Description Technical Field

The present invention relates to antibodies, antigens and immunogens, and more particularly to paratope-containing molecules that catalyze the hydrolysis of a preselected glycosidiσ bond.

Background of the Invention

Binding phenomena between ligands and receptors play many crucial roles in biological systems. Exemplary of such phenomena are the binding of oxygen molecules to deoxyhemoglobin to form oxyhemoglobin, and the binding of a substrate to an enzyme that acts upon it such as between a protein and a protease like trypsin. Still further examples of biological binding phenomena include the binding of an antigen to an antibody, and the binding of complement component C3 to the so-called CR1 receptor.

Many drugs and other therapeutic agents are also believed to be dependent upon binding phenomena. For example, opiates such as morphine are reported to bind to specific receptors in the brain. Opiate agonists and antagonists are reported to compete with drugs like morphine for those binding sites.

Ligands such as man-made drugs, like morphine and its derivatives, and those that are naturally present in biological systems such as endorphins and hormones bind to receptors that are naturally present in biological systems, and will be treated together herein. Such binding can lead to a number of the phenomena of biology, including particularly the hydrolysis of amide and ester bonds as where proteins are hydrolyzed into

constituent polypeptides by an enzyme such as trypsin or papain, or where a fat is cleaved into glycerine and three carboxylic acids, respectively.

Slobin, Biochemistry. 5:2836-2844 (1966) reported preparing antibodies to a p-nitrocarbobenzoxy conjugate of bovine serum albumin. Those antibodies were thereafter used to hydrolyze E~nitrophenyl acetate and epsilon-a inocaproate esters. The reaction of the acetate ester was described by a second-order rate constant and was said to appear to be nonspecific. The second-order rate constant obtained using normal gamma globulin was said to be about equal to that of the specially prepared antibodies. The presence of the specially prepared antibodies was said to inhibit the hydrolysis of the aminocaproate ester.

Kohen and coworkers also reported attempts using antibodies to catalyze esterolysis. The antibodies utilized by this group were, in each instance, raised to a portion of the' ultimately utilized substrate molecule that did not contain the bond to be hydrolyzed.

In their initial work rFEBS Letters. 100:137- 140 (1979) and Biochim. Biophvs. Acta. 629:328-337 (1980) ] anti-steroid antibodies were used to hydrolyze 7-umbelliferone (7-hydroxycoumerin) esters of a carboxyethyl thioether of a steroid. In each instance, an increase in hydrolytic rate was observed as compared to background or to a rate obtained with normal IgG. In both instances, turn over numbers were low (about one mole of substrates per mole of antibody per minute, or less) , and the reaction rates declined with time, reaching a plateau with saturation of the antibody. That slow down in rate was attributed to an irreversible binding of the steroidal acid product to the antibody.

Kohen et al. also reported hydrolysis of 7-[-N-(2,4-dinitrophenyl)-6-aminohexanoyl]-coumarin using monoclonal antibodies raised to the dinitrophenyl portions of that substrate molecule fFEBS Letters. 111:427-431 (1980)]. Here, a rate increase over background was also reported, but the reaction was said to be stoichiometric rather than catalytic. A decrease in rate that approached zero was reported as saturation of the antibody was reached. Again, the decrease was attributed to product inhibition caused by binding of the product acid to the antibody since some of the initial hydrolysis activity could be regenerated by chromatography of an antibody-substrate-product mixture. When strong antibody binding is directed to stable states of substrate molecules, the slow rate of dissociation of the complex will impede catalysis. Such is thought to be the situation for the results reported by Kohen and coworkers.

The above constructs, though interesting, are severely limited by the failure to address the mechanism of binding energy utilization which is essential to enzymes [W. P. Jencks, Adv. Enzvmol.. 43. 219 (1975)].

Those deficiencies can be redressed by using a transition state analog as the hapten to elicit the desired antibodies. This hapten (also referred to herein as an "analog-ligand") can assume the role of an inhibitor in the catalytic system.

Thus, immunological binding can be used to experimentally divert binding interactions to catalytic processes. For example, it was suggested that use of an antibody to a haptenic group that resembles the transition state of a given reaction should cause an acceleration in substrate reaction by forcing substrates to resemble the transition state. Jencks, W.P., Catalysis in Chemistry and Enzvmoloσv. page 288 (McGraw-

Hill, New York 1969) . Notwithstanding that broad suggestion, specific transition state haptens were not suggested, nor were specific reactions suggested in which the concept might be tested. Hydrolysis of amide and ester bonds is thought by presently accepted chemical theory to proceed in aqueous media by a reaction at the carbonyl carbon atom to form a transition state that contains a tetrahedral carbon atom bonded to (a) a carbon atom of the acid portion of the amide or ester, (b) two oxygen atoms, one being from the carbonyl group and the other from a hydroxy1 ion or water molecule of the medium, and (c) the oxygen atom of the alcohol portion of an ester or the nitrogen atom of the amine portion of an amide. Transition states of such reactions are useful mental constructs that by definition, cannot be isolated, as compared to intermediates, which are isolatable or observable.

Although the above hydrolytic transition states, or any transition state, cannot be isolated, a large amount of scientific literature has been devoted to the subject. Some of that literature is discussed hereinafter.

Recently, Lerner, Tramontano and Janda [Science, 234. 1566 (1986) ] reported monoclonal antibodies to hydrolyze esters in U.S. Patent No. 4,656,567. Pollack, Jacobs and Schultz [Science, 234. 1570 (1986) ] reported a myeloma protein denominated M0PC167 [Leon et al., Biochem., 10, 1424 (1971)] that catalyzes the hydrolysis of a carbonate.

In the two Lerner and Tramontano disclosures, the antibodies were raised to a phosphonate that was synthesized to represent a stable analog of the tetrahedral hydrolytic transition state of the carboxylic acid ester or carbonate ester. The Pollack

et al. antibody principally discussed was a myeloma protein that happened to bind to a phosphonate that was structurally analogous to the carbonate analog hydrolyzed. Thus, in the Lerner and Tramontano et al. work, the substrate to be hydrolyzed was preselected, with the immunizing analog and hydrolytic antibodies being synthesized in accordance with the desired product. Pollack et al. designed the substrate to be hydrolyzed once they knew the specificity of the myeloma protein. Pollack et al. also reported (above) the existence of a catalytic antibody, substrated and analog substrate system for carbonate hydrolysis similar in concept to that of Lerner et al. Work relating to that system is reported in Jacobs et al., J. Am. Chem Soc. _r 109, 2174 (1987).

