WO2005066639A2 - Affinity displacement avidity effect protein resolution methods - Google Patents

Abstract

Compositions and methods to predict the Kd values of a compound for protein targets within a ligand binding proteome are provided. More specifically, the method relates to measuring the avidity effect of a ligand upon the formation of a complex of a protein and a compound of interest, then determining the affinity of the compound of interest for the protein from the measurement.

Description

AFFINITY DISPLACEMENT AVIDITY EFFECT PROTEIN RESOLUTION METHODS

FIELD OF THE INVENTION The present invention relates generally to techniques for determining the affinity of a compound of interest for at least one protein or polypeptide target.

BACKGROUND OF THE INVENTION Conventional drug discovery involves single enzyme screens against small molecule libraries containing drug like molecules and has been highly successful for uncovering new drugs to treat specific diseases where the underlying molecular mechanisms of the disease processes are largely understood. Diseases such as cancer, however, do not fall into this category, since the current level of understanding of the molecular causes of these diseases is rudimentary. Approaches that sample more of the genome/proteome simultaneously during the drug library screening process are more likely to discover potent, selective drug candidates and their targets than single enzyme/protein screens. These screens may be particularly effective in the discovery of anticancer therapeutics exhibiting polypharmacology (drugs that target multiple gene family members simultaneously). Drugs exhibiting polypharmacology may be less likely to develop resistance in the treatment of aggressive malignances. Whether exhibiting polypharmacology or targeting a single protein, successful drug candidates will ideally interact with their target selectively. Affinity of a drug target for a candidate compound is often used as a proxy for the selectivity of a test compound. While traditional solution phase kinetics assays are useful for determining the selectivity of a compound for a single enzyme, such assays become time consuming where a large portion of the proteome is tested. Accordingly, there is a need for assays that allow the assessment of a compound's affinity for multiple proteins simultaneously. Non-selective or low affinity interactions between a drug candidate and proteins other than the desired target may manifest themselves as the phenomenon of drug partitioning. This partitioning effect is a consequence of an avidity effect exerted by high copy proteins with low affinity for a particular compound. Thus if a low affinity binder is expressed in a tissue that is away from the desired site of action, the drug is likely to accumulate in that tissue. Drug partitioning can impact the behavior of a drug in two ways; first, by causing undesirable distribution of the drug to non-targeted tissues where it can cause side effect issues and, second, by sequestration of the drug reducing bioavailability. Accordingly, there is a need for assays that allow the assessment of avidity effects likely to be experienced by a drug candidate in a given proteome. The first drug partitioning impact, undesirable distribution of a drug leading to side effect issues, is exemplified by the accumulation of antimalarial drug chloroquine (CQ) in the retina during treatment of arthritis and lupus. (Treatment of arthritis and lupus is a second indication for antimalarials of this class.) Over time, the accumulation of CQ in the retina results in the development of retinopathy leading to blindness. In Graves et al. (2002) Mol. Pharmacol, 62:1364-72, a weak-binding CQ target was identified using proteome mining technology. Although weak-binding (the target has low affinity for CQ), the target is present in high copy in retinal cells and constitutes a large portion of the total cellular protein in the retina. Over a lengthy course of treatment, the CQ becomes sequestered in the retina because the high concentration of low affinity, high-copy target exerts a considerable avidity effect causing CQ accumulation. For the short times and low therapeutic concentrations involved in CQ treatment of malaria this side effect is not an issue. However, in chronic treatment of arthritis and lupus at the concentrations and lengths of time required to bring about a therapeutic effect, retinopathy can result. Chronic treatment will also likely be necessary for treatment of most cancers using drugs targeted towards certain proteomes such as the receptor-tyrosine kinases. Consequently, undesirable distribution of drugs to non-targeted tissue due to avidity effects exerted by low affinity, high copy proteins is likely to become a therapeutic issue in cancer treatment. Any technology that can measure this effect may enable these issues to be avoided or at least evaluated to offset potential problems during therapy. The second impact of drug partitioning, drug sequestration, is an acute problem that affects bioavailability of the drug towards the desired target. Where a drug exhibits polypharmacology that includes a sub-set of desired targets with high affinity, but also binds one or more seemingly benign targets of high copy number and low affinity, the benign target may absorb the freely available drug, thereby reducing or eliminating its bioavailability for the desired targets through an avidity effect. Weak affinity proteins that are expressed many hundreds- to thousands-fold greater than the intended target will exert a considerable avidity effect causing the drug to be effectively sequestered reducing the effective therapeutic concentration. Measuring these weak affinities therefore becomes paramount if one wishes to offset the impact of the avidity effect exerted by expression of high copy proteins.

BRIEF SUMMARY OF THE INVENTION Compositions and methods to predict the Kd values of a compound for protein targets within a ligand binding proteome are provided. The method comprises saturating multiple ligand interaction sites with a protein population. The protein- bound ligand complexes are contacted with a compound of interest. Those proteins that interact with the compound forming a protein-compound complex are liberated from the ligand. The protein-compound complex is then contacted with unoccupied ligand interaction sites which exerts an avidity effect upon the complex. The rate of rebinding of the protein to the ligand is directly dependent upon affinity of the compound for its protein targets. Further, as the compound-protein complex migrates under flow through the additional amount of ligand, its retention will solely depend upon the relative affinities of the compound for the protein versus the protein for the ligand. If the compound is a weak binder, the protein is delayed through an increase in time the protein is bound to the ligand. Compounds that are highly potent will remain associated with their protein target and thus elute very early. By comparing avidity effect upon a compound of interest with the avidity effect upon a compound of known Kd, the Kd_app for the compound of interest can be determined. The true Kd for the compound of interest can be determined from the Kd_app. It is recognized that the method can be manipulated in various ways including varying the compound of interest, modulating the concentration of the ligand interaction sites in either stage, modulating the time of binding, varying the concentration of compound and or ligand, temperature, and the like. The ligand is chosen to provide native proteome sampling with minimal diffusional constraints. Further, the ligand has protein-ligand kinetics that mimic solution phase kinetics. In one aspect, the ligand can be linked to media and utilized with the methods of the present invention. Examples of such media can be found in United States Provisional Application No. 60/453,697, filed January 22, 2003, titled "Alkyl-Lmked Nucleotide Compositions," as well as U.S. Patent No. 5,536,822, filed March 4, 1994, incorporated herein in their entirety.

DETAILED DESCRIPTION OF THE INVENTION

Compositions and methods for measuring the affinity of a compound of interest to at least one protein target are provided. Generally, the method comprises the steps of measuring the avidity effect of a ligand upon the formation of a complex of a protein and a compound of interest, then determining the affinity of the compound of interest for the protein from the measurement. The method allows for measuring the affinity by reversibly binding the protein target to the ligand, allowing the drug candidate to reach equilibrium binding with the protein target, contacting the equilibrated state with an additional amount of the ligand, and measuring the retention time of the compound-protein complex within the additional amount of ligand. Weak affinity compounds allow the protein to be delayed through an increased time the protein is bound to the ligand, resulting in long retention times. Potent compounds remain associated with their protein target, resulting in short retention times.

Definitions By "compound of interest" is intended any compound that is organic or inorganic, naturally-occurring or non-naturally occurring, as will be appreciated by one of skill in the art. For example, the compound of interest can be a compound from a combinatorial library or a chemical library. Examples of methods for the synthesis of molecular libraries can be found in the art, for example in: DeWitt et al.

