EP0891548A1 - Sulphur ratio tagging method for combinatorial libraries - Google Patents

Abstract

Invented is a method of preparing combinatorial libraries and combinatorial libraries prepared thereby. Also invented is a method for identifying compounds having desired characteristics from a combinatorial library. Also invented is a method for encoding combinatorial libraries using beads encoded with sulfur identifiers.

Description

SULFUR RATIO TAGGING METHOD FOR COMBINATORIAL LIBRARIES

FIELD OF THE INVENTION The field of this invention concerns combinatorial chemistry which involves the syntheses of one or more encoded combinatorial libraries where large numbers of products having varying compositions are obtained. This invention also relates to methods of encoding combinatorial libraries.

BACKGROUND OF THE INVENTION In the continuing search for new chemical moieties that can effectively modulate a variety of biological processes, the standard method for conducting a search is to screen a variety of pre-existing chemical moieties, for example, naturally occurring compounds or compounds which exist in synthetic libraries or databanks. The biological activity of the pre-existing chemical moieties is determined by applying the moieties to an assay which has been designed to test a particular property of the chemical moiety being screened, for example, a receptor binding assay which tests the ability of the moiety to bind to a particular receptor site.

In an effort to reduce the time and expense involved in screening a large number of randomly chosen compounds for biological activity, several developments have been made to provide libraries of compounds for the discovery of lead compounds. The chemical generation of molecular diversity has become a major tool in the search for novel lead structures. Currently, the known methods for chemically generating large numbers of molecularly diverse compounds generally involve the use of solid phase synthesis, in particular to synthesize and identify peptides and peptide libraries. See, for example, Lebl et al., Int. J. Pept. Prot. Res., 41, p. 201 ( 1993) which discloses methodologies providing selectively cleavable linkers between peptide and resin such that a certain amount of peptide can be liberated from the resin and assayed in soluble form while some of the peptide still remains attached to the resin, where it can be sequenced; Lam et al., Nature, 354, p. 82 ( 1991 ) and (WO 92/00091 ) which disclose a method of synthesis of linear peptides on a solid support such as polystyrene or polyacrylamide resin; Geysen et al., J. Immunol. Meth., 102, p. 259 (1987) which discloses the synthesis of peptides on derivatized polystyrene pins which are arranged on a block in such a way that they correspond to the arrangement of wells in a 96-well microtiter plate; and Houghten et al., Nature, 354, p. 84 ( 1991 ) and WO 92/09300 which disclose an approach to de novo determination of antibody or receptor binding sequences involving soluble peptide pools. The major drawback, aside from technical considerations, with all of these methods for lead generation is the quality of the lead. Linear peptides historically have represented relatively poor leads for pharmaceutical design. In particular, there is no rational strategy for conversion of a linear peptide into a non-peptide lead. As noted above, one must resort to screening large databanks of compounds, with each compound being tested individually, in order to determine non-peptide leads for peptide receptors.

It is known that a wide variety of organic reactions can be carried out on substrates immobilized on resins. These include, in addition to peptide synthesis reactions which are well known to those of ordinary skill in the art, nucleophilic displacements on benzylic halides, halogenation, nitration, sulfonation, oxidation, hydrolysis, acid chloride formation, Friedel-Crafts reactions, reduction with LiAlH4, metallation, and reaction of the organometallic polymer with a wide variety of reagents. See, for example, N. K. Mathur et al., Polymers as Aids in Organic Chemistry, Academic Press, New York, p. 18 (1980). In addition, Farrall et al., J. Org. Chem., 41, p. 3877 (1976) describe the experimental details of some of these reactions carried out with resins.

Nonpeptidic organic compounds, such as peptide mimetics, can often surpass peptide ligands in affinity for a certain receptor of enzyme. An effective strategy for rapidly identifying high affinity biological ligands, and ultimately new and important drugs, requires rapid construction and screening of diverse libraries of non-peptidic structures containing a variety of structural units capable of establishing one or more types of interactions with a biological acceptor (e.g., a receptor or enzyme), such as hydrogen bonds, salt bridges, pi-complexation, hydrophobic effects, etc. However, work on the generation and screening of synthetic test compound libraries containing nonpeptidic molecules is now in its infancy. One example from this area is the work of Ellman and Bunin on a combinatorial synthesis of benzodiazepines on a solid support (J. Am. Chem. Soc. 114, 10997, (1992); see Chemical and Engineering News. January 18, 1993, page 33).