United States Patent No. 4,888,281 (Schochetman et al.) discusses the possible use of antibodies as catalysts, and presents data relating to the use of polyclonal serum in hydrolyzing o-nitrophenyl-beta-D-galactoside. The antibodies useful in that patent are said to be inducible by a reactant, a reaction intermediate or to an analog of the reactant, product or reaction intermediate. The term "analog" is there defined to encompass isomers, homologs or other compounds sufficiently resembling the reactant in terms of chemical structure that an antibody raised to an analog can participate in an immunological reaction with the reactant but will not necessarily catalyze a reaction of the analog. The data provided in that specification only indicate that some cleavage of the substrate (reactant) galactoside occurred over an eighteen hour time period using a relatively concentrated antibody preparation (1:10 and 1:20 dilutions). Although catalysis was alleged, catalytic activity was not shown since no turn

over of the allegedly catalytic antibody was shown, nor was there an indication of the percentage of substrate galactoside cleaved. The patent did indicate that beta-D-galactosidase cleaved about ten times as much substrate as did the polyclonal antibodies, presuming linearity of absorbance at the unnamed concentration of substrate studied.

From the data presented in that patent, it is possible that a nucleophilic replacement of the p_-nitrophenyl group occurred by a terminal amino group of a lysine residue of the antibody preparation used. Thus, the observed absorbance could have been due to formation of epsilon-amino lysinyl _^ -nitrophenyl aniline or to the formation of an epsilon-amino-lysinyl galactoside and p_-nitrophenol, either of which occurrences would not be catalytic since the antibody was consumed, rather than turning over.

In more recent work, bimolecular amide formation catalyzed by antibody molecules has been disclosed [Benkovic et al., Proc. Natl. Acad. Sci. USA. 85:5355 (1988)], as has an antibody-catalyzed Claisen rearrangement [Jackson et al., J. Am. Che . Soc.. 110:4841 (1988)]. None of that work, nor the previously discussed work, has contemplated the use of antibodies to catalyze any reaction in a stereospecifiσ manner.

Stereospecificity was shown in an antibody- catalyzed lactone-forming reaction [Napper et al. , Science. 237:1041 (1987)] and in an antibody-catalyzed Claisen reaction [Hilvert et al. , Proc. Natl. Acad. Sci. USA. 85:4955 (1988)].

A glycoside is an ether formed from a cyclic aldosugar (aldose) and an alcohol. The glycosidic bond is formed between the 1-position carbon atom (the C, carbon atom adjacent to the ring oxygen atom) and the carbon atom of the alcohol via an oxygen atom.

A glycosidic bond is typically stable in base, and hydrolyzable in acid. The transition state for the acid catalyzed glycosidic bond breaking (hydrolysis) is proposed to involve protonation of the ring oxygen atom and a half-chair conformation of the sugar ring containing that protonated ring oxygen atom. Partial double bond formation between the C carbon atom and the protonated ring oxygen has also been postulated. See, for example Kajimoto et al, J. Am. Che . Soc.. 113:6187- 6196 (1991).

Glycosides are present throughout nature, in both prokaryotes and eukaryotes. For example, bacterial cell walls contain glycosidic polymers, and human proteins such as antibodies and the familiar blood group antigens A, 0 and B include oligo- and polyglycosides. The repeating units of the polysaccharides pectin, a ylose, amylopectin, glycogen and cellulose are all bonded together by glycosidic bonds, as are the common disaccharides maltose, sucrose and lactose.

Brief Summary of the Invention

The present invention contemplates a receptor molecule that is a monoclonal antibody or paratope- containing portion thereof that catalytically hydrolyzes a preselected glycosidic bond of a glycoside reactant ligand, a hybridoma that secretes such a receptor, as well as methods for preparing such a hybridoma and for using the receptor molecule.

The paratope of a contemplated receptor molecule catalyzes glycosidic bond hydrolysis and binds to:

(a) a glycoside reactant ligand that includes a cyclic aldosugar (aldosaccharide) portion bonded to a ring compound portion by a preselected glycosidic bond, the cyclic aldosugar portion including a ring oxygen

atom that is bonded to the C, carbon atom, which carbon atom is also bonded to the oxygen atom of the preselected glycosidic bond; and

(b) an analog of the glycoside reactant ligand (an analog-ligand) that includes an analog cyclic aldosugar (aldosaccharide) portion and a ring portion that are linked by an a idine group, the a idine group including first and second nitrogen atoms and a carbon atom therebetween bonded to both nitrogen atoms, the first nitrogen, carbon and second nitrogen atoms of the amidine group occupying positions that are the same as the positions of the ring oxygen atom, and C, carbon atom and the oxygen atom of the glycosidic bond, respectively, of the cyclic aldosugar portion of the glycoside reactant ligand, the second nitrogen atom also linked to said ring compound portion.

A hybridoma that secretes an above receptor is also contemplated. A particularly preferred hybridoma is denominated 8D11. Also contemplated is a method of catalytically hydrolyzing a preselected glycosidic bond of a reactant ligand glycoside. This method comprises the steps of:

(a) admixing a catalytically effective amount of the monoclonal antibody molecules or paratope- containing molecules of described before with the reactant ligand glycoside in an aqueous medium to form a reaction mixture; and

(b) maintaining the admixture under biological reaction conditions and for a time period sufficient for the monoclonal antibody molecules or paratope-containing portions thereof to bind to the reactant ligand glycoside and to hydrolyze the preselected glycosidic bond.

A method of preparing the above glycosidic monoclonal receptor molecules is also contemplated.

Here, a before-described haptenic analog-ligand molecule containing a hydrolytic transition state analog is provided linked to a carrier as an immunogenic conjugate. The conjugate thus provided is dissolved or dispersed in a physiologically tolerably diluent to form an inoculum. The inoculum is introduced as by injection into a suitable, non-human mammalian host in an amount sufficient to induce antibodies to the haptenic analog- ligand. Immunoglobulin-producing cells such as those from the spleen of the immunized animal are collected and are fused with myeloma cells to form hybridoma cells. The hybridoma cells are grown in a culture medium and the supernatant medium from the growing hybridoma cells is assayed for the presence of antibodies that bind to the reactant ligand and hydrolyze the preselected glycosidic bond.

Hybridoma cells whose supernatant contains such reactive monoclonal antibodies "are collected. Those hybridoma cells are then cloned to provide the desired monoclonal antibodies from culture medium supernatant or from the ascites of a host mammal into which the hybridoma is introduced.

Detailed Description of the Invention I. Introduction

The present invention relates to molecules collectively referred to as receptors that are antibodies and idiotype-containing polyamide (antibody combining site or paratopic) portions induced by an analog of a reactant ligand glycoside that mimics the stereochemistry and conformation of the transition state in the reaction sequence for the hydrolysis of that reactant ligand glycoside. The receptor molecules (antibodies and antibody combining sites) bind to the

analog-ligand and to the reactant ligand substrate, and are thought to stabilize the hydrolytic transition state of a preselected glycosidic bond of the reactant ligand substrate and thereby exhibit catalytic properties that produce the hydrolyzed reactant ligand.

Antibodies and enzymes are both proteins whose function depends on their ability to bind specific target molecules. Enzymatic reactions differ from immunological reactions in that in an enzymatic reaction the binding of the enzyme to its substrate typically leads to chemical catalysis, whereas a non-catalytic complex is the usual result of antibody-antigen binding.

