin E. Biochemistry 24, 7 1(r720. Stanworth D. R. and Kuhns S. W. J. (1965) Quantitative studies on the assay of human sensitizing antibodies (reagins). 1. An examination of factors affecting the accuracy of the Prausnitz-Kustner (P-K) test. Zmmunology 8, 323-344. Stanworth D. R., Humphrey J. H., Bennich H. H. and Johansson S. G. 0. (1967) Specific inhibition of the Prausnitz-Kustner reaction by an atypical human myeloma protein. Lancer ii, 330-332. Woof J. M., Partridge L. J., Jeffries R. and Burton D. R. (1986) Localization of the monocyte-binding region on human immunoglobulin G. Molec. Immun. 23, 319-330.
0161-5890/87 $3.00 + 0.00 Pergamon Journals Ltd
1987
ANALYSIS OF THE INTERACTION BETWEEN RAT IMMUNOGLOBULIN E AND RAT MAST CELLS USING ANTI-PEPTIDE ANTIBODIES DAVID S. BURT,* GILLIAN Z. HASTINGS,JOHN HEALYand DENIS R. STANWORTH Rheumatology and Allergy Research Unit, Department of Immunology, University of Birmingham, Birmingham B15 2TJ, U.K. (Received
26 March 1986; accepted in revised form 26 September
1986)
Abstract-Polyclonal antisera with pre-determined specificities for a range of rat IgE epitopes were produced by immunizing rabbits with KLH-conjugates of five different synthetic peptides representing sequences 378-396, 414428, 491-503, 522-535 and 56G571 in the CH3 and CH4 domains of rat IgE. Each rabbit elicited peptide-specific antibodies which were capable of binding affinity-purified rat IgE (IR162) (titres l/lOOCrl/lO,OOO) and IgE in rat immunocytoma serum (IR162) either immobilized on microtitre-plates or in free-solution as assessed by ELISA. Heating a solution of rat IgE at 56°C for 1 hr, a treatment known to abolish the cytophilic activity of rat IgE and also induce irreversible conformational changes in the CH3 and CH4 domains, resulted in enhanced binding of the immunoglobulin to antibodies directed against IgE sequences represented by two of the synthetic peptides 414428 and 491-503, but not to the three other peptides. The five anti-peptide sera together with two previously studied antisera specific for rat IgE sequences 459472 and 542-557 were tested in functional assays designed to investigate the mode of interaction between rat IgE and its receptor on rat mast cells. Each anti-peptide serum was capable of inhibiting the binding of IgE to mast cells and furthermore, able to initiate the secretion of histamine from cells sensitized with rat IgE in an “anti-1gE”-induced manner. In view of the evidence implicating the CH3 and/or CH4 domains as the location of the mast cell receptor-site on rat IgE, we propose a model to describe the mode of interaction between IgE and its mast cell receptor.
INTRODUCTION Immunoglobulin E (IgE) plays an important role in type 1 immediate hypersensitivity reactions due to its ability to bind reversibily to high affinity receptors on mast cells and basophils (Stanworth et al., 1967; Metzger, 1983). Cross-linking of surface-bound IgE by antigen initiates a series of membrane and intracellular events terminating in the secretion of biologically active mediators. Results of limited studies suggest that IgE binds to mast cells via sites within its CH3 and/or CH4 domain (Dorrington and Bennich, 1978; Perez-Montfort and Metzger, 1982; Holowka and Baird, 1983). However, apart from one unconfirmed study (Hamburger, 1975) attempts to identify sub-domain sites within IgE which interact with its Fc receptor on mast cells have been unsuccessful. As an alternative approach to more conventional ways of tackling this problem, we have investigated the possibility of producing antibodies against selective linear sequences within IgE; which may be useful in delineating, precisely, the mast cell receptor binding site on IgE. Recently we reported that synthetic peptides corresponding to two surface accessible regions of the CH4 domain of rat IgE (His 542-Lys 557 and Tyr 459-Arg 472) were capable of eliciting
antibodies in rabbits which recognized specifically the same sequences in intact rat IgE (Burt et al., 1986). Furthermore, these anti-peptide antibodies were capable of differentiating between native rat IgE and IgE heated at 56°C for 1 hr; a treatment which induces irreversible conformational changes in the CH3 and CH4 domains of IgE (Dorrington and Bennich, 1973) and a reduction in its receptorbinding capacity (Dorrington and Bennich, 1978). These findings suggested that the heat-sensitive sequences against which the two antipeptide sera are directed may contribute to the mast cell receptor binding site on IgE. We have now extended this approach to study five additional CH4 and CH3 domain sequences of rat IgE. In this report we show that rabbit antisera produced against these peptides are also capable of binding specifically native rat IgE and that they demonstrate different capacities to distinguish between the native and heated immunoglobulin. Furthermore, these anti-peptide sera together with the two previously studied sera were shown to inhibit the binding of rat IgE to mast cells and also interact with cell-associated IgE. These results are discussed in terms of a model for the IgE-mast cell receptor interaction. MATERIALSAND METHODS
*Present and correspondence address: Division of Virology, National Institute for Medical Research, Mill Hill, London NW7 IAA, U.K.
Peptide
synthesis
Peptides were synthesied by the solid-phase pro379
DAVID S. BURT et al.
