EP1123317A1 - Immunoglobulin variants - Google Patents

Abstract

An IgE polypeptide, or effective fragment thereof, wherein said polypeptide contains a modification such that it is able to bind to its receptor at the cell surface of an immune cell but unable to bring about the release of immuno-active agents from said cell populations.

Description

IMMUNOGLOBULIN VARIANTS

The invention herein described relates to the identification, characterisation and use, of novel IgE variants

The body has developed many defences against invasion by bacteria and parasitic organisms.

A general term to cover a number of distinct cell types intimately involved in both a humoral and cellular defence mechanisms are the white blood cells. Each white blood cell type has a separate role to play in an immune system.

Specialised inflammatory cells include eosinophils, basophils and mast cells. Eosinophils and basophils are circulating white blood cells which carry substances ( ie histamine, proteases) that degrade infectious agents. When presented with an antigenic substance these cells undergo degranulation, which results in the release of these inflammatory chemicals and results in the symptoms associated with inflammation and allergy. Mast cells are non- circulating basophils and are found in the lungs, skin, tongue and epithelial linings of nose and intestinal tract. The mast cell is the major cell type responsible for allergy.

Moreover the body has mechanisms that allow it to react to the presence of a foreign substance (the allergen). For example, an allergen may be inhaled, eaten or injected (i.e. via a sting) into the body and results in a series of cellular and humoral immunological events that manifest in allergic response. An allergic response to a substance can be divided into 3 stages. Stage 1 involves the first contact of the allergic substance with the immune system, so called sensitisation. No allergic response is generated but the immune system is primed for a subsequent contact with the allergen. Sensitisation involves the degradation of the allergen by macrophages and the display of allergen peptide fragments, via major histo-compatibility complex, on T- lymphocytes. T-lymphocytes are activated to proliferate and produce interleukin-4 (a growth factor which promotes the maturation, proliferation of B-lymphocytes and class switching to IgE). B-lymphocytes secrete immunoglobulin E (IgE) antibodies specific to the allergen peptide epitope presented by the T-lymphocyte. The secreted specific IgE antibodies function as ligands for IgE-receptors on circulating basophils and tissue mast cells. These cells are found in close proximity to blood vessels.

Stage 2 involves a later encounter of the allergen with the immune system. When the allergen comes into contact with a specific IgE antibody presented by a mast cell or basophil a signal transduction pathway is activated which results in the release of substances (i.e. histamine, prostaglandins, proteases, chemokines) from granules. The release of these substances triggers the allergic response. Clearly any therapeutic treatment that can interfere or prevent these events will reduce the severity of an allergic attack.

Finally, stage 3 of the allergic response is characterised by prolonged immune activation. Mast cells synthesise a range of molecules that induce the migration of basophils and eosinophils to the site of inflammation to maintain an inflammatory state and can result in tissue damage during late phase. It is apparent that IgE is a pivotal molecule in bringing about an allergic response. Notably IgE binds to its corresponding receptors ( FcεRI and FcεRII) on cells of the immune system and then, in the presence of appropriate antigen (allergen), the occupancy of the IgE receptor with IgE, following aggregation of a number of said receptors, results in the initiation of an intracellular cascade culminating in the secretion of m |[ substances.

This two phase event: i) binding; and ii) intracellular response leading to secretion, is brought about by the interaction of IgE with its receptor in the presence of allergen.

Typically, an allergic response is an appropriate and desirable way in which the body responds to infection. However, there are instances when it is desirable to curb this response. For example, and not by way of limitation, where the relative seriousness of the infection is far outweighed by the presence and/or scale of the inflammatory response such as in the case of hay fever, or where the scale of the response is far more serious than the corresponding trigger such as in the case of asthma. In either of these two cases, for example, it is often important to be able to scale down, or switch off, the immune response. This could be undertaken by targetting the cells responsible for the inflammatory reponse with a suitable agent and then selectively destroying the cells.

In our studies, we have identified an agent that binds, selectively, and with high affinity, to a marker for the aforementioned cells and so we have identified an agent for use in the targeted destruction of said cells. The agent is a modified IgE molecule and the marker is its corresponding high affinity receptor.

It is therefore an object of the invention to provide a modified IgE molecule that is capable of binding exclusively to the high affinity receptor.

The receptor targets for IgE are presented by various cells that participate in an inflammatory response ( inflammatory cells). For example low-affinity receptors (FcεRII) are found on various inflammatory cells including macrophages, eosinophils and platelets. High affinity receptors (FcεRI) are found predominantly on mast cells and basophils. Most IgE molecules are cell bound to tissue mast cells and localised to the eyes, lungs, skin and intestine. As mentioned above, in stage 2, the interaction between allergen and IgE at the mast and/or basophil cell surface leads to a number of intra- cellular events which manifest themselves as the allergic response. Clearly if one could interfere with the binding of IgE to its cognate receptor and/or prevent the signal transduction pathway resulting from said binding, it would be possible to inhibit the release of molecules that involve an inflammatory response.

It is therefore an object of the invention to provide a modified IgE molecule that is capable of performing only one of the above two phases referred to in its mode of activation ( ie it is capable of binding to its corresponding high- affinity receptor but unable to recognise cells expressing the low-affinity receptor or to initiate the subsequent intracellular response).

In its broadest aspect there is provided according to the present invention there is provided an IgE polypeptide, or effective fragment thereof, wherein said polypeptide contains at least one modification such that the effect of its receptor binding is modified.

In a first embodiment of the invention there is provided an IgE polypeptide, or effective fragment thereof, wherein said polypeptide contains a modification such that it is able to bind to its receptor at the cell surface of an immune cell but unable to bring about the release of immuno-active agents from said cell populations.

In a second alternative or additional embodiment of the invention there is provided an IgE polypeptide, or effective fragment thereof, wherein said polypeptide contains a modification such that it is able to bind predominantly to a high affinity IgE receptor.

In a third alternative or additional embodiment of the invention there is provided an IgE polypeptide, or effective fragment thereof, wherein said polypeptide contains two modifications such that it is able to bind at the site of a first modification predominantly to a high affinity IgE receptor and is able to bind at the site of its second modification to its receptor at the cell surface of an immune cell and deliver a therapeutic agent.

In the instance where said polypeptide binds to both FcεRI and FcεRII, ideally said polypeptide binds to FcεRII has lost the affinity of its natural counterpart, due to has a modified (increased) kD for said FcεRII.

Reference herein to the word predominantly comprises reference to an IgE polypeptide that binds either only to the high affinity receptor, FcεRI, or preferentially to FcεRI and with little affinity to FcεRII, typically having regard to the normal binding constant for wild type IgE and FcεRII. The embodiments of the invention illustrate the highly specific nature of the modifications and their effect on receptors. In the case of plural modifications the combined effect can be different to the individual effects. In the third embodiment above, the combined construct can be employed for safe delivery of (immuno)-toxins to mast cells or basophils in order to destroy them, whereby it is useful in eg mast cell malignancies such as mastocytomas and mast cell or basophilic leukaemias. In addition this IgE variant has an application in the selective immuno-magnetic purification of cells expressing the high-affinity receptor without activating receptor mediated signalling.

In the first embodiment of the invention, preferably said IgE polypeptide, or effective fragment thereof, is modified by addition, deletion, substitution, or inversion, of at least part of said IgE polypeptide such that binding of said polypeptide to its receptor is not prevented but that release of agents, such as pro-inflammatory agents, from said cell is significantly reduced or inhibited.

Preferably said modification comprises the addition, deletion, or substitution of at least one amino acid residue.

More preferably said modification is the substitution of PRO 333 of human IgE by a glycine residue ( Human IgE, and the location of PRO 333 is described in Figure 6), Pro333 is invariant in all known species, it cannot be deleted, this destroys receptor binding.

More preferably said modification is the replacement of PRO 333 with a glycine amino acid residue or the replacement of PRO333 with a modified amino- acid. In an alternative or additional embodiment of the invention said IgE polypeptide, or said effective fragment thereof, is modified by substitution, or inversion of at least one amino acid residue of said IgE polypeptide such that binding of said polypeptide is predominantly to a high affinity receptor.

Preferably said modification is deletion or substitution of LYS 352 of human IgE ( human IgE and the location of LYS 352 is identified in Figure 7), or to the homologous amino acid residue in a non-human IgE.

More preferably said modification is a substitution of LYS 352 with a glycine amino acid residue; or is a substitution of LYS352 with a modified amino acid.

In an alternative or additional embodiment of the invention said IgE polypeptide, or said effective fragment thereof, is modified by substitution, of at least two amino acid residues of said IgE polypeptide such that binding of said polypeptide is predominantly to a high affinity receptor, and additionally that release of agents, such as pro-inflammatory agents, from said cell is significantly reduced or inhibited.

Preferably said modification is substitution of PRO333 and LYS 352 of human IgE, (the location of PRO333 and LYS 352 are identified in Figure 8), or to the homologous amino acid residue in a non-human IgE.

More preferably said modification is a substitution of PRO333 and LYS 352 with a glycine amino acid residue; or is a substitution of PRO333 and LYS352 with an alanine amino acid residue; or is a substitution of PRO333 and LYS352 with a modified amino acid. It will be appreciated that by the present invention there is a synergism in terms of the combined result in inhibiting allergic response and moreover in delivery of agents such as immuno-toxins to cells expressing high affinity receptors such as leukaemia cells.

It will be apparent to one skilled in the art that modified amino acids include, and not by way of limitation, 4-hydroxyproline, 5 -hydroxy lysine, N⁶ - acetyllysine, N⁶ - methyllysine, N⁶, N⁶ dimethyllysine, N⁶ N ⁶ N ⁶ trimethyllysine, cyclohexyalanine, D-amino acids, omithine. The incorporation of modified amino acids may confer advantageous properties on IgE polypeptides that bind receptors presented by inflammatory cells. For example, the incorporation of modified amino acids may increase the affinity of the IgE polypeptide for its binding site, or, the modified amino acids may confer increased in vivo stability on the polypeptide thus allowing a decrease in the effective amount of therapeutic IgE administered to a patient.

Preferably said receptor is a high affinity IgE receptor; more preferably said high affinity receptor is FcεRI.

It will be apparent to one skilled in the art that the IgE variant molecule of the invention has potential as a therapeutic agent to treat a range of antigen induced diseases. For example, and not by way of limitation, asthma, allergy (typically anaphylaxis, hayfever).

By way of explanation, we propose our modified IgE molecule inhibits the aggregation of neighbouring IgE/receptor complexes in response to antigen and so inhibits the formation of, what may be termed, an aggregation signal; since release of pro-inflammatory molecules only occurs when bound IgE aggregates in response to a specific antigen (either via the antigen, lectins or anti IgE antibodies). It will be evident that the polypeptides herein described, which predominantly prevent this aggregation event, inhibit the release of mediators of inflammation.