Enzymes are believed to catalyze the hydrolysis of proteins or oligo- or polysaccharides by combining with the protein to stabilize the transition state of the hydrolysis reaction. It is generally believed that the rate of an enzymatic reaction is increased relative to the rate of a non-enzymatic reaction because of the ability of the enzyme to stabilize the transition state of the reaction; i.e., to reduce the free energy of the transition state, and thus, the free energy of activation, of the reaction [Jencks, W.P., Adv. Enzvmoloσv. 43., 219 (1975) and Pauling, L., Aroer. Scientist. 36. 58 (1948)]. Support for this theory comes from the observation that substances that are thought to model the presumed transition states are often strongly bound to the enzymes as competitive inhibitors. Leinhard, G. , Science. 180. 149 (1973) and Wolfenden, R., Ace. Chem. Res.. 5_, 10 (1972) . It is further thought that the enzyme accomplishes this lowering of the reaction free energy by binding the transition state geometry of the reactant more strongly than it binds to the corresponding substrate(s) or product(s).

This means that the intrinsic binding energy of the enzyme is much greater than can be measured from the binding of substrates or products. Essentially, the binding energy of the enzyme is utilized to perform the chemical reaction [Jencks, W.P., XVII International Solvay Conference (November 1983)].

The converse proposition is that an antibody that is prepared to optimally bind a suitable analog of a transition state would function as a catalyst. The demonstration of this result by Lerner and co-workers and Schultz and co-workers in the previously cited patent and papers completes the correlation of enzyme function and antibody structure and provides a useful approach to devising artificial enzymes. The basic idea behind immunological hydrolysis described herein contemplates the use of analog-ligands in the preparation of antibodies of predetermined specificity that preferentially bind to and thereby stabilize the transition state of glycosidic bond hydrolysis upon binding to the specified reactant ligand substrate. An analog-ligand simulates the conformation of a high energy transition state in hydrolysis to induce the production of antibodies having the ability to bind related substrates and stabilize their hydrolyses.

Such preferential binding and stabilization results in a reduction in the activation energy for the hydrolysis reaction, thus meeting a criterion for catalysis. Antibodies that display this property can be obtained by immunization with synthetic analogs that are chemically modified to resemble the bonding characteristics of a substrate reactant ligand undergoing bond hydrolysis; i.e., by immunization with transition state analogs of the particular reaction.

The mechanism by which an antibody hydrolyzes an glycosidic bond of a bound reactant ligand can be thought of in terms of an "induced fit" model. As the loosely bound substrate distorts or rearranges to conform to the binding geometry of the antibody paratope, stress can be relieved by chemical reorganization of a predetermined (preselected) glycosidic bond such that this reorganization leads to the hydrolysis of the bond. The term "receptor" is used herein to mean a biologically active molecule that binds to a reactant ligand, inhibitor ligand, or analog-ligand. The receptor molecules of the present invention are antibodies, substantially intact antibodies or paratope- containing polyamide portions of an antibody.

Biological activity of a receptor molecule is evidenced by the binding of the receptor to its antigenic reactant ligand, inhibitor ligand or analog- ligand upon their admixture in an aqueous medium, at least at physiological pH values and ionic strengths. Preferably, the receptors also bind to an antigenic ligand within a pH value range of about 5 to 9, and at ionic strengths such as that of distilled water to that of about one molar sodium chloride. Idiotype-containing polyamide portions

(antibody combining sites or paratopes) of antibodies are those portions of antibody molecules that include the idiotype, and bind to the ligand or analog-ligand. Such portions include the Fab, Fab', Fv and F(ab') ₂ fragments prepared from antibodies by well-known enzymatic cleavage techniques. See for example, U.S. Patent No. 4,342,566 to Theofilopoulos and Dixon, generally, and specifically. Pollack et al., [Science, 234 _r 1570 (1987) ] who reported accelerated hydrolytic rates for Fab fragments were the same as those of the

native immunoglobulin. Inasmuch as the antibodies from which idiotype-containing polyamides are obtained are described as raised against or induced by immunogens (haptens) , idiotype-containing polyamide (antibody combining site-containing) receptors are discussed as being "raised" or "induced" with the understanding that a cleavage step is typically required to obtain an idiotype-containing polyamide from an antibody. Intact antibodies are preferred, however, and are utilized as illustrative of the receptor molecules of this invention.

The receptors useful in the present invention are monoclonal antibodies. A "monoclonal antibody" is a receptor produced by clones of a single cell called a hybridoma that secretes but one kind of receptor molecule. The hybridoma cell is fused from an antibody- producing cell and a myeloma cell or other self- perpetuating cell line.

Techniques for preparing the monoclonal antibodies of the present invention are well known. Such receptors were first described by Kohler and Milstein, Nature. 256. 495 (1975) , which is incorporated herein by reference. Monoclonal antibodies are typically obtained from hybridoma tissue cultures or from ascites fluid obtained from mammals into which the hybridoma tissue was introduced. Both methods are described herein.

A "ligand" is defined herein as a molecule that immunoreacts with or binds to a receptor molecule antibody combining site. Two types of ligand are contemplated herein. A first is termed an analog-ligand and is used as an immunogen (hapten) to induce preparation of receptor molecules and as an inhibitor of the receptor molecule-catalyzed reaction. The analog- ligand is substantially inert to undergoing the

catalyzed reaction. The second is referred to as the reactant ligand or substrate ligand and is a glycoside molecule that undergoes that catalyzed hydrolysis reaction. As described herein, chemical analogs of glycoside substrate ligands are synthesized that incorporate amidine moieties, and more particularly, amidinium moieties (protonated amidines) at specific, predetermined sites to mimic the conformation of the transition state in the hydrolysis of a glycosidic bond. Short polypeptide chains can induce the production of antibodies that recognize and bind to a homologous protein at a predetermined specific site. Phosphonates have been similarly used to induce antibodies that bind to and hydrolyze corresponding esters. The present invention carries the earlier work with polypeptides and esters a major step forward. Here, the antibodies (receptors) are induced by an immunizing haptenic first amidinium ion-containing molecule (the analog-ligand) , and recognize and bind not only to that first molecule, but also to a second, structurally similar molecule, the reactant ligand glycoside.

On binding that second molecule, the receptor causes hydrolysis of a preselected glycosidic bond that corresponds in topology to the topology of the amidinium ion-containing portion of immunizing, haptenic first molecule. The correspondence in topology; i.e., size, shape, stereochemistry and charge, provides a means for preselecting the site at which hydrolysis of the substrate ligand occurs. Inhibitor ligands that resemble the structure of an analog-ligand or a reactant ligand are also bound by receptor molecules.

Consequently, by synthesis of a relatively small, immunizing haptenic analog-ligand, one can induce the production of receptor molecules that recognize.

bind to and catalytically cleave a glycosidic bond in another molecule that can contain a plurality of glycosidic bonds. Thus, a receptor can be prepared that causes hydrolysis of a preselected (predetermined) glycosidic bond of a glycoside.

The implication of this result is that one can confer the activity of hitherto unknown glycosidases to im unoglobulins. Furthermore, the activity of the antibody combining site can be directed to any predetermined site at will by designating the glycosidic bond to be cleaved with the amidinium configuration in the haptenic analog-ligand used for immunization.