- A - (1993) Proc Natl. Acad. Sci. USA 90:6909; Erb et al. (1994) Proc. Natl. Acad. Sci. USA 91:11422; Zuckermann et al. (1994). J Med. Chem. 37:2678; Cho et al. (1993) Science 261:1303; Carrell et al. (1994) Angew. Chem. Int. Ed. Engl. 33:2059; Carell et al. (1994) Angew. Chem. Int. Ed. Engl. 33:2061; and Gallop et al. (1994) J. Med. Chem. 37:1233. The compound of interest can also be a compound extracted from a single cellular organism, a multicellular organism, or from an organ or a tissue of a multicellular organism. Examples of such organisms include, without limitation, bacteria, algae, fungi, plant, fish, amphibians, mammals, and the like. The compound of interest can be a single compound, or alternatively, a mixture of compounds. By "Kd" is intended affinity of the compound of interest for its protein target. By "Kd_app" is intended the apparent affinity of the compound of interest for its protein target. Typically, classical solution phase kinetics assays are performed at ligand concentrations that not in the physiological range. Where the AD AT method is carried out using ligand concentrations in the physiological range, the apparent affinity of the compound of interest will therefore differ from that determined by classical solution phase kinetics. See Equation 1. By "Km" is intended the affinity of the protein for the ligand. By "Kn _app" is intended the avidity effect of the ligand for the protein. By "ligand" is intended any compound that is organic or inorganic, naturally- occurring or non-naturally occurring, that is capable of binding with a protein or proteins of interest. A suitable ligand will possess a solvent interface that allows it to be connected to a linker while still binding the protein or proteins of interest. For instance, a ligand can be a nucleoside and a nucleotide. A nucleoside and a nucleotide, as referred to herein, can be naturally-occurring or non-naturally- occurring. Furthermore, the nucleoside or nucleotide can be biologically or synthetically-derived using techniques that are standard to one of skill in the art. Suitable nucleosides include, without limitation, adenosine (A), guanosine (G), cytidine (C), thymidine (T) and uridine (U), and derivatives and analogs thereof. Nucleotides are nucleosides with at least one phosphate group (or thiophosphate group), for example, a monophosphate, diphosphate or triphosphate group. The nucleotide can have phosphate groups, thiophosphate groups, or phosphate mimics, or a combination thereof. The number of phosphate or thiophosphate groups is at least one, and can be one, two, three or more in number. Such nucleotides are often referred to in abbreviation, for example, AMP, ADP, ATP, GMP, GDP, GTP, etc., as is understood by one of skill in the art. The isolation or synthesis of nucleoside derivatives and analogs are accomplished using techniques that are standard in the art, see for example, Guranowski et al. (1981) Biochemistry 20:110-15; Yaginuma et al. (1981) J. Antibiot. 23:359-66; Robins et al. (1983) J. Am. Chem. Soc. 105:4059-65; Borchardt et al. (1984) J. Biol Chem. 259: 5353-58; De Clercq et al. (1987) Biochem. Pharmacol. 36:2567-75; Nan Calenberg et al. (1994) Helv. Chim. Ada. 77:631-44; Picher et al. (1996) Biochem. Pharmacol. 51:1453- 601; Franchetti et al, (1994) J. Med. Chem. 37: 3534-3541; Cowart et al. (1999) J. Org. Chem. 64:2240-49. Fischer et al, (1999) J. Med. Chem. 42: 3636-3646; van Tilburg et al. (1999) J Med. Chem. 43:1393- 400; Halbfmger et al, (1999) J. Med. Chem. 42: 5325-5337; Gendron et al, (2000) J. Med. Chem. 43: 2239-2247; and Hernandez et al (2002) J Med. Chem. 45:4254-63 the teachings of all of which are incorporated herein by reference in their entirety. A ligand can also be a signaling molecule, such as a hormone, pheromone, or neurotransmitter such as cortisol, progesterone, estradiol, testosterone, thyroxine, retinoic acid peptide hormones, such as insulin, growth factors, and glucagon, small charged molecules, such as epinephrine and histamine, or lipophilic molecules that bind to cell-surface receptors, such as the prostaglandms. A ligand could also be a second messenger, such as cAMP, and IP3. Further, a ligand can be a conserved protein that acts in a signal-transduction pathways, including GTPase switch proteins (trimeric G proteins and monomeric Ras-like proteins), protein kinases, and adapter proteins. Thus, a ligand can be a naturally occurring molecule that interacts with, for instance, G-protein coupled receptors, ion-channel receptors, receptors linked to cytosolic tyrosine kinases, or receptors with intrinsic catalytic activity. A ligand may also be a compound, naturally occurring or synthetic, from a chemical library. A synthetic compound from a combinatorial library can serve as a ligand. Examples of methods for the synthesis of molecular libraries can be found in the art, as described above. A ligand can be the result of a rational design based upon crystallographic data. For instance, a ligand could be a molecule possessing a pharmacophore such as a biaryl (heteroaryl) ureas and amides, coupled to an appropriate linker. By "ligand interaction site" is intended the region or structure of the ligand which interacts with a target protein. By "media" or "medium" is intended a ligand linked to a solid support or a tag, such that the solid support or tag is suitable for the separation of the linked ligand, and optionally, compounds (such as proteins, for example) bound to the linked ligand, from unbound compounds. A solid support can be any suitable support, such as a resin, or a particulate material, such as a bead, or a particle. Alternatively, a solid support can be a continuous solid surface, such as a plate, chip, well, channel, column or a tube. The material of a solid support will be of any suitable substance, compound or polymer, as will be appreciated by one of skill in the art. Examples include, without limitation, acrylamide, agarose, methacrylate polymers, methacrylate copolymers, cellulose, nylon, silica, glass, ceramic, a magnetized particle or surface, nitrocellulose, polystyrene, thermoresponsive polymers, and derivatives thereof. The choice of a suitable solid support will be apparent to one of skill in the art from the known characteristics of a solid support and the method of use of that solid support. The preparation and use of solid supports, such as agarose, are well known in the art (see, for example, Cuatrecasas and Anfinsen, "Affinity Chromatography" in Ann. Rev. Biochem. Snell et al, eds. (CA: Annual Reviews Inc.), 40: 259-278 (1971), the teachings of which are incorporated herein by reference in their entirety). As used herein, a tag is an agent that provides for the specific detection or capture of the linked ligand. For example, and without limitation, a suitable tag is biotin, avidin, streptavidm, a hapten, a fluorophore or a chromophore. Detection or capture of the linked ligand employs techniques that are standard in the art. Visual detection methods for fluorophore or chromophore-tagged linked ligands are readily understood by one of skill in the art. A "linker" can be selected from any suitable alkyl, heteroalkyl, cycloalkyl, heterocycloalkyl, aryl, heteroaryl, substituted, non-substituted, linear or branched group, or a combination of same. In one aspect, the linker is a hydrophobic linker. In another aspect, the linker is a hydrophilic linker. For example, the linker can be a 3, 4, 5, 6, 7, 8, 9, 10, or longer carbon chain. Furthermore, a linker used in the invention can be highly hydrophobic or moderately hydrophobic, as will be understood by one of skill in the art. Furthermore, a linker, such as described above, can be attached (for example, via a condensation reaction) to another linker to form a larger and/or a longer linker. Media or medium suitable for use with the methods of the present invention, and methods for the synthesis of such media or medium, are disclosed in United States Provisional Application No. 60/453,697, titled Alkyl-Linlced Nucleotide Compositions, filed January 22, 2003, incorporated herein in its entirety. By "protein target" is intended a protein or polypeptide, as well as fragments, biologically active portions, and variants thereof that binds a compound of interest and a ligand. Protein targets may be mature or immature, glycosylated or unglycosylated, as will be understood by one of skill in the art. Protein targets may be produced by recombinant means or isolated from cells or tissue from an organism, specific organ, or subpart thereof. By "proteome" is intended a complex mixture of proteins that are derived from a common source, such as an extract isolated from a particular cell or tissue. For example a human proteome represents a mixture of proteins isolated from human cells. The category can be further defined by specifying a particular cell/tissue source for the proteome (i. e. a human myocardial tissue proteome represents all the proteins isolated from human myocardial tissue). The category can also be defined by a shared binding interaction with a particular ligand. For instance, one example of a binding proteome is the purine binding proteome. Purine binding proteins include over 510 distinct protein kinases, Manning et al. (2002) Science, 298:1912-34, as well as non-protein kinases. One can construct an artificial binding proteome by various methods, including recombinant expression of a particular class of binding proteins. Recombinantly produced binding proteins can be treated enzymatically to convert the recombinantly expressed protein to the inactive conformation (inactive conformer). Alternatively, a rational site directed mutagenesis program can be pursued to produce a mutein that, when expressed, assumes the inactive conformation. Affinity displacement avidity effect protein resolution. Affinity displacement avidity effect protein resolution technology (AD AT) comprises two elements, the true affinity of a compound of interest (such as a drug candidate) for its protein targets, and avidity, the affinity of protein for an immobilized ligand (such as a substrate) at varying density. By exploiting these two opposing affinities it is possible to accurately predict the affinity of a compound of interest for its target protein (called "the Kd"). When performing AD AT, a ligand is saturated with protein. Essentially all of the available (ligand) binding sites are filled. Under these conditions the behavior of any one protein captured by the ligand in the presence of a selective inhibitor (the compound of interest) is governed by the expression:

Equation 1. Kd_app = Kd(l+ [ligand) '/Km) Kd_app - the apparent affinity of the compound of interest for its target Kd - the true affinity of the compound of interest for its target Km - the dissociation constant for the ligand

If the affinity of the compound of interest for the protein (the Kd) is qualitatively greater than the affinity of the protein for the ligand (the Km) the protein is liberated. If the affinity of the protein for the ligand (the Km) is greater than the affinity of the compound of interest for the protein (the Kd) the protein will tend to remain bound to the ligand. This is the fundamental principle underlying proteome mining and is referred to as affinity displacement. After the compound-proteimligand-protein have reached equilibrium, the compound of interest-protein complex is allowed to interact with or contact an additional amount of the ligand with unoccupied binding sites. This now exerts an avidity effect upon the compound of interest-protein complex because the apparent affinity of the protein for the ligand is increased (qualitatively). The rate of rebinding, therefore, is directly dependant upon the affinity of the compound of interest for its protein target (Kd). Further, as the compound-protein complex migrates under flow through the additional amount of ligand, its retention will solely depend upon the relative affinities of the compound for the protein versus the protein for the ligand. If the compound is a weak binder, the protein is delayed through increased time the protein is bound to the ligand. Compounds that are highly potent will remain associated with their protein target and thus elute very early. By comparing avidity effect upon a compound of interest with the avidity effect upon a compound of known Kd the Kd_app for the compound of interest can be determined. As described above, the true Kd for the compound of interest can be determined from the Kd_app. In general, the AD AT methods can be performed using general chromatography techniques and apparatus. Chromatography techniques and apparatus compatible with the AD AT methods of the invention include, but are not limited to, high-performance liquid chromatography (HPLC), fast performance liquid chromatography (FPLC), thin-layer chromatography (TLC), and radial chromatography. Standard protocols and equipment for HPLC and FPLC are known in the art. See Practical HPLC Methodology & Applications, Wiley-Interscience, John Wiley & Sons, Inc., 1993; Cunico, R. L., Gooding, K. M., Wehr, T.; Basic HPLC and CE ofBiomolecules; Bay Bioanalytical Laboratory: Richmond, CA, 1998. Where AD AT is carried out by a liquid chromatography technique such as HPLC or FPLC, the ligand is linked to a solid media with minimal diffusional constraints wherein the protein-ligand kinetics mimic solution phase kinetics. Ideally, the ligand is connected to the media in such a manner that the media (including the linker that connects the ligand to the media), will not disrupt the interaction between the ligand and protein target. For instance, the media and the linker should neither bind, repel, or stearically hinder the target protein. Further, the media should not cause a significant diffusional constraint upon the target protein. Sample media and methods for its synthesis can be found in United States Provisional Application No. 60/453,697, filed January 22, 2003, titled "Alkyl-Linked Nucleotide Compositions," and U.S. Patent No. 5,536,822, filed March 4, 1994, the teachings of which are hereby incorporated by reference in their entirety. When carrying out AD AT by liquid chromatography, two media stages are utilized. Generally, the ligand is the same for both stages, although in certain AD AT techniques the ligand will differ between the two stages. The first stage media is saturated with a solution containing at least one protein, such as a cell extract, etc., until essentially all of the ligand interaction sites are filled. The compound of interest is then added to the first stage media and the combination of compound- protein:ligand-protein allowed to reach equilibrium. Once the combination has reached equilibrium, flow is applied. Compound-protein complex is therefore moved to a second stage media that contains ligand with unoccupied binding sites which exert an avidity effect upon the compound-protein complex. The rate of rebinding is directly dependent upon the Kd of the compound of interest for the protein. Under ideal conditions Kd can be determined from the retention time and peak shape of the compound of interest-protein complex as it emerges from the second stage of the AD AT apparatus. Where the affinity for a specific ligand of all proteins in a specific ligand binding proteome is very similar, use of the AD AT method with the ligand provides exquisite separation of protein targets for a given compound. Where affinity for the ligand is very similar across a proteome, the Kd, as determined from the general equation Kd_app = Kd(l+[ligand]/Km), can be determined for multiple protein targets simultaneously. AD AT therefore enables the actual partitioning of a compound of interest, such as a lead, to be accurately determined for any lead identified via proteome mining. This is particularly important for evaluating potential drug leads that exhibit polypharmacology. One non-limiting example of a proteome in which the affinity of all proteins in the binding proteome is very similar is the purine binding proteome. Purine binding proteins include over 510 distinct protein kinases, Manning et al. (2002) Science, 298:1912-34, as well as non-protein kinases. Several such molecules have proven successful drug targets, such as dihydrofolate reductase, Christopherson et al. (2002) Ace Chem Res., 35(11):961-71 and HSP90, Neckers et al. (2002) Trends MolMed. 8:S55-61. Due to the similarity in affinity, purine binding proteins will be similarly affected by the avidity effect and are suitable for use with the AD AT methods of the invention. Suitable affinity chromatography techniques for use with the present methods can be found in, for example, WO 00/63694, filed 12 April, 2000 and U.S. Patent No. 5,536,822, filed March 4, 1994, the teachings of which are incorporated herein by reference in their entirety. Manipulatins the avidity effect. In AD AT, the avidity effect can be manipulated in several ways, including increasing or decreasing (1) the ligand concentration of the second stage; (2) the dimensions of the second stage (longer stages at low density exert the same effect as increased density in a short stage); (3) the flow rate of the materials, i.e. proteins, protein-compound complex (the slower the flow the more influence on/off rates will have on rebinding); and (4) temperature (increased temperature enhances on/off rates). By performing a series of AD AT assays using compound-protein combinations with known Kd, where each combination differs by Kd, a spectrum of comparative standards is generated. Compounds with strong affinity will result in narrow, sharply peaked elution profiles and shorter retention times; those with low affinity will result in broader elution profiles and longer retention times. Ligand density effect. The ligand density effect, which is determined by the concentration of ligand in the second stage, as described by the following equation:

Equation 2. Kd_app - Kd(l + [Stationary Ligand_x]/Km) [Stationary Ligand_x] - the effect of changing the ligand density on Kd_app values determined using AD AT.