Alternatively or additionally, the IgE molecule of the invention has potential as a therapeutic agent to selectively deliver an agent, for example a toxin or a signal stimulating cell apoptosis to cells predominantly expressing high affinity receptors. For example and not by way of limitation, leukaemia cells.

An example of a mast cell malignancy is mastocytosis, which is a disorder of both children and adults and is caused by the overproduction of mast cells. Over production of mast cells can be of two forms, cutaneous and systemic. The former is the more common and occurs when mast cells infiltrate the skin. Systemic mastocytosis is caused by the accumulation of mast cells in the liver, spleen, bone marrow and small intestine. Mast cell overproduction results in excessive secretion of pro-inflammatory agents leading to bone pain, nausea , ulcers, skin lesions, and anaphylaxis. Currently, treatments for mastocytosis relate to alleviating these symptoms rather than treating the over-production of mast cells. In cases where mastocytosis is malignant, conventional chemotherapy is administered. It would therefore be desirable to target cells that predominantly express high affinity receptors with , for example a cytotoxin, and preferably, an immunotoxin, or a molecule inducing cell apoptosis, to reduce, for example, mast cell number. Furthermore, it can be utilised for the selective targeting of toxins or agents stimulating apoptosis to treat of malignancies of basophils such as basophilic leukaemias. Preferably said IgE, or said effective fragment thereof, is derived from a monoclonal antibody; more preferably said monoclonal antibody is humanised.

According to a further aspect of the invention there is provided an isolated DNA molecule, or fragment thereof, encoding an IgE polypeptide, or an effective fragment thereof, as herein before defined.

Preferably said DNA molecule is genomic DNA; preferably said molecule encodes human IgE, as represented in Figures 6, 7 and/or 8.

Preferably said DNA molecule is synthetically derived. Reference herein to the term synthetic comprises reference to an oligonucleotide manufactured using conventional DNA oligo-synthesizing technology.

In yet a further preferred embodiment of the invention said DNA molecule is modified by substitution, of at least one nucleic acid base pair.

It will be apparent to one skilled in the art that conventional genetic engineering techniques may be undertaken to produce said modification. For example, and not by way of limitation, any of the following techniques may be used: restriction digestion may be undertaken using selected restriction enzymes; and/or polymerase chain reaction methods may be undertaken to amplify selected regions of DNA molecules encoding said IgE polypeptides; or incoφoration of point-mutations may be undertaken using both PCR methodology and/or conventional methods to introduce point- mutations and/or deletions up-stream of amino acid residue 342. In yet a third aspect of the invention there is provided a vector containing a DNA molecule encoding an IgE polypeptide according to any preceding aspect or embodiment of the invention.

In a preferred embodiment of the invention said vector is provided with means to recombinantly manufacture the IgE polypeptide of the invention.

It will be apparent to one skilled in the art that said vector will be provided with promoter sequences that facilitate the constitutive and/or regulated expression of the DNA sequence encoding said IgE polypeptide. Further said promoter sequences will be selected such that expression in eukaryotic and/or prokaryotic cells is facilitated. In addition said vector is provided with polyadenylation signals and/or termination signals that optimise expression of said vector in either a eukaryotic cell and/or prokaryotic cell.

Advantageously the above described vectors are provided with necessary selectable markers that will facilitate their selection in a eukaryotic or prokaryotic cell(s).

In yet a further aspect of the invention there is provided a method to recombinantly manufacture IgE polypeptides according to the invention comprising:

i. providing a cell transformed transfected with a vector according to the invention;

ii. growing said cell in conditions conducive to the manufacture of said polypeptide; and iii. purifying said polypeptide from said cell, or its growth environment, by conventional means.

In a preferred method of the invention said vector encodes, and thus said recombinant polypeptide is provided with, a secretion signal to facilitate purification of said polypeptide.

In yet still a further aspect of the invention there is provided a therapeutic composition comprising an IgE polypeptide according to the invention including an excipient, diluant or carrier. Typically, said composition is for use in the treatment of allergen mediated disease.

In a further aspect of a further embodiment of the invention there is provided a therapeutic composition comprising: an IgE polypeptide according to the invention; including, in association therewith, ideally coupled or joined thereto, a cytotoxic agent, and further comprising, an excipient, diluent or carrier. Typically, said composition is for use in the treatment of a blood cell disorder that would benefit from exposure to said cytotoxic or apoptotic agent.

Preferably said cytotoxic agent is an immunotoxin or an agent stimulating apoptosis.

Preferably a composition is in unit dosage form.

In yet still a further aspect of the invention there is provided a method to treat an allergen mediated disease or blood disorder comprising:

i. providing a therapeutic composition according to the invention ; ii. administering said composition to a human/animal; and

iii. monitoring and modulating said therapeutic treatment according to the response of said human/animal, to same.

It will be apparent to one skilled in the art that the therapeutic composition can be provided in the form of an oral or nasal spray, an aerosol, suspension, emulsion, and/or eye drop fluid. Alternatively the therapeutic composition may be provided in tablet form or intra-venous infusion.

Alternative delivery means include inhalers or, nebulisers; or syringe whereby direct intravenous injection of the composition either at the site of inflammation or at a distance from the site of inflammation, may be undertaken.

It will also be apparent that the therapeutic composition may be effective at preventing and/or alleviating allergic conditions in animals other than humans, for example, and not by way of limitation, family pets, livestock, horses.

In a further aspect of the invention there is provided a therapeutic veterinary composition for the treatment of animals comprising an IgE polypeptide according to the invention including an excipient, diluant or carrier. Typically, said composition is for use in the treatment of allergen mediated disease of e.g. dogs or horses.

In a further aspect of the invention there is provided a therapeutic veterinary composition for the treatment of animals comprising an IgE polypeptide according to the invention including an excipient, diluant or carrier and a cytotoxic agent. Typically said composition is for use in the treatment of a blood cell disorder which would benefit from exposure to said cytotoxic/apoptotic agent. Preferably said cytotoxic agent is an immunotoxin

According to a further aspect of the invention there is provided a kit for treatment of allergen mediated disease or blood disorders comprising: said therapeutic IgE composition; and delivery means to facilitate the administration of the therapeutic composition.

Methods of delivery of said therapeutic composition are common in the art and include nasal spray devices, eye drop applicators, inhalers, nebulisers, intravenous infusion and direct injection via hypodermic needles.

It will be apparent to those skilled in the art that the therapeutic composition can be used either prophylactically or curatively. In the former instance, the therapeutic composition can be administered before a predictable allergic response is shown with a view to blocking a subsequent allergic reaction; whereas in the latter instance, the therapeutic composition can be administered after an allergic response has been initiated with a view to preventing the further release of inflammatory agents. In yet a further aspect of the invention there is provided a method for the selection of cells expressing high affinity receptors for IgE peptides comprising;

i) contacting a cell sample with an IgE polypeptide according to the invention; ii) incubation of said cell sample for a sufficient period to allow selection of cells expressing high affinity receptors; and iii) isolating the IgE/ cell complex by conventional methods.

It will be apparent that the method provides the means to isolate cells expressing high affinity receptors for IgE. Conventional methods include but are not limited to Fluoresence Activated Cell Sorting ( FACS) or immuno- magnetic purification.

In a preferred method of the invention said cells are inflammatory cells, preferably mast cells and/or basophils and/or a subclass of eosinophils expressing the high affinity receptor.

Embodiments of the invention will now be described, by example only, and with reference to the following figures, tables, materials and methods wherein:

Figure 1 is a graphical representation of human IgE identifying selected residues in the Cε3 region;

Figure 2 shows the interaction of wild type and mutant IgE with sFcεRIα. (A) shows the binding of IgE to immobilised monoclonal antibody LE27 followed by binding of sFcε Rlα. (B) interaction of IgE with varying sFcεRIα concentration.(C) Shows sensograms of the association phase of sFcεRIα to the surface of bound IgE. (D) shows sensograms of the dissociation phase of the FcεRIα/IgE interaction;

Figure 3 (A) and (B) represents the association constants of IgE variants with FcεRIα; Figure 4A represents the fractional occupancy of IgE binding sites of hlgE variants. Figure 4B represents the affinity and dissociation of hlgE variants for FcεRIα;

Figure 5 represents the ability of IgE variants to mediate antigen induced pro-inflammatory release from RBL-J41 cells;

Figures 6, 7 and 8 represent part of human IgE indicating the locations of PRO333 and LYS352;

Figure 9 shows the effect of mutations within the binding profiles of hlgE Fe on the IgE FcεRII interactions;

Figure 10 is a graphical representation of the involvement of various inflammatory cells in an allergic response;

Figure 11 is a flow diagram showing the establishment of a stable cell line expressing human FcεRIα;

Figure 12 represents a deletion map of human IgE defining the minimal binding domain to FcεRI and FcεRII;

Figure 13 represents a peptide modelled on the A-B loop of Human Cε3;

Figure 14 represents structural models of IgE domains;

Table 1 shows rationale for site directed mutagenesis of residues in the hlgE Cε3 domain. Engineered mutations are listed (in regions targeted) including additional mutations* introduced during generation (see text); and Table 2 shows kinetic constants for the human IgE-FcεRIα interaction. Kinetic constants, calculated using the pseudo 1^st order and empirical SPR analysis, are shown. A summary of previously determined values is included for comparison. Pseudo 1^st order analysis of the kinetics of native hlgE binding to the cellular receptor and SPR data produced values that are in excellent agreement and also applicable to IgE mutants P333A/G. More complex binding kinetics became apparent following a detailed analysis of SPR data (see text). Also included are the kinetic constants for several key hlgE Fc fragments that have been expressed as GST fusion proteins.

Materials and Methods

(a) Gene constructs and site directed mutagenesis. The numbering scheme for the and FcεRIα a.a. has been maintained (12-16). The sequences of rodent and hlgE are based on our own data, which agree with those published by Kabat et al. (35). The procedure for the construction of mutant recombinant IgE molecules has been described previously (14); mutagenic primers were designed to incoφorate the minimum base pair substitutions for the generation of the desired a.a. substitution(s), primers ranged between 15-55bp in length.

For the construction of the soluble form of FcεRIα chain extracellular (EC) domains a modified Pichia pastoris pPIC9 expression vector (Invitrogen) was used. This contained the Tn903 kan^r gene inserted into the Nael site of pPIC9 which conferred resistance to G418 (designated pPIC9K). The gene fragment coding for the extracellular region of FcεRIα (residues 1- 172) was generated by PCR as described above using the FcεRIα cDNA cloned into pGEM3Zf as a template (5' Ria-1 GGCCGGGAΔ1ICATGGTCCCTCAGAAACCTAAG, 3' Ria-172 GGATCCGCGGCCGCTCAAGCTTTTATTAC A GTA ATGTT) The resultant 542bp fragment was blunt cloned into the Smal site of pUC18 for gene sequencing and then subcloned into the pPIC9K expression vector using EcoRI and Notl to generate pPIC9k-αEC. The gene was sequenced again prior to transformation into the Pichia pastoris strain GS115 to ensure the DNA was in frame for expression. Bacterial strains JM109 and TGI were used for the propagation of plasmid DNA.