II. Transition State of Glycoside Cleavage and Hapten (Analoσ-Liσandϊ Desiσn

Design of the analog-ligand flows backward from the structure of the product to be formed through the transition state for bond cleavage to be mimicked, and then to the analog-ligand. Reactions that involve glycosidic cleavage provide an illustrative example of the general concept and are utilized herein as exemplary for a glycosidic cleavage.

Compounds with a tetrahedral configuration such as those of a phosphonate, phosphate and phosphonamidate have been used to inhibit enzymes that hydrate amide and ester bonds in a similar substrate.

Other phosphonate and phosphonamidate compounds have been useful im unogens in preparing antibodies having esterase and amidase activity. Similarly, a glucose analog having an amidinium ion group in the positions occupied by the ring oxygen, C, carbon and 1-position hydroxyl group and having a half-chair conformation has been found to be a potent inhibitor of sweet almond fS-glucosidase; that amidinium ion compound was also reported to be unstable in aqueous solution, and

therefore not practically useful as an inhibitor [Kajimoto et al., J. Am. Chem. Soc.. 113:6187-6196 (1991); Tong et al., J. Am. Chem. Soc.. 112:6137-6139 (1990)]. The structures of the analog-ligands and reactant ligands for this investigation were selected according to certain criteria. These included the availability and stability of the amidine precursors, the corresponding glycoside reactant ligand, the convenience of the chemical synthesis for its preparation, and the adaptability to diverse schemes for immunological presentation.

An analog-ligand includes an analog cyclic aldosugar portion that is linked to a ring portion. The analog cyclic aldosugar portion, by its name is an analog of an aldose, as compared to a ketose, and includes hydroxyl groups, or substituted hydroxyl groups as are commonly found on sugars.

Exemplary substituent groups for the sugar hydroxyl groups include C ₁ -C ₄ acyl such as formyl, acetyl, propionyl and iso-butryl, as well as benzoyl and benzyl. Hydrogen can also be substituted for a hydroxyl group, thereby making the aldosugar analog an analog of a deoxy sugar such as 3-deoxyglucose or 6-deoxymannose (rhamnose) .

The corresponding reactant ligand utilizes an aldosugar portion that is the same as the sugar whose analog is present in the analog-ligand. Thus, if the analog-ligand contains an analog of mannose as the aldosugar whose glycosidic bond is broken, the reactant ligand used as a substrate for the catalyzed glycosidic cleavage reaction also contains a correspondingly situated mannoside bond.

The analog cyclic aldosugar portion includes an amidine (or amidinium ion) group in the positions

occupied by the ring oxygen, C, carbon and glycosyl oxygen atom (the l-position hydroxyl of a monosaccharide) of the cyclic aldosugar.

In the amidine-containing analog ligand, one nitrogen atom (the first nitrogen) occupies the position of the ring oxygen atom of the cyclic aldosugar. That first nitrogen atom is bonded to two ring carbon atoms, one of which is at the l-position, the C _τ carbon atom, and which carbon atom is doubly-bonded to the first nitrogen atom. The other ring carbon is typically at the 4- or 5-position depending upon whether an aldofuranose or aldopyranose is being mimicked. The other amidine nitrogen atom (the second nitrogen atom) of the analog-ligand occupies the position of the 1- position hydroxyl group of a monosaccharide, or more specifically, the oxygen atom of the preselected glycosyl bond that is to be hydrolyzed in the reactant ligand. That second nitrogen atom is also linked (bonded) to the ring compound portion that is included in the analog-ligand. Upon protonation of an amidine, the unsaturation is normally considered and depicted to be spread over the two nitrogens and the carbon of the resulting amidinium ion, and is usually so shown herein. It is also noted that strict structural congruence between the ring compound portion of the analog-ligand and reactant ligand is not required. Thus, for example, where the substrate ligand contains a mono- or oligosaccharide group, or a single or fused ring system such as a phenyl or 7-hydroxy-coumarin (methylumbelliferyl) group, respectively, the ring portion of the analog-ligand can also be a single aromatic ring such as a benzyl group.

Completely specific recognition and binding are not required here and indeed, binding specificity of the paratope can be so great that the products of

glycosidic bond cleavage inhibit turnover of the catalyst. Thus, whereas antibody-antigen affinity constants on the order of 10 ⁸ or more are normally used for immunoassays, affinity constants on the order of about 10 ⁴ to about 10 ⁷ for the present receptor-reactant ligand are useful herein.

An antibody paratope can accommodate two to about five saccharide rings. As the number of rings present in an analog-ligand or reactant ligand increases, binding affinity and hydrolytic specificity can increase.

The table below lists exemplary analog-ligands and reactant ligands that in which both the aldosugar and ring compound portions of the molecules are saccharides. Standard saccharide abbreviations are utilized and are shown in parentheses for the portions whose structures are depicted.

Λpatog-Ugapςi Reactant Liαand

GlcNAc-GalNAc-Gal-Glu

The ring compound portion to which the cyclic analog aldosugar or cyclic aldosugar portions are bonded via the amidine in the analog-ligand or via glycosidic bond of the reactant ligand, respectively, contains one 6-membered ring. That 6-membered ring can be a single ring as in the case of a phenyl group or a saccharide. group, or can be a part of a fused ring system as is present in the glucosyl-7-hydroxycoumerin (umbelliferyl glucoside; Compound 7) utilized herein. The single saccharide ring that comprises the ring compound portion can also itself be linked to one or more of the same or different saccharides as are illustrated in the table above.

By including an additional carboxylic acid, alcohol, mercaptan or amine substituent in the cyclic analog aldosugar or ring compound portions of the analog-ligand, the analog-ligand can be provided with a functional appendage for coupling to an antigenic (immunogenic) carrier protein. Thus, an analog-ligand is itself typically haptenic and is coupled to an antigenic carrier protein to provide immunogeneity. The appendage and accompanying linking atoms (a linking or spacing group) can also be present in the reactant ligand, particularly where the reactant ligand is relatively small so that the antibody combining site can be relatively filled with the ligand.

An analog-ligand and its carrier protein linking atoms (linking or spacing group) that provides the necessary features for inducing catalytic receptor molecules with glycosidic activity is shown below as Compound 13.

13

The o-amino group of Compound 13 provides the appendage for linking to a carrier.

A compound used herein as a reactant ligand is illustrated below as Compound I as a mixture of α- and 3-anomers. Only the α-anomer was found to react in the present studies.