A key unsolved problem in the area of generation and use of nonpeptide libraries is the generation and use of nonpeptide libraries is the elucidation of the structure of molecules selected from a library that show promising biological activity. An attempt to uncover the structures of peptides selected from a library using unique nucleotide sequence codes, which are synthesized in tandem with the peptide library, has been described by Brenner and Lerner (Brenner, S. and Lerner, R.A. Proc. Nat'l. Acad. Sci. USA. 1992 89 . 5381-5383). The nucleotide sequence of the code attached to each peptide must be amplifiable via the polymerase chain reaction (PCR). However, nucleotide synthesis techniques are not compatible with all of the synthetic techniques required for synthesis of many types of molecular libraries. Furthermore, the close proximity of nucleotide and synthetic test compound in the library, which can result in interactions between these molecules interfering with the binding of the ligand with a target receptor of enzyme during the biological assay, also limits this approach. The nucleotide component of the library can also interfere during biological assays in a variety of other ways. Kerr et al. (J. Am. Chem. Soc. 1993, 115, 2520-2531) reported synthesizing solution phase libraries of peptides, containing non-natural amino acid residues, in parallel with peptide coding strands. The peptide ligand and its coding strand in this library are covalently joined together, which allows isolation and sequence determination of pairs of synthetic test compound and corresponding code. However, as with the nucleic-acid-encoded library described by Brenner and Lerner, above, the coding peptide may interfere with the screening assay.

PCT/US93/09345 describes a method of identifying actives in a combinatorial library by attaching multiple tags in a predetermined binary coding system. PCT/HU93/0030 describes fluorescently labeled sub-library peptide kits for use in peptide synthesis.

PCT/US94/06078 describes methods of encoding combinatorial libraries using polymeric sequences.

Shearer et al. (Journal of Chromatographic Science. Vol. 31. March 1993 and Anal. Chem. 1992, 64, 2192-2196) describes the application of gas chromatography and flameless sulfur chemiluminescence detection to the analysis of petroleum products.

Many of the disadvantages of the known methods as well as many of the needs not met by them are addressed by the present invention which, as described more fully hereinafter, provides numerous advantages over the known methods.

SUMMARY OF THE INVENTION This invention relates to a method of encoding combinatorial libraries which comprises utilization of sulfur identifiers. This invention also relates to a method of encoding each choice of a combinatorial library with sulfur identifiers and combinatorial libraries encoded thereby. This invention also relates to beads with sulfur identifiers attached thereto. This invention also relates to beads with varying ratios of sulfur identifiers attached thereto.

Detailed Description of the Invention

As used herein, the term "beads" means any solid support material capable of providing a base for combinatorial syntheses, such as 1 to 2% crosslinked polystyrene, polyacrylamide, polyethylene glycol polystyrene co-polymer, preferably polystyrene modified to incoφorate a polyethylene glycol side chain. As used herein, the term "tag", unless otherwise indicated, means an encoding characteristic of a bead or group of beads, such as differences in size, differences in material composition or differences in flow properties using varying ratios of sulfur identifiers.

As used herein, the term "sulfur identifier" or "identifier" means a coding label attached to a bead or group of beads by adding varying ratios of sulfur moieties, preferably two, different sulfur moieties in varying ratios.

As used herein, the term "intensity-differentiated" means an identifier (as used herein) in which varying ratios of different sulfur moieties are added to a bead or group of beads. As used herein, the term "choice" means the alternative variables for a given stage in a combinatorial synthesis (not limited to peptide chemistry), such as reactant, reagent, reaction conditions, and combinations thereof. Where the term "stage" corresponds to a step in the sequential synthesis of a compound or ligand; the compound or ligand being the final product of a combinatorial synthesis. The term "registry", as used herein, has the same meaning as the term "stage" as indicated above.

The development of elegant mass spectral methods for structural determination of compounds from bead libraries has made reaction identification tags less necessary for modest size libraries. However, as large libraries are developed mass redundancy becomes unavoidable making MS less useful.