380
cedure (Merrifield, 1963) as described previously (Burt et al., 1986). Peptides were characterized by quantitative amino acid analysis (following hydrolysis in 6 M HCl for 24 hr at llOC), their homogeneity being checked by reverse-phase HPLC and TLC. Fractions giving the highest peptide yield and amino acid composition consistent with that expected from their sequences were used in subsequent studies. Peptides were stored in a desiccator at 4°C over P,Os. Antisera production
Purified synthetic peptides were coupled to KLH as previously described (Burt et al., 1986). Essentially, 5 mg peptide in 0.1 M phosphate buffer (pH 7) was coupled to 10 mg KLH using 1 ml 12 mM glutaraldehyde. Female New Zealand White rabbits (Buxted Rabbit Co., Sussex, U.K.) were immunized S.C.with 500 pg peptide-conjugate in Freund’s complete adjuvant. Injections of 500 pg conjugate were repeated on day 44 and 74. Animals were bled 30 days after the primary immunization and 14 days after each subsequent immunization. Serum was separated and stored at -20°C. Polyclonal rabbit serum specific for the Fc region of rat IgE was prepared as described by Batchelor and Stanworth, 1980. Purification of IgE
Rat IgE (IR162) (serum, a gift from Dr H. Bazin, University of Louvain, Brussels, Belgium) and human IgE (WT) (plasma by courtesy of Dr C. West, Walton Hospital, Liverpool, U.K.) were purified by ammonium sulphate precipitation, gel filtration and affinity chromatography as previously described (Burt et al., 1986). The resultant purified rat immunocytoma (IR162) IgE (prIgE)i showed positive reactivity with rabbit antisera against rat IgE(Fc) and rat light chain, but failed to react with antisera raised against: whole rat serum proteins, rat albumin, rat IgM(Fc) and rat IgM(Fc) and rat IgG(Fc). Purified human (WT) IgE (phIgE) reacted positively with goat anti-human IgE(Fc) and a sheep anti-human IgE(Fc), but not with mouse monoclonal antibodies specific for human IgG and IgM. Rat IgE was radiolabelled with I-125 using chloramine-T (Hunter, 1970) and isolated by chromatography on a sephadex-G25 column. The resultant material had a sp. act. of approx. 3.7 x 106cpm/pg protein.
was incubated at 56°C in a water bath for 1 hr, whilst the other was maintained at 20°C for the same timeperiod. Unused prIgE (56°C) was maintained at 4°C and used within 34 days. Rat IgE anti-ovalbumin serum
Hooded Lister rats (Bantin and Kingman Ltd, Hull, UK) were immunized i.p. with 10 pg ovalbumin mixed with AI( (4.63 mg) in a total vol of 0.6 ml. Animals were bled 14 days later and the serum isolated and pooled. The ability of the serum to passively sensitize rat mast cells was determined as outlined below. Antisera characterization
The abilities of the rabbit antisera to bind peptide and whole IgE, and the specificities of these reactions were determined by indirect and solution-phase inhibition enzyme-linked immunosorbent assays (ELISA). The assay conditions used have been described fully elsewhere (Burt et al., 1986). Essentially they involve the addition of serial dilutions of test antiserum or pre-immune serum to microtitre plate wells pre-coated with either free-peptide (2.5pM), prIgE or phIgE (1 ,ug/ml) or whole rat immunocytoma serum (IR162) (1 in 1000 dilution). Peptidereactive rabbit antibodies were identified by adding a 1 in 1000 dilution of goat anti-rabbit IgGhorseradish peroxidase conjugate followed by pphenylendiaminehydrogen peroxide substrate. The coloured reaction was measured at 492nm (O.D. 492) in an automated plate reader. Inhibition ELISA was performed as above, except that the rabbit anti-peptide sera were first preincubated at 4°C overnight with various concns of either free-peptide or prIgE or rMyelE prior to addition to the antigen-coated microtitre-plates. Binding of anti-peptide sera to I-125 rat IgE
Preparation of heated rat IgE [prZgE (56”C)]
Various dilutions of anti-peptide sera or preimmune serum were incubated with approx. 0.08 pg (300,000 cpm) I-125 rat IgE in PBS in a final vol of 30 ~1 for 2 hr at room temp. Fifty microlitres of 10% v/v Staphylococcus aureus Protein A bacterial adsorbent (Miles Laboratories, Slough, U.K.) was then added for 15 min at room temp followed by the addition of 1 ml PBS and centrifugation at 750g for 10 min. The radioactivity in each sedimented residue was counted after washing twice in PBS by repeated resuspension and centrifugation.
Immediately prior to use, a solution of prIgE (1 mg/ml) was divided into two equal aliquots. One
Isolation and purification of mast cells
TAbbreviations: prIgE,
Rat serosal mast cells were obtained by washing the peritoneal and pleural cavities of male Wistar rats (Bantin and Kingman, Hull, U.K.), with 25 ,uM Tris-buffered medium containing 123 mM NaCl, 2.7 mA4 KCl, 1 mM MgCl,, 5.6 mM glucose and 0.1% gelatin, pH 7.2, as described elsewhere (Burt and Stanworth, 1983). For histamine release studies
affinity-purified rat IgE; prIgE (56”Q affinity-purified rat IgE heated at 56°C for 1 hr; phIgE, affinity-purified human IgE; rMye1 E, serum from rats secreting IgE immunocytoma protein IR162; KLH, keyhole limpet hemocyanin.