The ligand density effect can be manipulated by utilizing affinity media synthesized with a particular ligand density. Varying ligand density will impact the apparent affinity of each protein for the stationary ligand. Increasing ligand density will lower the numerical value for the apparent Km (i.e., increasing the apparent affinity of the protein for the ligand). In the presence of a fixed concentration of compound of interest this will progressively lengthen retention time and broaden peak shape. Potentially denser media could have better resolving power, however, excessive peak broadening will result in sample dilution. Lower density media will increase the numerical value of the apparent Km, causing peak sharpening but a reduced resolution. It is recognized that the optimal ligand density can be determined by conducting trial runs for each ligand to identify a ligand density that is optimal for protein resolution and peak shape for the most potent inhibitors. General methods for comparing and assessing elution peaks are provided in Bidlingmeyer et al. (1984) Analytical Chem., 56:1583-96. A compromise between high vs. low ligand density is to utilize a second stage gradient density media. A gradient density media is synthesized within a prepacked column by introducing the ligand to be immobilized at the top of the media under slow flow. The ligand cross links with upper most part of the media, which rapidly becomes saturated. As the ligand progresses through the media a gradient of ligand cross-linking is established from high to low density. Media of various ligand densities can be created in a systematic manner without employing a different resin by incubating the activated polymeric resin with varying ratios of the linker molecule and a capping compound. For example, where the linker molecule contains diamines, one would use a capping compound that possesses a single nucleophilic amine. The free amines in capping reagents such as ethanolamine or glycine will therefore compete with the amino groups in the linker for reactive sites on the resin. Both of these capping reagents limit the ligand density but preserve a hydrophilic resin environment. Where ATP or a similar molecule is the ligand, replacement of a significant fraction of the ligand with a simple capping group may result in significant changes in the gross physical properties of the media. If so, an alternative approach is to employ a mixture of ATP with ATP analogs known to be too large to bind into the normal

ATP site. N-6-Benzyl-ATP is an example of such a compound, it has been shown not to bind to the ATP site in native kinases and will only work in kinases that have been mutated to accept larger analogues. See Bishop et al. (2002) Trends Cell Biol 11:167-172 and Bishop et al (2000) Nature 407:395-401. Non-limiting examples of acceptable ligand densities for media vary several orders of magnitude, from 1 fmol/ml to 500 μmol/ml. Accordingly acceptable ligand densities for the second stage media include at least about lnM, 50 nM, 100 nM, 150 nM, 200 nM, 250 nM, 300 nM, 350 nM, 400 nM, 450 nM, 500 nM, 550 nM, 600 nM, 650 nM, 700 nM, 750 nM, 800 nM, 850 nM, 900 nM, 950 nM, 1 μM, 50 μM, lOOμM, 150 μM, 200μM, 250 μM, 300uM, 350 μM, 400μM, 450 μM, 500μM, 550 μM, 600μM, 650 μM, 700 μM, 750μM, 800μM, 850 μM, 900μM, 950 μM, 1 mM, 50 mM, 100 mM, 150 mM, 200 mM, 250 mM, 300 mM, 350 mM, 400 mM, 450 mM, or 500 mM. Standard density media of about 5 μmol/ml to about 15 μmol/ml may be used for the first stage, since all available ligand interaction sites will be saturated with protein. Nonetheless, acceptable ligand densities for the first stage media include at least about lOOμM, 150 μM, 200μM, 250 μM, 300uM, 350 μM, 400μM, 450 μM, 500μM, 550 μM, 600μM, 650 μM, 700 μM, 750μM, 800μM, 850 μM,

900μM, 950 μM, 1 mM, 50 mM, 100 mM, 150 mM, 200 mM, 250 mM, 300 mM, 350 mM, 400 mM, 450 mM, or 500 mM.

Dimensional effect. Generally, increasing the length of the second stage will have a similar effect to increased ligand density, causing an apparent numerical decrease in the Km (i.e., higher apparent affinity) of the protein for the ligand. A longer second stage column at lower ligand density may be an optimal solution for resolution of both high and low molecular weight proteins. A high-density media may exert tremendous avidity effect, giving good resolution of proteins during elution. However, densely packed ligands may cause molecular exclusion of high molecular weight proteins. A compromise therefore is to use a longer column at a lower density. Since proteins will take longer to migrate through the second stage, this should compensate for a weaker avidity effect. However, longer columns are generally prone to cause peak broadening. Nonlimiting examples of acceptable column lengths include 5mm, 10mm,

20mm, 30mm, 40mm, 50mm, 60mm, 75mm, 100mm, 125mm, 150mm, 200mm, 250mm, 300mm or more. Non-limiting examples of acceptable diameters include 1.0mm, 1.1mm, 1.2mm, 1.3mm, 1.4mm, 1.5mm, 1.6mm, 1.7mm, 1.8mm, 1.9mm and 2.0mm.

Flow rate. Flow rate impacts protein resolution and peak broadening in two ways. First, slower flow rates increase occupancy time enabling the protein to re- equilibrate with the ligand. The impact of slow flow rates therefore is the same as increased density, although this is offset by diffusional effects which will cause peak broadening and loss of resolution. Factors such as on rates (rate the protein binds to the ligand) versus off rate (rate of protein release) can impact interpretation of flow rate. On /off rates between the compound of interest and protein are also an important variable. If the compound of interest off rate is very slow then the protein is likely never to interact with the ligand again following its liberation from the first stage and flow rate will have little impact other than increasing peak width through diffusion if it is too slow. Therefore, for each ligand, trials are carried out to determine optimal flow rates that minimize diffusional effects while retaining protein resolution in cases where a particular compound targets more than one protein. General liquid chromatography methods for analyzing migration and elution of compounds are described in Snyder et al. (1983) Analytical Chem., 55:1412-30. Temperature. Temperature will effect the thermodynamics of the on/off rates, effectively making them reach equilibrium faster as temperature rises. At high temperature protein stability will be an issue. Low temperatures could improve protein resolution at constant flow by slowing equilibrium rates, i.e., by favoring binding of a compound to its protein partner. For low affinity compounds, lower temperatures may favor their resolution from the second stage column. Therefore, for each ligand, trials are carried out to detennine optimal temperature to maximize protein stability and cbromatographic resolution. Exemplary temperatures include 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37°C.

Identification of Compounds of Interest for use with AD AT Compounds of interest may be known drug candidates or other such compounds and molecules. Alternatively, compounds of interest can be identified by screening libraries of test compounds. For instance, compounds of interest can be identified utilizing Proteome Mining. See, e.g., International Application No. WO 00/63694, incorporated herein by reference in its entirety. In brief, proteome mining involves three general steps: First, the reversible capture of an entire binding proteome from a tissue of interest. This step is conducted under conditions that allow both high and low copy proteins to be captured efficiently. Proteome mining can be carried out in a format comprising 96-micro column arrays pre-packed with affinity media at an effective ligand concentration similar to that of naturally occurring ligand. See, e.g., United States Serial No. 10/022,931, filed December 18, 2001 for "Integrated Protein AffinityCapture Centrifugation Device," the teachings of which are incorporated by reference in their entirety. Sufficient biomass is passed through each channel in the array to ensure recovery of 1 fmol of any binding protein. The reversibly captured native proteins are washed to remove non-specifically bound proteins. The second step is treatment of individual wells is in parallel with test compounds. Test compounds can include pharmaceuticals, hormones, chemotherapeutics, nucleic acids, and the like. They can be naturally occurring or obtained from a chemical library, such as a combinatorial library. If the compounds compete for the binding site on any captured protein, they will elute that protein from the media. The eluents from each channel are collected in parallel and subjected to the third step of analysis. The third step of analysis involves separation of the proteins by standard analytical techniques. The protein bands are then analyzed by mass spectrometry. Sequence information from the MS/MS spectra is compared against both DNA and protein databases to identify the proteins. A compound of interest may be identified for further analysis on the basis that it selectively interacts with a protein or proteins, or because it interacts with a protein or proteins found in transformed versus normal tissue, etc.

