Microplate thermal shift assay apparatus for ligand development and multi-variable protein chemistry optimization

[0002] This application is a continuation of application Ser. No. 08/853,459, filed May 9, 1997, now U.S. Pat. No. 6,036,920, which claimed benefit of U.S. provisional application Ser. No. 60/017,860 May 9, 1996.

[0004] 1. Field of the Invention

[0005] The present invention relates generally to the screening of compound and combinatorial libraries. More particularly, the present invention relates to a method and apparatus for performing assays, particularly thermal shift assays.

[0006] 2. Related Art

[0007] In recent years, pharmaceutical researchers have turned to combinatorial libraries as sources of new lead compounds for drug discovery. A combinatorial library is a collection of chemical compounds which have been generated, by either chemical synthesis or biological synthesis, by combining a number of chemical “building blocks” as reagents. For example, a combinatorial polypeptide library is formed by combining a set of amino acids in every possible way for a given compound length (i.e., the number of amino acids in a polypeptide compound). Millions of chemical compounds can theoretically be synthesized through such combinatorial mixing of chemical building blocks. Indeed, one investigator has observed that the systematic, combinatorial mixing synthesis of 100 million tetrameric compounds or 10 billion pentameric compounds (Gordon, E. M. et al., J. Med. Chem.37:1233-1251 (1994)).

[0008] The rate of combinatorial library synthesis is accelerated by automating compound synthesis and evaluation. For example, DirectedDiversity® is a computer based, iterative process for generating chemical entities with defined physical, chemical and/or bioactive properties. The DirectedDiversity® system is disclosed in U.S. Pat. No. 5,463,564, which is herein incorporated by reference in its entirety.

[0009] Once a library has been constructed, it must be screened to identify compounds which possess some kind of biological or pharmacological activity. To screen a library of compounds, each compound in the library is equilibrated with a target molecule of interest, such as an enzyme. A variety of approaches have been used to screen combinatorial libraries for lead compounds. For example, in an encoded library, each compound in a chemical combinatorial library can be made so that an oligonucleotide “tag” is linked to it. A careful record is kept of the nucleic acid tag sequence for each compound. A compound which exerts an effect on the target enzyme is selected by amplifying its nucleic acid tag using the polymerase chain reaction (PCR). From the sequence of the tag, one can identify the compound (Brenner, S. et al., Proc. Natl. Acad. Sci. USA89:5381-5383 (1992)). This approach, however, is very time consuming because it requires multiple rounds of oligonucleotide tag amplification and subsequent electrophoresis of the amplification products.

[0010] A filamentous phage display peptide library can be screened for binding to a biotinylated antibody, receptor or other binding protein. The bound phage is used to infect bacterial cells and the displayed determinant (i.e., the peptide ligand) is then identified (Scott, J. K. et al., Science249:386-390 (1990)). This approach suffers from several drawbacks. It is time consuming. Peptides which are toxic to the phage or to the bacterium cannot be studied. Moreover, the researcher is limited to investigating peptide compounds.

[0011] In International Patent Application WO 94/05394 (1994), Hudson, D. et al., disclose a method and apparatus for synthesizing and screening a combinatorial library of biopolymers on a solid-phase plate, in an array of 4×4 to 400×400. The library can be screened using a fluorescently labeled, radiolabeled, or enzyme-linked target molecule or receptor. The drawback to this approach is that the target molecule must be labeled before it can be used to screen the library.

[0012] A challenge presented by currently available combinatorial library screening technologies is that they provide no information about the relative binding affinities of different ligands for a receptor protein. This is true whether the process for generating a combinatorial library involves phage library display of peptides (Scott, J. K. et al., Science249:386-390 (1990)), random synthetic peptide arrays (Lam, K. S. et al., Nature354:82-84 (1991)), encoded chemical libraries (Brenner, S. et al., Proc. Natl. Acad. Sci. USA89:5381-5383 (1992)), the method of Hudson (Intl. Appl. WO 94/05394), or most recently, combinatorial organic synthesis (Gordon, E. et al., J. Med. Chem.37:1385-1399 (1994)).

[0013] To acquire quantitative binding data from the high throughput screening of ligand affinities for a target enzyme, researchers have relied on assays of enzyme activity. Enzymes lend themselves to high throughput screening because the effect of ligand binding can be monitored using kinetic assays. The experimental endpoint is usually a spectrophotometric change. Using a kinetic assay, most researchers use a two-step approach to lead compound discovery. First, a large library of compounds is screened against the target enzyme to determine if any of the library compounds are active. These assays are usually performed in a single concentration (between 10⁻⁴-10⁻⁶) with one to three replicates. Second, promising compounds obtained from the first screen (i.e., compounds which display activity greater than a predetermined value) are usually re-tested to determine a 50% inhibitory concentration (IC₅₀), an inhibitor association constant (K_i), or a dissociation constant (K_d). This two-step approach, however, is very labor intensive, time-consuming and prone to error. Each re-tested sample must either be retrieved from the original assay plate or weighed out and solubilized again. A concentration curve must then be created for each sample and a separate set of assay plates must be created for each assay.

[0014] There are other problems associated with the biochemical approach to high throughput screening of combinatorial libraries. Typically, a given assay is not applicable to more than one receptor. That is, when a new receptor becomes available for testing, a new assay must be developed. For many receptors, reliable assays are simply not available. Even if an assay does exist, it may not lend itself to automation. Further, if a K_iis the endpoint to be measured in a kinetic assay, one must first guess at the concentration of inhibitor to use, perform the assay, and then perform additional assays using at least six different concentrations of inhibitor. If one guesses too low, an inhibitor will not exert its inhibitory effect at the suboptimal concentration tested.

[0015] In addition to the drawbacks to the kinetic screening approach described above, it is difficult to use the kinetic approach to identify and rank ligands that bind outside of the active site of the enzyme. Since ligands that bind outside of the active site do not prevent binding of spectrophotometric substrates, there is no spectrophotometric change to be monitored. An even more serious drawback to the kinetic screening approach is that non-enzyme receptors cannot be assayed at all.

[0016] Thermal protein unfolding, or thermal “shift,” assays have been used to determine whether a given ligand binds to a target receptor protein. In a physical thermal shift assay, a change in a biophysical parameter of a protein is monitored as a function of increasing temperature. For example, in calorimetric studies, the physical parameter measured is the change in heat capacity as a protein undergoes temperature induced unfolding transitions. Differential scanning calorimetry has been used to measure the affinity of a panel of azobenzene ligands for streptavidin (Weber, P. et al., J. Am. Chem. Soc.16:2717-2724 (1994)). Titration calorimetry has been used to determine the binding constant of a ligand for a target protein (Brandts, J. et al., American Laboratory22:30-41 (1990)). The calorimetric approach, however, requires that the researcher have access to a calorimetric device. In addition, calorimetric technologies do not lend themselves to the high throughput screening of combinatorial libraries, **three thermal scans per day are routine.

[0017] Like calorimetric technologies, spectral technologies have been used to monitor temperature induced protein unfolding (Bouvier, M. et al., Science265:398-402 (1994); Chavan, A. J. et al., Biochemistry33:7193-7202 (1994); Morton, A. et al., Biochemistry1995:8564-8575 (1995)). The single sample heating and assay configuration, as conventionally performed, has impeded the application of thermal shift technologies to high throughput screening of combinatorial libraries. Thus, there is a need for a thermal shift technology which can be used to screen combinatorial libraries, can be used to identify and rank lead compounds, and is applicable to all receptor proteins.

[0018] Thermal shift assays have been used to determine whether a ligand binds to DNA. Calorimetric, absorbance, circular dichroism, and fluorescence technologies have been used (Pilch, D. S. et al., Proc. Natl. Acad. Sci. U.S.A.91:9332-9336 (1994); Lee, M. et al., J. Med. Chem.36:863-870 (1993); Butour, J.-L. et al., Eur. J Biochem.202:975-980 (1991); Barcelo, F. et al., Chem. Biol. Interactions74:315-324 (1990)). As used conventionally, however, these technologies have impeded the high throughput screening of nucleic acid receptors for lead compounds which bind with high affinity. Thus, there is a need for a thermal shift technology which can be used to identify and rank the affinities of lead compounds which bind to DNA sequences of interest.

[0019] When bacterial cells are used to overexpress exogenous proteins, the recombinant protein is often sequestered in bacterial cell inclusion bodies. For the recombinant protein to be useful, it must be purified from the inclusion bodies. During the purification process, the recombinant protein is denatured and must then be renatured. It is impossible to predict the renaturation conditions that will facilitate and optimize proper refolding of a given recombinant protein. Usually, a number of renaturing conditions must be tried before a satisfactory set of conditions is discovered. In a study by Tachibana et al., each of four disulfide bonds were singly removed, by site-directed mutagenesis, from hen lysozyme (Tachibana et al., Biochemistry33:15008-15016 (1994)). The mutant genes were expressed in bacterial cells and the recombinant proteins were isolated from inclusion bodies. Each of the isolated proteins were renatured under different temperatures and glycerol concentrations. The efficacy of protein refolding was assessed in a bacteriolytic assay in which bacteriolytic activity was measured as a function of renaturing temperature. The thermal stability of each protein was studied using a physical thermal shift assay. In this study, however, only one sample reaction was heated and assayed at a time. The single sample heating and assay configuration prevents the application of thermal shift technologies to high throughput screening of a multiplicity of protein refolding conditions. Thus, there is a need for a thermal shift technology which can be used to rank the efficacies of various protein refolding conditions.

[0020] Over the past four decades, X-ray crystallography and the resulting atomic models of proteins and nucleic acids have contributed greatly to an understanding of structural, molecular, and chemical aspects of biological phenomena. However, crystallographic analysis remains difficult because there are not straightforward methodologies for obtaining X-ray quality protein crystals. Conventional methods cannot be used quickly to identify crystallization conditions that have highest probability of promoting crystallization (Garavito, R. M. et al., J. Bioenergtics and Biomembranes28:13-27 (1996)). Even the use of factorial design experiments and successive automated grid searches (Cox, M. J., & Weber, P. C., J. Appl. Cryst.20:366-373 (1987); Cox, M. J., & Weber, P. C., J. Crystal Growth90:318-324 (1988)) do not facilitate rapid, high throughput screening of biochemical conditions that promote the crystallization of X-ray quality protein crystals. Moreover, different proteins are expected to require different conditions for protein crystallization, just as has been the experience for their folding (McPherson, A., In: Preparation and Analysis of Protein Crystals,Wiley Interscience, New York, (1982)). Conventional methods of determining crystallization conditions are cumbersome, slow, and labor intensive. Thus, there is a need for a rapid, high throughput technology which can be used to rank the efficacies of protein crystallization conditions.

[0021] Rapid, high throughput screening of combinatorial molecules or biochemical conditions that stabilize target proteins in thermal shift assays would be facilitated by the simultaneous heating of many samples. To date, however, thermal shift assays have not been performed that way. Instead, the conventional approach to performing thermal shift assays has been to heat and assay only one sample at a time. That is, researchers conventionally 1) heat a sample to a desired temperature in a heating apparatus; 2) assay a physical change, such as absorption of light or change in secondary, tertiary, or quaternary protein structure; 3) heat the samples to the next highest desired temperature; 4) assay for a physical change; and 5) continue this process repeatedly until the sample has been assayed at the highest desired temperature.

[0022] This conventional approach is disadvantageous for at least two reasons. First, this approach is labor intensive. Second, this approach limits the speed with which thermal shift screening assays can be performed and thereby precludes rapid, high-throughput screening of combinatorial molecules binding to a target receptor and biochemical conditions that stabilize target proteins. Thus, there is a need for an apparatus capable of performing rapid, high-throughput thermal shift assays that will be suitable for all receptors, including reversibly folding proteins.

[0023] The present invention provides a multi-variable method for ranking the efficacy of one or more of a multiplicity of different molecules or different biochemical conditions for stabilizing a target molecule which is capable of denaturing due to a thermal change. The method comprises contacting the target molecule with one or more of a multiplicity of different molecules or different biochemical conditions in each of a multiplicity of containers, simultaneously heating the multiplicity of containers, measuring in each of the containers a physical change associated with the thermal denaturation of the target molecule resulting from heating, generating a thermal denaturation curve for the target molecule as a function of temperature for each of the containers, comparing each of the denaturation curves to (i) each of the other thermal denaturation curves and to (ii) the thermal denaturation curve obtained for the target molecule under a reference set of biochemical conditions, and ranking the efficacies of multiplicity of different molecules or the different biochemical conditions according to the change in each of the thermal denaturation curves.

[0024] The present invention provides a multi-variable method for optimizing the shelf life of a target molecule which is capable of denaturing due to a thermal change. The method comprises contacting the target molecule with one or more of a multiplicity of different molecules or different biochemical conditions in each of a multiplicity of containers, simultaneously heating the multiplicity of containers, measuring in each of the containers a physical change associated with the thermal denaturation of the target molecule resulting from heating, generating a thermal denaturation curve for the target molecule as a function of temperature for each of the containers, comparing each of the denaturation curves to (i) each of the other thermal denaturation curves and to (ii) the thermal denaturation curve obtained for the target under a reference set of biochemical conditions, and ranking the efficacies of multiplicity of different molecules or the different biochemical conditions according to the change in each of the thermal denaturation curves.

[0025] The present invention also provides a multi-variable method for ranking the affinity of a combination of two or more of a multiplicity of different molecules for a target molecule which is capable of denaturing due to a thermal change. The method comprises contacting the target molecule with a combination of two or more different molecules of the multiplicity of different molecules in each of a multiplicity of containers, simultaneously heating the multiplicity of containers, measuring in each of the containers a physical change associated with the thermal denaturation of the target molecule resulting from the heating, generating a thermal denaturation curve for the target molecule as a function of temperature for each of the containers, comparing each of the thermal denaturation curves with (i) each of the other thermal denaturation curves obtained for the target molecule and to (ii) the thermal denaturation curve for the target molecule in the absence of any of the two or more different molecules, and ranking the affinities of the combinations of the two or more multiplicity of different molecules according to the change in each of the thermal denaturation curves.

[0026] The present invention also provides a multi-variable method for ranking the efficacies of one or more of a multiplicity of different biochemical conditions to facilitate the refolding of a sample of a denatured protein. The method comprises placing one of the refolded protein samples in each of a multiplicity of containers, wherein each of the refolded protein samples has been previously refolded according to one or more of the multiplicity of conditions, simultaneously heating the multiplicity of containers, measuring in each of the containers a physical change associated with the thermal denaturation of the protein resulting from heating, generating a thermal denaturation curve for the protein as a function of temperature for each of the containers, comparing each of the denaturation curves to (i) each of the other thermal denaturation curves and to (ii) the thermal denaturation curve obtained for the native protein under a reference set of biochemical conditions, and ranking the efficacies of the multiplicity of different refolding conditions according to the change in the magnitude of the physical change of each of the thermal denaturation curves.

[0027] The present invention also provides a further multi-variable method for ranking the efficacies of one or more of a multiplicity of different biochemical conditions to facilitate the refolding of a sample of a denatured protein, which comprises determining one or more combinations of a multiplicity of different conditions which promote protein stabililty, folding the denatured protein under said one or more combinations of biochemical conditions that were identified as promoting protein stabilization, asseessing folded protein yield, ranking the efficacies of said multiplicity of different refolding conditions according to folded protein yield, and repeating these steps until a combination of biochemical conditions that promote optimal protein folding are identified.

[0028] Using the microplate thermal shift assay, one can determine one or more biochemical conditions have an additive effect on protein stability. Once a set of biochemical conditions that facilitate an increase in protein stability have been identified using the thermal shift assay, the same set of conditions can be used in protein folding experiments with recombinant protein. If the conditions that promote protein stability in the thermal shift assay correlate with conditions that promote folding of recombinant protein, conditions can be further optimized by performing additional thermal shift assays until a combination of stabilizing conditions that result in further increase protein stability are identified. Recombinant protein is then folded under those conditions. This process is repeated until optimal folding conditions are identified.

[0029] The present invention also provides a multi-variable method for ranking the efficacy of one or more of a multiplicity of different biochemical conditions for facilitating the crystallization of a protein which is capable of denaturing due to a thermal change. The method comprises contacting the protein with one or more of the multiplicity of different biochemical conditions in each of a multiplicity of containers, simultaneously heating the multiplicity of containers, measuring in each of the containers a physical change associated with the thermal denaturation of the protein resulting from the heating, generating a thermal denaturation curve for the protein as a function of temperature for each of the containers, comparing each of the denaturation curves to (i) each of the other thermal denaturation curves and (ii) to the thermal denaturation curve obtained using a reference set of biochemical conditions, and ranking the efficacies of the multiplicity of different biochemical conditions according to the change in each of the thermal denaturation curves.

[0030] The present invention also provides a method for ranking the affinity of each of a multiplicity of different molecules for a target molecule which is capable of denaturing due to a thermal change. The method comprises contacting the target molecule with one molecule of a multiplicity of different molecules in each of a multiplicity of containers, simultaneously heating the containers, measuring in each of the containers a physical change associated with the thermal denaturation of the target molecule resulting from heating, generating a thermal denaturation curve for the target molecule as a function of temperature in each of the containers, comparing each of the thermal denaturation curves with the thermal denaturation curve obtained for the target molecule in the absence of any of the molecules in the multiplicity of different molecules, and ranking the affinities of each molecule according to the change in each of the thermal denaturation curves.

[0031] The present invention also provides a method for assaying a pool or collection of a multiplicity of different molecules for a molecule which binds to a target molecule which is capable of denaturing due to a thermal change. The method comprises contacting the target molecule with a collection of at least two molecules of a multiplicity of different molecules in each of a multiplicity of containers, simultaneously heating the multiplicity of containers, measuring in each of the containers a physical change associated with the thermal denaturation of the target molecule resulting from heating, generating a set of thermal denaturation curves for the target molecule as a function of temperature for each of the containers, comparing each of the thermal denaturation curves with the thermal denaturation curve obtained for the target molecule in the absence of any of the molecules in the multiplicity of different molecules, ranking the affinities of the collections of different molecules according to the change in each of the thermal denaturation curves, selecting the collection of different molecules which contains a molecule with affinity for the target molecule, dividing the selected collection into smaller collections of molecules in each of a multiplicity of containers, and repeating the above steps until a single molecule responsible for the original thermal shift in the multiplicity of molecules is identified.

[0032] This invention also provides an improved method for generating lead compounds which comprises synthesizing a multiplicity of compounds and testing the ability of each compound to bind to a receptor molecule. The improvement comprises contacting the receptor molecule with one compound of a multiplicity of different compounds in each of a multiplicity of wells in a microplate, simultaneously heating the wells, measuring in each of the wells a physical change, resulting from heating, associated with the thermal denaturation of the receptor molecule, generating a thermal denaturation curve for the receptor molecule as a function of temperature in each of the wells, comparing each of the thermal denaturation curves with the thermal denaturation curve obtained for the receptor molecule in the absence of any of the compounds in the multiplicity of different compounds, and ranking the affinities of each compound according to the change in each of the thermal denaturation curves.

[0033] The present invention also provides a product of manufacture which comprises a carrier having a multiplicity of containers therein, each of the containers containing a target molecule which is capable of denaturation due to heating, and at least one molecule selected from a multiplicity of different molecules present in a combinatorial library, wherein each of the different molecules are present in a different one of the multiplicity of containers in the carrier.

[0034] Optimization of protein stability, ligand binding, protein folding, and protein crystallization are multi-variable events. Multi-variable optimization problems require large numbers of parallel experiments to collect as much data as possible in order to determine which variables influence a favorable response. For example, multi-variable optimization problems require large numbers of parallel experiments to collect as much data as possible in order to determine which variables influence protein stabililty. In this regard, both protein crystallization and quantitative structure activity relationship analyses have greatly benefited from mass screening protocols that employ matrix arrays of incremental changes in biochemical or chemical composition. Thus, in much the same way that quantitative structure activity relationships are constructed to relate variations of chemical functional groups on ligands to their effect on binding affinity to a given therapeutic receptor, the methods and apparatus of the present invention facilitate the construction of a quantitative model that relates different biochemical conditions to experimentally measured protein stability, ligand specificity, folded protein yield, and crystallized protein yield.

[0035] The present invention offers a number of advantages over previous technologies that are employed to optimize multi-variable events such as protein stabilization, ligand binding, protein folding, and protein crystallization. Foremost among these advantages is that the present invention facilitates high throughput screening. Further, the present invention offers a number of advantages over previous technologies that are employed to screen combinatorial libraries. Foremost among these advantages is that the present invention facilitates high throughput screening of combinatorial libraries for lead compounds. Many current library screening technologies simply indicate whether a ligand binds to a receptor or not. In that case, no quantitative information is provided. No information about the relative binding affinities of a series of ligands is provided. In contrast, the present invention facilitates the ranking of a series of compounds for their relative affinities for a target receptor. With this information in hand, a structure-activity relationship can be developed for a set of compounds. The ease, reproducibility, and speed of using ligand-dependent changes in midpoint unfolding temperature (T_m) to rank relative binding affinities makes the present invention a powerful tool in the drug discovery process.