(b) Gene expression/cell culture. Plasmids containing mutant N_NpHε constructs (approx. 25 μg) were linearised using Pvul and transfected into the J558L plasmocytoma cell line by electroporation. J558L cells were cultured in Dulbecco's Modified Eagles Medium (DMEM) (10% fetal calf serum, penicillin (100 U/ml)/streptomycin (100 μg/ml) and gentamicin (50 μg/ml)). Exponentially growing cells were isolated by centrifugation, washed twice with DMEM in the absence of any supplements, and resuspended to give 10⁷ cells/0.8 ml. The linearised DΝA was mixed with the 0.8 ml cell suspension, transferred into a 0.4 cm gap electroporation cuvette (Biorad) and placed on ice for 10 min. A single 250 V, 960 μF pulse was applied (Biorad Genepulser), and the cells were returned to the ice for 10 min. 20 ml of fresh medium was added and the cells were plated into 24 well plates (0.5 ml/well). After 48 hours spent medium was carefully aspirated and selection medium was added (containing mycophenolic acid (1 μg/ml), xanthine (250 μg/ml) and hypoxanthine (15 μg/ml)). This procedure was repeated every 3-4 days, and clones were visible after 15 days. Cloning, using limiting dilution and screening by a sandwich enzyme linked immunosorbant assay (ELISA), isolated high secreting clones. A mouse anti- human IgE capture antibody LE27, (anti-Cε4, gift of Prof. B. Stadler, Univ. Bern, Switzerland) and a commercial HRP-anti IgE detection antiserum (Dako) were used for this procedure. Protein expression was also analyzed by SDS-PAGE and Western blotting using the same commercial HRP conjugated anti-human ε chain antibody (Dako). High secretion variants were grown in stirred, aerated vessels (1-4 litres of culture) until the cells were confluent. The supernatant was collected, filtered and concentrated 10 x using a Millipore Minitan System with a 30 kDa MW cut off.

For the production of soluble, secreted sFcεRIα the pPIC9K-αEC vector (10-20μg) was linearized using Sad and transformed by electroporation into the Pichia pastoris strain GS115 (Invitrogen). The transformation conditions used were as recommended by Invitrogen, and involved a single pulse of 1500 N, 25 μF, 200 Ohms applied to 90 μl of DΝA/cell suspension in a 0.2 cm gap electroporation cuvette (Biorad) (36). The transformed cells were first selected by plating onto MD plates in order to detect HIS⁺ transformants, and then high copy number clones were isolated by plating the HIS+ transformants onto YPD plates containing increasing concentrations of G418 (0.5, 1.0, 1.5, 2.0 mg/ml). G418 resistant colonies were isolated and sFcεRIα expression was assessed by small scale expression with methanol induction using BMGY and BMMY media in 250 ml baffled flasks. Supernatants were analysed by SDS-PAGE and ECL detection using a monoclonal anti hFcεRIα chain antibody 15.1 (gift of J. P. Kinet, Harvard, Boston, USA). For the large-scale production of sFcεRIα the fermentation was scaled up into 2.5 1 baffled flasks containing 500 ml of media. After 2-3 days of methanol induction, the supernatant was isolated by centrifugation, filtered and concentrated 10-20 x using a stirred ultra- filtration cell with a 10 kDa MW cut off (Amicon). In an analogous way, a control cell line was generated by transformation with the pPIC9K vector in the absence of the FcεRIα gene.

(c) Detection of Recombinant IgE by ELISA. A sandwich ELISA was used for the detection and semi-quantification of recombinant IgE present in cell supernatants. On the day prior to analysis, a 96 well assay plate (Falcon) was coated with a mouse monoclonal anti hlgE Cε4 antibody (LE27) at a concentration of 10 μg/ml in 0.1 M carbonate/bicarbonate buffer pH 9.6 (100 μl/well). The plate was incubated overnight at room temp, then washed three times with wash buffer (0.05% Tween/PBS) followed by 200 μl of a blocking solution (5% milk powder/0.05% Tween PBS). After a 2 hour incubation at 37°C, the plate was washed and a number of dilutions of the samples were added, 100 μl/well. A standard graph was generated using commercially available recombinant IgE (Serotec, 0.001-50 μg/ml). Following a second 2 hour incubation at 37°C the plate was washed and 100 μl anti-hlgE HRP conjugated antiserum (DAKO, 1/500 dilution in blocking buffer) were added to each well. Following a further 2 hour incubation at 37°C the plate was washed and HRP activity was detected by adding 100 μl/ml of OPD buffer (0.4 mg 0-phenylenediamine (OPD), 12 μl H₂O₂ in 0.2 M citrate/phosphate buffer pH 5.0). After 5-10 minutes the color reaction was stopped by adding 50 μl/well 2M H₂SO . The absorbance (OD) at 492nm was determined using an ELISA plate reader (Milenia Kinetic Analyzer), and test samples were calculated from the constructed absorbance vs. cone, graph. Appropriate controls were included at each stage of the assay in order to correct for background.

(d) Recombinant/chimeric IgE and sFcεRIα purification Each recombinant IgE protein was purified from tissue culture supernatant employing established procedures (Brueggemann et al., 1987). Purified antibodies were dialyzed extensively against PBS to remove hapten from the variable region, filtered through a 0.45 μm filter (Amicon) and concentrated using ultrafiltration (Amicon Centriplus 50). sFcεRIα was affinity adsorbed onto rat IgE coupled to Sepharose

(Sepharose, Pharmacia). The sFcεRIα was bound to the resin slurry by incubation overnight on a rotating platform at 4°C and eluted with 0.2 M glycine buffer pH 2.8, followed by immediate neutralisation using 1 M Tris- HC1 pH 8.0.

The IgE variants and sFcεRIαwere quantified on the BIAcore 2000 system using BIAconcentration software. Known concentrations of IgE (Serotec) and an affinity purified preparation of FcεRIα were employed to construct a dose/response curve on the biosensor. The samples were analyzed for a number of dilutions and the concentrations were calculated. A description of the BIAcore system is described in section (f) below.

(e) IgE mediated stimulus secretion coupling in rat basophilic leukemia cells transfected with the α-chain of the human high affinity receptor. Recombinant IgE molecules were assessed for their ability to bind to and aggregate the hFcεRIα chain constitutively expressed on the surface of the RBL-J41 cell line, which also expresses the subunits of rat FcεRI (36). RBL-J41 and the parental RBL-2H3.1 (37) cell lines were used to evaluate the ability of recombinant IgE variants to support the release of cellular mediators, monitored by the secretion of [³H] 5-hydroxytryptamine [5-HT] as described previously (28). Since peak secretion via human and rat receptors varied by up to 10 % between experiments, the data set was normalised to the peak secretion effected via triggering through the endogenous rodent receptor, which was included as an internal control in each experiment. Each hlgE variant (1 μg/ml) was assessed for its ability to mediate stimulus secretion coupling via the hFcεRIα chain using a hapten crosslinking agent (NIP-HSA 0.1-10000 ng/ml), or a commercial anti-ε- chain antiserum (DAKO, dil. 1/100-1/5000).

(f) Ligand binding studies using Surface Plasmon Resonance (SPR). A biosensor-based analytical system (BIAcore Inc, Uppsala, Sweden) was used to analyze the kinetics of interaction between IgE variants and sFcεRIα. Binding events to surface immobilized sites are detected in real-time by change in local refractive index, which is measured in arbitrary units (termed resonance units, RU). A mouse monoclonal anti hlgE antibody LE27 [-13, 000 resonance units (RU) (13 ng/mm²)] was immobilized to the CM5 sensor surface via amino groups and used to capture hlgE. Binding studies were carried out at 25°C. For each IgE variant analysed, seven experimental cycles were performed, including the control cycle. Each cycle consisted of binding a constant amount of the IgE variant followed by a series of fixed concentrations of sFcεRIα. The IgE variants were diluted in HBS (10 mM HEPES pH 7.4, 3.4 mM EDTA, 150 mM NaCl, 0.005% (v/v) surfactant P20) to a final concentration of 10 μg/ml and bound to the capture antibody to provide between 600 and 3300 (0.6 - 3.3 ng/mm²) of surface-bound ligand. sFcεRIα, at concentrations 100 to 600 nM were passed over the captured IgE variant at a flow rate of 15 μl/ml for 3 min. Dissociation was monitored for a further 3 hours. 10 μl 10 mM glycine-HCl pH 2.2 was used to remove IgE- sFcεRIα complexes and to prepare the surface for the next analytical cycle. The sFcεRIα was added as secreted supernatant from tissue culture, and medium from the control cell line was included in a control experiment. All sensorgrams were corrected for dissociation of IgE from the LE27 antibody by subtraction of the control curve. (g) Binding analysis based on an assumed 1:1 pseudo-first order interaction. The apparent rate constants k_a and kj for the IgE/sFcεRIα interaction were calculated using BIAevaluation software (BIAcore, Uppsala, Sweden) in two alternative analysis strategies. First, for the determination of effective apparent rate and equilibrium constants for the overall binding process, BIAevaluation v2.1 was used (38). The apparent dissociation rate constant was determined by selection of the linear parts of a logarithmic plot In (R(to)/R) vs. t of the dissociation sensorgrams (where R(t) denotes the sensor signal). Based on the relationship In (R(to)/R) = k_dx(t-t₀), the slope of the data subset gives an apparent dissociation rate constant. Similarly, an apparent association rate constant was determined by taking the slope of a time derivative plot dR/dt vs. t of the association phase, which obeys k_axc + kj in a pseudo-first order reaction (38). For a more detailed global kinetic analysis with the 1:1 pseudo-first order model, BIAevaluation v3.0 was used. The complete set of dissociation phases over the entire 8000 sec of observation time were fitted globally with the pseudo-first order rate equation under the constraint to give a single apparent dissociation rate constant. The set of association phase data were fitted over the entire observation time using the pseudo-first order rate equation, calculating the apparent association rate constant.

(h) Empirical comparative analysis of the binding of different IgE variants. The observed binding kinetics are biphasic, (i.e. the dissociation curves show initially a phase with a signal decrease steeper than the best-fit single exponential, followed by a phase of slower dissociation). Since this does not appear to be due to mass transport limitations or artifacts of immobilization (see below), this indicates a more complex interaction scheme. In addition, dissociation and rebinding processes of IgE from the capturing antibody within the gel matrix and along the sensor surface causes small signal drifts, complicating the observed kinetic process. In particular, the correction for leakage of the captured IgE and IgE-sFcεrlα is not precise and reproducible enough to allow a reliable kinetic analysis on a level detailed enough to discriminate between possible binding mechanisms underlying the biphasic kinetics. Therefore, an empirical and robust semi- quantitative analysis was performed, which describes general trends of the association and dissociation kinetics, under conditions used in the experiments. Since the following analysis obviously does not describe the detailed process of the binding of IgE to its receptor, the obtained estimates cannot be inteφreted as bimolecular rate constants. However, they allow a comparison of the effects of mutations on the overall characteristics of the binding process. Three empirical measures for affinity and binding kinetics were applied.

