US20030068652A1

US20030068652A1 - Positive identification of phospho-proteins using motif-specific, context-independent antibodies coupled with database searching

Info

Publication number: US20030068652A1
Application number: US10/174,105
Authority: US
Inventors: Michael Comb; Yi Tan; Hui Zhang
Original assignee: Cell Signaling Technology Inc
Current assignee: Cell Signaling Technology Inc
Priority date: 1998-09-04
Filing date: 2002-06-18
Publication date: 2003-04-10
Also published as: DE69936389T2; JP2006036784A; WO2000014536A1; EP2325204A2; EP1110086B1; EP1850131A1; US7344714B2; DE69936389D1; US6982318B1; EP2325204B1; US20020132988A1; EP1110086A1; CA2341143A1; JP3732408B2; US7259022B2; JP2002524092A; US20020168684A1; JP2016210783A; US6441140B1; EP1110086A4

Abstract

The invention provides a method for producing antibodies that selectively recognize short, modified amino acid motifs substantially independent of the surrounding amino acid context in which the motif occurs. A novel class of motif-specific, context-independent antibodies is also provided. The invention encompasses modified motifs consisting of single modified amino acids, for example phosphotyrosine or acetylated lysine, as well other modified motifs of multiple amino acids, such as kinase consensus substrate motifs and protein-protein binding motifs relevant to cell signal transduction. Also provided are methods of profiling large and diverse protein populations on a genome-wide basis by utilizing the antibodies of the invention, and methods for the positive identification of cellular phosphoproteins using one or more motif-specific, context-independent antibodies of the invention coupled with protein database searching.

Description

RELATED APPLICATIONS

This application is a continuation-in-part of, and claims priority to, U.S. Ser. No. 09/535,364, filed Mar. 24, 2000, which itself is a continuation-in-part of U.S. Ser. No. 09/148,712, filed Sep. 4, 1998, both presently pending, the disclosures of which are hereby incorporated herein by reference.[0001]

FIELD OF THE INVENTION

The invention relates generally to antibodies, and more specifically to activation-state specific antibodies and their uses.

BACKGROUND OF THE INVENTION

The activation of proteins by modification represents an important cellular mechanism for regulating most aspects of biological organization and control, including growth, development, homeostasis, and cellular communication. For example, protein phosphorylation plays a critical role in the etiology of many pathological conditions and diseases, including cancer, developmental disorders, autoimmune diseases, and diabetes. In spite of the importance of protein modification, it is not yet well understood at the molecular level. The reasons for this lack of understanding are, first, that the cellular modification system is extraordinarily complex, and second, that the technology necessary to unravel its complexity has not yet been fully developed.

The human genome encodes, for example, over 520 different protein kinases, making them the most abundant class of enzymes known. See Hunter, Nature 411: 355-65 (2001). Many kinases have been shown to phosphorylate specific serine, threonine, or tyrosine residues located within distinct amino acid sequences, or kinase consensus substrate motifs, contained within different protein substrates. Most kinases phosphorylate many different proteins: it is estimated that one-third of all proteins encoded by the human genome are phosphorylated, and many are phosphorylated at multiple sites by different kinases. See Graves et al., Pharmacol. Ther. 82: 111-21 (1999). Many of these phosphorylation sites regulate critical biological processes and may prove to be important diagnostic or therapeutic targets for molecular medicine. For example, of the more than 100 dominant oncogenes identified to date, 46 are protein kinases. See Hunter, supra. Understanding which proteins are modified by these kinases will greatly expand our understanding of the molecular mechanisms underlying, e.g., oncogenic transformation.

Although a few targets of protein phosphorylation have been identified, most remain unknown, particularly those that regulate cell growth and differentiation. For example, the MAP kinase cascade is known to play an important role in the regulation of cell growth (Lewis et al., Adv. Cancer Res. 74:49-139 (1998), Crowley et al., Cell 77:841-852 (1994)). However, beyond a handful of substrates, few protein targets responsible for the diverse actions of the MAP kinase cascade have been identified (Fukunaga and Hunter, EMBO 16(8):1921-1933 (1997), Stukenberg et al., Curr. Biol. 7:338-348 (1997)).

Another example of cell signaling proteins are the 14-3-3 proteins, which represent a conserved family of phosphoserine binding proteins involved in the regulation of cell survival, apoptosis, proliferation and checkpoint control. These proteins represent a large fraction of total brain protein and are known to bind a wide variety of signaling molecules including: ras, raf, bad, cdc25, and many others (Yaffe et al., Cell 91:961-971 (1997)). Recently, it has been shown that 14-3-3 proteins bind specifically to phosphorylated sites on proteins with the following motif: RXRSXS*XP where S* is phosphoserine and X represents any amino acid (Muslin et al., Cell 84:889-897 (1996), Yaffe et al. supra(1997)).

Similarly, histones have long been known to be modified by acetylation at specific lysine residues. Acetylation of lysine in histones is thought to reduce protein-DNA interactions and serve to open chromatin in regions undergoing transcription (Struhl, Genes & Development, 12:599-606 (1998)). Recently, other proteins associated with transcription complexes have been shown to be acetylated on lysine although the functional significance is unclear (Imhof et al., Curr. Biol. 7:689-692 (1997), Struhl supra (1998)). It is clear, therefore, that new reagents and methods are needed to enable the detection and identification of the particular modified proteins that are relevant to normal, or aberrant, signal transduction in a given cell.

Antibodies have been a reagent of choice for the detection of protein modifications. Among such reagents, antibodies against phosphotyrosine have proven to be of great value in identifying and characterizing intracellular signaling mechanisms (Ross et al., Nature 294:654 (1981), Kozma et al., Method. Enzymol. 201:28 (1991), White and Backer, Method. Enzymol. 201:65 (1991), Kamps, Method. Enzymol. 201:101 (1991); Frackelton et al., Method. Enzymol. 201:79 (1991); Wang, Method. Enzymol. 201:53 (1991); Glenney, Method. Enzymol. 201:92 (1991)). Phosphotyrosine antibodies discriminate on the basis of whether or not a protein is tyrosine phosphorylated, and thus react with a large variety of different phosphotyrosine containing proteins. Attempts to produce similar phosphoserine and phosphothreonine antibodies have met with limited success. (Heffetz et al., Method. Enzymol. 201:44 (1991)).

Site-specific phosphoserine and phosphothreonine antibodies that bind unique phosphorylated sequences on a particular protein were first described by Nairn et al.. in 1982 and have proven to be highly useful tools to study protein phosphorylation (Czernik et al., Method. Enzymol. 201:264 (1991), Czernik et al., Neuroprot. 6:56-61 (1995)). One drawback of this type of antibody is that a different antibody needs to be produced for each site of interest. Phosphorylation site-specific antibodies are, therefore, not suitable for the detection of multiple proteins within a genome that contain a common modified motif of interest, such as a kinase consensus substrate motif.

Accordingly, there remains a need in the art for a new class of antibodies that are capable of recognizing short modified motifs common to a plurality of cell signaling peptides or proteins within a genome, in a manner that is independent of the context in which the motifs occur. Such antibodies would be useful, for example, in identifying new targets of cellular signaling activity (e.g. targets of 14-3-3 binding), and as reagents for the detection of enzymatic modifications like phosphorylation in vitro, e.g. in high throughput kinase assays for drug screens, as a single antibody can be used to recognize many different phosphorylated substrates containing the motif of interest.

Motif-specific, context-independent antibodies would be particularly useful in, and would enable, new methods for profiling genome-wide changes in protein levels and protein modification, and positively identifying modified proteins based, in part, upon the presence of the detected motif. For example, the use of motif-specific, context-independent antibodies and 2D gel electrophoresis to profile genome wide changes in protein phosphorylation (Patterson and Garrels, Cell Biology: A Laboratory Handbook 249-257 (1994), Academic Press) as the result of drug treatment or overexpression of a particular protein will undoubtedly prove useful in identifying potential drug-protein interactions and suggest new downstream targets for overexpressed proteins.

Despite an ever-increasing amount of protein phosphorylation and modification information in worldwide protein databases, it has largely not been possible to efficiently determine the identity of particular modified proteins in a cell at a given point in time. The traditional 2D phosphopeptide mapping is time consuming and not suited for identification of phosphorylation profiles in high throughput fashion (see e.g., Affolter et al., Anal. Biochem. 223: 74-81 (1994)). Immuno-precipitation of proteins using a sequence-specific antibody followed by western blot with general phospho-antibody, such as phosphotyrosine antibody, has been used to determine if a candidate protein is phosphorylated (Ostergaard et al., J. Biol. Chem. 273: 5692-96 (1998)). This methodology requires prior knowledge of the candidate protein and its phosphorylation sites, despite the fact that such sites are usually undetermined.

Immunoprecipitation of phosphoproteins using general phospho-antibodies, such as phosphotyrosine antibody followed by gel electrophoresis has been also used to identify phosphoproteins on a gel. See Yanagida et al., Electrophoresis 21(9): 1890-98 (2000). However, the method tediously requires that the protein band be excised from the gel and the protein subsequently identified by mass spectrometry. Significantly, this method can only identify those phosphoproteins abundant enough to be visualized by staining methods. Often, the actual phosphorylation site or residue remains undetermined because the phosphopeptides may not be detected after protease digest.

Accordingly, there remains a need for new, more sensitive methods for the positive identification of modified cellular proteins, such as phosphoproteins, and in particular, methods that are suitable for high-throughput profiling and identification of cellular protein modifications.

SUMMARY OF THE INVENTION

In accordance with the present invention, in part, there is provided a method of producing antibodies that selectively recognize specified short amino acid motifs in a manner substantially independent of the surrounding amino acid, peptide, or protein sequences. The method allows the production of antibodies that recognize modified single amino acids, for example phosphorylated serine, threonine, and tyrosine, or acetylated lysine, as well other unmodified or modified motifs of one or more amino acids, such as kinase consensus substrate motifs and protein-protein binding motifs. Preferred embodiments, advantages, and uses of the novel class of motif-specific, context-independent antibodies provided by the invention are described in detail in U.S. Ser. No. 09/148,712, Comb et al., (co-pending), the disclosure of which is incorporated herein by reference. Among such embodiments, are methods of profiling large and diverse protein populations on a genome-wide basis or scale by utilizing motif-specific, context-independent antibodies against modified motifs conserved on such cell signaling proteins.

The invention also provides methods for the positive identification of cellular phospho-proteins and their phosphorylation sites using one or more motif-specific, context-independent antibodies of the invention coupled with protein database searching. Detection of modified proteins, such as phosphoproteins, using one or more motif-specific, context-independent antibodies to detect proteins separated by molecular weight and/or isoelectric point (PI) provides initial information about modified motif sequences (specifically bound by the antibody) that are present in the detected proteins of known molecular mass and/or PI (as determined, e.g., by 1D SDS gel electrophoresis or PI (2D electrophoresis)). This information can then be used with a search program/software, such as Scansite, to identify potential phosphoproteins in databases that have substantially matching molecular weights and/or PI's and whose sequence comprises the required motif(s) specifically bound by the antibody. Proteins containing modifications other than phosphorylation, e.g. acetylation, may similarly be identified using motif-specific, context-independent antibodies directed to such modified motifs.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1[0017] a—is a table depicting the specificity of the affinity-purified, polyclonal antibodies produced against a phosphorylated threonine peptide library in Example I, when tested against specific peptides (SEQ ID NOs: 1-13).
FIG. 1[0018] b—is a table depicting the specificity of the phosphothreonine antibodies of Example I when tested against various phosphopeptide libraries (SEQ ID NOs: 14-29).
FIG. 1[0019] c—is a Western analysis depicting the reactivity of the phosphothreonine antibodies of Example I against cell extracts from cells treated with and without okadaic acid and against other phosphoproteins.
FIG. 1[0020] d—is a table depicting the context-independence of the anti-phosphothreonine antibodies of Example I as shown by immobilized grid.
FIG. 2[0021] a—is a table depicting the specificity of the affinity-purified, polyclonal antibodies produced against a phosphorylated PXS*P peptide library in Example II (SEQ ID NOs: 30-32).
FIG. 2[0022] b—is a Western analysis depicting the reactivity of the phospho-PXS*P antibodies of Example II against cell extracts from cells treated with and without okadaic acid and against other phosphoproteins.
FIG. 3[0023] a—is a table depicting the lack of reactivity of the affinity-purified, polyclonal 14-3-3 antibodies of Example III when tested against non-phosphopeptides or phosphopeptides lacking the motif (SEQ ID NOs: 33-40).
FIG. 3[0024] b—is a Western analysis depicting the reactivity of the phospho-14-3-3 antibodies of Example III against cell extracts from cells transfected with GST-Bad and with TPA.
FIG. 4[0025] a—is a table depicting the specificity of the monoclonal antibodies produced against a phosphorylated PXT*PXR library in Example IV (SEQ ID NOs: 41-44).
FIG. 4[0026] b—is a Western analysis depicting the reactivity of the CDK consensus site monoclonal antibodies of Example IV against phosphorylated and nonphosphorylated RB protein.
FIG. 5[0027] a—is a Western analysis depicting the specificity of the acetylated-lysine antibodies of Example V against acetylated BSA.
FIG. 5[0028] b—is a Western analysis depicting the reactivity of the acetylated-lysine antibodies of Example V against various proteins present in C6-cell extracts when antibodies are preincubated with nonacetylated peptide library.
FIG. 5[0029] c—is a Western analysis depicting the reactivity of the acetylated-lysine antibodies of Example V against various proteins present in C6-cell extracts when antibodies are preincubated with acetylated peptide library.
FIG. 5[0030] d—is a Western analysis depicting the reactivity of the acetylated-lysine antibodies of Example V against the control acetylated BSA when antibodies are preincubated with acetylated peptide library.
FIG. 6—shows the signal to noise ratio of ELISA readings using phospho-Akt substrate antibody with phospho-peptides of Akt substrates vs. non-phospho-peptides of Akt substrates (SEQ ID NOs: 48-61). [0031]
FIG. 7—is a Western analysis of calyculin A-treated A431 cells using phospho-Akt substrate antibody. [0032]
FIG. 8—shows the signal to noise ratio of ELISA reading using phospho-PKA substrates antibody against peptides have arginine or lysine at −3 position (SEQ ID NOs: 62-77). [0033]
FIG. 9—is a Western analysis of calyculin A-treated A431 cells using phospho-PKA substrates antibody. [0034]
FIG. 10—is a Western analysis of A431 cell extracts phosphorylated by protein kinase A, ERK2 and CDC2/cyclinA in vitro using phospho-PKA substrate antibody. [0035]
FIG. 11—shows the signal to noise ratio of ELISA reading using phospho-serine/threonine phenylalanine antibody against the peptides containing phenylalanine, tyrosine or tryptophan (SEQ ID NOs: 78-87). [0036]
FIG. 12—is a Western analysis of calyculin A-treated A431 cells using phospho-serine/phenylalanine substrates antibody. [0037]
FIG. 13—is a graphic plot indicating the relative preference values versus amino acid selected with respect to the binding of phospho-Akt consensus substrate motif antibody to a first test phospho-peptide library, AxxxxxxxT*xxxxAKKv (SEQ ID NO: 90). [0038]
FIG. 14—is a graphic plot indicating the relative preference values versus amino acid selected with respect to the binding of phospho-Akt consensus substrate motif antibody to a second test phospho-peptide library, AxxxRxxT*xGGGAKK (SEQ ID NO: 91) [0039]
FIG. 15—is a Western blot indicating the specific detection of RxRxxS*/T* motif containing phosphoproteins in cell lysate from IGF1-treated PDK1 null or wild type ES cells probed with antibodies against consensus Akt substrate motif. [0040]
FIG. 16—is a Western blot analysis of IGF1 treated wild type or PDK1−/−ES cells, using a Phospho-(Ser) PKC substrate motif antibody, phospho-specific antibodies against threonine at activation loops of PKC ζ/λ (Thr410), PKCζ (Thr403), PKCθ (Thr538), and PKCδ (Thr505), a phospho-specific antibody against the phosphorylation site in linker sequence between the activation loop and hydrophobic motif of PKC/II (Thr638/641), and a phospho-specific antibody against the hydrophobic motifs of PKC isoforms, P-PKC (pan). [0041]
FIG. 17—is a Western blot analysis of wild type or PDK1−/−ES cells treated with TPA in the presence or absence of PKC inhibitor, Ro31.8220, using a Phospho-(Ser) PKC substrate motif antibody. [0042]
FIG. 18—shows the list of potential matching proteins resulting from searching the Swiss-Prot database for putative proteins matching both Akt and PKC substrate motif antibody sequences and the requisite molecular weight between 28-32 kDa (SEQ ID NOs: 191-195). [0043]
FIG. 19—depicts the probing of the Western blot described in FIG. 15 with an antibody against S6 ribosomal protein; this antibody detects a band at the same size as detected by the phospho-Akt motif. [0044]
FIG. 20—depicts the specific immunoprecipitation of positively identified phospho-S6 ribosomal protein from wild type ES cells by phospho-(Ser/Thr) Akt motif antibody. [0045]
FIG. 21—is a table indicating the specificity of Akt substrate motif antibody for various consensus Akt substrate motif-containing peptides as determined by ELISA. Specificity is presented as a percentage of the ELISA reading for each peptide relative to that for phospho-GSK3 peptide (SEQ ID NOs: 92-152). [0046]
FIG. 22—is a table indicating the specificity of PKC substrate motif antibody for various consensus PKC substrate motif-containing peptides as determined by ELISA. Specificity is presented as a percentage of the ELISA reading for each peptide relative to that for phospho-AFX peptide (SEQ ID NOs: 153-190).[0047]