[0036] Typically, the conventional kinetic screening approach requires at least six additional well assays at six different concentrations of inhibitor to determine a K_i. Using the present invention, throughput is enhanced ˜6 fold over the enzyme based assays because one complete binding experiment can be performed in each well of a multiwell microplate. The kinetic screening approached are even further limited by the usual compromise between dilution and signal detection, which usually occurs at a protein concentration of about 1 nM. In this regard, the calorimetric approaches, either differential scanning calorimetry or isothermal titrating calorimetry, are at an even worse disadvantage since they are limited to solitary binding experiments, usually 1 per hour. In contrast, the present invention affords a wide dynamic range of measurable binding affinities, from ˜10⁻⁴to 10⁻¹⁵M, in a single well.

[0037] The present invention does not require radioactively labeled compounds. Nor does it require that receptors be labeled with a fluorescent or chromophoric label.

[0038] A very important advantage of the present invention is that it can be applied universally to any receptor that is a drug target. Thus, it is not necessary to invent a new assay every time a new receptor becomes available for testing. When the receptor under study is an enzyme, researchers can determine the rank order of affinity of a series of compounds more quickly and more easily than they can using conventional kinetic methods. In addition, researchers can detect ligand binding to an enzyme, regardless of whether binding occurs at the active site, at an allosteric cofactor binding site, or at a receptor subunit interface. The present invention is equally applicable to non-enzyme receptors, such as proteins and nucleic acids.

[0039] In a further aspect of the present invention, an assay apparatus is provided that includes a heating means for simultaneously heating a plurality of samples, and a receiving means for receiving spectral emission from the samples while the samples are being heated. In yet a further aspect of the present invention, an assay apparatus is provided that includes a temperature adjusting means for simultaneously adjusting a temperature of a plurality of samples in accordance with a pre-determined temperature profile, and a receiving means for receiving spectral emission from the samples while the temperature of the samples is adjusted in accordance with the temperature profile.

[0040] In yet a further aspect, the present invention also provides an assay apparatus that includes a movable platform on which are disposed a plurality of heat conducting blocks. The temperature of the heat conducting blocks, and their samples, are adjusted by a temperature adjusting means. Each of the plurality of heat conducting blocks is adapted to receive a plurality of samples. A light source is provided for emitting an excitatory wavelength of light for the samples. While the temperature of the samples is being adjusted, a sensor detects the spectral emission from the samples in response to the excitatory wavelength of light. The movable platform is moved between heat conducting blocks to sequentially detect spectral emission from the samples in each of the plurality of heat conducting blocks.

[0041] The assay apparatus of the present invention affords the artisan the opportunity to rapidly screen molecules and biochemical conditions that affect protein stability. Samples are simultaneously heated over a range of temperatures. During heating, spectral emissions are received. The assay apparatus of the present invention also provides the artisan with an opportunity for conveniently and efficiently carrying out the methods of the present invention. The assay apparatus of the present invention is particularly adapted for carrying out thermal shift assays of molecules and biochemical conditions that stabilize target proteins.

[0042] Because the apparatus of the present invention comprises both a heating means and a spectral emission receiving means, the apparatus of the present invention obviates the need to heat samples in one apparatus and transfer the samples to another apparatus prior to taking spectral emission readings. As a result, the apparatus of the present invention facilitates changing temperature in accordance with a pre-determined temperature profile, rather than incremental temperature increases and intermediate cooling steps. Thus, more data points can be collected for a given sample and more accurate information can be obtained.

[0043] Further, because the assay apparatus of the present invention comprises both a heating means and a spectral emission receiving means, spectral measurements can be taken from the samples while they are being heated. Thus, using the assay apparatus of the present invention, the artisan can study both irreversibly unfolding proteins and reversibly folding proteins.

[0044] Further features and advantages of the present invention are described in detail below with reference to the accompanying drawings.

[0045] The present invention is described with reference to the accompanying drawings. In the drawings, like reference numbers indicate identical or functionally similar elements. Additionally, the left-most digit(s) of a reference number identifies the drawing in which the reference number first appears.

[0046] FIG. 1 shows the results of a microplate thermal shift assay for ligands which bind to the active site of human α-thrombin (with turbidity as the experimental signal).

[0047] FIG. 2 shows the results of a microplate thermal shift assay for ligands which bind to acidic fibroblast growth factor (aFGF) (with turbidity as the experimental signal).

[0048] FIG. 3 shows the results of a microplate thermal shift assay for ligand binding to the active site of human α-thrombin (with fluorescence emission as the experimental signal). The lines drawn through the data points represent non-linear least squares curve fits of the data using the equation shown at the bottom of the figure. There are five fitting parameters for this equation of y(T) vs. T: (1) y_f, the pre-transitional fluorescence for the native protein; (2) y_u, the post-transitional fluorescence for the unfolded protein; (3) T_m, the temperature at the midpoint for the unfolding transition; (4) ΔH_u, the van't Hoff unfolding enthalpy change; and (5) ΔC_pu, the change in heat capacity upon protein unfolding. The non-linear least squares curve fitting was accomplished using KaleidaGraph™ 3.0 software (Synergy Software, Reading Pa.), which allows the five fitting parameters to float while utilizing Marquardt methods for the minimization of the sum of the squared residuals.

[0049] FIG. 4 shows the result of a microplate thermal shift assay of ligands which bind to the D(II) domain of human FGF receptor 1 (D(II) FGFR1) (with fluorescence emission as the experimental signal). The lines drawn through the data points represent non-linear least squares curve fits of the data using the equation shown at the bottom of the figure, as described for FIG. 3.

[0050] FIG. 5 shows the results of a miniaturized microplate thermal shift assay for Factor D in the absence of any ligands.

[0051] FIG. 6 shows the results of a microplate thermal shift assay for Factor Xa in the absence of any ligands.

[0052] FIG. 7 shows the results of a miniaturized microplate thermal shift assay of a ligand that binds to the catalytic site of human α-thrombin.

[0053] FIGS. 8A and 8B show the results of a miniaturized microplate thermal shift assay of aprosulate binding to the D(II) domain of human FGF receptor 1.

[0054] FIG. 9 shows the results of a miniaturized microplate thermal shift assay for urokinase in the presence of glu-gly-arg chloromethylketone.

[0055] FIG. 10 shows the results of a miniaturized microplate thermal shift assay of human α-thrombin in which the assay volume is 2 μl. Thermal denaturation curves for three experiments are shown.

[0056] FIG. 11 shows the results of a miniaturized microplate thermal shift assay of human α-thrombin in which the assay volume is 5 μl. Thermal denaturation curves for five experiments are shown.

[0057] FIG. 12 shows the results of a single temperature microplate thermal shift assay of human α-thrombin in the presence of four different compounds in four separate experiments.

[0058] FIG. 13 shows the results of a microplate thermal shift assay of the intrinsic tryptophan fluorescence of human α-thrombin. In this assay, blank well fluorescence was not subtracted from sample fluorescence.

[0059] FIG. 14 shows the results of a microplate thermal shift assay of the intrinsic tryptophan fluorescence of human α-thrombin. In this assay, blank well fluorescence was subtracted from sample fluorescence.

[0060] FIG. 15 shows the results of microplate thermal shift assays of single ligand binding interactions to three different classes of binding sites for human α-thrombin.

[0061] FIG. 16 shows the results of microplate thermal shift assays of multi-ligand binding interactions for human α-thrombin.

[0062] FIGS. 17A-D show the results of microplate thermal shift assays of the effect of pH and various sodium chloride concentrations on the stability of human α-thrombin. In FIG. 17A, the fluorophore is 1,8-ANS. In FIG. 17B, the fluorophore is 2,6-ANS. In FIG. 17C, the fluorophore is 2,6-TNS. In FIG. 17D, the fluorophore is bis-ANS.

[0063] FIG. 18 shows the results of microplate thermal shift assays of the effect of calcium chloride, ethylenediaminetetraacetic acid, dithiothreitol, and glycerol on the stability of human α-thrombin.

[0064] FIG. 19 shows the results of microplate thermal shift assays of the effect of pH and sodium chloride concentration of the stability of human D(II) FGF receptor 1.

[0065] FIG. 20 shows the results of microplate thermal shift assays of the effect of various biochemical conditions on the stability of human D(II) FGF receptor 1.

[0066] FIG. 21 shows the results of microplate thermal shift assays of the effect of various biochemical conditions on the stability of human D(II) FGF receptor 1.

[0067] FIG. 22 shows the results of microplate thermal shift assays of the effect of various biochemical conditions on the stability of human D(II) FGF receptor 1.

[0068] FIG. 23 shows the results of microplate thermal shift assays of the effect of various biochemical conditions on the stability of human D(II) FGF receptor 1.

[0069] FIG. 24 shows the results of microplate thermal shift assays of the effect of various biochemical conditions on the stability of human D(II) FGF receptor 1.

[0070] FIG. 25 shows the results of microplate thermal shift assays of the effect of various biochemical conditions on the stability of human urokinase.

[0071] FIG. 26 is a schematic diagram of a thermodynamic model for the linkage of the free energies of protein folding and ligand binding.

[0072] FIG. 27 is a schematic diagram of a method of screening biochemical conditions that optimize protein folding.

[0073] FIG. 28 shows the results of microplate thermal shift assays of human α-thrombin stability using various fluorophores.

[0074] FIG. 29 shows a schematic diagram of one embodiment of an assay apparatus of the present invention.

[0075] FIG. 30 shows a schematic diagram of an alternate embodiment of the assay apparatus of the present invention.

[0076] FIG. 31 shows a schematic diagram of the assay apparatus according to another embodiment of the present invention.

[0077] FIGS. 32A-E illustrate one embodiment of a thermal electric stage for the assay apparatus of the present invention. FIG. 32A shows a side view of the thermal electric stage. FIG. 32B shows a top view of the thermal electric stage. FIGS. 32C-E show three configurations of inserts that can be attached to the thermal electric stage. In one embodiment, inserts accommodate a microtitre plate. In such an embodiment, assay samples are contained within the wells of the microtitre plate.

[0078] FIG. 33 is a schematic diagram illustrating a top view of another embodiment of the assay apparatus of the present invention.

[0079] FIG. 34 is a schematic diagram illustrating the top view of the embodiment of the assay apparatus shown in FIG. 33 with a housing installed.

[0080] FIG. 35 is a schematic diagram illustrating a side view of the embodiment of the assay apparatus shown in FIGS. 33 and 34.

[0081] FIGS. 36A and 36B illustrate a temperature profile and how the temperature profile is implemented using the automated assay apparatus of the present invention.

[0082] FIG. 37 shows an exemplary computer system suitable for use with the present invention.

[0083] FIG. 38 shows a flow diagram illustrating one embodiment for implementation of the present invention.

[0084] FIG. 39 shows a flow diagram illustrating an alternate embodiment for implementation of the present invention.

[0085] FIG. 40 shows a comparison of the results of microplate thermal shift assays of human α-thrombin denaturation performed using a fluorescence scanner and a CCD camera.

[0086] FIGS. 41A and 41B show photographs of microplate thermal shift assay of human α-thrombin denaturation performed using a CCD camera. FIG. 41A: V-bottom well microplate. FIG. 41B: dimple microplate.

[0087] FIG. 42 shows a comparison of the results of microplate thermal shift assays of human α-thrombin denaturation performed using a fluorescence scanner and a CCD camera.

[0088] In the following description, reference will be made to various terms and methodologies known to those of skill in the biochemical and pharmacological arts. Publications and other materials setting forth such known terms and methodologies are incorporated herein by reference in their entireties as though set forth in full.

[0089] Overview of the Methods of the Present Invention

[0090] The present invention provides a method for ranking a multiplicity of different molecules in the order of their ability to bind to a target molecule which is capable of unfolding due to a thermal change. In one embodiment of this method, the target molecule is contacted with one molecule of a multiplicity of different molecules in each of a multiplicity of containers. The containers are then simultaneously heated, in intervals, over a range of temperatures. After each heating interval, a physical change associated with the thermal denaturation of the target molecule is measured. In an alternate embodiment of this method, the containers are heated in a continuous fashion. A thermal denaturation curve is plotted as a function of temperature for the target molecule in each of the containers. Preferably, the temperature midpoint, T_m, of each thermal denaturation curve is identified and is then compared to the T_mof the thermal denaturation curve obtained for the target molecule in the absence of any of the molecules in the containers. Alternatively, an entire thermal denaturation curve can be compared to other entire thermal denaturation curves using computer analytical tools.

[0091] The term “combinatorial library” refers to a plurality of molecules or compounds which are formed by combining, in every possible way for as given compound length, a set of chemical or biochemical building blocks which may or may not be related in structure. Alternatively, the term can refer to a plurality of chemical or biochemical compounds which are formed by selectively combining a particular set of chemical building blocks. Combinatorial libraries can be constructed according to methods familiar to those skilled in the art. For example, see Rapoport et al., Immunology Today 16:43-49 (1995); Sepetov, N. F. et al., Proc. Natl. Acad. Sci. U.S.A. 92:5426-5430 (1995); Gallop, M. A. et al., J. Med. Chem. 9:1233-1251 (1994); Gordon, E. M. et al., J. Med. Chem. 37:1385-1401 (1994); Stankova, M. et al., Peptide Res. 7:292-298 (1994); Erb, E. et al., Proc. Natl. Acad. Sci. U.S.A. 91:11422-11426 (1994); DeWitt, S. H. et al., Proc. Natl. Acad. Sci. U.S.A. 90:6909-6913 (1993); Barbas, C. F. et al., Proc. Natl. Acad. Sci. U.S.A. 89:4457-4461 (1992); Brenner, S. et al. Proc. Natl. Acad. Sci. U.S.A. 89:5381-5383 (1992); Lam, K. S. et al., Nature 354:82-84 (1991); Devlin, J. J. et al., Science 245:404-406 (1990); Cwirla, S. E. et al., Proc. Natl. Acad. Sci. U.S.A. 87:6378-6382 (1990); Scott, J. K. et al., Science 249:386-390 (1990). Preferably, the term “combinatorial library” refers to a DirectedDiversity library, as set forth in U.S. Pat. No. 5,463,564. Regardless of the manner in which a combinatorial library is constructed, each molecule or compound in the library is catalogued for future reference.

[0092] The term “compound library” refers to a plurality of molecules or compounds which were not formed using the combinatorial approach of combining chemical or biochemical building blocks. Instead, a compound library is a plurality of molecules or compounds which are accumulated and are stored for use in future ligand-receptor binding assays. Each molecule or compound in the compound library is catalogued for future reference.

[0093] The terms “multiplicity of molecules,” “multiplicity of compounds,” or “multiplicity of containers” refer to at least two molecules, compounds, or containers.

[0094] The term “multi-variable” refers to more than one experimental variable.

[0095] The term “screening” refers to the testing of a multiplicity of molecules or compounds for their ability to bind to a target molecule which is capable of denaturing.

[0096] The term “ranking” refers to the ordering of the affinities of a multiplicity of molecules or compounds for a target molecule, according to the ability of the molecule or compound to shift the thermal denaturation curve of the target molecule, relative to the thermal denaturation curve of the target molecule in the absence of any molecule or compound.

[0097] The term “ranking” also refers to the ordering of the efficacies of a multiplicity of biochemical conditions in optimizing protein stabilization, protein folding, protein crystallization, or protein shelf life. In the context of optimization of protein stabilization, optimization of protein folding, optimization of protein crystallization, and optimization of protein shelf life, the term “ranking” refers to the ordering of the efficacies of one or more combinations of biochemical conditions to shift the thermal denaturation curve of the target molecule, relative to the thermal denaturation curve of the target molecule under a reference set of conditions.

[0098] The term “reference set of conditions” refers to a set of biochemical conditions under which a thermal denaturation curve for a target molecule is obtained. Thermal denaturation curves obtained under conditions different than the reference conditions are compared to each other and to the thermal denaturation curve obtained for the target molecule under reference conditions.

[0099] As discussed above, ranking molecules, compounds, or biochemical conditions according to a change in the T_mof a thermal denaturation curve is preferable. Alternatively, molecules, compounds, or biochemical conditions can be ranked for their ability to stabilize a target molecule according to the change in entire thermal denaturation curve.

[0100] The term “lead molecule” refers to a molecule or compound, from a combinatorial library, which displays relatively high affinity for a target molecule. The terms “lead compound” and “lead molecule” are synonymous. The term “relatively high affinity” relates to affinities in the K_drange of from 10⁻⁴to 10⁻¹⁵M.

[0101] The term “target molecule” encompasses peptides, proteins, nucleic-acids, and other receptors. The term encompasses both enzymes and proteins which are not enzymes. The term encompasses monomeric and multimeric proteins. Multimeric proteins may be homomeric or heteromeric. The term encompasses nucleic acids comprising at least two nucleotides, such as oligonucleotides. Nucleic acids can be single-stranded, double-stranded or triple-stranded. The term encompasses a nucleic acid which is a synthetic oligonucleotide, a portion of a recombinant DNA molecule, or a portion of chromosomal DNA. The term target molecule also encompasses portions of peptides, proteins, and other receptors which are capable of acquiring secondary, tertiary, or quaternary structure through folding, coiling or twisting. The target molecule may be substituted with substituents including, but not limited to, cofactors, coenzymes, prosthetic groups, lipids, oligosaccharides, or phosphate groups. The term “capable of denaturing” refers to the loss of secondary, tertiary, or quaternary structure through unfolding, uncoiling, or untwisting. The terms “target molecule” and “receptor” are synonymous.

[0102] Examples of target molecules are included, but not limited to those disclosed in Faisst, S. et al., Nucleic Acids Research 20:3-26 (1992); Pimentel, E., Handbook of Growth Factors, Volumes I-III, CRC Press, (1994); Gilman, A. G. et al., The Pharmacological Basis of Therapeutics, Pergamon Press (1990); Lewin, B., Genes V, Oxford University Press (1994); Roitt, I., Essential Immunology, Blackwell Scientific Publ. (1994); Shimizu, Y., Lymphocyte Adhesion Molecules, R G Landes (1993); Hyams, J. S. et al., Microtubules, Wiley-Liss (1995); Montreuil, J. et al., Glycoproteins, Elsevier (1995); Woolley, P., Lipases: Their Structure Biochemistry and Applications, Cambridge University Press (1994); Kuijan, J., Signal Transduction: Prokaryotic and Simple Eukaryotic Systems, Academic Press (1993); Kreis, T., et al., Guide Book to the Extra Cellular Matrix and Adhesion Proteins, Oxford University Press (1993); Schlesinger, M. J., Lipid Modifications of Proteins, CRC Press (1992); Conn, P. M., Receptors: Model Systems and Specific Receptors, Oxford University Press (1993); Lauffenberger, D. A. et al., Receptors: Models For Binding Trafficking and Signaling, Oxford University Press (1993); Webb, E. C., Enzyme Nomenclature, Academic Press (1992); Parker, M. G., Nuclear Hormone Receptors; Molecular Mechanisms, Cellular Functions Clinical Abnormalities, Academic Press Ltd. (1991); Woodgett J. R., Protein Kinases, Oxford University Press (1995); Balch, W. E. et al., Methods in Enzymology, 257, Pt. C: Small GTPases and Their Regulators: Proteins Involved in Transport, Academic Press (1995); The Chaperonins, Academic Press (1996); Pelech, L., Protein Kinase Circuitry in Cell Cycle Control, R G Landes (1996); Atkinson, Regulatory Proteins of the Complement System, Franklin Press (1992); Cooke, D. T. et al., Transport and Receptor Proteins of Plant Membranes: Molecular Structure and Function, Plenum Press (1992); Schumaker, V. N., Advances in Protein Chemistry: Lipoproteins, Apolipoproteins, and Lipases, Academic Press (1994); Brann, M., Molecular Biology of G-Protein-Coupled Receptors: Applications of Molecular Genetics to Pharmacology, Birkhauser (1992); Konig, W., Peptide and Protein Hormones: Structure, Regulations, Activity—A Reference Manual, VCH Publ. (1992); Tuboi, S. et al., Post-Translational Modification of Proteins, CRC Press (1992); Heilmeyer, L. M., Cellular Regulation by Protein Phosphorylation, Springer-Verlag (1991); Takada, Y., Integrin: The Biological Problem, CRC Press (1994); Ludlow, J. W., Tumor Suppressors: Involvement in Human Disease, Viral Protein Interactions, and Growth Regulation, R G Landes (1994); Schlesinger, M. J., Lipid Modification of Proteins, CRC Press (1992); Nitsch, R. M., Alzheimer's Disease: Amyloid Precursor Proteins, Signal Transduction, and Neuronal Transplantation, New York Academy of Sciences (1993); Cochrane, C. G. et al., Cellular and Molecular Mechanisms of Inflammation, Vol. 3: Signal Transduction in Inflammatory Cells, Part A, Academic Press (1992); Gupta, S. et al., Mechanisms of Lymphocyte Activation and Immune Regulation IV: Cellular Communications, Plenum Press (1992); Authi, K. S. et al., Mechanisms of Platelet Activation and Control, Plenum Press (1994); Grunicke, H., Signal Transduction Mechanisms in Cancer, R G Landes (1995); Latchman, D. S., Eukaryotic Transcription Factors, Academic Press (1995).

[0103] The term “target molecule” refers more specifically to proteins involved in the blood coagulation cascade, fibroblast growth factors, and fibroblast growth factor receptors, urokinase, and factor D.

[0104] The term “molecule” refers to the compound which is tested for binding affinity for the target molecule. This term encompasses chemical compounds of any structure, including, but not limited to nucleic acids and peptides. More specifically, the term “molecule” encompasses compounds in a compound or a combinatorial library. The terms “molecule” and “ligand” are synonymous.