Preparation Of Sample From Tissue For Use With AD AT To ensure identification of low copy proteins by mass spectrometry, sufficient tissue/cell mass must be utilized such that any one protein expressed at 100 copies per cell is recovered by the media. For example, using ABI QSTAR fitted with nano source and ABI MALDI TOF-TOF, a minimum of 1 fmol of protein is required for positive identification. Therefore, the minimal cell mass required to recover any one single protein expressed at 100 copies per cell is 10 cells (equivalent to ~lmg w.w. tissue). To allow for loss of protein recovery during extract processing, etc., minimal tissue mass should be roughly 10 mg per experiment. Proteins can be extracted utilizing the following nonlimiting protocol. Tissue or cells are obtained and frozen. The frozen tissues/cells are ground into a powder in liquid nitrogen and weighed. Extracts are prepared by mild homogenization (using a dounce) in the presence of isotonic pH 7.4 buffers at low ionic strength. Extracts (plasma membrane and cytosol) are clarified by centrifugation and mixed with 100 μl of each media to be tested. Sufficient extract will be added to the media to ensure complete saturation of all of the available ligand sites. Following incubation for a fixed interval (usually 5 minutes at 25°C), non-specifically associated proteins are removed by packing the loose media into micro column arrays. The media is eluted with several column volumes (100X) homogenization buffer containing high salt (1M NaCl). The media are washed into homogenization buffer at physiological ionic strength. Other techniques for protein extraction are known in the art and may be substituted for the preceding protocol. See Sambrook et al, Molecular Cloning - A Laboratory Manual, 2^nd ed., Vol. 1-3 (1989); and Current Protocols in Molecular Biology, F.M. Ausubel et al, Eds., Current Protocols, a joint venture between Greene Publishing Associates, Inc. and John Wiley & Sons, Inc. (1994). AD AT apparatus. One non-limiting example of apparatus for performing AD AT via liquid chromatography comprises a HPLC equipment with an automated sampler, photodiode array detector, and fraction collector. All tubing is protein compatible peak microbore tubing. First and second stage media are contained within a single glass micro column. Typical micro columns are l-2mm in diameter. See Simpson et al (1987) J. Chromatogr.400:297. Such columns are known in the art and may be purchased from vendors such as Pharmacia™. Various densities and lengths of stage are packed into the glass column in equilibration buffer. Satisfactory equilibration buffers include phosphate buffered saline (physiological pH and ionic strength). Following saturation with a targeted proteome, a fixed volume of first stage media is gently layered on the top of the second stage (see Preparation of Sample, above). The column is sealed and attached to the HPLC. The column is washed with equilibration buffer at various rates of flow and temperature. Various compounds of interest are introduced using the auto sampler and elution of proteins monitored by absorbance. Methods of analyzing elution by absorbance are known. See McClure (1994) Analytical Chem., 66:44. Fractions will be collected for analysis by SDS- PAGE and mass spectrometry. Gels are stained with a fluorescent stain such as sypro ruby (a highly sensitive fluorescent protein stain that can readily detect less than 1 fmol of total protein, i.e., less than 0.04ng for a 40kDa protein). The gels will be imaged using a standard flat bed gel imager. Analysis of Eluted Proteins from ADAT Analysis To identify a protein eluted from an ADAT column, various methods of mass spectrometry can be used. MALDI TOF-TOF mass spectrometry can be used to generate a peptide mass fingerprints (PMF). See, e.g., Karas et al. (1988) Anal. Chem. 60:2299-2301. The PMF can then be used to interrogate a protein database. Individual tryptic peptides within each PMF spectra can be sampled to derive MS/MS spectra to further interrogate the databases to identify the protein. Alternatively, electrospray mass spectrometry spectra can be used to derive de novo amino acid sequence to identify the protein. Data bases can then be searched with the T/FASTS algorithm, available on the internet at fasta.bioch.virginia.edu (note that no www prefix is needed). See Mackey et α/. (2002) Mol CellProteomics, 1 :139-47.98. See also Cone (Ed.) Handbook ofProteomic Methods, Humana Press, Totowa, New Jersey (2003).

Usins ADAT to determine Kd. As explained above, ADAT analysis is purely governed by the K _app (the avidity effect of the ligand for the protein) versus the Kd (the affinity of the protein for the compound of interest). These two factors alone will govern the migration time through the second stage. Therefore migration time and peak shape can be used to determine true Kd values for any compound tested by ADAT analysis. In the ADAT methods described above, so long as an excess of stationary ligand as compared to the compound of interest is used, the amount of compound of interest added does not impact the analysis. This is because the retention of the protein in the second phase under flow will be governed purely by the Km of the protein for stationary ligand versus affinity of the compound of interest for the protein (Kd) irrespective of protein concentration. In other words, although a nM concentration of compound of interest may only displace 1% of the total protein from the first stage, as the compound of interest-protein complex migrates through the second stage under conditions of flow its retention will solely depend upon its relative affinities for stationary versus bound ligand. If the compound of interest is a weak inhibitor, the protein migration is likely to be delayed through binding exchanges between the compound of interest phase and the stationary ligand phase. Compounds of interest that are highly potent will remain associated with their protein partner and thus elute very early. Increasing the concentration of compound of interest in the assay is unlikely to effect retention time, since the free unbound ligand will always be in significant excess over the actual compound of interest concentration (μM compound of interest versus mM stationary ligand). ADAT experiments are repeated with a defined proteome composed of at least one protein: compound pair for which the Kd is known. These trials allow the construction of a standard retention time vs. Kd elution table. Because retention time is directly dependent upon Kd_app, Kd is related to retention time by the expression Kd_apP = Kd(l+[ligand]/Km). Accordingly, the elution table can be used to accurately predict the Kd of any compound of interest-protein combination, at concentrations within the parameters discussed above tested (provided the eluted protein is detectable). Alternatively, a relative determination can be made by conducting an ADAT experiment with one compound of known affinity: By running a compound of interest in one ADAT trial and a known compound (that has the same protein target) in a duplicate trial, the compound that elutes the protein target first has the greater affinity. In the case of the purine binding proteins, exemplary pairs of protein: compounds suitable to generate a standard elution table are set forth in Table 1.

Generally, ADAT assays are carried out at l-500μM compound of interest concentrations. More particularly, ADAT assays are carried out with at least about 1 μM, 10 μM, 20 μM, 30 μM, 40 μM, 50 μM, 60 μM, 70 μM, 80 μM, 90 μM or 100 μM, 110 μM, 120 μM, 130 μM, 140 μM, 150 μM, 160 μM, 170 μM, 180 μM, 190 μM or 200 μM, 210 μM, 220 μM, 230 μM, 240 μM, 250 μM, 260 μM, 270 μM, 280 μM, 290 μM or 300 μM, 310 μM, 320 μM, 330 μM, 340 μM, 350 μM, 360 μM, 370 μM, 380 μM, 390 μM or 400 μM, 410 μM, 420 μM, 430 μM, 440 μM, 450 μM, 460 μM, 470 μM, 480 μM, 490 μM or 500 μM of compound. The ligand concentration is generally more than one order of magnitude higher. Where the ligand is derived from a naturally occurring cofactor, a physiological range is generally used. In the case of the purine binding proteome, for instance, where the immobilized ligand can be ATP, the media ligand density is physiological, i.e., 5-15 μmol/ml (5-15 mM). Under these conditions, a low copy protein that is targeted by a compound of interest with nM affinity would be detectable in the assay. In one embodiment, the ADAT method of the invention can be utilized to determine the Kd of a compound of interest for its target proteins utilizing purine ligand-linked media described in United States Provisional Application No. 60/453,697, filed January 22, 2003. Suitable apparatus for carrying out the method comprises an existing microbore HPLC apparatus with automated sampler, photodiode array detector, and fraction collector is modified for carrying out ADAT with a micro-column 5 mm long by 2 mm wide. In this embodiment, two media stages are used. For the first stage, standard density media of 100 μM to 500 mM is used, such as at least 150 μM, at least 200 μM, at least 300 μM, or at least 400 μM. Second stage media is used with a ligand density of between about 1 μM to about 500 mM, such as at least 5 μM, at least 10 μM, at least 25 μM, at least 50 μM, at least 100 μM, at least 150 μM, at least 200 μM, at least 250 μM, at least 300 μM, or at least 350 μM. A standard retention time vs. Kd elution table is constructed as described above. The proteome is extracted from a tissue of interest using 100 mg w.w. of cells. The cells are frozen and ground into a powder. Extracts are then prepared by mild homogenization in a suitable isotonic buffer at low ionic strength with a pH of about 7.4. Extracts are clarified by centrifugation and the supernatant is mixed with roughly 100 μl of media. Sufficient extract is added to first-stage media to ensure complete saturation of all of the available ligand sites. Following incubation for about 5 minutes at about 25°C, non-specifically associated proteins are removed by packing the loose media into micro column arrays. The media is eluted with several hundred column- volumes homogenization buffer containing about 1M NaCl. The media are washed into homogenization buffer at physiological ionic strength. The first stage media is then gently layered on the top of the second stage media. The column is sealed and attached to the HPLC. The column is washed with a suitable equilibration buffer. The compound of interest (mobile phase) added at fixed concentration using the auto injector. A volume of a solution containing the compound of interest equal to the first stage volume is added. Once equilibrium is reached, a constant flow is applied. Elution of proteins is monitored by absorbance. Peak shape and time of elution are noted. Fractions are collected for analysis by SDS-PAGE and mass spectrometry. Gels are stained with a fluorescent stain such as sypro ruby and imaged using a standard flat bed gel imager. Identification of the eluted proteins is carried out using ABI QSTAR fitted with nano source and ABI MALDI TOF-TOF. The elution time and peak profile for each identified protein are noted and stored in a database.