Rat immunoglobulin E and rat mast cell interaction analysis
these samples were pooled, washed by repeated centrifugation at 400 g for 5 min and the mixed preparation of cells consisting of approx. 3-5% mast cells used directly. For I-125 rat IgE-binding studies mast cells were purified through 65% Percoll (Pharmacia). Pooled cells from one to two rats were resuspended in 1 ml T&buffered medium and layered onto 3 ml 65% Percoll. After centrifugation at 490 g for 15 min the supernatant was aspirated and the pellet comprising approximately 8&95% mast cells was washed three times in medium. Histamine release studies Binding of anti-peptide antibodies to cell-associated ZgE. The ability of antibodies in the antipeptide sera
to bind receptor-occupied rat IgE was assessed by their capacity to cross-link IgE on mast cells and subsequently mediate the secretion of histamine in an “anti-1gE”-induced manner (Humphrey et al., 1963). Pooled, washed mixed serosal cells from four rats were pre-incubated with 0.1 ml rat IgE immunocytoma serum (IR162) in a final vol of 1 ml for 1 hr at 37°C in order to saturate the mast cell IgEreceptors. Cells were washed and aliquots containing 5 x lo4 mast cells in Tris-medium supplemented with 1 mM CaCl, and 20 pg/ml phosphatidylserine (PS) (Sigma, London, U.K.), were challenged for 30 min at 37°C in a total incubation vol of 1 ml with various dilutions of anti-peptide sera or pre-immune serum previously decomplemented by heating at 56°C for 30 min. In some experiments cells were incubated with a mixture of 10e3 M 3-deazaadenosine (Southern Research Laboratories Inc., Birmingham, Alabama)/ 10d4 M L-homocysteine thiolactone (Sigma, London, U.K.), known inhibitors of IgE-mediated mast cell histamine secretion, for 5 min at 37°C prior to addition of anti-peptide serum. After the incubations, cells were sedimented by centrifugation at 400g for 5 min and each supernatant separated from its corresponding cell pellet. Percentage of the total histamine released into the supernatant was determined for each sample using an automated fluorimetric assay (Evans et al., 1973) and expressed as % histamine release histamine in supernatant x 100 = histamine in cell pellet ’ + histamine in supernatant Znhibition of passive sensitization. Various dilutions
of anti-peptide sera or pre-immune serum (50 ~1) were pre-incubated with 25 ~1 of a 1 in 2 dilution of rat IgE anti-ovalbumin serum. After an overnight incubation at 4”C, 50~1 of washed rat serosal cells comprising approx. 5 x lo4 mast cells were added and the tubes incubated at 37°C for a further 2 hr. Cells were then washed three times by repeated addition of cell medium and centrifugation at 400 g for 5 min. To
381
the cell pellet was added 0.5 ml Tris medium containing 20pg/ml PS and 1 mM CaCl, with, or without, 5 pg/ml ovalbumin for 15 min at 37°C. The cells were then processed for the measurement of histamine release as described above. Inhibition of binding of I-125 rat IgE to mast cells.
To approx. 0.08 pg of I-125 rat IgE in T&buffered medium were added various dilutions of anti-peptide sera or pre-immune serum in a final vol of 200 ~1. After a 2 hr incubation at room temp 5 x lo5 purified rat mast cells (100 ~1) were introduced for a further 2 hr at 37°C. Cell-associated I-125 rat IgE was separated from unbound labelled immunoglobulin according to the method of Kulczycki et al., 1974, by layering 75 ~1 of each incubation mixture onto 250 ~1 feotal calf serum in microcentrifuge tubes and sedimenting the cells by centrifugation for 30 set in a high-speed microcentrifuge. Supernatants were aspirated and the radioactivity in each pellet counted. Results were expressed as a percentage of the cpm I-125 rat IgE bound in the presence of the same dilution of pre-immune serum after subtraction of non-specific binding determined in the presence of 50 pg prIgE. RESULTS
Selection of rat ZgE heavy chain sequences for synthesis
Amino acid sequences predicted to be located on the surface regions of the intact rat IgE Fc-dimer (and therefore possibly accessible for interaction with cellular receptors) were selected from the primary structure of the rat IgE heavy chain deduced by Hellman et al., (1982). The criteria of peptide selection have been outlined elsewhere (Burt et al., 1986). The schematic structure of the IgE CH3 and CH4 domains and the tentative location of the peptides chosen for solid-phase synthesis (Table l), are shown in Fig. 1. In native rat IgE the sequences represented by peptides 378-396 and 560-571 possess putative glycosylation sites at Asn 381 and Asn 561 (Hellman et al., 1982). One of these sites (Asn 381) has also been identified in human (ND) IgE (Max et al., 1982), whilst Asn 561 is located in the additional CH4 C-terminal decapeptide present in rat, but absent in human IgE (Hellman et al., 1982). Amino acid analysis of the prepared peptides showed their compositions to be in good agreement with expected values (Table 2). Anti-peptide responses
Rabbits were immunized in duplicate with each KLH-peptide conjugate. All animals produced antibodies which recognized their own immunizing peptide giving titres of between 1 in 12,500 and 1 in 312,500 as assessed by indirect ELISA. The peptide specificity profiles of the most potent antisera from each pair of animals, taken after the second booster immunization, are shown in Fig. 2. These sera were
382
DAVID S. BURT et Table
1. Rat IgE primary
sequences
al.
selected for synthesis
Amino acid sequence” 378
396
,NH Glu Ser Glu Glu Asn Ile Thr Val Thr Trp Val Arg Glu Arg Lys Lys Ser Ile Gly NH, 414 428 ,NH Tyr Ser Ile Leu Pro Val Asp Ala Lys Asp Trp Ile Glu Gly Glu Gly NH, 491 503 ,NH Tyr Leu Gin Asp Ser Lys Leu Ile Pro Lys Ser Gin His Ser NH, 522 535 ,NH Tyr Arg Leu Glu Val Thr Lys Ala Leu Trp Thr Gin Thr Lys Gin NH, 571 560 ,NH Gly Asn Thr Ser Leu Arg Pro Ser Gin Ala Ser Met NH, “Amino acids are numbered according to their position in the primary sequence of IgE (IR2) heavy chain (Hellman et al., 1982). Peptides 491-503, 522-535 and 416428 were synthesized with additional N-terminal tyrosines not present in native rat IgE for future radio-iodination purposes. All peptides were amidated at their C-terminus to avoid a C-terminal negative charge not normally present in the intact IgE molecule.
used, in all subsequent experiments. Antisera produced against peptides 414-428, 491-503, 522-535 and 560-571 exhibited a strict specificity for their own immunizing sequence. Surprisingly, antibodies raised against peptide 378-396 consistently recognized solid-phase peptide 560-571 as well as their own immunizing peptide [Fig. 2(A) and (D)]. This crossreactivity was observed in sera from two different rabbits and for every bleed including the pre-boost bleeds. This result cannot be explained on the basis of sequence homology (Fig. 1) and furthermore, the lack of recognition of peptide 378-396 by anti560-571 antibodies suggests that these antibodies are not recognizing common sequences or conformations.