DETAILED DESCRIPTION OF THE INVENTION

The present invention is based, in part, upon the concept that the concentration of any individual sequence in a peptide library used as antigen is extremely low and hence will be insufficient to drive an immune response in a host. The only antigenic determinants of sufficiently high concentration to drive the immune response are thus the fixed residues common to each sequence, as well as the peptide backbone itself. [0048]
Immunizing a host with peptide libraries representing many or all 20 amino acids at each degenerate position will produce antibodies tolerant to many, or all, amino acids at the variable positions surrounding one or more fixed residues. Such antibodies will then react with the antigenic determinant in the context of the broadest possible range of surrounding amino acid, peptide, or protein sequences. The fixed residue(s) of the motif may be a single unmodified or modified amino acid, such as a phosphorylated or unphosphorylated residue, or may be multiple unmodified or modified amino acids, such as a consensus recognition site. [0049]
The ability of this novel class of antibodies to recognize a particular modified motif common to a plurality of different proteins or peptides within a genome enables new methods for the positive identification of modified proteins, e.g. phosphoproteins, based upon their molecular weight and/or PI and the known modified sequence specifically bound by one or more motif-specific, context-independent antibodies of the invention. Preferred embodiments of such methods and other advantages of the invention are further described in detail below. [0050]
As used herein, “antibodies” means polyclonal or monoclonal antibodies, including Fc fragments, Fab fragments, chimeric antibodies, or other antigen-specific antibody fragments. “Antibodies” and “antibody” are used interchangeably herein. [0051]
As used herein, “motif-specific, context-independent antibody” means an antibody that is specific against one or more fixed amino acid residues that comprise an amino acid motif in the context of variable surrounding peptide or protein sequences (flanking the motif); such antibody specificity is thus highly independent of the context in which the antigenic motif occurs, the antibody thus recognizing a plurality of peptides or proteins within a genome that contain the motif. As used herein, “substrate” means any target molecule, including peptides or proteins, which an enzyme specifically recognizes and acts upon. [0052]
Also as used herein, “preparation” or “cellular preparation” means a preparation of proteins or peptides from one or more cells, tissues, or biological fluids of an organism, whether unpurified or slightly purified, for example a crude cell extract or serum, or a partially purified extract; and “phosphoprotein” or “phosphopeptide” means a protein or peptide, respectively, comprising at least one, but alternatively more than one, phosphorylated amino acid. [0053]
The teachings of all references cited in this specification are hereby incorporated herein by reference. Further aspects, advantages and uses of the invention are described in more detail below. [0054]
Production and Characterization of Motif-Specific, Context-Independent Antibodies [0055]
The general method by which motif-specific, context-independent antibodies are produced in accordance with the present invention is as follows: [0056]
(1) Motif-specific antibodies that react with any protein or peptide containing specific target residues independently of the surrounding amino acids may be obtained by synthesizing a highly degenerate peptide library. In one preferred embodiment, the library comprises XXXXXXJ*XXXXXXC where X=all 20 amino acids except cysteine and J*=a modified (*) amino acid (J), for example, phosphothreonine (T*) or acetylated-lysine (K*). It will be appreciated that the specific target residue may be unmodified and that a shorter or longer library may be generated and less than all of the surrounding amino acids may be varied. In one preferred embodiment, the peptide library is about 6 to 14 residues long. While the preferred embodiment utilizes one fixed amino acid (either modified or unmodified) in a varied surrounding context, other preferred embodiments may utilize a motif comprising several fixed amino acids. Likewise, the surrounding sequence of the library may be varied at more than one position simultaneously, or, as in the preferred embodiment, varied at only one surrounding sequence position per degenerate molecule, such that a library is produced which is completely degenerate at every position except the fixed residue(s). The peptide library can be synthesized by standard F-Moc solid phase peptide synthesis using an ABI peptide synthesizer and using mixtures of each amino acid during degenerate coupling reactions. [0057]
The incorporation of modified amino acids at fixed positions should not be limited to phosphorylation or acetylations as other modified protected amino acids can also be incorporated, for example, amino acids modified with lipids (e.g. farnesylated, isoprenylated) or protected O-linked or N-linked sugars (e.g. glycosylated), methylated, or ribosylated amino acids, or nucleotides, polymers of nucleotides, nucleosides, or amino acids such as ubiquitin, or amino acid analogues. [0058]
The incorporation of unmodified amino acids at fixed positions may be selected to mimic conserved motifs, for example zinc fingers or repeating arginine residues. [0059]
(2) In order to produce as equal a representation of each amino acid as possible at each degenerate position, several rounds of altering the amino acid composition, synthesizing, and peptide sequencing are conducted. Amino acid sequence analysis at several different positions along the peptide is conducted to verify a random amino acid representation at each position and that the random representation is maintained throughout the synthesis. It will be recognized by one of skill in the art that the number of rounds may vary in order to achieve an equal distribution of all amino acids at each position. [0060]
(3) The highly diverse peptide library is used as an antigen, preferably by covalent coupling to a carrier. In a preferred embodiment, keyhole limpet hemocyanin (KLH) emulsified in Freund's adjuvant is used as the coupling agent, and the coupled peptide library injected intradermally into a host, such as female New Zealand white rabbits. Booster injections may be given in incomplete Freund's adjuvant until an immune response is obtained. Antibody titre is measured by a suitable method, such as ELISA against the motif-specific peptide libraries. Antisera raised in this manner may be used in both crude or purified preparations, as outlined below. [0061]
(4) Antisera from the most promising hosts are purified, for example over protein A, and adsorbed over a J (non-modified) peptide library column. In the preferred embodiment, the nonadsorbed fraction (flow through) is then applied to a J* column, eluted at suitable pH, dialyzed and tested for J* specificity by a suitable method, such as ELISA using J* and J as antigen. [0062]
(5) Antibodies affinity purified in this fashion recognize the J* peptide library but do not react with the J library and exhibit a high degree of specificity for J*. These antibodies may be further tested for lack of reactivity against the unmodified form of the target modified amino acid, J*, or a J* homologue, utilizing a suitable method, such as ELISA. [0063]
(6) Antibodies may be further tested by western blotting, or another suitable method, using cell extracts prepared from cells treated with and without a selected protein modification enzyme inhibitor, such as protein phosphatase inhibitor okadaic acid. Treatments that increase protein modification will increase the number of antibody reactive proteins as well as the intensity of reactivity. The J* specific antibodies will react with a relatively small number of proteins from control extracts but will react with a very large number following treatment with the selected inhibitor. The antibodies will show no reactivity with the inactive-non-modified versions of these proteins, demonstrating a high degree of J* specificity and suggesting broad cross-reactivity to many different modified-target containing proteins. [0064]
(7) The degree of context-independence may be more carefully examined, for example, by ELISA analysis against individual J* peptides that are mixed together or tested individually. Such analysis can indicate if poor reactivity occurs with certain motifs, such as when J* is followed by proline, for example. [0065]
(8) The context-dependence of J* antibody recognition may be further examined, as in the preferred embodiment, using a immobilized grid of modified-peptide libraries. In addition to a fixed target residue, J*, each different library is synthesized to contain an additional fixed amino acid at different positions relative to J* but with all other positions containing all 20 amino acids except cysteine. Each peptide library is coated, for example, on the bottom of an ELISA well and exposed to the J* antibodies. Antibodies that do not react with a particular spot (peptide library) on the grid do not bind when the specified amino acid is present at the specified position. This analysis determines whether or not a particular amino acid at a particular position relative to J* will allow or block binding. [0066]
Alternatively, purified antibodies can be linked to beads, allowed to bind the modified or unmodified library, unbound sequences washed away, and bound sequences recovered and subject to amino acid sequencing to determine the amount of each amino acid present at each position in the library. This information will indicate what amino acids are tolerated at each position. [0067]
(9) Monoclonal antibodies may be prepared, as in one form of the preferred embodiment, by coupling the J* peptide library to a suitable carrier, such as KLH, and injected into a host, such as BalbC mice. The J* peptide-KLH conjugate may be emulsified in Freund's adjuvant and booster injections in incomplete Freund's adjuvant may be carried out every other week until a response is obtained. [0068]
(10) Antibody titre is measured by a suitable method, such as ELISA against J* and non-J* peptide libraries. Sera from hosts showing high-titre responses are adsorbed with immobilized non-J* peptide and the nonadsorbed fraction tested by, for example, western blotting. [0069]
(11) Spleens from hosts showing J*-specific responses are fused to myeloma cells and hybridoma clones are selected and screened. Supernatants from individual clones are screened first for their ability to bind the J*-peptide library. Positive clones are next screened for their cross-reactivity against the non-J* library. Clones showing the highest degree of J*-specificity are chosen for further analysis as described above in steps (5) through (8). [0070]
(12) Overproduction of monoclonal antibodies resulting from step (11) above may be carried out, for example, by harvesting ascites, culturing selected hybridoma clones, or cloning into a host organism, such as [0071] E. coli.
Further Binding Characterization [0072]
Motif-specific, context-independent antibodies produced and characterized as described above may be further characterized by oriented peptide analysis, in order to determine their precise binding specificity. Such determination is preferably conducted prior to the use of the antibodies in the positive phosphoprotein identification methods described herein. The oriented peptide technique has been described for the determination of optimal motifs phosphorylated by a particular kinase. See U.S. Pat. No. 5,532,167, Cantley et al., Issued Jul. 2, 1996. [0073]
For oriented peptide analysis, a desired motif-specific, context-independent antibody is used to select, from a target peptide mixture in a diverse library, those modified peptides with binding sequences preferred by the antibody, in order to determine the precise specificity of this antibody (i.e. the particular motif residues required in addition to the phosphorylated residue(s)). The, binding preference of the antibody may not include all of residues of the target motif employed in the degenerate peptide library used to generate the antibody. Therefore, determining precise binding preference allows for a maximally stringent database search, as described below. [0074]
The test antibody is immobilized, for example by mixing with a slurry of pre-swelled protein A beads, and the mixture incubated and washed according to standard methods. The beads with bound antibody may then be transferred to an appropriate column, for example a microspin column (BioRad) and a highly diverse peptide library containing the putative target motif (e.g. as described above) applied to the column. The column is incubated to allow for peptide binding, and then washed and eluted according to standard protocols. See e.g., Yaffe, [0075] Methods in Enzymology 328: 157-170 (2000). Peptides bound by, and eluted from, the immobilized motif-specific, context-independent antibody are evaporated to dryness and the pellet then resuspended to allow for peptide sequencing by standard methods on an automated sequencer. Peptide library screening is also done with appropriate controls, such as an irrelevant antibody (e.g. an antibody that recognizes a motif not present in the peptide library or an antibody specific for a phosphorylated residue not present in the library) and/or protein A beads alone.
The recovered peptides are analyzed by automated amino acid sequencing according to standard protocols. The abundance of each amino acid at a given cycle in the sequence of the peptide mixture bound by the motif-specific, context-independent antibody is divided by the abundance of the same amino acid in the same cycle of the starting peptide library. These raw preference values are then summed and normalized to the total number of amino acids in the particular degenerate position (e.g. 18, if all amino acids except Cysteine and Tryptophan are included at each position in the peptide library). These preference values may be plotted if a visual representation of amino acid preference at each antigen position is desired. From these data, the strongest selection of a given amino acid at a particular position by the test antibody may be determined. For example, for the phospho-(Ser/Thr) Akt consensus substrate motif antibody disclosed herein, it was determined that the antibody prefers strongly prefers arginine at −3 position (see Example VI (b)). An optimal binding motif for the antibody may then be estimated. [0076]
The absolute requirement or strong preference for other putative antigen residues in the motif may further be determined by construction of a second peptide library with the strongest selected amino acid from the first peptide library selection fixed to determine the preference for additional amino acids in other positions, and the same procedure outlined above repeated to determine binding requirements at these putative positions flanking the phosphorylated residue of the motif. For example, further specificity testing with the phospho-(Ser/Thr) Akt antibody of the invention using a secondary peptide library with arginine fixed at position −3 indicates that the antibody further selects for arginine at the −6, −5, and −4 positions relative to the phosphorylated residue. In this way, the precise binding specificity or preference of the motif-specific, context-independent antibodies disclosed herein may be determined. [0077]
Alternatively, ELISA analysis of synthetic peptides using an antibody of the invention may also be employed to determine the optimal binding motif sequence for the antibody. Such ELISA analysis may be used to confirm the binding specificity or preferences determined by oriented peptide analysis, as described above. ELISA analysis in this matter can indicate whether the binding requirements of the antibody for a particular residue are more stringent than another required residue, or whether alternative residues are permitted. Again, by way of example, ELISA analysis for the phospho-(Ser/Thr) Akt consensus substrate motif antibody indicated an absolute requirement for arginine at −3 position, a somewhat weaker requirement for arginine at the −5 position, some preference for hydrophobic amino acids at the +1 position, and the fact that lysine can substitute arginine to some extent. [0078]
The relative preference values (via oriented peptide analysis) and ELISA analysis for antibody binding specificity can be used to generate sequence profile motif matrices in which favorable or unfavorable contributions from individual amino acids are weighed quantitatively to predict the likelihood of a sequence recognized by a particular motif-specific, context-independent antibody. These analyses are particularly desirable where these antibodies will be used, in conjunction with database searching, to positively identify cellular phosphoproteins, as described below. The determination of the precise binding preference (e.g. specificity) of the motif-specific antibodies enables the conduct of stringent database searching in order to minimize the number of “false match” sequences that comprise target motif residues not, in fact, bound by (required by) the antibody. [0079]
Accordingly, in a preferred embodiment of the invention, the method for the positive identification of phosphoproteins disclosed below comprises the step of determining the precise binding preference of the motif-specific, context-independent antibody prior to conducting the database searching. [0080]
Detection Methods [0081]
The motif-specific, context-independent antibodies produced by the method of the invention may be used to identify an unknown substrate of an enzyme. Such antibodies are first generated against a motif that is recognized by the enzyme of interest, for example, a consensus site. These antibodies are then used to screen a sample for the presence of other, unknown substrates that contain the same motif. This method enables the rapid detection of important new substrates in a variety of cascades that involve conserved substrate motifs. For example, antibodies that selectively recognize a wide variety of proteins only when phosphorylated at the MAPK consensus phosphorylation site would greatly facilitate the detection of new MAP kinase targets. MAP kinase could be overexpressed in cell culture, activated by growth factors, and target substrate proteins identified by western blotting using antibodies that selectively recognize the phosphorylated substrate proteins (Stukenberg et al., [0082] Curr. Biol. 7:338-348 (1997). Alternatively, MAPK could be used to phosphorylate cDNA expression libraries in vitro and MAPK consensus-site antibodies used to identify cDNA clones expressing MAPK phosphorylated substrates (Funkunaga and Hunter, EMBO 16(8):1921-1933 (1997).
