CA2236339A1 - A fluorescence-based screening method for identifying ligands - Google Patents

Abstract

A novel method for screening chemical compounds for potential pharmaceutical effectiveness is provided. The disclosed method identifies possible therapeutic test ligands by placing them in the presence of target proteins and determining the ability of test ligands to prevent the partial unfolding of the target protein by measuring the fluorescence emission of a conformation-sensitive probe. This differs significantly from known methods of novel pharmaceutical testing in that the biochemical function of the target protein need not be known and the existence of any known ligands of the target protein is unnecessary.

Description

CA 02236339 1998-0~-20 r r 2 PCT~US96/19698 "A FLUORESCENCE-BASED SCREENING METHOD FOR tDENTlFYlNG LIGANDS"

Field of the Invention lS This invention pertains to novel fluorescence-based methods for high-throughput screening for pharmaceutical compounds, in particular those that bind to proteins involved in pathogenesis of disease or in regulation of a physiological function.
Bac~ G~,d of the Invention Pharmaceuticals can be developed from lead compounds that are identified through a random screening process directed towards a target, such as a receptor. ~arge scale screening approaches can be complicated by a number of factors. First, many assays are laborious or expensive to perform. Assays may involve experimental ~nlmAls~ cell lines, or tissue cultures that are dif~icult or expensive to acquire or maintain. They may require the use of radioactive materials, and thus pose safety and disposal problems. These considerations often place practical limitations on the number of compounds that reasonably can be screened. Thus, those employing random screening methods are frequently forced to limit their search to those compounds for which some prior knowledge suggests that 3S the compounds are likely to be effective. This strategy limits the range of compounds tested, and many useful drugs may be overlooked.
Furthermore, the specificity of many biochemical assays may exclude a wide variety of useful chemical compounds, CA 02236339 1998-0~-20 W097t2095Z PCT~S96/19698 because the interac~ions between the ligand and the receptor protein are outside the scope of the assay. For example, many proteins have multiple functions, whereas most assays are capable of monitoring only one such activity. With such a specific assay, many potential pharmaceuticals may not be detected.
Finally, in most existing biochemical screening approaches to drug discovery, the activity of the target protein must be de~ined. This requires that the system in question be well-characterized before screening can begin.
Even when a protein sequence is known, as in, e.g., a newly cloned gene, the specific functions of the protein may not be revealed simply by analysis of its sequence. Consequently, biochemical screening for therapeutic drugs directed against many target proteins must await detailed biochemical characterization, a process that generally requires extensive research.
Thus, there is a need in the art for a rapid, cost-effective, high-throughput assay that enables the screening of large numbers of compounds for their ability to bind therapeutically or physiologically relevant proteins.
Furthermore, there is a need in the art for screening methods that are independent of the biological activity of the target proteins, and that will detect compounds that bind regions o~
the target proteins other than biologically active domains.

Summ~rY of the In~ention The present invention provides a fluorescence-based method for identifying a ligand that binds a target protein.
According to the method, compounds not known to bind the target protein are selected as test ligands. The target protein is incubated with each of the test ligands individually to produce a test combination, and in the absence of a test ligand to produce a control combination. The test and control combinations are contacted with a conformation-sensitive ~luorescence probe, i.e., a probe that binds preferentially to the folded, unfolded, or molten globule state of the protein or whose fluorescence properties are in any way a~ected by the CA 02236339 1998-0~-20 folding status of the target protein. The combinations are treated to cause a detectable fraction of the target protein to exist in a partially or totally unfolded state. Then, the extent to which the target protein occurs in a folded state, an unfolded state, a molten globule state, or combinations thereof, in the test and control combinations is measured by monitoring the fluorescence emission intensity of the probe.
When a dif~erence in fluorescence intensity or other fluorescence property between the test and control combinations indicates that the target protein is present in the folded state to a greater or lesser extent in the test combination than in the control combination, the test ligand is a ligand that binds the target protein. In a preferred em~odiment, the steps of the method are repeated in a high-throughput mode using a plurality of test compounds until a ligand is identified.

Brief DescriPtion of the Drawin~s Figure 1 shows an SDS-polyacrylamide gel profile of carbonic anhydrase after proteolysis in the absence and presence of increasing concentrations of acetazolamide.
Figure 2 shows an SDS-polyacrylamide gel profile of carbonic anhydrase after proteolysis in the absence and presence of 1.0 mM acetazolamide, in the absence and presence of a fungal extract.
Figure 3 shows a graph representing a titration of the binding of radiolabelled human neutrophil elastase to nitrocellulose filters after proteolysis in the absence and presence of increasing concentrations of elastatinal.
Figure 4 shows a graph representing a titration of the ELISA detection of human neutrophil elastase after proteolysis in the presence of increasing concentrations of IC
200,355.
Figure 5 shows a graph representing the distribution of test ligands identified as ligands of human neutrophil - elastase.
Figure 6 shows a graph representing the titration of a ligand for human neutrophil elastase.

CA 02236339 1998-0~-20 t W097~0952 PCT~US96/19698 Figure 7 shows a graph representing the titration of five ligands for their ability to inhibit the enzymatic activity of human neutrophil elastase.
Figure 8 shows a graph representing a titration of the binding of human hemoglobin to nitrocellulose filters after proteolysis in the absence or presence of increasing concentrations of 2,3-diphosphoglycerate.
Figure 9 shows a graph representing a titration of the ELISA detection of human hemoglobin after proteolysis in the presence of increasing concentrations of 2,3-diphosphoglyceraate.
Figure 10 show6 a graph representing the distribution o~ test ligands identified as ligands of human hemoglobin S.
Figure 11 shows a graph representing the titration of a ligand for human hemoglobin.
Figure 12 shows a graph illustrating the effects of increasing concentrations of guanidinium hydrochloride (GCl) on the fluorescence emission of bis-l-anilino-8-naphthalene sulfonate (bis-ANS) measured in the absence and presence of carbonic anhydrase I.
Figure 13 shows a graph illustrating the time-dependent change in fluorescence emission intensity of bis-l-anilino-8-naphthalene sulfonate ~bis-ANS) measured in the presence of carbonic anhydrase I and the indicated concentrations of ~uanidinium hydrochloride (GCl).
Figure 14 shows a graph illustrating the effect of increasing concentrations of a carbonic anhydrase ligand, acetazolamide, on the P~h~ncement of fluorescence emission intensity of bis-l-anilino-8-naphthalene sulfonate measured in the presence of carbonic anhydrase I and 2M guanidinium hydrochloride.

~etailed Dencri~tion o~ the Invention All patent applications, patents, and literature references cited in this specification are hereby incorporated by reference in their entirety. In case of conflict, the present description, including definitions, will prevail.
-CA 02236339 1998-0~ 20 t W O 97/20952 PCT~US96/19698 l~efini ~ion~
A~ used herein, the term "ligand" refers to an agent that binds a target protein. The agent may bind the target protein when the target protein is in its native conformation, or when it is partially or totally unfolded or denatured.
According to the present invention, a ligand is not limited to an agent that binds a recognized functional region of the target protein e.g. the active site of an enzyme, the antigen-combining site of an antibody, the hormone-binding site of a receptor, a cofactor-binding site, and the like. In practicing the present invention, a ligand can al50 be an agent that binds any ~urface or internal sequences or conformational ~o~;nR of the target protein. Therefore, the ligands of the present invention encompa~s agents that in and of themselves may have no apparent biological ~unction, beyond their ability to bind to the target protein in the m~nner described above.
AB used herein, the term "test ligand" re~ers to an agent, comprising a compound, molecule or complex, which is being tested for its ability to bind to a target protein. Test ligands can be virtually any agent, including without limitation metals, peptides, proteins, lipids, polysaccharides, nucleic acids, small organic molecules, and combinations thereof. Complex mixtures of substances such as natural product extracts, which may include more than one test ligand, can also be tested, and the component that binds the target protein can be purified from the mixture in a subsequent step.
Compounds suitable as test ligands may be found in, for example, natural product libraries, fermentation libraries (encompassing plants and microorganisms), combinatorial libraries, compound files, and synthetic compound libraries.
For example, synthetic compound libraries are commercially available from Maybridge Chemical Co. (Trevillet, Cornwall, UK), Comgenex (Princeton, NJ), Brandon Associates (Merrimack, NH), and Microsource ~New Milford, CT~. A rare chemical library is available from Aldrich Chemical Company, Inc.
(Milwaukee, WI). Alternatively, libraries of natural compounds in the form of bacterial, fungal, plant and ~n~m~l extracts are available from, for example, Pan Laboratories (Bothell, WA) or -CA 02236339 1998-0~-20 t W097/20952 PCT~96/19698 MycoSearch (NC), or are readily producible. Additionally, natural and synthetically produced libraries and compounds are readi~y modified through conventional chemical, physical, and biochemical means (Blondelle et al., TibTech 14:60, 1996).
preferably using automated equipment, to allow for the simultaneous screening of a multiplicity of test compounds.