[0105] The terms “thermal change” and “physical change” encompass the release of energy in the form of light or heat, the absorption of energy in the form or light or heat, changes in turbidity and changes in the polar properties of light. Specifically, the terms refer to fluorescent emission, fluorescent energy transfer, absorption of ultraviolet or visible light, changes in the polarization properties of light, changes in the polarization properties of fluorescent emission, changes in turbidity, and changes in enzyme activity. Fluorescence emission can be intrinsic to a protein or can be due to a fluorescence reporter molecule (below). For a nucleic acid, fluorescence can be due to ethidium bromide, which is an intercalating agent. Alternatively, the nucleic acid can be labeled with a fluorophore (below).

[0106] The term “contacting a target molecule” refers broadly to placing the target molecule in solution with the molecule to be screened for binding. Less broadly, contacting refers to the turning, swirling, shaking or vibrating of a solution of the target molecule and the molecule to be screened for binding. More specifically, contacting refers to the mixing of the target molecule with the molecule to be tested for binding. Mixing can be accomplished, for example, by repeated uptake and discharge through a pipette tip. Preferably, contacting refers to the equilibration of binding between the target molecule and the molecule to be tested for binding. Contacting can occur in the container (infra) or before the target molecule and the molecule to be screened are placed in the container.

[0107] The target molecule may be contacted with a nucleic acid prior to being contacted with the molecule to be screened for binding. The target molecule may be complexed with a peptide prior to being contacted with the molecule to be screened for binding. The target molecule may be phosphorylated or dephosphorylated prior to being contacted with the molecule to be screened for binding.

[0108] A carbohydrate moiety may be added to the target molecule before the target molecule is contacted with the molecule to be screened for binding. Alternatively, a carbohydrate moiety may be removed from the target molecule before the target molecule is contacted with the molecule to be screened for binding.

[0109] The term “container” refers to any vessel or chamber in which the receptor and molecule to be tested for binding can be placed. The term “container” encompasses reaction tubes (e.g., test tubes, microtubes, vials, etc.). Preferably, the term “container” refers to a well in a multiwell microplate or microtiter plate. The term “sample” refers to the contents of a container.

[0110] A “thermal denaturation curve” is a plot of the physical change associated with the denaturation of a protein or a nucleic acid as a function of temperature. See, for example, Davidson et al., Nature Structure Biology2:859 (1995); Clegg, R. M. et al., Proc. Natl. Acad. Sci. U.S.A.90:2994-2998 (1993).

[0111] The “midpoint temperature, T_m” is the temperature midpoint of a thermal denaturation curve. The T_mcan be readily determined using methods well known to those skilled in the art. See, for example, Weber, P. C. et al., J. Am. Chem. Soc. 116:2717-2724 (1994); Clegg, R. M. et al., Proc. Natl. Acad. Sci. U.S.A. 90:2994-2998 (1993).

[0112] The term “fluorescence probe molecule” refers to a fluorophore, which is a fluorescent molecule or a compound which is capable of binding to an unfolded or denatured receptor and, after excitement by light of a defined wavelength, emits fluorescent energy. The term fluorescence probe molecule encompasses all fluorophores. More specifically, for proteins, the term encompasses fluorophores such as thioinosine, and N-ethenoadenosine, formycin, dansyl derivatives, fluorescein derivatives, 6-propionyl-2-(dimethylamino)-napthalene (PRODAN), 2-anilinonapthalene, and N-arylamino-naphthalene sulfonate derivatives such as 1-anilinonaphthalene-8-sulfonate (1,8-ANS), 2-anilinonaphthalene-6-sulfonate (2,6-ANS), 2-aminonaphthalene-6-sulfonate,N,N-dimethyl-2-aminonaphthalene-6-sulfonate, N-phenyl-2-aminonaphthalene, N-cyclohexyl-2-aminonaphthalene-6-sulfonate, N-phenyl-2-aminonaphthalene-6-sulfonate, N-phenyl-N-methyl-2-aminonaph-thalene-6-sulfonate, N-(o-toluyl)-2-aminonaphthalene-6-sulfonate, N-(m-toluyl)-2-aminonaphthalene-6-sulfonate, N-(p-toluyl)-2-aminonaphthalene-6-sulfonate, 2-(p-toluidinyl)-naphthalene-6-sulfonic acid (2,6-TNS), 4-(dicyanovinyl) julolidine (DCVJ), 6-dodecanoyl-2-dimethylaminonaphthalene (LAURDAN), 6-hexadecanoyl-2-(((2-(trimethylammonium)ethyl)methyl)amino)naphthalenechloride(PATMAN), nile red, N-phenyl-1-naphthylamine, 1,1-dicyano-2-[6-(dimethylamino) naphthalen-2-yl]propene (DDNP), 4,4′-dianilino-1,1-binaphthyl-5,5-disulfonic acid (bis-ANS), and Dapoxyl™ derivatives (Molecular Probes, Eugene, Oreg.). Preferably for proteins, the term refers to 1,8-ANS or 2,6-TNS.

[0113] A double-stranded oligonucleotide may be used in fluorescence resonance energy transfer assays. One strand of the oligonucleotide will contain the donor fluorophore. The other strand of the oligonucleotide will contain the acceptor fluorophore. For a nucleic acid to “contain” a donor or an acceptor fluorophore, the fluorophore can be incorporated directly into the oligonucleotide sequence. Alternatively, the fluorophore can be attached to either the 5′- or 3′-terminus of the oligonucleotide.

[0114] A donor fluorophore is one which, when excited by light, will emit fluorescent energy. The energy emitted by the donor fluorophore is absorbed by the acceptor fluorophore. The term “donor fluorophore” encompasses all fluorophores including, but not limited to, carboxyfluorescein, iodoacetamidofluorescein, and fluorescein isothiocyanate. The term “acceptor fluorophore” encompasses all fluorophores including, but not limited to, iodoacetamidoeosin and tetramethylrhodamine.

[0115] The term “carrier” encompasses a platform or other object, of any shape, which itself is capable of supporting at least two containers. The carrier can be made of any material, including, but not limited to glass, plastic, or metal. Preferably, the carrier is a multiwell microplate. The terms microplate and microtiter plate are synonymous. The carrier can be removed from the heating element. In the present invention, a plurality of carriers are used. Each carrier holds a plurality of containers.

[0116] The term “biochemical conditions” encompasses any component of a physical, chemical, or biochemical reaction. Specifically, the term refers to conditions of temperature, pressure, protein concentration, pH, ionic strength, salt concentration, time, electric current, potential difference, concentrations of cofactor, coenzyme, oxidizing agents, reducing agents, detergents, metal ion, ligands, or glycerol.

[0117] The term “denatured protein” refers to a protein which has been treated to remove secondary, tertiary, or quaternary structure. The term “native protein” refers to a protein which possesses the degree of secondary, tertiary or quaternary structure that provides the protein with full chemical and biological function. A native protein is one which has not been heated and has not been treated with denaturation agents or chemicals such as urea.

[0118] The term “denatured nucleic acid” refers to a nucleic acid which has been treated to remove folded, coiled, or twisted structure. Denaturation of a triple-stranded nucleic acid complex is complete when the third strand has been removed from the two complementary strands. Denaturation of a double-standed DNA is complete when the base pairing between the two complementary strands has been interrupted and has resulted in single-stranded DNA molecules that have assumed a random form. Denaturation of single-stranded RNA is complete when intramolecular hydrogen bonds have been interrupted and the RNA has assumed a random, non-hydrogen bonded form.

[0119] The terms “folding,” “refolding,” and “renaturing” refer to the acquisition of the correct secondary, tertiary, or quaternary structure, of a protein or a nucleic acid, which affords the full chemical and biological function of the biomolecule.

[0120] The term “efficacy” refers to the effectiveness of a particular set of biochemical conditions in facilitating the refolding or renaturation of an unfolded or denatured protein.

[0121] The terms “spectral measurement” and “spectrophotometric measurement” refer to measurements of changes in the absorption of light. Turbidity measurements, measurements of visible light absorption, and measurement of ultraviolet light absorption are examples of spectral measurements.

[0122] The term “polarimetric measurement” relates to measurements of changes in the polarization properties of light and fluorescent emission. Circular dichroism and optical rotation are examples of polarization properties of light which can be measured polarimetrically. Measurements of circular dichroism and optical rotation are taken using a spectropolarimeter. “Nonpolarimetric” measurements are those that are not obtained using a spectropolarimeter.

[0123] The term “collection” refers to a pool or a group of at least one molecule to be tested for binding to a target molecule or receptor.

[0124] A “host” is a bacterial cell that has been transformed with recombinant DNA for the purpose of expressing protein which is heterologous to the host bacterial cell.

[0125] The thermal shift assay is based on the ligand-dependent change in the thermal denaturation curve of a receptor, such as a protein or a nucleic acid. When heated over a range of temperatures, a receptor will unfold. By plotting the degree of denaturation as a function of temperature, one obtains a thermal denaturation curve for the receptor. A useful point of reference in the thermal denaturation curve is the temperature midpoint (T_m), the temperature at which the receptor is half denatured.

[0126] Ligand binding stabilizes the receptor (Schellman, J., Biopolymers 14:999-1018 (1975)). The extent of binding and the free energy of interaction follow parallel courses as a function of ligand concentration (Schellman, J., Biophysical Chemistry 45:273-279 (1993); Barcelo, F. et al., Chem. Biol. Interactions 74:315-324 (1990)). As a result of stabilization by ligand, more energy (heat) is required to unfold the receptor. Thus, ligand binding shifts the thermal denaturation curve. This property can be exploited to determine whether a ligand binds to a receptor: a change, or “shift”, in the thermal denaturation curve, and thus in the T_m, suggests that a ligand binds to the receptor.

[0127] The thermodynamic basis for the thermal shift assay has been described by Schellman, J. A. (Biopolymers15:999-1000 (1976)), and also by Brandts et al. (Biochemistry 29:6927-6940 (1990)). Differential scanning calorimetry studies by Brandts et al., (Biochemistry 29:6927-6940 (1990)) have shown that for tight binding systems of 1:1 stoichiometry, in which there is one unfolding transition, one can estimate the binding affinity at T_mfrom the following expression: $\begin{array}{cc}{K}_L^{T_m}=\frac{\begin{array}{c}\exp \{-\frac{\Delta H_u^{T_0}}{R}\left[\frac{1}{T_m}-\frac{1}{T_0}\right]+\\ \frac{\Delta C_{\mathrm{pu}}}{R}\left[\ln \left(\frac{T_m}{T_0}\right)+\frac{T_0}{T_m}-1\right]\}\end{array}}{\left[L_{T_m}\right]}& \text{(equation 1)}\end{array}$

[0128] where

[0129] K_L^T=the ligand association constant at T_m;

[0130] T_m=the midpoint for the protein unfolding transition in the presence of ligand;

[0131] T₀=the midpoint for the unfolding transition in the absence of ligand;

[0132] ΔH_u^T=the enthalpy of protein unfolding in the absence of ligand at T₀;

[0133] ΔC_pu=the change in heat capacity upon protein unfolding in the absence of ligand;

[0134] [L_T]=the free ligand concentration at T_m; and

[0135] R=the gas constant.

[0136] The parameters ΔH_uand ΔC_puare usually observed from differential scanning calorimetry experiments and are specific for each receptor. To calculate the binding constant from equation 1, one should have access to a differential scanning calorimetry instrument to measure ΔH_uand ΔC_pufor the receptor of interest. One can also locate these parameters for the receptor of interest, or a receptor closely related to it, in the literature. In these situations, equation (1) will allow the accurate measurement of K_Lat T_m.

[0137] It is also possible to calculate the ligand association constant at any temperature, K_Lat T, using equation 2. To use equation 2, calorimetry data for the binding enthalpy at T, ΔH_L, and the change of heat capacity upon ligand binding, ΔC_pLmust be known (Brandts et al., Biochemistry29:6927-6940 (1990)). $\begin{array}{cc}{K}_L^T=K_L^{T_m}\exp \left\{-\frac{\Delta H_L^T}{R}\left[\frac{1}{T}-\frac{1}{T_m}\right]+\frac{\Delta C_{\mathrm{pL}}}{R}\left[\ln \left(\frac{T}{T_m}\right)-\frac{T}{T_m}+1\right]\right\}& \text{(equation 2)}\end{array}$

[0138] where

[0139] K_L^T=the ligand association constant at any temperature T;

[0140] K_L^T=the ligand association constant at T_m;

[0141] T_m=the midpoint for the protein unfolding transition in the presence of ligand;

[0142] ΔH_L^T=the enthalpy of ligand binding in the absence of ligand at T;

[0143] ΔC_pL=the change in heat capacity upon binding of ligand; and

[0144] R=the gas constant.

[0145] The second exponential term of equation 2 is usually small enough to be ignored so that approximate values of K_Lat T can be obtained using just the first exponential term: $\begin{array}{cc}{K}_L^T=K_L^{T_m}\exp \left\{-\frac{\Delta H_L^T}{R}\left[\frac{1}{T}-\frac{1}{T_m}\right]\right\}& \text{(equation 3)}\end{array}$

[0146] One need not, however, calculate binding constants according to equations 1-3 in order to rank the affinities of a multiplicity of different ligands for a receptor. Rather, the present invention provides a method for ranking affinities of ligands according to the degree to which the thermal denaturation curve is shifted by the ligand. Thus, it is possible to obtain estimates of K_Lat T_m, even in the absence of accurate values of ΔH_u, ΔC_pu, and ΔH_L.

[0147] The present invention is particularly useful for screening a combinatorial or a compound library. To achieve high throughput screening, it is best to house samples on a multicontainer carrier or platform. A multicontainer carrier facilitates the heating of a plurality of samples simultaneously. In one embodiment, a multiwell microplate, for example a 96 or a 384 well microplate, which can accommodate 96 or 384 different samples, is used as the carrier.

[0148] In one embodiment, one sample is contained in each well of a multi-well microplate. The control well contains receptor, but no molecule to be tested for binding. Each of the other samples contains at least one molecule to be tested for binding. The thermal denaturation curve for the receptor in the control well is the curve against which curves for all of the other experiments are compared.

[0149] The rate of screening is accelerated when the sample contains more than one molecule to be tested for binding. For example, the screening rate is increased 20-fold when the sample contains a pool of 20 molecules. Samples which contain a binding molecule must then be divided into samples containing a smaller collection of molecules to be tested for binding. These divided collections must then be assayed for binding to the target molecule. These steps must be repeated until a single molecule responsible for the original thermal shift is obtained.

[0150] Receptor denaturation can be measured by light spectrophotometry. When a protein in solution denatures in response to heating, the receptor molecules aggregate and the solution becomes turbid. Thermally induced aggregation upon denaturation is the rule rather than the exception. Aggregation generally complicates calorimetric experiments. Aggregation, however, is an advantage when using a spectrophotometric technology, because changes in turbidity can be measured by monitoring the change in absorbance of visible or ultraviolet light of a defined wavelength.

[0151] Denaturation of a nucleic acid can be monitored using light spectrophotometry. The change in hyperchromicity, which is the increase in absorption of light by polynucleotide solutions due to a loss of ordered structure, is monitored as a function of increase in temperature. Changes in hyperchromicity is typically assayed using light spectrophotometry.

[0152] In another embodiment, however, fluorescence spectrometry is used to monitor thermal denaturation. The fluorescence methodology is more sensitive than the absorption methodology.

[0153] The use of intrinsic protein fluorescence and fluorescence probe molecules in fluorescence spectroscopy experiments is well known to those skilled in the art. See, for example, Bashford, C. L. et al., Spectrophotometry and Spectrofluorometry: A Practical Approach, pp. 91-114, IRL Press Ltd. (1987); Bell, J. E., Spectroscopy in Biochemistry, Vol. I, pp. 155-194, CRC Press (1981); Brand, L. et al., Ann. Rev. Biochem. 41:843 (1972).

[0154] If the target molecule or receptor to be studied is a nucleic acid, fluorescence spectrometry can be performed using an ethidium bromide displacement assay (Lee, M. et al., J. Med. Chem. 36:863-870 (1993)). In this approach, ligand binding displaces ethidium bromide and results in a decrease in the fluorescent emission from ethidium bromide. An alternative approach is to use fluorescence resonance emission transfer. In the latter approach, the transfer of fluorescent energy, from a donor fluorophore on one strand of an oligonucleotide to an acceptor fluorophore on the other strand, is monitored by measuring the emission of the acceptor fluorophore. Denaturation prevents the transfer of fluorescent energy.

[0155] The fluorescence resonance emission transfer methodology is well known to those skilled in the art. Fore example, see Ozaki, H. et al., Nucleic Acids Res. 20:5205-5214 (1992); Clegg, R. M. et al., Proc. Natl. Acad. Sci. U.S.A. 90:2994-2998 (1993); Clegg, R. M. et al., Biochemistry 31:4846-4856 (1993).

[0156] The element upon which the sample carrier is heated can be any element capable of heating samples rapidly and in a reproducible fashion. In the present invention, a plurality of samples is heated simultaneously. The plurality of samples can be heated on a single heating element. Alternatively, the plurality of samples can be heated to a given temperature on one heating element, and then moved to another heating element for heating to another temperature. Heating can be accomplished in regular or irregular intervals. To generate a smooth denaturation curve, the samples should be heated evenly, in intervals of 1 or 2° C. The temperature range across which the samples can be heated is from 25 to 110 ° C. Spectral readings are taken after each heating step. Samples can be heated and read by the spectral device in a continuous fashion. Alternatively, after each heating step, the samples may be cooled to a lower temperature prior to taking the spectral readings. Preferably, the samples are heated continuously and spectral readings are taken while the samples are being heated.

[0157] Spectral readings can be taken on all of the samples in the carrier simultaneously. Alternatively, readings can be taken on samples in groups of at least two at a time. Finally, the readings can be taken one sample at a time.

[0158] In one embodiment, thermal denaturation is monitored by fluorescence spectrometry using an assay apparatus such as the one shown in FIG. 29. The instrument consists of a scanner and a control software system. The system is capable of quantifying soluble and cell-associated fluorescence emission. Fluorescence emission is detected by a photomultiplier tube in a light-proof detection chamber. The software runs on a personal computer and the action of the scanner is controlled through the software. A precision X-Y mechanism scans the microplate with a sensitive fiber-optic probe to quantify the fluorescence in each well. The microplate and samples can remain stationary during the scanning of each row of the samples, and the fiber-optic probe is then is moved to the next row. Alternatively, the microplate and samples can be moved to position a new row of samples under the fiber-optic probe. The scanning system is capable of scanning 96 samples in under one minute. The scanner is capable of holding a plurality of excitation filters and a plurality of emission filters to measure the most common fluorophores. Thus, fluorescence emission readings can be taken one sample at a time, or on a subset of samples simultaneously. An alternate embodiment of the assay apparatus is shown in FIG. 33. The assay apparatus of the present invention will be described in more detail below.

[0159] The present invention is also directed to a product of manufacture which comprises a carrier having a multiplicity of containers within it. The product of manufacture can be used to screen a combinatorial library for lead compounds which bind to the receptor of interest. The combinatorial library can be screened using the method according to the present invention.

[0160] In the product of manufacture, each of the containers contains a uniform amount of a receptor of interest. In addition, each of these containers contains a different compound from a combinatorial library at a concentration which is at least 2-fold above the concentration of the receptor. Preferably, the product of manufacture is a multiwell microplate or a multiplicity of multiwell microplates. If the receptor is a protein, each container may further contain a fluorescence probe molecule. If the receptor is a nucleic acid, each container may further contain ethidium bromide. Alternatively, the nucleic acid may be labeled with a fluorophore.

[0161] Prior to use, the product of manufacture can be stored in any manner necessary to maintain the integrity of the receptor of interest. For example, the product of manufacture can be stored at a temperature between −90° C. and room temperature. The receptor and compound can be stored in lyophilized form, in liquid form, in powdered form, or can be stored in glycerol. The product of manufacture may be stored either in the light or in the dark.

[0162] The heat conducting element or block upon which the sample carrier is heated can be any element capable of heating samples rapidly and reproducibly. The plurality of samples can be heated on a single heating element Alternatively, the plurality of samples can be heated to a given temperature on one heating element, and then moved to another heating element for heating to another temperature. Heating can be accomplished in regular or irregular intervals. To generate a smooth denaturation curve, the samples should be heated evenly, in intervals of 1 or 2° C. The temperature range across which the samples can be heated is from 25 to 110° C.

[0163] In the present invention, a plurality of samples is heated simultaneously. If samples are heated in discrete temperature intervals, in a stairstep fashion, spectral readings are taken after each heating step. Alternatively, after each heating step, the samples may be cooled to a lower temperature prior to taking the spectral readings. Alternatively, samples can be heated in a continuous fashion and spectral readings are taken during heating.

[0164] Spectral readings can be taken on all of the samples on a carrier simultaneously. Alternatively, readings can be taken on samples in groups of at least two at a time.

[0165] The present invention also provides an improved method for generating lead compounds. After a compound or a combinatorial library of compounds has been screened using the thermal shift assay, compounds which bind to the target receptor are chemically modified to generate a second library of compounds. This second library is then screened using the thermal shift assay. This process of screening and generating a new library continues until compounds that bind to the target receptor with affinities in the K_drange of from 10⁻⁴to 10⁻¹⁵M are obtained.