Similarly, antibodies produced by the method of the instant invention may be used to identify an enzyme that modifies a known substrate motif. Such antibodies, whether specific for modified (e.g. phosphorylated) or unmodified (e.g. zinc finger) motifs, can be used to detect whether a certain enzyme of interest has modified a substrate, which contains that motif. This method allows for the rapid detection of important new proteins that act on known classes of substrates containing contain conserved motifs, for the example MAPK consensus site. [0083]
The motif-specific, context-independent antibodies of the instant invention may also be used in vitro as reagents in high-throughput assays, such as drug screens, to detect the enzymatic modification of certain substrates containing a conserved motif. For example, antibodies specific for a certain phosphorylated motif enable the rapid detection of inhibitors of the enzyme that act at that motif. In the case of a drug screen, a single motif-specific antibody can be used to assay the activity of a wide range of enzymes acting at many diverse sequence motifs. Phosphotyrosine antibodies are currently employed in high throughput kinase assays to screen for selective, high affinity tyrosine kinase inhibitors. Compounds or drugs that block enzyme activity are detected by their ability to inhibit kinase activity as determined by a reduction of phosphotyrosine antibody binding to phosphorylated substrate. Similar assays can be set up to screen for pharmaceutically useful compounds using antibodies produced as described above for phosphoserine, phosphothreonine, or antibodies detecting other protein modifications. [0084]
Antibody based detection of protein kinase activity has several advantages over radioactive assays for use in automated high throughput kinase assays. First, radioactive assays are difficult to automate because they employ transfer of 32-P gamma-labeled ATP to a peptide substrate. The phosphopeptide is then separated from labeled ATP using phosphocellulose filters and several washing steps, and finally, phosphorylation is quantitated by liquid scintillation methods. Together these steps are time consuming and difficult to automate. Antibody detection allows a wide variety of ELISA-type assays that are well suited for automation and high throughput screens. [0085]
Second, radioactive assays require low levels of ATP to insure high levels of 32-P incorporation for maximal sensitivity. Low levels of ATP in the kinase assay bias the search for inhibitors towards compounds that compete with ATP binding in the protein kinase catalytic cleft. Such screens consistently yield competitive inhibitors at the ATP binding site, which due to the highly conserved nature of this binding site results in inhibitors with poor selectivity. [0086]
Current high-throughput kinase assays typically utilize biotinylated peptide substrates immobilized on the bottom of a 96 or 386 well plate that is subsequently incubated together with the desired protein kinase, ATP, and the appropriate kinase buffer. Kinase activity is measured using a fluorescently labeled phosphospecific-antibody that reacts only with the phosphorylated peptide substrate. These assays come in two formats homogeneous (not involving wash steps and heterogeneous (involving wash steps). Homogeneous fluorescent assays typically utilize lanthanide-labeled phosphoantibody binding to a phosphorylated peptide substrate that has linked to it an energy acceptor, for example allophycocyanin. Binding of the phosphoantibody the phosphorylated peptide substrate brings the two fluorophores close enough together to allow fluorescence resonance energy transfer to occur shifting the frequency of the emitted signal, indicating the presence of a biomolecular complex. Different compounds are added to each well and the ability of the compound to inhibit substrate phosphorylation is determined by inhibition of fluorescence energy transfer. This format is similar to the scintillation proximity assay commonly used in radioactive assays. Other homogeneous assays involve the use of fluorescence polarization to measure the binding of phosphoantibody to phosphorylated substrate. [0087]
The key feature in the homogeneous assays is the limited number of steps and the ease in automation. A large variety of heterogeneous kinase assays based upon ELIZA formats are also currently in use. These assays typically utilizing fluorescently labeled phosphoantibodies binding phosphorylated peptide substrates that are immobilized in 96 or 386 well formats. In this case wash steps are required to separate bound from unbound antibody. Fluorescently labeled antibody retained in the well is then detected using time resolved fluorescence. [0088]
The motifs used to generate antibodies for such modification screening assays may be either modified or unmodified substrate motifs. Antibodies generated against unmodified motifs will not bind if the substrate has been subsequently modified by an enzyme. Similarly, antibodies generated against modified motifs can detect increases in modified substrate concentrations owing to enzymatic activity. [0089]
Similar approaches may be applied to study a variety of other enzymatic modifications, and are not limited to the protein kinase or acetyltransferase activities discussed below. For example, the approach could be used to generate antibodies that recognize many other types of protein modification, including, but not limited to, the addition of sugars, methyl groups, carboxyl groups, the addition of various lipids, or the addition of nucleotides, or polymers of nucleotides, nucleosides, or amino acids such as ubiquitin. [0090]
Genome-wide Profiling [0091]
Likewise, motif-specific, context-independent antibodies of the invention may be used on a genome-wide scale to simultaneously profile large and diverse protein populations that contain conserved (i.e. shared) motifs. A specific two or three amino acid binding site, for example consecutive arginine residues, should appear (based upon a random distribution of amino acids) once every 400 or 8000 residues, respectively, (equating to approximately once per protein, or once every 20 proteins, respectively, (assuming the average protein is 400 amino acids)). Thus, an antibody specific for such a motif independent of the context in which it occurs allows for the rapid screening of a great number of proteins. [0092]
Phosphorylation specific antibodies allow genome wide profiling of changes in phosphorylation of proteins as a result of drug treatment or the overexpression of specific genes/proteins as a result of such treatment. Such antibodies also facilitate the profiling of expression of specific proteins in sequenced genomes. [0093]
For example, suppose that a drug is developed which inhibits the cell-cycle dependent protein kinase cdc2. The drug has been shown to inhibit cdk2 with high affinity, but the specificity of the compound needs to be further tested to examine whether other protein kinases are inhibited and if so, which ones. As an early step in this process cell lines may be treated with the drug and the effects on total cell protein phosphorylation monitored using a panel of motif-specific and general phosphoantibodies to examine the nature of the phospho-substrates inhibited by the compound or lead drug. [0094]
Total protein from cell extracts prepared from control or drug treated cells may be fractionated using, for example, 2-dimentional gels (isoelectric focusing in the first dimension and standard SDS-polyacrylamide molecular weight fractionation in the second dimension), transferred to nitrocellulose membranes, and analyzed by western blotting using, in this hypothetical case, kinase consensus site-specific phosphoantibodies. [0095]
In this case, global analysis of total cell proteins using a cdc2 consensus substrate motif-specific antibody would provide information regarding the ability of the drug to block phosphorylation at all potential cdc2 site substrates. The pattern of inhibition at other non-cdc2 substrates (i.e. the degree of specificity) could also be examined using antibodies to different kinase consensus sites, or using antibodies to phosphotyrosine to determine whether the inhibitor also acts to block tyrosine kinases. [0096]
Positive Identification of Cellular Phosphoproteins [0097]
Motif-specific, context-independent antibodies of the invention will prove highly useful in enabling new methods for the positive and facile identification of modified proteins, such as phosphoproteins (and/or their modification sites) within a given cell or tissue at a particular point in time (e.g. during progression of a disease). Until now, such facile identification of particular phosphorylated proteins in a cell has not been possible. Currently, for mammalian cells, the identity of the majority of protein “spots” visualized on 2-D gels are unknown. However, as all human genes are identified and sequenced and the corresponding proteins characterized and “spots” identified, analysis by protein profiling in accordance with the present invention will become even more powerfully informative. The identity of the proteins inhibited will not only confirm the drug specificity but the identity of additional “nonspecific” proteins inhibited will also suggest possible side effects. Identical analysis can be carried out in simpler, completely sequenced organisms, such as yeast where many of the protein “spots” on 2-D gels have already been identified. [0098]
In accordance with the present invention, it is now possible to positively identify cellular phosphoproteins (or their phosphorylation sites) detected, for example, by 2-D gel analysis (separation on basis of mass in one direction and isoelectric point in the other) of a proteinaceous sample, e.g. a cellular extract, using one or more motif-specific, context-independent antibodies of the invention. The novel class of antibodies disclosed herein enables the initial identification of particular phosphorylated proteins which contain one or more specific, modified motifs recognized by one or more motif-specific, context-independent antibodies. If these bound proteins have also been separated on the basis of mass/molecular-weight (for example by 1-D gel analysis) and/or isoelectric point (PI), then the parameters of both mass and/or PI and presence of particular phosphorylated motif sequence(s) are initially known for a particular fractionated protein. [0099]
These parameters can then be used in a search program (software), such as ScanSite (http://scansite.mit.edu/), which is suitable for searching publicly-accessible protein databases (e.g. SwissProt, NCBI none-redundant protein sequence databasegen, etc.) to identify reported proteins known to contain the necessary sequence(s) (i.e. the modified motif(s)) and the requisite mass and/or PI. The ability to initially identify a plurality of different proteins containing a common phosphorylated motif is particularly powerful since a great number of phosphorylated proteins are present in a typical cell, many of which will contain conserved phosphorylated signal transduction motifs. Present methods for attempting to determine the identity of particular phosphorylated proteins are tedious and time-consuming, and thus are unsuitable for high-throughput screens, etc. [0100]
Preferably, the precise binding preferences/specificities of the motif-specific, context-independent antibodies employed have first been determined as described above (see “Additional Characterization” section). Thus, the precise modified sequence motif recognized and bound by the antibody is known, and, therefore, phosphoproteins bound by the antibody necessarily contain such motif(s). Determination of the precise binding specificity of the antibody prior to carrying out the database searching allows for a maximally stringent database search, in order to identify only putative proteins containing the sequence necessarily bound by the antibody. Identification of “false positive” proteins in the database—which contain variations of the target motif sequence having residues not actually required or preferred by the antibody—may therefore be avoided, resulting in a more accurate identification of phosphoproteins. [0101]
In one embodiment, the invention provides a method for the positive identification of cellular phosphoproteins using motif-specific, context-independent antibodies coupled with database searching, the method comprising the steps of: (a) fractionating a preparation comprising a plurality of phosphoproteins on the basis of mass (molecular weight) and/or isoelectric point (PI); (b) contacting the fractionated preparation of step (a) with at least one motif-specific, context-independent antibody that binds a phosphorylated motif of known sequence, thereby to detect phosphoproteins comprising the motif; and (c) utilizing a search program to search at one or more protein database(s) for reported proteins having a sequence comprising the motif and having a mass and/or PI substantially matching that determined in step (a), thereby to positively identify at least one cellular phosphoprotein in the preparation as a reported protein of known sequence and mass and/or PI. In a preferred embodiment, the precise target sequence preferred by the motif-specific, context-independent antibody employed in step (b) has first been determined, thereby allowing a more stringent search in step (c) for reported proteins having a sequence comprising the preferred target sequence (as well as the requisite mass and/or PI). In another embodiment of the method, step (b) comprises contacting the fractionated preparation of step (a) with a plurality of motif-specific, context-independent antibodies that each bind a different phosphorylated motif of known sequence, thereby to detect at least one phosphoprotein comprising two or more of the different motifs, and the search of step (c) is for reported proteins having a sequence comprising the two or more different motifs detected in step (b). [0102]
In another embodiment, the invention provides a method for the positive identification of cellular phosphoproteins using motif-specific, context-independent antibodies coupled with database searching, said method comprising the steps of: (a) fractionating a cellular preparation comprising a plurality of phosphoproteins on the basis of mass and/or PI; (b) contacting the fractionated preparation of step (a) with a plurality of motif-specific, context-independent antibodies that each bind a different phosphorylated motif of known sequence, wherein the precise target sequences preferably bound by said antibodies have been determined, thereby to detect at least one phosphoprotein comprising two or more of said precise target sequences; and (c) utilizing a search program to search one or more protein database(s) for reported proteins having a sequence comprising the two or more precise target sequences detected in step (b) and having a mass substantially matching that determined in step (a), thereby to positively identify at least one cellular phosphoprotein in said preparation as a reported protein of known sequence and mass and/or PI. [0103]
The disclosed identification method represents a significant advance over conventional methods for identifying modified cellular proteins. In particular, the method is more sensitive than traditional methods of phosphopeptide identification since it based on western blot or immunoprecipitation, which are the most sensitive means detect the exist of proteins (less than fmole level). In contrast, traditional methods rely on the purification of phosphopeptides, followed by sequence identifications, which require at least 10 fmole level of proteins. The sequencing step is time consuming and costly. Moreover, the requirement of traditional methods for higher amounts of phosphoprotein is an important limitation, since most phosphopeptides relevant to signal transduction and signaling molecules are present in cells at low abundance. The present method has the further advantages of being readily practiced by simple laboratories that do not have HPLC, peptide sequence instruments, and/or mass spectrometry instrumentation. [0104]
The positive identification methods disclosed herein are not limited to the identification of phosphoproteins, but may be applied to identify other modified proteins containing conserved signaling motifs, such as acetylated, nitrosylated, glycosylated, or methylated motifs. In such cases, motif-specific, context-independent antibodies against the desired modified motif(s) are produced as provided herein. [0105]
Protein Preparations: [0106]
Biological preparations suitable for use in the method of the invention may be any phosphoprotein-, or phosphopeptide-, containing preparation from a cell, tissue, or biological fluid, etc. of interest. For example, the preparation may be obtained from bacteria, yeast, worms, amphibia, fish, plants, parasites, insects, or mammals. In a preferred embodiment, the organism is a mammal. In another preferred embodiment, the mammal is a human. The method can be applied to a cellular preparation from one or more cell types or fluid samples derived from any organism. Cellular preparations may be obtained, for example, by growing cells in tissue culture according to standard methods, harvesting the cells from culture media by centrifugation, and lysing the cells by sonication or other standard means of opening cells. [0107]