First, the fractional occupation of IgE binding sites after an injection of 600 nM sFcεRIα for 180 sec was measured. To eliminate contributions from non-specific binding and sample refractive index, the signal increase due to bound sFcεRIα was measured immediately after injection of sFcεRIα. This value was divided by the signal increase during capture of IgE. This ratio is a measure of the fractional occupation of binding sites. Since the signal of the 600 nM injection after 180 sec is close to steady state plateau binding, the fractional binding can be regarded as a gross empirical measure of the affinity of IgE/sFcεRIα interaction, and can be compared for the different mutations. Second, the association kinetics during the first 100 sec during the injection of sFcεRIα was fitted to a single exponential in order to obtain an empirical binding rate Ic^_s dependent on the sFcεRIα concentration. These curves k_{o S}(c) can serve as a model-free empirical measure for kinetic properties of the interaction in the association phase. Third, the first 300 sec of the dissociation kinetics were analysed using a single exponential model R(t) = R(to)^χexp[k_d*x (t-tn)]. The empirical dissociation rate constant kj* may not be identical to a true molecular dissociation rate constant, but gives an empirical comparative measure of the dissociation properties of the mutants. In principle, this empirical approach could be extended to a purely descriptive double exponential analysis. However, the reproducibility of IgE leakage, the errors involved in double exponential modeling, due to cross-correlation of the parameters and selection of the data subset included in the analysis, prohibit the empirical comparison of small effects of the mutations on a level of detail higher than in the single exponential approach described here.

i) Generation of rodent cell lines expression the ligand binding domain of the high-affinity receptor for human IgE. Rat basophilic leukaemia cell lines (RBL) expressing the human (h) α-chain of the FcεRI complex were engineered using as a host cell line a high secreting variant of the rat RBL 2H3 cell line [8], which expresses a functional receptor complex for rodent IgE. The h FcεRI α-chain gene was subcloned from pUC19 into the multiple cloning site of the vector pcDNA3 which supports constitutive expression of recombinant proteins in mammalian cells. Correct insertion was confirmed by gene sequencing. The plasmid containing the h FcεRI α- chain gene was transfected by electroporation into the RBL-2H3 cells [9] and is expressed as a functional unit with the rodent receptor on the cell surface. The generation and characterisation of the RBL 2/2/C cell line, which supports dexamethazone inducible expression of h FcεRI α and the characterisation of IgE binding and secretory responses in native and transfected cells, has been described in earlier publications [8,9, 11,12].

j) Identification of the high- and low-affinity receptor binding site in h IgE. The methodology has been described in earlier publications [15,14].

FACS Analysis of the hlgE variant FcεRII interaction

The RPMI-8866 cell lme was maintained as described previously (Meisher et al 1994) On the day pnor to ana sis the cells were seeded m fresh medium (RPMI 1640, 10% FCS, Penicillin (100 U/ml)/Streptomycm (100 μg ml). Gentamicin (50 μg/ml)) in order to isolate cells m die exponential phase of growth on the day of the assa\ This procedure was to standardise die surface expression level of hFcεRII which has been shown to van' with cell cycle The cells were isolated by centnfugauon and washed 3 times using wash buffer (1% FCS/0.1% Sodium Azide D-PBS). this any influence of soluble FcεRII which is released mto the supernatant during the culture of this cell line Tlie cells were resuspended to a density of 5 x 10° cells/ml, to 100 μl of ceil suspension 20 μl of recombinant hlgE. control (mouse IgE (SPE-7, Sigma)). MHM6 ami liFcεRII mouse monoclonal anubodv (Dako) or the mouse IgGi control (Dako) was added Tlie 20 μl contained recombinant chimeπc hlgE. concentration range 0 25-2 5 μg, or the other antibodies at 1 μg The cells were incubated for 30 minutes on ice. isolated by centnfugauon and ashed with 1 ml of wash buffer, this procedure was repeated a further 2 times and tlie cells were resuspended in lOOμl wash buffer Fluorescent labellmg was achieλ ed using FITC labelled anti mouse λ light cham (1/ 100) (for hlgE chmiera and mouse IgE) and FiTC labelled anti mouse IgGi (1/200) (for MHNI6 and the IgGi lsot pe control) Following 30 minutes on ice the cells were washed as preuoush and resuspended in 200 μl of wash buffer for analysis using a Coulter Eucs Elite

flow cvtometcr and Datamate (Dako) software Unstained and isotvpe control cells were examined ui order to determine tlie le\el of background fluorescence Tlie median of log channels as used for a comparison of the FcεRII binding lc\els for tlie different hlgE molecules the binding was evaluated in three independent experiments and the mean calculated -/- SEλl RESULTS

(a) Design of mutant forms of IgE. The goal of this study was to gain information concerning residues in human IgE that influence the interaction with FcεRI and FcεRII in the context of a model we proposed for a planar and bent conformation of IgE (15). Table 1 provides an overview of the rationale for each of the substitutions. Additional mutations to those engineered were also introduced in some constructs, due to lack of proof reading ability of the polymerase, and their effects were evaluated.

(b) Expression of variant forms of IgE. Plasmids encoding native and mutant form of IgE-N_NP-Hε were stably expressed in J558L plasmocytoma cells using established procedures (14). High secreting clones were selected by cloning at limiting dilutions and screening by ELISA. Expression levels varied between 0.1 - 30 μg/ml. Purified antibodies displayed identical characteristics in SDS-PAGE/ECL analysis shown in previous studies (30, data not shown).

(c) Expression of sFcεRIα in yeast Pichia pastoris. The data for sFcεRIα production in the methylotropic yeast Pichia pastoris will be presented elsewhere (Cain et al, manuscript in preparation). Following affinity purification, a diffuse band of 45-55 KDa, corresponding to the heterogeneously glycosylated form of FcεRIα was observed. This band reacted with a mouse anti-hFcεRIα monoclonal antibody (15-1) and protein sequence analysis confirmed that this was the soluble extracellular domain of human FcεRIα (data not shown).

(d) Kinetic analysis of the IgE variant-sFcεRIα interaction. Figure 2 shows a typical sensorgram used for the determination of kinetic constants for the wild type recombinant IgE/sFcεRIα interaction. This procedure was employed for all other IgE variants. The sensorgrams obtained were highly reproducible and clearly demonstrate capturing of the IgE to the immobilized antibody as well as its interaction with sFcεRIα Kinetic analysis applying a pseudo-first order model as outlined by Karlsson (38) leads to apparent rate constants for each concentration used, as shown in Table 2. The k_a and k_<j values obtained for native IgE shown in Table 2 compare well with previously published determinations (14,19,39,41,42) where the binding of hlgE to cell surface FcεRI was measured. These data are however based on the assumption that IgE/FcεRI interaction is a simple 1 : 1 pseudo-first order interaction.

The results of a more refined kinetic analysis applying global fitting techniques are shown in Figure 2, panels C and D. The relatively large deviations from the best fit of a pseudo- first order model to the data clearly show that the kinetic binding data of IgE/FcεRI interaction provided by the SPR biosensor does not fit a simple 1 : 1 binding model. This is also reflected in the nonlinear dependence of k_{o s} on concentration, as shown in Figure 3. For a pseudo-first order reaction, the relationship of k_obS(c) is linear, and the slope gives the chemical on-rate constant, while an extrapolated rate 1^(0) is equal to the dissociation rate constant (43). For the IgE/FcεRI interaction this is clearly not fulfilled (Figure 3), as is evident from the biphasic characteristics of the sensorgrams.

The observed biphasic kinetics can be empirically modeled by double exponentials (data not shown). Such kinetics can only be meaningfully inteφreted on the basis of a detailed model for the interaction, which is currently unavailable. In addition, baseline subtraction to account for leakage of IgE from the surface was found to be not precise enough for a detailed kinetics analysis employing a complex model. For this reason a simple and robust model free descriptive analysis was used for the analysis of the different IgE variants. While the obtained values are not identical to bimolecular rate constants, the observed relationship k^c) still contains crucial information on the association properties and shows clear effects of the mutations (Figure 3). Similarly, the empirical dissociation rate constant k_d* and the fractional binding after completion of the 600 nM injection, demonstrate the distinct effects of mutations on equilibrium properties and dissociation process of the formed IgE/FcεRI complex (Figure 4). The dissociation rate constant kj* during the first 300 sec is a lower limit for the true bimolecular dissociation rate constant. It may, however, be a quantity compounded by, for example, the kinetics of possible conformational changes.

Based on this analysis, all samples of native IgE show curves k_obs(c) relatively consistent to each other and to most of the mutants (Figure 3). The largest effect was found for S341I/R342P*, which showed significantly slower association kinetics and an order of magnitude increased dissociation kinetics, consistent with an order of magnitude reduced fractional saturation. IgE variants with P333A and P333G mutations also showed significantly increased dissociation, a slightly slower initial association process, but little effect on fractional saturation. IgER16 showed the fastest association process and highest fractional saturation, while its dissociation rate is nearly identical to that of the wildtype. N371T exhibits slightly slower association kinetics, dissociation kinetics, and fractional binding. The N394T variants (N394T, N371/394T*) showed no detectable binding activity.

(e) Establishment of a rat basophilic leukemia (RBL) cell line supporting constitutive expression of the α-chain of human FcεRI and stimulus secretion coupling by native and mutant forms of human IgE. The RBL-J41 constitutively expresses -10,000 human α-chains per cell and was employed to assess the coupling of IgE-mediated receptor activation to degranulation (36) and to detect any changes in the effector functions of IgE variants. Data for native IgE and the 13 IgE variants generated in this study are summarized in Figure 5. Panel (A) illustrates that the ability of the IgER16 variant (14) to stimulate mediator secretion following sensitization and challenge with NIP-HSA or anti ε-chain antibody is essentially indistinguishable from native hlgE. Suφrisingly, replacement of P333 by A or G has only a modest effect on binding to the cellular receptor (Table 2) but profoundly influences the capability of the mutant ligands to induce mediator secretion following crosslinking by antigen or anti-IgE. The P333A* mutant showed an approximately 50% reduction in its ability to couple a crosslinking stimulus to mediator secretion compared to native IgE. This indicates that this mutation inhibited the ability of the molecule to aggregate hFcεRIα in response to an antigen or anti-IgE-mediated crosslinking stimulus (Figure 5A). It is unlikely that this effect is influenced by an additional mutation (F321L) since the single P333G point mutation is associated with a complete loss of stimulus secretion coupling.

