EP1234184A2 - Multiblot kinase analysis - Google Patents

Abstract

Methods, apparatus, media and signals for identifying associated cell signaling proteins are disclosed. The method involves producing and storing a comparison value for each pair of the cell signaling proteins in response to data values representing physical properties of respective cell signaling proteins. The method further involves identifying cell signaling protein pairs having comparison values satisfying a condition indicative of an association between the cell signaling proteins.

Description

MULTIBLOT KINASE ANALYSIS

Background of the Invention

The human genome is believed to contain about 120,000 genes, which are present in each of the 50 trillion nucleated cells of the body. At any given moment in each cell, however, only about 20% of these genes are actively transcribed for the production of the proteins that they encode. The unique complement of proteins that each cell expresses is referred to as its "proteome". It is critical that the levels of expression and activity of the proteins in a cell are tightly regulated. This is achieved through a subset of about 10% of these proteins, which are dedicated to cell communications and control One of the largest classes of proteins involved in cell signalling are enzymes called protein kinases. Protein kinases control other proteins by catalyzing their phosphorylation, which is a process that can be reversed by protein phosphatases. Often protein kinases operate within signalling pathways that are further integrated into networks. The unique complement of protein kinases expressed in a cell is referred to herein as its "kineome".

There are different amino acids that can be phosphorylated by protein kinases. Most commonly, this occurs on serine and threonine, and to a much lesser extent on tyrosine. For example, in skin cells called fibroblasts, over 99% of measurable protein kinases catalyze the serine and threonine phosphorylation of proteins, whereas a different class of related enzymes generally carry out tyrosine phosphorylation. Therefore, protein- serine/threonine kinases are responsible for most protein phosphorylation events in cells. For convenience, these kinases are referred to herein as protein-serine kinases, although they also phosphorylate proteins on threonine.

Approximately fifty of the hundred or so known genes that have been directly linked to induction of cancer (i.e. oncogenes) encode protein kinases. The remainder of the oncogenes specify proteins that either activate kinases or are phosphorylated by kinases. Most of the oncogene-encoded protein kinases are tyrosine-specific, but several are protein-serine kinases such as protein kinase C, Rafl, Akt, ILK-1, Tpl2, and Mos. Although the findings are less direct, aberrant cell signalling through protein kinases has also been linked to cardiovascular disease, diabetes, inflammation, arthritis and other immune disorders, and neurological disorders such as Alzheimer's disease. Over 400 human diseases have been linked to defective signalling through protein kinases.

Essentially all signalling proteins, if they are not already protein kinases, appear to be regulators of protein kinases or their substrates. As signal transduction networks govern and co-ordinate all cellular functions, including cell structure, metabolism, reproduction, adaptation, differentiation and death, knowledge of the structure of signalling networks will permit a complete understanding of how the cell operates under a diversity of conditions. Kineome analysis represents a key step in this process.

Kineome analysis will yield many practical benefits. The presence and state of activity of diverse protein kinases and their pathways are indicators of how a cell perceives its internal and external environments and how it is responding. Therefore, by monitoring the kineome, it will be feasible to obtain a molecular diagnosis of a disease condition. Moreover, by inhibiting or activating the appropriate protein kinases by pharmacological intervention, antisense or gene therapy, it would be possible to "reprogram" the kineome to better treat the disease condition. For example, in cancer, the gain of function of one of over fifty different oncogene-encoded protein kinases may be pivotal for neoplastic transformation of cells. Inhibition of the appropriate kinase or its downstream effectors could block the improper proliferative signalling and initiate apototic processes leading to programmed death of the tumour cells.

The genome sequencing projects for man, mouse and other organisms will permit the rapid identification of protein kinases within the next few years. Elucidation of the connections between these protein kinases in different cells will allow kineome analysis to then reach its full potential. Over two thousand different protein kinases are thought to be encoded by the human genome and several hundred are likely to be expressed within any given cell. All of these protein kinases will have to be tracked for a complete elucidation of the architecture of kinase networks. This is feasible through the employment of kinase specific probes.

Most eukaryotic protein kinases are evolutionarily related, i.e. the genes of almost all protein-serine/threonine and protein-tyrosine kinases display sequence identity. In particular, there are 16 amino acid residues located in 10 subdomains in the catalytic region of protein kinases that are highly conserved. These amino acids allow unambiguous identification of novel protein kinases following analysis of the primary structures of proteins as revealed by the nucleotide sequences of genes. Complete sequencing of the genomes of humans and other species, will facilitate identification of the protein kinases.

From knowledge of the primary structure of a protein, it is feasible to produce nucleotide or antibody probes that are specific for that protein. Antibodies can be generated against the full length-expressed protein or portions. An effective strategy is to identify a region of about 10 to 20 amino acids that are extremely well conserved in that protein in diverse species, but which does not appear in other proteins. Antibodies generated, for example in a rabbit or mouse, against a synthetic peptide based on this amino acid sequence will cross-react with the full-length protein that contains this sequence, with little or no cross-relativity with other proteins. With the knowledge of the primary structures of all the protein kinases and other proteins encoded by mammalian genomes, specific antigen peptides can be designed to elicit the production of antibodies against any protein kinase.