Thus, the present invention generally relates to monoclonal receptors that catalytically hydrolyze a preselected glycosidic bond of a reactant ligand glycoside. The paratope of the monoclonal receptor binds to:

(a) a glycoside reactant ligand that includes a cyclic aldosugar (aldosaccharide) portion bonded to a

ring compound portion by the preselected glycosidic bond. The cyclic aldosugar portion includes a ring oxygen atom that is bonded to the C, carbon atom, which carbon atom is also bonded to the oxygen atom of the preselected glycosidic bond; and

(b) An analog of the glycoside reactant ligand that includes an analog cyclic aldosugar (aldosaccharide) portion and a ring portion that are linked together by an amidine group. The amidine group includes first and second nitrogen atoms and a carbon atom therebetween bonded to both nitrogen atoms, the first nitrogen, carbon and second nitrogen atoms of the amidine group occupying positions that are the same as the positions of the ring oxygen atom, and C, carbon atom and the oxygen atom of the glycosidic bond, respectively, of the cyclic aldosugar portion of the glycoside reactant ligand. The second nitrogen atom is also linked to the ring compound portion. The analog cyclic aldosugar and the cyclic aldosugar of the reactant ligand are the same except for the above changes.

As noted before, the structure of an analog- ligand is analogous and not congruent with the structure of a reactant ligand. That lack of structural congruence includes the replacement of the ring oxygen, the C _t carbon atom and glycosidic bond oxygen of the reactant ligand cyclic aldosugar with the amidine first nitrogen, carbon and second nitrogen atoms, respectively, and inclusion of the group used for linking to the antigenic carrier. In addition, the ring compound portions need not contain the same rings. The radical containing the linking group can also be somewhat different from the analogous radical in the substrate ligand with that difference typically being in the length of a chain or group that includes the linking

group. Regardless of that lack of structural congruity, the reactant ligand and immunizing, analog-ligand are structurally similar enough (analogs of each other) so that the induced antibody molecules bind to both, as noted before.

An inhibitor ligand is also often used when studying the properties of a catalytic receptor. An inhibitor ligand is typically identical to an analog- ligand except that a linking group that would have an ionic charge in water at the pH values of the study is sometimes made to be free of ionic charge. For example, where the linking group of the analog-ligand is a carboxylic acid, the corresponding inhibitor ligand contains an ester or amide group of that carboxylic acid. Similarly, if the linking group is an amine as here, the inhibitor can have an amide prepared from that amine. The inhibitor ligand is free from ionic charge so that it more closely resembles the reactant ligand substrate that is also free of ionic"charge. A desired reactant ligand can be an oligo- or polysaccharide found in nature such as are noted previously. In addition, recently published work such as the disclosures of Ichikawa et al. , J. Am. Che . Soc.. 113:6300-6302 (1991) as to the enzymatic preparation of oligosaccharides and those of Danishefsky and co-workers working with chemical syntheses ΓJ. Am. Chem. Soc.. 6656-6675 (1989)] as well as the older, well known synthetic methods have pointed the way to the ready preparation of desired oligosaccharides. Further, older synthetic procedures for glycoside preparation can be found in Wyss et al., Helv. Chim. Acta. 58:1847-1864 (1975); Lemieux et al.. Can. J. Chem.. 42:532-538 (1964); Paulsen, Anσew. Chem. Int. Ed. Enσl.. .21:155-173 (1982); Schmidt, Anσew. Chem. Int. Ed. Enσl.. 25:212-235 (1986) ; and Thiem, in ACS Symposium Series 386. Horton

et al. eds., American Chemical Society (1989) pages 131-149.

An analog-ligand can be prepared in a manner generally similar to that used to prepare the reactant ligand, except for the aldosugar-amidine-linked-ring compound portion. The syntheses of a specific analog- ligand is discussed hereinafter. A more generalized synthesis is described below.

The general synthetic method utilized reacts a primary amine of the ring compound portion with a second analog of the aldosugar portion, whose ring oxygen is replaced by an -NH- group and whose C ₁ carbon atom and glycosyl oxygen are replaced by a >C=S group, thereby making the second analog aldosugar ring a thionolactam. This synthetic scheme is discussed specifically hereinafter and disclosed in Tong et al., J. Am. Chem. Soc.. 112:6137-6139 (1990). The following discussion of analog-ligands will center upon materials whose ring compound portions contain an oligosaccharide, as their syntheses are typically less straight forward than are the syntheses of other analog-ligands.

As is well known in the art, an oligo- or polysaccharide contains what are referred to as a non- reducing end, usually shown on the left, and a reducing end, usually shown on the right. The reducing end is so named regardless of whether the terminal saccharide unit is in fact a reducing sugar or not.

Where a desired analog-ligand includes one or more saccharide units linked to the non-reducing end of the cyclic analog aldosugar portion of the compound, desired glycosidically linked saccharide units can be added by synthetic chemical means following the procedures of Danishefsky and co-workers or others, or following the examples illustrated by Wong and co-workers in Ichikawa et al., J. Am. Chem. Soc..

113:6300-6302 (1991) using enzymatic methods. All that is required of the cyclic aldosugar analog portion is a free, non-sterically hindered hydroxyl group in the position desired for glycosylation. Additional enzymatic glycosylation systems are disclosed in Toone et al.. Tetrahedron. 4J5:5365 (1989) and Beyer et al. Adv. Enzymol.. 5.2:23 (1981).

Where one or more saccharide rings are desired glycosidically linked at the reducing end of ring compound portion of the analog-ligand, it is usually preferable to add those rings prior to formation of the amidine group that links both parts of the analog-ligand molecule. Again, one of the before-described oligoglycosyl synthetic methods can be used to construct the desired oligosaccharide. Those syntheses typically proceed from reducing end toward non-reducing end.

The non-reducing end terminal saccharide of the oligosaccharide ring compound portion is prepared to include a free or protected primary amino group, typically at the 4- or 6-position (for a pyranose) . When present, the amine protecting group is removed prior to reaction with the thiolacta . Typical amine blocking groups that are readily removed are cyclic imides formed from maleic, succinic, 2-methyl-succinic or phthalic acids. The cyclic imide protecting groups are readily removed with hydrazine to provide the free primary amine.

METHODS In another embodiment, this invention relates to a method of catalytically hydrolyzing a preselected glycosidic bond in a glycoside reactant ligand molecule. The method comprises the steps of: (a) admixing a catalytically effective amount of one of the foregoing receptors with glycoside reactant ligand molecules in an

aqueous medium; and (b) maintaining the admixture under biological reaction conditions and for a period of time sufficient for the reactant ligand molecules to bind to the receptors and for the receptor molecules to hydrolyze the preselected glycosidic bond of the reactant ligand. The products of that hydrolysis can be thereafter recovered, if desired.

A hydrolytic method of this invention utilizes an aqueous medium as a portion of the reaction admixture. That medium typically contains water and buffer salts. In addition, the medium can contain other salts such as sodium chloride, as well as water-soluble calcium and magnesium salts as are frequently found in protein-containing media. Organic solvents such as methanol, ethanol, acetonitrile, dimethyl sulfoxide, dioxane, hexamethylphosphoramide and N,N-dimethylforamide can also be present. Surface active agents that emulsify the reactant ligand and receptor molecule can also be present. The critical feature of ingredients present in the aqueous medium is that those ingredients not substantially interfere with or inhibit the catalytic reaction as by denaturation of the receptor molecule. Additionally, the aqueous medium is substantially free from salt, proteins generally, and enzymes, specifically, that inhibit the bond-breaking reaction catalyzed by the receptor molecule.