Data in Figure 5 (B) show secretion levels obtained following the analysis of a series of A-B loop variants; D347N, D347E, P345A, R351K and L348I. Interestingly, none of the AB loop variants altered the hFcεRIα binding and aggregation activity compared to native IgE. Figure 5 (C) shows the dose/response challenge data for RBL-J41 cells sensitised with the K352G*, N371T, N394T and N371T/N394T*. It demonstrates that IgE mediated secretion levels for the K352G* and N371T variants are in excellent agreement with the control values obtained for native IgE. IgE mutants that contain the N394T (N394T, N371T/N394T*) substitution demonstrated no detectable binding activity and do not support secretion of cellular mediators. f) Binding of hlgE variants to FcεRII.

In order to evaluate the influence of the site directed mutagcnesis of hlgE Fc on the IgE-FcεRlI interaction, we employed a scmi-quaniiiaiive Fluorescence-Activated Cell Sorting (FACS) assay using the human EBV-transformcd B-ccll line. RPMI-8S66 ( intncr &. Sugden 1981). The RPMI-SS66 cell line expresses high levels of hFcεPJI and l as been used extensively in hlgE stπicrure/function studies (Hook et al 1991. Kissini et al 1993. Meishcr ct al 1994. Helm et al 1996). hFcεRII expression was assessed in each experiment using a mouse monoclonal anti liFcεRII antibody (MHM6) followed by a FITC labelled anti IgGi mouse monoclonal antibody. Expression levels were relatively consistent and varied +/- 1-10% between experiments indicating a good standardisation of assay conditions. The background fluorescence was determined by examining unstained cells and isotype control labelled cells, the species specificity of the liIgE-FcεRII interaction was utilised for the IgE variants analysis by using mouse IgE as a negative control and a commercial IgGi was used as a control for the MHM6 antibody. These levels were found to be low and consistent between the different controls and experiments which is common in FACS analysis using specific monoclonal antibodies.

The hlgE variants w^:ere analysed by quantifying tlie FcεRTI binding in a dose response manner (0.25- 2.5 μg of IgE variant/ 5 x 10" cells). The hlgE-FcεRϋ complex was visualised by probing with an FITC labelled anti-mouse λ light chain antibody utilising tlie chimeric nature of the antibody.

Tlie effect of mutations within the hlgE Fc on tlie IgE-FcεRTI interaction are illustrated in Figure 3. Included in each figure is tlie wild type hlgE level of binding and tlie level of background fluorescence observed For all experiments tlie wild type molecule binding profile was consistent with saturation binding curves observed for other ligand/receptor interactions and demonstrated that the analysis was performed under pre- saturation conditions. Figure 3 Panel (A) shows the analysis of the Cε2-3 interface variants, P333 to A and G. It can be seen that wild type and P333 A* show essentially identical binding profiles whereas the P333G appears to have an enhanced binding activity compared to the wild type molecule. This is of particular interest the removal of structural restraints associated with the mutation of P to G may introduce greater inter-domain flexibility

promoting the FcεRII interaction. Also shown in Panel (A) is the analysis of two A-B loop variants. P345A and D347E, both of these variants show increased FcεRII binding activity compared to wild type hlgE under these conditions. The P345A shows the most pronounced effect which is similar in magnitude to the P333G variant, again suggesting a possible role of P in the restriction of structure influencing the interaction, its removal having an enhancing effect Figure 3 (B) illustrates the analysis of three A-B loop variants; the R16, S341I/R342P* and R351K, the R16 mutation has had a dramatic effect on tlie FcεRTI binding activity of the molecule destroying the IgE-FcεRII interaction completely (fluorescence intensity to background levels). These data confirm our previous findings examining this variant (Helm et al 1996) and do suggest a role for the A-B loop i tlie FcεRII interaction. The S341I/R342P* variant shows diminished FcεRII binding activity, however interpretation of these data in complicated by the multiple substitutions associated with this variant. The R35 IK variant showed essentially wild type hlgE-FcεRTI binding characteristics.

Figure 3 (C) illustrates the analysis of a further three A-B loop substitutions and tlie mutation of the glycosylation site at N371. The D347N and L348I variants show essentially wild type FcεRII binding characteristics. The N371T variant is associated with a significant increase in the FcεRII binding activity compared to the wild type molecule. This is an important result and confirms previous studies which demonstrated that the removal of glycosylation from the hlgE molecule is associated with an increase in FcεRII binding activity (Vercelli et al 1989). The K352G* variant is also of considerable interest, this substitution is associated with a significant decrease in the FcεRII binding activity of the molecule. The interpretation is complicated by the presence of the additional mutation E270G in tlie Cε2 domain. It seems unlikely that this mutation would influence tlie interaction, therefore K352 may be a class specific effector residue or be involved in the maintenance of the A-B loop conformation required for FcεRII docking.

Figure 3 (D) illustrates the analysis of the t o N394T variants (N394T. N371/394T*). In comparison to tlie wild type and background controls there does appear to be binding but at a lower level. Tlie curve is suggestive of non specific binding showing only a limited progression to saturation, although the use of mouse monoclonal antibodies should eliminate any non specific interactions, as seen for tlie R16 variant were the FcεRII interaction was destroyed (Figure 3 (A)). It is significant that the removal of tl e single glycosylation site at N394 is associated with a greater decrease in FcεRII-binding compared to the removal of both tlie N371 and N394 sites, this is consistent with the data for the N371 variant which demonstrated an enhanced binding level

following deglycosylation at this residue. This, and the data for the R16 variant provide compelling evidence that the binding is specific and these data is of interest in the context of the IgE-FεRI/II interactions. g) Strategies for the development of therapeutic interventions in blood cell disorders

An approach to identifying the regions in IgE responsible for IgE/receptor recognition focuses on the nature of the complementary binding site between IgE and its receptors.

Identification of the FcεRI binding site in h IgE:

Fig. 12 shows that the sequences common to all FcεRI fragments capable of recognising FcεRI comprise Pro 343-Ser353 in the Cε3 domain.

As our study shows, this IgE epitope has an application as an immunogen in the therapy of all IgE-mediated allergies through active immunisation irrespective of the nature of the allergen [rev. in ref.l].

DISCUSSION

Kinetic analysis. In the current investigation measurements for the binding of hlgE to RBL cells transfected with the α-chain of the hFcεRI were complemented by a biosensor based analytical system, which is a relatively new technique applied to the study of ligand/receptor interaction (4). Although our data are indicative of a biphasic binding mode, we demonstrate that when equal criteria, i.e. pseudo first order kinetics are applied to the interaction, identical kinetic constants are obtained from cell binding and SPR studies. Similar observations were made when the kinetics of binding of the P333A/G mutants to the cellular receptor were compared with SPR measurements (Table 2). Our analysis, therefore, does not indicate that additional receptor subunits like the γ-chain increase the affinity between IgE and FcεRI.

Mutations of residues in the A-B loop of Cε3. The assignment of residues in the A-B loop in Cε3 of hlgE as a major structural determinant in IgE receptor interaction was based on previous observations, which showed that the peptide sequence Pro343-Ser353 is an essential structural determinant common to all hlgE-Fc derived fragments that engage FcεRI (14). Replacement of residues conserved between rodents and man, including P345A, D347N or E and L348I, show little influence on binding. This renders it unlikely that these residues make a direct contribution to IgE/FcεRI interaction.

Collectively, these findings indicate that the overall conformation of this region, computed to form a loop constrained by invariant P residues proximal to the Cε4 domain, is critical for the correct docking of the ligand into the receptor. While N-and C-terminal truncations of the loop by only one residue effect loss of FcεRI recognition, no individual residue appears essential for binding. Others demonstrated that replacement of the A-B loop by the homologous sequence from IgG₂, which involves the insertion of an extra residue, reduces IgE/FcεRI interaction to background levels (3% positive cells) (26), indicating that mutations that alter the loop size inhibit receptor recognition. Further support for this notion comes from our previous observation that a disulfide bond constrained A-B loop peptide, but not a random peptide encompassing this sequence, blocks the binding of hlgE to FcεRI with an IC₅₀ in the μmolar range (15). Similarly, chemokine- receptor interaction also depends on the recognition of a loop region in the ligand (52). In addition to the IgE loop as the principal structural determinant in IgE/receptor interaction, an important role for additional residues in the Cε3-terminal domain is apparent, since C-terminal deletion of residues up to Pro354 is associated with a >3000 fold increase in the dissociation of truncated Fcε from cells expressing the receptor as shown in Table 2. (14,15,17,19). delete, not relevant

Mutations of P333 in the hinge link region between Cε2-3. A role for the Cε2-3 interface (residues 330-335) in the IgE/FcεRI interaction has been suggested by several groups (reviewed inl5). By homology with IgG, we suggested that the receptor-binding site in IgE might be found in this region. Biological activity of this sequence was initially based on inhibition of PCA by IgE-derived peptides, which can give questionable inteφretations (15). Our recent studies, which show that N-terminal Cε3 sequences can be deleted to L340 with only a 30 fold increase in the rate of dissociation of the ligand from the receptor (Table 2), rule out a major contribution of these residues in FcεRI interaction. We focused on the role of P333 since FRET measurements suggested that rat IgE bends out of plane in this region (54). Structures with proline tend to be more inflexible since there are fewer possible conformational states. If IgE interacts with the receptor in a bent conformation, P in the hinge might be critical. To determine whether this residue contributes to the interaction by a restriction on structure, P333 was replaced by A and G. As shown in Figure 4B, SPR kinetic analysis showed only a modest increase in the dissociation rate for both mutants compared to the native IgE molecule, and both ligands recognize cellular hFcεRI with high affinity (Table 2).

In contrast, a striking decrease in the ability of the P333 mutated IgE molecules to support aggregation of FcεRI was observed. P333A reduced mediator release by 50%, while P333G could not mediate degranulation. (Figure, Panel A). A comparison of NIP-HSA and anti-hlgE mediated aggregation demonstrated (Figure 5, Panel A) that this failure is not influenced by the nature of the crosslinking stimulus. These observations are unexpected in view of the small effects of the P333 mutations on the kinetics of IgE/FcεRI interaction and difficult to reconcile with the notion that, once receptor sensitization occurs, secretory responses can be initiated by clustering adjacent receptors via cognate antigen or anti-IgE (55,56). They disagree with earlier suggestions that the persistence of FcεRI-mediated signal transduction is directly related to the intrinsic affinity of the ligand for the receptor (57). It has, however, been suggested that orientational constraints can affect the mobility and state of aggregation of the receptor, and that this determines secretory responses. Of particular interest in this respect is the work by Pecht and collaborators (49,58,59-61), who compared IgE-mediated stimulus secretion coupling in response to monoclonal antibodies. Their work suggests that the capacity to induce mediator release is not a function of the affinity of the crosslinker for the ligand, but depends on the assembly of aggregated receptors into immobile species (59). Their observation, that the rigidity of IgE/anti-IgE correlates with the capacity to trigger mast cell exocytosis, is directly relevant to our findings, which demonstrate that a progressive decrease in the capacity of hlgE variants to mediate mast cell degranulation parallels the relief of structural constraints computed to occur as a result of exchange of P by A and G. It is interesting to note that others conclude that P333 plays no role in IgE/FcεRI interaction (23), since its mutation is associated with a <7fold increase in dissociation. No information was published relating to the ability of this variant to mediate degranulation. In contrast, P345A mutation does not alter effector functions of IgE, high-lighting the critical role of P333A/G mutations in the ligand' s ability to couple a crosslinking stimulus to mediator secretion. The effects we observe point to a progressive loss of entropic constraints, which have been computed to introduce a bend into the structure of the interdomain region between Cε2/3 (21). It can therefore be seen that our modified IgE has a role to play in modulating the response of the immune system to it can be used to control the presence/degree of inflammatory response.