As used herein, the term "target protein~' refers to a peptide, protein or protein complex for which identification of a ligand or binding partner is desired. Target proteins include without limitation peptides or proteins known or believed to be involved in the etiology of a given disease, condition or pathophysiological state, or in the regulation of physiological function. Target proteins may be derived from any living organism, such as a vertebrate, particularly a m~t~t~l and even more particularly a human. For use in the present invention, it is not necessary that the protein' 5 biochemical function be specifically identified. Target proteins include without limitation receptors, enzymes, oncogene products, tumor suppressor gene products, viral proteins, and transcription factors, either in purified form or as part of a complex mixture of proteins and other compounds. Furthermore, target proteins may comprise wild type proteins, or, alternatively, mutant or variant proteins, including those with altered stability, activity, or other variant properties, or hybrid proteins to which foreign amino acid sequences e.g. sequences that facilitate purification have been added.
As used herein, "test combination" refers to the combination of a test ligand and a target protein. "Control combination" refers to the target protein in the absence of a test ligand.
As used herein, the "folded state" of a protein refers to the native or lln~n~tured form of the protein as it is present in its natural environment, or after isolation or purification, i.e. before exposure to denaturing conditions.
Similarly, the "unfolded state" refers to a situation in which the polypeptide has lost elements of its secondary and/or CA 02236339 1998-0~-20 W O 97/20952 PCT~US96/19698 tertiary structure that are present in its "folded state." It will be recognized by those skilled in the art that it is difficult to determine experimentally when a polypeptide has become completely unfolded i.e. has lost all elements of S secondary and tertiary structure. Thus, the term ~unfolded state" as used herein encompasses partial or total unfolding.
As used herein, "detectable fraction" refers to a ~uantity that is empirically determined and that will vary depending upon the method used to distinguish folded from unfolded protein. For example, when fluorescent probes are used to monitor the degree of target protein folding, conditions are chosen so that changes in fluorescence are of a readily detecta~le magnitude. When protease sensitivity is used to monitor folding, conditions are chosen (e.g. by adjusting temperature or adding denaturants) so that approximately 80 of the target protein is digested within a convenient incubation period. When antibodies specific to the folded or unfolded state of a target protein are used as the detection method, conditions are chosen so that a sufficient amount of anti~ody is bound to give a detectable signal.
The present invention encompasses high-throughput screening methods for identifying a ligand that binds a target protein. If the target protein to which the test ligand binds is associated with or causative of a disease or condition, the ligand may be useful for diagnosing, preventing or treating the diseaRe or condition. A ligand identified by the present method can also be one that is used in a purification or separation method, such as a method that results in purification or separation of the target protein from a mixture. The present invention also relates to ligands identified ~y the present method and their therapeutic u~es (for diagnostic, preventive or treatment purposes) and uses in purification and separation methods.
According to the present invention, a ligand for a target protein is identified by its ability to influence the extent of folding or the rate of folding or unfolding of the target protein. 3xperimental conditions are chosen so that the target protein is su~jected to unfolding, whether reversible CA 02236339 1998-0~-20 or irreversible. If the test ligand binds to the target protein under these conditions, the relative amount o~
folded:unfolded target protein or the rate of folding or unfolding of the target protein in the presence of the test ligand will be dif~erent, i.e. higher or lower, than that observed in the ab6ence of the test ligand. Thus, the present method encompasses incubating the target protein in the presence and absence of a test ligand, under conditions in which (in the absence of ligand) the target protein would partially or totally unfold. This is followed by analysis of the absolute or relative amounts of folded vs. unfolded target protein or of the rate of folding or unfolding of the target protein.
An important feature of the present invention is that it will detect any compound that binds to any sequence or domain of the target protein, not only to sequences or ~Aln~
that are intimately involved in a biological activity or function. The binding sequence, region, or domain may be present on the surface of the target protein when it is in its folded state, or may be buried in the interior of the protein.
Some binding sites may only become accessible to ligand binding when the protein is partially or totally unfolded.
In practicing the present invention, the test ligand is combined with a target protein, and the mixture i8 maintained under appropriate conditions and for a sufficient time to allow binding of the test ligand to the target protein.
Experimental conditions are determined empirically for each target protein. When testing multiple test ligands, incubation conditions are chosen so that most ligand:target protein interactions would be expected to proceed to completion. In general, the test ligand is present in molar excess relative to the target protein. The target protein can be in a soluble ~orm, or, alternatively, can be bound to a solid phase matrix.
The matrix may comprise without limitation beads, membrane 3S filters, plastic surfaces, or other suitable solid supports.
For each target protein, appropriate experimental conditions, e.g. temperature, time, pH, salt concentration, and additional components, are chosen 80 that a detectible fraction CA 02236339 1998-0~-20 WO 97/2~952 PCT~US~6/19698 of the protein is present in an unfolded form in the absence o~ test ligand. For a target protein that unfolds irreversibly, preferred experimental conditions allow a detectable amount of the protein to unfold during a convenient incubation period in the absence of test ligand. To adjust or optimize the ratio of folded:unfolded protein or the rate of fo~ding or unfolding, denaturing conditions may be required, including the use of elevated temperatures, the addition of chaotropes or denaturants such as urea or guanldinium or guanidium salts such as guanidinium thiocyanate, detergents, or combinations thereo~. Furthermore, mutant proteins that contain stabilizing or destabilizing amino acid substitutions relative to the wild-type version of the protein may be used to manipulate the folded:un~olded ratio.
The time necessary for binding of target protein to ligand will vary depending on the test ligand, target protein and other conditions used. In some cases, binding will occur instantaneously (e.g., essentially simultaneous with combination of test ligand and target protein), while in others, the test ligand-target protein combination is maintained for a longer time e.g. up to 12-16 hours, before binding is detected. When many test ligands are employed, an incu~ation time is chosen that is sufficient for most protein:ligand interactions. In the case of target proteins that unfold irreversibly, the rate of unfolding must also be taken into consideration in determining an appropriate time for b; n~; ng of test ligand.
B; n~; ng of a test ligand to the target protein is assessed by comparing the absolute amount of folded or unfolded target protein in the absence and presence of test ligand, or, alternatively, by determining the ratio of folded:un~olded target protein or the rate of target protein folding or unfolding in the absence and presence of test ligand. If a test ligand binds the target protein (i.e., if the test ligand is a ligand for the target protein), there may be significantly more folded, and less unfolded, target protein (and, thus, a higher ratio of ~olded to un~olded target protein) ~han is present in the absence of a test ligand. Alternatively, CA 02236339 1998-0~-20 binding of the test ligand may result in significantly less folded, and more unfolded, target protein than is present in the absence of a test ligand. Similarly, binding of the test ligand may cause the rate of target protein folding or unfolding to change significantly.
In either case, determination of the absolute amounts of folded and unfolded target protein, the folded:unfolded ratio, or the rates of folding or unfolding, may be carried out using one of the known methods as described below. These methods include without limitation measuring the fluorescence emission of conformation-speci~ic probes, proteolysis of the target protein, binding of the target protein to appropriate surfaces, binding of specific antibodies to the target protein, binding of the target protein to molecular chaperones, binding of the target protein to immobilized ligands, and measurement of aggregation o~ the target protein. Other physico-chemical techniques may also be u~ed, either alone or in conjunction with the above methods; these include without limitation measurements of circular dichroism, intrinsic ultraviolet and fluorescence spectroscopy, and calorimetry. A preferred embodiment involves comparing the fluore~cence emission of a conformation-specific probe in the presence of a target protein following incubation in the absence and presence of a test ligand. Typically, this involves contacting the test and control combinations with the fluorescence probe prior to the treatment (such as, e.g., elevated temperature) that is employed to manipulate the degree of folding. ~owever, it will be recognized by those skilled in the art that each target protein may have unique properties that make a particular detection method most suitable for the purposes of the present invention.
For the purposes of high-throughput screening, the experimental conditions described above are adjusted to achieve a threshold proportion of test ligands identified as "positive"
compounds or ligands from among the total compounds screened.
This threshold is set according to two criteria. First, the number of positive compounds should be manageable in practical terms. Second, the number of positive compounds should re~lect CA 02236339 1998-0~-20 W O 97/20952 PCT~US96/19698 ligands with an appreciable a~finity towards the target protein. A preferred threshold is achieved when 0.1~ to 1~
of the total test ligands are shown to be ligands o~ a given target protein.
5Binding to a given protein is a prerequisite for pharmaceuticals intended to modify directly the action of that protein. Thus, if a test ligand is shown, through use of the present method, to bind a protein that reflects or affects the etiology of a condition, it may indicate the potential ability of the test ligand to alter protein function and to ~e an effective pharmaceutical or lead compound for the development of such a pharmaceutical. Alternatively, the ligand may serve as the basis for the construction of hybrid compounds containing an additional component that has the potential to alter the protein's function. In this case, b; n~; ng of the ligand to the target protein serves to anchor or orient the additional component so as to effectuate its pharmaceutical effects. For example, a known compound that inhibits the activity of a family of related enzymes may be rendered specific to one member of the family by conjugation of the known compound to a ligand, identi~ied by the methods of the present invention, that binds specifically to that member at a different site than that recognized by the known compound.
The fact that the present method is based on physico-chemical properties common to most proteins gives it widespread application. The present invention can be applied to large-scale systematic high-throughput procedures that allow a cost-effective screening of many thousands of test ligands. Once a ligand has been identified by the methods of the present invention, it can be further analyzed in more detail using known methods specific to the particular target protein used.
For example, the ligand can be tested for b;n~;ng to the target protein directly e.g. by incubating radiolabelled ligand with unlabelled target protein, and then separating protein-bound and unbound ligand. Furthermore, the ligand can be tested for its ability to influence, either positively or negatively, a known biological activity of the target protein.