Recognition of rat IgE The ability of the various anti-peptide sera to bind affinity purified rat IgE (prIgE) immobilized on microtitre-plate wells is illustrated in Fig. 3(A). Compared with pre-immune serum, each anti-peptide serum demonstrated positive reactivity with prIgE. Minimum ELISA titres for the reactions were in the range 1 in 1000-l in 10,000, several orders of magnitude lower than those observed for reactivity with homologous peptide (Fig. 2). Antibodies in serum produced against the C-terminal sequence 560-571 exhibited the highest titre (> 1 in 10,000) and the greatest absolute optical density reading. The species specificity of these interactions are demonstrated in
CH 3
CH 4
Fig. 1. Schematic representation of rat heavy-chain showing the CH3 and CH4 domains, adapted from the computer-graphic derived representation of human IgG CH2 and CH3 domains (Dr B. Sutton, Molecular Biophysics, University of Oxford), based on the Deisenhofer (1981) X-ray crystallographic co-ordinates. The presumed location of the synthetic peptides are shown together with their sequence number according to their position in rat IgE (IR2) heavy-chain as deduced by Hellman et al. (1982).
Rat imrnunoglobulin E and rat mast cell interaction analysis
383
Fig. 3(B), which shows that antibodies in sera raised against peptide 414-428, 491-503 and 522-535 were able to recognize affinity purified human IgE (phIgE). However, these anti-~ptide antisera reacted more strongly with rat IgE. Tbis cross-reactivity can be explained on the basis of varying degrees of homology between rat and human IgE in sequences represented by these peptides (Hellman et al., 1982); antibodies against peptide 414-428, which reflects greater than 60% sequence homology recognized phIgE to the greatest extent [Fig. 3(B)]. When tested for their capacity to bind IgE in rat immunocytoma serum (IRl62), the anti-peptide sera demonstrated identical patterns of reactivity as those for their interaction with prIgE (data not shown). Results from inhibition ELISA studies showed that the IgEreactive fraction of each antiserum resided in the peptide-reactive population of antibodies; since, IgEcontaining immunocytoma serum inhibited the binding of anti-peptide antibodies to their homologous peptides, or alternatively, pre-incubation with homologous peptide (but not other peptides) inhibited the binding of anti-peptide antibody to immobilized rat IgE (data not shown). Binding of anti-peptide sera to rat IgE heated at 56°C
The interactions between each anti-peptide serum and solid-phase prIgE were inhibited to various extents in a dose-de~ndent manner by prIgE (Fig. 4). These findings indicate that in each of the antisera there is a population of anti-peptide antibodies which are capable of binding rat IgE in free-solution. Heated (56°C for 1 hr) rat IgE, previously demonstrated to show reduced binding to antibodies specific for the Fc region of rat IgE (Burt et al., 1986), proved to be considerably more effective compared to prIgE at inhibiting the binding of anti-414428 and anti-491-503 sera to solid-phase prIgE [Fig. 4(B) and (C)l. PrIgE (10 pg/ml) inhibited the anti-414-428-prIgE interaction by 10.5%, whilst the same concn of prIgE (56°C) gave 30% inhibition [Fig. 4(B)]. For the ~ti-491-5O~IgE interaction [Fig. 4(C)], prIgE was only inhibitory at concns greater than SOpg/ml; whilst, prIgE (56°C) showed significant inhibition (2@47%) over a much lower concn range of O.l-lOpg/ml. As was previously shown (Burt et al., 1986), prIgE (56°C) was also more effective than prIgE at inhibiting the binding of antibodies raised against peptide 459472 to solidphase prIgE [Fig. 4(D)]. Conversely, antisera directed against sequences 378-396, 522-535 and 560-571 were unable to distinguish between the two forms of IgE, as shown in Fig. 4(A), (E) and (F), where the inhibition curves produced by prIgE and prIgE (56°C) were superimposable. Inhibition of binding of ZgE to rat mast cells
Z-125 rat ZgE. Antibodies in a polyclonal rabbit serum specific for rat IgE(Fc) bound greater than
DAVID S. BURT et al.
384
f
. A&-378-396 A II 414-428 cl 1' 491-503 0 II 560-571 .
0”
II 522-535
CD)
(El
2
I
0
103
104
105
106
103
104
I05
106
102
IO'
104
105
Fig. 2. Binding of anti-peptide antibodies to various peptides. Serial dilutions of each antiserum were added to microtitre plates coated with 2.5 nM free peptide as described in the Materials and Methods section. Peptide coating: (A) 378-396; (B) 414428; (C) 491-503; (D) 560-571; and (E) 522-535. Each point is the mean of duplicate determinations. 3-
(A) 0
ti
: 0"
1
.
Ix\ \
‘L_ 0
’ IO
\
I 50
-x
9
1
2-
I ---x-_-x_ 100
2%L_*_~, , 500
1000
5000
~0000
Dilution-’
Figure 3. Binding of anti-peptide antibodies to affinity-purified IgE. Serials dilutions of pre-immune serum (X) or sera raised against peptides 378-396 (A); 414428 (A); 491-503 (0); 5222535 (0) or 56&571 (0) were added to microtitre plates coated with (A) 1 pg/ml prIgE or (B) 1 pg/ml phIgE. Each point is the mean of duplicate determinations.
Rat immunoglobulin
E and rat mast cell interaction
r 0
01
051
5
IO
analysis
385
(E)
50100
Fig. 4. Binding of anti-peptide sera to rat IgE heated at 56°C for 1 hr. Antisera raised against sequence (A) 378-396 (1 in 100); (B) 414428 (1 in 200); (C) 491-503 (1 in 100); (D) 459472 (1 in 1000); (E) 522-535 (1 in 200); and (F) 56tL571 (1 in 1000) were pre-incubated with various concns of either prIgE (--) or prIgE (56°C) (---) prior to addition to microtitre-plates coated with 1 ng/ml prIgE. Reactions were processed as described in Materials and Methods. Each point is the mean & SEM for three determinations.