Discussion of hlgE - FcεRII interaction

The new IgE variant IgE Gly 333/352 no longer recognises cells expressing FcεRII , while binding to cells expressing FcεRI with the same affinity as wild type human IgE. The distinct advantage of this construct relates to the ability of this molecule to engage cells expressing FcεRI without inducing cellular responses.

Therapeutic applications

The development of stable cell lines, which express the ligand binding domain of h FcεRI and which respond to a h IgE-mediated antigenic stimulus with mediator release, led to the identification of the minimum sequence requirements for the binding to both receptors and provided important information regarding residues/structural variables necessary for cell activation and secretion of mediators of the allergic response.

The engineering of a variant form of IgE (IgE LYS 352 to GLY) which selectively recognises cells expressing the high-affinity receptor, but which does not bind to FcεRI 1/CD23 has potential therapeutic applications in the treatment of systemic mast cell and basophil malignancies when linked to a (immuno)toxin, radioactive isotope or agent stimulating apoptosis. In addition, it can be used for the selective isolation of cells expressing FcεRI for functional studies. This offers a distinct advantage compared to current methods, which utilise c-Kit (stem cell factor) ligand, which stimulates post- receptor responses in mast cells and basophils, which culminate in degranulation of cellular mediators

Discussion

In comparison to the MgE-FcεRI interaction comparatively little is known about the structural basis of the hlgE-FcεRII interaction The interaction lias been proposed to ha\e a 2 1 bmdmg stoichiometn . with the two lectin homology head regions of the receptor interacting with the two ε chains of an indmdual IgE molecule (Sutton & Gould 1993). dimeπsaϋon of the hlgE molecule is required (Vercel et al 1989. Helm et al 1996) This has recently been confirmed using a soluble form of the FcεRII lectin homology head region and a hlgE Fcε3-4 fragment (Shi et al 1997) As for the IgE-FcεRI interaction it is suggested that the IgE molecule adopts a bent receptor bmdmg conformation, this was oπgmal based on FRET measurements examining rat IgE (Zheng et al 1991) The mam bmdmg regιon(s) on each molecule ha\e been assigned to tlie Cε3 domain of hlgE (see mtroducuon) and the lectin homology head region of the receptor (Bettler et al 1992). although still outstanding is the identification of class specific effector residues within the hlgE Fc

In the current investigation we aimed to further define the structural basis of the hlgE-FcεRII mteracuon by identif ing class specific effector residues within the gE-Fc To this end we engineered a large number of single and multiple substitutions mto tlie hlgE Fc region The Cε2-3 interface. Cε3 A-B loop and Cε3 glycosylation sites were targeted in order to determine the role of these regions m the bmdmg interaction or m the maintenance of the bmdmg conformation of the hlgE molecule The effects of the substitutions on tlie mteracuon were assessed using a semi -quantitative FACS analysis utilising the RPMI-8866 cell line which expresses high of liFcεRII Previously we have been crmcal of the use of FACS in order to examine IgE- Fc receptor interactions (Helm et al 1996). however the technique provided a convement method for the identification of gross changes m the FcεRII bindmg activit} of the hlgE vaπants Caution was maintained in the interpretation of these data due to the use of non equilibrium conditions, and the fact that the short incubation times and wash steps may not affect all vaπants m the same manner This could be due to the influence of altered kinetics of bindmg of each vaπant The method did provide evidence for the role of a number of regions of the hlgE molecule invoh ed in the IgE -FcεRII interaction and identified class specific effector residues

Mutation of Glycosylation sites

The FACS analysis shows that mutation of N371 results m a significant enhancement in the level of FcεRII receptor bmdmg compared to the wild tvpe molecule (Figure 3(C)) This is m excellent agreement with previous studies w Inch demonstrated that enzymaticallv deglv cosv lated IgE (PS) and recombinant Fcε fragments expressed m E colt exhibited an enhanced FcεRII bindmg activity (Vercelh et al 1989) It is interesting to speculate that this removal of the gh cosv lation at N371 has relieved structural restraints associated with the carbohydrate moiety which influence tlie IgE -FcεRII interaction The direct bmdmg influence of hlgE glvcosylation has been dismissed though the FcεRII shows homologv to lectins which bmd their hgands Λia carbohydrate due to the activity of £ colt synthesised Fcε fragments (Vercelh et al 1989, Helm et al 1996) The glvcosylation site at N371 is in close proximity to sequences of the Cε3 implicated as being imolved in the IgE-FcεRII interaction, residues 367-376 (Cretien et al 1988), residues 367-370 (Vercelh et al 1989) and residues 364-383, 401-415 (Ghadeπ & Stanworth 1993) It is therefore tempting to suggest that the remoΛal of glycosv lation at N371 has made these regions more accessible to the complimentary site on FcεRTI, confirming their role The predicted surface exposed position within our model of the hlgE Fc (Figure 1, Padlan & Davies 1986, Helm et al 1991) is concurrent with these suggestions

In contrast to the N371 vaπant the N394 vaπants (N394T, N371/394T*) demonstrated a significant reduction m FcεRII bmdmg activity (Figure 3(D)) The nature of tlie bmdmg curves for these vaπants show onlv a limited coπelation with typical saturation bmdmg curves (see wild tvpe control) and are suggesttv e of nonspecific bmdmg However a close examination of the data for tlie vaπants argues against this interpretation, tlie use of monoclonal antibodies facilitates low non specific bindmg, e g. the R16 vaπant shows no activity and gives background levels of fluorescence (Figure 3(A)), this is m contrast to the data for the N394 vaπants In addition the single mutation at N394 is associated with a greater reduction m FcεRTI bmdmg activity compared to the double gh cosv lation aπant with the additional mutation at N371. this enhanced bmdmg associated with the N371 mutation is m excellent agreement with tlie single N371 vaπant and therefore is suggestive of a specific interaction Tlie results were not as predicted it was postulated that the substitution of the N394, predicted to be buπed in the Cε3 domam structure (Figure 1, Padlan & Davies 1986. Helm et al 1991) would result m a loss of doma conformation leaduig to the total removal of FcεRII-receptor recognition This was based on the observations of our recent study which showed that these vaπants could no longer engage FcεRI (Sa ers et al , paper on preparation) These data suggest that mutation of N394. which is conserved m other Ig classes and is thought to be cπtical for the formation of domam structure, has a role m the formation of an active hlgE molecule m eukaryotes The apparent difference in the dependence of the two FcεR receptors on Cε3 domam conformation is of particular mterest, the data of this study and the smdy on the FcεRI interaction suggest that the FcεRI mteracuon is more sensitive to changes in the overall domam conformation m agreement with the concept that the complimentary bmdmg sιte(s) on hlgE for FcεRI involve multiple non continuous sites across the whole Cε3 domam (Helm et al 1996. McDonnell et al 1997, Henry et al 1997. Savers et aL Paper m preparation) In contrast tlie FcεRTI may mvolv e more restπcted interaction sιte(s) within the hlgE Cε3 domam

This is m agreement with the studv of Nissim and co-workers who demonstrated tlie importance of domam conformation for both FcεR interactions bv the introduction of rodent residues mto tlie N- and C terminal regions of hlgE Cε3 (Nissim et al 1993) Tins interpretation seems to be the most likely in tlie absence of definiti e structural data, the direct involvement of N394 glycosv lation m either FcεR interaction has previously been dismissed (Vercelh et al 1989, Helm et al 1996, Savers et al . Paper m preparation)

Mutation ofCε2-3 Interface

The effect of the P333 to A and G substitutions are shown m figure 3(A), it can be seen that the more conservative change of P333A* has had no effect on the FcεRII interaction whereas the substitution to G is associated with an apparent increase m FcεRTI bmdmg The mutation of P to A removes the fixed bend while maintaining the hvdrophobicity and size of tlie residue In contrast the mutation to G mvolves the substitution of the most conformationally restπctive to most conformationally unrestπctive residue, therefore this enhanced FcεRII binding may occur due to an indirect mechanism, with the removal of conformational restraints associated with the P to G substitution. This interpretation appears to be the most likely as a direct role for the Cε2-3 interface (aa330-335) m the FcεRII interaction is thought to be unlikelv, substitution of the rodent residues at 330-346 in the human Cε3 onlv reduced hFcεRII bindmg by 5 fold (Nissim et al 1993) Using Fcε fragments expressed m E colt e have deleted to residue 340 with no major effect on the interaction (Helm et al 1996) One study using monoclonal antibody mapping has suggested a direct role for the region (Takemoto et al 1994), although the interpretation of antibody mappmg studies is complicated by steπc hindrance These data suggest an indirect influence on the FcεRII interaction due to conformational changes, possiblv by the removal of restraints associated with the relative position of the Cε2 domam m the bent receptor bmdmg conformation of native hlgE (see Figure 1). It is of mterest that these substitutions had a profound effect on the FcεRI-mediated functions of the molecule which was interpreted as a result of changes m the nature/maintenance of the bent IgE Fc structure (Savers et al , Paper in preparation) Together these data give an insight mto tlie nature of the IgE- FcεR interactions, implying that inter-domain conformation is cπtical for the mamtenance of the FcεRI-mediated function and conversely its removal can even enhance FcεRII-binding In a related studv an anti Cε2 antibodv was able to enhance the hlgE-FcεRTI interaction, illustrating that conformational changes distal to the mam bindmg sιte(s) can influence the mteracuon (Miesher et al 1994) The concept of additional structural/conformational requirements to the mam bmdmg sιte(s) within tlie Cε3 domam has been demonstrated by the 2'-4 Fcε fragment in the study of Vercelh and co-workers (Vercelh et al 1989) This fragment was composed of the 30 C-teπiunal residues of Cε2 and the entire Cε3 and 4 domains, m compaπson to tlie Cε3-4 fragment this 2 ^"-4 Fcε fragment showed a significant increase m FcεRII bindmg activitv (similar to the Fes fragment. Cε2-4) implying tlie role of Cε2 residues m the mamtenance of a receptor bmdmg conformation (Vercelh et al 1989)