CA 02236339 1998-0~-20 W O 97/20952 PCT~US96/19698 In a preferred embodiment of the present invention, binding of test ligand to target protein iB detected through the use of molecular probes whose fluorescence emission characteristics are sensitive to target protein con~ormation.
Certain fluorescent compounds exhibit only a weak ~luorescence emission when free in aqueous solution (Semisotnov, et al., Biopolymers 31~ , 1991), but fluoresce much more strongly when bound to organized hydrophobic surfaces. Binding of these compounds to fully folded globular proteins is typically weak, since hydrophobic residues are predominantly buried in the interior of the protein.
Furthermore, binding of these compounds to random coil conformations (as found in fully unfolded or denatured polypeptides) is also disfavored, because in these conformations hydrophobic residues, though exposed, are not sufficiently well organized to support high affinity binding of the probes. The probes, however, typically bind with higher affinity and stoichiometry to compact unfolded protein conformations, often referred to as "molten globules", which are characterized by compactness relative to random coil unfolded states, the presence of substantial ~econdary structure, and the lack of a unique overall conformation.
Molten globule states were initially identified as compact unfolded states that occur stably for a variety o~ proteins under specialized conditions such as low pH, moderate denaturant concentrations, and heat ~Ptitsyn et al., Mol.Biol.
17:569, 1983). The molten globule has ~een identifled as a common intermediate state in the process of protein folding (Ptitsyn et al., FEBS Letts. 262:20, 19903. Without wishing to be bound by theory, it is believed that molten globule states display sufficiently organized hydrophobic surfaces to support binding of useful fluorescent probes with affinities in the micromolar range.
Fluorescent molecules useful in practicing the present invention (re~erred to hereina~ter as "probes") include without limitation 1-anilino-8-naphtha~ene sulfonate (ANS), bis-1-anilino-8-naphthalene sulfonate (bis-ANS) and 6--CA 02236339 1998-0~-20 W097/20952 PCT~S96/19698 propionyl-2-(N,N-dimethyl)-~m;no~hthalene (Prodan) (Molecular Probes, Eugene, OR).
It will be understood that any fluorescent compound may be used that has substantially altered ~luorescent properties upon binding to protein and that preferentially binds to molten globule or other unfolded forms of the particular target protein. The only limitations are that the relative binding affinities and stoichiometry of binding of the pro~e must be of a magnitude to ensure that the fluorescence change observed upon conformational change between the native and molten globule states can be readily detected.
Preferably, the concentration of the probe used in practicing the invention is low enough to avoid substantially destabilizing the folded form. For example, native DnaK
protein binds bis-ANS with a stoichiometry of 1:1 and a dissociation constant of 7.0 ~M; while DnaK in the molten globule state binds bis-ANS with a stoichiometry of 3:1 and dissociation constants of 2.0, 5.1, and 39 ~M (Shi et al., Biochemistry 33 :7536, 1994) . Exposure of 1 ~ DnaK protein to ~M bis-ANS results in conformational changes in DnaK
consi~tent with conversion from a native to a molten globule conformation. Lower concentrations of bis-ANS, however, can be used as an indicator of conformational change according to the present invention. For example, addition of 1 ~M DnaK to 1 ~M bis-ANS enhances the bis-ANS fluorescence nearly 30-fold as the result of binding to the molten globule state of DnaK.
Addition of a ligand, ATP, to the mixture reduces the fluorescence enhancement about five-fold.
In this embodiment, a target protein is contacted with a probe in the presence or absence of a test compound (i.e., in both test and control combinations), under conditions in which at least a portion of the target protein i8 partially unfolded. After an appropriate period o~ incubation, usually one to ten minutes, the test and control combinations are irradiated with light of an appropriate wavelength to excite the probe, and the fluorescence emission of the probe is measured at a wavelength appropriate for the particular probe.

CA 02236339 l998-0~-20 WO 97/20952 PCT~US96/19698 If the test compound is a ligand of the target protein, the amount o~ protein in the molten globule (or folding intermediate) state should be reduced by the presence of the test compound, which should be reflected in a decrease in binding of the probe and a corresponding reduction in the intensity of the fluorescence emission.
In another embodiment, binding of test ligand to target protein is detected through the use of proteolysis.
This assay is based on the increased susceptibility of unfolded, denatured polypeptides to protease digestion relative to that of folded proteins. In this case, the test ligand-target protein combination, and a con~rol combination lacking the test ligand, are treated with one or more proteases that act preferentially upon unfolded target protein. After an appropriate period of incubation, the level o~ intact i.e.
unproteolysed target protein is assessed using one of the methods described below, e.g., ~el electrophoresis and/or ; mmllnoassay .
There are two possible outcomes that indicate that the test ligand has bound the target protein. Either a significantly higher, or significantly lower, absolute amount of intact or degraded protein may be observed in the presence of ligand than in its absence.
Proteases useful in practicing the present invention include without limitation trypsin, chymotrypsin, V8 protease, elastase, carboxypeptidase, proteinase K, thermolysin and subtilisin (all of which can be obtained from Sigma Chemical Co., St. Louis, M0). The most important criterion in selecting a protease or proteases for use in practicing the present invention is that the protease(s) must be capable of digesting the particular target protein under the chosen incu~ation conditions, and that this activity be pre~erentially directed towards the unfolded form of the protein. To avoid "false positive" results caused by test ligands that directly inhibit the protease, more than one protease, particularly proteases with different enzymatic mechanisms of action, can be used simultaneously or in parallel assays. In addition, cofactors that are required for the activity of the protease~s) are CA 02236339 1998-0~-20 WO 97/20952 PCT~US96/19698 provided in excess, to avoid false positive results due to test ligands that may sequester these factors.
Typically, a purified target protein is first taken up to a final concentration of 2-100 ~g/ml in a buffer containing 50mM Tris-HCl, pH 7.5, 10 mM calcium acetate and ~ 0.034 mg/ml bovine serum albumin. Proteinase K and thermolysin, proteases with distinct mechanisms of action, are then added to a final concentration o~ 2-10 ~g/ml. Parallel incubations are then per~ormed ~or different time periods ranging from 5 minutes to one hour, at temperatures ranging from 20~C to 65~C. Reactions are terminated by addition of phenylmethylsulfonyl chloride (PMSF) to a final concentration of 1 mM and ethylenediaminotetraacetic acid (EDTA) to a final concentration of 20 mM. The amount of intact protein re~ ~n;ng in the reaction mixture at the end o~ the incubation period is then assessed by any of the following methods: polyacrylamide gel electrophoresis, ELISA, or binding to nitrocellulose filters. The above protocol allows the selection of appropriate conditions that result in digestion of approximately 80~ of the target protein, indicating that a significant degree of unfolding has occurred. If a known ligand for the target protein is available, the ligand is included in the reaction mixture at a concentration o~ 20-200~M, and the experiment is repeated. Typically, at least a two-fold increase or decrease in the level of intact target protein is observed, indicating that binding of a known ligand changes the ratio of folded:unfolded target protein and/or the rate of folding or unfolding.
Once conditions are established for high-throughput screening as described above, the protocol is repeated simultaneously with a large number o~ test ligands at concentrations ranging from 20 to 200 ~M. Observation of at least a two-~old increase or decrease in the level of intact protein signifies a "hit" compound i.e. a ligand that binds the target protein. Preferred conditions are those in which between 0.1 and 1~ of test ligands are identified as "hit"
compounds using this procedure.

CA 02236339 1998-0~-20 WO 97/209~2 PCTAJS96/19698 In another embodiment, the relative amount of folded and unfolded target protein in the presence and absence of test ligand is assessed by measuring the relati~e amount of target protein that binds to an appropriate surface. This method takes advantage of the increased propensity of unfolded proteins to adhere to surfaces, which i9 due to the increased surface area, and decrease in m~:5k~ng of hydrophobic residues, that results from unfolding. If a test ligand binds a target protein (i.e., is a ligand of the target protein), it may stabilize the folded form of the target protein and decrease its binding to a solid surface. Alternatively, a ligand may stabilize the unfolded form of the protein and increase its binding to a solid surface.
In this embodiment, the target protein, a test ligand and a surface that preferentially binds unfolded protein are combined and maintained under conditions appropriate for binding of the target protein to a ligand and binding of unfolded target protein to the surface. Alternatively, the target protein and test ligand can be pre-incubated in the absence of the surface to allow binding. Surfaces suitable for this purpose include without limitation microliter plates constructed from a variety of treated or untreated plastics, plates treated for ti6sue culture or for high protein binding, nitrocellulose filters and PVDF filters.
Determination of the amount o~ surface-bound target protein or the amount of target protein r~ n;ng in solution can be carried out using standard methods known in the art e.g.
determination of radioactivity or i~mllnoassay~ If significantly more or less target protein is surface bound in the presence of a test ligand than in the absence of the test ligand, the test ligand is a ligand of the target protein. Similarly, the ratio of surface-bound:soluble target protein will be significantly greater or smaller in the presence of a test ligand than in its absence, if a test ligand is a ligand for the target protein.
In another embodiment, the extent to which folded and unfolded target protein are present in the test combination is assessed through the use of antibodies specific for either the -CA 02236339 1998-0~-20 W097/20952 PCT~S96/19698 un~olded state or the folded state of the protein i . e.
denatured-specific ("DS"), or nature-specific ("NS") antibodies, respectively. (Breyer, 1989, ~. Biol. Chem., 264 (5):13348-13354).
Polyclonal and monoclonal DS and NS antibodies specific for particular target proteins can be prepared by methods that are well known in the art (E. Harlow & D. Lane, ANTIBODIES: A LABORATORY MANUA~, Cold Spring Harbor Laboratory, 1988; Zola, Monoclonal ~ntibodies: A M~n~ of Techniques, CRC
Press, Inc., Boca Raton, Florida,1987). For DS antibodies, ~n;~l S can be imml]n;zed with a peptide from a region of the protein that i9 buried in the interior of the protein when it i8 in the native state. If the three-~imensional structure of the protein is unknown, antibodies are prepared against several peptides and then screened for pre~erential binding to the denatured state. For NS antibodies, intact non-denatured protein is used as an immunogen, and the resulting antibodies are screened for preferential bindiny to the native protein and purified for use in the present invention.
DS or NS antibodies can be utilized to detect a ligand-induced change in the level of folded target protein, unfolded target protein, the folded:unfolded ratio, or the rate of folding or unfolding.
In one approach, a test combination containing the DS antibody, the target protein, and the test ligand is exposed to a solid support e.g. a microliter plate coated with the denatured target protein or a peptide fragment thereof, under conditions appropriate for b; n~; ng of the target protein with its ligand and binding of the DS antibody to unfolded target protein. A control combination, which i8 the same as the test combination except that it does not contain test ligand, is processed in the same manner as the test solution. By comparing the amount of antibody bound to the plate or the amount rem~;n;ng in solution in the test and control combinations, the difference in target protein folding is detected. The amount of antibody bound to the plate or rem~; n; ng in solution can be measured as described below.