Most protein kinases appear to be activated as a consequence of their own phosphorylation by upstream kinases or by self-phosphorylation (termed autophosphorylation). Phosphorylation cannot be monitored by nucleic acid-based approaches. However, phosphorylation of a protein can produce marked changes in its mobility on electrophoresis gels which are designed to act as a molecular sieve. Of such techniques, sodium dodecyl sulphate (SDS) - polyacrylamide gel electrophoresis (SDS-PAGE) has become the standard method for separation of proteins on the basis of their size for analytical and preparative purposes. This technique relies of the sieving effect of the gel when proteins coated with negatively-charged detergent (e.g. SDS) are drawn through the gel in an electric field. Smaller sized proteins are able to migrate through the gel faster than larger sized proteins. Proteins that differ by as little as a few hundred Daltons can be resolved by this method. Protein staining methods permit the visualization of discreet proteins in the gel as individual bands in a bar code like pattern. When these proteins are transferred from the gel onto a nitrocellulose membrane, the locations of specific proteins can be identified with antibodies in an immunoblotting procedure referred to as Western blotting (see: Towbin; U.S. Patent No. 4,452,901). Most proteomic analytical methods are based on two dimensional (2D) gel electrophoresis by the standard method of Dr. Patrick O'Farrell described nearly 20 years ago. The 2D gel technique initially involves the separation of proteins in a first dimension based on their intrinsic charge in a pH gradient within a isoelectric focusing gel (typically a tube gel). Proteins migrate through the isoelectric focusing gel in the presence of an electric field until they encounter a pH at which the protein no longer possesses an electric charge. This pH is known as the isoelectric point of a protein, and it is a distinguishing characteristic. Following electrophoresis in the first dimension, the isoelectric focusing gel is applied length-wise to the top of a molecular sieve gel such as a SDS-PAGE gel, and electrophoresis is continued into the second dimension. When the 2D gel is stained with sensitive-dyes (e.g. based on silver reagent), the various proteins inside a cell can be visualized as resolved spots. The greater amount of a given protein within a cell sample, the larger and darker its specific spot appears. Several thousand proteins can be distinguished from one and another by this technique. If the protein samples have been obtained from cells that have been incubated with radioactive ³²P-phosphate, then the 2D gel can be exposed to x-ray film, and the phosphoproteins can be specifically detected. The more that a protein is phosphorylated or prevalent, the larger and more intense the spot on the x-ray film. The silver-staining of a 2D gel can be used to track the expression of proteins and their covalent modification by phosphorylation.

Public databases have been created that allow the identification of over a thousand different proteins on 2D gel proteomic maps. However, the positions of very few protein kinases are available. This reflects the fact that protein kinases are present at very minute levels in cells, and are often undetectable by even such sensitive protein dyes as silver- stain. Typically, transduction proteins are expressed at a hundred- to a thousand-fold lower levels than structural proteins and metabolic pathway enzymes. Therefore, it has often been necessary to incorporate selective enrichment techniques such as antibody- based purification as a preliminary step prior to 2D gel electrophoresis. Proteomic analysis of kinases has evolved to date in much the same fashion as techniques for resolution of other proteins. One dimensional immunoblot techniques have been employed in which proteins from a smooth muscle homogenate are separated on an SDS-PAGE gel and then subjected to Western blotting followed by detection using a single polyclonal or monoclonal (MAB) anti-kinase antibody per blot (see: H. Togashi et al. (1997) "Quantitative Immunoblot Analysis of PKC Isoforms Expressed in Airway Smooth Muscle" Am. J. Physiol. 272 (Lung Cell. Mol. Physiol. 16): L603-L607). Conventional wisdom dictates that simultaneous use of multiple antibodies on a single immunoblot necessitates the use of MAB's (see: Coates, S.R. et al.; EP 025384 published January 27, 1988). Furthermore, 2D gel electrophoresis is thought to be the preferable technique for resolving complex protein mixtures. For example, Sanchez, J.C. et al. reported the use of a mixture of nine MAB's for detection of different proteins indicative of oncogene expression (including the kinase MEK-1) on a single immunoblot produced from 2D electrophoresis ("Simultaneous Analysis of Cyclin and Oncogene Expression Using Multiple Monoclonal Antibody Immunoblots" (1997) Electrophoresis 18:638-641).

While it is possible to visualize some protein kinases on 2D gels by immunoblotting techniques, I have determined that in most cases, a maximum of only four or five protein kinases can be detected at a time by Western blotting of 2D gels with mixtures of protein kinase-specific antibodies. Furthermore, I have found that recovery of many protein kinases from a first dimension pH gradient gel, is less than 10%. That means that 90% or more of these protein kinases do not enter the second dimension gel and are therefore unresolved. Moreover, if approximately 10,000 different genes are expressed in a typical mammalian cell, and each of these translated proteins undergo extensive post-translation modification (at more than 10 sites per average protein), then there may be over 100,000 protein species present in a cell lysate. 2D gel electrophoresis can resolve at best 7500 protein species, so there is extensive overlap of protein spots on 2D gels that can confound interpretation of silver-stained 2D gels.

Finally, if 32-P radiolabelling of proteins in intact cells is undertaken prior to 2D gel electrophoresis, the results from such studies after autoradiography of these gels are highly suspect. The depletion of phosphate and the incubation of cells with milliCurie amounts of [32-P]phosphate induces stress responses in cells that can down-regulate the effects of growth factors and mask the effects of stress factors. To view the coordinate regulation of proteins in the proteome or its kineome subset, it is necessary to avoid perturbations of model systems that influence the analyses and employ probes that are high selective for the proteins of interest.

Summary of Invention

This invention provides a method for detection of multiple kinases or multiple kinase substrates, whereby the presence and phosphorylation state of a large number kinases and/or kinase substrate proteins may be tracked in a single sample.

This invention includes electrophoretically separating proteins in a sample to be tested for kinase or kinase substrate content to produce an array of proteins so separated; contacting the array with two or more antibodies selected from anti-kinase and anti-kinase substrate antibodies; and, detecting the presence of antibodies bound to kinase or kinase substrate moieties in the array. This method may also comprise recording one or more values representative of a location for each of the detected antibodies bound to proteins in the array indicative of a location of a kinase or kinase substrate in the array. The method of this invention may also comprise measuring intensity of a signal representative of an amount of antibody detected at a location on the array and recording values representative of said intensity and said location. This invention also provides a record of a value or values representing a location or a location and intensity at the location, produced by the method of this invention. Such a record may be used for comparison to a record so produced from another sample and such a comparison may be employed to produce a comparison value or values relating to a particular kinase or kinase substrate in each example.