Examination of MS fragments may alleviate some of the problem, but this is time consuming, and is not always be successful. As disclosed herein, tags which reveal the reaction history of a particular bead either by themselves, or if not all reaction steps are labeled, in conjunction with MS are particularly useful. Recently several convenient selective sulfur GC detectors (flameless sulfur chemiluminescence detectors) have been marketed at least one of which has a >5% accuracy starting at low picogram levels and extending over 3 to 4 orders of magnitude. These detectors do not respond to carbon or nitrogen thus the analyses are relatively free of interferences. Since the detector measures only the sulfur content of the peak, it is independent of the structure of the sulfur containing compounds. Thus, the present invention provides for the identification of a large number of events on resin beads by the sulfur ratios and retention times with as few as 2 or 3 tags.

A typical large bead library will contain beads containing about 15 nmol of functionality for attaching ligands. The amounts of 1 to 1000 pg of sulfur can currently be determined accurately enough so that factors of 2 can be distinguished. Further, the individual identifiers will have very similar reactivity for attachment and subsequent cleavage, thus, by varying two identifiers, 21 unique ratios of tags are available with a ratio range of 0.001 to 1000. Using 3 identifiers in one step gives 126 identifiable ratios. Thus in a symmetrical 2 step sublibrary synthesis this gives information on 15,876 (126 x 126) reaction histories using 6 tags. These sublibraries can be mixed and a third step carried out with numerous reagents to give very large sublibraries. For example, 100 third step reagents could give a tagged library of 1,586,600 (15,876 x 100) compounds using only 6 tags.

To obtain 1000 pg of sulfur on the column and allowing for wastage would require about 2000 pg of sulfur on the bead. If fewer ratios are needed, the required amount of tag is greatly reduced. Thus 15 ratios need a maximum of 250 pg (0.008 nmole) of sulfur, and 3 tags can identify 90 reaction histories.

Thus, 3 tags including an allowance for wastage would require 1000 pg (0.032 nmole) of sulfur. This compares to about 15 nmol of functionality on a 250 micron bead and 0.15 nmol on a typical commercial polyethylene glycol (PEG) - polystyrene resin. For a symmetrical library tagging only the first 2 steps, 6 tags could identify the reaction histories of 8100 sublibrary compounds, and for example, 81,000 compounds after a third step using only ten differentiating reagents.

An alternative strategy to decrease the required number of tags is to tag only beads which will carry products with redundant molecular weights which cannot be differentiated by mass spectral analysis.

In the case where the final ligands will be cleaved from the resin before analysis, the first tag may be attached to the resin by the same linker system as used by the ligand. The tag will be in the ligand bioassay solution but at its low concentration will not interfere with the bioassay, and since the tag detector is sulfur selective and retention time selective, the ligand will generally not interfere with tag analysis. In carrying out the synthesis to prepare a library, one may initially begin with a large number of beads, usually at least 10^, more usually at least 10^, and desirably at least 10^, while generally not exceeding at least 10^, more usually not exceeding at least 10 0, characterized in that the beads are separated into groups, the beads within each group being untagged or similarly tagged with each group being uniquely tagged by an identifier, or one group being untagged and each of the remaining groups being uniquely tagged by an identifier. The number of groups of beads will correspond to the number of choices in the first differentiating step. The entirety of each group is entered into a separate container. One can use microtiter well plates, flasks, Merrifield synthesis vessels, etc. The beads will usually be divided up into groups of at least one bead each, usually a plurality of beads, generally 1000 or more, and may be 10^ or more depending on the total number of registries involved in the library and the number of beads desired for each final compound. One would then add the appropriate agents to each of the individual containers to process them in stages (or "registries" as used herein). If the beads in the first stage are untagged, each pool is then reacted with a unique identifier which encodes the reagent and stage, followed by combining all of the beads into a single mixture and then separating the beads according to the number of choices for the next registry. If the beads in the first stage are tagged, the pools are combined into a single mixture and then separated according to the number of choices for the next registry. The procedure of dividing beads, followed by a synthetic stage (to form a registry) including a tagging step, and then recombining beads is iterated until the combinatorial library is completed. In some instances, the same reaction may be carried out in a manner to enhance the proportion of product having a particular substituent in a particular registry as compared to the other choices. In other instances, one or more of the registries may involve a portion of the beads being set aside and undergoing no reaction, so as to enhance the variability associated with the final product. In other situations, batches may be taken along entirely different synthetic pathways. The library thus prepared will contain tagged beads which identify the reaction sequence of each choice therein.