The aqueous medium typically has a pH value of about 5 to about 9, and preferably about pH 6.0 to about 8.0. pH Values greater and less than those recited values can also be utilized so long as the catalyzed reaction is again not substantially interfered with or inhibited.

The catalytic reactions are typically carried out at ambient room temperature; i.e., at about 20 to about 25°C or at 37°C, and at an ambient atmospheric

pressure; i.e., at about one atmosphere. However, temperatures down to about the freezing point of the aqueous medium and up to about the boiling point of the medium at atmospheric pressure can also be used. As is known, proteins such as the receptor molecule tend to denature at elevated temperatures such as those at which an aqueous medium boils, e.g., at about 100°C and thus temperatures below about 40°C are preferred. As is also well known, reactions that follow multimolecular kinetic expressions decrease in rate as the temperature decreases. Thus, a minimal temperature of about 15°C is preferred.

The above conditions of aqueous medium, temperature, pH value, and atmospheric pressure constitute "biological reaction conditions".

The reactant ligand is present in a reaction mixture in an amount up to its solubility in the aqueous medium. A two phase system that includes insoluble reactant ligand can also be used, but normally is not so used. Normally used concentrations of the reactant ligand are about 0.1 micromolar (μM) to about 10 millimolar (mM) , with that amount also being a function of the solubility of the reactant ligand in the solvent medium. Where the product is desired, per se, relatively higher concentrations are used as compared to lower concentrations where a reaction mechanism or reaction kinetics are to be studies.

An effective amount of the receptor molecule is also present. That effective amount is typically a catalytic amount; i.e., the receptor is used at a molar ratio to the reactant ligand of about 1:2 to about 1:10,000, with a molar ratio of about 1:10 to about 1:100 being preferred. The ratio of receptor molecule to reactant ligand typically depends upon the specific activity of the receptor molecule toward the reactant

ligand and the purpose of the user in running the reaction.

Thus, where the product is desired, a relatively higher concentration of receptor and higher receptor to reactant ligand ratio are used. Where the reaction mechanism or kinetics of the reaction are being studied, a lower concentration and ratio are typically used. A stoichiometric amount of receptor or more can also be used, but since the receptor is a catalytic molecule, use of even a stoichiometric amount can be wasteful. Thus, at least a catalytic amount of the receptor is utilized.

The admixture formed from mixing receptor molecules and reactant ligand in an aqueous medium is maintained for a time period sufficient for the binding and reaction to occur. The duration of that maintenance period is a function of several parameters including the receptor and reactant ligand selected, their concentrations, pH value, and temperature, as well as what is being sought from the reaction.

Thus, where kinetic studies are being carried out, maintenance times of minutes to hours are frequently encountered. Where the reaction products are desired, maintenance times of hours to days are more usual.

III. Preparation of Coniuσates and Inocula

Conjugates of haptenic analog-ligand molecules with antigenic (immunogenic) protein carriers such as bovine serum albumin (BSA) can be prepared, for example, by activation of the carrier with a coupling agent such as MBS (m-maleimidobenzoyl-N-hydroxy succini ide ester) , and coupling to the thiol group of the analog-ligand. See, for example, Liu et at., Biochem.. 80. 690 (1979). As is also well known in the art, it is often beneficial

to bind a compound to its carrier by means of an intermediate, linking group.

Useful carriers are well known in the art and are generally proteins themselves. Exemplary of such carriers are keyhole limpet hemocyanin (KLH) , edestin, thyroglobulin, albumins such as bovine serum albumin or human serum albumin (BSA or HSA, respectively) , red blood cells such as sheep erythrocytes (SRBC) , tetanus toxoid, cholera toxoid as well as polyamino acids such as poly(D-lysine:D-glutamic acid), and the like. The choice of carrier is more dependent upon the ultimate intended use of the antigen than upon the determinant portion of the antigen, and is based upon criteria not particularly involved in the present invention. The carrier-hapten conjugate is dissolved or dispersed in an aqueous composition of a physiologically tolerable diluent such as normal saline, PBS, or sterile water to form an inoculum. An adjuvant such as complete or incomplete Freund's adjuvant or alum can also be included in the inoculum. The inoculum is introduced as by injection into the non-human mammal such as a laboratory mouse or rat used to raise the antibodies in an amount sufficient to induce antibodies, as is well known. In an exemplary procedure, 2.5 mg of a reaction product of haptenic analog-ligand containing an added alcohol or amine group for linking purposes bonded to a reacted succinimidyl adipoyl anhydride or succinimidyl glutaroyl anhydride in 250 μl of dimethylfor amide is slowly added to 2 mg of KLH in 750 μl of 0.01 M sodium phosphate buffer at a pH value of 7.2. A temperature of 4°C is utilized and the resulting admixture is stirred for about one hour to form the hapten-linked KLH conjugate. The conjugate reaction product so formed is thereafter purified by usual means.

In the present work Compound 13 was conjugated to BSA (cBSA "SuperCarrier", Pierce Chemical Co.) following the procedures of Bauminger et al. , Methods♦ Enz. _r 70.:151 (1980). The resulting conjugate was used to immunize mice as discussed below.

IV. Preparation of Monoclonal Receptors

Balb/c mice (6-8 weeks) were immunized with 13-cBSA (Cl antigen, cBSA is sold under the trade name of "SuperCarrier" by Pierce Chemical Company, Rockford, IL) conjugate (Cl-conjugate; 1 mg/ml) emulsified 1:1 with complete Freund's adjuvant (200 μl, subcutaneous injection) . Boosts were administered with incomplete Freund's (intraperitoneal injection) at intervals of three and five weeks from the primary immunization. A final boost of Cl-conjugate (100 μg) in 200 μl phosphate buffered saline (PBS) was injected intravenously at seven weeks. The mouse was sacrificed four days later and the spleen was removed aseptically. Spleen cells were fused with myeloma (FO) cells in a 3:1 ratio in the presence of PEG 8000 [Harlow, et al., Antibodies - A Laboratory Manual. Cold Spring Harbor Laboratory (1988)]. The fusion products were diluted into 180 ml of medium [DMEM-high glucose, 20 percent fetal bovine serum (FBS), 2x OPI,

L-glutamine], and plated on 15X96-well tissue culture plates at 100 μl/well. After 24 hours, an additional 100 μl of the same medium supplemented with 2x HAT were added to each well. One week later, the medium was removed by aspiration and replaced with fresh medium supplemented with lx HT and gentamycin (50 μg/ml) . After five days, the medium was removed by aspiration and replaced with phenol-red deficient DMEM-high glucose, lx OPI, L-glutamine and 10 percent FBS that was previously heated at 56°C for one hour (200 μl/well) .

This step was repeated once more 24 hours later. The cell culture supernatants were harvested 1-2 days later and used directly in the assay for glycosidase activity. Cultures were expanded into tissue culture flasks in standard medium (DMEM-high glucose/10 percent FBS, L-glutamine, and genta ycin) .