Cellular preparations may also be obtained directly from tissue samples. In a preferred embodiment, the tissue sample is a biopsy sample. These small pieces of living tissue, typically weighing less than 500 milligrams, are taken directly from an organism and used directly without growth in tissue culture. The use of such living tissue allows direct analysis of the biological state of the tissue without introducing artifacts that may arise as a consequence of growth in culture. Any desired cell type from a given organism may be utilized. For example, tumor cells (e.g. from breast, prostate, etc.) may be cultured or obtained by biopsy to study proteins with roles in cancer. Neural cells lines are available to characterize proteins involved in neurotransmission. Fat cells can be cultured or obtained by biopsy to study proteins involved in the hormonal mechanisms of fat deposition. Cellular preparations from tissue samples may contain peptides or proteins from multiple cell lines or types. In addition, cell lines with specific, desirable features could be engineered genetically, e.g., to overexpress a protein thought to have an important regulatory role in a specific pathway, e.g. cell lines overexpressing Akt protein. In other preferred embodiments, cellular preparations are obtained from bodily fluids, such as serum, urine, spinal fluid, or synovial fluid. Preparations from blood samples may also be employed, whether certain cells, e.g. erythrocytes, are first removed. [0108]
Preparations are obtained by standard methods, e.g. for cells and tissues, by sonication, homogenization, abrasion, enzymatic digestion, or chemical solubilization. Generally the method used to lyse cells will be the one most commonly used for that specific cell type, e.g., enzymatic lysis for bacteria, abrasion for plant cells, and sonication for animal cells, but other desired methods may be suitably employed. Cellular preparations for use in the method of the invention need not be extensively purified prior to practice of the positive identification methods disclosed herein. For example, urine samples or serum samples may be directly analyzed. This allows less sample processing, which increases the likelihood of identifying low-level modifications and makes it less likely that fractionation methods will bias or skew the profile of experimentally assigned modifications. [0109]
The mixture can be a crude cell lysate (for example, from tissue culture or a biopsy, or serum), a partially fractionated lysate (for example, a highly purified membrane or organelle), or a known and well-defined composition (for example, a in vitro modification reaction, that is, a protein modification enzyme allowed to react with one or more substrate proteins). However, if desired, simple purifications may be carried out to remove non-protein elements and/or non-signaling, structural proteins by standard methods, e.g. by centrifugation to remove erythrocytes, ultracentrifugation to remove cellular debris and cytoskeletal proteins, or by treatment with class-specific enzymes such as nucleases to remove DNA and RNA. In preferred embodiments, the cellular preparation is a crude cell extract or fluid, which has not been extensively purified, or a partially-purified cell extract. [0110]
Preferably, cellular preparations are obtained so as to reflect the baseline, in vivo activation state, e.g. phosphorylation state, of proteins in a given cell, e.g. a breast cancer cell. However, preparations may be obtained from cells or organisms pre-treated with inducers. For example, cells grown in tissue culture can be exposed to chemicals such as calyculin or okadaic acid, which broadly elevate cellular phosphoprotein levels by inhibiting cellular phosphatases. Alternatively, a considerably narrower and more specific set of phosphoproteins in pathways can be induced by treatment with hormones, such as epidermal growth factor, that activate certain signaling pathways. Organisms can also be treated with drugs or infectious agents, and the effects of these treatments can be evaluated by isolating and analyzing specific tissues or fluids from the organism, both before and after treatment with a test compound. [0111]
Accordingly, in one preferred embodiment of the disclosed method, the cellular preparation is from a diseased cell or a cell susceptible to disease. In another preferred embodiment, the disease is cancer. The method of the invention will particularly useful in identifying changes in the phosphorylation of particular proteins (e.g., signal cascade proteins) resulting from treatment with a test compound, for example a drug targeting a particular kinase activated in a particular disease. Accordingly, in another preferred embodiment, the cellular preparation is from a cell treated with a test compound. In a preferred embodiment, the test compound is a kinase inhibitor. [0112]
Fractionation by Mass and/or PI [0113]
Preparations containing phosphoproteins and/or phosphopeptides to be identified may be fractionated on basis of the molecular weight (mass) and/or isoelectric point (PI) of the proteins or peptides by any suitable method well known in the art. Typically, as in a preferred embodiment, proteins will be separated by gel electrophoresis. This fractionation step may be done singly, or may be coupled with additional fractionations on the basis of other protein characteristics, such as isoelectric point. In a preferred embodiment of the method, the fractionation step (a) comprises fractionation on the basis of mass and on the basis of isoelectric point, for example by 2-dimensional gel electrophoresis. Other well-known mass fractionation methods may also be suitably employed, such as size-exclusion column chromatography, mass spectrometry of proteins and/or peptides. [0114]
Initial fractionation by mass and/or PI provides a significant decrease in the universe of possible matching proteins ultimately identified in the database search, and thus contributes to the efficiency of the method. For example, if the total human database has 90,000 proteins, the possible candidate proteins will reduce by about a factor of 10 using the mass constraint alone. The possible protein matches are further constrained (e.g. to one unique protein) by additionally selecting for, e.g. a three amino acid motif present in detected proteins of known mass. This is true because each amino acid constraint will reduce the total search hits by a factor of 20. Hence, the total number of protein hits (i.e. candidate phosphoproteins) ultimately resulting from the database search is represented by 90000*1/10*1/20*1/20*1/20 (or the total human proteome, narrowed 10-fold by mass, and a further 20-fold for each amino acid in the imposed motif constraint(s)). [0115]
In a preferred embodiment of the disclosed method, the phosphoprotein-containing preparation is first fractionated on the basis of mass and/or PI, and then contacted with at least one motif-specific, context-independent antibody to detect phosphoproteins containing the target motif. However, these steps may alternatively been done in reverse, i.e. the cellular preparation may first be reacted with one or more motif-specific antibodies in order to detect phosphoproteins or peptides containing the target motif, and then separated on the basis of mass, for example, as done with immunoprecipitation followed by separation on gel. This is particularly useful to confirm that a protein with a given mass that is seen by several motif-specific antibodies is the same protein. This can be accomplished by immunoprecipitation with one motif-specific antibody and separation on gel followed by western blot using other motif antibodies. [0116]
In another preferred embodiment, motif-containing cellular phosphopeptides may first be isolated from crude peptide mixtures, such as protease-digested cell lysates, using the immunoaffinity isolation methods described in U.S. Ser. No. 60/337,012, Rush et al., (co-pending), the disclosure of which is incorporated herein in its entirety. Briefly, Rush et al. provide a method for isolating a modified peptide from a complex mixture of peptides (such as exists in a cell extract digest) by the steps of: (a) obtaining a proteinaceous preparation from an organism, in which modified peptides from two or more different proteins are present; (b) contacting the proteinaceous preparation with at least one immobilized modification-specific antibody (e.g. a motif-specific, context-independent antibody); and (c) isolating at least one modified peptide specifically bound by the immobilized antibody. Phosphopeptides specifically isolated by this method contain a known phospho-motif sequence, thus providing a sequence parameter as provided in step (b) of the present invention. In this case, the mass of the protease digested peptides are accurately determined by mass spectrometry. Using a plurality of different motif-specific, context-independent antibodies for immunoaffinity isolation of phosphopeptides according to the method of Rush et al. would thus provide multiple sequence parameters which, together with the molecular weight parameter identified by mass spectrometry, may then be used for the database screening of step (c) as described herein, in order to positively identify the parent cellular phosphoprotein(s) from which the identified phosphopeptides are derived. [0117]
Accordingly, in another embodiment, the invention provides a method for the positive identification of cellular phosphoproteins using motif-specific, context-independent antibodies coupled with database searching, said method comprising the steps of: (a) fractionating a cellular preparation comprising a plurality of phosphopeptides by immunoaffinity isolation using at least one motif-specific, context-independent antibody that binds a phosphorylated motif of known sequence, wherein the mass of bound peptides is also determined; and (b) utilizing a search program to search one or more protein database(s) for reported proteins having a sequence comprising said motif and having a mass substantially matching that determined in step (a), thereby to positively identify at least one cellular phosphoprotein in said preparation as a reported protein of known sequence and mass. [0118]
In certain preferred embodiments of the method, the motif-specific, context-independent antibody binds a motif comprising a kinase consensus substrate motif or a protein-protein binding motif. In another preferred embodiment, the kinase consensus substrate motif or protein-protein binding motif comprises one or more phosphorylated amino acids. Particularly preferred antibodies and motifs are described below. [0119]
Antibodies: [0120]
Any desired motif-specific, context-independent antibody may be used in the practice of the disclosed methods to positively identify cellular phosphoproteins that contain a modified (e.g. phosphorylated) motif sequence specifically bound by the antibody. One or more different motif antibodies may be employed, either simultaneously, or in series, to screen for phosphoproteins containing multiple (e.g. two or more) known motifs required by each motif-specific antibody. Preferably, the antibodies used have previously been further characterized to determine precise binding specificities, as described above. In a preferred embodiment of the present method, a plurality of motif-specific, context-independent antibodies are used in step (b), each antibody specifically binding a different phosphorylated motif, so as to initially detect phosphoproteins that contain two or more distinct phosphorylated motifs, and wherein the search of step (c) is for reported proteins comprising two or more (or all) of the different motifs detected in step (b). This type of initial screening with multiple motif-specific antibodies enables the more efficient subsequent searching of protein databases, since fewer “leads” will be obtained that contain both necessary sequence elements and the required mass. [0121]
As described at length above, the novel class of motif-specific, context-independent antibodies disclosed herein includes virtually any short, conserved signaling motif comprising one or more modified amino acids, such as phosphorylated, methylated, or acetylated amino acids. In preferred embodiments of the phosphoprotein identification method, the motif-specific, context-independent antibody binds a motif selected from the group consisting of a kinase consensus substrate motif and a protein-protein binding motif. Particularly preferred kinase consensus substrate motifs include, but are limited to MAPK consensus substrate motifs, CDK consensus substrate motifs, PKA consensus substrate motifs, bulky rind-directed kinase consensus substrate motifs, and Akt consensus substrate motifs. Particularly preferred protein-protein binding motifs include, but are not limited to, 14-3-3 binding motifs, PDK1 docking motif. Other preferred motifs include signal transduction (i.e. cell signaling) motifs comprising at least on acetylated amino acid, such as acetylated lysine. [0122]
The detection (in step (b)) of motif-containing phosphoproteins present in the cellular preparation can be carried out simultaneously, for example by carrying out a single western blot using different fluorescence dyes conjugated to each of multiple primary motif-specific, context-independent antibodies. Motif and kinase substrate antibodies can be formulated according to standard techniques so as to optimize the single western analysis. [0123]
Database Searching: [0124]
Any software program or algorithm suitable for the automated searching of protein databases to identify putative matches with published proteins based on mass and/or PI and sequence parameters (i.e. inputs) may be employed in the practice of the disclosed methods. Exemplary programs include, but are not limited, to, the programs Scansite or Prosite, which search sequence databases based on short liner amino acid sequences or motifs present in proteins. The invention is not limited to these exemplary search programs, but includes in its scope equivalent programs either presently available or subsequently developed. For example, it would be desirable to develop new programs to search sequence databases using multiple motifs and mass and/or PI constraints in accordance with the present invention. Such developed software is within the scope of the present invention. In a preferred embodiment of the method, the search program of step(c) comprises Scansite search software. [0125]
Any public or private database compiling sequence information on proteins, and most desirably phosphoproteins, may be searched. Preferably, large publicly accessible databases such as Swissprot or NCBI none-redundant protein sequence databasegene are searched. However, private or more limited protein or peptide databases are also within the scope of the invention. The searching of multiple databases is most preferable, in order to maximize the number of possible matching proteins. [0126]
As discussed above, searching may be most efficiently conducted if multiple different motif-specific, context-independent antibodies have been employed to screen for the presence of phosphoproteins comprising each of the sequence motifs recognized by each of the different motif antibodies. Using multiple antibodies increases the stringency of the mass/PI plus sequence parameters, thereby resulting in fewer “hits” or matches resulting from the database search and more accurate identifications. Accordingly, in some preferred embodiments, a plurality of motif-specific, context-independent antibodies that each bind a different phosphorylated motif of known sequence are employed to detect at least one phosphoprotein comprising two or more of the different motifs. The subsequent search step (c) is then conducting for reported proteins having a sequence comprising the two or more different motifs (and the requisite mass and/or PI). [0127]
Following the identification of reported proteins (in the database(s)) having a sequence comprising the selected motif(s) and having a mass and/or PI substantially matching that determined in step (a)—thereby positively identifying at least one cellular phosphoprotein in the preparation as a reported protein of known sequence and mass and/or PI—suitable verification may optimally be carried out to confirm the phosphoprotein identification. Accordingly, in a preferred embodiment, the disclosed method further comprises the step of (d) verifying the identity of the phosphoprotein as determined in step (c) by contacting the phosphoprotein with at least one antibody specific for the protein in phosphorylated or unphosphorylated form. By way of example, if the identified phosphoprotein is determined to be phosphorylated EGFR, this identity may be verified by detecting (e.g. by Western blot) EGFR or phospho-EGFR using a suitable EGFR-specific antibody. Preferably, the verification will be carried out by a second Western blot using the original gel in which the cellular preparation was fractionated (in this example, the EGFR would be present at a particular band, previously identified with a motif-specific antibody). The identity of phosphoprotein as EGFR can be also determined by immunoprecipitation using EGFR specific antibody followed by western blotting by the motif-specific antibody. In a preferred embodiment, immuno-depletion can be used to confirm the phosphoprotein identified by the motif antibody coupled with database searching. In the example of EGFR, the cell lysates used in identification of the phosphoproteins can be immuno-depleted by EGFR specific antibody, and the supernatant of the depleted lysates can be fractionated by gel and probed with motif-specific antibodies. The specific mass corresponding to EGFR will be eliminated. [0128]
The disclosed methods for the positive identification of cellular phosphoproteins are expected to be of great value, inter alia, in profiling the phosphorylation (activation) status of particular phosphoproteins in cells under particular circumstances or at a particular point in time. For example, phosphorylation status of particular proteins relevant in one type of cancer may be identified as relevant in another type of cancer. Similarly, phosphorylation status (or phosphoprotein level) of particular proteins may be examined before and after treatment with a test compound, for example a kinase inhibitor, to examine the effect of treatment on particular phosphoproteins, or to identify other phosphoproteins activated or inhibited by such treatment. High throughput screens employing the method of the invention may be designed to rapidly determine the identity of particular phosphoproteins in diseased cells or tissues of interest, and to examine the status of such phosphoproteins at a particular point in time or in response to a particular treatment or stimulus. [0129]
The Examples presented below are only intended to exemplify the invention in specific preferred embodiments and are not intended to limit the scope of the invention except as provided in the claims herein. The present invention encompasses modifications and variations of the methods taught herein which would be obvious to one of ordinary skill in the art. [0130]