[0166] A fluorescence emission imaging system can be used to monitor the thermal denaturation of a target molecule or a receptor. Fluorescence emission imaging systems are well known to those skilled in the art. For example, the AlphaImager™ Gel Documentation and Analysis System (Alpha Innotech, San Leandro, Calif.) employs a high performancd charge coupled device camera with 768×494 pixel resolution. The charge coupled device camera is interfaced with a computer and images are anlayzed with Image analysis software™. The ChemiImager™ (Alpha Innotech) is a cooled charge coupled device that performs all of the functions of the AlphaImager™ and in addition captures images of chemiluminescent samples and other low intensity samples. The ChemiImager™ charge coupled device includes a Pentium processor (1.2 Gb hard drive, 16 Mb RAM), AlphaEase™ analysis software, a light tight cabinet, and a UV and white light trans-illuminator. For example, the MRC-1024 UV/Visible Laser Confocal Imaging System (BioRad, Richmond, Calif.) facilitates the simultaneous imaging of more than one fluorophore across a wide range of illumination wavelengths (350 to 700 nm). The Gel Doc 1000 Fluorescent Gel Documentation System (BioRad, Richmond, Calif.) can clearly display sample areas as large as 20×20 cm, or as small as 5×4 cm. At least two 96 well microplates can fit into a 20×20 cm area. The Gel Doc 1000 system also facilitates the performance of time-based experiments.

[0167] A fluorescence thermal imaging system can be used to monitor receptor unfolding in a microplate thermal shift assay. In this embodiment, a plurality of samples is heated simultaneously between 25 to 110° C. A fluorescence emission reading is taken for each of the plurality of samples simultaneously. For example, the fluorescence emission in each well of a 96 or a 384 well microplate can be monitored simultaneously. Alternatively, fluorescence emission readings can be taken continuously and simultaneously for each sample. At lower temperatures, all samples display a low level of fluorescence emission. As the temperature is increased, the fluorescence in each sample increases. Wells which contain ligands which bind to the target molecule with high affinity shift the thermal denaturation curve to higher temperatures. As a result, wells which contain ligands which bind to the target molecule with high affinity fluoresce less, at a given temperature above the T_mof the target molecule in the absence of any ligands, than wells which do not contain high-affinity ligands. If the samples are heated in incremental steps, the flourescence of all of the plurality of samples is simultaneoulsy imaged at each heating step. If the samples are heated continuously, the fluorescent emission of all of the plurality of samples is simultaneously imaged during heating.

[0168] A thermal shift assay can be performed in a volume of 100 μL volumes. For the following reasons, however, it is preferable to perform a thermal shift assay in a volume of 10 μL. First, approximately 10-fold less protein is required for the miniaturized assay. Thus, only ˜5 to 40 pmole of protein are required (0.1 μg to 1.0 μg for a 25 kDa protein) for the assay (i.e. 5 to 10 μL working volume with a target molecule concentration of about 1 to about 4 μM). Thus, 1 mg of protein can be used to conduct 1,000 to 10,000 assays in the miniaturized format. This is particularly advantageous when the target molecule is available in minute quantities.

[0169] Second, approximately 10-fold less ligand is required for the miniaturized assay. This advantage is very important to researchers when screening valuable combinatorial libraries for which library compounds are synthesized in minute quantities. In the case of human α-thrombin, the ideal ligand concentration is about 50 μM, which translates into 250 pmoles of ligand, or 100 ng (assuming a MW of 500 Da) of ligand per assay in the miniaturized format.

[0170] Third, the smaller working volume allows the potential of using larger arrays of assays because the miniaturized assay can fit into a much smaller area. For example, a 384 well (16×24 array) or 864 well (24×36 array) plates have roughly the same dimensions as the 96 well plates (about 8.5×12.5 cm). The 384 well plate and the 864 well plate allows the user to perform 4 and 9 times as many assays, respectively, as can be performed using a 96 well plate. Alternatively, 1536 well plates (32×48 arrays; Matrix Technologies Corp.) can be used. A 1536 well plate will facilitate sixteen times the throughput afforded by a 96 well plate.

[0171] Thus, using the 1536 well plate configuration, the assay speed can be increased by about 16 times, relative to the speed at which the assay can be performed using the 96 well format. The 8×12 assay array arrangement (in a 96-well plate) facilitates the performance of 96 assays/hr, or about 2300 assays/24 hours. The 32×48 array assay arrangement facilitates the performance of about 1536 assays hr., or about 37,000 assays/24 hours can be performed using a 32×48 assay array configuration.

[0172] The assay volume can be 1-100 μL. Preferably, the assay volume is 1-50 μL. More preferably, the assay volume is 1-25 μL. More preferably still, the assay volume is 1-10 μL. More preferably still, the assay volume is 1-5 μL. More preferably still, the assay volume is 5 μL. Most preferably, the assay volume is 1 μL or 2 μL.

[0173] Preferably, the assay is performed in V-bottom polycarbonate plates or polycarbonate dimple plates. A dimple plate is a plate that contains a plurality of round-bottom wells that hold a total volume of 15 μL.

[0174] One alternative to taking spectral readings over a temperature range around the T_mof the therapeutic target to obtain a full thermal unfolding curve for the ligand/target complex, in order to identify shifts in T_m, is to perform the assay at a single temperature near the T_mof the target molecule. In this embodiment, samples that emit less fluorescence, relative to a control sample (containing a target molecule, but no candidate ligand) indicate that the candidate ligand binds to the target molecule.

[0175] In this embodiment, the magnitude of a physical change associated with the thermal denaturation of a target molecule resulting from heating is determined by generating a thermal denaturation curve for the target molecule as a function of temperature over a range of one or more discrete or fixed temperatures. The physical change associated with thermal denaturation, for example, fluorescence emission, is measured. The magnitude of the physical change at the discrete or fixed temperature for the target molecule in the absence of any ligand is noted. The magnitude of the physical change in the presence of each of a multiplicity of different molecules, for example, combinatorial compounds, is measured. The magnitude of the physical change associated with thermal denaturation of the target molecule in the presence of each of the multiplicity of molecules is compared to magnitude of the physical change obtained for the target molecule at the discrete or fixed temperature in the absence of any of the multiplicity of different molecules. The affinities of the multiplicity of different molecules are ranked according to the change in the magnitude of the physical change.

[0176] The discrete or fixed temperature at which the physical change is measure can be any temperature that is useful for discriminating shifts in thermal stability. Preferably, the discrete or fixed temperature is the midpoint temperature T_mfor the thermal denaturation curve for the target molecule in the absence of any of the multiplicity of different molecules tested for binding to the target molecule.

[0177] The single temperature configuration is particularly advantageous if one is interested in assaying a series of relatively high affinity ligands, which are the preferred compounds for candidates in clinical testing. In cases where a less stringent requirement for binding affinity is preferred, however, one may increase the ligand concentration to 500 μM in order to identify ligands with K_d's of 2.5 μM or higher affinity.

[0178] The single temperature embodiment offers a number of advantages. First, assay speed is increased by a factor often fold. Thus, as the 96 well plate (8×12 array) assay facilitates about 96 assays per hour, the single temperature variation will facilitate about 1000 assays per hr. Using a 1536 well plate (32×48 array), as long as sample aliquoting can be effected at the same rate for the 32×48 array system as in the 8×12 array system, about 15,000 assays can be performed per hour.

[0179] Another alternative method for detecting the thermal unfolding transitions for the microplate thermal shift assay is through the intrinsic tryptophan (Trp) fluorescence of the target protein. Most fluorescence emission plate readers, such as the CytoFluor II, use tungsten-halogen lamps as their light source. These lamps do not give off enough light near 280 nm to allow excitation of the intrinsic Trp residues of proteins which absorb maximally near 280 nm. However, the Biolumin 960 (Molecular Dynamics) uses a Xenon-Arc lamp. The Xenon-Arc lamp affords excitation at 280 nm and the measurement of emission at 350 nm.

[0180] The methods and assay apparatus of the present invention are not limited to assaying ligand-protein interactions. The methods and the assay apparatus can be used to rapidly assay any multi-variable system related to protein stabilization. For example, the methods and the assay apparatus of the present invention can be used to simultaneously assay the binding of more than one compound or ligand to a target molecule. Using this approach, the additive effect of multiple-ligand binding can be assessed. Positive and negative cooperativity can be determined. To accomplish this method, thermal shift assays are performed for a target molecule, such as a protein, in the absence of any ligands, in the presence of a single ligand, and in the presence of two or more ligands. A thermal denaturation curve is generated for the protein alone and for each combination of protein and ligand(s). The midpoint temperature T_mis then determined for each curve. Each T_mis then compared to each of the other T_m's for the other curves. Alternatively, each entire thermal denaturation curve is compared to each of the other thermal denaturation curves. In either of these manners, the additive contribution of more than one ligand to binding interaction or to protein stability can be determined.

[0181] In a similar fashion, the additive contributions of one or more biochemical conditions to protein stability can be determined. Thus, the present invention can be used to rapidly identify biochemical conditions that optimize protein stabililty, and hence shelf-life of a protein.

[0182] Further, the methods and the assay apparatus of the present invention can be used to rank the efficacies of various biochemical conditions for refolding or renaturing an unfolded or denatured protein. This embodiment addresses the need in the art for a reliable method for screening for effective refolding or renaturing conditions.

[0183] For example, expression of recombinant DNA in a bacterial cell usually results in the sequestration of recombinant protein into bacterial inclusion bodies (Marston, F. A. O., Biochem. J. 240:1-12 (1986)). Although other expression systems can be used instead of bacterial expression systems, expression in bacterial cells remains the method of choice for the high-level production of recombinant proteins (Rudolph, R., Protein Engineering: Principles and Practices, pp. 283-298, John Wiley & Sons (1995)). In many cases, recovery of recombinant protein requires that protein be isolated from inclusion bodies. Protein purification from inclusion bodies process necessitates the denaturation of recombinant protein. As a result, recombinant protein must be renatured or refolded under conditions suitable to generate the protein in its native, fully functional form.

[0184] In each of these cases, denatured protein must be renatured or refolded in order to be useful for further study or use. Unfortunately, one cannot easily predict the exact conditions under which a given protein or fragment of the protein should be renatured. Each protein is different. One must always resort to testing a number of different combinations of renaturing conditions before one can know which set of conditions is optimal. Thus, it is desirable to have a reliable and rapid method for ranking the efficacies of various renaturing conditions.

[0185] Recombinant DNA technology has allowed the biosynthesis of a wide variety of heterologous polypeptides of interest in relatively large quantities through the recruitment of the bacterial protein expression apparatus. However, the promise of cheap and abundant supplies of correctly folded rare human proteins of high therapeutic value expressed in E. colihas foundered due to the overwhelmingly predominant aggregation of unfolded or partially unfolded target proteins into insoluble protein inclusion bodies. For recent reviews, see Rudolph, R., & Lilie, H., FASEB J.10:49-56 (1995); Sadana, A., Biotechnology & Bioengineering48:481-489 (1995); Jaenicke, R., Phil. Trans. Royal Soc. London Ser. B-Biol. Sci.348:97-105 (1995)). Reasons for the prevailing self aggregation reaction in E. colihave centered on the relatively high concentration of the heterologous protein (as high as 30% of the weight of the cell) found to various degrees in partially unfolded states. Thus, at the elevated protein concentrations of an overexpressing E. colistrain, the exposed hydrophobic residues of unfolded proteins are more likely to encounter other molecules with similarly exposed groups (inter-molecular reaction) than they are to sample self collapsed polypeptide conformations where these hydrophobic residues are packed in a proper orientation (intra-molecular transition state) for proceeding to the fully folded native state (see FIG. 26). From this perspective, the insoluble protein inclusion bodies are seen as kinetically trapped side reaction products that thwart the preferred protein folding process.

[0186] Techniques for isolating inclusion bodies, purifying recombinant protein from inclusion bodies, and techniques for refolding or renaturing protein are well known to those skilled in the art. For example, see Sambrook, J. et al., Molecular Cloning: a Laboratory Manual, pp. 17.37-17.41, Cold Spring Harbor Laboratory Press (1989); Rudolph, R. et al., FASEB J. 10:49-56 (1995).

[0187] Another impediment to producing large quantities of correctly folded proteins in E. coliis that the reducing redox potential of the E. colicytosol impedes the formation of disulfide bonds in vivo. The formation of disulfide bonds is an important co- and post-translational event in the biosynthesis of many extracellular proteins that is often coupled to protein folding. In addition, the cis-trans proline isomerization reaction has been demonstrated to be a rate determining step for correct folding of certain proteins (Lin, L.-N., & Brandts, J. F., Biochemistry22:564-573 (1983)). As a result, partially folded intermediates accumulate in sufficient quantity in vivo that they aggregate and precipitate into protein masses.

[0188] Cells employ a class of host proteins called molecular chaperonins that assist in vivo protein folding by apparently preventing many of the unproductive side reactions discussed above with regard to inclusion body formation, i.e. aggregation and improper disulfide bond formation. However, the E. colichaperonin machinery, which is comprised in part by the proteins, GroEL and GroES, presumably becomes overwhelmed by massive overexpression. Despite many attempts to correct this chaperonin deficit by co-expression of molecular chaperoning with the protein of interest (Rudolph, R., & Lilie, H., The FASEB J. 10:49-56 (1995)) positive results have been reported in only one case (Goloubinoff, P., et al., Nature342:884-889 (1989)).

[0189] Two hypotheses have been promoted to explain how GroEL and GroES assist in vivo protein folding. Under the first hypothesis, the Anfinsen cage hypothesis, the function of a molecular chaperonin is to provide a protected environment where folding of the protein to its native state can proceed without interference by pro-aggregation conditions in the cell (Martin, et al, Nature352:36-42 (1991); Ellis, R. J., Current Biology4:633-635 (1994)). Under the second hypothesis, the “iterative annealing” hypothesis, the function of the chaperonin is to partly unfold misfolded proteins (that is, kinetically trapped intermediates) with some of the energy of ATP hydrolysis being channeled into the conformational energy of the substrate polypeptide, forcing the polypeptide into a higher energy state from which it could once again attempt to refold correctly after being released into solution (Todd, M. J. et al., Science265:659-666 (1994); Jackson, et al., Biochemistry32:2554-2563 (1993); Weissman, J. S., et al., Cell78:693-702 (1994); Weissman, J. S., & Kim, P. S., Science253:1386-1393 (1991)).

[0190] The in vivo results discussed above are in many ways consistent with the more recent experiences with in vitro refolding of recombinant heterologous proteins expressed in E. coli.That is, while the primary amino acid sequence of a protein may contain sufficient information to determine its native folded conformation (Anfinsen, C. B., Science181:223-230 (1973)), the biochemical conditions in which the folding reaction takes place can strongly influence the partitioning between unfolded, aggregated, and correctly folded forms.

[0191] For example, pH can be understood to influence the folding reaction by its effect on the long range electrostatic interactions summed in the fourth term of the equation (4).

[0192] ΔG_fold=ΔG_conf+ΣΔg_i,int+ΣΔg_i,s+ΔW_el+(ΔG_bind)  Equation(4)

[0193] where

[0194] ΔG_conf=conformational free energy (order/disorder term);

[0195] Δg_i,int=short range interactions (H-bonds, van der Walls interactions, salt bridges, cofactor binding, etc.);

[0196] Δg_i,s=short range interactions with solvent (hydrophobic effect, hydration of ions, etc.); and

[0197] ΔW_el=long range electrostatic interactions.

[0198] ΔG_bind=ligand binding free energy

[0199] As the pH of a protein solution is lowered below the pI for the protein, functional groups on the polypeptide become increasingly protonated, to the point where the electrostatic repulsion between functional groups eventually out balances the other terms in the free energy equation (equation (4)), and the protein is no longer able to adopt the native conformation.

[0200] Another important biochemical parameter for protein folding is the solvent, water, which repels aliphatic and aromatic side chains (and possibly the main chain to some extent) to minimize their exposed surface area. The influence of solvent over the folding reaction is summed in the third term of the free energy equation (equation (4)). Certain salts are known to increase the hydrophobic interaction among protein side chains in water solutions. The effect depends upon the nature of the ions following the Hofmeister series: Cations: Mg²⁺>Li⁺>Na+>K⁺>NH₄⁺. Anions: SO₄²⁻>HPO₄²⁻>acetate>citrate>tartrate>Cl⁻>NO₃−>ClO₃⁻>I⁻>ClO₄⁻>SCN⁻. Stabilizing Hofmeister anions, such as SO₄²⁻ and HPO₄²⁻ at 0.4 M have been found to increase the yield of correctly folded proteins (Creighton, T. E., In: Proteins: Structures and Molecular Properties,Freeman, N.Y., (1984)). This favorable outcome for the native conformation of the protein has been attributed to the cations' and anions' “salting out” effect which leads to the preferential hydration of the protein (Creighton, T. E., In: Proteins: Structures and Molecular Properties,Freeman, N.Y., (1984)).

[0201] Glycerol alters the solvation properties of water to favor the native conformation of proteins. The mechanism by which this occurs is the co-solvent exclusion and preferential hydration of the protein, not unlike the effect of salts of the salts of the Hofmeister series (Timasheff & Arakawa, In: Protein Structure, A Practical Approach,T. E. Creighton, ed., IRL Press, Oxford, UK (1989), pp. 331-354).

[0202] Another example of how the environment influences protein folding is the effect that known ligands and cofactors have on the yield of folded protein. Ligand binding has the effect of shifting the equilibrium from an unfolded state to a native-ligand complex through a coupling of the binding free energy to that of the folding reaction. The role of metal ions in the refolding of bovine carbonic anhydrase II has been described (Bergenhem & Carlsson, Biochim. Biophys. Acta998:277-285 (1989)). Other biochemical parameters that have been shown to affect protein folding are: protein concentration, temperature, glutathione redox buffers (GSH, GSSG), the presence of detergents, and the presence of other additives, such as glycerol, arginine-HCI, polyethylene glycol (PEG), and organic solvents.

[0203] During incubation under refolding conditions, recombinant proteins can be immobilized to solid phase support. This configuration resembles the “Anfinsen cage” hypothesis for the function of GroEL and GroES where an unfolded protein becomes temporarily immobilized in a protected environment where folding to the native state can proceed without interference from competing aggregation reactions. Confirmation of protein folding on solid supports has now come from two recent reports in the literature. A poly-histidine tagged TIMP-2 protein could be refolded by dialysis while still bound to a metal chelate column (Negro, A. et al., FEBS Lett.360:52-56 (1995)). A polyionic fusion peptide attached to the amino or carboxyl terminus of α-glucosidase allowed folding while bound to heparin-Sepharose resin at about 5 mg/mL (Rudolph & Lilie, FASEB J.10:49-56 (1995)). A polyionic arginine tag metholdology for immobilizing and refolding α-glucosidase is disclosed in Stempfer, G. et al., Nature Biotechnology14:329-334 (1996).

[0204] In the present invention, the thermal shift assay is used to rank the efficacy of various refolding or renaturing conditions. Each of a multiplicity of aliquots of a protein of interest, which has been incubated under a variety of different biochemical folding conditions, are placed in a container in a multicontainer carrier. An aliquot of the native, fully functional protein of known concentration is placed in the control container. The samples can be placed in any multicontainer carrier. Preferably, each sample can be placed in a well of a multiwell microplate.

[0205] In considering the many biochemical variables that can influence the outcome of the protein folding reaction, optimization of protein folding is a multi-variable optimization problem, not unlike protein crystallization and quantitative structure activity relationships (QSAR) in drug discovery. Multi-variable optimization problems require large numbers of parallel experiments to collect as much data as possible in order to influence a favorable response. In this regard, both protein crystallization and QSAR analyses have greatly benefited from mass screening protocols that employ matrix arrays of incremental changes in biochemical or chemical composition.

[0206] The present invention can be used to rank the efficacies of refolding or renaturing conditions. Such conditions include, but are not limited to, the concentration of glycerol, the concentration of protein, the use of agents which catalyze the formation of disulfide bond formation, temperature, pH, ionic strength, type of solvent, the use of thiols such as reduced glutathione (GSH) and oxidized glutathione (GSSG), chaotropes such as urea, guanidinium chlorides, alkyl-urea, organic solvents such as carbonic acid amides, L-arginine HCl, Tris buffer, polyethylene glycol, nonionic detergents, ionic detergents, zwitterionic detergents, mixed micelles, and a detergent in combination with cyclodextrin. The present invention can be used regardless of whether a denaturation agent is removed from the protein using dialysis, column chromatographic techniques, or suction filtration.

[0207] Using a spectral thermal shift assay, the conditions which facilitate optimal protein refolding can be determined rapidly. In this embodiment, the renatured protein samples and a control protein sample (i.e., a sample of native protein in its fully functional form) are heated over a temperature range. At discrete temperature intervals, a spectral reading is taken. Alternatively, spectral readings can be taken during a continuous, pre-determined temperature profile. A thermal denaturation curve is constructed for each sample. The T_mfor the thermal denaturation curve of the native, fully functional protein is determined. The relative efficacies of the refolding conditions are ranked according to the magnitude of the physical change associated with unfolding at the T_mof the native, fully functional protein, relative to the magnitude of the physical change of a known quantity of the native, fully functional protein at that T_m. The magnitude of physical change used to measure the extent of unfolding (reflected on the ordinate, or y-axis, of a thermal denaturation curve) corresponds to the amount of correctly folded protein.