The lymphocytes employed to form the hybridomas of the present invention can be derived from any mammal, such as a primate, rodent (e.g., mouse or rat) , rabbit, guinea pig, cow, dog, sheep, pig or the like. As appropriate, the host (here a mouse) can be sensitized by injection of the immunogen, in this instance a.haptenic analog-ligand coupled to a carrier as a conjugate, followed by a booster injection, and then isolation of the spleen.

It is preferred that the myeloma cell line be from the same species as the lymphocytes. Therefore, fused hybrids such as mouse-mouse hybrids [Shulman et al.. Nature. 276. 269 (1978)] or rat-rat hybrids [Galfre et al.. Nature, 272, 131 (1979)] are typically utilized. However, some rat-mouse hybrids have also been successfully used in forming hybridomas [Goding, "Production of Monoclonal Antibodies by Cell Fusion," in Antibody as a Tool. Marchalonis et al., eds., John Wiley & Sons Ltd., p. 273 (1982)]. Suitable myeloma lines for use in the present invention include MPC-11 (ATCC CRL 167), P3X63-Ag8.653 (ATCC CRL 1580), Sp2/0-Agl4 (ATCC CRL 1581), P3X63Ag8U.l (ATCC CRL 1597), Y3-Agl.2.3. (deposited at Collection Nationale de Cultures de Microorganisms, Paris, France, number 1-078) and

P3X63Ag8 (ATCC TIB 9) . The non-secreting murine myeloma line Sp2/0 or Sp2/0-Agl4 is preferred for use in the present invention.

A monoclonal receptor of the present invention is preferably produced by introducing, as by injection.

the hybridoma into the peritoneal cavity of a mouse. Preferably, syngeneic or se i-syngeneic mammals such as mice are used, as in U.S. Patent 4,361,549, the disclosure of which is incorporated herein by reference.

The introduction of the hybridoma causes formation of antibody-producing hybridomas after a suitable period of growth, e.g. 1-2 weeks, and results in a high concentration of the receptor being produced that can be recovered from the bloodstream and peritoneal exudate (ascites) of the host mouse. Although the host mice also have normal receptors in their blood and ascites, the concentration of normal receptors is typically only about five percent that of the monoclonal receptor concentration.

Monoclonal receptors are precipitated from the ascitic fluids, purified by anion exchange chromatography, and dialyzed against three different buffers. The resulting solutions containing isolated Ig fractions are typically prepared into stock solutions of receptor at 1-20 mg/ml using an appropriate buffer such as 50 mM Tris-HCl or sodium phosphate containing 0.01 percent sodium azide.

Of the original fifteen plates of anti- Compound 13 monoclonal receptors prepared one hybridoma whose secreted monoclonal antibody catalyzed the hydrolysis of reactant ligand Compound I in the later- discussed Glycosidase Activity Assay was selected for further study as being the most active. The hybridoma that produces the catalytic monoclonal receptor, given laboratory designation 8D11, was deposited at the American Type Culture Collection, 12301 Parklawn Drive, Rockville, MD on October 2, 1991 and was given ATCC accession number HB 10890.

The present deposit was made in compliance with the Budapest Treaty requirements that the duration of the deposit should be for 30 years from the date of deposit or for five years after the last request for the deposit at the depository or for the enforceable life of a U.S. patent that matures from this application, whichever is longer. The hybridoma will be replenished should it become non-viable at the depository.

A Fab fragment of a monoclonal receptor can be prepared from the purified receptor using predigested papain in a 0.1 M sodium acetate buffer, at a pH value of 5.5, at 37°C, followed by reaction with iodoacetamide. The Fab fragment is typically further purified by anion exchange chromatography, dialysis, and DEAE anion exchange chromatography, and its homogeneity is judged by gel electrophoresis.

For another preparation of the receptor molecules, the gene that encodes an antibody combining site-forming fragment can be obtained from any cell that produces an antibody molecule that immunoreacts as discussed herein. A preferred cell is a hybridoma cell.

For examples of general recombinant DNA cloning methods, see Molecular Cloninσ. Maniatis et al.. Cold Spring Harbor Lab., N.Y., 1982; DNA Cloninσ. Glover, ed. , IRL Press, McLean VA (1985). For the genomic cloning and expression of immunoglobulin genes in lymphoid cells, see Neuberger et al.. Nature. 312:604-8 (1984); Ochi et al., Proc. Natl. Acad. Sci. USA. £0:6351-55 (1987); and Oi et al., Proc. Natl. Acad. Sci. USA. JQ.825-29 (1983). For cloning of immunoglobulin genes from hybridoma cells and expression in Xenopus oocytes, see Roberts et al. , Protein Enσineerinσ. 1:59-65 (1986), and see Wood et al. for expression in yeast. Nature. 3_14.:446-9 (1985).

Synthesis of an Exempt«τ Tmmunoσenic Analoσ-Liσand

2,3,4,6-Tetra-O-benzyl-α-D-glucopyranose (Compound 1) was prepared from methyl-D-glucoside according to the procedure of Perrine et al., J. Qrσ. Chem.. 22:664 (1967). This material is oxidized with DMS0/Ac ₂ 0 to 2,3,4,6-tetra-O-benzyl-D-gluconolactone (Compound 2) as described by Kazuhara et al., J. Qrσ. Chem.. 3_2:2531 (1967) . Lactone (Compound 2) was ring- opened with dimethylamine in dry diethyl ether and then oxidized to form 2,3,4,6-tetra-0-benzyl-N,N-dimethyl-D- xylo-5-hexulosonamide (Compound 3) with DMS0/Ac ₂ 0 [Kazuhara et al., J. Qrσ. Chem.. 3_2.:2535 (1967)]. (In the following structural formulas, Bn is benzyl. Me is methyl and Ac is acetyl.)

The 5-ketogluconamide (Compound 3) was then converted to 2,3,4,6-tetra-O-benzyl-N,N-dimethy1-D-xylo- 5-hexulosonamide oxime (Compound 4) by refluxing Compound 3 (10 mmol) with hydroxylamine hydrochloride (11 mmol) and sodium methoxide (11 mmol) in methanol for four hours. After filtration and removal of the solvent by rotary evaporation. Compound 4 was purified by chromatography on silica gel (20:1 CH ₂ C1 ₂ :CH ₃ 0H) to give white solid. Rf=0.28 (20:1 CH ₂ C1 ₂ :CH ₃ 0H) ; chars ochre yellow; FAB m/z 597 (M+H ⁺ ) ; 77 percent yield from Compound 1. Compound 4 can be recrystallized with little loss in 1:1 hexane:ethyl acetate (white flakes).