ADAT And Drus Partitioning. The likelihood that a drug candidate will undergo partitioning in a particular organ can be tested using the ADAT method. Tissue from a particular organ or organs is obtained and extracted, as described above. ADAT is carried out as described above. Proteins are eluted and monitored by absorbance. Fractions will be collected for analysis by SDS-PAGE and mass spectrometry. Gels are stained with a fluorescent stain such as a fluorescent stain such as sypro ruby. Alternatively, proteins can be identified by mass spectroscopy. The identification of a protein known to be present at high copy number is indicative that sequestration will likely occur.

This invention is further illustrated by the following examples which should not be construed as limiting. EXPERIMENTAL EXAMPLE 1 Determining Optimal Conditions For ADAT With Specific Media Utilizing ADAT in conjunction with an existing microbore HPLC apparatus with automated sampler, the Ki of a known inhibitor is determined against a defined proteome using ATP as the stationary ligand. Suitable ATP media is disclosed in United States Provisional Application No. 60/453,697, filed January 22, 2003. Defined proteome preparation: Expressions of all 58 RTKs from the human genome in Sf9 cells. The completion of the human genome has enabled all the coding regions for all 58 RTKs to be identified. The coding regions for the soluble cytoplasmic domains of various RTKs are cloned by PCR. Template cDNA pools is prepared from RNA isolated from human placenta and various human cell lines and the relevant portion of each RTK cloned by standard PCR techniques. See Sambrook et al, Molecular Cloning - A Laboratory Manual, 2" ed., Vol. 1-3 (1989); and Current Protocols in Molecular Biology, F.M. Ausubel et al, Eds., Current Protocols, a joint venture between Greene Publishing Associates, Inc. and John Wiley & Sons, Inc. (1994). The PCR products are cloned into pFastbac-HT plasmid from the Bac-to-Bac Baculovirus Expression System™ from Invitrogen™ according to the manufacturer's protocol. The sequences of the clones verified by DNA sequencing. The pFastbac- HT constructs are used to produce recombinant baculoviruses. High-titer baculoviruses are produced and used to infect Sf9 cells for protein production. Cells are harvested 60-72 hours post-infection, according to the kit manufacturer's protocol. Cell pellets are resuspended in hypotonic buffer, lyzed by sonication, and centrifuged for one hour at 100,000 x g. The supernatant is mixed with ATP affinity media and washed to remove non-specific bound proteins.

ADAT HPLC Protocol An existing microbore hplc apparatus with automated sampler, photodiode array detector and fraction collector is modified for carrying out ADAT. All metal tubing is replaced with protein compatible peak microbore tubing. Initially first and second stage media will be contained within a single glass micro column (Pharmacia™). First stage media is saturated with the defined proteome, and a fixed volume of first stage resin is gently layered on the top of the second stage media. The column is sealed and attached to the HPLC. The column is washed with equilibration buffer and inhibitors are introduced using the auto sampler at fixed concentrations successively to the flow using the auto injector. Sufficient volume is added to saturate the entire Stage 1 and Stage 2 resins.

Optimization. Iterative ADAT trials are carried out with inhibitors of known Ki using the HPLC protocol and defined proteome described in the preceding sections to optimize the parameters of Ki effect, ligand density, column length, flow rate, and temperature.

Ki Effect. Ki effect is optimized using several well characterized protein kinase inhibitors including Iressa and Gleevec, both RTK inhibitors; BIRB796, SB2976, which inhibit p38mapk; methotrexate, which inhibits dihydrofolate reductase; radicicol, which inhibits HSP90; and chloroquine, which inhibits QR2. Since Ki values for their respective protein targets are well documented in solution phase enzymatic assays, these compounds are used as comparative standards to optimize the parameters listed above.

Ligand Density Effect. To optimize the ligand density parameter, a range of media that vary in ligand density over several orders of magnitude will be synthesized (fmol/ml to μmol/ml). Standard density media of 1-15mM will be used for Stage 1, since all available ligand sites will be saturated with the artificial recombinant proteome. Each protein within the tested artificial proteome will be represented equally to minimize the effect of enzyme copy upon drug molecules. A fixed volume of proteome-saturated media is applied to a fixed volume of second stage media of varying density. Constant flow is applied and the inhibitor standards added at fixed concentrations 10-fold the reported Ki_app for each of the respective protein targets. Sufficient drug volume is added to saturate the entire fist and second stage media. Elution of the various proteins will be monitored by absorbance and by electrophoresis. The peak shape and retention time for a fixed ligand density are recorded. Potent inhibitors show short retention times with narrow very asymmetric peaks. Moderate inhibitors demonstrate slower retention times with peak broadening, whereas weak inl ibitors will yield long retention times and very broad peaks. Using an inhibitor concentration of drug 10 times above the reported Ki_app will eliminate any general partitioning effect that may the result through the collective influence of multiple very low affinity purine binding sites within the artificial proteome. Typically concentrations of inhibitor will be 100 to 200 μM. Optimal elution times for each inliibitor protein pair will be determined for a fixed inhibitor concentration. Ligand densities exhibiting optimal peak shapes and retention times for the most potently inhibited targets will be developed further. Ligand densities that give optimal resolution of protein targets at fixed column length are of special interest. Gleevec and Iressa have been reported to exhibit polypharmacology and target various RTK family members with differing potencies (e.g. Gleevec inhibits several RTKs including Kit and PDGF in addition to BCRAbl, Iressa targets various members of the Her2 family including EGFr, VEGF and Her2). Accordingly, these compounds will be contrasted in ADAT analysis with methotrexate, BIRB796 and radicicol which are highly selective for dihydrofolate reductase, p38mapk and HSP90. SDS-PAGE of column fractions and quantitation by sypro orange staining will be used to determine retention times with compounds that are anticipated to elute mixtures of proteins. Peptide sequencing by mass spectrometry will be used to identify each protein in the gels. Column Length. After the optimal ligand density is identified, column length is optimized using with the artificial proteome at equimolar protein concentrations at fixed temperature and flow. Performance will be evaluated as described for optimal ligand density. Flow Rate. Flow rate is then optimized to minimize diffusional effects while retaining protein resolution in cases where a particular compound targets more than one protein in the artificial recombinant protein with different affinities. Temperature. Various temperatures are tested using optimal conditions of density, length and flow determined above. Demonstrating ADAT Using Compounds With a Known Ki To utilize ADAT to determine the Ki for each ofthe drugs of Example 1, first stage media is saturated with the artificial proteome. The saturated proteome is applied to the second stage media using optimal conditions of ligand density, length, flow and temperature established above. Compounds for which the Ki has been determined by enzymatic (solution phase) assay are then introduced into the flow at fixed concentrations. Each protein target of interest is examined individually during this analysis. For example for p38mapk, the first stage is charged with the artificial proteome, layered onto the second stage, and the ADAT column eluted with InM ofthe drug. The retention time ofthe eluted protein is determined. First stage resin is then replaced with freshly charged proteome and the next concentration applied. The experiment will be repeated with increasing concentrations until all ofthe recombinant p38mapk is eluted. After each elution at each concentration, the area under the curve will be determined and used to measure the amount of protein that is released at each concentration tested. This protein concentration is plotted against compound concentration to determine the actual concentration required to elute 50% ofthe bound protein (Ki). This value is compared with Ki values determined in enzymatic (solution phase) assays to confirm that Ki values obtained in ADAT analysis are within a factor of two of enzymatic assay. This process is repeated for other compounds that target other members ofthe artificial proteome. Beyond confirming that actual Ki is predicted from retention time alone using the ADAT approach irrespective of concentration, these analyses facilitate the construction of a standard retention-time elution table that can be used to accurately predict the Ki of any compound - protein combination, at any compound concentration tested (provided the protein is detectable). Many modifications and other embodiments ofthe inventions set forth herein will come to mind to one skilled in the art to which these inventions pertain having the benefit ofthe teachings presented in the foregoing descriptions and the associated drawings. Therefore, it is to be understood that the inventions are not to be limited to ■ the specific embodiments disclosed and that modifications and other embodiments are intended to be included within the scope ofthe appended claims. Although specific terms are employed herein, they are used in a generic and descriptive sense only and not for purposes of limitation. All publications and patent applications mentioned in the specification are indicative ofthe level of those skilled in the art to which this invention pertains. All publications and patent applications are herein incorporated by reference to the same extent as if each individual publication or patent application was specifically and individually indicated to be incorporated by reference. Those skilled in the art will recognize, or be able to ascertain using no more than routine experimentation, many equivalents to the specific embodiments ofthe invention described herein. Such equivalents are intended to be encompassed by the following claims.