EXAMPLE I

Context-Independent Phosphothreonine Antibodies

Synthesis of Peptide Library Antigens: [0131]
Phospho-specific antibodies that react with any protein containing phosphorylated threonine residues, i.e. that bind phosphothreonine independently of the surrounding amino acids, were obtained by synthesizing a highly degenerate peptide library XXXXXXThr*XXXXXXC where X=all 20 amino acids except cysteine and Thr*=phosphothreonine. [0132]
The phosphothreonine peptide library was synthesized by standard F-Moc solid phase peptide synthesis using an ABI peptide synthesizer and using mixtures of each amino acid during degenerate coupling reactions. Degenerate peptides were synthesized using an ABI model 433A peptide synthesizer, using FastMoc chemistry (Fields et al., [0133] Pept. Res. 4:95-101 (1991), hereby incorporated by reference herein) at a scale of 0.085 mmol. Fmoc/NMP chemistry utilizing HBTU amino acid activation (Dourtoglou et al., Synthesis 1984: 572-574 (1984), Knorr et al., Tetra. Let. 30:1927-1930 (1989), Knorr et al., in Peptides 1988 37-129 (1989), Walter de Gruter & Co.) was employed for all cycles. Preloaded Fmoc-Cys(Trt) HMP (p-hydroxymethylphenoxymethyl) polystyrene resin functionalized at 0.5 mmol/g was used for each degenerate pool of peptides. Peptides were synthesized using single coupling during each cycle, although coupling times were extended at each position containing a phosphorylated amino acid. The final Fmoc was removed during synthesis. Utilization of preloaded HMP resin along with final Fmoc group removal yields peptides having both free amino and carboxy termini after cleavage and deprotection.
In order to produce as equal a representation of each amino acid as possible at each degenerate position several rounds of altering the amino acid composition, synthesizing, and peptide sequencing were conducted. The desired peptide pools were to contain an equimolar mix of 19 amino acids (all standard amino acids except Cys) at each “degenerate” site. Because the rate of reactivity of each protected amino acid differs, simply mixing equimolar amounts (each at approximately 5.26% of total) does not result in a population of peptides that is equimolar at each position. In order to maximize degeneracy at each residue, peptide synthesis was first done using equimolar “mixes” at each position. Phenylthiocarbamyl-amino acid analysis was performed therefore allowing assessment of relative amino acid content at each position. Based on amino acid analysis the molar amounts of each amino acid in the “mix” were adjusted to compensate for different reaction rates, in order to ensure equal representation of each amino acid at each degenerate position. Several rounds of peptide synthesis followed by amino acid analysis were necessary to optimize the amino acid mix, which resulted in a totally degenerate peptide. The optimized amino acid mix arrived at was as follows: G (4.6%); A (5.6%); V (3.3%); L (2.5%); I (4.25%); S (4.4%); T (8.4%); F (2.25%); Y (6.0%); W (6.8%); M (2.9%); P (2.5%); D (5.8%); N (9.5%); E (6.2%); Q (9.4%); K (6.1%); R (6.4%); H (3.5%). [0134]
Cleavage of the degenerate peptides from the resin along with removal of side chain protecting groups occurs simultaneously upon treatment with TFA. The cleavage mixture (Perkin Elmer, Emeryville, Calif. (1995)) consists of the following: 0.75 g phenol, 0.125 ml methyl sulfide, 0.25 [0135] ml 1,2-ethanedithiol, 0.5 ml milliQ H2O, 0.5 ml thioanisol, 10 ml TFA. The entire mixture was added to the peptide resin (approx. 300 mg). The resin was flushed with nitrogen and gently stirred at room temperature for 3 hours. The resin was then filtered allowing the peptide to be precipitated into cold (0° C.) methyl-t-butyl ether. The ether fraction was centrifuged allowing collection of the precipitate. The peptide precipitate was vacuum dried, analyzed by mass spectroscopy, and HPLC purified.
A sample of the peptide was dissolved in acetonitrile/water (50:50, v/v) and analyzed on a Perceptive Biosystems (Framingham, Mass.) MALDI-TOF mass spectrometer using 2,4,6-trihydroxyacetophenone plus ammonium citrate as the matrix. As expected, the peptide mixture did not show a homogeneous product. MALDI-TOF analysis demonstrated that the peptide pool was degenerate, showing an average mass and the expected statistically normal curve of peptide mass. [0136]
Peptides were purified using a Waters HPLC system consisting of a Lambda-Max Model 481 Multiwavelength detector, 500 series pumps, and Automated gradient controller. A Vydac semi-preparative C18 column was used for reverse-phase purification. A 60 min. linear gradient, 10%-100% B, was used at a flow rate of 2 ml/minute. Buffer A consisted of 0.1% TFA/H[0137] ₂O (v/v) while buffer B consisted of 0.1% TFA/60% CH₃CN/40% H₂O (v/v/v). Detection was at 214 nm.
Because the peptide pool was degenerate (as demonstrated by mass spectroscopy) HPLC purification was not expected to yield a homogeneous product. Base-line separation of peptide mixtures was not achieved by this method and it was only intended as a crude purification/desalting step. Mass spectroscopy was performed and all fractions whose mass was within the theoretical range were pooled and lyophilized. [0138]
Amino acid sequence analysis at several different positions along the peptide indicated a random amino acid representation at each position and that the random representation was maintained throughout the synthesis. The results indicated the production of highly diverse peptide libraries that would serve as suitable antigens. [0139]
Production of Rabbit Polyclonal Antibodies: [0140]
All peptides synthesized contained C-terminal cysteine residues allowing conjugation to the carrier protein (KLH) using the heterobifunctional cross-linking reagent m-maleimidobenzoyl-N-hydroxysuccinimide ester (MBS). The conjugation procedure used was as described by the manufacturer (Pierce), although the amount of peptide coupled to KLH was increased to 10 mg in order to provide increased material for immunization and boosting of animals. Scale-up required use of a larger desalting column (Bio-Rad 10 DG (Cambridge, Mass.)) to remove the excess MBS after reaction to N-termini and the ε-amino group of KLH Lysine residues. [0141]
The phosphothreonine peptide library was covalently coupled to keyhole limpet hemocyanin (KLH) (250 μgrams), emulsified in Freund's adjuvant and injected intradermally into female New Zealand white rabbits. Booster injections (200 μgrams) in incomplete Freund's adjuvant were carried out every other week until a response was obtained. Rabbit sera was screened at three week intervals for the presence of phosphopeptide specific immunoreactivity by ELISA using both the phosphothreonine and nonphosphothreonine peptide libraries. When the titre of antibody against phosphopeptide reached 10[0142] ⁵, rabbits were put on a production bleed schedule with bleeds collected every two weeks. When 40 ml of high titre serum were obtained, purification of phosphospecific antibodies was initiated, as described below.
Antisera from the most promising rabbit was purified over protein A and passed over a nonphospho Thr/Ser peptide library column. The nonadsorbed fraction (flow through) was applied to a phosphothreonine column, eluted at low pH, dialyzed and tested for phosphospecificity by ELISA using phospho- and nonphosphopeptides. Antibodies affinity-purified in this fashion recognized the phosphorylated threonine peptide library but did not react with the nonphosphothreonine/serine library, indicating a high degree of specificity for phosphothreonine (see FIG. 1[0143] a). ELISA results also indicated that the antibodies also reacted specifically with a mixture of 18 different phosphothreonine peptides but showed no reactivity with any of the corresponding nonphosphopeptides (FIG. 1b). The antibodies also exhibited a strict preference for phosphothreonine, showing no reactivity with a mixture of 38 different phosphoserine peptides (FIG. 1b) or peptides containing phosphotyrosine.
We next tested the antibodies by western blotting using cell extracts prepared from cells treated with and without the protein phosphatase inhibitor okadaic acid. As shown in FIG. 1[0144] c the phosphothreonine antibodies react with a relatively small number of proteins from control extracts but react with a very large number following treatment with okadaic acid (see the smear of high Mol Wt. reactive proteins in FIG. 1c, lane 2). The antibodies also reacted specifically with the active forms of MAPK (ERK1) and MKK3 only when phosphorylated at threonine residues at their respective activation loops. The antibodies showed no reactivity with the inactive-nonphosphorylated versions of these proteins (FIG. 1c, lanes 3-6). These results demonstrate a high degree of phosphothreonine specificity and suggest broad cross-reactivity to many different threonine-phosphorylated proteins and peptides.
To examine more carefully the degree of context-independence, ELISA analysis was conducted against individual threonine phosphorylated peptides that were mixed together in the previous experiment. As shown in FIG. 1[0145] a, the phosphothreonine antibody reacts well with all phosphopeptides except those where phosphothreonine is immediately followed by proline, for example the c-Myc and APP1 phosphopeptides (FIG. 2b). These results indicate that purified rabbit antibodies reacted in a phosphospecific manner with a wide variety of phosphothreonine but react only poorly with phosphopeptides where the phosphorylated threonine is followed by proline.
The context-dependence of phosphothreonine antibody recognition was further examined using a immobilized grid of phosphopeptide libraries. In addition to a fixed phosphothreonine, each different library was synthesized to contain an additional fixed amino acid at the −4, −3, −2, −1, +1, +2, +3 positions relative to phosphothreonine but with all other positions containing all 20 amino acids except cysteine. Each peptide library was coated on the bottom of an ELISA well and exposed to the phosphothreonine antibodies. Antibodies that do not react with a particular spot (peptide library) on the grid do not bind when the specified amino acid is present at the specified position. This analysis determines whether or not a particular amino acid at a particular position relative to phosphothreonine will allow or block binding (FIG. 1[0146] d).
Results confirmed that the phosphothreonine antibodies tolerated all amino acids in the −1, −2, −3, −4, and +2, +3 position, and bound equally well to every amino acid except proline at the +1 position (see FIG. 1[0147] d, first row). The reactivity as defined by this binding profile indicates that the antibodies will bind all phosphothreonine-containing sequences except those followed immediately in −1 position by proline. Further analysis using a variety of specific phosphothreonine containing peptides confirmed these results.
Phosphothreonine specific antibodies from several other rabbits immunized with the same peptide library antigens were further purified and characterized. Antibodies purified from sera obtained from two other rabbits also produced broadly cross-reacting phosphothreonine antibodies as determined by ELISA. One rabbit produced antibodies that react equally well with peptides containing proline following the phosphothreonine. Taken together, these results demonstrate the broad context-independence of the phosphothreonine response obtained when combinatorial peptide libraries are used as immunogens. [0148]