CA 02236339 1998-0~-20 WO 97nog52 PCT~US96/19698 In a second approach, a test combination cont~in;ng the DS antibody, the test ligand, and the target protein is exposed to a solid support coated with a second antibody, re-ferred to as a solid phase antibody, which cannot bind to the target protein simultaneously with the DS antibody, and is ~pecific for the target protein, but is either specific for ~he folded state (NS antibody) or unable to differentiate between the native and denatured states ("non-differentiating" or "ND"
antibody). The resulting test combination or solution is maintained under conditions appropriate for b; n~i ng of the target protein with a ligand of the target protein and for binding of the antibodies to the proteins they recognize. A
control combination, which is the same as the te6t solution except that it does not contain test ligand, is processed in the same manner as the test solution. In both combinations, denatured (unfolded) target protein binds the DS antibody and is inhibited from binding the solid phase antibody. The ability of the test ligand to bind the target protein can be gauged by determining the amount of target protein that binds to the solid phase antibody in the test solution and comparing it with the extent to which target protein binds to the solid phase antibody in the absence of test ligand, which in turn reflects the amount of target protein in the folded state. The amount of target protein bound to the plate via the second antibody or rem~in;ng in solution can be detected by the methods described below. This approach may be used in a comparable manner with NS antibody as the soluble antibody and DS or ND antibody on ~he solid phase.
In a third approach, a test solution cont~;ning the target protein and the test ligand is exposed to a solid support e.g. a microliter plate that has been coated with a DS
or NS antibody and maintained under conditions appropriate for binding of target protein to its ligand and for binding of the antibody to target protein. Alternatively, the antibody can be present on the surfaces of beads. The ability of the test ligand to bind the target protein i~ gauged by determining the extent to which target protein r~; n~ in solution (unbound to the antibody) or on the solid surface (bound to the antibody), CA 02236339 1998-0~-20 W 097/20952 PCTnUS96/19698 or the ratio of the two, in the presence and in the absence of test ligand. Alternatively, the antibody can be present in ~olution and the ~arget protein can be attached to a solid phase, such as a plate surface or bead surface.
In another embodiment, molecular chaperones are u~ed to assess the relative levels of folded and unfolded protein in a test combination. Chaperones encompass known proteins that bind unfolded proteins as part o~ their normal physiological function. They are generally involved in assembling oligomeric proteins, in ensuring that certain proteins fold correctly, in facilitating protein localization, and in preventing the ~ormation of proteinaceous aggregates during physiological stress (Hardy, 1991, Science, 251:439-443). These proteins have the ability to interact with ~any unfolded or partially denatured proteins without specific recognition of defined sequence motifs.
One molecular chaperone, found in E. col i, is a protein known as SecB. SecB has a demonstrated involvement in export of a subset of otherwise unrelated proteins.
Competition experiments have shown that SecB binds tightly to all the unfolded proteins tested, including proteins outside of its particular export subset, but does not appear to interact with the folded protein. Other chaperones suitable for use in the present invention include without limitation heat shock protein 70s, heat shock protein 90s, GroEI and GroES
(Gething et al., Nature 355:33, 1992).
In this embodiment, a test com~ination containing the test ligand and the target is exposed to a solid support e.g.
microliter plate or other suitable surface coated with a molecular chaperone, under conditions appropriate for b;n~;ng of target protein with its ligand and binding of the molecular chaperone to unfolded target protein. The unfolded target protein in the solution will have a greater tendency to bind to the molecular chaperone-covered surface relative to the ligand-stabilized folded target protein. Thus, the ability of - the test ligand to bind target protein can be determined by determining the amount of target protein r~m~;n;ng unbound, or the amount bound to the chaperone-coated surface.

CA 02236339 1998-0~-20 W097/20952 PCT~S96/19698 Alternatively, a competition assay for binding to molecular chaperones can be utilized. A test combination containing purified target protein, the test ligand, and a molecular chaperone can be exposed to a solid support e.g. a microliter well coated with denatured (unfolded) target protein, under conditions appropriate for binding target protein with its ligand and binding of the molecular chaperone to unfolded target protein. A control combination, which is the same as the test combination except that it does not contain test ligand, is processed in the same m~nnpr.
Denatured target protein in solution will bind to the chaperone and thus inhibit its binding to the denatured target protein bound to the support. Binding of a test ligand to the target protein will result in a difference in the amount of unfolded target protein, and, thus, more or less chaperone will be available to bind to the solid-phase denatured target protein than is the case in the absence of binding of test ligand.
Thus, binding of test ligand can be determined by assessing chaperone bound to the surface or in solution in the test combination and in the control combination and comparing the results. In this assay, the chaperones are generally not provided in excess, so that competition for their binding can be measured.
Alternatively, a test combination containing the target protein, the test ligand and a molecular chaperone can be exposed to a solid support e.g. a microliter well that has been coated with antisera or a monoclonal antibody specific for the folded target protein (NS antibody) and unable to bind the target protein bound to the chaperone. Unfolded target protein will bind chaperone in solution and thus be inhibited from binding the solid phase antibody. By detecting target protein in the solution or bound to the well walls and comparing the extent of either or both in an appropriate control (the same combination without the test ligand), the ability of the test ligand to bind target protein can be determined. If the test - ligand is a ligand for the target protein, more or less target protein will be bound to the antisera or monoclonal antibody bound to the container surface in the test combination than in CA 02236339 1998-0~-20 W O 97~09S2 PCT/US96/19698 the control combination, and correspondingly more or les~
target protein will be present unbound (in solution~ in the test combination than in the control combination.
In another embodiment, a known ligand, cofactor, ~ubstrate, or analogue thereof of the target protein is used to assay for the presence of folded target protein. The higher the fraction of protein in the folded form, the greater the amount of protein that i5 available to bind to a ligand that binds exclusively to the folded state. Consequently, if a protein has a known ligand, it is po~sible to increase or decrease the binding of the protein to the known ligand by adding a test ligand that binds another site on the protein.
For example, binding of dihydrofolate reductase to methotrexate, a folic acid analogue, can be used to assess the level of folding of this enzyme.
In this approach, the ligand, cofactor, substrate, or analogue thereof known to bind to the target protein is immobilized on a solid substrate. A solution cont~; n ing the target protein and test ligand is then added. An increase or decrease in the amount of target protein that binds to the immobilized compound relative to an identical assay in the absence of test ligand indicates that the test ligand binds the target protein. The amount of target protein bound to the solid substrate can be assessed by sampling the solid substrate or by sampling the solution.
In another embodiment, the amount of unfolded target protein in a test combination is assessed by measuring protein aggregation. For proteins that unfold irreversibly, unfolded protein often forms insoluble aggregates. The extent of protein aggregation can be measured by techniques known in the art, including without limitation light scattering, centrifugation, and filtration.
In this approach, target protein and test ligand are incubated and the amount of protein aggregation is mea~ured over time or after a fixed incubation time. The extent of ~ protein aggregation in the test mixture is compared to the same measurement for a control assay in the absence of test ligand.
If a test ligand binds a target protein, the rate of unfolding CA 02236339 1998-0~-20 W097/209~2 PCT~S96/19698 of target protein will be lower or higher than in the absence of test ligand. For measurements over time, the rate of appearance of aggregated protein will be lower or higher if the test ligand i8 a ligand for the target protein than if it is not. For measurements at a fixed time, there will be more or less unfolded protein and correspondingly less or more aggregated protein if the test ligand is a ligand for the target protein than if it is not. Thus, the ability of a test ligand to bind a target protein can be determined by assessing the extent of protein aggregation in the presence and absence of test ligand.
The embodiments described above are summarized in the following table.