[0208] The present invention provides a method for screening biochemical conditions that facilitate and optimize protein folding. To screen conditions for a given protein, it is first necessary to determine the thermal unfolding profile for a protein of interest. This is accomplished by generating a denaturation curve using the microplate thermal shift assay. Various conditions can be optimized, including pH optimum, ionic strength dependence, concentration of salts of the Hofmeister series, glycerol concentration, sucrose concentration, arginine concentration, dithiothreitol concentration, metal ion concentration, etc.

[0209] Using the microplate thermal shift assay, one can determine one or more biochemical conditions have an additive effect on protein stability. Once a set of biochemical conditions that facilitate an increase in protein stability have been identified using the thermal shift assay, the same set of conditions can be used in protein folding experiments with recombinant protein. See FIG. 27. If the conditions that promote protein stability in the thermal shift assay correlate with conditions that promote folding of recombinant protein, conditions can be further optimized by performing additional thermal shift assays until a combination of stabilizing conditions that result in further increase protein stability are identified. Recombinant protein is then folded under those conditions. This process is repeated until optimal folding conditions are identified. Protein stability is expected to correlate with improved yields of protein folding. Yield of correctly folded protein can be determined using any suitable technique. For example, yield of correctly folded protein can be calculated by passing refolded protein over an affinity column, for example, a column to which a ligand of the protein is attached, and quantifying the amount of protein that is present in the sample. In this way, folding conditions can be assessed for their additive contributions to correct folding. The transition state for the protein folding reaction resembles the native form of the protein more than the denatured form. This has been demonstrated to be the case for may proteins (Fersht, A. R., Curr. Op. Struct. Biol.7.3-9 (1997)).

[0210] The methods and the apparatus of the present invention provide a rapid, high throughput approach to screening for combinations of biochemical conditions that favor the protein folding. The method does not require cumbersome and time consuming steps that conventional approaches to protein folding require. For example, using the method of the present invention, it is not necessary to dilute protein to large volumes and low protein concentrations (˜10 μg/mL) in order to avoid aggregation problems associated with conventional methods of recombinant protein refolding. Suppression of protein aggregation will allow for screening biochemical parameters that shift the protein folding equilibrium (between the unfolded and the folded forms of proteins) to the correct native conformation.

[0211] Like protein stabilization, protein folding, ligand selection, and drug design, selection of conditions that promote protein crystallization is another multi-variable optimization problem that is solved using the methods and the apparatus of the present invention.

[0212] The methods and the assay apparatus of the present invention are also useful for determining conditions that facilitate protein crystallization. The crystallization of molecules from solution is a reversible equilibrium process, and the kinetic and thermodynamic parameters are a function of the chemical and physical properties of the solvent system and solute of interest (McPherson, A., In: Preparation and Analysis of Protein Crystals,Wiley Interscience (1982); Weber, P. C., Adv. Protein Chem.41:1-36 (1991)) 1991). Under supersaturating conditions, the system is driven toward equilibrium where the solute is partitioned between the soluble and solid phase instead of the unfolded and native states. The molecules in the crystalline phase pack in ordered and periodic three dimensional arrays that are energetically dominated by many of the same types of cohesive forces that are important for protein folding, i.e. van der Waals interactions, electrostatic interactions, hydrogen bonds, and covalent bonds (Moore, W. J., In: Physical Chemistry,4th Ed., Prentice Hall, (1972), pp. 865-898).

[0213] Thus, in many ways protein crystallization can be viewed as a higher level variation of protein folding where whole molecules are packed to maximize cohesive energies instead of individual amino acid residues. Moreover, for both protein crystallization and protein folding, the composition of the solvent can make very important contributions to the extent of partitioning between the soluble (unfolded) and crystalline (native) forms. The cohesive interactions present in protein macromolecules and the role played by solvent in modulating these interactions for both protein folding and protein crystallization are complex and not fully understood at the present time. In this regard, biochemical conditions that promote protein stabililty and protein folding also promote protein crystallization.

[0214] For example, biochemical conditions that were found to increase the stability of D(II) FGF receptor 1 (FIGS. 19-24) correlate with the conditions that facilitated the crystallization of x-ray diffraction quality protein crystals. Conditions that were employed to obtain crystals of D(II) FGFR1 protein (Lewankowski, Myslik, Bone, R. Springer, B. A. and Pantoliano, M. W., unpublished results (1997)) are shown in Table 1. Protein crystals were obtained in the pH range 7.4 to 9.2 in the presence of the Hofmeister salt Li₂SO₄(65 to 72%). These crystallization conditions correlated with the pH optimum of about 8.0 in FIG. 23. Other salts of the Hofmeister series such as Na₂SO₄, (NH₄)₂SO₄and Mg₂SO₄were also found useful as additives for lowering the amount of Li₂SO₄required as the precipitant. Clearly, these conditions for successful D(II) FGFR1 crystallization correlate closely with the stabilizing conditions that were identified using the microplate thermal shift assay.

[0215] Conditions that were identified as facilitating human α-thrombin stabilization also facilitate human α-thrombin protein crystallization. FIGS. 17A-D and 18 show the results of microplate thermal shift assays of conditions that facilitate human α-thrombin stability. Table 2 contains a summary of the conditions identified by three different investigators that facilitate crystallization of x-ray diffraction quality human α-thrombin crystals (Bode, W., et al, Protein Sci.1:426-471 (1992); Vijayalakshmi, J. et al., Protein Sci.3:2254-22271 (1994); and Zdanov, A. et al., Proteins: Struct. Funct. Genet.17:252-265 (1993)).

[0216] The conditions summarized in Table 2 correlate closely with the conditions identified in the microplate thermal shift assay as facilitating human α-thrombin stability. Crystals formed near a pH optimum of about 7.0. Furthermore, there is a clear preference for the presence of 0.1 to 0.5 M NaCl (50% of the conditions) or 0.1 to 0.2 M NaHPO₄. This is consistent with the recently discovered Na⁺ binding site (Dang et al., Nature Biotechnology15:146-149 (1997)) and microplate thermal shift assay results in FIGS. 17A-D and 18. All of the human α-thrombin samples described in Table 2 that have yielded good crystals are complexed with a ligand, thereby further stabilizing the native structure of this protein beyond that acquired from the biochemical conditions.

[0217] Protein crystallization is a slow and tedious process that has historically been the rate determining step for the x-ray diffraction determination of protein and nucleic acid structures. The method and apparatus of the present invention facilitate the rapid, high-throughput elucidation of conditions that promote the stability of a given protein and thus the formation of X-ray quality protein crystals.

[0218] When a protein is more stable, it has fewer thermodynamic motions that inhibit packing into a crystal lattice. With fewer motions, the protein fits better into a crystal lattice. Using conventional crystallization methods, crystallization experiments are set up at room temperature for weeks at a time. Over time, protein unfolding occurs. Using the methods of the present invention, conditions that stabilize a protein are examined over a temperature range. Conditions that shift the thermal unfolding curve to higher temperature will lower extent of unfolding that occurs while the crystallization process occurs.

[0219] Overview of Assay Apparatus

[0220] The assay apparatus of the present invention is directed to an automated temperature adjusting and spectral emission receiving system that simultaneously adjusts the temperature of a multiplicity of samples over a defined temperature range and receives spectral emission from the samples. The assay apparatus of the present invention is particularly useful for performing microplate thermal shift assays of protein stability. The assay apparatus of the present invention can be used to practice all of the methods of the present invention.

[0221] The assay apparatus of the present invention replaces separate heating devices and spectral emission receiving devices. In contrast to other devices, the assay apparatus of the present invention can be configured to simultaneously adjust the temperature of a multiplicity of samples and receive spectral emissions from the samples during adjustment of temperature in accordance with a predetermined temperature profile.

[0222] After heat denaturation, reversibly folding proteins partially or fully refold after heat denaturation. Refolding precludes meaningful measurements in a thermal shift assay. Using the assay apparatus of the present invention, however, one can assay reversibly folding proteins in a thermal shift assay. That is because in the assay apparatus of the present invention, spectral measurements are taken while the protein is being heated. And protein refolding does not occur.

[0223] In such a configuration, the assay apparatus of the present invention includes a sensor which is positioned over a movable heat conducting block upon which an array of samples is placed. A relative movement means, such as a servo driven armature, is used to move the sensor so that the sensor is sequentially positioned over each sample in the array of samples. The sensor receives spectral emissions from the samples.

[0224] The assay apparatus of the present invention can be configured so that it contains a single heat conducting block. Alternatively, the assay apparatus can be configured so that it contains a plurality of heat conducting blocks upon a movable platform. The platform may be a translatable platform that can be translated, for example, by a servo driven linear slide device. An exemplary linear slide device is model SA A5M400 (LAI America, Torrance, Calif.). In this embodiment, the sensor receives spectral emissions from each of the samples on a given heat conducting block. The platform is then translated to place another heat conducting block and its accompanying samples under the sensor so that it receives spectral emissions from each of the samples on that heating block. The platform is translated until spectral emissions are received from the samples on all heat conducting blocks.

[0225] Alternatively, the platform may by a rotatable platform that may be rotated, for example, by a servo driven axle. In the latter embodiment, the sensor receives spectral emissions from each of the samples on a given heat conducting block. The platform is then rotated to place another heat conducting block and its accompanying samples under the sensor so that it receives spectral emissions from each of the samples on that heating block. The platform is rotated until spectral emissions are received from the samples on all heat conducting blocks.

[0226] System Description

[0227] FIG. 29 shows a schematic diagram of one embodiment of an assay apparatus 2900 of the present invention. Assay apparatus 2900 includes a heat conducting block 2912 that includes a plurality of wells 2920 for a plurality of samples 2910. Heat conducting block 2912 is composed of a material that has a relatively high coefficient of thermal conductivity, such as aluminum, stainless steel, brass, teflon, and ceramic.

[0228] Thus, heat conducting block 2912 can be heated and cooled to a uniform temperature but will not be thermally conductive enough to require excess heating or cooling to maintain a temperature.

[0229] Assay apparatus 2900 also includes a light source 2906 for emitting an excitatory wavelength of light, shown generally at 2916, for the samples. Light source 2906 excites samples 2910 with excitatory light 2916. Any suitable light source can be used. For example a tungsten-halogen lamp can be used. Alternatively, a Xenon-arc lamp, such as the Biolumin 960 (Molecular Dynamics) can used.

[0230] Alternatively, a high pressure mercury (Hg) Lamp can be used. High pressure mercury lamps emit light of higher intensity than Xenon (Xe) lamps. The intensity of light from a high pressure mercury lamp is concentrated in specific lines, and are only useful if the Hg lines are at suitable wavelengths for excitation of particular fluorophores.

[0231] Some fluorescent plate readers employ lasers for excitation in the visible region of the electromagnetic spectrum. For example, the FluorImager™ (Molecular Dynamics, Palo Alto, Calif.) is such a device. This technology is useful when using fluorescent dyes that absorb energy at around 480 nm and emit energy at around 590 nm. Such a dye could then be excited with the 488 nm illumination of standard argon, argon/krypton lasers. For example, 1,1-dicyano-2-[6-(di-methylamino)naphthalen-2-yl]propene (DDNP) is such a dye. The advantage in using a laser is that a laser is characterized by very high intensity light, which results in an improved signal to noise ratio.

[0232] Excitatory light 2916 causes a spectral emission 2918 from samples 2910. Spectral emission 2918 can be electromagnetic radiation of any wavelength in the electromagnetic spectrum. Preferably, spectral emission 2918 is fluorescent, ultraviolet, or visible light. Most preferably, spectral emission 2918 is fluorescence emission. Spectral emission 2918 is received by a photomultiplier tube 2904. Photomultiplier tube 2904 is communicatively and operatively coupled to a computer 2914 by an electrical connection 2902. Computer 2914 functions as a data analysis means for analyzing spectral emission as a function of temperature.

[0233] As discussed above, the spectral receiving means or sensor of the assay apparatus of the present invention can comprise a photomultiplier tube. Alternatively, the spectral receiving means or sensor can include a charge coupled device or a charge coupled device camera. In still another alternative, the spectral receiving means or sensor can include a diode array.

[0234] An alternate embodiment of the assay apparatus of the present invention is shown in FIG. 30. In the embodiment shown in FIG. 30, a charge coupled device (CCD) camera 3000 is used to detect spectral emission 2918 from samples 2910. CCD camera 3000 can be any CCD camera suitable for imaging fluorescent emissions. For example, suitable CCD cameras are available from Alpha-Innotech (San Leandro, Calif.), Stratagene (La Jolla, Calif.), and BioRad (Richmond, Calif.). For measuring fluorescent emission in the microplate thermal shift assay, one alternative to a fluorescent plate reader is a charge coupled device (CCD). For example, high resolution CCD cameras can detect very small amounts of electromagnetic energy, whether it originates from distance stars, is diffracted by crystals, or is emitted by fluorophores. A CCD is made of semi-conducting silicon. When photons of light fall on it, free electrons are released. As an electronic imaging device, a CCD camera is particularly suitable for fluorescence emission imaging because it can detect very faint objects, affords sensitive detection over a broad spectrum range, affords low levels of electromagnetic noise, and detects signals over a wide dynamic range—that is, a charge coupled device can simultaneously detect bright objects and faint objects. Further, the output is linear so that the amount of electrons collected is directly proportional to the number of photons received. This means that the image brightness is a measure of the real brightness of the object, a property not afforded by, for example, photographic emulsions.

[0235] When a fluorescence imaging camera or a CCD camera is used, excitatory light 2916 can be a suitable lamp that is positioned over the plurality of samples 2910. Alternatively, excitatory light 2916 can be a suitable lamp that is positioned under the plurality of samples 2910. In another alternative embodiment, excitatory light 2916 can be delivered to each sample 2910 by a plurality of fiber optic cables. Each fiber optic cable is disposed through one of a plurality of tunnels in conducting block 2912. Thus, each of samples 2910 receives excitatory light 2916 through a fiber optic cable.

[0236] As shown in FIG. 30, source 2906 excites samples 2910 with excitatory light 2916. Excitatory light 2916 causes spectral emission 2918 from samples 2910. Spectral emission 2918 is filtered through an emission filter 3002. Emission filter 3002 filters out wavelengths of spectral emission 2918 that are not to be monitored or received by CCD camera 3000. CCD camera 3000 receives the filtered spectral emission 2918 from all of samples 2910 simultaneously. For simplicity and ease of understanding, only spectral emissions form one row of samples 2910 are shown in FIG. 30. CCD camera 3000 is communicatively and operatively coupled to computer 2914 by electrical connection 2902.

[0237] With reference now to FIG. 31, one embodiment of assay apparatus 2900 is shown in more detail. As shown in FIG. 31, many apparatus components are attached to a base 3100. A heat conducting block relative movement means 3128 is used to move heat conducting block 2912 in directions 3150 and 3152. Heat conducting block relative movement means 3128 is communicatively and operatively connected to a servo controller 3144. Activation of heat conducting block relative movement means 3128 by servo controller 3144 moves heat conducting block 2912 in directions 3150 and 3152. Servo controller 3144 is controlled by a computer controller 3142. Alternatively, computer 2914 could be used to control servo controller 3144.

[0238] A sensor is removably attached to a sensor armature 3120. An exemplary sensor is a fiber optic probe 3122. Fiber optic probe 3122 includes a fiber optic cable capable of transmitting receiving excitatory light 2916 to samples 2910, and a fiber optic cable capable of receiving spectral emission 2918 from samples 2910. Electromagnetic radiation is transmitted from excitatory light source 2906 to fiber optic probe 3122 by excitatory light input fiber optic cable 3108. In one embodiment of the present invention, a spectral receiving means comprising photomultiplier tube 2904 is used to detect spectral emission from samples 2910. In this embodiment, electromagnetic radiation is transmitted from fiber optic probe 3122 to photomultiplier tube 2904 by fiber optic cable 3110. In an alternative embodiment of the present invention, CCD camera 3002 is used to detect spectral emission from samples 2910. In this embodiment, fiber optic cable 3110 is not required.

[0239] A temperature sensor 3124 is removably attached to sensor armature 3120. Temperature sensor 3124 is communicatively and operably linked to a temperature controller 3162. Temperature sensor 3124 monitors the temperature of heat conducting block 2912 and feeds temperature information back to temperature controller 3162. Temperature controller 3162 is connected to heat conducting block 2912 by a thermoelectric connection 3164. Under the action of temperature controller 3162, the temperature of heat conducting block 2912 can be increased, decreased, or held constant. Particularly, the temperature of heat conducting block 2912 can be changed by temperature controller 3162 in accordance with a pre-determined temperature profile. Preferably, temperature computer controller 3162 is implemented using a computer system such as that described below with respect to FIG. 37.

[0240] As used herein, the term “temperature profile” refers to a change in temperature over time. The term “temperature profile” encompasses continuous upward or downward changes in temperature, both linear and non-linear changes. The term also encompasses any stepwise temperature change protocols, including protocols characterized by incremental increases or decreases in temperature during which temperature increases or decreases are interrupted by periods during which temperature is maintained constant. In the apparatus of the present invention, the temperature profile can be pre-determined by programming temperature computer controller 3162. For example, temperature profiles can be stored in a memory device of temperature controller 3162, or input to temperature controller 3162 by an operator.

[0241] A sensor armature relative movement means 3130 is used to move sensor armature 3120 in directions 3154 and 3156. A sensor armature servo controller 3118 is fixedly connected to excitatory light filter housing 3160. Activation of sensor armature servo controller 3118 moves fiber optic probe 3122 in directions 3154 and 3156. It would be readily apparent to one of ordinary skill in the relevant art how to configure servo controllers to move heat conducting block 2912 and sensor armature 3120. It should be understood that the present invention is not limited to the use of servo controllers for movement of heat conducting block 2912 and sensor armature 3120, and other suitable means known to one of skill in the art can also be used, such as a motor.

[0242] Servo controllers 3118 and 3144 are both communicatively and operatively connected to computer controller 3142. Computer controller 3142 controls the movement of sensor armature 3120 in directions 3154 and 3156. In addition, computer controller 3142 controls the movement of heat conducting block relative movement means 3128 in directions 3150 and 3152.

[0243] In the assay apparatus of the present invention, excitatory light source 2906 is used to excite samples 2910. Excitatory light source 2906 is communicatively and operably connected to excitatory light filter 3104, which is contained within excitatory light filter housing 3160. Excitatory light filter 3104 filters out all wavelengths of light from excitatory light source 2906 except for the wavelength(s) of light that are desired to be delivered by fiber optic probe 3122 to samples 2910. An excitatory light filter servo controller 3106 controls the aperture of excitatory light filter 3104. Excitatory light source 2906 and excitatory light filter servo controller 3106 are communicatively and operatively connected to excitatory light computer controller 3102. Computer controller 3102 controls the wavelength of excitatory light transmitted to samples 2910 by controlling excitatory light filter servo controller 3106. Excitatory light 2916 is transmitted through excitatory light input fiber optic cable 3108 to fiber optic probe 3122 for transmission to samples 2912.

[0244] Spectral emission 2918 from samples 2910 is received by fiber optic probe 3122 and is transmitted to a spectral emission filter 3114 by output fiber optic cable 3110. Spectral emission filter 3114 is contained within a spectral emission filter housing 3166. Spectral emission filter housing 3166 is disposed on photomultiplier tube housing 3168. Photomultiplier tube housing 3168 contains photomultiplier tube 2904. A spectral emission servo controller 3112 controls the aperture of spectral emission filter 3114, thereby controlling the wavelength of spectral emission 2918 that is transmitted to photomultiplier tube 2904. Spectral emission servo controller 3112 is controlled by a computer controller 3170.

[0245] Spectral emission 2918 from samples 2910 is transmitted from photomultiplier tube 2904. Electrical output 3140 connects photomultiplier tube 2904 to electric connection 2902. Electric connection 2902 connects electrical output 3140 to computer 2914. Driven by suitable software, computer 2914 processes the spectral emission signal from samples 2910. Exemplary software is a graphical interface that automatically analyzes fluorescence data obtained from samples 2910. Such software is well known to those of ordinary skill in the art. For example, the CytoFluor™II fluorescence multi-well plate reader (PerSeptive Biosystems, Framingham, Mass.) utilizes the Cytocalc™ Data Analysis System (PerSeptive Biosystems, Framingham, Mass.). Other suitable software includes, MicroSoft Excel or any comparable software.

[0246] FIGS. 32A-C illustrate one embodiment of a thermal electric stage or heat conducting block for the assay apparatus of the present invention. FIG. 32A shows a side view of heat conducting block 2912 and a heat conducting wire 3206. FIG. 32B shows a top view of heat conducting block 2912 and heat conducting wire 3206. Heat conducting wire 3206 is a temperature adjusting element that adjusts the temperature of heat conducting block 2912. By means readily known to one of skill in the art, temperature controller 3162 causes heat conducting wire 3206 to increase or decrease in temperature, thereby changing the temperature of heat conducting block 2912. For example, an exemplary temperature controller is a resistance device that converts electric energy into heat energy. Alternatively, the heating element can be a circulating water system, such as that disclosed in U.S. Pat. No. 5,255,976, the content of which is incorporated herein by reference. In another alternative, the temperature adjusting element can be a heat conducting surface upon which heat conducting block 2912 is disposed. Particularly, the temperature of heat conducting wire 3206 can be changed by temperature controller 3162 in accordance with a pre-determined temperature profile. Temperature controller 3162 is preferably implemented using a computer system such as that described below with respect to FIG. 37. Alternatively, computer 2914 could be used to implement temperature controller 3162. An exemplary set of specifications for temperature controller 3162 and heat conducting block 2912 is as follows:

[0247] Temperature controller 3162 changes temperature in accordance with a temperature profile as discussed below with respect to FIGS. 36A and 36B.