The method of this invention may comprise:

(a) obtaining a sample to be tested for kinase or kinase substrate content;

(b) optionally performing one or more of: (i) addition of at least one protein phosphatase to the sample to dephosphorylate proteins in the sample; (ii) inactivating protein phosphatase in the sample; (iii) addition of at least one kinase and ATP to the sample; and (iv) inactivating protein kinase in the sample. (c) performing SDS-PAGE on the sample in a single dimension to produce an array of separated kinase or kinase substrate moieties from the sample;

(d) transferring the array to a membrane;

(e) contacting the membrane with a panel comprising two or more anti-kinase or anti-kinase substrate antibodies; and (f) detecting the presence of antibodies from the panel bound to kinases or kinase substrate moieties on the membrane.

While (c) and (d) above in combination, is similar to standard Western blotting procedure, it is preferable that the electrophoresis gel be constructed to increase the likelihood that proteins will exhibit "band shift" between phosphorylated and dephosphorylated states. Typically, a protein will display reduced migration during SDS- PAGE when the protein is in a phosphorylated state. The reduced mobility may be as much as 1-5 kDa and this separation is enhanced by using a gel with a higher than normal acrylamide content and a lower than normal bisacrylamide content. While content of these gel components is normally adjusted to suit electrophoretic conditions and the average size of proteins to be separated, a 12.5% acrylamide/0.4% bisacrylamide gel is often suitable to achieve separation of phosphorylated and dephosphorylated kinases. When the method of this invention is employed to detect kinases in a sample either to elucidate the kinase profile of a tissue or cell type or to identify novel kinases, it is preferable that the panel of anti-kinase antibodies comprise polyclonal antibodies rather than MAB's. This departure from conventional wisdom increases the likelihood that new kinase proteins will be detected in the sample since polyclonal anti-kinase antibodies generally exhibit greater cross-reactivity to kinases as compared to anti-kinase MAB's. Furthermore, polyclonal antibodies are likely to successfully perform in a wide range of animal species. Despite the use of polyclonal antibodies, the panel may comprise from 2 to about 100 antibodies.

Brief Description of the Drawings

Figure 1, represents SDS-PAGE gels presenting comparison of multi-kinase immunoblotting patterns (arrays) produced by this invention from different rat tissues. Electrophoresis of detergent solubihzed ly sates prepared from rat brain (A), heart (B) and skeletal (C) muscle was performed, and the positions of various protein kinases was visualized by ECL detection. Protein kinases of smaller size migrated correspondingly closer to the bottom of the gels. Each of the 18 strips shown at (A), (B) and (C) were derived from an SDS-PAGE gel and were probed with different panels of protein kinase antibodies.

Figure 2, represents a SDS-PAGE gel demonstrating effects of anti-IgM treatment for 5 min on protein kinases in the human Ramos B cell line. Electrophoresis of detergent solubilized lysates prepared from Ramos cells untreated (-) or exposed (+) to anti-IgM polyclonal antibody for 5 min was performed in alternating lanes, and the positions of various protein kinases was visualized by ECL detection. Each of the 14 paired strips derived from two SDS-PAGE gels were probed with different panels of protein kinase antibodies.

Figure 3. Fig. 3 A represent gels showing differential effects of kinase inhibitors on band shifting of selected protein kinases. Electrophoresis of detergent solubilized lysates prepared from human ovarian surface epithelial cells were untreated (Lane 1) or exposed to 20 ng/ml of human hepatocyte growth factor (HGF) in the absence (Lane 2) or presence of PD98059 (Lane 3), SB203580 (Lane 4), LY294002 (Lane 5) or rapamycin (Lane 6) was performed. The effects of these treatments on the positions of Erkl and Erk2 (Panel A), p38 Hog MAP kinase (Panel B), PKB1 (Panel C), PKB2 (Panel D) and S6 kinase (Panel E) as visualized by ECL detection are shown. The phosphorylated and band shifted forms of these kinases are denoted with a "p" before their name. The partial structures of the protein kinases pathways in which these enzymes operate and the known sites of action of these drugs are shown in Fig. 3B.

Figure 4, represents gels showing detection of known kinases and putative kinases in normal and tumour breast tumour biopsy samples of four human patients. Detergent solubilized lysates prepared from tumour (T) and adjacent control (C) breast tissue were subjected to the method of this invention. In panel (A), the increased levels of p38 MAP kinase, protein kinase B-α(PKBα), casein kinase 2 (CK2), protein kinase G (PKG) and cyclin-dependent kinase 8 (Cdk8) in the tumour samples is evident. Five of 12 proteins that were demonstrated to be elevated in tumours and not yet known for their identity are shown on panel (B).

Figure 5 represents a Western Blot showing separation of Erkl, Erk2 and protein kinase C-β by 2D gel electrophoresis. Detergent solubilized rat brain extract (1 mg protein) was subjected to isoelectric focusing and SDS-PAGE. In the left most lane, 200 μg of the brain extract was directly applied to the same SDS-PAGE gel. Following 2D gel electrophoresis, the proteins were transferred to a nitrocellulose membrane, which was probed with antibodies for Erkl, Erk2 and PKC-β.

Figure 6 represents a Western Blot showing gels presenting a comparison of patterns obtained by use of this invention with whole brain samples prepared from 1 and 50 day old rats. Mini-SDS-PAGE electrophoresis of the detergent solubilized lysates prepared from the brains of 1 day (A) and 50 day (B) old rats was performed, and the proteins were transferred to nitrocellulose membranes. The membranes were probed with a panel of mixtures of kinase antibodies using a 20 lane immunoblotter from Immunetics (Cambridge, MA).