W097/20952 PCT~S96/19698 TABLE
DETER~l~l~ FO~DED AND UNFOhDED TARGET PROTEIN
- . ... .~ ~ od ~ ~ ¦ s t Obser ed rf Test L gand.B'nds T get Protein::
rluu.es~,"..ce C- ' specific riuolu~,.ll probes are used ¦ More or less fluorescence emission in test ~,oll.l,il~"Lio I than in eontrol C~ ;01 Proteolvsis Protease that ~ .f.,.~.-lially hydrolyzes unfolded target ¦ More or less intact target protein in test c~ ti~ -- than protein is used I in control culllb;.~ali Surfaee Bindin~
Surface that ~ ,f~.. ,nli~.lly binds unfolded target protein is ¦ More or less target protein unbound (in solution) to used . ¦ surface in test cu.. l,i.. ~liv-- than in control ~;u~l~l)hlaLio Antibodv Bindin~
DS antibody in solution/unfolded target protein or peptide More or less DS antibody bound to unfolded target fragment thereof on surfaee protein or peptide fragment thereof on surface in test c~ L -~n than in controi cvll~b;.lalivll DS antibody in solution/ antibody that .eco "i,~ folded More or less target protein bound to antibody on surface target protein on surfaee in test C< IIIb ' than in control L ' ~ ,n NS antibody in solution/ antibody that l~.O"lli~ folded More or less target protein bound to antibody on surface target protein on surface in test collllJ .Iàlic.l~ than in control uullllJillaLiull DS antibody on surface More or less target protein bound to DS antibody on surface in test co,.,l.;..-li.,., than in control co.llLilla~iu 2 0 NS antibody on surfaee More or less target protein bound to NS antibody on surface in test cv",l, than in control cv~ inaLion Molecular CL~ .u.._~
Chaperone on surface More or less target protein bound to cha~ vlle on surface in test co---l,~ liu.. than in control cv---v C"" ,~ Assay Unfolded target protein on solid phase, target protein in More or less ~ .vne bound to unfolded target prooein 2 5 solution on solid phase in test cu,nl, la~ivn than in control .,u,.,~, ' Chaperone in solution/antibody that ~~co~ folded More or less target protein bound to surface-bound target protein on surface antibody in test c.,.-,~ ~- than in control cv",l~
Differential l~indin~ to L~ v~ d Li~and Target protein in solution, known ligand of target protein More or less target protein bound to surface bound ligand 3 0 attached to surface in test combination than in control co. .
Protein A~ ,k..Liùl-rv....alion of ..~ t~,d protein by i..~vc.~il,le prooein More or less agg.,,gat~d protein ~igher or lower rate of unfolding formation of aggregated proteiD) in test cv---b than in control 3 5 Protein De~ect;on Methods The embodiments described above require a final step for detecting and/or quanLiryillg the level of target protein or digestion products thereof, or antibodies, in order SUBSTITUTE SHEET (RULE 26) to ~luallLiry the relative amounts of folded and unfolded target protein after exposure to test ligands. In practicing the present invention, methods known in the art are used to detect the presence or absence of protein, small peptides or free amino acids. The method used will be cle~ P,d by the product (prote..ls, peptides, free amino acids) to be ~letP-ctf~l For 5example, terhniql~os for detectin~ protein size can be used to ~ i..P the extent of proteolytic degradation of the target protein e.g. gel electrophoresis, capillary electrophoresis, size exclusion chromatography, high-lx,.ro~ allce liquid chromatography, and the like. Mea~ l of r~lio~ctivity, fluol~scence, or el~ylllaLic activity can detect the ~l-,sellce or absence of products, either in solution or on a solid support. Tmmllnological 10methods inrln~lin~ e.g. ELISA and radioi-------~ y can detect ~he ~lesellce or absence of a known target protein in solution or on a substrate. The above mPth~ are described in e.g.
Harlow, E. and D. Lane, Antibodies: A Laboratory Manual, Cold Spring Harbor Labc,l~.lolies, 1988; S.F.Y. Li, Capillary Electrophoresis, Elsevier Press, 1993;
Bi~llingrn~yer, Practical HPLC Methodology and Applications, John Wiley and Sons, Inc., 151992; and Cantor, C.R. and P.R. ~ himmPl, Biophysical Chemistry, WH Freeman and Co., 1980.
In a prer~,~ed emboflimpnt~ gel electrophoresis is used to detect the pl~e.lce or absence of protein, and can further be used to detect the size of the protein. This latter method is especially useful in co~ o~ with proteolysis, as the l)lesel~ce of a greater or 201esser amount of undigested target protein in the test combination than in the control con~ alion in~lir~t~s that the test ligand bound to the target proteirl.
The following examples are intPn(l~l to illustrate the invention without limiting it thereof.

25~,Y~mp'~ 1: Methol,t:A~I~ Rin~lin~ I~ole~l~ D;}lylllofolate 112e~ ct~e (DHE~) From Proteolytic D;~e~Xoi. by Prot~in~e K
The following were combined and inf~nb~ted at 54~C for 5 ~ rs: DHFR
(100 ,ug/ml)? P.otei~se K (80 ~g/ml), 0.1 M Tris-HCl pH 7.5, and Methotrexate at 10-~~ to Samples were removed and undigested DHFR was ~ riPd by ELISA as follows:

W 097/20952 PCT~US96/19698 (a) Protease inrllb~tion~ were diluted 50-fold wi~ Tris-buffered saline (TBS);
(b) 50 ,ul diluted samples were I-A.~r~. ~ed to the wells of an ELISA plate and infu~ated 60 minntes at room temperature;
(c) the plate wells were thoroughly washed with TBS plus 0.1% Tween-20 (TBST);
(d) 50 ,ul anti-DHFR rabbit serum diluted 250-fold into TBST plus 5 % nonfat dry miLk was added to each well and inrllh~ted 30 mimltPs at room L~ peldLulc;;
(e) plate wells were washed as in (c) above;
(f~ 50 ~1 of goat anti-ra~bit IgG ~1k~1in~ phosphatase conjugate diluted 500-fold in TBST plus 5% milk was added to each well and inr~lbated 30 .,.i",.~es at room (g) plate wells were washed as in (c); and (h) 0.1 ml of 1.0 mg/mlp-nitrophenylphosphate in 0.1~ dieth~n~-lamine was 5added. Color development is plopc,l~ional to ~lk~linP phosphatase antibody conjugate bound.
The ELISA analysis showed that ~LIwLlc;~Le l)loLe~;~ DHFR from digestion at concentrations of 10-8M and higher. By the same mrth~ds nicotin~mi-le ~riPninP
imlrl~Potide phosphate (NADPH) and dihy~ rolate at col~cellL,aLions of 10-5M and higher were shown to inhibit proteolysis of DHFR in sepala~ "ents.

FYq n~ 2: Methol~ ~A~le, NADPH and Dihydrofolate R:mlin~Dihydro~olate R~lln~t~ce (DHF~) From Proteolytic Digestion by I'~ e K in the I~es~,.ce of a Mixture of Amino Acids The following were combined and inr~lbated at 54' C for 5 .. ,;.,.. ~s: DHFR
(2.1 ~g/ml), Pl~L~illdse K (80 ,~g/ml), O.lM Tris-HCl (pH 7.5), 10-5M of all 20 common amino acids and either 0 or 10-5M ligand. The ligands used were the inhibitor Methol,c~aL~;
and the ~ dihydrofolate and NADPH.
Samples were removed and undigested DHFR was quantified by ELISA as 3 of ollows:
(a) Protease inrllhatinns were diluted 50 fold with Tris-buffered saline (TBS);
(b) 50 ul diluted samples were L~ re~led to the wells of an ELISA plate and inr~lb~tPd 60 mimltes at room Lel~ dlUlCi;

(c) the plate wells were thoroughly washed with TBS plus 0.1% Tween-20 (TBST);
(d) 50 ,ul anti-DHFR rabbit serum diluted 250 fold into TBST plus 5 % nonfat dry miL~c was added to each well and inrnh~tP-1 30 ..l~ s at room ~ )e1dLU11~; (e) plate wells were washed as in (c) above;
(f) 50 pl of goat anti-rabbit IgG ~lk~lin~ phosphatase conjugate diluted 500 fold in TBST plus 5% miLk was added to each well and inr~lb~te~l 30 mimltPs at room temperature;
(g) plate wells were washed as in (c); and (h) 0.1 ml of 1.0 mg/mlp-nitrophenylphosphate in 0.1% rli~th~nolamine was added. Color development is pLopolLional to ~lk~line phosphatase antibody conjugate bound.
The EIISA analysis showed that meth~ a~ and the ~ub~L~Ies protect DHFR
from digestion relative to the absence of ligands that bind to DHFR. Thus, specific binding can be ~letecte~l in the presence of a complex ~ uie of compounds that do not bind to the 5target protein.

F,YD-nr1~ 3 Meth~r~ R:n~ling ~nh;h;t~ Binding o~ DHE~ to Microliter Plates The following were combined in a volume of 60 ~l and inr~b~tPd in a Falcon 2 o3072 "tissue-culture treated" microliter plate at 20 or 47~C: lO0 mg DHFR, 50 MM Tris-Cl (pH 7.5), and MethoLIG~aLe 10-l~ to lO~M.
50 ~1 of each sample was then Ll~l~r~ ,d to the wells of an ELISA plate, and the DHFR that le.~ )ts(l in solution was q~l~ntifiP-l by ELISA as follows:
(a) The 50 ,ul samples were inr~ te~1 ~or 60 ...i..~es at room tem~el~Lure;
2 5 (b) the plate wells were thoroughly washed with TBS plus 0. l % Tween-20 (TBST);
(c) 50 ,ul anti-DHFR rabbit serum diluted 250-fold into TBST plus 5% nonfat dry miLk was added to each well and inr~lb~te~l 30 ~ s at room L~ Lu (d) plate wells were washed as in (c) above;
3 o (e) 50 ~l of goat anti-rabbit IgG ~lk~lin~ phosphatase conjugate diluted 500-fold in TBST plus 5% miLk was added to each well and inr~lb~tr~l 30 ..~ s at room temperature;
(f) plate wells were washed as in (b); and W O 97/20952 PCT~US96/19698 (g) 0.1 ml of 1.0 mg/ml D-nitrophenylphosphate in 0.1% diethanolamine was added. Color development is proportional to ~lk~line phosph~t~e antibody conjugate bound.
The EIlSA analysis revealed that m~;LlloL~ te inhibits DHFR binding to the Falcon 3072 plate at con~e~ it)ns of 1~7M and above.

~mpl~ 4: Inhibition or ~nh~-.r~ ' of Unfolded-Specific Antibody Binding (1) ELISA plates are coated by inrllb~tion for 60 ~.-il---l~s with the followingmixture: 4 ,ug/ml hl~v~ibly del aLuled target protein or peptide fr~gmPnt~ thereof in Tris-buffered Saline (lO mM Tris-Cl, pH 7.5, 0.2M NaCl; TBS).
(2) The plates are washed 3 times with TBS plus 0.1% Tween-20 (TBST).
(3) The following ~ Lulc (total volume 50 ~l) is inrllb~t~(1 in the coated wells vf the microliter plate for 60 ..-i..~ s:
(a) Antibody specifir- for the unfolded state of the target protein at a sufficient conrçntr~tion to give 50% of m~xim~l binding ~in the absence of competing 5target protein).
(b) Target protein at a co.~r~~ ion sllfflriçnt to achieve 9~%
inhibition of antibody binding to the plate. The a~ ia~ target protein collcenLldLion differs for each target protein. The collce~ dtion depends, in part, on the stability of the folded form of the target protein. In some cases it may be desirable to reduce the stability 2 oof the target protein by elevated ~ alul~" inrhl~ n of r~Pmir~l protein-~ie~ agents, or introduction of destabilizing amino acid substitutions in the target protein. (c) 10-9 to ~0-5M test ligands (d) 5 % nonfat dry milk in TBST

(4) The plates are washed 3 times with TBST.