[0248] The temperature of heat conducting block 2912 can be controlled such that a uniform temperature is maintained across the heat conducting block. Alternatively, the temperature of heat conducting block 2921 can be controlled such that a temperature gradient is established from one end of the heat conducting block to the other. Such a technique is disclosed in U.S. Pat. Nos. 5,255,976 and 5,525,300, the entirety of both of which is incorporated by reference.

[0249] Heat conducting block 2912 is preferably configured with plurality of wells 2929 for samples 2910 to be assayed. In one embodiment, each of wells 2929 is configured to receive a container containing one of plurality of samples 2910. Alternatively, heat conducting block 2912 is configured to receive a container containing plurality of samples 2910. An exemplary container for containing plurality of samples 2910 is a microtiter plate.

[0250] In yet a further alternate embodiment, heat conducting block 2912 is configured to receive a heat conducting adaptor that is configured to receive a container containing one or more of samples 2910. The heat conducting adaptor is disposed on heat conducting block 2912, and the container containing samples 2910 fits into the heat conducting adaptor. FIGS. 32C-E show three exemplary configurations of a heat conducting adaptor. An adaptor 3200 is configured with round-bottomed wells. An adaptor 3202 is configured with square-bottom wells. An adaptor 3204 is configured with V-shaped wells. For example, adaptor 3200 can receive a plurality of round-bottom containers, each containing one sample. Similarly, adaptor 3202 can receive a plurality of square-bottom containers, and adaptor 3204 can contain a plurality of V-shaped bottom containers. Adaptors 3200, 3202, and 3204 can also receive a carrier for a multiplicity of round-bottom containers. An exemplary carrier is a microtitre plate having a plurality of wells, each well containing a sample. When heat conducting block 2912 is heated, heat conducting adaptors 3200, 3202, or 3204 are also heated. Thus, the samples contained in the containers that fit within adaptors 3200, 3202, or 3204 are also heated. Adaptors 3200, 3202, and 3204 can accept standard microplate geometries.

[0251] Another embodiment of the assay apparatus of the present invention is shown in FIG. 33. In this embodiment, a plurality of heat conducting blocks 2912 is mounted on a rotatable platform or carousel 3306. Alternativeley, the platform can be a translatable platform. Platform or carousel 3306 can be composed of a heat conducting material, such as the material that heat conducting block 2912 is composed of. Although six heat conducting blocks are shown in FIG. 33, this number is exemplary and it is to be understood that any number of heat conducting blocks can be used. As shown in FIG. 33, an axle 3308 is S rotatably connected to base 3100. Rotatable platform 3306 is axially mounted to rotate about axle 3308. Rotation of axle 3308 is controlled by a servo controller 3312. Servo controller 3312 is controlled by a computer controller 3314 in a manner well known to one of skill in the relevant arts. Computer controller 3314 causes servo controller 3312 to rotate axle 3308 thereby rotating rotatable platform 3306. In this manner, heat conducting blocks 2912 are sequentially placed under fiber optic probe 3122.

[0252] Each of the plurality of heat conducting blocks 2912 can be controlled independently by temperature controller 3162. Thus, the temperature of a first heat conducting block 2912 can be higher or lower than the temperature of a second heat conducting block 2912. Similarly, the temperature of a third heat conducting block 2912 can be higher or lower than the temperature of either first or second heat conducting block 2912.

[0253] In a manner similar to that described above for FIG. 31, relative movement means 3130 is also used to move sensor armature 3120 in directions 3150 and 3152 so that fiber optic probe 3122 can be moved to detect spectral emission from samples 2910. A second sensor armature relative movement means 3316 is used to move sensor armature 3120 in directions 3154 and 3156.

[0254] The temperature of heat conducting blocks 2912 is controlled by temperature controller 3162. Temperature controller 3162 is connected to rotatable platform 3306 by connection 3164 to heat conducting blocks 2912. Under the action of temperature controller 3162, the temperature of heat conducting blocks 2912 can be increased and decreased. Alternatively, temperature controller 3162 can be configured to adjust the temperature of rotatable platform 3306. In such a configuration, when rotatable platform 3306 is heated, heat conducting blocks 2912 are also heated. Alternatively, the temperature of each of heat conducting blocks 2912 can be controlled by a circulating water system such as that noted above.

[0255] In a manner similar to that illustrated in FIG. 31, excitatory light source 2906 is used to excite samples 2910. Excitatory light source 2906 is communicatively and operably connected to excitatory light filter 3104, which is contained within excitatory light filter housing 3160. Excitatory light filter 3104 filters out all wavelengths of light from excitatory light source 2906 except for the wavelength(s) of light that are desired to be delivered by fiber optic probe 3122 to samples 2910. An excitatory light filter servo controller 3106 controls the aperture of excitatory light filter 3104. Excitatory light source 2906 and excitatory light filter servo controller 3106 are communicatively and operatively connected to excitatory light computer controller 3102. Computer controller 3102 controls the wavelength of excitatory light transmitted to samples 2910 by controlling excitatory light filter servo controller 3106. Excitatory light 2916 is transmitted through excitatory light input fiber optic cable 3108 to fiber optic probe 3122 for transmission to samples 2912.

[0256] Spectral emission 2918 from samples 2910 is received by fiber optic probe 3122 and is transmitted to spectral emission filter 3114 by fiber optic cable 3110. Spectral emission servo controller 3112 controls spectral emission filter 3114 aperture and thus controls the wavelength of spectral emission that is transmitted to photomultiplier tube 2904. In a manner similar to that explained for FIG. 31, spectral emission servo controller 3112 is controlled by computer controller 3170.

[0257] The assay apparatus of the present invention can detect spectral emission from samples 2910 one sample at a time or simultaneously from a subset of samples 2910. As used herein, the term “subset of samples” refers to at least two of samples 2910. To detect spectral emission simultaneously from a subset of samples in an embodiment of the assay apparatus of the present invention comprising photomultiplier tube 2904, a plurality of excitatory light filters 3104, excitatory light input fiber optic cables 3108, emission light output fiber optic cables 3110, and emission light filters 3114 must be used.

[0258] The spectral emission signal is transmitted from photomultiplier tube 2904 to computer 2914. Photomultiplier tube 2904 is communicatively and operatively coupled to computer 2914 by electrical connection 2902. Connection 2902 is connected to photomultiplier tube 2904 through electrical output 3140. Computer 2914 functions as a data analysis means for analyzing spectral emission as a function of temperature.

[0259] FIG. 34 illustrates a top view of the assay apparatus shown in FIG. 33 with a housing 3400 that covers the apparatus. A door 3402 opens to reveal samples 2910. Door 3402 can be a hinge door that swings open. Alternatively, door 3402 can be a sliding door that slides open. A side view of the assay apparatus shown in FIGS. 33 and 34 is illustrated in FIG. 35. Cover 3400 is disposed on top of base 3100. Cover 3400 can be made of any suitable material. For example, cover 3400 can be made of plexiglass, fiberglass, or metal.

[0260] FIGS. 36A and 36B illustrate a temperature profile and how the temperature profile is implemented using the assay apparatus of the present invention. FIG. 36A illustrates a temperature profile 3600 that shows the temperature of heat conducting blocks 2912 as a function of time. Heat conducting blocks 2912 and samples 2910 are heated in a continuous fashion in accordance with temperature profile 3600. Alternatively, rotatable platform 3306 can be heated along with heat conducting blocks 2912. Preferably, temperature profile 3600 is linear, with temperatures ranging from about 25° C. to about 110° C.

[0261] Alternatively, temperature profile 3600 can be characterized by incremental, stair step increases in temperature, in which heat conducting blocks 2912 and samples 2910 are heated to a predetermined temperature, maintained at that temperature for a predetermined period of time, and than heated to a higher pre-determined temperature. For example, temperature can be increased from 0.5° C. to 20° C. per minute. Although the temperature range from about 25° C. to about 110° C. is disclosed, it is to be understood that the temperature range with which a given target molecule, for example, a protein, is to be heated to generate a thermal denaturation curve can readily be determined by one of ordinary skill in the art. The length of time over which temperature profile 3600 is accomplished will vary, depending on how many samples are to be assayed and on how rapidly the sensor that receives spectral emission 2918 can detect spectral emission 2918 from samples 2910. For example, an experiment in which each of six heat conducting blocks 2912 holds a total of 96 samples 2910 (for a total of 576 samples), and in which samples are scanned using a fluorescent reader device having a single fiber optic probe, and in which the temperature profile is from 38° C. and 76° C., would take approximately 38 minutes to perform using the apparatus shown in FIG. 33.

[0262] While heating in accordance with temperature profile 3600, spectral emission 2918 from each sample 2910 in a first heat conducting block 2912 is received through fiber optic probe 3122. As illustrated in FIG. 36B, after emissions from all of samples 2910 in first heat conducting block 2912 have been received, platform 3306 is rotated to move the next heat conducting block 2912 under fiber optic probe 3122 and spectral emission 2918 from samples 2910 is received by fiber optic probe 3122. This process is continued until reception of spectral emissions from all samples in all heat conducting blocks 2912 is complete. Spectral emission from samples 2910 on each heat conducting block 2912 can be received one at a time, simultaneously from a subset of samples, simultaneously from one row of samples at a time, or all of the samples at one time.

[0263] Computer Program Implementation of the Preferred Embodiments

[0264] The present invention may be implemented using hardware, software, or a combination thereof, and may be implemented in a computer system or other processing system. A flowchart 3800 for implementation of one embodiment of the present invention is shown in FIG. 38. Flowchart 3800 begins with a start step 3802. In a step 3804, temperature profile 3600 is initiated. For example, temperature controller 3162 causes the temperature of heat conducting block 2912 to increase. In a step 3806, a sensor such as fiber optic probe 3122 or CCD camera 3000 is moved over a sample 2910, row of samples 2910, or all of samples 2910. In a step 3808, excitatory light is transmitted to sample(s) 2910 using excitatory light source 2906. In a step 3810, spectral emission is received by the sensor from sample(s) 2910. In a decision step 3812, it is determined whether spectral emission 2918 has been received from all of the samples, rows of samples, in one heat conducting block 2912. If spectral emission 2918 has not been received from all of the samples or rows of samples, the sensor is moved over the next sample or row of samples in a step 3814. Processing then continues at step 3808 to transmit excitatory light 2916. Processing then continues to a step 3810 to receive spectral emission 2918 from sample(s) 2910.

[0265] If spectral emission 2918 has been received from all of samples or rows of samples, processing continues to a decision step 3816. In decision step 3816, it is determined whether spectral emission 2918 has been received from samples in all heat conducting blocks. If not, rotatable platform 3306 is rotated in a step 3818 to place the next heat conducting block 2912 and samples 2910 contained therein under the sensor. Steps 3806 through 3818 are followed until spectral emission 2918 has been received from all of the samples in all of heat conducting blocks 2912. Processing then continues to a step 3820, in which temperature profile 3600 is completed and processing ends at a step 3822.

[0266] A flowchart 3900 for implementation of an alternate embodiment of the present invention is shown in FIG. 39. In this embodiment, a sensor for simultaneously receiving spectral emission 2918 from all of samples 2910 on heat conducting block 2912, such as CCD camera 3000, is positioned over heat conducting block 2912. Flowchart 3900 begins with a start step 3902. In a step 3904, temperature profile 3600 is initiated. For example, temperature controller 3162 causes the temperature of heat conducting block 2912 to increase. In a step 3906, excitatory light is transmitted to sample(s) 2910 using excitatory light source 2906. In a step 3908, spectral emission is received by CCD camera 3000 from sample(s) 2910. In a decision step 3910, it is determined whether spectral emission 2918 has been received from all of heat conducting blocks 2912. If not, rotatable platform 3306 is rotated in a step 3912 to place the next heat conducting block 2912 and samples 2910 contained therein under CCD camera 3000. Steps 3906 through 3912 are followed until spectral emission 2918 has been received from samples 2910 in all of heat conducting blocks 2912. Processing then continues to a step 3914. In step 3914, temperature profile 3600 is completed and processing ends at a step 3916.

[0267] As stated above, the present invention may be implemented using hardware, software, or a combination thereof, and may be implemented in a computer system or other processing system. An exemplary computer system 3702 is shown in FIG. 37. Computer controllers 3102, 3142, 3162, 3170, or 3314, can be implemented using one or more computer systems such as computer system 3702.

[0268] After reading this description, it will become apparent to a person skilled in the relevant art how to implement the invention using other computer systems and/or computer architectures. Computer system 3702 includes one or more processors, such as processor 3704. Processor 3704 is connected to a communication bus 3706.

[0269] Computer system 3702 also includes a main memory 3708, preferably random access memory (RAM), and can also include a secondary memory 3710. The secondary memory 3710 can include, for example, a hard disk drive 3712 and/or a removable storage drive 3714, representing a floppy disk drive, a magnetic tape drive, an optical disk drive, etc. The removable storage drive 3714 reads from and/or writes to a removable storage unit 3716 in a well known manner. Removable storage unit 3716 represents a floppy disk, magnetic tape, optical disk, etc. which is read by and written to by removable storage drive 3714. As will be appreciated, the removable storage unit 3716 includes a computer usable storage medium having stored therein computer software and/or data.

[0270] In alternative embodiments, secondary memory 3710 may include other similar means for allowing computer programs or other instructions to be loaded into computer system 3702. Such means can include, for example, a removable storage unit 3718 and an interface 3720. Examples of such can include a program cartridge and cartridge interface (such as that found in video game devices), a removable memory chip (such as an EPROM, or PROM) and associated socket, and other removable storage units 3718 and interfaces 3720 which allow software and data to be transferred from the removable storage unit 3718 to computer system 3702.

[0271] Computer system 3702 can also include a communications interface 3722. Communications interface 3722 allows software and data to be transferred between computer system 3702 and external devices. Examples of communications interface 3722 can include a modem, a network interface (such as an Ethernet card), a communications port, a PCMCIA slot and card, etc. Software and data transferred via communications interface 3722 are in the form of signals 3724 which can be electronic, electromagnetic, optical or other signals capable of being received by communications interface 3722. These signals 3724 are provided to communications interface via a channel 3726. This channel 3726 carries signals 3724 and can be implemented using wire or cable, fiber optics, a phone line, a cellular phone link, an RF link and other communications channels. In the assay apparatus of the present invention, one example of channel 3726 is electrical connection 2902 that carries signal 3724 of spectral emission 2918 to computer 2914.

[0272] In this document, the terms “computer program medium” and “computer usable medium” are used to generally refer to media such as removable storage device, 3716 and 3718, a hard disk installed in hard disk drive 3712, and signals 3724. These computer program products are means for providing software to computer system 3702.

[0273] Computer programs (also called computer control logic) are stored in main memory 3708 and/or secondary memory 3710. Computer programs can also be received via communications interface 3722. Such computer programs, when executed, enable the computer system 3702 to perform the features of the present invention as discussed herein. In particular, the computer programs, when executed, enable the processor 3704 to perform the features of the present invention. Accordingly, such computer programs represent controllers of the computer system 3702.

[0274] In an embodiment where the invention is implemented using software, the software may be stored in a computer program product and loaded into computer system 3702 using removable storage drive 3714, hard drive 3712 or communications interface 3722. The control logic (software), when executed by the processor 3704, causes the processor 3704 to perform the functions of the invention as described herein.

[0275] In another embodiment, the invention is implemented primarily in hardware using, for example, hardware components such as application specific integrated circuits (ASICs). Implementation of the hardware state machine so as to perform the functions described herein will be apparent to persons skilled in the relevant art(s).

[0276] In yet another embodiment, the invention is implemented using a combination of both hardware and software.

[0277] The assay apparatus of the present invention is particularly suited for carrying out the methods of the present invention. To conduct a microplate thermal shift assay using the method and apparatus of the present invention, samples are placed in a heat conducting block, heated according to a predetermined temperature profile, stimulated with an excitatory wavelength of light, and the spectral emission from the samples is detected while the samples are being heated in accordance with the pre-determined temperature profile.

[0278] It is to be understood that the assay apparatus of the present invention is not limited to use with the methods of the present invention or limited to conducting assays on biological polymers, proteins, or nucleic acids. For example, the assay apparatus of the present invention can be used to incubate samples to a predetermined temperature. Alternatively, the assay apparatus of the present invention can be used to perform polymerase chain reaction, thermal cycling steps for any purpose, assaying thermal stability of a compound, such as a drug, to determine conditions that stabilize a compound, or to determine conditions that facilitate crystallization of a compound.

[0279] Having now generally described the invention, the same will become more readily understood by reference to the following specific examples which are included herein for purposes of illustration only and are not intended to be limiting unless otherwise specified.

[0280] Using the computer controlled process DirectedDiversity® (see U.S. Pat. No. 5,463,564), scientists at 3-Dimensional Pharmaceuticals, Inc. have generated a combinatorial library of compounds directed at the active site of human α-thrombin. Approximately 400 compounds were synthesized and assayed by a conventional spectrophotometric kinetic assay in which succinyl-Ala-Ala-Pro-Arg-p-nitroanilide (Bachem, King of Prussia, PA) served as substrate. Five of these compounds, which are characterized by K_i's that span almost four orders of magnitude in binding affinity, were used to test the range and limits of detection of the thermal shift assay. These five proprietary compounds are listed in Table 3, along with the K_ifor each respective compound, as measured by the kinetic assay (last column). K_i's for these compounds ranged from 7.7 nM for 3 dp-4026 to 20.0 μM for 3 dp-3811.

[0281] A stock human α-thrombin solution (1.56 mg/mL) from Enzyme Research Labs was first diluted to 0.5 mg/mL (11 μM) with 50 mM Hepes, pH 7.5, 0.1 M NaCl (assay buffer, unless mentioned otherwise), and stored on ice. The five ligands (recrystallized solids characterized by mass spectrometry and NMR) were accurately weighed out to be 1.5 to 2.0 mg and dissolved in 1.0 mL of 100% DMSO so that the concentration was between 1.8 and 3.8 mM. A 96 well V-bottom Costar microplate was then set up such that 100 μL of the 11 μM human α-thrombin solution was pipetted into wells A1 through A6. This was followed by the addition of 2 μL of 3 dp-3811 into well A2, 2 μL of 3 dp-3959 into well A3, 2 μL of 3 dp4077 into well A4, 2 μL of 3 dp-4076 into well A5, 2 μL of 3 dp4026 into well A6, and 2 μL of 100% DMSO into control well A1. The contents were mixed by repeated uptake and discharge using a 100 μL pipette tip. Finally, one drop of mineral oil (Sigma, St. Louis, Mo.) was added on top of the wells to reduce evaporation of samples at elevated temperatures. The microplate was then placed on heating block 4 of a RoboCycler Gradient 96 Temperature Cycler (Stratagene, La Jolla, Calif.), set at 25° C., for 1 minute. The plate was then placed into a SPECTRAmax™ 250 spectrophotometer (set to 30° C.) and the absorbance at 350 nm was measured for each sample. This reading served as the blank or reference from which all the other readings at higher temperatures were compared. The assay was initiated by setting heating block 1 to 38° C., programming the temperature cycler to move the microplate to heating block 1, and keeping the microplate there for 3 minutes. Following the equilibration at 38° C., the plate was moved to the 25° C. block (Block 4) for 30 seconds, inserted in the spectrophotometer, and absorbance was read at 350 nm. The microplate was then put back into the temperature cycler and was moved to heating block 2, which had been pre-equilibrated at 40° C. After 3 minutes at 40° C., the plate was returned to 25° C. (on block 4) for 30 seconds, and was returned to the spectrophotometer for a measurement of absorbance at 350 nm. This process was repeated 18 more times until the temperature had been raised to 76° C. in 2° C. increments. After subtraction of the blank absorbance (A₃₅₀at 25° C.), turbidity, reflected in the absorbance value, was plotted as a function of temperature. The thermal denaturation curves for this experiment are shown in FIG. 1.

[0282] The control (in well A1), which contained only 11 μM human α-thrombin in 2% DMSO, was found to undergo a thermal denaturation transition starting at ˜50° C., as reflected in the large increase in A₃₅₀. The midpoint in this transition was observed to be ˜55° C. This result was consistent with differential scanning calorimetric measurements for bovine prothrombin 1, which revealed a denaturation transition at T_m=58° C. (Lentz, B. P et al., Biochemistry 33:5460-5468 (1994)). The thermal denaturation curves for all of the tested inhibitor compounds displayed a shift in the transition towards higher temperatures. 3 dp-4026 showed the largest shift in T_m: ˜9° C. This result is consistent with the fact that, among the compounds tested, 3 dp-4026 exhibited the greatest binding affinity, as judged by kinetic measurements with succinyl-Ala-Ala-Pro-Arg-p-nitroanilide as substrate. Indeed, the rank order of shifts in T_m, shown in FIG. 1, paralleled the order of binding affinity as measured by conventional enzymology. These results indicate that by simply observing the shift in T_mfor a series of compounds relative to the control, one can easily and correctly rank a series of compounds in increasing order of binding affinity to the protein of interest.