Mutation ofCε3 A-B Loop

The replacement of residues 341-356 (A-B loop) m the hlgE Fc for the homologous rat IgE residues (vaπant R16) resulted m tlie complete loss of hFcεRII bmdmg (Figure 3(B)) This is m agreement with our previous study (Helm et al 1996) and tlie study of Nissim and co-workers which replaced residues within the N- terminal of the hlgE Cε3 domain for the homologous mouse residues (Nissim et al 1993) Replacement of residues 330-346 (vaπant C3BX) maintained the interaction, however replacement of residues 330-356 (vaπant C3HD) dramaticall reduced activity suggesting a role for this region m the FcεRfl^" mteracuon (Nissim et al 1993) It is considered unlikelv that the substitutions generated within the R16 vaπant have effected the overall Cε3 domam conformation leading to this loss of FcεRII bmdmg activitv because when the FcεRI mteracuon was examined essentially wild tvpe interaction characteπsucs were observed although there was a slight increase m the association process (Savers et al , Paper m preparation)

In order to rationalise the data for the R16 vaπant further A-B loop vaπants ere examined vaπants R351K. D347N and L348I showed essentialh wild tvpe FcεPJI-binding characteπstics suggestmg that these residues do not have a direct role in the interaction Analvsis of the S341I R342P* vaπant (mutation to rat residues) (Figure 3(B)) showed that the FcεRII bmdmg activity was significantly decreased m companson to the wild type molecule Interpretation of this result is difficult due to the multiple additional mutations contamed within this vaπant (A338V, L345P. T434A) It is interesting that the decreased activitv of this vaπant demonstrates that residues S341 and R342 mav be. at least m part responsible for the R16 (contains R341I/R342P substitutions) vaπant loss of FcεRJI bindmg This is speculative due to the probable effect of the additional mutations on the Cε3 domam conformation, particularly the introduction of a P The additional mutations within this vaπant are predicted to be m close proximity to the glycosv lation site at N371 previously implicated in tlie IgE-FcεRII interaction (Cretien et al 1988. Vercelh et al 1989. Ghaden & Stanworth 1993) (see Figure 1) This and the data for the N371 mutation seem to provide evidence for the role of this region m addition to the A-B loop m the hlgE-FcεRII lnteraction.

Mutation of P345 to A resulted m an mcrease m the FcεRTI bindmg activitv compared to the wild type molecule (Figure 3(A)). this is of interest due to the possible role of P345 m the restπction of structure m order to mamtam the conformation of the A-B loop This removal of defined structure mav mcrease the flexibility of the loop leading to the apparent enhancement of the docking process of hlgE and FcεRII, it is also suggestiv e that the P345 residue is not cπtical for the overall mamtenance of the A-B loop FcεRlI-binding conformation.

Another vaπant that demonstrated an increased level of FcεRII-binding contamed the D347E substitution (Figure 3(A)). this was of interest because mutation of D347 to N generated a molecule with essentially wild type bmdmg characteπstics (Figure 3(C)). although there was a small reduction m bmdmg level Both substitutions are highly conservative, it is interesting that the enhanced bmdmg is associated with the mamtenance of charge of the residue although the side chain length has increased It is tempting to speculate concerning the reason for this increase in bmdmg level although the D347N data suggests that we have not identified a class specific effector residue

Of all of the A-B loop vaπants generated witluii this studv the most prominent effect was observed for the K352G* substimtion which resulted m a significant reduction m FcεRϋ bmdmg compared to the wild type molecule (Figure 3(C)) Interpretation is complicated by the additional mutation within tlie Cε2 domain. E270G although it is considered that this distal mutation would not result m the dramatic effect observed even though a supporting role for Cε2 m the hlgE-FcεRII interaction has been suggested previously (Vercelh et al 1989, Nissim et al 1993) We mterpret the effect of this substimtion as identifying K352 as a class specific effector residue that directlv contπbutes to FcεRII bindmg possibly via an electrostatic interaction, and may account, at least m part for the loss of FcεRII bindmg associated with the R16 vaπant which contained this substimtion.

At present, limited structural information is available concerning the interaction of hlgE and its FcεR receptors Modelling of these components using homology to known structures has greatly facilitated the interpretation of prote engineering data although caution is still required In the cuπent investigation we identity kev regions of the MgE molecule mvolved in the FcεRII mteraction and give new insislits into the structural nature of this interaction. Acquiring information concerning the structure of complimentary binding sites on IgE and its Fc receptors holds great potential for the rational design of anti-allergic drugs which can modulate tlie allergic response.

REFERENCES

1. Capron, A., Dessaint, J.P., Capron, M., Joseph, M., Ameisen, J.C. & 5 Tonnel, A.B. (1986) Immunology Today. 7 15-18.

2. Hagan, P., Blumenthal, U.J., Dunn, D., Simpson, A.J.G. & Wilkins, H.A. (1991 Nature 349: 243-245.

3. Holgate, S. T. (1997) Lancet 350 (suppl. II), 5-9.

4. Stanworth, D.R., Humphrey, J.H. & Bennich, H.H. (1968) Lancet. 2 : 10 17-18.

5. Stanworth, D.R., Jones, V.M., Lewin, I.V. & Nayyar, S. (1990) Lancet 336 : 12791281.

6. Ishizaka, K.; Ishizaka-T. & Lee-E.H. (1970) Immunochemistry. 7 : 687-702.

15 7. Kenten, J., Helm, B., Ishizaka, T., Cattini, P. & Gould, H. (1984) Proc. Natl. Acad. Sci. U.S.A. 81, 2955-2950.

8. Lui, F.T., Albrandt, K.A., Bry, C.G. & Ishizaka, T. (1984) Proc. Natl. Acad. Sci. U.S.A 81 : 5369-5373.

9. Geha, R.S; Helm, B. & Gould, H (1985) Nature. 315 : 577-578.

20 10. Coleman, J.W., Helm, B.A., Stanworth, D.R. & Gould, HJ (1985) Eur. J Immunology. 15 : 966-969.

11. Ishizaka, T., Helm, B., Hakimi, J., Niebyl, J., Ishizaka, K. & Gould, H. (1986) PNAS USA. 83 : 8323-8327.

12. Helm, B., Marsh, P., Vercelli, D., Padlan, E., Gould, H. & Geha, R 25 (1988) Nature 331 : 181-183.

13. Helm, B.A, Kebo, D., Vercelli, D., Glovsky, M.M., Gould, H.J., Ishizaka, K., Geha, R., and Ishizaka, T. (1989) Proc. Natl. Acad. Sci. U.S.A. 186, 9465-9469. 14. Helm, B.A., Sayers, I., Higginbottom, A., Machado, D.C., Ling, Y., Ahmad, K., Padlan, E.A. & Wilson, A.P.M (1996) J. Biol. Chem. 271 : 7494-7500.

15. Helm B.A., Spivey, A.C. & Padlan, E.A. (1997) Allergy 52 : 1155- 1169.

16. Vercelli, D., Helm, B., Marsh, P., Padlan, E., Geha, R.S. & Gould, H (1989) Nature. 338 : 649-651.

17. Weetall, M., Shopes, B., Holowka, D. & Baird, B. (1990) J. Immunol. 145 : 3849-3854.

18. Nissim, A. & Eshhar, Z (1991) International Archives of Allergy and Applied Immunology. 94 : 93-95.

19. Nissim, A. & Eshhar, Z (1992) Molecular Immunology. 29 : 1065- 1072. 20. Nissim, A., Schwarzbaum, S., Siraganian, R. & Eshhar, Z (1993) J. Immunol. 150 : 1365-1374.

21. Helm, B.A., Ling, Y., Teale, C, Padlan, E.A. & Bruggemann, M. (1991) Eur. J. Immunol. 21 : 1543-1548.

22. Duncan, A.R., Woof, J.M., Partridge, L.J., Burton, D.R. & Winter, G. (1988) Nature. 332 : 563-564.

23. Beavil, A.J., Beavil, R.L., Chan, C.M.W., Cook, J.P.D., Gould, H.J., Henry, A.J., Owens, R.J., Shi, J., Sutton, B.J. & Young, R.J (1993) Biochem. Soc. Trans. 21 : 968-972.

24. Sutton, B.J. & Gould, H.J (1993) Nature. 366 : 421-428. 25. Henry, A.J., Cook, P.D., McDonnell, J.M., Mackay, G.A., Shi, J., Sutton, B.J. & Gould, H.J. (1997) Biochem. 36 : 15568-15578.

26. Presta, L., Shields, R., OConnell, L., Lahr, S., Porter, J., Gorman, C.

& Jardieu, P (1994) J. Biol. Chem. 269 : 26368-26373.

27. Marsh, P., Helm, B.A. & Gould, H.J. (1989) Johansson S.G.O, Editor. Pharmacia Allergy Research Foundation, Uppsala, Sweden, 26-31. 28. Wilson, A.P., Pullar, C, Camp, A. & Helm, B.A. (1993). Eur. J Immunol. 23 : 240-244.

29. Lindstrom-Bahr, H. (1974) Ph.D Thesis, University of Uppsala.

30. Neuberger, M.S., Williams, G.T., Mitchell, E.B., Jouhal, S.S., Flanagan, J.G. & Rabbitts, T.H. (1985) Nature 314 : 268-270.

31. Brueggemann, M., Williams, G., Bindon, C, Clark, M., Walker, M., Jefferis, R., Waldmann, H. & Neuberger, M. (1987) J. Exp. Med. 166 : 1351-1361.

32. Padlan, E.A. & Davies, D.R (1986) Molecular Immunology. 23 : 1063-1075.

33. Deisenhofer, J. (1981) Biochem. 20 : 2361-2370.

34. Kulczycki, A., Isersky, C. & Metzger, H. (1974) J. Exp. Med. 139 : 600-616.

35. Kabat, E.A., Wu, T.T., Reid-Miller, M., Perry, H..M. & Gottesmann, KS. (1991) Sequences of Proteins of Immunological Interest., 5^th ed.,

U.S. Department of Health and Human Services.

36. Helm, B.A., Sayers, I.; Swan, J., Smyth, L.J.C., Cain, S.A., Suter, M., Machado, D.C., Spivey, A.C. & Padlan, E.A. (1998) Technology & Health Care, in press. 37. Bingham, B.R., Monk, P.N. & Helm, B.A. (1994) J. Biol. Chem..

269 : 19300-19306.

38. Karlsson, R., Michaelsson, A., Mattsson, L. (1991) J. Immunological Methods.145 : 229-240.

39. Miller, L., Blank, U., Metzger, H. & Kinet, J.P. (1989) Science 244 : 434-437.

40. Hakimi, J., Seals, C, Kondas, J.A., Pettine, L., Danho, W. & Kochan, J. (1991) J. Biol. Chem. 265 : 22079-22081. 41. Basu, M., Hakimi, J., Dharm, E., Kondas, J.A., Tsien, W.H., Pilson, R.S., Lin, P. & Gilfillan, A., Haring, P., Braswell, E.H., Nettleton, M.Y. & Kochan, J.P (1993) J. Biol. Chem. 268 : 13118-13127.

42. Young, R.J., Owens, R.J., Mackay, G..A., Chan, C.M.W., Shi, J., Hide, M., Francis, D.M., Henry, A.J., Sutton, B.J & Gould, H.J.

(1995) Protein Engineering. 8 : 193-199.