Detailed Description

In this specification, the term kinase refers to those members of the class of enzymes that catalyze a chemical reaction in which a phosphate group is transferred from adenosine triphosphate (ATP) to a recipient protein. Throughout this specification, such a recipient protein is termed a "kinase substrate". This chemical reaction is called protein phosphorylation and is a reversible process with dephosphorylation being catalyzed by enzymes known as protein phosphatases. Kinases may be found in all organisms.

In this specification, "testing for kinase" or "testing for kinase substrate" content means determming the presence of at least one kinase or kinase substrate in a sample. Preferably the phosphorylation state of the kinase or kinase substrate will also be determined.

In this specification, a "kinase or kinase substrate moiety" is a protein having the characteristics of a kinase or kinase substrate which occupies a single position after electrophoresis in a SDS-PAGE gel.

In this specification, reference to electrophoresis of proteins or electrophoretically separating proteins means the use of an electric field applied to a substrate to separate proteins on the substrate. Typically, the substrate is a gel or other collodial substrate containing a liquid phase in which the proteins are dispersed or dissolved. The proteins so separated form an "array" or "pattern" of individual protein moieties distributed along the substrate. Performing electrophoresis in one dimension or providing an "array" in one dimension means that the moieties are distributed along a single axis. Thus, the array has a single axis along which moieties are located as spots or bands transecting the axis.

An "array" or "pattern" as referred to herein may exist on an electrophoresis substrate or may exist on another substrate (such as a Western Blot membrane) to which the array is transferred from an electrophoresis substrate. Such an array may also be indirectly represented as a recorded pattern or series of values representative of the array (e.g. representative of the location and/or quantity of moieties in the array). Means for production of such a record of a pattern or a record of values representative of the array are well known in the art and such records include any graphic, photographic, xerographic or radiographic representation of an array or any signal produced by apparatus adapted to scan such an array and produce an electronic signal, components of such a signal being an analog or digital data structure representative of values correlating with a location on an array and/or the quantity of material detected at a location on an array. Typical apparatus used for such "scanning" are adapted to measure the location of an optically detectable event on the array and to measure the intensity of an optical signal associated with such an event. Examples of apparatus adapted for such recording are described below and in the Examples.

Samples used in the method of this invention may be any cell or tissue homogenate, extract or other such sample which has been processed to purify or partially purify kinases or kinase substrates in the sample. This invention is particularly suitable for testing patient biopsy samples. Such samples may be manipulated to increase prevalence of desired cell types in the sample. The sample will be typically prepared for electrophoresis using standard techniques, employing appropriate buffers which may contain various inhibitors or enzymes. For example, protease inhibitors may be present to reduce protein degradation on the sample. When the sample is to be tested for kinase substrate content, it may be desirable to add one or more protein phosphatases to dephosphorylate substrates which may already exist in a phosphorylated state in the sample. The protein phosphatase will then be inactivated with an appropriate phosphatase inhibitor (e.g. β-glycerophosphate, sodium fluoride or sodium orthovanadate) and a selected protein kinase or mixture of protein kinases is then added to the sample to phosphorylate those substrates present which are specific to kinase added to the sample. Alternatively, endogenous kinases in the sample may be relied upon to phosphorylate dephosphorylated substrates in the sample. The SDS-PAGE employed in this invention is gel electrophoresis performed in a single dimension, typically using a slab shaped gel or a series of tube gels. The gel may be constructed and used employing standard methods, electrophoresis buffers and electrophoresis equipment. The gel may comprise a stacking gel and a separation gel. Commercial kits and equipment are available for performing SDS-PAGE.

Preferably, the makeup of the separation gel will range from 10% to 15% (acrylamide) and 0.2 to 2% (bisacrylamide). For optimum separation of kinases and most kinase substrates, an electric current will typically be applied to the gel until proteins with a molecular mass of less than about 25-27 kDa are eluted from the bottom of the gel. This is because protein kinases and most substrates of interest do not have a molecular mass less than the latter amounts. Optimizing separation of kinases or substrates increases ability to resolve phosphorylated and unphosphorylated states.

Once electrophoresis is complete resulting in a pattern of separated protein moieties in the gel, the pattern is transferred to any membrane (e.g. nitrocellulose, PVDF, nylon, etc.) that is suitable for use in the Western Blotting technique. Transfer is typically done by standard electro-transfer techniques. Once the pattern is transferred to the membrane, the membrane may be cut into strips each of which will typically contain a pattern separated from a single sample (e.g. a test sample or a control sample). Alternatively, a multiblotting apparatus (e.g. as available from Bio-Rad Laboratories or Immunetics, Inc.) may be used whereby the membrane is left intact.

Once the electrophoresis and Western blotting aspects of this invention are complete, the resulting membrane or membrane strips are probed with a panel of different antibodies that react with distinct categories, subsets, isoforms, etc. of protein kinases or kinase substrates. In this specification, such antibodies are termed "anti-kinase antibodies" or "anti-kinase substrate antibodies". The panel may be applied in one step as a mixture of antibodies or, the antibodies may be applied sequentially to the membrane. Binding of such antibodies to moieties present on the membrane is then detected using any suitable immunoassay procedure (e.g. see: Stites and Terr (eds) "Basic and Clinical Immunology", (7 ed) 1991). A particularly suitable procedure is to treat the antibodies in the panel as primary antibodies in a "sandwich" type assay. Unbound primary antibodies are washed away or otherwise removed. The membrane is then treated with secondary antibodies which are reactive with the primary antibodies. The secondary antibody may be bound to a detectable label or fused with an enzyme. Secondary antibody bound to primary antibody is detected by observing the label or the activity of the fused enzyme. Suitable labels and enzymes are known in the art and include magnetic or coloured beads, fluorescent dyes, radiolabels. horseradish peroxidase, alkaline phosphatase, etc. The enzyme linked sandwich type assay (ELISA) is a particularly suitable methodology for use in this invention.