[0283] It was possible to take the microplate thermal shift assay one step further and estimate the binding affinity of each ligand at T_m. This was done by substituting T₀, T_m, ΔH_uand ΔC_puinto equation (1). If ΔH_uand ΔC_pucannot be measured because a calorimetric device is not available, one can make educated guesses at ΔH_uand ΔC_pufor the therapeutic target. In the case of human α-thrombin, it was possible to use ΔH_u=200.0 kcal/mol, a value measured for the closely related protein bovine prothrombin 1 (Lentz, B. R. et al., Biochemistry 33:5460-5468 (1994)). A value of ΔC_pu=2.0 kcal/mol-° K was used to calculate K_Lat T_msince similar proteins of this size have been shown to yield similar values. The binding affinities at T_mof the five test ligands closely paralleled the K_i's measured with a spectrophotometric substrate (Table 3).

[0284] Assays for ligands that bind to the heparin binding site of human α-thrombin are more difficult to perform than assays for ligands that bind to the active site of human α-thrombin. At the heparin binding site, no substrate is hydrolyzed, so no spectrophotometric signal can be amplified for instrumental detection. Heparin activity is usually estimated in biological clotting time assays. Alternatively, heparin binding affinity for human α-thrombin can be determined by laboriously conducting 15 to 20 single point assays, in which the concentration of low MW heparin is varied over two logs, and monitoring the quenching of the fluorescent probe, p-aminobenzamidine, bound to the active site of human α-thrombin (Olson, S. T. et al., J. Biol. Chem.266:6342-6352 (1991)). Thus, heparin binding to human α-thrombin represents the kind of challenge encountered with the vast majority of non-enzyme receptor/ligand binding events, which are commonly observed for hormone/receptor interactions, repressor/DNA interactions, neurotransmitter/receptor interactions, etc. Several heparin-like sulfated oligosaccharides and sulfated naphthalene compounds were assayed by the microplate thermal shift assay. Using the thermal shift assay, it was possible to use a single compound per well to quickly rank the compounds in order of increasing binding affinity, with K_d's ranging over three orders of magnitude (see Table 4). Like the experiment in Example 1, the thermal shift assay results agreed closely with the results obtained through an alternative method, which required a series of laborious (15 to 20 single determinations) fluorescence quench assays over a wide range of concentrations of low MW heparin (Olson, S. T. et al., J. Biol. Chem. 266:6342-6352 (1991)). These results confirm that by simply observing the shift in T_mfor a series of compounds, relative to the control, one can easily and correctly rank a series of compounds in increasing order of binding affinity for the protein of interest.

[0285] A search of the literature did not locate alternatively measured binding results for the other ligands, which may attest to the difficulty of these experiments. However, the literature did reveal that pentosan polysulfate (PSO₄) (Sigma, St. Louis, Mo.), dextran SO₄(Sigma, St. Louis Mo.), and suramin (CalBiochem, LaJolla, Calif.) have been observed to have anticoagulant properties. Indeed, pentosan polysulfate and suramin were tested previously in clinical trials for anti-angiogenic activity, but were discounted due to toxic effects, many of which were described as coagulation anomalies (Pluda, J. M. et al., J. Natl. Cancer Inst. 85:1585-1592 (1993); Stein, C. A., Cancer Res. 53:2239-2249 (1993)). The affinities of pentosan PSO₄and suramin at T_m, as measured by the thermal shift assay, were found to be 7-fold and 5700-fold higher, respectively, than the affinity of heparin 5000 (Table 4). These results suggested that these ligands may alter clotting rates by interfering with the heparin mediated binding of human α-thrombin to anti-thrombin III (AT III), a protein co-factor for human α-thrombin activity.

[0286] The results in Table 4 revealed another advantage of the microplate thermal shift assay for screening compound libraries: the process is blind and unbiased in the sense that it detects ligand binding regardless of whether it is at the active site, an allosteric cofactor binding site, or at a protein subunit interface. The ability to detect ligands that bind with high affinity to sites outside an enzyme's active site will greatly facilitate discovery of lead molecules.

[0287] The second therapeutic receptor tested in the microplate thermal shift assay was acidic fibroblast growth factor (aFGF), a growth factor that plays a key role in angiogenesis (Folkman, J. et al., J. Biol. Chem. 267:10931-10934 (1992)). A synthetic gene for this protein was purchased from R&D Systems (Minneapolis, Minn.), and was cloned and expressed in E. coliusing methods similar to those described for basic fibroblast growth factor (bFGF) (Thompson, L. D. et al., Biochemistry 33:3831-3840 (1994); Pantoliano, M. W. et al., Biochemistry 33:10229-10248 (1994); Springer, B. A. et al., J. Biol. Chem. 269:26879-26884 (1994)). Recombinant aFGF was then purified by heparin-sepharose affinity chromatography as described (Thompson, L. D. et al., Biochemistry 33:3831-3840 (1994)). aFGF is also known to bind heparin/heparan, which is a cofactor for mitogenic activity. Heparin-like molecules, such as pentosan PSO₄and suramin, inhibit the growth factor's biological activity. A microplate thermal assay of these compounds was set up in a way similar to that described above for human α-thrombin. The change in turbidity, as a function of temperature, for each of the ligands suramin, heparin 5000, and pentosan PSO₄, is shown in FIG. 2. The results are summarized in Table 5. The affinity constants covered a fairly broad range of binding affinities, with pentosan PSO₄showing the highest affinity. The order of ligand binding affinity of pentosan PSO₄, heparin 5000 and suramin paralleled that found for bFGF, as measured using isothermal titrating calorimetry (Pantoliano, M. W. et al., Biochemistry 33:10229-10248 (1994)). The lack of alternatively measured binding affinities for these compounds probably attests to the difficulty of making these measurements using assays which do not monitor physical, temperature-dependent changes.

[0288] The results in Table 5 are consistent with the results in Tables 3 and 4. Simply observing the shift in T_mfor a series of compounds relative to the control, one can easily and correctly rank a series of compounds in increasing order of binding affinity to the protein of interest.

[0289] The microplate thermal shift assay was used to assess ligands for binding to the heparin binding site of basic fibroblast growth factor (bFGF). The gene for bFGF was purchased from R&D Systems and was cloned and expressed in E. colias previously described (Thompson, L. D. et al., Biochemistry 33:3831-3840 (1994); Pantoliano, M. W. et al., Biochemistry 33:10229-10248 (1994); Springer, B. A. et al., J. Biol. Chem. 269:26879-26884 (1994)). It was found that pentosan PSO₄and s bound to bFGF with binding affinities of 55 nM and 3.5 μM, respectively. This result for PSO₄compared very well with the affinity of 88 nM observed for PSO₄binding to bFGF, as determined by isothermal titrating calorimetry.

[0290] Because fluorescence measurements are more sensitive than absorbance measurements, a fluorescence thermal shift assay was used to assess ligand binding to human α-thrombin. The fluorescence emission spectra of many fluorophores are sensitive to the polarity of their surrounding environment and therefore are effective probes of phase transitions for proteins (i.e., from the native to the unfolded phase). The most studied example of these environment dependent fluorophores is 8-anilinonaphthalene-1-sulfonate (1,8-ANS), for which it has been observed that the emission spectrum shifts to shorter wavelengths (blue shifts) as the solvent polarity decreases. These blue shifts are usually accompanied by an increase in the fluorescence quantum yield of the fluorophore. In the case of ANS, the quantum yield is 0.002 in water and increases to 0.4 when ANS is bound to serum albumin.

[0291] ANS was used as a fluorescence probe molecule to monitor protein denaturation. In the fluorescence assay, the final concentration of human α-thrombin was 0.5 μM, which is 20-fold more dilute than the concentrations used in the turbidity assays. This concentration of human α-thrombin is in the range used for the kinetic screening assays.

[0292] ANS was excited with light at a wavelength of 360 nm. The fluorescence emission was measured at 460 nm using a CytoFluor II fluorescence microplate reader (PerSeptive Biosystems, Framingham, Mass.). The temperature was ramped up as described above for the turbidity assays (see Example 1). The plot of fluorescence as a function of temperature is shown in FIG. 3 for human α-thrombin alone, and for the 3 dp-4026/human α-thrombin complex. The denaturation transition for human α-thrombin was clearly observed at 57° C., a temperature which is only slightly higher than that observed in the turbidity experiment. The result from the fluorescence assay is, nonetheless, in close agreement with the T_mof 58° C. observed for prothrombin 1 from differential scanning calorimetry experiments. Importantly, 3 dp-4026 (at 67 μM) was found to shift the denaturation transition to ˜66° C. to give a shift in T_mof 9° C., which is identical to that found using turbidity as the detection signal (Table 3).

[0293] The results in FIG. 3 and Table 4 illustrate several important points. First, at least a 20-fold increase in sensitivity can be gained by switching from an absorbance to a fluorescence emission detection system. This can be critical for those receptor proteins for which supplies are limited.

[0294] Second, in the fluorescence assays, the denaturation transition signal is much cleaner than the signal in the turbidity assays. In the turbidity assays, higher concentrations of protein led to precipitation of denatured protein. Precipitated protein contributed to the noisy signal.

[0295] Third, shifts in T_mmeasurements from the microplate thermal shift assays are reproducible from one detection system to another.

[0296] The microplate thermal-shift assay was employed to test the binding of heparin 5000 and pentosan PSO₄to the known heparin binding site in the D(II) domain of fibroblast growth factor receptor 1 (FGFR1). D(II) FGFR1 is a 124 residue domain which is responsible for most of the free energy of binding for bFGF. D(II) FGFR1 was cloned and expressed in E. coli.Recombinant D(II) FGFR1 was renatured from inclusion bodies essentially as described (Wetmore, D. R. et al., Proc. Soc. Mtg., San Diego, Calif. (1994)), except that a hexa-histidine tag was included at the N-terminus to facilitate recovery by affinity chromatography on a Ni²⁺ chelate column (Janknecht, R. et al., Natl. Acad. Sci. USA 88:8972-8976 (1991)). D(II) FGFR1 was further purified on a heparin-sepharose column (Kan, M. et al., Science 259:1918-1921 (1993); Pantoliano, M. W. et al., Biochemistry 33:10229-10248 (1994)). Purity was >95%, as judged by SDS-PAGE. The D(II) FGFR1 protein was concentrated to 12 mg/mL (˜1 mM) and stored at 4° C.

[0297] The D(II) FGFR1 protein was dissolved in an ANS solution to a concentration of 1.0 mg/mL (70 μM). The quantum yield for ANS bound to the denatured form of D(II) FGFR1 was lower than the quantum yield for ANS bound to human α-thrombin. Because ANS fluorescence is very environment dependent (see Lakowicz, I. R., Principles of Fluorescence Spectroscopy, Plenum Press, N.Y. (1983)), the quantum yield observed for the denaturation of different proteins will vary. For D(II) FGFR1, the signal for the turbidity version of the assay, however, was nearly undetectable. Despite the decreased sensitivity for D(DI) FGFR1, ANS rescued this system for the microplate assay. A similar result was obtained for Factor Xa, except that the fluorescence quantum yield for ANS bound to denatured Factor Xa was almost as good as it was for human α-thrombin. It was found that the fluorescence quantum yield for ANS bound to denatured bFGF was as high as the quantum yield for ANS binding to human α-thrombin.

[0298] The results of D(II) FGFR1 binding experiments, as determined by the microplate thermal shift assay, are shown in FIG. 4 and Table 6. As was previously demonstrated for all of the other receptor proteins described above, the microplate thermal shift assay facilitated the ranking of ligand binding affinities for D(II) FGFR1.

[0299] In order to further demonstrate the cross target utility of the microplate thermal shift assay, another enzyme, Factor D, was tested for its ability to undergo thermal unfolding transitions. Factor D is an essential serine protease involved in the activation of the alternative pathway of the complement system, the major effector system of the host defense against invading pathogens. Factor D was purified from the urine of a patient with Fanconi's syndrome (Narayana et al., J. Mol. Biol.235:695-708 (1994)) and diluted to 4 μM in assay buffer (50 mM Hepes, pH 7.5, 0.1 M NaCl). The assay volume was 10 μL and the concentration of 1,8-ANS was 100 μM. The experiment was carried out using 15 μL round bottom dimple plates (an 8×12 well array). The protein was heated in two degree increments between 42° C. to 62° C., using a Robocycler™ temperature cycler. After each heating step, and prior to fluorescence scanning using the CytoFluor II™ fluorescence plate reader the sample was cooled to 25° C. (see Example 1). The non-linear least squares curve fitting and other data analysis were performed as described for FIG. 3. The results of the microplate thermal shift assay of Factor D is shown in FIG. 5 and reveal a thermal unfolding transition that occurs near 324 K (51 ° C.) for the unliganded form of the protein. No reversible ligands of significant affinity are known for Factor D. The results in FIG. 5 show that the microplate thermal shift assay can be used to screen a library of compounds for Factor D ligands. The results in FIG. 5 also show that the microplate thermal shift assay is generally applicable to any target molecule.

[0300] Human Factor Xa, a key enzyme in the blood clotting coagulation pathway, was chosen as yet another test of the cross target utility of the microplate thermal shift assay. Factor Xa was purchased from Enzyme research Labs (South Bend, Ind.) and diluted to 1.4 μM in assay buffer (50 mM Hepes, pH 7.5, 0.1 M NaCl). The assay volume was 100 μL and the concentration of 1,8-ANS was 100 μM. The protein was heated in two degree increments between 50 ° C. to 80° C. using a Robocycler™ temperature cycler. After each heating step, prior to fluorescence scanning using the CytoFluor II™ fluorescence plate reader, the sample was cooled to 25° C. (see Example 1). The results of a microplate thermal shift assay of Factor Xa is shown in FIG. 6. A thermal unfolding transition was observed at 338 K (65° C.). Data analysis was described as described for FIG. 3. The results in FIG. 6 show that the microplate thermal shift assay of protein stability is generally applicable to any target molecule.

[0301] A miniaturized form of the microplate thermal shift assay was developed to minimize the amount of valuable therapeutic protein and ligands required for the assay. In the first attempt at decreasing the assay volume, the assay volume was decreased from 100 μL to 50 μL without adversely affecting the fluorescent signal. When the assay volume was reduced further by a factor of ten, to 5 μL, favorable results were obtained for human α-thrombin. As shown in FIG. 7, the human α-thrombin unfolding transition could be easily observed at its usual T_m. More importantly, an active site inhibitor was observed to shift the T_mof the unfolding transition by 8.3° K to yield an estimate of the K_dof 15 nM at the T_m. The K_aat T_mwas calculated using the relationship: $\begin{array}{cc}{K}_L^{T_m}=\frac{\begin{array}{c}\exp \{-\frac{\Delta H_u^{T_0}}{R}\left[\frac{1}{T_m}-\frac{1}{T_0}\right]+\\ \frac{\Delta C_{\mathrm{pu}}}{R}\left[\ln \left(\frac{T_m}{T_0}\right)+\frac{T_0}{T_m}-1\right]\}\end{array}}{\left[L_{T_m}\right]}& \text{(equation 1)}\end{array}$

[0302] where

[0303] K_L^T=K_aat T_m(ligand associate constant at T_m)

[0304] T_m=332.2° K (midpoint of the unfolding transition in the absence of a ligand)

[0305] T₀=323.9° K

[0306] ΔH_u^T=200.0 kcal/mol (enthalpy of unfolding for pre thrombin observed by Lentz et al., 1994)

[0307] ΔC_pu=2.0 kcal/mol (estimated change in heat capacity of unfolding for human α-thrombin)

[0308] L_T=50.0 μM

[0309] The Kd at temperatures near 25 or 37° C. will be of higher affinity if the enthalpy of binding, ΔH_b, is negative for this ligand. Using a spectrophotometric assay, an apparent K_iof approximately 8 nM was observed at 37° C. (310° K).

[0310] The measurements shown in FIG. 7 were obtained using the CytoFluor II fluorescence plate reader (PerSeptive Biosystems, Framingham, Mass.). In the experiment, the excitation wavelength of light was 360 nm and the emission was measured at 460 nm. The microplates employed for this miniaturized assay were either the conventional polycarbonate V-bottom 96 well plate (Stratagene, or Costar) or polycarbonate plates that contain 15 μL dimples in an 8×12 array (Costar plate lids). In the reaction, the concentration of human α-thrombin was μM in assay buffer (50 mM Hepes, pH 7.5, 0.1 M NaCl). The assay volume was 5 μL and the concentration of 1,8-ANS was 100 μM. The protein was heated in two degree increments between 44° C. to 64° C. using a Robocycler™ temperature cycler. After each heating step, and prior to fluorescence scanning using the CytoFluor II™ fluorescence plate reader the sample was cooled to 25 ° C. for 30 seconds (see Example 1). The non-linear least squares curve fitting and other data analysis were performed as described for FIG. 3.

[0311] Recombinant D(II) FGFR1 was purified from inclusion bodies and purified by affinity chromatography on heparin sepharose. A stock solution of D(II) FGFR1 (15 mg/mL; 1.1 mM) was diluted to 50 μM in assay buffer (50 mM Hepes, pH 7.5, 0.1 M NaCl). The assay volume was 10 μL and the concentration of 1,8-ANS was 250 μM. The unfolding transition in the absence of ligands was found to be about 312 K (39° C.) as shown in FIG. 8. In the presence of the heparin mimic aprosulate (300 uM), the unfoding transition was observed to increase by about 8 K to about 320 K. Using this temperature midpoint T_m, it is possible to estimate the binding affinity of aprosulate to D(II)FGFR1 to be about 18 μM at the T_m(Table 6). These results demonstrate the ability of the microplate thermal shift assay to estimate ligand binding affinity to a non-enzyme target molecule.

[0312] Another target molecule analyzed was human urokinase-type plasminogen activator (u-PA). U-PA enzymatically converts plasminogen into the active protease plasmin. U-PA is involved in tissue remodeling, cellular migration and metastases. The gene for u-PA was obtained from ATCC (Rockville, Md.) and modified to appropriately express active enzyme in E. coli.u-PA was cloned, overexpressed in E. coli,and purified using procedures similar to those described by Winkler et al. (Biochemistry25:4041-4045 (1986)). The last step of u-PA purification was performed in the presence of the active site inhibitor glu-gly-arg-chloromethylketone (CMK) and hence the u-PA utilized for the u-PA utilized for this experiment was the CMK-u-PA complex. The experiment was performed in the miniaturized format in 5 μL well volume. One μL of concentrated CMK-u-PA (13 g/L, 171.4 μM) was added to 4 μL of 62.5 mM MOPS, pH 7, 125 mM NaCl, and 250 μM 1,8-ANS, in multiple wells of a 96-well polycarbonate V-bottom microtiter plate. A thermal denaturation curve was generated as previously described for thrombin, aFGF, D(II)FGFR1, Factor D, and Factor Xa, by incremental heating of the microplate followed by a fluorescence reading after each temperature increase. Analysis and non-linear least squares fitting of the data for this experiment show that the T_mfor CMK-u-PA under these conditions is 81° C., which is considerably higher than that seen for thrombin, aFGF, D(II)FGFR1, Factor D, and Factor Xa (55, 44, 40, 51, 55, and 65° C., respectively). This experiment demonstrates the utility of the current invention in determining the T_mfor relatively thermostable proteins or proteins stabilized by the high affinity binding of ligand(s) and further demonstrates the ability to perform such an experiment in a miniaturized format.

[0313] A stock thrombin solution was diluted to 1 μM in 50 mM Hepes, pH 7.5, 0.1 M NaCl and 100 μM 1,8-ANS. An electronic multi-channel pipettor was used to dispense either 2 μL or 5 μL of diluted thrombin solution into wells of a 96-well polycarbonate microtiter plat. The plate was subjected to 3 minutes of heating in a thermal block capable of establishing a temperature gradient across the microplate, followed by 30 seconds cooling to 25° C., and subsequent reading in the CytoFluor II fluorescence plate reader. Data were analyzed by non-linear least squares fitting and plotted as shown in FIGS. 10 and 11. Each curve represents a replicate experiment. Standard deviations for T_mdeterminations were very good for experiments utilizing either 5 μL or 2 μL volumes (+/−1.73 and +/−0.90 K, respectively), demonstrating the ability of the current invention to operate at very low volumes. In fact, the volume which one could employ in the current invention seems to be limited only by the technology available to dispense small volumes accurately.

[0314] The assay volume was reduced to 2 μL, as shown for human α-thrombin (1.0 μM) in FIG. 11. Reproducible pipetting of 2 μL in a 96 well array requires the employment of specialized pipetting tools such as the multi-channel pipettor available from Matrix Technologies Corp. (Lowell, Mass.) which has ±2.0% or 0.15 μL precision and +2.5% or 0.15 μL accuracy for volumes 0.5 to 12.5 μL.

[0315] Results of a single temperature assay are shown in FIG. 12. The compounds 3DP-3811, 3DP-3959, 3DP4076, and 3DP-4660 bind to the active site of human α-thrombin. The K_i's (enzymatically determined) of these four compounds for human α-thrombin are of 20,000 nM, 250 nM, 25nM, and 8 nM, respectively. Each of these four compounds were equilibrated with human α-thrombin in separate 5 μl assay volumes in a 96 well plate. The final ligand concentration was 50 μM.