EXAMPLE II

Protein Kinase Consensus Site-Specific Phosphoantibodies: MAPK-consensus Recognition Sites: PXS*P

A peptide library of the preferred site for MAPK phosphorylation PXS*P was synthesized (FIG. 2[0149] a) substantially as described in Example I. In addition to an equimolar mix of phosphoserine and threonine, amino acids at two other positions were also fixed; proline at −2 and proline at +1. This library was coupled to KLH and injected into rabbits as described for phosphothreonine. IgG from the most promising rabbit was protein A purified and passed over a nonphospho-Thr/Ser peptide library column. The nonadsorbed fraction (flow through) was applied to a phospho-PXS*P column, eluted at low pH, dialyzed and tested for phosphospecificity by ELISA using phospho- and nonphosphopeptides.
Antibodies affinity purified in this fashion reacted strongly with the phosphorylated PXS*P peptide library but did not react with the nonphosphothreonine/serine library (see FIG. 2[0150] a). ELISA results also indicated that the antibodies also reacted specifically with a mixture of 18 different phosphothreonine peptides but showed no reactivity with any of the corresponding nonphosphopeptides (FIG. 2a). In addition to being phosphospecific, the antibodies exhibited a preference for proline at the −2 and +1 positions and showed no reactivity with phosphorylated peptides that lack proline at this position (FIG. 2a). The antibodies reacted strongly with the RB and cdk4 phosphopeptides but showed no reactivity with the MKK3, PKCalpha, or p70S6 phosphopeptides that lack proline at the +1 position (FIG. 2a). These antibodies do react with some peptides lacking proline at −2, for example the cdk4 phosphopeptide, suggesting that proline at this position is not absolutely necessary.
PXS*P antibodies were further tested by western blotting using cell extracts prepared from cells treated with and without the protein phosphatase inhibitor okadaic acid. Binding of the PXS*P antibodies to cell extracts from [0151] RS 4;11 cells was strongly enhanced following treatment with okadaic acid (smear of high Mol Wt. proteins in FIG. 2b, lane 2). The antibodies also reacted specifically with ATF-2 phosphorylated in vitro with MAP kinase but not the nonphosphorylated form of this protein (FIG. 2b, lanes 3 and 4), demonstrating a high degree of phospho-specificity and broad cross-reactivity to many different phosphorylated proteins and peptides.
The specificity of PXS*P antibody recognition was also examined using an immobilized grid of phosphopeptide libraries. As described above, in addition to a fixed phosphothreonine or phosphoserine, each different library was synthesized to contain an additional fixed amino acid at the −1, +1, +2 positions relative to phosphothreonine but with all other positions containing all 20 amino acids except cysteine. [0152]
The PXS*P antibody reacted weakly with peptide libraries where proline was fixed at the −1 position and reacted strongly with libraries where proline was fixed at both the −2 and +1 positions. The reactivity as defined by this binding profile indicates that the PXS*P antibodies strongly bind only sequences containing the PXS*P motif, as expected, but that the antisera still contain some residual reactivity to S*P (as a result of impurities), which could be removed by further purification using immobilized S*P peptide library. [0153]

EXAMPLE III

Protein Kinase Consensus Site-Specific Phosphoantibodies: 14-3-3 Binding Site: RSXS*XP

Antibodies that identify 14-3-3 targets were obtained by synthesizing a peptide library: XXXXRSXS*XPXXXXC where S* is phosphoserine and X represents any amino acid and C is cysteine. The above 14-3-3 phosphopeptide library was synthesized by standard F-Moc solid phase peptide synthesis using an ABI peptide synthesizer and mixtures of each amino acid except cysteine during degenerate coupling reactions, as discussed in Example I. [0154]
The 14-3-3 phosphopeptide library was coupled to KLH and injected into rabbits as described above for phosphothreonine and PXS*P. Antisera from the most promising rabbit was purified over protein A and adsorbed over a nonphospho-14-3-3 peptide library column. The flow-through of this column was applied to a phospho-14-3-3 column eluted at low pH, dialyzed and tested for phosphospecificity by ELISA using phospho-and nonphospho-14-3-3 peptide libraries. These affinity purified phospho-14-3-3 antibodies recognized the phosphorylated 14-3-3 peptide library but not the nonphospho-14-3-3 library, indicating a high degree of specificity for phospho-14-3-3 (see FIG. 3[0155] a). The antibodies also reacted strongly with several different peptides containing the 14-3-3 motif including; phospho-Bad-Ser136, cdc25-Ser216, and more weakly with phospho-Bad-Ser112 which contains a slight variant motif. The antibodies showed no reactivity with the corresponding nonphospho-peptides (FIG. 3a) or with many other phosphopeptides that did not contain the motif.
Phospho-14-3-3 antibodies were further tested by western blotting using cell extracts prepared from cells transfected with a GST-Bad fusion protein and treated with and without the phorbol ester TPA. The antibodies reacted with a small number of proteins from control extracts (see FIG. 3[0156] b). Bad was detected in extracts prepared from transfected cells but not control cells. Since the basal level of Bad phosphorylation is high it was difficult to see increased phosphorylation with TPA, although TPA did induce the phosphorylation of several higher molecular weight proteins (arrow in FIG. 3b). These results indicate that the phospho-14-3-3 antibodies can detect phosphorylated Bad and other TPA stimulated phospho-proteins.
ELISA analysis against the previously described grid of serine/threonine phosphorylated peptide libraries was also conducted. As expected, the phospho-14-3-3 antibodies have an absolute requirement for proline at the +2 position. [0157]

EXAMPLE IV

Production of Mouse Monoclonal Antibodies: CDK Consensus Phosphorylation Site PXT*PXR

The PXT*/S*PXR sequence represents a consensus phosphorylation site for many of the cell cycle-dependent proteins kinases (cdks). Antibodies that recognize this phosphorylated motif would be useful to identify new cdk substrates important in controlling cell cycle progression. The PXT*/S*PXR peptide library shown in FIG. 4[0158] a was coupled to KLH and injected into Balb/c mice. The phosphopeptide-KLH conjugate (50 μgrams) emulsified in Freund's adjuvant was injected IP. Booster injections (12.5 to 25 μgrams) in incomplete Freund's adjuvant were carried out every three weeks until a response was obtained. Antibody titre was measured by ELISA against the immunized phosphopeptide library. Sera from mice showing high-titre responses were adsorbed with immobilized nonphospho Thr/Ser peptide and the nonadsorbed fraction tested by western blotting (data not shown).
Splenocytes from a mouse showing phosphospecific responses were fused to myeloma X63Ag8.635 cells (Kearney et al., [0159] J. Immunol. 123:1548-1550 (1979)) and approximately 1,100 hybridoma clones were selected and screened. Supernatants from individual clones were screened first for their ability to bind the immunized phosphopeptide library and next for their cross-reactivity against the non-phosphopeptide library. Two different clones showing the highest degree of phospho-specificity were chosen for further analysis. The specificity of clones 6B8 and 5A9 were further characterized using the phosphopeptide libraries and phosphopeptides shown in FIG. 4a. Both clones reacted specifically with phosphothreonine containing libraries and individual peptides but did not significantly react with phosphoserine containing peptides, indicating that phosphothreonine selective clones had been identified. Both clones reacted strongly with peptide libraries where proline is fixed in the −2 and +1 positions relative to phosphothreonine. Reactivity against T*P and PXT*P libraries does not indicate relaxed specificity since one of 400 and one of 20 peptides in the respective libraries will have the appropriate amino acids at the fixed positions. Both clones reacted strongly with a single RB phosphothreonine peptide containing each of the fixed positions present in the immunized library but did not react significantly with the corresponding nonphosphopeptide.
Western analysis shows that okadaic acid treatment of cultured cells dramatically increases the reactivity with both clones 6B8 and 5A9 (FIG. 4[0160] b). Clone 6B8 is also shown to detect cdc2 phosphorylated RB by western blotting (FIG. 4b) but does not react with nonphosphorylated RB protein. Clone 5A9 was deposited in accordance with the terms and conditions of the Budapest Treaty on Sep. 4, 1998 with the American Type Culture Collection under ATCC Accession No. HB12563.

EXAMPLE V

Acetylated Lysine Specific Antibodies

Antibodies specifically reactive against acetylated lysine but not reactive against non-acetylated lysine were obtained by synthesizing the following acetylated lysine peptide library: XXXXXXK*XXXXXXC where K* is acetylated and X represents any amino acid except cysteine and C is cysteine. The acetylated lysine peptide library was synthesized as described previously by standard F-Moc solid phase peptide synthesis using commercially available fully protected acetylated lysine. [0161]
The peptide library was coupled to KLH and injected into rabbits. The K*-peptide-KLH conjugate (250 μgrams) was used as immunogen as described for the other phosphopeptide libraries. Antisera from the most promising rabbit were purified over protein A and adsorbed over a non-acetylated lysine peptide library column. The flow through of this column was applied to an acetylated lysine column, eluted at low pH, dialyzed and tested for phosphospecificity by ELISA. [0162]
Acetylated-lysine antibodies, affinity purified as described above, recognized the acetylated lysine peptide library but not the non-acetylated library, indicating a high degree of specificity for acetylated lysine as measured by ELISA. The antibodies also reacted specifically with as little as 0.5 ng of acetylated bovine serum albumin (BSA) but showed no reactivity with up to 10 μgrams of nonacetylated BSA (see FIG. 5[0163] a).
The antibodies were further examined by western blotting using cell extracts prepared from cells treated with and without anisomycin. The antibodies react with a number of different proteins present in the C6-cell extracts (FIG. 5[0164] b). In panels b and c, antibodies were preincubated with 1 μgram of nonacetylated peptide library (FIG. 5b) or 1 μgram of acetylated peptide library (FIG. 5c). Preincubation with nonacetylated peptide library had little effect on antibody reactivity with acetylated control protein or bands visualized in the cell extract (FIG. 5c, lanes 5-8). However, preincubation of the antibodies with the acetylated lysine peptide library completely blocked antibody binding to control acetylated BSA as well as binding to many proteins present in the cell extract (FIG. 5d, lanes 9-12). These results demonstrate a high degree of specificity for acetylated lysine and indicate that the antibodies recognize a broad spectrum of different sized proteins that contain acetylated lysine in a variety of surrounding sequence contexts (compare FIG. 5c and d, lanes 1, 2).

EXAMPLE VI

Phosphoantibody to the Substrate Consensus Sequence for Akt: RXRXXT*

A. Production of Akt Antibody and Initial Characterization [0165]
The Akt protein kinase is an important regulator of cell survival and insulin signaling, but very few of its in vivo targets have been identified. Studies with synthetic peptide substrates of Akt (D. R. Alessi et al. FEBS Lett. 399:333-338 (1996)) as well as the analysis of known Akt phosphorylation sites on GSK-3 (T. F. Franke et al. Cell 88:435-437 (1997)), Bad (M. Pap et al. J. Biol. Chem. 273:19929-19932 (1998); Datta et al. Cell 91:231-241 (1997)), FKHR Brunet et al. Cell 96:857-868 (1999)), and Caspase-9 (M. H. Cardone et al. Science 282:1318-1321 (1998)) indicate that Akt phosphorylates its substrates only at a serine or threonine in a conserved motif characterized by arginine at positions −5 and −3. [0166]
To study and discover new Akt targets, an antibody was developed that specifically recognizes the phosphorylated form of the Akt substrate consensus sequence RXRXXT*. This antibody was raised against the following synthetic peptide antigen, where X represents a position in the peptide synthesis where a mixture of all twenty amino acids were used, and Thr* represents phospho-threonine: Cys-X-X-X-Arg-X-Arg-X-X-Thr*-X-X-X-X (SEQ ID NO: 45). The synthetic phospho-peptide was conjugated KLH (keyhole limpet hemocyanin) and injected into rabbits. Test bleeds were collected and characterized by ELISA on phospho and non-phospho versions of the peptide antigen. [0167]
Once rabbits started to show high phospho-specific titers, 40 ml production bleeds were obtained. Bleeds were dialyzed overnight in 0.025M NaAcetate, 0.01M NaCl pH=5.2 at 4° C., then spun at 11,200 rpm at 4° C. for 30 min to precipitate serum lipids. Serum supernatant was then purified by Protein A chromatography on a Pharmacia ÄKTA FPLC to isolate the IgG antibody fraction. Affinity chromatography is then performed using peptide coupled to SulfoLink resin from Pierce (#20401; coupling directions according to manufacturer). Phospho-Akt Substrate Antibody was found to be already highly phospho-specific as crude serum, so that a subtraction step on a column containing the non-phospho peptide was not necessary and the elution from the Protein A column could be used directly for affinity chromatography on a phospho-peptide-containing column. Protein A eluate was incubated with phospho-peptide resin by rotation in a sealed column at room temperature for one hour. Column was then drained, washed twice with PBS, and eluted with 0.1M Glycine, pH 2.7 and pooled fractions neutralized with 1M Tris-HCl, pH 9.5 (˜1-2% of fraction volume). The eluted phospho-specific antibody was then dialyzed overnight in PBS at 4° C. [0168]
The resulting antibody is highly specific for peptides that contain phospho-threonine/serine preceded by arginine at positions −5 and −3 (FIG. 6). Some cross-reactivity is observed for peptides that contain arginine at positions −3 and −2. (FIG. 6) also shows that this antibody is highly phospho-specific and recognizes these motifs only when phosphorylated (signal to noise ratios were determined as a ratio of reactivity with the phospho-peptide to reactivity with the corresponding non-phospho-peptide). This antibody does not recognize other phospho-threonine/serine containing motifs. (FIG. 7) indicates that in mammalian cells there are many phosphoproteins recognized by this antibody. [0169]
B. Further Characterization of Binding Specificity [0170]
In order to precisely determining the binding specificity of the above-described Akt motif-specific antibody prior to its use to positively identify cellular phosphoproteins (described in Example X, below), the antibody was further characterized using oriented peptide library analysis and ELISA. [0171]

A first test peptide library having the general sequence MAXXXXXT*XXXXAKKK (SEQ ID NO: 88), where X stands for mixture of 18 amino acid (omitted W and C) and T*=phosphothreonine, was synthesized (substantially as described in Example 1). Sequence analysis of this peptide library indicated that it contains all 18 amino acids at each degenerate position, with less than three folds variation for each amino acid (Table 1, below).