Intensity-differentiated sulfur-labeled beads can be prepared by derivatizing pools of beads with varying amounts of the sulfur moieties. Additionally, multiple, preferably 2 or 3, sulfur tags can be used in varying ratios to encode beads. This is preferably implemented, for example, by varying the stoichiometry of a first sulfur tag (A) and a second sulfur tag (B), such as A:B = 1:1, 1 :2, 2: 1, 1:4, 4: 1, etc., in the tagging steρ(s).

As used herein, intensity-differentiated sulfur-labeled beads can be prepared by the method outlined in Scheme 1 below.

Scheme 1

substrate

R=Lower alkyl; R 1 = Lower alkyl, halogen, -OC substrate = bead on which ligand is being synthesized As used in Scheme 1 a sample of beads, preferably a polyethylene glycol modified polystyrene resin, 100 - 300 micron particles, is used. The beads are divided into N pools. The members of each pool are derivatized with an identifier unique to that pool. The R and R groups of 2 are chosen such that the final 4- hydroxbenzenesulfonamides of 4 are separable by capillary GC either directly or after derivatization. Solutions of known concentrations of two or three different examples of 2 are prepared and aliquots are mixed to give solutions whose component ratios are precisely known. These solutions are used to tag the various batches of resin by generating the carbenes using, for example, Rh(TFA)2- The reaction history of individual beads of interest is determined by removing the tags from the beads by irradiation using a 366 nm UV light source to give 4, and determining the ratio of the components by GC using a flameless sulfur chemiluminescence detector (Sievers Instruments, Inc., Boulder, Colorado).

For characterization of a large number of beads the tags are introduced into the GC by autosampler thus allowing unattended high throughput. Another type of tag for use herein is indicated in scheme 1 as 5 (which can be used on place of 3) which releases the GC tag by oxidation with Ce(NH_i)2(NO3)6- Other linker strategies which could be used in the presently invented tagging methods are discussed in PCT US93/09345.

In a particularly preferred aspect of the invention a combinatorial library is prepared, each choice therein being encoded by a tag using sulfur identifiers, and tested for biological activity.

An additionally preferred aspect of this invention relates to combinatorial libraries prepared using beads encoded by sulfur identifiers and to pharmaceutically active compounds identified by such combinatorial library.

Without further elaboration, it is believed that one skilled in the art can, using the preceding description, utilize the present invention to its fullest extent.

While the preferred embodiments of the invention are illustrated by the above, it is to be understood that the invention is not limited to the precise instructions herein disclosed and that the right to all modifications coming within the scope of the following claims is reserved.

Claims

What is claimed is:

1. A method of tagging combinatorial libraries by incorporation of a sulfur identifier.

2. The method of claim 1 wherein the sulfur identifier is prepared by varying the ratio of 2 or more sulfur constructs at specified stages during the synthesis of the combinatorial library.

3. A combinatorial library prepared as in claim 1.

4. A method for identifying compounds in a combinatorial library having desired characteristics which comprises: a) preparing a combinatorial library wherein the beads are tagged with sulfur identifiers prepared by varying the ratio of 2 or more sulfur constructs at specified stages during the synthesis of the combinatorial library; b) testing the library in a bioassay which identifies compounds having desired characteristics; c) identifying beads or solubilized derivatives from individual beads which show desired activity in the bioassay; and d) identifying theoretically active ligand by determining the ratio of the sulfur constructs using a highly sulfur selective analytical system to identify and quantify the identifiers relative to each other.

5. Beads encoded with sulfur identifiers.

6. A combinatorial library in which each choice therein is encoded by a unique sulfur identifier.

7. A pharmaceutically active compound identified by a combinatorial library of claim 6.

8. Combinatorial libraries prepared as in claim 1 or 6 where the reaction history of a beads determined at least in part by the ratios of 2 of more sulfur containing ligands.