70% of the total cpm I-125 rat IgE added at 1 in 60 dilution (Table 3). In the same assay, compared with pre-immune serum, anti-sera directed against sequences 378-396,522-535 and 560-571 bound considerable amounts of radiolabelled immunoglobulin. Anti-peptide antibodies in sera specific for rat IgE sequences 414-428, 459472, 491-503 and 542-535 bound low, but significant levels of radioiodinated rat IgE. Pre-incubation of I-125 rat IgE with various dilutions of anti-peptide sera directed against sequences 378-396, 522-535 and 56&571 resulted in the inhibition of binding of radiolabelled prIgE to mast cells (Fig. 5). Antibodies speciffc for the CH4, C-terminal sequence, 560-571 showed the greatest activity, reducing IgE-binding to 68.5% of the control value at a dilution of 1 in 20. Not surprisingly antibodies directed against sequences 414-428, 459472, 491-503 and 542-557 in rat IgE, which
Table 3. Bindine
bound low levels of I-125 rat IgE (Table 3), had no effect on the binding of I-125 rat IgE to mast cells. Passive sensitization. Since four of the seven antipeptide sera failed to bind appreciable amounts of I-125 rat IgE, the ant&era were tested for their capacity to interfere with the ability of native IgE, in a rat anti-ovalbumin serum, to bind to rat mast cells. As Fig. 6 shows pre-treatment of rat anti-ovalbumin serum with various dilutions of each of the antipeptide sera prior to addition of rat mast cells resulted in a dose-dependent inhibition of passivesensitization as assessed by the subsequent reduced ability of the cells to release histamine in response to ovalbumin. Antibodies in the same dilution of preimmune serum had no significant effect on passive sensitization compared with cells sensitized in the presence of buffer alone, indicating that the observed inhibition with the anti-peptide sera was not due to
of anti-DeDtide “‘I-rat
Antiserum specificity Non-immune 378-396 414428 459472 491-503 522-535 542-557 560-571 Rat IgE (Fc)
l/2 1942 f 21.7 69,408 f 3057 2052 + 2800 + 2147k 73,058 f 4119 k 117,224 i -
32.1 108 116 517 287.5 6672
antibodies
to ‘2SI-rat IaE
IgE - Bound f SEM Serum dilution
l/6 1664 f 26.9 17,926 + 1774 2327 2792 2596 21,478 3555 90,733 193,118
+ f + + + + +
90.4 201.5 429 2392 122.4 3831 12,651
l/60 -
235,511 f 10,715
Antipeptide sera or non-immune serum at final dilutions indicated were incubated with 324,000 cpm I-125 rat IgE in a total vol of 30~1 for 2hr at 20°C. I-125 IgE bound to rabbit antibodies was precipitated with 50 ~1 10% v/v Staphylococcus aureus Protein A as described in Materials and Methods. Values represent the mean + SEM of triplicate determinations.
DAVID S. BURT et
386
al.
immunoglobulins in the rabbit sera competing with IgE antibodies for IgE(Fc) receptors on the mast cells. These results suggest that antibodies directed against the surface sequences of the CH3 and CH4 domains of rat IgE can inhibit the binding of IgE to its mast cell receptor. Interaction between anti-peptide receptor-occupied IgE
0-0
AntIserum
dilutbon-’
Fig. 5. Inhibition of binding of I-125 rat IgE to mast cells. Approximately 300,OOOcpm I-125 rat IgE were incubated with various dilutions of anti-peptide sera for 2 hr at 20°C. Purified rat mast cells (5 x 10s) were added for a further 2 hr at 37°C and cell-associated I-125 IgE determined as described in Materials and Methods. Results are expressed as the percentage of CPM I-125 IgE bound in presence of the same dilution of pre-immune serum (13,494 f 469.3) after subtraction of non-specific binding (901.7 f 12.1). Each point is the mean + SEM for triplicate determinations. Anti-459472 (x ), anti-542-557 (B). Other symbols as in Fig. 2.
0
I I
I
2 AntIserum
I
4 dilution-’
‘+----A
Fig. 6. Inhibition of passive sensitization of mast cells. Twenty-five microlitres of a 1 in 2 dilution of rat antiovalbumin serum was preincubated with 50~1 of various dilutions of anti-peptide serum overnight at 4°C. Mixed rat serosal cells (50 ~1) were added for a further 2 hr at 37°C. Cells were challenged with 0.5 pg/ml ovalbumin and processed for histamine release as already described. Results are expressed as percentage histamine released compared to that observed for cells sensitized in the presence of the same concn of pre-immune serum. Spontaneous histamine release = 6.05 f 1.1%. Histamine release for cells sensitized in presence of buffer alone = 2 1.7 f 0.1%. Each point is the mean + SEM for triplicate determinations. Symbols as in Figs 2 and 5.