TABLE 1


Pico mole amount of each amino acid at different position of
P-Thr peptide library

M

A

X

T*

X

A

K

D	4.39	3.27	9.92	11.5	11.8	11.2	8.98	4.81	7.89	7.92	8.08	7.83	2.91	2.94	2.75
N	1.52	1.78	9.53	11.2	11.3	11.1	9.57	3.42	6.32	6.84	6.83	7.84	2.79	1.98	1.64
S	6.84	2.72	7.28	6.8	6.49	6.35	5.87	2.59	4.83	4.56	3.81	6.12	1.85	1.34	1.23
Q	0.81	1.79	9.06	9.28	9.35	8.71	9.31	4.25	7.7	8.2	7.24	8.01	3.84	2.76	2.38
T	3	1.58	6.35	5.44	4.97	5.06	4.43	3.8	4.28	3.51	2.83	4.03	1.57	1.18	1.21
G	9.34	8.8	14.7	16	16.9	16.2	12.6	10.6	13.3	13.9	12.7	12.4	10.7	9.96	9.79
E	2.18	2.33	11.9	12.5	12.8	11.4	11.1	4.81	9.16	9.34	8.31	6.66	4.02	2.77	2.3
H	8.18	3.47	6.55	8.4	6.99	6.76	7.46	2.9	4.78	7.45	5.41	5.64	2.72	1.66	1
A	6.56	201	27.9	18.6	19	16.7	15.8	7.26	13.7	14	12.2	11.6	29.5	14.4	7.38
R	3.56	2.17	6.8	9.81	8.64	9.51	11.3	5.42	6.9	9.69	7.71	7.75	4.75	3.77	3.02
Y	2.61	2.51	12.1	12.7	12.6	11.3	9.16	4.3	8.47	8.72	7.47	5.47	3.28	2.24	1.75
P	3.92	2.64	16.3	15.8	15.7	13.5	14.5	6.45	8.91	8.23	6.1	4.54	2.77	1.93	1.65
M	187	6.55	12	12.2	11.9	10.8	10.1	3.24	7.83	7.45	6.03	3.9	1.76	0.82	0.66
V	4.12	2.39	8.38	8.13	7.32	7.34	7.48	3.44	5.41	5.23	4.38	3.43	2.48	2.01	1.9
F	1.55	1.99	10.8	11.5	11.1	10.4	8.83	3.38	7.68	7.5	6.25	4.5	2.48	1.97	1.6
I	3.04	3.2	9.38	8.8	7.95	8.11	8.22	3.91	6.25	5.94	5.33	5.57	2.57	3.43	3.16
K	1.05	0.82	8.61	8.66	8.37	7.77	8	3.1	6.09	6.43	5.09	3.98	2.45	9.63	7.41
L	2.92	3.1	14.1	14.3	13.8	12.4	12.5	5.94	10.3	10	8.46	6.31	3.8	3.86	3.51
Total pmoles	252	252	202	202	197	185	175	83.6	140	145	124	116	86.2	68.7	54.3

This phosphothreonine-containing test peptide library was then tested against the phospho-(Ser/Thr) Akt consensus substrate motif antibody in order to determine the precise binding specificity of this antibody. 1 mg of Phospho-(Ser/Thr) Akt Substrate Antibody was mixed with 200 [0173] μl 50% slurry of pre-swelled protein A beads, and the mixture was incubated at 4° C. overnight with gentle agitation. The beads were then washed three times with 1 ml of PBS with 0.5% of NP-40, followed by twice of 1 ml of PBS only. The beads were transferred to a microspin column (BioRad) and washed three times with 1 ml of PBS. 1 mg of peptide library (30 mg/ml) was loaded to the column. The column was incubated at room temperature for 10 minutes and 4° C. for 1.5 hours. The column was rapidly washed twice with 1 ml of ice-cold PBS +0.5% NP-40, and twice with 1 ml of ice-cold PBS only.
The bound peptides were eluted with 30% acetic acid at room temperature for 10 minutes. The peptides were separated from antibody by passing the elution through a centrocon (10 kDa cut off) twice with 0.4 ml of 30% acetic acid. The solution was evaporated to dryness on a SpeedVac apparatus. The pellet was resuspended in 80 μl of water, and 40 μl aliquots were used for sequencing. [0174]
Control experiments were run simultaneously by performing library screening on protein A beads alone, and protein A beads immobilized with another irrelevant antibody (phospho-tyrosine specific antibody), which does not bind the peptides in the library. 1 μl of elution from the Akt motif antibody column, and control phospho-Tyr antibody and protein A columns, as well as the flow through from each column, was analyzed by MALDI-TOF mass spectrometry according to standard methods. As expected, only the elution from the phospho-(Ser/Thr) Akt consensus substrate motif antibody column contains peptides, indicating the specific binding of the peptides to the motif-specific antibody. [0175]
The recovered (eluted) peptides were analyzed by automated amino acid sequencing (Applied Biosystems) by standard methods. The abundance of each amino acid at a given cycle in the sequence of the bound peptide mixture was divided by the abundance of the same amino acid in the same cycle of the starting peptide library. These raw preference values were then summed and normalized to the total number of amino acids in the degenerate position (18, as Cys and Trp are omitted from the library). FIG. 14 depicts the graphic plots of relative preference values versus amino acid for each sequencing cycle. From these data, we found the strongest binding selection of this antibody is for peptides with arginine at the −3 position relative to the required phospho-threonine (FIG. 13). [0176]
To further determine the binding requirement of this antibody at other positions besides the requirement for R at −3 position, a second test peptide library was synthesized with fixed R at −3 position, having the general sequence MAXXXRXXT*XGGGAKK (SEQ ID NO: 89), where T*=phospho-threonine. The same procedure described above was repeated with this secondary test library, and the relative preference of each amino acid at different positions is shown in FIG. 14. This data indicates that the phospho-Akt motif-specific, context-independent antibody further selects for arginine at the −6, −5, and −4 positions relative to the phosphothreonine. [0177]
The determinations of the precise amino acid requirements for the binding of this motif-specific, context-independent antibody were confirmed by ELISA analysis using the antibody and synthetic peptides containing phospho- and non-phospho threonine or serine Akt motifs. ELISA reactivity relative to phospho-GSK3 peptide are shown in FIG. 21. Akt consensus substrate motif antibody bound only to phospho-peptides. For phosphothreonine-containing peptides, arginine at position −3 is required, although lysine can substitute for arginine with weaker binding. For phosphoserine-containing peptides, arginine appears to be required at position −3 and at position −5 or −2. Good antibody binding was associated with hydrophobic amino acids at position +1, small non-charged residues at position −1, and either small residues or arginine/lysine at position −2. The relative preference value and ELISA analysis data can be used to generate sequence profile motif matrices in which favorable or unfavorable contributions from individual amino acids are weighed quantitatively to predict the likelihood of a sequence recognized by this antibody. This information is useful for subsequent database searching of higher stringency using the motif matrices determined for this antibody, rather than simply searching using the Akt consensus substrate motif sequence itself. [0178]

EXAMPLE VII

Phosphoantibody to the Substrate Consensus Sequence for PKA: RRXT*

cAMP-dependent Protein Kinase A (PKA) is an important kinase for regulating a striking number of physiologic processes, including intermediary metabolism, cellular proliferation and neuronal signaling, by altering basic patterns of gene expression (M. Montminy, Annual Rev. Biochem. 66:807-822 (1997)). Studies with synthetic peptide substrates have established a consensus phosphorylation site for PKA, namely serine or threonine with arginine at the −2 and −3 positions (Z. Songyang et al. Current Biology 4:973-982 (1994)). [0179]
To identify and study new in vivo substrates of PKA, an antibody was developed that specifically recognizes the phosphorylated form of the PKA substrate consensus sequence RRXT*. The following synthetic phospho-peptide peptide antigen was used, where X represents a position in the peptide synthesis where a mixture of all twenty amino acids were used, and Thr* represents phospho-threonine: Cys-X-X-X-X-X-Arg-Arg-X-Thr*X-X-X-X (SEQ ID NO: 46). The synthetic phospho-peptide was conjugated KLH (keyhole limpet hemocyanin) and injected into rabbits. Test bleeds were collected and characterized by ELISA on phospho and non-phospho versions of the peptide antigen. [0180]
Once rabbits started to show high phospho-specific titers, 40 ml production bleeds were obtained. Bleeds were dialyzed overnight in 0.025M NaAcetate, 0.01M NaCl pH=5.2 at 4° C., then spun at 11,200 rpm at 4° C. for 30 min to precipitate serum lipids. Serum supernatant was then purified by Protein A chromatography on a Pharmacia (Piscataway, N.J.) ÄKTA FPLC to isolate the IgG antibody fraction. Affinity chromatography was then performed using peptide coupled to SulfoLink resin from Pierce (#20401; coupling directions according to manufacturer). Both phospho-peptide-containing resin and the corresponding non-phospho-peptide resin were prepared. Protein A eluate was first incubated with non-phospho-peptide resin by rotation in a sealed column at room temperature for one hour, in order to remove antibodies reactive with the non-phospho version of the protein antigen. This resin was then drained and the flow-through then incubated with phospho-peptide resin. This column was drained, washed twice with PBS, phospho-specific antibody eluted with 0.1M Glycine, pH 2.7 and pooled fractions neutralized with 1M Tris-HCl, pH 9.5 (˜1-2% of fraction volume). The eluted phospho-specific antibody was then dialyzed overnight in PBS at 4° C. [0181]
FIG. 8 shows that the resulting antibody is highly specific for peptides or proteins containing phospho-threonine with arginine at the −3 position. The antibody also recognizes some proteins containing phospho-serine with arginine at the −2 and −3 position. It does not recognize the non-phosphorylated version of these motifs (as shown by the signal to noise ratios in FIG. 8 which were determined as a ratio of reactivity with the phospho-peptide to reactivity with the corresponding non-phospho-peptide); nor does the antibody recognize other phospho-serine/threonine containing motifs. FIG. 9 indicates that in mammalian cells there are many phosphoproteins recognized by this antibody, while FIG. 10 shows that this antibody specifically detects many PKA protein substrates in a cell but will not recognize substrates of the ERK2 or CDC2 kinases, which have different substrate specificities. [0182]

EXAMPLE VIII

Phosphoantibody to the Substrate Consensus Sequence for Bulky Ring-Directed Kinases: [F/Y][T/S]* or [S/T]*F

Some important classes of protein kinases are regulated by phosphorylation of a specific serine or threonine flanked by either phenylalanine or tyrosine. For example, Akt, which plays a central role in regulating cell survival, is activated by phosphorylation at Ser473, a site flanked by phenylalanine and tyrosine (D. R. Alessi et al. EMBO J. 15:6541-6551 (1996)). RSK1 (Ser381) and the PKC's also contain this consensus site, phosphorylation of which is required for their activity (K. N. Dalby et al. J. Biol. Chem. 273:1496-1505 (1998); L. M. Keranen et al. Curr. Biol. 5:1395-1403 (1995)). [0183]
To help study signaling pathways regulated by phosphorylation at these key regulatory sites we developed an antibody that detects phospho-serine and phospho-threonine only when preceded by tyrosine, tryptophan or phenylalanine or when followed by phenylalanine. This antibody was raised against the following synthetic peptide antigen, where X represents a position in the peptide synthesis where a mixture of all twenty amino acids were used, and Ser* or Thr* represents phospho-serine or phospho-threonine: X-X-X-X-F-X-X-F-[S*/T*]-[FN]-X-X-X-X-C (SEQ ID NO: 47). This synthetic phospho-peptide was conjugated to KLH and injected into rabbits. Test bleeds were collected and characterized by ELISA on phospho and non-phospho versions of the peptide antigen. [0184]
Once rabbits stared to show high phospho-specific titers, 40 ml production bleeds were obtained. Bleeds were dialyzed overnight in 0.025M NaAcetate, 0.01M NaCl pH=5.2 at 4° C., then spun at 11,200 rpm at 4° C. for 30 min to precipitate serum lipids. Serum supernatant was then purified by Protein A chromatography on a Pharmacia (Piscataway, N.J.) ÄKTA FPLC to isolate the IgG antibody fraction. Affinity chromatography was then performed using peptide coupled to SulfoLink resin from Pierce (#20401; coupling directions according to manufacturer). Both phospho-peptide-containing resin and the corresponding non-phospho-peptide resin were prepared. Two rounds of subtractive purification were performed using the non-phospho-peptide resin: Protein A eluate was incubated with non-phospho-peptide resin by rotation in a sealed column at room temperature for one hour, in order to remove antibodies reactive with the non-phospho version of the protein antigen. The column was drained and the flow-through (containing the desired antibody) incubated with fresh non-phospho-peptide resin. The flow-through from this second subtractive step was finally positively purified by incubation with phospho-peptide resin. After the phospho-peptide column was drained and washed twice with PBS, phospho-specific antibody (bound to the resin) was eluted with 0.1M Glycine, pH 2.7 and pooled fractions were neutralized with 1M Tris-HCl, pH 9.5 (˜1-2% of fraction volume). The eluted phospho-specific antibody was then dialyzed overnight in PBS at 4° C. [0185]
The resulting antibody is highly specific for phosphorylated [F/Y][T/S]- or [S/T]F-containing peptides (FIG. 11). It does not recognize non-phosphorylated [F/Y][T/S] or [S/T]F motifs or other phospho-serine/threonine containing proteins and peptides (signal to noise ratios were determined as a ratio of reactivity with the phospho-peptide to reactivity with the corresponding non-phospho-peptide). This antibody does not recognize other phospho-threonine/serine containing motifs. FIG. 12 indicates that in mammalian cells there are many phosphoproteins recognized by this antibody. [0186]

EXAMPLE IX

Binding Specificity Characterization: Phospho-PKC Consensus Substrate Motif Antibody