(5) 50 ~l of goat anti-IgG ~lk~linP ph~ sph~t~e conjugate at an a~ru~lidLe dilution are added in TBST plus 5% nonfat dry miLk and i.~ le~l for 30 ~-.i-.~es at room lt~ lalul~.

(6) Plates are washed 3 times with TBST.
- (7) 0.1 ml of 1.0 mg/ml D-nitrophenylphosphate in 0.1% ~lip~h~nnlamine 30are added and the amount of color development recorded by means of an ELISA plate reader.

WO 97/20952 PCT~US96/19698 28 ELISA analysis will reveal more or less antibody bound to the plate when sllrce~sful test ligand-target protein binding has occurred than in the absence of such binding.

Example 5: Inhibition or ~.nhqn~mPnt of Chaperone Binding (1) ~LISA plates are coated by il~ ~Ib~ n for several hours with 4 ,ug/ml chape~ e in TBS.
(2) The plates are washed 3 times with TBST.
(3) The following nlLx~u~c (total volume 50 ,ul) is then inrull~terl in the coated wells of the microliter plate 10 for 60 ~ s:
(a) Target protein at a collcellL~aLion sl~ffi~ient to sd~uldle about 50%
of the available binding sites present on the ch~erolle ~loLei--s. Del~ulhlg conditions may be used in cases where the folded form of the target protein is otherwise too stable to permit d~reciable binding to chaperones.
(b) 10-9 to 1~5M test ligands in TBST
(4) Aliquots of the well solutions are ~ld~i~ll~,d to wells of a new ELISA
plate and inrllb~te~l for 60 ~-~ s at room lt;lllpeldLUle.
(5) The plate wells are washed 3 times with TBST.
(6) 50 ,ul antibody specific for the target protein at the appr~lid~e dilution in TBST, plus 5% nonfat dry milk, are added to each well and inr~lb~ted 30 ...i..~ s at room 2 o ~Ul~l dlUl~ .

(7) The plate wells are washed 3 times with TBST.

(8) 50 ~1 of goat anti-rabbit IgG ~lk~lin~ phosph~t~e conjugate at an a~plupliale dilution in TBST plus 5% nonfat dry miL~c are added to each well and incubated 30 min ltes at room ~w~ alulc.

(9) The plate wells are washed 3 times with TBST.

(10) 0.1 ml of 1.0 mg/mlp-nitropl1~lyl~ho~hale in 0.1% ~ th~nr)lamine will be added. Color development (proportional to ~lk~lin~ phosphatase antibody conjugate bound) is monitored with an ELISA plate reader.
ELISA analysis will reveal target protein in the solution at higher or lower 3 Ocol~cc;,lL,~lion when test ligand-target protein binding has occurred than when it has not.

CA 02236339 l998-05-20 WO 97/20952 PCT~US96/19698 ~,Y~.nrl~ 6: F'.nh~ or l~hibition of R;n(lin~ to a Known T.i~
(1) The following mixture (total volume 50 ,ul) is in~ t~l in the coated wells of the microliter plate for 60 ...;..~ s:
(a) Ligand known to bind to the target protein, covalently ~ h~d to 5solid beads such as Sephadex. This ligand can be a small molecule or a macromolecule.
(b) Target protein at a collcellLlalion well below saturation of the ligand and such that only 10% of the protein binds to the ligand sites. The solution conditions are such that most of the target protein is present in the denatured state.
(c) 10-9 to 10-sM test ligands (d) in TBST plus ~-oce~ dellalu~ , such as urea.
(2) Aliquots of the well ~u~ free of beads) are Ll~l~Ç.,ll~d to wells of a new ELISA plate and inrnh~ted for 60 minlltes at room temperature.
(3) The plate wells are washed 3 times with TBST.
(4) 50 ,ul antibody specific for the target protein 5at the a~pr~l;ale /lillltil~n in TBST, plus 5% nonfat dry miL~c, are added to each well and in~llb~t~(l 30 ~,-i--~ s at room ~ clalult.
(5) The plate wells are washed 3 times with TBST.
(6) 50 pl of goat anti-rabbit IgG alk~linlo phosph~t~ce conjugate at an ~r~liale dilution in TBST plus 5% miLk are added to each well and in~llb~tt~d 30 mimltes 20at room Lt~ ,e~al~
(7) The plate wells are washed 3 times with TBST.
(8) 0.1 ml of 1.0 mg/mlp-nitrophenylphosphate in 0.1% ~lieth~nolamine are added. Color development (~r~olLional to alk~linP phosphatase antibody conjugate bound) is monitored with an ELISA plate reader.
ELISA analysis will reveal a higher or lower concentration of target protein in the solution when sllrceccful test ligand-target protein binding has occurred.

~Y~mpl~ 7: Low Illlo~ l Assay for Carbonic Anl~y~ ,se T,i~n~l~
~ Ligand binding to carbonic anhydrase I (Sigma) was tested using proteolysis 3 oas a probe of target protein folding, and d~ . ;..g gel electrophoresis was used as a method for detection of intact protein rem~ining after digestion with ~roL~,ases.

To validate the assay, acetazolamide, a known ligand of carbonic anhydrase, was tested. Though acetazolamide is a known inhibitor of carbonic anhydrase activity, these experiments make no use of that ~r~elly, and do not measure the enzymatic activity of the protein. In addition, the se~iliviLy of the method to illL~lÇ~.ellce by a natural product extract 5was e~min~d.
rtio~ lul~,s co~t~inPd 13.3 ,ug/ml carbonic anhydrase, 0.05 M Tris-HCl pH 7.5, 0.01 M calcium acetate, 2.5 ~bglml proLei~se K, 10% DMSO and acetazolamide (Sigma) in concentrations rAngin~ from 0.0 to 1.0 mM. The reactions were inr~lb~t~d at 54~C for 15 ~ les~ and then chilled on ice. Phenyl methyl sulfonyl fl~lQri(le (PMSF) was Othen added from a 20 mM stock solution in ethanol to a final conce~ alion of 1 mM, and EDTA was added from a O.5M stock solution to a final collc~nllalion of 20 mM. 0.01 ml of SDS loading buffer (10% sodium dodecyl sulfate (SDS), 0.5 M Dithiothreitol, 0.4 M Tris-HCl buffer, pH 6.8, 50% Glycerol? was added and samples were heated at 95~C for 3 ill--le,s. Samples were analyzed by SDS-polyacrylamide gel electrophoresis using a 4-15%
spolyacrylamide (BioRad) gr~ nt gel, which was then stained with Coornassie Blue dye.
As shown in Figure 1, binding of the known ligand acetazolamide to carbonic allhy~ se resulted in stabilization of carbonic anhydrase against proteolysis by ~l~,L~inase K
at 1 X 10-5M acetazolamide. The dissociation co~il~llL for this interaction has been reported to be 2.6 X 10~M (~tCllmnto~ K. et. al. (1989), Chem. Pharm. Bull, 37:1913-1915).
A fungal ..~ ..ol extract was included in reactions that were otherwise irlentir~l to that described above such that the f~al collcellLlcl~ion of an added small molecule would be equal to its co~cenlldlion in the source culture. The presence of extract neither ..re~l a false signal nor ~ l,r~ the ,~ol~e to 1.0 mM açet~7ol~mi~lç (Figure 2.) 25~Y~mp'e 8: Low Throl~hr~t Assay for lHIV Rev I~leill Reaction mixtures (0.03 ml total volume) contained 30 ,ug/ml HIV Rev protein that had been produced in E. coli, 0.05M Tris-HCl, pH 7.5, 0.01M ç~lcillm acetate, 2.5,ug/ml pl ~L~ulase K, 10% DMSO, and varying amounts of tRNA as a known ligand. The reactions were in~ h~tP~l on ice for 15 .~ ules. After ~d~liti~n of PMSF and EDTA as 3 odescribed in Example 7 above, samples were prepared for gel electrophoresis and analyzed as described in Example 7.

W O 97/20952 PCT~US96/19698 The results showed that in the absence of tRNA, Rev protein is almost completely degraded by ~Lo~ei.,dse K under these conditions. In the l)Lese,lce of tRNA, however, a lower-molecular weight fragment of the protein is stabilized against proteolysis.
Thus, binding of a known ligand to HIV Rev protein is detect~ble using the methods of the 5present invention.

~,Y~mr'- 9: High~ ou~llput S~ g of T,ig~n~l~ for ~lml?n Neutrophil ~l~ct~ce In practicing the present invention, the ability to pelr~ the binding assay on olarge llu~ of compounds is critical to its utility in discovering compounds with potential ph~ relltir~l utility. Two dirr.,.~.lL approaclles have been s~lccP,ccfillly implenllont~d in a high-throughput s~;le~ g mode and each of these has been applied to two target ploL~illS:
human l~uLIol)hil elastase (E~NE) and human hemoglobin, both hemoglobin A (HbA) and hemoglobin S (HbS) (described in Example 10 below).
Notably, these target proteills differ from one another in a number of il~Ol L~Lr~s~e.;L~: HbS is an intracellular, lel~ Jic protein that contains a prosthetic group critical to its f~mCtio~ It is known to exist in two cv~ onC with diff~ structural and functional pr~e.lies. In contrast, HNE is monomeric, lacks a prosthetic group, and is secreted. HNE has an el~yllldLic activity (proteolysis) and does not appear to undergo any 2 oglobal cc llro.lllational çh~es.
For high-throughput screening with both of these target proteins, proteolysis is used as the probe of target protein folding. The two high-throughput modes differ in the mPthorlc used for detPcti-n of residual target protein following proteolysis. The two detPction m~thorls are 1) capture of radiolabeled protein on nitrocellulose fflters followed by 2 sq~l~ntit~til~n of bound r~tlio~ctivity and 2) measurement of protein by enzyme linked immlln-)sorbent assay (EIISA.) Each of these mPth~ rls was used sl~ccescfully with both hemoglobin and HNE.