Claims

WHAT IS CLAIMED:

1. A method for determining the affinity of a compound of interest for a protein comprising the steps of: (a) contacting a ligand with a protein under conditions in which the protein can form a complex with the ligand; (b) contacting the protein and the ligand with a compound of interest under conditions in which a complex ofthe protein and the ligand and a complex ofthe protein and the compound of interest can form; (c) contacting the complex ofthe protein and the compound of interest with an excess ofthe ligand under conditions in which a complex ofthe protein and the ligand and a complex ofthe protein and the compound of interest can form; (d) measuring the avidity effect ofthe ligand upon the formation of a complex ofthe protein and the compound of interest; and (e) determining the affinity of the compound of interest for the protein from the measurement.

2. The method of claim 1 wherein the ligand is linked to a solid support.

3. A method for determining the affinity of a compound of interest for at least one member of a plurality of proteins comprising the steps of: (a) contacting an amount of a ligand with a plurality of proteins under conditions in which at least one member ofthe plurality of proteins can form a complex with the ligand; (b) contacting the plurality of proteins and the amount ofthe ligand with an amount of a compound of interest under conditions in which a complex of at least one member ofthe plurality of proteins and the ligand and a complex of at least one member ofthe plurality of proteins and the compound of interest can form; (c) contacting the complex of at least one member of the plurality of proteins and the compound of interest with an excess ofthe ligand under conditions in which a complex of at least one member ofthe plurality of proteins and the ligand and a complex of at least one member ofthe plurality of proteins and the compound of interest can form; (d) measuring the avidity effect ofthe ligand upon the formation of a complex of at least one member ofthe plurality of proteins and the compound of interest; and (e) determining the affinity of the compound of interest for at least one member ofthe plurality of proteins from the measurement.

4. The method of claim 3 wherein the ligand is linked to a solid support.

5. A method for determining the affinity of a compound of interest for a protein comprising the steps of: (a) contacting a stationary ligand with a protein under conditions in which the protein can form a complex with the stationary ligand; (b) contacting the protein and the stationary ligand with a compound of interest under conditions in which a complex ofthe protein and the stationary ligand and a complex ofthe protein and the compound of interest can form; (c) contacting the complex ofthe protein and the compound of interest with an excess ofthe stationary ligand under conditions in which a complex ofthe protein and the stationary ligand and a complex ofthe protein and the compound of interest can form; (d) measuring the avidity effect ofthe stationary ligand upon the formation of a complex ofthe protein and the compound of interest; and (e) determining the affinity ofthe compound of interest for the protein from the measurement.

6. A method for determining the affinity of a compound of interest for at least one member of a plurality of proteins comprising the steps of: (a) contacting an amount of a stationary ligand with a plurality of proteins under conditions in which at least one member ofthe plurality of proteins can form a complex with the stationary ligand; (b) contacting the plurality of proteins and the amount ofthe stationary ligand with an amount of a compound of interest under conditions in which a complex of at least one member ofthe plurality of proteins and the stationary ligand and a complex of at least one member ofthe plurality of proteins and the compound of interest can form; (c) contacting the complex of at least one member ofthe plurality of proteins and the compound of interest with an excess ofthe stationary ligand under conditions in which a complex of at least one member ofthe plurality of proteins and the stationary ligand and a complex of at least one member ofthe plurality of proteins and the compound of interest can form; (d) measuring the avidity effect ofthe stationary ligand upon the formation of a complex of at least one member ofthe plurality of proteins and the compound of interest; and (e) determining the affinity ofthe compound of interest for at least one member of the plurality of proteins from the measurement.

7. A method for determining the affinity of a compound of interest for a protein comprising the steps of: (a) contacting a first stage media comprising a stationary ligand with a protein under conditions in which a complex ofthe protein and the stationary ligand can form; (b) contacting the protein and the first stage media with a compound of interest under conditions in which a complex ofthe protein and the stationary ligand and a complex ofthe protein and the compound of interest can form; (c) moving a complex of the protein and the compound of interest through a second stage media under conditions in which a complex ofthe protein and the stationary ligand and a complex ofthe protein and the compound of interest can form, wherein the second stage media comprises the stationary ligand; (d) measuring the avidity effect ofthe stationary ligand upon the formation of a complex of the protein and the compound of interest; and (e) determining the affinity ofthe compound of interest for the protein from the measurement.

8. The method of claim 7 wherein the avidity effect is manipulated by at least one technique selected from the group consisting of: (a) increasing the ligand density ofthe second stage media; and (b) decreasing the ligand density ofthe second stage media.

9. The method of claim 7 wherein the avidity effect is manipulated by at least one technique selected from the group consisting of: (a) increasing the rate at which the complex ofthe protein and the compound of interest moves through the second stage media; and (b) decreasing the rate at which the complex ofthe protein and the compound of interest moves through the second stage media.

10. The method of claim 7 wherein the avidity effect is manipulated by at least one technique selected from the group consisting of: (a) increasing the temperature ofthe second stage media; and (b) decreasing the temperature ofthe second stage media.

11. The method of claim 7 wherein steps (c)-(d) are carried out by a chromatographic technique.

12. The method of claim 11 wherein the chromatographic technique is column chromatography and the avidity effect is manipulated by at least one method selected from the group consisting of: (a) increasing the length ofthe chromatography column; and (b) decreasing the length ofthe chromatography column.