Antibodies for use in this invention may be obtained commercially or prepared using standard techniques. A variety of anti-kinase and anti-kinase substrate polyclonal antibodies are commercially available from various sources, including the following: Biomol Research Laboratories, Inc. (Plymouth Meeting, Pennsylvannia)

Biosource International, Inc. (Camarillo, California)

Promega Corporation (Madison, Wisconsin)

Santa Cruz Biotechnology (Santa Cruz, California)

Sigma (Saint Louis, Missouri) StressGen Biotechnologies Corp. (Victoria, British Columbia)

Transduction Laboratories (Lexington, Kentucky)

Upstate Biotechnology Inc. (Lake Placid, New York)

Zymed Laboratories Inc. (South San Francisco, California)

Antibodies to new kinases or kinase substrates may be prepared as described below. Typically, new kinases are partially purified by techniques such as column chromatography and SDS-PAGE. Microsequencing of partially purified kinases permits comparison to known kinases and possible development of immunological techniques for recovery of more of the new kinase by making use of cross reactivity with known antibodies. Antibodies can be raised against protein kinases or substrates in various host animals, including but not limited to cattle, horses, rabbits, goats, sheep and mice.

Polyclonal antibodies can be obtained from immunized animals and tested for specificity using standard techniques. Alternatively, monoclonal antibodies may be prepared using any technique that provides for production of antibody molecules by continuous cell lines in culture, including the hybridoma technique of Kohler and Millstein, the human B-cell hybridoma technique, and the EBV-hybridomain technique. Alternatively, techniques for the production of single chain antibodies and antibody fragments that contain specific binding sites for a protein kinase or substrate may be generated by known techniques and employed in this invention. Such fragments include F(ab')₂ fragments that may be generated by digestion of an intact antibody molecule and Fab fragments that may be generated by severing dissulfide bridges in F(ab') ₂ fragments or through the use of Fab expression libraries. Preferably, none of the antibodies in a given mixture to be used as a panel in this invention will cross-react with proteins that overlap in size. This may compromise interpretation of the result. Each antibody panel mixture should be blended to avoid such overlaps. Furthermore, every mixture should be adjusted for the concentration of each antibody so that there is optimal detection of the individual target kinases in diverse cell and tissue samples.

Following incubation of intact membranes (in combination with a multiblotter apparatus) or cut strips with different mixtures of primary antibodies, the strips or membranes are incubated with a secondary antibody (e.g. a goat antibody that recognize rabbit antibody) that reacts with the primary antibody. The secondary antibody is fused with an enzyme (e.g. alkaline phosphatase or horse radish peroxidase) to facilitate detection of the positions of the primary antibody, to which it binds by producing a light emission in an enzymatic reaction. Strips may then be reassembled to appear in the order of the original membrane. The intact membrane or reassembled membrane may be subjected to enhanced chemiluminescence (ECL) and exposure to x-ray film or detected by a phosphoimager (e.g. Fluor-S Max Multi-imager from Bio-Rad Laboratories). In this indirect manner, the original positions of resolved protein kinases or kinase substrates can be visualized as dark bands on a transparent background. The intensity of the bands can be quantitated by densitiometric analysis. In many cases, quantitation of the amounts of a given protein kinases in the upper, phosphorylated form and the lower dephosphorylated form can provide an accurate measurement of how much of the kinase is in the inactive and active states.

The intensity of the signal that is generated for a protein kinase band may be readily quantified using known technologies. For example, quantification may be done by using a Fluor-S Max Multi-imager from Bio-Rad Laboratories Canada Ltd., of Mississauga, Ontario, Canada. This equipment can quantize changes in band intensity in range of 1:100,000 but 1:1,000 is more typical. Multiple exposures of X-ray films to ECL for detection of immunoreactive bands to compensate for any non-linearity of response of the film prior to quantization by densitometric analysis could be performed as an alternative method. Each immunoreactive band may be assigned a set of parameters that includes its relative optical density, molecular weight and immunoreactivity. The relative optical density (R.O.D.) value of an immunoreactive protein band is based on the ratio of the intensity of that protein relative to the intensity of a protein kinase band that serves as an internal control. For example, the mitogen-activated protein (MAP) kinase Erkl in 50 μg of rat brain cytosolic protein detected with Erkl-CT antibody could serve as such an internal control. Erkl has been found to be one of the most uniformly expressed protein kinases in different rat tissues and diverse organisms. Alternative standards could be the zeta isoform of protein kinase C or the alpha isoform of p38 Hog MAP kinase. If a protein has the same intensity on a Western blot as Erkl in rat brain, then it has an R.O.D. value of 100. Another parameter may be the molecular mass of an immunoreactive protein band, which is based on its migration on the SDS-PAGE gel relative to known molecular mass marker proteins such as phosphorylase, bovine serum albumin, ovalbumin, glyceraldehyde 3-phosphate dehydrogenase and lysozyme. A further parameter may be the immunoreactivity of a protein band, which is somewhat selective, and particularly appropriate when the immunogen to which the antibody was originally developed is considered. For example, an antibody developed against the C-terminal 40 amino acids of the rat brain Erkl isoform would be expected to immunoreact with the full-length 44 kDa form of Erkl on Western blots of rat brain cytosol.