A) Ni~ocellulose binding of ro~iol~helle~ HNE:
0.1 mg HNE (Elastin Products) was labelled by reaction with l2sI-Sodium Iodide (Amersham) in the presence of Iodogen (Pierce) according to m~mlf~ lrer's protocols (Pierce). Reaction mixtures were prepared in a final volume of 0.05 ml cont~ining radiolabelled HNE (20,000 cpm, coll~,*)olldihlg to approximately 10 ~g), 0.025 mg/ml WO 97t20952 PCT~USg6/19698 Bovine Serum Albumin, 50 mM Tris-HCl, pH 7.5, 10 mM calcium acetate, 2.5 ,~cg/mlthermolysin (Boeringer Ma~ heil,l), 2.5 ~g/ml y~L~hlase K (Merck), 10% DMSO, and the test compound at a concentration of 200 ~4M. Control mixtures were i<lçntir~l7 except that the test compound was omittrc~
The ~ ul~s were ;.1~ ~I1J~I~(1 at 20~C for 15 ...i..~ s, then at 65~C for 30 S, after which they were placed on ice. 0.12 ml 50 mM sodium acetate buffer, pH
4.5, was then added to each mixture. After an additional 15 minute incubation on ice, the samples were filtered through nitrocellulose membrane sheets using the Schleicher and Schuell Minifold. Each well of the appdlalus was then washed once with 0.2 ml 50 mM
0sodium acetate buffer, pH 4.5, and twice with 0.5 ml 50 mM sodium phosphate, pH 5.5, cont~inin~ 2.0% SDS and 1.0% Triton X-100. After drying the filter, bound r~lio~r~ y was dele~ hled by scintill~ti~ n counting using the Wallac MicroBeta a~"~
To validate the assay, a known ligand for HNE, c~ l, was inrlll-1r~1 in the assay at cc ~re~ iol-~ ranging from 1-5 mM. As shown in Figure 3, inclusion of sel~t~tin~l increased the retçnti- n of labelled HNE on the nitrocellulose filters, in~lir~tin~ that it l"oL~-;Led HNE from proteolysis.

B) ELISA Q~n7lfitnt;o.. of HNE:
Reaction 111L1~IU1~S in final volume of 0.05 ml eollldilled 2 ~Lglml HNE, 0.020 20mg/ml Bovine Serum Albumin, 50 rnM Tris-HCl, pH 7.5, 10 mM calcium acetate, 7.5 ,ug/ml therinolysin (Boeringer Mannheim), 7.5 ,ug/ml l)ruLei~dse K (Merck), 10% DMSO, and the test compound at 20 or 200 ~uM concentration. Control llli~Lules were j<1Pntir~l except that the test compound was omittecl The lllL~Ul~,S were inr~ trcl at 20~C for 15 es, then at 63~C, 30 ~ es then placed on ice.
2 5 0.1 ml of rabbit anti HNE antibody (Calbiochem) at a dilution of 1:10,000 in TBST (10 mM Tris-HCl, pH 7.5, 0.15 M NaCl, 0.05% Tween-20) c~ g 5% nonfat dry miLk (l~rn~tinn) was then added to each reaction. After 10 lllillu~es incubation at room L~ c.dLul~, the l~ Lures were t~dl~r~ d to 96-well Immulon~ plates (Dynatech) that had been coated with HNE by overnight inrnb~tinn with 0.1 ml per well of 0.2 ,ug/ml HNE in 3 o50 mM Sodium Borate buffer, pH 8.5, and 3 mM sodium azide and then washed thoroughly with TBST. The plates were then inr~lb~tPcl at room temperature for one hour, after which they were thoroughly washed with T33ST. 0.1 ml of ~lk~linP phosphatase-conjugated goat W 097/20952 PCT~US96/19698 33 anti-rabbit IgG antibody (Calbiochem) diluted 1:1000 in TBST cont~inin~ 5 % nonfat dry miLk was added to each well, and the plates were in~ atlo~l at room temperature for 1-2 hours.
The plates were then washed thoroughly with TBST and finally with TBST lacking Tween.
0.1 ml per ml of p-lliL.opllellylphosphate (0.5 mg/ml) in lX diethanolamine s~-hstr~te buffer 5(Pierce) was added to each well. Plates were in~lb~t~l at room L~ d~ulci until color developed, after which the absoll,ance of each well at 405 nm was measured using a BioRad 3550-UV microplate reader.
To validate the assay, a known ligand for HNE, ICI 200,3~5, was included in the assay at concellLI-dlions ranging from 0.01-10 ,uM. As shown in Figure 4, inciusion 0of the ligand caused an inhibition of antibody binding to the plate, inr1i-~tin~ an increased level of immlln-reactive HNE in the reaction mi~ul~s.

C) Results of High-17zroug*put SC~
3,600 compounds have been screened for interaction with HNE using sproteolysis and EL[SA ~letecti-n as above (Figure 5). Of these, 24 irlhibited proteolysis of HNE by p~ ~se K to an extent of 50% or more when assayed at a con-~entration of 20 ~M (positive hit compounds.) An additional 6 compounds were found to ill~lease the extent of proteolysis at least two-fold when tested at 20 ,uM (llegd~ e hit compounds.) The collcellL.d~ion depen~1en~e of the effects of hit compounds was measured. Hit compounds 20showed half m~xim~l effects at conce.llldlions as low as 8 ~4M; one example is shown in Figure 6. ~xim~l inhibition was usually, but not always, nearly 100%.
The hit compounds were assayed for their ability to inhibit the e~yllla~ic activity of HNE. Since compounds i(lPntifi~l in the binding assay may bind ~lywhele on the protein surface, only a small fraction would be expected to inhibit the el~ylll~lic activity 250f HNE. The compounds were tested as inhibitors of the proteolysis of Suc-(Alah-pNA
(Elastin Products), a chromogenic synthetic sul~sLl~le~ according to the method of Bieth, J, Spiess, B. and Werm~lth, C. G. (1974, Biochemical Medicine, 11:350-357.) Two positive hit compounds and one negative hit compound inhibit the proteolytic activity of HNE
.$i~nifi~ ~ntly in these assays (Figure 7).

W O 97/20952 PCT~US96/19698 s~mrle 10: High~ put S~ g of ~ i~nrl,~ for ~n~n HPmo~lokin A) Nitrocellulose Ri~ g of ~n~iol~helle~ Hemoglobin:
0.2 mg HbS or HbA (Sigma) was radiola~elled by reaction with 1 mCi l2sI-Bolton-Hunter reagent (Al,~ n) in 100 mM sodium borate buffer, pH 8.5, on ice for one 5hour. Labelling was stopped by addition of borate buffer cn..~ 200 mM glycine. The mixture was then fractionated by size on an execellulose GF-5 column (Pierce) in 50 mM
sodium phosphate buffer, pH 7.5, co..l;1i,.i-.~ 0.25% gelatin.
For the binding assay, reaction ~ ules in a final volume of 0.05 ml cont~inP-~l radiolabelled hemoglobin (20,000 CPM3, 0.063 mg/ml unlabelled hemoglobin, 100.034 mg/ml Bovine Serum Albumin, 50 mM Tris-HCl, pH 7.5, 10 mM calcium acetate, 2.5 ~g/ml thermolysin (Boeringer Marmheim), 2.5 ~4g/ml ploL~ ase K (Merck), 10%
DMSO, and test compound. Control mixtures were j~lçntir~l, except that the test compound was omitte~l The ll~LuleS were i~ d at 20~C for 15 mimltçs, then 40~C for 30 .~ s and then placed on ice. 0.12 ml 50 mM sodium acetate buffer, pH 4.5, was then added to 5eaCh llli~Ul~. After an additional 15 minute inrllh~tinn on ice, the samples were filtered through nitrocellulose membrane sheets using the Schleicher and Schuell Minifold. Each well of the a~palaLus was ~hen washed once with 0.2 ml 50 mM sodium acetate buffer, pH
4.5, twice with 0.5 ml of 50 mM sodium phosphate buffer, pH 5.5, cont~inin~ 2.0% SDS
and 1.0% Triton X-100. After drying the filter, bound radioactivity was ~leterminPd by 20scintill~tion counting using the Wallac MicroBeta a~pdldlus.
To validate the assay, a known ligand for hemoglobin, 2,3-diphosphoglycerate, was inrlll~ t in the reaction IlliXlUl-, at concentrations r~n~in~ from 10~ to 10-1M. As shown in Figure 8, 2,3-diphosphoglyc~laLe ~ignifir~ntty increased the filter rete~tinn of hemoglobin.