antibodies
and
The capacity of antibodies in each of the antipeptide sera to recognize sites on receptor-occupied IgE was assessed by their ability to cross-link receptor-associated IgE and initiate the secretion of histamine (Table 4). Compared with pre-immune serum, each of the anti-peptide sera were capable of releasing significant amounts of histamine from IgE-sensitized mast cells in a dose-related manner. Anti-56CL571 serum was by far the most effective, initiating beween 3&60% histamine release between 1 in 2 and 1 in 16 dilution. Antisera comprising antibodies specific for peptides 378-396,459472 and 542-557 induced the release of more than 10% of the total histamine content of the cells at the lowest dilutions whilst antipeptide antibodies directed against sequences 414428 and 491-503 induced the release of low (2% above controls), but significant levels of histamine. Interestingly, these results demonstrate that the same antipeptide antibodies which impair the binding of rat IgE to mast cells, presumably by masking the receptor-site on IgE can also interact with receptor-occupied rat IgE. It has been shown that 3-deazaadenosine (in the presence of L-homocysteine thiolactone) and phosphatidylserine (PS) inhibit and enhance respectively the secretion of histamine from mast cells mediated by the cross-linking of membrane-associated IgE; but not mediated by other non-IgE-dependent mechanisms (Morita et al., 1981; Goth et al., 1971). In this present study, histamine secretion induced by a 1 in 100 dilution of a rabbit anti IgE(Fc) serum (76.7%) was reduced to 56.7 and 28.7% in the presence of 100 PM 3-deazaadenosinei 10 mML-homocysteine thiolactone and absence of PS, respectively (Fig. 7). In the same assay, histamine secretion induced by the most active anti-peptide sera was likewise suppressed in the presence of 3-deazaadenosine, and apart from the anti 459-472 mediated reaction, enhanced by PS. These results suggest that the anti-peptide sera and the rabbit anti-rat IgE(Fc) serum mediate the secretion of histamine from mast cells by similar mechanisms, involving the cross-linking of membranebound rat IgE. DISCUSSION The results of this present study extend and confirm our previous findings, which demonstrated that antibodies produced against peptides corresponding to various regions of the heavy chain of rat IgE are capable of binding to the intact, native IgE molecule
387
Rat immunoglobulin E and rat mast cell interaction analysis Table 4. Anti-peptide
sera-induced
secretion
of histamine
from rat mast cells
% Histamine release Serum dilution Serum specificity
112
114
118
l/l6
2.2 k 0.06 8.5 + 0.3
2.2 * 0.03 5.6 k 0.3
1.4+0.1 3.6 f 0.2
Pre-immune 378-396
3.1 * 0.3 11.1 f0.6
414-428 459-472 491-503 522-535
5.1 f 0.1 11.6kO.7 4.9 f 0.3 20.3 + 0.9
3.8 8.6 3.8 17.0
542-557 560-571 Rat IgE (Fc)
13.0 * 0.5 60.1 kO.6
7.9 * 0.3 60.1 + 0.08
Sensitized rat containing described “P < 0.05. NS alone was
k * + *
0.2” 0.03 0.07” 0.2
2.9 4.8 2.8 12.7
f f f f
0.3(NS) 0.2 0.2(NS) 0.5
6.9 + 0.3 51.8 k 0.9 -
3.4 4.3 2.0 9.0
+ + f f
l/100
l/l000
-
0.10 0.4 O.l(NS) 0.2
5.0 f 0.4 37.2 f 0.6
74.4 k 0.8
58.8 f 0.6
serosal cells were incubated at 37°C for 30min with various dilutions of anti-peptide sera or pre-immune in Tris-buffer 1 mM CaCI, and 20 pg/ml PS in a total vol of I ml. Samples were processed for the determination of histamine release as in Materials and Methods. Values represent the mean + SEM of triplicate determinations. = not significant. P values were between 0.001 and 0.005 for all others. Percentage histamine released in presence of buffer 2.9 + 0.3%.
(Burt et al., 1986). In common with other reports (Neurath et al., 1982; Burt et al., 1986), the antipeptide sera recognized their homologous peptides to a greater extent than the intact IgE protein. The minimum ELISA titres for anti-peptide antibody binding to homologous peptide and rat IgE were 1 in 62,500 to 1 in 312,500 and 1 in 1000 to 1 in 10,000, respectively (Figs 2 and 3). As discussed elsewhere (Sachs et al., 1972) such differences in affinities may be expected since sequences represented by synthetic peptides would in the native protein have structures which are more ordered than those of the peptides themselves in free-solution. Interestingly, antibodies raised against sequence 560-571 demonstrated the greatest ability to bind rat IgE as expressed in terms of both intensity of interaction and minimum ELISA titre of the antiserum. In native IgE this sequence would probably possess the greatest conformational freedom compared with other nonterminal sequences.
Heating IgE at 56°C markedly reduces its cytophilicity (Stanworth and Kuhns, 1965; Dorrington and Bennich, 1973). When each anti-peptide antiserum was tested for its capacity to bind native rat IgE and IgE heated at 56°C for 1 hr in a solutionphase inhibition assay, antibodies directed against peptides 414428 and 491-503 demonstrated enhanced binding to heated rat IgE as shown by a shift in the IgE-inhibition curve to the left (Fig. 4). These data support the previous findings of Dorrington and Ben&h, 1978, suggesting that irreversible conformational changes occur within areas of both the CH3 and CH4 domains of IgE as a result of heating at 56°C. Within these domains, such alterations in conformation appear to be localized to sequences 414428 and 491-503 according to findings from this study, and sequences 45!9472 and 542-557 from our previous work (Burt et al., 1986). Conversely, sequences represented by peptides 378-396, 522-535 and 560-571 appear to remain unaffected by heating 1
0
Control 3-DZA/homo.th~o
0
Minus
PS
3
0
0 NRS
378-396
459-472
AntIserum
522-535
speclf uty
542-557
)f
I RatIgE(Fc)
Fig. 7. Effect of 3-deamadenosine (3-DZA) and phosphatidylserine (PS) on anti-peptide-induced secretion of histamine. Rat serosal cells were either challenged with a 1 in 2 dilution of each anti-peptide serum or a 1 in 500 dilution of a rabbit anti-rat IgE(Fc) serum in the presence (open bars) or absence (shaded bars) of 20yg/ml PS; or after a pre-incubation with 10-j M/3-DZA (in the presence of lo-’ ML-homocysteine thiolactone) and challenged in the presence of 20 ug/ml PS (dark bars). Conditions as described in Table 3. Results are expressed as mean percentage histamine released &-SEM for three determinations.