Oxime Compound 4 was catalytically reduced to an epimeric mixture of 5-amino sugars by stirring Compound 4 (15 mmol) in ethanol (120 mL) containing Raney Nickel (Aldrich W-2, 10 mL) and 2 g KOH under H ₂ for ten days at room temperature. Following removal of the catalyst by filtration (celite) and concentration in vacuo, the crude mixture of 2,3,4,6-tetra-0-benzyl-N,N- dimethyl-D-5-aminogluconamide (Compound 5) and 2,3,4,6- tetra-O-benzyl-N,N-dimethyl-D-5-aminoidonamide (Compound 6) was redissolved in methylene chloride, washed with aqueous NaOH (1 mM) and purified by flash chromatography on silica gel (100:5:1 CH ₂ Cl ₂ :CH ₃ OH:Et ₃ N) to afford a clear oil. R _f =0.23 (100:5:1 CH ₂ Cl ₂ :CH ₃ OH:Et ₃ N) ; chars dark red; FAB m/z 537 (M+H ⁺ ) .

To effect transamidation with ring closure, aminoamide Compounds 5 and 6 were heated to 120°C under reduced pressure (2 mm Hg) for six hours to afford 2,3,4,6-tetra-O-benzyl-D-glucono-l,5-lactam (Compound 7) and 2,3,4,6-tetra-0-benzyl-D-idono-l,5-lactam (Compound 8) . The epimeric lacta s were separated by silica gel chromatography (1:1 hexane:ethyl acetate) into Compound 7 (R _f =0.32) and Compound 8 (R _f =0.25) in ratio of 1:5 (7:8) and 18 percent combined yield from Compound 4. Catalytic hydrogenation in acetic acid (10 percent palladium/C, H ₂ (1 atmosphere) , room temperature, two days) afforded D-glucono-1,5-Lactam (Compound 9). R _f =0.20 (20:4:1 CH ₃ CN:H ₂ 0:CH ₃ C00H) ; chars purple; FAB m/z 178 (M+H ⁺ ) ; in quantitative yield from Compounds 7 and 8.

¹ H NMR specta of the tetra-O-acetyl derivatives of Compounds 9 and 10 (Compounds 9a and 10a) are fully consistent with the proposed structures.

¹ H NMR of Compound 9a:£ (CDC1 ₃ ) 2.07 (s, 3H) , 2.10 (s, 3H), 2.12 (s, 3H) , 2.15 (s, 3H) , 3.80 (ddd, IH, J=2.8, 6.7, 9.4 Hz) , 3.96 C<* , IH, J=6.7, 11.7 Hz) , 4.29 (dd, IH, J=2.8, 11.7), 5.09 (d, IH, J=9.4 Hz), 5.23 (dd, IH, J=9.4, 9.4 Hz), 5.58 (dd, IH, J=9.4, 9.4 Hz), 6.07 (broad s, IH) .

IH NMR of Compound 9b: (CDC1 ₃ ) 2.10 (s, 3H) , 2.12 (s, 3H), 2.13 (s, 3H) , 2.16 (s, 3H) , 4.00 (m, IH) ,

4.14 (dd, IH, J=7.2, 11.5 Hz) , 4.32 (dd, IH, J=3.7, 11.5), 5.26 (dd, IH, J=4.4, 7.3 Hz), 5.26 (d, IH, J=7.3 Hz), 5.46 (dd, IH, J=7.3, 7.3 Hz), 6.35 (broad s, IH) .

Gluconolactam Compound 9 was converted to the corresponding thionogluconolactam (Compound 11) according to the procedure of Tong et al. , J. Am. Chem. Soc.. 112:6137-6138 (1990). Compound 11 (52 mmol) was then treated with 2-aminobenzylamine (900 mmol) in dry methanol (0.5 L) containing 4λ sieves to, afford after chromatography on silica gel (20:4:1 CH ₃ CH:H ₂ 0:CH ₃ C00H) the 2-aminobenzylamidine Compound 12, in the form of its acetate salt. R _f =0.09 (20:4:1 CH ₃ CN:H ₂ 0:CH ₃ C00H) ; chars brown; FAB m/z 282 (M ⁺ ) . H NMR indicated that Compound 12 is a mixture of cis and trans isomers with respect to the benzyl substituent.

12

¹ H NMR of Compound 12:5 (D ₂ 0) 2.04 (s) , 3.72 (m, 2H) , 3.92 (dd, 2H, J=7.5, 12.1 Hz), 4.00 (dd, IH, J=1.6, 8.4 Hz), 4.08 (dd, IH, J=4.2, 12.1 Hz) , 4.09 (dd, IH, J=4.2, 12.1 Hz), 4.21 (dd, IH, J=3.6, 3.6 Hz), 4.27 (dd, IH, J=3.6, 5.0 Hz), 4.35 (dd, IH, J=1.6,' 5.4 Hz), 4.69 (d, IH, J=3.6 Hz), 4.90 (d, IH, J=3.6 Hz), 4.93 (d, 4H, J=9.3 Hz), 7.18 (d, IH, 7.7 Hz), 7.19 (d, IH, 7.7 Hz), 7.26 (d, 2H, J=7.3 Hz), 7.36 (t, 2H, J=7.3 Hz), 7.42 (t, 2H, J=7.3 Hz).

A six carbon spacer (linking) group with terminal carboxyl group was attached by stirring Compound 12 with adipic anhydride in water and isolating the product (Compound 13) by chromatography on silica gel. R _f =0.28 (20:4:1 CH ₃ C :H ₂ 0:CH ₃ C00H) ; chars brown; FAB m/z 410 (M ⁺ ) . Conjugation of Compound 13 to the BSA carrier protein (to produce the "C1 ^M antigen) was done according to the dicyclohexylcarbodiimide/N- hydroxysuccinimide method of Bauminger et al.. Methods Enz.. 70.:151 (1980).

13

Glycosidase Activity Assay:

Tissue culture supernatants from the original 15 plates of hybridomas were pooled in groups of eight (100 μl each) , and 300 μl of each pool were diluted with 300 μl of assay buffer (200 mM MES, 200 mM NaCl, 0.02 percent sodium azide, pH 6.1) containing a mixture of a- and _/ 9-methylumbelliferyl glucosides (1 mM final concentration of each; avaiable from Sigma Chemical Co.) . The mixtures were incubated at room temperature in a polystyrene fluorimeter cuvette. Fluorescence

readings (λ _eχ 365, λ^, 440) were recorded over a 24 hour period and a rate of hydrolysis was inferred from the fluorescence change over time.

Individual hybridoma cultures from a pool showing the greatest increase over the background were reassayed individually. One culture (8D11) in this group showed significant activity above the rest. A group of 20 randomly chosen hybridoma culture supernatants were also assayed in the same way. The mean rate for these samples was 1.5 x 10 ^*6 M/hr, ranging between 1.2-1.7 x 10 ^"6 M/hr. Medium alone that was not incubated over hybridoma cultures gave a similar rate of 0.7 x 10 ^"6 M/hr.

The rate from the unique culture (8D11) showed a rate of 3.4 x 10 ^*6 M/hr or better than double the background rate. The activity was shown to be specific for the α-glucoside, and no acceleration over background was observed for any clone with the ,9-glucoside.

The foregoing is intended as illustrative of the present invention but not limiting. Numerous variations and modifications can be effected without departing from the true spirit and scope of the invention.