[0316] For the ligands that bind to human α-thrombin with higher affinity, low levels of fluorescence emission were observed, relative to the control reaction (human α-thrombin alone) at 55° C. The result for the sample containing the weakly binding ligand 3DP-3811 was little different from the result obtained for the control sample. The decrease in fluorescence emission for 3DP4076 was not as large as expected, given its high affinity (K_iof 25 nM) for human α-thrombin. This result could be due in part to the lower solubility of chloride salts of this compound.

[0317] The data in FIG. 12 clearly demonstrate the utility of the single temperature embodiment of the microplate thermal shift assay for quickly identifying ligands with binding affinities (K_d's) of 250 nM or better when the ligand concentration is 50 μM.

[0318] The intrinsic Trp fluorescence of human α-thrombin was assayed in a microplate thermal shift assay. 100 μL samples contained 2 μM human α-thrombin. The samples were exposed to light from a Xenon-Arc lamp at 280 nm. Emission was detected at 350 nm using the BioLumin 960 (Molecular Dynamics). Temperature cycling, between 44° C. and 66° C., was performed as described in previous examples. The results of the assay are shown in FIGS. 13 and 14. A small increase in fluorescence emission was observed at 350 nm with increasing temperature. However, this increase in fluorescence emission was barely detectable above the level of fluorescence in the blank wells that contained no protein (FIG. 13). Subtracting an average blank improved the signal to noise ratio (FIG. 14), but the observed unfolding transition was different from that typically observed in assays employing 1,8-ANS. In contrast to the transition observed using 1,8-ANS, the transition in FIG. 14 appears broader and has a midpoint temperature T_mat 334.4±5.1° K, some five degrees higher than the T_mobserved for human α-thrombin in assays performed with 1,8-ANS.

[0319] As previously demonstrated, the thermal shift assay can be used for the screening of ligands for binding to single sites on target proteins. In light of the underlying physical principles upon which the microplate thermal shift assay is based, the near additivity of the free energy of ligand binding and protein unfolding, it is possible to employ the microplate thermal shift assay for analyzing multi-ligand binding interactions with a target protein. If the free energy of binding of different ligands binding to the same protein are nearly additive, then one can analyze multi-ligand binding systems, whether the ligands bind in a cooperative (positive) fashion or a non-cooperative (negative) fashion.

[0320] Multiple ligand binding to human α-thrombin was assayed in a microplate thermal shift assay. Human α-thrombin it has at least four different ligand binding sites: (1) the catalytic binding site; (2) the fibrin binding site (exosite I); (3) the heparin binding site (exosite II); and (4) the Na⁺ binding site, located ˜15 Å from the catalytic site. First, independent binding of three individual ligands was assayed: 3DP-4660, Hirugen (hirudin 53-64) (Bachem), and heparin 5000 (CalBiochem). These ligands bind to the catalytic site, the fibrin binding site and the heparin binding site, respectively.

[0321] A stock thrombin solution was diluted to 1 μM in 50 mM Hepes, pH 7.5, 0.1 M NaCl, 1 mM CaCl₂, and 100 μM 1,8-ANS. Each thrombin ligand was included singly and in various combinations to 1 μM thrombin solutions at final concentrations of 50 μM each, except for heparin 5000, which was 200 μM. 100 μL of thrombin or thrombin/ligand(s) solution was dispensed into wells of a 96-well V-bottom polycarbonate microtiter plate. The plate was subjected to 3 minutes of heating in a thermal block capable of establishing a temperature gradient across the microplate, followed by 30 seconds cooling at 25° C., and subsequent reading in a fluorescence plate reader. Data were analyzed by non-linear least squares fitting.

[0322] The results of these individual binding reactions are shown in FIGS. 15 and 16. The rank order of binding affinity was 3DP-4660>Hirugen>heparin 5000, corresponding to K_dvalues of 15 nM, 185 nM and 3434 nM, respectively, for the ligands binding at each T_m(see Equation (4)).

[0323] The results reveal thermal unfolding shifts that are slightly smaller than would be expected if the free energies of binding were fully additive. For example, Hirugen alone displays a ΔT_mof 5.8° C., and 3DP-4660 alone displays a ΔT_mof 7.7° C. In combination, however, Hirugen and 3DP-4660 display a ΔT_mof 12.2° C. This result means that the binding affinity of one or both ligands is diminished when both ligands are bound, and is an example of negative cooperativity in binding between the fibrin and catalytic binding sites. Such a negatively cooperative effect is consistent with the human α-thrombin literature, in which the kinetics of hydrolysis of various chromogenic substrates were found to depend upon ligands binding to exosite I. Indeed, a 60% decrease in K_mfor the hydrolysis of D-phenylalanylpipecolyl arginyl-p-nitroanilide was observed when Hirugen was present (Dennis et al., Eur. J. Biochem.188:61-66 (1990)). Moreover, there is also structural evidence for cooperativity between the catalytic site and exosite I. A comparison of the isomorphous structures of human α-thrombin bound to PPACK (a human α-thrombin catalytic site inhibitor) and Hirugen revealed conformational changes that occur at the active site as a result of Hirugen binding at the exosite I (Vijayalakshmi et al., Protein Science3:2254-2271 (1994)). Thus, in the microplate thermal shift assay, the apparent cooperativity observed between the catalytic center and the exosite I is consistent with functional and structural data in the literature.

[0324] Similarly, when the binding of all three ligands was assayed, a ΔT_mof 12.9° C. was observed (FIG. 16). If the free energies of binding were fully additive, one would expect to observe a ΔT_mof 17.7° C. The observed result means that further negative cooperativity occurs via ligand binding at all three protein binding sites. This result is consistent with the literature. In a ternary complex with heparin and fibrin monomer, human α-thrombin has decreased activity toward tri-peptide chromogenic substrates and pro-thrombin (Hogg & Jackson, J. Biol. Chem.265:248-255 (1990)), and markedly reduced reactivity with anti-thrombin (Hogg & Jackson, Proc. Natl. Acad. Sci. USA86:3619-3623 (1989)). Also, recent observations indicate that ternary complexes also form in plasma and markedly compromise heparin anticoagulant activity (Hotchkiss et al., Blood84:498-503 (1994)). A summary of these multi-ligand binding results is shown in Table 7.

[0325] The results in FIG. 15, FIG. 16, Table 7 illustrate the following advantages of using the microplate thermal shift assay to perform multi-variable analyses. First, the same microplate thermal shift assay can be used to simultaneously detect the binding of multiple ligands at multiple binding sites in a target protein. Second, the microplate thermal shift assay can be used to detect the same ligand binding to two or more sites in a therapeutic target. Third, the microplate thermal shift assay affords the detection of cooperativity in ligand binding. Information about ligand binding cooperativity can be collected and analyzed very quickly. Thus, multi-ligand binding experiments that would take months to perform using alternative technologies take only hours to perform using the microplate thermal shift assay.

[0326] The microplate thermal shift assay was used, with four different fluorophores, to simultaneously screen the effects of multiple pH values, sodium chloride concentrations, and reduction-oxidation compounds on human α-thrombin stability. Thrombin solution was diluted to 1 μM in 50 mM Hepes, pH 7.5, NaCl at either 0.1 M or 0.5 M, 10 mM EDTA, 10 mM CaCl₂, 10 mM dithiothreitol, 10 1 mM CaCl₂, and 100 μM 1,8-ANS, 10% (v/v) glycerol, or 0.1% (w/v) polyethylene glycol (PEG) 6000. Reaction volume was 100 μL.

[0327] The results of these multi-variable experiments are shown in FIGS. 17A-D and FIG. 18. FIGS. 17A-D summarize the stability data collected in a single 96 well plate for human α-thrombin. In FIG. 17A, the fluorophore is 1,8-ANS. In FIG. 17B, the fluorophore is 2,6-ANS. In FIG. 17C, the fluorophore is 2,6-TNS. In FIG. 17D, the fluorophore is bis-ANS. The results in FIGS. 17A-D show a pH optimum of about 7.0 and an increase in stability with increasing NaCl concentration. A ΔT_mof about 12° C. was observed when the NaCl concentration was increased from 0 to 0.5 M. FIG. 18 shows a stabilizing effect of 10% glycerol and a destabilizing effect of dithiothreitol. From FIGS. 17A-D and 18 is evident that the flourophores 1,8-ANS and 2,6-TNS are most effective in the microplate thermal shift assay.

[0328] The stabilizing effect of NaCl is particularly interesting since there are recent reports in the literature of a weak Na⁺ binding site (K_dof 30±3 mM in 5 mM Tris buffer pH 8.0, 0.1% PEG, 25° C.) approximately 15 Å from the catalytic center of thrombin (Dang et al., Nature Biotechnology15:146-149 (1997)). Using equation (1), it is possible to estimate the NaCl binding to be ˜6 mM near the T_m(53° C.) in 50 mM Hepes pH 8.0 buffer (zero and 0.10 M NaCl).

[0329] The additional stabilization that occurs at a NaCl concentration of greater than 0.10 M may come from additional Na⁺ and/or Cl⁻ binding events summed over the entire structure of human α-thrombin. Alternatively, the source of this further stabilization may come from less specific salting out effect that is usually observed at 0.5 to 2 M NaCl and is due to the preferential hydration of proteins induced by salts (Timasheff & Arakawa, In: Protein Structure, A Practical Approach,T. E. Creighton, ed., IRL Press, Oxford, UK (1989), pp. 331-354)).

[0330] The stabilizing effect of glycerol on proteins has been attributed to a balance between the preferential exclusion of glycerol (i.e. preferential hydration of proteins) and the specific binding to polar regions on the surface of proteins (Timasheff & Arakawa, In: Protein Structure, A Practical Approach,T. E. Creighton, ed., IRL Press, Oxford, UK (1989), pp. 331-354)).

[0331] The microplate thermal shift assay was used to simultaneously screen the effects of multiple biochemical conditions on D(II) FGF receptor 1 stability. The assays were performed by mixing 1 μL of D(II) FGFR1 (from a 500 μM concentrated stock in 50 mM HEPES pH 7.5) with 4 μL of each biochemical condition in wells of a 96-well polycarbonate microtiter plate. Final protein concentration after mixing was 100 μM and final 1,8-ANS concentration was 200 μM. Biochemical conditions were tested as follows: The pH's tested were 5 (Na acetate), 6 (MES), 7 (MOPS), 8 (HEPES), and 9 (CHES), with final buffer concentrations of 50 mM.

[0332] The salt concentrations tested were 0.1 or 0.5 M NaCl. Additives were tested in 50 mM MOPS, pH 7, 0.1 M NaCl, at final concentrations of 1 mM (EDTA, dithiothreitol), 10 mM (CaCl₂, MgCl₂, MgSO₄, NiSO₄), 50 mM (arginine), 100 mM ((NH₄)₂SO₄, LiSO₄, Na₂SO₄, ZnSO₄), 5% w/v (polyethylene glycol 6000), and 10% v/v glycerol.

[0333] Thermal denaturation profiles were generated as previously described for thrombin, aFGF, Factor D, and Factor Xa, by incremental heating of the microplate followed by a fluorescence reading after each temperature increase. Data were analyzed by non-linear least squares fitting as described previously.

[0334] The results of these multi-variable experiments are shown in FIGS. 19-24. As shown in FIG. 19, stability increased with increasing NaCl concentration. A ΔT_mof about 5° C. was observed as NaCl concentration was increased from 0.1 to 0.5 M. As shown in FIG. 20, both MgSO₄and arginine stabilized the protein. As shown in FIG. 21, 10% glycerol stabilized the protein. Further, salts of the Hofmeister series such as Li₂SO₄, Na₂SO₄, (NH₄)₂SO₄and Mg₂SO₄all had stabilizing effects (FIG. 21). As shown in FIG. 22, dithiothreitol destabililzed the protein. These results are not very different form that of human α-thrombin. As shown in FIG. 23, a pH optimum of about 8.0 was observed. The relative stabilizing effects of EDTA, CaCl₂, MgCl₂, MgSO₄, arginine, (NH₄)₂SO₄, Li₂SO₄, Na₂SO₄, glycerol, polyethylene glycol 6000, and dithiothreitol are shown in FIG. 24.

[0335] The microplate thermal shift assay was used to simultaneously screen the effects of multiple biochemical conditions on human urokinase stability. This experiment was performed by mixing 1 μL of urokinase (from a 371 5M concentrated stock in 20 mM Tris pH 8) with 4 μL of each biochemical condition in wells of a 96-well polycarbonate microtiter plate. Final protein concentration after mixing was 74 μM and final 1,8-ANS concentration was 200 μM. Biochemical conditions were tested as follows: The pH's tested were 5 (acetate), 6 (MES), 7 (MOPS), 8 (HEPES), and 9 (CHES) with final buffer concentrations of 50 mM. The salt concentrations tested were 0.1 or 0.5 M NaCl. Glycerol was tested at 10% v/v in 50 mM MOPS, pH 7, 0.1 M NaCl.

[0336] Thermal denaturation profiles were generated as previously described for thrombin, aFGF, Factor D, D(II) FGFR1, and Factor Xa, by incremental heating of the microplate followed by a fluorescence reading after each temperature increase. Data were analyzed by non-linear least squares fitting as described previously.

[0337] The results of these multi-variable experiments are shown in FIG. 25. A pH optimum of about 7.0 was observed. Increasing concentrations of sodium chloride stabilized the protein. 10% glycerol also stabilized the protein. These results are consistent with the results reported in the literature (Timasheff & Arakawa, In: Protein Structure, A Practical Approach,T. E. Creighton, ed., IRL Press, Oxford, UK (1989), pp. 331-354).

[0338] FIGS. 17-25 illustrate the advantage of using the microplate thermal shift assay to simultaneously screen for multi-variable biochemical conditions that optimize protein stability. Using the methods and apparatus of the present invention, one can rapidly screen large arrays of biochemical conditions for conditions that influence the stability of proteins. Thus, the present invention can be used to rapidly identify biochemical conditions that optimize protein shelf-life.

[0339] Factorial experiments were performed to identify biochemical conditions that increased the yield of correctly folded His₆-D(II)-FGFR1. His₆-D(II)-FGFR1 is recombinant D(II) FGF receptor 1 protein, to which a polyhistidine tag is attached to the N-terminus. The results are summarized in Table 8. When the final guanidinium hydrochloride concentration was 0.38 M, a refolded protein yield of 13.5±0.2% was obtained at pH 8.0 and 0.5 M NaCl. This yield could be increased to 15.5±0.3% if glycerol was present at 7% (v/v). A further increase in His₆-D(II)-FGFRI refolding yield to about 18% was observed when the pH was increased to 8.9. In fact, increasing the pH from 8.0 to 8.9 improved the yields in all experiments. These results demonstrate that a pH between 8 and 9, and 7% glycerol, are two important conditions that facilitate D(II)-FGFRI folding. Each of these conditions increased the folded protein yield by about 15 to 20% over the starting conditions at pH 8.0 and 0.5 M NaCl.

[0340] Importantly, the effects of pH and glycerol appear to be nearly additive. The increased yield of refolded protein at pH 8.9 and 7% glycerol was found to be 17.8%, 32% higher than the yield obtained at a pH 8.0 and 0.5 M NaCl (13.5±0.2% yield). The near additivity of refolding determinants has important consequences since it suggests that the small individual free energy components that comprise the overall free energy of folding can be incrementally combined to optimize the yield of folded protein.

[0341] Results of a second round of refolding experiments at a final Gdn-HCl concentration of 0.09 M revealed that the Gdn-HCl is an even more important factor affecting the folding of His₆-D(II)FGFR1 (Table 9). At pH 8.0 and 0.5 M NaCl, decreasing the Gdn-HCl concentration to 0.09 M doubled the refolded protein yield, relative to the yield obtained at pH 8.0, 0.5 M NaCl, and 0.38 M Gdn-HCl (Table 9). In accordance with the results obtained at a Gdn-HCl concentration of 0.38 M, the yield of refolded His₆-D(II)-FGFR1 in 0.09 M Gdn-HCl was also increased in the presence of glycerol. These results suggest that the improved yield of refolded His₆D(II)-FGFR1 in glycerol (5 to 10%) and lower Gdn-HCl concentration are additive. Further, the results in Table 9 reveal that the Hofmeister salt Na₂SO₄increases the yield of refolded protein almost as well as 5 to 10 % glycerol.

[0342] Upon comparison of the biochemical conditions that increase the yield of refolded Ni²⁺NTA bound His₆-D(II)-FGFR1 (Tables 8 and 9) and those conditions that increase the overall protein stability of His₆-D(II)-FGFR1 (FIGS. 19-24), it is clear that there is a strong correlation between the protein folding results and the protein stability results. Glycerol, salts of the Hofmeister series, and pH 8.5 to 8.9 improve protein folding yield and overall protein stability of His₆-D(II)-FGFR1.

[0343] These results are consistent with the model of protein folding in FIG. 26. If the aggregation of unfolded His₆-D(II)-FGFR1 is suppressed when immobilized to Ni²⁺NTA, and a simple two state equilibrium exists between U and N, then the factors that influence the relative position of the equilibrium between U and N should be the same whether one starts from U (in the refolding experiment) or start from N (in the microplate thermal shift assay protein stability screen). Since thermodynamics are path independent, only the initial and final states of this reaction should be important. Since similar biochemical conditions facilitate protein stability and folded protein yield, the simple model for protein folding depicted in FIG. 26 is accurate for this protein. Thus, the microplate thermal shift assay can serve as a rapid and general method for screening biochemical conditions that optimize protein folding.

[0344] FIG. 28 shows the results of microplate thermal shift assays of using each of four fluorescence probe molecules: bis-ANS, 2,6-TNS, 1,8-TNS, and 2,6-ANS. Thrombin solution was diluted to 1 μM in 50 mM Hepes, pH 7.5, and 0.1 M NaCl.

[0345] A Gel Documentation and Analysis System (Alpha Innotech Corp., San Leandro, Calif.) was used to perform a microplate thermal shift assay. This system uses a CCD camera to detect fluorescence emission from stained gels, dot blot assays, and 96 well plates. The excitatory light source was a long wavelength UV trans-illumination box located directly below the CCD camera. The 96 well plate to be assayed was placed on the trans-illumination box within the focal viewing area of the CCD camera (21×26 cm).

[0346] A 2 μM solution of human α-thrombin was prepared in 50 mM Hepes, pH 7.5, 0.1 M NaCl by diluting a 34 μM stock solution (1:17) of purified human α-thrombin (Enzyme Research Labs, Madison, Wis.). The human α-thrombin solution also contained 100 μM 1,8-ANS. 100 μL of the human α-thrombin-1,8-ANS solution was aliquoted into each of twelve wells of a single row (row A) of a V-bottom polycarbonate microplate (Costar). A gradient block (RoboCycler™, Stratagene) was used to heat the twelve samples, from 44 to 66° C., across the rows of the microplate. i.e. a temperature gradient of 2° C. per well was established. Thus, well A1 was at 66° C. and well A12 was at 44° C. The control solution that contained 100 μM 1,8 ANS in the same buffer (no protein) was placed in each of wells B1 to B12. After adding a drop of mineral oil to each well to prevent evaporation, the plate was heated on the gradient block for 3 min. The contents of each well were then allowed to reach room temperature and transferred to a flat bottom microplate. In this experiment, no filters were employed to narrow the excitatory wavelength to ˜360 nm and the emission wavelength to ˜460 nm, which are optimal wavelengths for the 1,8 ANS fluorophore. The flat bottom plate was then placed on the near UV transillumination box and the CCD camera was used to measure the amount of emitted light. The plate was also read using a conventional fluorescence plate reader (CytoFluor II), in order to compare the results obtained by the two different detection methods. The results for the two detection methods are plotted in FIG. 40. The results in FIG. 40 show that the CCD camera is useful as a fluorescence emission detector for monitoring the unfolding of a protein in the microplate thermal shift assay.

[0347] An emission filter was used to block out all stray light outside the region of the emission region for 1,8-ANS (˜460 nm). In addition, the 5 μL miniaturized form of the microplate thermal shift assay was employed to test the CCD camera detection method in this configuration. Both the polycarbonate V-bottom and dimple plates were tested. The experiment was essentially the same as described in Example 16, except that the volume of the assay was 5 μL in either the V-bottom or dimple 96 well plates. The temperature range was 44 to 66° C. (right to left) for the V-bottom plate, and 46 to 70° C. (right to left) for the dimple plate. Photographs of the CCD images are shown in FIG. 41. The V-bottom well microplate image is shown in FIG. 41A. The dimple plate image is shown in FIG. 41B. The results obtained from the plate in FIG. 41A is shown in FIG. 42. The results in FIG. 42 show that data obtained using a CCD camera compare very well with data obtained using a fluorescence plate reader that employs a photo-multiplier tube (PMT) for fluorescence detection.

[0348] All publications and patents mentioned hereinabove are hereby incorporated in their entireties by reference.

[0349] While the foregoing invention has been described in some detail for purposes of clarity and understanding, it will be appreciated by one skilled in the art from a reading of this disclosure that various changes in form and detail can be made without departing from the true scope of the invention and appended claims.