388
DAVID S. BURT et al.
under the same conditions. Hence, heating rat IgE at 56°C induces localized structural changes in the CH3 and CH4 domains of IgE rather than a generalized change in conformation. If as we predict, the peptides used in these studies represent surface accessible regions of the CH3 and CH4 domains, our findings suggest that one or more of the IgE-Fc sequences represented by amino acids 414428, 459472, 491-503 and 542-557 may be part of the heatsensitive cytophilic site on rat IgE. Indeed recent experiments have shown that synthetic peptides corresponding to these four rat IgE sequences, but not others, are able to inhibit the binding of I-125 rat IgE to rat mast cells (Burt and Stanworth, submitted). Antibodies in sera specific for each of the seven IgE sequences studied were able to block, to various extents the binding of IgE to mast cells (Figs 5 and 6). The capacity of antibodies directed against any part of the exposed surface of the CH3 and CH4 domains of rat IgE to inhibit IgE receptor binding suggests that this region of rat IgE is involved in the receptor interaction. This proposal is consistent with data from other laboratories based on the localization of heat-sensitive domains on IgE (Dorrington and Bennich, 1978); the relative enzymatic susceptibility of IgE in solution and bound to its receptor (Perez-Montfort and Metzger, 1982); and the mapping of distances between various regions on receptor-occupied IgE and the mast cell membrane surface (Holowka and Baird, 1983). Furthermore, a mouse monoclonal anti-rat IgE(Fc) recently shown to recognize an epitope in the CH2 domain (Holowka et al., 1985) was unable to inhibit the binding of rat IgE to rat basophilic leukemic cells (Conrad et al., 1983), arguing against the involvement of the CH2 domain in receptor-binding. However: whereas the anti-peptide antisera were able to differentiate non-heat-sensitive between heat-sensitive and sequences (Burt et al., 1986; Fig. 4 in this paper), they were unable to distinguish between sites which may be involved in receptor-binding and other nonreceptor-interacting sites. This result may be explained in terms of stearic effects associated with the binding of anti-peptide antibodies to sequences adjacent to, but not corresponding to the mast cell Similar observations have been receptor-site. reported by Baniyash and Eshhar (1984) using a variety of mouse monoclonal antibodies specific for non-identified conformation determinants within the Fc region of rat IgE. The possibility of stearic interference of the IgE-receptor sites by antibodies in this present study could be tested by using isolated F(ab); and Fab’ fragments of the intact antipeptide antibodies. However, other investigators have found that both intact IgE(Fc)-reactive monoclonal antibodies and their Fab’ fragments are equally effective at inhibiting the binding of IgE to its receptor (Baniyash and Eshhar, 1984). An alternative explanation for the generalised inhibitory effect observed may be that binding of anti-peptide antibodies at
distant sites may induce conformational changes at the receptor-site. This possibility remains to be investigated. Each peptide-specific serum used in this present study was also capable of interacting with receptoroccupied IgE as assessed by their ability to induce the secretion of histamine from IgE-sensitized mast cells (Table 4). The rank order of effectiveness of the anti-peptide sera in stimulating secretion was similar to their efficacies in binding free rat IgE (Burt et al., 1986; Fig. 3(A) in this paper), supporting the proposal that the anti-peptide antibodies in these sera are cross-linking surface IgE on the mast cell membrane. More importantly, these results suggest that surface regions of the CH3 and CH4 domains are exposed on receptor-occupied rat IgE. It could be argued that sequences within these two domains of rat IgE which are not involved in receptor-binding would be accessible to specific antibodies. In such cases, the ability of the same anti-peptide antibodies to also inhibit the binding of rat IgE to mast cells could be explained in terms of the stearic or conformational effects already alluded to. A similar dual activity for monoclonal anti-rat IgE(Fc) antibodies has also been reported (Baniyash and Eshhar, 1984). However, the ability of antibodies directed against sequences, based on our studies to be putative receptor-interacting sites, to interact with receptor-occupied rat IgE has important implications in relation to the orientation of the IgE heavy-chain when bound to its specific receptor on mast cells. Our finding that antibodies directed against most of the CH3-CH4 surface, including those specific for the terminal sequence of the last domain (56&571) can bind to receptor-occupied IgE argues against the existence of a “pocket” receptor for rat IgE, where the CH3-CH4 domains are masked and the IgE molecule sits perpendicular to the membrane surface. That the Fc region of receptor-occupied rat IgE is accessible to proteolytic enzymes (Perez-Montfort and Metzger, 1982) and monoclonal antibodies (Baniyash and Eshhar, 1984) supports this view. In common with other immunoglobulin classes, the structure of rat IgE is such that any sequences comprising the receptor-site (which we predict to involve regions 414428, 459-472, 491-503 and 542-557) would exist in duplicate with corresponding sites on each heavy-chain rotated 180” with respect to one another (Metzger, 1983). Furthermore, the antipeptide antibodies described in this present study would recognize linear sequences which are likewise duplicated on each identical heavy-chain, unlike conventional antibodies which normally recognize conformational determinants which may comprise amino acid residues contributed by both heavychains. This latter property of the anti-peptide antibodies is crucial to the model that we now propose for the IgE-receptor interaction. On the basis of our own data together with that of other groups, we suggest that only one of the identical pair of se-
Rat immunoglobulin
E and rat mast cell interaction
quences postulated to comprise receptor-site residues actually interacts with the receptor, thereby leaving the symmetry-related set accessible and available for combination with its specific anti-peptide antibody. This may occur in one of two ways; either, the receptor contact residues are provided by both heavychains, or alternatively, all receptor contact residues may be derived from a single heavy-chain. In conclusion, irrespective of which of these arrangements is correct, the proposed model predicts a lateral interaction between the receptor and rat IgE along its two-fold axis of symmetry. This idea is consistent with results from Holowka and Baird (1983), and Holowka et al. (1985) which imply that the terminal end of the heavy-chain of rat IgE is closely associated with the mast cell membrane, whilst the remainder of the molecule is not fully extended from the membrane, but rather, disposed at an angle to the surface with the CH2 domains approx. 35-52A away from the membrane. This proposal necessitates that the IgE heavy-chain has some degree of flexibility at the CH2-CH3 interface; a prediction supported by at least one study (Slattery et al., 1985). A similar type of model has recently been suggested for human IgGl bound to its receptor on monocytes (Woof et al., 1986). Acknowledgements-The authors would like to thank Dr John Fox, Macromolecular Service Laboratory, Department of Chemistry, University of Birmingham for assistance with peptide synthesis and Janet Hadley for typing this manuscript. This work was supported by the Wellcome Trust.