Existing software may be used to produce digital data values representing the physical properties of kinases, such as their amounts in phosphorylated and dephosphorylated states respectively, for example. Such software may produce such digital data and stores it (e.g. on a floppy diskette) or other medium, in a MICROSOFT EXCEL™ spreadsheet table format readable by EXCEL 98™ software from Microsoft Corporation, Redmond, Washington, USA, however, other formats may be substituted. An example of suitable software for densitometric quantization of chemiluminescence generated by ECL from a Western blot of target proteins is the QUANTITY ONE™ software from Bio-Rad Laboratories Canada Ltd., of Mississauga, Ontario, Canada. This invention offers advantages over standard 2D gel proteomic methods. This technique can be applied to any cell or tissue sample. No prelabelling with radioisotopes is necessary, because kinase detection is based on immunoreactivity. The technology may be adapted for wide scale diagnostic applications because patterns of protein kinase expression are stable for periods of up to six hours before an organ is subjected to fractionation and freezing, providing the organ is stored during this time over ice. This procedure may be carried out within two days from start to finish. By contrast, the 2D gel electrophoresis approach is extremely laborious, much more difficult to render and takes at least twice the time. This invention provides the ability to compare multiple samples side by side. Whereas two or more samples can be analyzed on the same ID gel, a 2D gel can only be used for a single sample. It is more difficult to compare two different samples by the 2D gel route, because of potential variations in the setting up, running and analysis of separate 2D gels.

One of the reasons why 2D gel electrophoresis has become the industry standard for proteomic analysis is the remarkable resolving power of the method with potentially thousands of spots being distinguishable on a 2D gel. Most of these spots, however, are "fuzzy" in appearance and may be overlapping. The method of this invention provides much tighter protein bands with as much as a 2- to 4-fold better resolution in the SDS- PAGE size-separation dimension. With detection based on immunoreactivity, the background of metabolic enzymes and structural proteins is essentially eliminated. This background is problematic even for 2D gel maps of phosphoproteins, since a third of all the proteins inside of cells appear to be phosphorylatable. In one exemplary embodiment of this invention, about 50 μg of a control cell extract from untreated or healthy cells is loaded on to a SDS-PAGE gel in odd numbered lanes. In adjacent, even numbered lanes, equivalent amounts of experimental extracts are deposited. The latter samples are from cells that have been treated with a hormone or drug or have been obtained from diseased tissue. The extracts may be prepared by homogenizing cells in buffer containing a detergent such as 0.5% Triton X-100™ and protein phosphatase inhibitors (to preserve the state of protein phosphorylation in the sample). The extracts are then subjected to ultracentrifugation to remove insoluble matter.

To optimize the detection of protein band shifts, the SDS-PAGE gel is precast with a higher than normal concentration of acrylamide and a lower than normal concentration of bisacrylamide. An electric current is applied to the slab gel until proteins with a molecular mass less than 27,000 Dalton are eluted from the bottom of the gel. The proteins remaining on the slab gel are then electro-transferred on to a nitrocellulose or PVDF membrane that traps the proteins. The membrane is cut into separate strips that each contain samples of the resolved proteins from both control and experimental cell extracts. Each strip is probed with a different mixture of primary antibodies (e.g. from rabbit) that react with a distinct subset peptide or protein substrate by the protein kinase of interest. Each reaction is conducted in a separate tube or well of a microtitre plate.

One application of this invention is for the discovery of novel protein kinases. The following strategy would permit rapid acquisition of protein kinase drug targets. The objective of this approach is to identify those protein kinases that demonstrate increased expression or phosphorylation in association with a disease state or in response to an extracellular signal such as mitogen, drug or stress factor. The approach is based on the following:

1. Antibodies developed for one protein kinase can cross-react with structurally related protein kinases. 2. A band shift of a cross-reactive protein on an immunoblot is commonly due to phosphorylation, and increased phosphorylation is usually associated with activation of the kinase. Greater than 90% of the known protein kinases are phosphorylated in their active states. One of the exceptions is glycogen synthase kinase-3, which is inhibited when it is phosphorylated on serine by protein kinase B. However, activation of glycogen synthase kinase-3 is still dependent on tyrosine phosphorylation of this kinase.

3. Proteins that cross-react with protein kinase antibodies and also bind to gamma- ATP-agarose beads have a very high probability of being protein kinases. This resin will capture many ATP binding proteins in addition to protein kinases but this procedure can purify kinase by up to 200-fold. 4. Proteins that autophosphorylate with [γ-³²P]ATP following immunoprecipitation with protein kinase antibody are likely to be protein kinases. Most antibodies are unsuitable for immunoprecipitation of proteins, and may require partial denaturation of the proteins. Denatured kinases would have little or nor autophosphorylating activity.

5. A combination of gamma- ATP-agarose, immunosorbent and fast protein liquid chromatophy column steps followed by SDS-PAGE permits rapid purification of an immunoreactive protein to allow for its identification by sequencing.

6. There is a likelihood that a protein kinase detected with antibodies is novel. Of the 2000 or so kinases expected to exist, only about a third have been fully sequenced. Nevertheless, partial cDNA sequences for most protein kinases are available in public and private EST cDNA sequence databases. Once a portion of the cDNA structure of the protein kinase gene is available, it is straightforward to obtain the complete nucleotide and amino acid structures of the gene and its protein with standard methodologies.

One of the beneficial outcomes of this invention is that unknown proteins which can cross-react with the kinase-specific antibodies are detected. Those unidentified proteins that change in their abundance or their phosphorylation state in response to a disease condition or treatment are worthy of closer analysis. If such proteins can be shown to bind to ATP-agarose or capable of autophosphorylation with radioactively labelled ATP, then there is a high probability that they are protein kinases. Moreover, with the antibody that was originally used to detect a putative kinase, it is possible to rapidly purify the protein so that it can be sequenced by the Edman degradation method or identified by mass spectroscopy of trypsin digested fragments of the protein. If any part of the protein has been previously sequenced, it would be available in public or private protein sequence databases. A partial sequence in the human EST sequence database may be available. From this information, a full length cDNA sequence for the protein could be rapidly obtained using PCR-based techniques. This would be worthwhile if the cDNA sequence contained conserved kinase catalytic subdomain sequences. In this manner, novel protein kinases that display desirable characteristics (e.g. increased expression in solid tumour relative to adjacent, normal tissue) can be detected and identified. If the inappropriate activity of such protein kinase is shown to contribute to the development of the disease, then they would be most valuable drug targets.