13. A method for determining the affinity of a compound of interest for at least one member of a plurality of proteins comprising the steps of: (a) contacting a first stage media comprising a stationary ligand with a plurality of proteins under conditions in which a complex of at least one member ofthe plurality of proteins and the stationary ligand can form; (b) contacting the plurality of proteins and the first stage media with a compound of interest under conditions in which a complex of at least one member ofthe plurality of proteins and the stationary ligand and a complex of at least one member ofthe plurality of proteins and the compound of interest can form; (c) moving a complex of at least one member of the plurality of proteins and the compound of interest through a second stage media under conditions in which a complex of at least one member ofthe plurality of proteins and the stationary ligand and a complex of at least one member ofthe plurality of proteins and the compound of interest can form, wherein the second stage media comprises the stationary ligand; (d) measuring the avidity effect ofthe stationary ligand upon the formation of a complex of at least one member ofthe plurality of proteins and the compound of interest; and (e) determining the affinity ofthe compound of interest for at least one member of the plurality of proteins from the measurement.

14. The method of claim 13 wherein the avidity effect is manipulated by at least one technique selected from the group consisting of: (a) increasing the ligand density ofthe second stage media; and (b) decreasing the ligand density ofthe second stage media.

15. The method of claim 13 wherein the avidity effect is manipulated by at least one technique selected from the group consisting of: (a) increasing the rate at which the complex of at least one member ofthe plurality of proteins and the compound of interest moves through the second stage media; and (b) decreasing the rate at which the complex of at least one member ofthe plurality of proteins and the compound of interest moves through the second stage media.

16. The method of claim 13 wherein the avidity effect is manipulated by at least one technique selected from the group consisting of: (a) increasing the temperature ofthe second stage media; and (b) decreasing the temperature ofthe second stage media.

17. The method of claim 13 wherein steps (c)-(d) are carried out by a chromatographic technique.

18. The method of claim 17 wherein the chromatographic technique is column chromatography and the avidity effect is manipulated by at least one method selected from the group consisting of: (a) increasing the length ofthe chromatography column; and (b) decreasing the length ofthe chromatography column.

19. The method of claim 7 wherein the stationary ligand exhibits solution phase kinetics.

20. The method of claim 13 wherein the stationary ligand exhibits solution phase kinetics.

21. The method of claim 7 wherein a standard curve is generated using a series of compounds of known Kd.

22. The method of claim 13 wherein a standard curve is generated using a series of compounds of known Kd.

23. The method of claim 7 wherein the ligand density ofthe second stage media is between about InM and 500 mM.

24. The method of claim 13 wherein the ligand density ofthe second stage media is between about InM and 500 mM.

25. The method of claim 7 wherein the ligand density ofthe second stage media is between about lμM and 15 mM.

26. The method of claim 13 wherein the ligand density ofthe second stage media is between about lμM and 15 mM.

27. The method of claim 7 wherein the ligand density of the first stage media is between about 100 μM and 500 mM.

28. The method of claim 13 wherein the ligand density ofthe first stage media is between about 100 μM and 500 mM.

29. The method of claim 12 wherein the avidity effect is measured by peak shape and retention time.

30. The method of claim 18 wherein the avidity effect is measured by peak shape and retention time.

31. The method of claim 7 wherein a saturating amount of protein is added to the first stage media.

32. The method of claim 13 wherein saturating amounts ofthe plurality of proteins are added to the first stage media.

33. The method of claim 31 wherein a fixed volume of protein-saturated first stage media is applied to a fixed volume of second stage media.

34. The method of claim 32 wherein a fixed volume of protein-saturated first stage media is applied to a fixed volume of second stage media.

35. The method of claim 33 wherein a constant flow is applied and the compound of interest is added at fixed concentration.

36. The method of claim 34 wherein a constant flow is applied and the compound of interest is added at fixed concentration.

37. The method of claim 35 wherein the concentration ofthe compound of interest is at least ten times greater than the reported dissociation constant for the compound of interest.

38. The method of claim 36 wherein the concentration of the compound of interest is at least ten times greater than the reported dissociation constant for the compound of interest.

39. The method of claim 35 wherein the concentration ofthe compound of interest is between about 1 nM to 10 mM.

40. The method of claim 36 wherein the concentration of the compound of interest is between about 1 nM to 10 mM.

41. The method of claim 35 wherein the elution ofthe protein is monitored by electrophoresis.

42. The method of claim 36 wherein the elution of at least one member of the plurality of proteins is monitored by electrophoresis.

43. The method of claim 35 wherein the identity ofthe eluted protein is determined by mass spectrometry.

44. The method of claim 36 wherein the identity of at least one eluted protein is determined by mass spectrometry.

45. The method of claim 7 wherein the second stage media is a gradient density media.

46. The method of claim 13 wherein the second stage media is a gradient density media.

47. The method of claim 11 wherein the second stage media is a low- density media and the column is a longer column.

48. The method of claim 17 wherein the second stage media is a low- density media and the column is a longer column.

49. A method for determining the affinity of a compound of interest for a protein comprising the steps of: (a) measuring the retention time of a protein exposed to a stationary ligand and a mobile compound of interest and (b) determining the affinity ofthe compound of interest for the protein from the retention time.

50. A method for determining the affinity of a compound of interest for a protein comprising the steps of: (a) contacting a ligand with a protein under conditions in which the protein can form a complex with the ligand; (b) washing the protein and ligand with a buffered solution; (c) contacting the protein and the ligand with a compound of interest under conditions in which a complex ofthe protein and the ligand and a complex ofthe protein and the compound of interest can form; (d) contacting the complex ofthe protein and the compound of interest with an excess ofthe ligand under conditions in which a complex ofthe protein and the ligand and a complex ofthe protein and the compound of interest can form; (e) measuring the avidity effect ofthe ligand upon the formation of a complex ofthe protein and the compound of interest; and (f) determining the affinity of the compound of interest for the protein from the measurement.

51. A method for determining the affinity of a compound of interest for at least one member of a plurality of proteins comprising the steps of: (a) contacting an amount of a ligand with a plurality of proteins under conditions in which at least one member ofthe plurality of proteins can form a complex with the ligand; (b) washing the protein and ligand with a buffered solution; (c) contacting the plurality of proteins and the amount ofthe ligand with an amount of a compound of interest under conditions in which a complex of at least one member ofthe plurality of proteins and the ligand and a complex of at least one member ofthe plurality of proteins and the compound of interest can form; (d) contacting the complex of at least one member ofthe plurality of proteins and the compound of interest with an excess ofthe ligand under conditions in which a complex of at least one member ofthe plurality of proteins and the ligand and a complex of at least one member ofthe plurality of proteins and the compound of interest can fonn; (e) measuring the avidity effect ofthe ligand upon the formation of a complex of at least one member ofthe plurality of proteins and the compound of interest; and (f) determining the affinity ofthe compound of interest for at least one member ofthe plurality of proteins from the measurement.

52. A method for determining the likelihood that a compound of interest will undergo a partitioning effect comprising the steps of: (a) contacting a first stage media comprising a stationary ligand with a plurality of proteins under conditions in which a complex of at least one member ofthe plurality of proteins and the stationary ligand can form; (b) contacting the plurality of proteins and the first stage media with a compound of interest under conditions in which a complex of at least one member ofthe plurality of proteins and the stationary ligand and a complex of at least one member ofthe plurality of proteins and the compound of interest can form; (c) moving a complex of at least one member of the plurality of proteins and the compound of interest through a second stage media under conditions in which a complex of at least one member ofthe plurality of proteins and the stationary ligand and a complex of at least one member ofthe plurality of proteins and the compound of interest can form, wherein the second stage media comprises the stationary ligand; (d) identifying at least one member ofthe plurality of proteins that forms a complex with the compound of interest; (e) measuring the avidity effect ofthe second stage media upon the formation of the complex between each identified member of the plurality of proteins and the compound of interest; (f) determining the affinity ofthe compound of interest for each identified member ofthe plurality of proteins from the measurement; and (g) determining the copy number of each identified member of the plurality of proteins wherein the identification of at least one member ofthe plurality of proteins that possesses an affinity for the compound of interest and that is present in high copy number is indicative of a likelihood that the compound of interest will undergo a partitioning effect.