43. Schuck, P. & Minton, A.P. (1996a) Trends in Biochemical Sciences. 252 : 458-460.

44. Schuck, P. (1997) Ann. Rev. Biophys. Biomol. Struct. 26 : 541-566. 45. Schuck, P. & Minton, A.P. (1996b) Anal. Biochem. 240 : 262-272.

46. O'Shannessy, D.J. & Winzor, D.J. (1996) Anal. Biochem. 236 : 275- 283.

47. Minton, A.P. (1980) Mol. Pharmacol. 19, 1-14.

48. Schuck, P, Milar, D.B. And Kort A.A. (1998) Analytical Biochemistry, in press

49. Ortega, E., Schweitzer-Stenner, R. & Pecht, I. (1991) Biochemistry 30, 3473-3483.

50. Sechi, S., Roller, P.P., Willette-Brown, J. & Kinet, J.P. (1996) J. Biol. Chem. 271, 19256-19263. 51. Nettleton, M.Y. & Kochan, J.P. (1995) Int. Arch. Allergy Immunol. 107 : 328-329. 52. Clark-Lewis, I., Kim, K.S., Rajarathnam, K., Gong, J.H., Dewald, B., Moser, B., Baggiolini, M. & Sykes, B.D. (1995) J. Leukocyte Biol. 57 : 703-711. 53. M^cDonnell, J. M., Beavil, A. J., Mackay, G. A., Henry, A. J., Cook, J. P. D., Gould, H. J. & Sutton, B. J. (1997) Biochem. Soc Trans. 25 387-392. 54. Zheng, Y., Shopes, B., Holowka, D. & Baird, B (1991) Biochemistry (USA). 30 : 9125-9132. 55. Ishizaka. T. & Ishizaka, K. (1984) Prog. Allergy. 34 : 188-235.

56. Metzger, H., Alcaraz, G., Hohman, R., Kinet, J.P., Pribluda, V. & Quarto, R. (1986) Ann. Rev. Immunol. 4 : 419-470.

57. Kent, U.M. Mao, S.Y., Wofsy., C, Goldstein, B., Ross, S. & Metzger, H. (1994) Proc. Natl. Acad. Sci. U.S.A. 91, 3087-3091.

58. Ortega, E., Schweitzer-Stenner, R. & Pecht, I. (1988) EMBO J. 7 : 4101-4109.

59. Pecht, I., Ortega, E. & Jouvin, T.M. (1991) Biochem. 30 : 3450-3458.

60. Schweitzer-Stenner, R. & Pecht, I. (1994) Biochem. 33 : 8813-8825. 61. Schweitzer-Stenner , R., Tamir, I. & Pecht, I. (1997) Biophys. J. 72 :

2470-2478.

Table 1 hlgE vaπant Rationale for mutation

C 2-3 Interface Pro333-Ala Pro333 is conserved m human, rat and mouse IgE. Mutation to Ala may remo (*F321L) fixed bend while maintaining the hydrophobicity and size of the residue. Pro333-Gly Mutation may remove additional conformational constraints and introduce flexibility associated with a Gly residue.

A-B Loop

Homologous rat residues (aa341-356) grafted into the human Fc region to replace

R16 loop. Can the mutation confer binding to the rodent high affinity receptor, which not engage human IgE?

Replacement of human residues by the homologous rat ammo acids in IgE Fc.

Ser341/Arg342- the mutation confer binding to rodent FcεRI? Ileu/Pro

(*A338V, L425P, T434A) Pro345-Ala Pro345 is conserved m human, mouse and rat IgE. Mutation to Ala remove potentially fixed bend while maintaining the hydrophobicity and size of the residue

Asp347-Asn Asp347 is conserved m human, mouse and rat IgE. Mutation to Asn mamtams the of the ammo acid side cham but alters the charge.

Asp347-Glu Mutation to Glu mamtams the charge of the residue but increases the length of side chain.

Leu348-Ileu Leu348 is conserved in human, mouse and rat IgE. Mutation to lieu is hi conservative, changing only the position of a methyl group on the side chain.

Arg351-Lys Arg351 is not conserved m mouse and rat IgE. The rodent homologue is

Mutation to Lys maintains the hydrophilic nature of the residue while decreasmg length of the side cham.

Lys352-Gly Lys352 is not conserved, the rodent homologue is Gly. (*E270G)

Glycosylation Sites

Asn371 is conserved in human and mouse IgE, the rat homologue is Thr. Mutatio

Asn371-Thr Thr may change the glycosylation from type N to O or inhibit glycosylation. Asn394-Thr Asn394 is conserved in human, rat, mouse IgE and m other Ig classes Mutatio Thr may alter glycosylation from type N to O or inhibit glycosylation.

Asn371/394-Thr Double glycosylation site mutation. (*G368R)

Wild Type Produced as an internal control for the system.

Table 2

Ligand Assay kaM-ls"¹ kds-1 KAM-I Author

Current

Determinations

Recombinant SPR25°C 7.4x10^* 2.7 x 10^"° 2.78xlθ'^υ -

IgE Pseudo 1^st Order

Recombinant SPR25°C 7.2 x 10⁴ 3.1 x 10-⁶ 2.28 x 10¹⁰ -

IgE (Serotec) Pseudo 1^st Order

Recombinant SPR25°C -4-5 x 10⁴ -lo-⁴ -10⁹ -

IgE Empirical Recombinant Cellular lxlO⁴ 2 x 10^"5 5xl0⁹

IgE P333A 0

25 C

Recombinant SPR 25°C 4.85 x 10⁴ 1.02 xlO^"5 4.5 x 10⁹ -

IgE P333A Pseudo 1^st Order

Recombinant Cellular 8.2 xlO⁴ 2.8 x 10-⁵ 3 xlO⁹ -

IgE P333G 25°C

Recombinant SPR25°C 4.36 x 10⁴ 1.4 lO^-5 3.12 xlO⁹ -

IgE P333G Pseudo 1^st Order

Previous

Determinations

Human IgE Cellular l.,

0 8.0 x 10⁴ 9.0 xlO^"6 0.90 xlO¹⁰ Miller et a

25 C 1989

Myeloma IgE Cellular 1.1 xlO⁵ 1.3 xlO^"5 0.85 x 10¹⁰ Hakimi et 25°C al., 1990

Recombinant Cellular 10

1.8 xlO⁴ 3.0 xlO^"6 0.62x10 Nissim et al., IgE 25°C 1992

IgE-PS Cellular 10

1.1 xlO⁵ 1.1 x 10^"5 1.00x10 Basu et al., 25°C 1993

Myeloma IgE Cellular 10

0 3.1 x 10⁵ 9.0 x 10^"6 3.40 x 10 Young et al.,

25 C 1995

Recombinant Cellular 10

3.1 xlO⁵ 1.3 xlO^"5 2.40x10 Helm et al., IgE 22°C 1996

GST-Fcε 326- Cellular 4.1 x 10⁵ 2.95 x 10-⁵ 1.37 xlO¹⁰ Helm et al.,

547 22°C 1996

GST-Fcε 340- Cellular 4.2 x 10⁵ 9.8 x lo^-5 4.35 x 10⁹ Helm et al.,

547 22°C 1996 GST-Fcε 226- Cellular 4.3 x 10' 6.02 x 10° 7.14x10' Helm et al.,

354 22°C 1996

Myeloma IgE SPR 25°C k_al 3.5 x 10⁵ kji 1.4 x 10^"2 K_A12.5 x 10⁷ Henry et al,

Biphasic k^δ.όxlO⁴ 1^3.6 xlO^-4 K_A22.4xl0⁸ 1997

Claims

1. An IgE polypeptide, or effective fragment thereof, wherein said polypeptide contains a modification such that it is able to bind to its receptor at the cell surface of an immune cell but unable to bring about the release of immuno-active agents from said cell populations.

2. An IgE polypeptide, or effective fragment thereof, wherein said polypeptide contains a modification such that it is able to bind predominantly to a high affinity IgE receptor

3. An IgE polypeptide, or effective fragment thereof, wherein said polypeptide contains two modifications such that it is able to bind at the site of a first modification predominantly to a high affinity IgE receptor and is able to bind at the site of its second modification to its receptor at the cell surface of an immune cell and deliver a therapeutic agent

4. An IgE polypeptide according to any of Claims 1 to 3 wherein said modification is the substitution of PRO 333 of human IgE by a glycine residue

5. An IgE polypeptide according to any of Claims 1 to 4 wherein the replacement of

PRO 333 with a glycine amino acid residue or the replacement of PR0333 with a modified amino acid.

6. An IgE polypeptide according to any of Claims 1 to 5 wherein said modification is deletion or substitution of LYS 352 of human IgE, or to the homologous amino acid residue in a non-human IgE.

7. An IgE polypeptide according to any of Claims 1 to 6 wherein said modification is a substitution of LYS 352 with a glycine amino acid residue; or is a substitution of LYS352 with a modified amino acid.

8. An IgE polypeptide according to any of Claims 1 to 7 wherein said modification is substitution of PR0333 and LYS 352 of human IgE or to the homologous amino acid residue in a non-human IgE.

9. An IgE polypeptide according to any of Claims 1 to 8 wherein said modification is a substitution of PR0333 and LYS 352 with a glycine amino acid residue; or is a substitution of PR0333 and LYS352 with an alanine amino acid residue; or is a substitution of PR0333 and LYS352 with a modified amino acid.

10. An IgE polypeptide according to any of Claims 1 to 9 wherein said receptor is a high affinity IgE receptor; more preferably said high affinity receptor is FcεRI.

11. An IgE polypeptide according to any of Claims 1 to 10 wherein the IgE molecule of the invention has potential as a therapeutic agent to selectively deliver an agent, for example a toxin or a signal stimulating cell apoptosis to cells predominantly expressing high affinity receptors preferably, leukaemia cells.

12. An IgE polypeptide according to any of Claims 1 to 11 wherein the modified IgE molecule inhibits the aggregation of neighbouring IgE/receptor complexes in response to antigen and so inhibits the formation of, what may be termed, an aggregation signal; since release of pro-inflammatory molecules only occurs when bound IgE aggregates in response to a specific antigen (either via the antigen, lectins or anti IgE antibodies).

13. An isolated DNA molecule, or fragment thereof, encoding an IgE polypeptide, or an effective fragment thereof, as herein before defined.

14. A vector containing a DNA molecule encoding an IgE polypeptide according to any preceding aspect or embodiment of the invention.

15. A method to recombinantly manufacture IgE polypeptides according to the invention.

16. Therapeutic composition comprising: an IgE polypeptide according to the invention; optionally including, in association therewith, ideally coupled or joined thereto, a cytotoxic agent, and further comprising, an excipient, diluent or carrier.

17. A method to treat an allergen mediated disease or blood disorder 18. A kit for treatment of allergen mediated disease or blood disorders comprising: said therapeutic IgE composition; and delivery means to facilitate the administration of the therapeutic composition. 19. A method for the selection of cells expressing high affinity receptors for IgE peptides