The binding specificity/preference of a motif-specific, context-independent antibody that recognizes phospho-(Ser)-PKC consensus substrate motif was precisely determined, as described above, preliminary to the use of this antibody for the positive identification of cellular phosphoproteins (as described in Example X). The PKC motif antibody was produced substantially as described in Examples I-VI above (data not shown here; production described in U.S. Ser. No. 10/014,485, Comb et al., co-pending). [0187]
ELISA analysis of the synthetic peptide library described in Example VI(B) was also used to determine the precise amino acid binding requirement for the phospho-(Ser) PKC consensus substrate motif antibody (see FIG. 22). ELISA readings for each peptide are presented in FIG. 22 as a percentage relative to phospho-AFX peptide. The data indicate that this antibody binds only phospho-serine containing peptides where phospho-serine is followed by arginine or lysine at position +2. In addition, the antibody appears selective for hydrophobic amino acids at position +1. [0188]
The relative preference value and ELISA analysis data can be used to generate sequence profile motif matrices in which favorable or unfavorable contributions from individual amino acids are weighed quantitatively to predict the likelihood of a sequence recognized by this antibody. This information is useful for subsequent database searching of higher stringency using the motif matrices determined for this antibody, rather than simply searching using the PKC consensus substrate motif sequence itself. [0189]

EXAMPLE X

Positive Identification of Phosphoproteins in Growth Factor-Treated Cells Using Akt and PKC Consensus Motif Antibodies

Motif-specific, context-independent antibodies to phospho (Ser/Thr)-Akt consensus substrate motif and phospho(Ser)-PKC consensus substrate motif, respectively, were employed in the method of the invention to positively identify cellular phosphoproteins containing the respective motifs. Cellular phosphoproteins were initially fractionated by mass and the presence of phosphoproteins containing target motifs detecting using these motif antibodies, followed by database searching using the search program/program, Scansite. [0190]
A. Preliminary Characterization by Motif Sequence and Mass [0191]
PDK1 is the kinase that phosphorylates Akt, PKCs, as well as many other members of the AGC family of kinases. The phosphorylation by PDK1 occurs at the activation loops of the substrate kinases (for Akt, the site is Thr308), and this phosphorylation is required for kinase activity. The AGC family of kinases can phosphorylate serine or threonine in a short motif containing basic amino acids surrounding the serine or threonine. Several members of AGC family of kinases, such as Akt, p70S6 kinase, and RSK can phosphorylate substrates with a similar consensus substrate motif, RXRXXT*/S*. [0192]
The known binding preferences of the disclosed phospho-(Ser/The) Akt consensus substrate motif antibody (as determined in Example VI(B)) and phospho-(Ser) PKC consensus substrate motif (as determined in Example IX) make these motif-specific, context-independent antibodies highly useful reagents for the identification of the phosphoproteins that have been phosphorylated by these kinases, e.g. upon growth factor stimulation. As detailed below, the phosphorylation only occurs in wild type ES cells and not PDK1 null ES cells; these phosphoproteins are presumably the substrates of the AGC family of kinases, such as Akt or PKC. [0193]
10 μg of cell lysate from IGF1-treated PDK1 null or wild type ES cells was run on 12% SDS gel and probed by antibodies against total Akt, phospho-Akt at Ser473 and Thr308, respectively, and the phospho-(Ser/Thr) Akt consensus substrate and phospho-(Ser) PKC consensus substrate motif antibodies described herein. As shown in FIG. 15, the phosphoproteins that are induced by growth factor only in wild type ES cells can be specifically identified, initially, with these antibodies. [0194]
As expected, phosphorylation of Akt at Thr308 and p70S6K at Thr389 was not detected in PDK1 −/− cells. Phosphorylation of Thr308 and Thr389 are required for its activities in PDK1 null cells. The loss of phosphorylation at Akt Thr308 and p70S6K The 389 is correlated with the loss of signal in PDK1 null ES cells when probed with the phospho-(Ser/Thr) Akt consensus substrate motif antibody. There is a strong band located between [0195] molecular marker 28 kDa and 32 kDa in wild type ES cells only. This protein is also specifically recognized by the phospho-(Ser) PKC consensus substrate motif antibody, as shown in FIGS. 16 and 17. Accordingly, this screen initially identifies certain phospho-proteins that have a known mass and are now known to contain sequence matching the precise target sequences (within the motif) bound by each of these motif-specific, context-independent antibodies. These parameters were then used as inputs in the Scansite search program to search publicly-accessible protein databases for described proteins that match the mass and motif sequence parameters, as detailed below.
B. Identification of Phosphoproteins by Scanning Protein Database with Determined Motif and Mass Parameters [0196]
The Swiss-Prot database was searched with both Scansite (http://scansite.mit.edu/) and Prosite (http://us.expasy.org/prosite/) programs using the sequence and molecular weight parameters identified in part (A) above. When searching the database with matrices (mass and determined precise target sequence bound by the motif-antibodies) for either the phospho-(Ser/Thr) Akt consensus substrate antibody or phospho-(Ser) PKC consensus substrate antibody alone, both ScanSite and Prosite programs give many hits (i.e. putative protein matches). However, the stringency of the search is improved and only a few putative matching proteins stand out when matrices for both motif-specific, context-independent antibodies were used in the database searching. FIG. 18 shows the list of potential matching proteins resulting from searching the Swiss-Prot database with the Scansite program using both Akt and PKC substrate motif antibody matrices within the molecular weight range between 28-32 kDa. [0197]
S6 ribosomal protein has been reported to undergo phosphorylation after growth factor treatment. The same Western blot previously probed with phospho-(Ser/Thr) Akt consensus substrate motif antibody was probed with an antibody against S6 ribosomal protein. As shown in FIG. 19, antibody against S6 ribosomal protein detected a band at the same size as detected by the phospho-Akt motif antibody in both wild type and PDK1 null ES cells. Antibody against the phospho-S6 ribosomal protein S236 site detected the same band as the phospho-Akt motif antibody and the phospho-(Ser) PKC motif Antibody (FIG. 19). [0198]
The identification of the ˜30 kDa band as S6 ribosomal protein was further confirmed by immunoprecipitation followed by Western blot. 100 μg of total cell lysates from PDK1 null or wild type ES cells were immunoprecipitated with phospho-(Ser/Thr) Akt consensus substrate motif antibody. The original cell lysate, supernatant from immunoprecipitation, and the elution from immunoprecipitation were run on 12% SDS gel followed by Western blotting with S6 ribosomal protein. As shown in FIG. 20, phospho-(Ser/Thr) Akt motif antibody can specifically immunoprecipitate a 30 kDa protein from wild type ES cells that can be identified by S6 ribosomal specific antibody. [0199]
It was therefore concluded that the ˜30 kDa band detected by both phospho-(Ser/Thr) Akt motif antibody and phospho-(Ser) PKC motif antibody was positively identified as S6 ribosomal protein. [0200]

EXAMPLE XI

Identification of Effects of a Kinase Inhibitor on Phosphoprotein Levels in Cell Extracts

The effects of test compounds (e.g. a drug) on cell signal transduction and particular phosphoprotein activation may now be examined using the methods of the present invention. Phosphoproteins involved in signaling mediated by a particular protein target, e.g. a kinase, may be monitored before and after treatment with an inhibitor of the protein target using motif-specific, context-independent antibodies that recognize substrates phosphorylated by the targeted protein. [0201]
For example, the effects of MEK1 inhibitors may examined using MAPK motif-specific antibodies to positively identify phosphoprotein substrate levels following treatment with the inhibitor. COS cells, for example, may be transfected with pCMV-MKK3 for 22 hours, pretreated with different amounts of a MEK1 Inhibitor, such as PD98059, for 1 hour, and then treated with anisomycin for 30 minutes. Cell extracts are blotted, e.g. with Phospho-MKK3/MKK6 (Ser189/207) and a motif-specific antibody to MAPK-consensus recognition motif (PXS*P) (see Example II), to identify phosphoproteins of known mass which contain the motif bound by this antibody. [0202]
The unique proteins detected only in cell extract not treated with PD98059 are immunoprecipitated by MAPK consensus substrate motif antibody, followed by western blot by several other motif-specific context-independent antibodies (such as described in Examples III, VI, VIII, IX) to determine if the proteins recognized by MAPK consensus motif antibody also contain other modified motifs recognized by the additional, different motif antibodies. [0203]
The determined mass and sequence/motif parameters are then used (using Scansite or Prosite) to search protein databases for sequence and mass matches with described proteins. The positive identification of the closest matches (hits) are further confirmed by specific antibodies (as described in Example X). Accordingly, changes in particular phosphoproteins resulting from PD98059 treatment can be examined. [0204]

Claims

What is claimed is:

1. A method for the positive identification of cellular phosphoproteins using motif-specific, context-independent antibodies coupled with database searching said method comprising the steps of:

(a) fractionating a preparation comprising a plurality of phosphoproteins on the basis of mass and/or isoelectric point (PI);

(b) contacting the fractionated preparation of step (a) with at least one motif-specific, context-independent antibody that binds a phosphorylated motif of known sequence, thereby to detect phosphoproteins comprising said motif; and

(c) utilizing a search program to search one or more protein database(s) for reported proteins having a sequence comprising said motif and having a mass and/or PI substantially matching that determined in step (a), thereby to positively identify at least one cellular phosphoprotein in said preparation as a reported protein of known sequence and mass and/or PI.

2. The method of claim 1, wherein the fractionation of step (a) comprises gel electrophoresis.

3. The method of claim 1, wherein the fractionation of step (a) comprises 2-dimensional gel electrophoresis separation on the basis of both mass and isoelectric point (PI).

4. The method of claim 1, wherein said search program of step (c) comprises Scansite and/or Prosite search software.

5. The method of claim 1, wherein said preparation comprises a crude cell extract or a partially purified cell extract.

6. The method of claim 5, wherein said preparation is obtained from a diseased cell or tissue or from a cell or tissue susceptible to disease.

7. The method of claim 6, wherein said disease is cancer.

8. The method of claim 1, wherein said preparation is obtained from a cell, tissue or organism treated with a test compound.

9. The method of claim 8, wherein said test compound is a kinase inhibitor.

10. The method of claim 1, further comprising the step of (d) verifying the identity of said phosphoprotein as determined in step (c) by contacting said phosphoprotein with at least one antibody specific for said protein in phosphorylated or unphosphorylated form, thereby verifying the protein's identity.

11. The method of claim 1, wherein a plurality of motif-specific, context-independent antibodies, each recognizing a different phosphorylated motif, is contacted with said fractionated preparation in step (b), and wherein the search of step (c) is for reported proteins comprising two or more of the different motifs.

12. The method of claim 1, wherein said motif-specific, context-independent antibody binds a motif comprising a kinase consensus substrate motif or a protein-protein binding motif.

13. The method of claim 12, wherein said motif comprises one or more phosphorylated amino acids.

14. The method of claim 12, wherein said kinase consensus substrate motif is selected from the group consisting of a MAPK consensus substrate motif, a CDK consensus substrate motif, a PKA consensus substrate motif, a bulky ring-directed kinase consensus substrate motif, and an Akt consensus substrate motif, and wherein said protein-protein binding motif comprises a 14-3-3 binding motif or a PDK1 docking motif.

15. The method of claim 1, wherein the precise target sequence preferably bound by the antibody employed in step (b) has first been determined, wherein the phosphoproteins detected in step (b) comprise said precise target sequence, and wherein the search of step (c) is for reported proteins having a sequence comprising said precise target sequence.

16. The method of claim 1, wherein the fractionation of step (a) comprises immunoprecipitation.

17. A method for the positive identification of cellular phosphoproteins using motif-specific, context-independent antibodies coupled with database searching, said method comprising the steps of:

(a) fractionating a cellular preparation comprising a plurality of phosphoproteins on the basis of mass and/or PI;

(b) contacting the fractionated preparation of step (a) with a plurality of motif-specific, context-independent antibodies that each bind a different phosphorylated motif of known sequence, thereby to detect at least one phosphoprotein comprising two or more of the different motifs; and

(c) utilizing a search program to search one or more protein database(s) for reported proteins having a sequence comprising the two or more different motifs detected in step (b) and having a mass and/or PI substantially matching that determined in step (a), thereby to positively identify at least one cellular phosphoprotein in said preparation as a reported protein of known sequence and mass and/or PI.

18. The method of claim 17, wherein said preparation comprises a crude cell extract or a partially purified cell extract.

19. The method of claim 17, wherein said motif-specific, context-independent antibody binds a motif comprising a kinase consensus substrate motif or a protein-protein binding motif.

20. The method of claim 17, further comprising the step of (d) verifying the identity of said phosphoprotein as determined in step (c) by contacting said phosphoprotein with at least one antibody specific for said protein in phosphorylated or unphosphorylated form, thereby verifying the protein's identity.

21. The method of claim 17, further comprising the step of (d) verifying the identity of said phosphoprotein as determined in step (c) by immuno-depletion of said phosphoprotein using at least one antibody specific for said protein in phosphorylated or unphosphorylated form.

22. A method for the positive identification of cellular phosphoproteins using motif-specific, context-independent antibodies coupled with database searching, said method comprising the steps of:

(a) fractionating a cellular preparation comprising a plurality of phosphoproteins on the basis of mass and/or isoelectric point (PI);

(b) contacting the fractionated preparation of step (a) with a plurality of motif-specific, context-independent antibodies that each bind a different phosphorylated motif of known sequence, wherein the precise target sequences preferably bound by said antibodies has first been determined, thereby to detect at least one phosphoprotein comprising two or more of said precise target sequences; and

(c) utilizing a search program to search one or more protein database(s) for reported proteins having a sequence comprising the two or more precise target sequences detected in step (b) and having a mass and/or PI substantially matching that determined in step (a), thereby to positively identify at least one cellular phosphoprotein in said preparation as a reported protein of known sequence and mass and/or PI.

23. A method for the positive identification of cellular phosphoproteins using motif-specific, context-independent antibodies coupled with database searching, said method comprising the steps of:

(a) fractionating a cellular preparation comprising a plurality of phosphopeptides by immunoaffinity isolation using at least one motif-specific, context-independent antibody that binds a phosphorylated motif of known sequence, wherein the mass of bound peptides is also determined; and

(b) utilizing a search program to search one or more protein database(s) for reported proteins having a sequence comprising said motif and having a mass substantially matching that determined in step (a), thereby to positively identify at least one cellular phosphoprotein in said preparation as a reported protein of known sequence and mass.