Measurement of the activation of a protein kinase by this invention is dependent on the detection of its band shift on SDS-PAGE gels. A limited number of protein kinases do not exhibit a band shift change when they are activated. However, their in vivo substrates can display band shifts upon their phosphorylation. This can be exploited for the development of in vitro and in vivo substrate assays. Current approaches for high throughput screening of protein kinase inhibitors in vitro involve the use of radioactive [γ-³²]ATP and measurement of the incorporation of the radioactive phosphate into a peptide or protein substrate by protein kinase of interest. Each reaction is conducted in a separate tube or well of a microtitre plate generating high volumes of radioactive garbage. There are many examples of proteins that are highly specific substrates of particular protein kinases; examples include glycogen phosphorylase for phosphorylase kinase, myosin light chain for myosin light chain kinase, elF2α for PKR, MARCKS for protein kinase C, Erkl and Erk2 for Mekl and Mek2. Antibodies are commercially available for many of these substrates or may be produced as described above. Such antibodies may be used to probe for the phospho-states of the substrates, as revealed by their mobility on immunoblots of SDS-PAGE gels. These substrates would not have to be purified from crude cellular extracts for use in the protein kinase assays. However, since many of the substrates may already exist in phosphorylated forms in cell extracts, it may be necessary to incubate the extracts with active preparations of protein phosphatases, which can be subsequently inactivated with phosphatase inhibitors prior to the kinase assays. With this method, a crude mixture of active protein kinases may be added to the phosphatase-treated cellular extract in a single tube, and the phosphorylation reaction can commence with the inclusion of non-radioactive ATP. Only catalytic amounts of protein kinases will be necessary, so any phosphorylated substrates that contaminate the preparation of protein kinases will be relatively minor compared to the amounts of the substrates in the phosphatase-treated cell extracts. Any kinases that contaminate the phosphatase-treated cell extracts would not be a concern, since they are actually desirable.

It may be necessary to add a kinase preparation after phosphatase treatment of the substrate extracts, because many protein kinases are inhibited when they are dephosphorylated. After a short suitable incubation time, the reactions can be terminated by addition of SDS-PAGE sample buffer. Such substrate analysis can be performed as described above for protein kinases, except that panels of antibodies for the kinase substrates will be used in place of the kinase antibody panels. By this approach, the decreased mobility of the kinase substrates will be evident as band shift on the immunoblots in the absence of kinase inhibitors. The presence of specific protein kinase inhibitors would be revealed by the inhibition of the appearance of the upper bands.

In vitro substrate analysis according to this invention would be ideal for the further characterization of compounds that have already been shown to display inhibitor activity toward a kinase and the selectivity of these compounds is in question. A distinct advantage of this method is that it would be easy to compare the findings with a substrate analysis in vivo assay performed using the same blends and concentrations of kinase substrate antibodies that work in the in vitro kinase assay. However, the analysis would be performed on extracts from cells that have been incubated with agonists that stimulate the kinases of interest. These cells would also be exposed to the compounds that exhibit inhibitory activity towards kinase in vitro. In this manner, the efficacy of these inhibitors could be evaluated in living cells.

Another important application of the invention is to identify known proteins and detect novel proteins that may bind with high affinity to a target protein. Such a target protein, which could be another protein kinase or another type of protein, could be expressed as a recombinant fusion protein, for example with glutathione S-transferase or a poly histidine tag. The recombinant protein could be adsorbed on to agarose beads, in these examples with either glutathione or a nickel-containing moeity. A crude tissue lysate could then be incubated with the agarose-beads with the attached target protein, and tissue proteins can be affinity purified by binding to the target protein. The adsorbed proteins would be subjected to the method of this invention to detect retained proteins that can be visualized with the panel of antibodies employed. The ratio of the ECL signal detected for the immunoreactive protein after affinity chromatography divided by the ECL signal detected for the immunoreative protein in the unfractionated tissue lysate may be a relative measure of its affinity for the target protein. For novel immunoreactive proteins detected by this method, their identification would be relatively easy, since the antibody used to detect this protein is a powerful probe to monitor its purification sufficiently for identification from protein sequencing or MALDI MS/MS mass spectrometry. An advantage of this approach over other methods to examine protein-protein interactions such as the yeast two-hybrid method, is that it can detect interactions that are affected by the state of post-translation regulation of these proteins, such as their phosphorylation state. While this invention can provide quantitative information about the expression levels of proteins and their relative states of covalent modification, the method can also be exploited in a qualitative manner to generate a pattern that is distinctive for a given perturbation of a model system, for example with a drug or toxin. It is possible to generate a library of gel immunoblotting patterns for wide range of hormones, cytokines, drugs and toxins of known and specific mechanisms of action. The gel pattern for a new agent of unknown mechanism of action could be produced using the exact same biological model system. If the immunoblot pattern generated with the new agent precisely matched that of a known compound, then it can be assumed that the two have similar if not equivalent mechanisms of action.

By way of illustration, examples from different cell types are given to demonstrate the present invention. The following examples are not intended to be limiting of the invention.

Example 1

In this example, the method of this invention was used to probe for the presence of over 45 different protein kinases in soluble extracts prepared from the whole brain, heart and skeletal muscle of adult male Sprague-Dawley rats. The results demonstrate large differences in kinase expression patterns between these tissues.