2SB) ELISA Ql~n~t~rff~n of Hemoglobin:
Reaction ~ lul~s in a final volume of 0.05 ml contained 0.063 mg/ml Hemoglobin, 0.034 mg/ml Bovine Serum Albumin, 50 mM Tris-HCl, pH 7.5, 10 mM
calcium acetate, 7.5 ~g/ml thermolysin (Boe.h~,l Mamlh~u,,), 7.5 ,ug/ml ~n)~e"~ase K
(Merck), 10% DMSO, and the test compound at 20 or 200 ,uM cullce,~ Lion. Control3 Oreactions were identical, except that the test compound was omitted.
The mixtures were inc~lb~te(l at 20~C for 15 mimltes, then at 44~C for 30 minnt~S, and then placed on ice. To each mixture was then added 0.05 ml O.lM sodillm WO 97/20952 PCTAUS96tl9698 borate buffer cont~inin~ 20mM EDTA and lmM PMSF. After 10 minlltes incubation on ice, the mi~ul~s were llal~r~ ,d to uncoated 96-well Tmmmlton-4 plates (Dynatech). The plates were then inrub~t~(l at 4~C overnight to allow binding of the protein to the plate. The plates were washed thoroughly with TBST, and 0.1 ml of rabbit anti-human hemoglobin antibody 5(Calbiochem) dilute 1:500 was added to each well. The plates were inr~lb~t~cl at room yel~Lul~, for one hour, then thoroughly washed with TBST. Next 0.1 ml of ~lk~linr phosph~t~e conjugated goat anti rabbit IgG antibody (Calbiochem) diluted 1:1000 in TBST
plus 5% nonfat dry miLI~ was added to each well and the plates were inr~lb~tP-1 at room lel~elaLul~ 1-2 hours. The plates were then washed thoroughly with TBST and finally with 10TBST lacking Tween. 0.1 ml per ml of p-nitrophenylphosph~t~ (0.5 mg/ml) in lX
th~n-~lamine substrate buffer (Pierce) was added to each well. Plates were inrllb~tr~l at room temperature until color developed and the absorbance of each well at 405 nm was measured using a BioRad 3550-W microplate reader.
To validate the assay, a known ligand for hemoglobin, 2,3,-15diphosphoglyc~ was inr~ ed in the reaction. As shown in Figure 9, this compound hl~ileased the (letrcti- n of i.. ~ .eactive hemoglobin.

C) Results of High-throughput Screening 4,()00 compounds have been screened for intrr~ction with HbS using 2 oproteolysis and ELISA ~letectiQn as above (Figure 10). Of these, 23 were found to inhibit proteolysis to an extent of 20% or more when assayed at a concentration of 20 ~M (positive hit c~ ullds.) The col-re~ t;~ n depenllPnre of the effects of hit compounds was measured.
25Hit compounds showed half m~im~1 effects at concentrations ranging as low as 2.0 L4M
(For example, see Figure 11).

~nr~e 11: Low-Throughput Fluo. ~ Based Assay for Carbonic Anhydrase ~;g~
Ligand binding to carbonic anhydrase I was tested using conformation-specific fluorescell~ probes as in-lir~tors of target protein folding. Reaction mixtures in a final volume of 0.1 ml co,llailled 2 ,uM human carbonic anhydrase (Sigma), 50 mM Tris-HCl, pH

W097/20952 PCT~US96/19698 7.6, 50 mM NaCl, 2.0 ~M bis-1-anilino-8-naphth~l~n~ sulfonate (bis-ANS) (Molecular Probes, Inc., Eugene, Ol?). Fluoresc~on~e emission of bis-ANS was measured at 450 nm after excitation at 365 mn. Mea~urclllents were performed using a Dynatech fluorescence microplate reader.
First, the effects of h~cl-,asillg coLlcellllalions of gl~ni-linillm hydrochloride (GCI) on bis-ANS fluorescence were çYAmin.od Figure 12 shows the fluorescence hllellsily of bis-ANS (in a~ laly units) measured 3 Illhlu~s after addition of GCl. In l~ Lul~S
lacking carbonic anhydrase, t_e fluorGscGIll yields of bis-ANS were low and were not affected by the plGsellce of GCl (Figure 12, filled squares). By contrast, carbonic allhydlase enhAnreA the fluol~scç~ e emission of bis-ANS in a manner that was sensilive to the co~-re..~ lion of GCl. In the absence of GCl, the fluolcscç~ e was enhAnt~ed 5-fold by carbonic anhydrase. Addition of between 0.5M and 2M GCI increased the fluol.,scellce proportionately, to a mAximllm of 15-fold ~h~rc~ l at 2M GCl. Further addition of GCI
caused fluol~scç~-re to decline. Figure 13 shows the time depçn-l~nre of GC~in(1~lced e ~ Anr~-.-Pnt of fluolGscellce in the Illh~lures described above.
These data inr~ir~te that the fluu.~,~ece.~e emie~io~ Gu~iLy of bis-ANS is .~reA by the folding state of carbonic anllyd~ase.
To validate the use of this assay in SCIGG11U1g for C&fl)CjI~~C anhydrase ligands, acetazolamide, a known ligand of carbonic al~ydlase, was tested for its ability to inflllen~e the fluoLesce~.re emission of bis-ANS in the Illi~Ul~,S described above. Figure 14 shows bis-ANS emission one minute after exposure of carbonic al~hydlase to 2M GCl and hlclGasi-lg co~r~ )ne of acetazolamide. The fluorescence emission was reduced by acetazolamide in a conce~ a~ion-dependent fashion. These data in-lir~t~ that binding of acetazolamide to carbonic anhydrase ~rev~ the GCI-in~ ced conversion of the folded fonn of the protein to a molten globule-folding int~rrnP(1i~te, and that this effect can be used as a measure of ligand binding.

F.Y~mrl~ 12: ~Iigh-Throughput Fluo.es~ e-Based Assay for Ligands Cc,l~o~ a~ion-selective fluol~sce~e probes are used in a high-throughput screening format to measure ligand binding according to the present invention. Mixtures of a fluorescelll probe, a target protein, and test compounds (and parallel control wells lacking test compounds) are provided in individual wells of 96-well microhter plates. After an ayyl.,yliate inrllh~tir)n period, the fluorescence in each well is ~l~p~terminpd using a ~ fluorescence plate reader (such as, e.g., Dynatech, Chantilly, VA). An increase or decrease in fluoresc~nre hllel~siLy in a well relative to that in a control well lacking test compounds inrlir~tt-s that binding of the test compound has occurred to the folded or unfolded state, respectively, of the target protein.
The conditions for each target protein are Apt-prminr~l by ~y~ tir~lly monitoring the change in probe fluolcscellce as conditions are varied from stabilizing to destabilizing. A substantial portion of the observed fluol- sce,-re illlellsi~y must be due to intPr~ction of the probe with the mo}ten globule state of the target protein in order for a measurable change in fluorescence to occur upon stabili7~tif)n of the folded state by- a ligand.
If nrcç~ry, rle~ conditions such as elevated lcl-lpel~lu-e or the addition of urea, ~ui..-i-li..~, or organic solvents are used to increase the fraction of target protein present in the molten globule state. If .-rcçc~.y, the presence of a molten globule state is verified by biophysical mea~uicll.cllL~ including NMR, viscometry, intrin~ir fluorescence, and size exclusion chromatography.
Under ayyl~oyli~ con~litionC, increasing fluorescenre is observed as the target protein is converted from a native to a molten globule state, while decreasing fluolescellce is observed upon conversion from a molten globule state to a random coil. In the case of some target ~r~leills, the molten globule state may pre~ even under "stabilizing"
conditions (i.e., in the ~bsenre of ~l~"~ g conditions listed above). In such cases, relatively high e~h~rr~ l of probe fluor~sce-~re by the target protein is observed even under stabilizing conditions, and decreasing fluolc;sce~.re is observed as conditions become more destabilizing. The ylesel~ce of a molten globule state can be verified as ~escrihecl above.
The reversibility of the native-to-molten globule col,Çol",a~ional change is also char~ct~Pri7~od. If the tr~n.~ition iS reversible, assays may be established under equilibrium conditions. In this case, an ill~;ul)aLion time is chosen that is long enough to allow binding ~ of the probe to the molten globule state of the target protein (in the absence of test colllyvullds) to reach equilibrium. If the tr~n.~ition is irreversible, an inrllh~tion time is chosen such that a mP~llr7.hle, but not complete, change in fluorescel~ce enh~nremPnt occurs;
such conditions may also be used in the case of reversible collr~".,.~tir,~l changes.

Claims (8)

What is claimed is:

1. A fluorescence-based screening method to identify a ligand that binds to a predetermined target protein, comprising the steps of:
(a) selecting as test ligands a plurality of compounds not known to bind to the target protein;
(b) incubating the target protein with each of said test ligands to produce a test combination, and in the absence of a test ligand, to produce a control combination;
(c) contacting said test and control combinations with a conformation-sensitive fluorescence probe;
(d) treating said test and control combinations under conditions that cause the target protein to unfold to an appropriate extent;
(e) measuring the fluorescence emission of said conformation-sensitive fluorescence probe in said test and control combinations; and (f) comparing the measurement made in step (d) between the test and control combinations, wherein if the fluorescence emission intensity of said probe is greater or lesser in the test combination than in the control combination, the test ligand is a ligand that binds to the target protein.

2. The method of claim 1 further comprising repeating steps (b)-(f) with a plurality of said test ligands until a ligand that binds to the target protein is identified.

3. The method of claim 1, wherein said fluorescence probe binds preferentially to the folded, unfolded, or molten globule state of the protein;

4. The method of claim 3, wherein said probe is selected from the group consisting of 1-anilino-8-naphthalene sulfonate (ANS), bis-1-anilino-8-naphthalene sulfonate (bis-ANS) and 6-propionyl-2-(N,N-dimethyl)-aminonaphthalene (Prodan).

5. The method of claim 4, wherein said fluorescence probe is bis-ANS.

6. The method of claim 1, wherein said treating comprises raising the temperature to which said test and control combinations are exposed, contacting said test and control combinations with a protein denaturant, or combinations thereof.

7. The method of claim 1, wherein said target protein contains stabilizing or destabilizing mutations relative to the wild-type version of said protein.

8. The method of claim 1, wherein said test ligand is selected from the group consisting of metals, peptides, proteins, lipids, polysaccharides, nucleic acids, small organic molecules, and combinations thereof.