REFERENCES
Baniyash M. and Eshhar Z. (1984) Inhibition of IgE binding to mast cells and basophils by monoclonal antibodies to murine IgE. Eur. J. Immun. 14, 799-807. Batchelor K. W. and Stanworth D. R. (1980) A comparison of the histamine-releasing properties of rat pleural and peritoneal mast cells. Immunology 41, 271-278. Burt D. S., Hastings G. Z. and Stanworth D. R. (1986) Use of synthetic peptides in the production and characterization of antibodies directed against predetermined specificities in rat immunoglobulin E. Molec. Immun. 23, 181-191. Burt D. S. and Stanworth D. R. (1983) The effect of ribose and purine modified adenosine analogues on the secretion of histamine from rat mast cells induced by ionophore A23 187. Biochem. Pharmac. 32, 2729-2732. Conrad D. H., Studer E., Gervasoni J. and Mohanakumar T. (1983) Properties of two monoclonal antibodies directed against the Fc and Fab’ regions of rat IgE. Int. Archs Allergy appl. Immun. 70, 352-360. Deisenhofer J. (198 I) Crystallographic refinement and atomic models of a human Fc fragment and its complex with fragment B of protein A from Staphylococcus aureus at 2.9 and 2.8 A resolution. Biochemistry 20, 2361-2370. Dorrington K. J. and Bennich H. H. (1973) Thermally induced structural changes in immunoglobulin E. J. biol. Chem. 248, 8378-8384. Dorrington K. J. and Bennich H. H. (1978) Structurefunction relationships in human immunoglobulin E. Immun. Rev. 41, 3-25.
analysis
389
Evans D. I’., Lewis J. A. and Thompson D. S. (1973) An automated fluorimetric assay for the rapid determination of histamine in biological fluids. Life Sci. 12, 327-336. Goth A,, Adams H. R. and Knoohuizen M. (1971) Phosphatidylserine: selective enhancer of histamine release. Science U.S.A. 173, 103&1035. Hamburger R. N. (1975) Peptide inhibition of PrausnitzKustner reaction. Science 189, 389-390. Hellman L., Pettersson U., Engstrom A., Karlsson T. and Bennich H. H. (1982). Structure and evolution of the heavy-chain from rat immunoglobulin E. Nucl. Acid Res. 10, 6041-6049. Holowka D. and Baird B. (1983) Structural studies on the membrane-bound immunoglobulin E receptor complex. 2. Mapping of distances between sites on IgE and the membrane surface. Biochemistry 22, 3475-3484. Holowka D., Conrad D. H. and Baird B. (1985) Structural of membrane-bound immunoglobulin Emapping receptor complexes: use of monoclonal anti-IgE antibodies to probe the conformation of receptor-bound IgE. Biochemistry 24, 626&6267. Humphrey J. H., Austen K. F. and Rapp H. J. (1963) In vitro studies of reversed anaphylaxis with rat cells. Immunology 6, 226245. Hunter R. (1970) Standardization of the Chloramine-T method of protein iodination. Proc. Sot. exp. Eiol. Med. 133, 989-992. Kulczycki A., Jr, Isersky C. and Metzger H. (1974) The interaction between IgE and rat basophilic leukemia cells. I. Evidence for specific binding of IgE. J. exp. Med. 139, 600-616. Max E. E., Battey J., Ney R., Kirsch I. R. and Leder P. (1982) Duplication and deletion in the human immunoglobulin E genes. Cell 29, 691-699. Merrifield R. E. (1963) Solid phase peptide synthesis. 1. The synthesis of a tetrapeptide. J. Am. them. Sot. 85, 2149-2154. Metzger H. (1983) The receptor on mast cells and related cells with high affinity for IgE. In Contemporary Topics in Molecular Immunology (Edited by Inman F. P. and Kindt T. J.), Vol. 9, pp. 115-145. Plenum Press, New York. Morita Y., Chiang P. K. and Siraganian P. (1981) Effects of inhibitors of transmethylation on histamine release from human basophils. Biochem. Pharmac. 30, 785-791. Neurath A. R., Kent S. B. H. and Strick N. (1982) Specificity of antibodies elicited by a synthetic peptide having a sequence in common with a fragment of a virus protein, the hepatitis B surface antigen. Proc. Natn. Acad. Sci. U.S.A. 79, 7871-7875. Perez-Montfort R. and Metzger H. (1982) Proteolysis of soluble IgE-receptor complexes: localization of sites on IgE which interact with the Fc receptor. Molec. Immun. 19, 113-125. Sachs D. H., Schechter A. N., Eastlake A. and Anifinsen B. (1972) An immunologic approach to the conformational equilibria of peptides. Proc. Nain. Acad. Sci. USA. 69, 3790-3794. Slattery J., Holowka D. and Baird 8. (1985) Segmental flexibility of receptor-bound immunoglobulin E. Biochemistry 24, 7 1(r720. Stanworth D. R. and Kuhns S. W. J. (1965) Quantitative studies on the assay of human sensitizing antibodies (reagins). 1. An examination of factors affecting the accuracy of the Prausnitz-Kustner (P-K) test. Zmmunology 8, 323-344. Stanworth D. R., Humphrey J. H., Bennich H. H. and Johansson S. G. 0. (1967) Specific inhibition of the Prausnitz-Kustner reaction by an atypical human myeloma protein. Lancer ii, 330-332. Woof J. M., Partridge L. J., Jeffries R. and Burton D. R. (1986) Localization of the monocyte-binding region on human immunoglobulin G. Molec. Immun. 23, 319-330.