Materials Affinity-purified rabbit polyclonal or monoclonal antibodies and immunizing peptides used to raise these antibodies are listed in Table 1. These antibodies were either prepared or obtained commercially. Goat anti-rabbit IgG conjugated to alkaline phosphatase (AP) was obtained from Calbiochem (San Diego, CA). Enhanced chemiluminescence (ECL) detection reagents for immunoblotting were obtained from Amersham Pharmacia Biotech, Inc. (Baie dUrfe, Quebec). Other reagents were obtained from Sigma- Aldrich (St. Louis, MI), unless otherwise stated.

Preparation of rat tissue extracts Brains, hearts and hind leg tibial skeletal muscles from 50-day old male Sprague-

Dawley rats were rapidly excised, after induction of anesthesia by intraperitoneal injection of pentobarbital (60 mg/kg). The tissues were cut, rinsed with phosphate buffered saline at 4 °C, frozen in liquid nitrogen, and stored at -70 °C until use. The tissues were pulverized with 5 strokes of a liquid nitrogen-cooled hand French press and re-suspended in 10 volumes of ice-cold homogenization buffer containing: 20 mM MOPS, 15 mM EGTA, 2 mM Na EDTA, 1 mM NasVO4, 1 mM dithiothreitol, 75 mM β- glycerophosphate, 0.1 mM phenylmethanesulfonyl fluoride, 1 μg/ml aprotinin, 0.7 μg/ml pepstatin, 1 μg/ml leupeptin, and 1 % Triton X-100. This was then sonicated with a Branson Probe Sonicator at 4°C with 3 x 30 s bursts. The homogenates were ultracentrifiiged at 100,000 rpm (240,000 x g) for 15 min in a Beckman TLA-100.2 ultracentrifuge at 4 °C. The supernatants were immediately frozen at -70 °C until subsequent analysis.

Table 1. Sources of antibodies.

Kinase Ab Name Cone. Source & Catalog No.

( (μμgg//mmll)) or immunizing peptide

1. Calmoduim-dependent kinase kinase UaMK -Cl ^~D.2 AStressGen KAP-CA001

2. Calmodulin-dependent kinase 4 CaMPK4-NT 1.0 StressGen KAP-CA003

3. Cyclin-dependent kinase 1 (cdc2) Cdc2-CT 0.7 StressGen KAP-CC001

4. Cyclin-dependent kinase 2 Cdk2 1.6 StressGen KAP-CC007E

5. Cyclin-dependent kinase 5 Cdk5-CT 2.4 Upstate 06-258

6. Cyclin-dependent kinase 6 Cdk6 0.8 StressGen KAP-CC006E

7. Cyclin-dependent kinase -7 M015-PCT 0.5 StressGen KAP-CC010E

8. Cyclin-dependent kinase 8 Cdk8-NT 2.0 StressGen KAP-CC008E

9. Casein kinase 1 « and ε CK1 SG 1.0 StressGen KAP-ST103E

10. Casein kinase 2α CK2α-III 1.2 StressGen KAP-ST010

11. Cot (Tpl2) Cot-PCT 0.5 EESEMLKRQRSLYIDGC² (SEQ ID No: 1)

12. p43 and p45 MAP kinase homologues Erkl-ffl 0.6 Upstate 06-183

13. Erkl and Erk2 MAP kinases Erkl-CT 0.3 Upstate 06-182

14. Erk5 MAP kinase (Bmk) Erk5-PNT 4.0 SAEPPAREGRTRPHRC² (SEQ ID NO: 2)

15. Glycogen synthase kinase β GSK3β-XI 2.0 StressGen KAP-ST002E

16. Integrin linked kinase 1 ILK1 SG 1.0 StressGen KAP-ST203

17. Kkialre Cdk-like kinase Kkialre-CT 3.0 StressGen KAP-CC003

18. Kinase-suppressor of Ras Ksrl-CT 1.0 EKLPKLNRRLSHPGHFWKSC² (SEQ ID NO: 3)

19. MAP kinase-activated kinase 2 MAPKAPK2-PCT 1.0 StressGen KAP- A015E

20. MAP kinase kinase 1 Mekl-XI 2.0 Upstate 06-235

21. MAP kinase kinase 3 Mkk3-CT 1.0 Upstate 06-615

22. MAP kinase kinase 4 Mkk4-XI 1.6 Upstate 06-281

23. MAP kinase kinase 5 Mek5-PNT 1.6 StressGen KAP-MA003

24. MAP kinase kinase 6 Mek6-SG 0.8 StressGen KAP-MA014E

25. MAP kinase kinase kinase 1 Mekkl-PNT 1.2 StressGen KAP-SA010

26. MAP kinase kinase kinase 3 Mekk3 SG 0.5 StressGen KAP-MA013E

27. Mos Mos-III 1.0 StressGen KAP-MA004

28. p38 MAP kinase p38 Hog-CT 0.2 StressGen KAP-MA009E

29. p21 -activated kinase & Pak (C-19) 0.8 SantaCruz sc-881

30. Piml Piml-T 1.4 StressGen KAP-ST004

31. Protein kinase A (cAMP-dep. kinase) PKA-NT 3.0 StressGen KAP-PK001

32. Protein kinase C β PKC-β M7 mAb 0.5 Gift from Susan Jaken

33. Protein kinase C ε nPKC-ε (C-15) 0.5 SantaCruz sc-214

34. Protein kinase C ζ PKC-ζ (C-20) 0.5 SantaCruz sc-226

35. Protein kinase G (cGMP-dep. kinase) PKG1-CT 2.6 StressGen KAP-PK005

36. Protein kinase B α PKB-CT 1.8 Upstate 06-276

37. Protein kinase B β PKB2-PCT 2.0 CRYDSLGLLEDQRT² (SEQ ID NO: 4)

38. RafB RafB-CT 3.0 StressGen KAP-MA006

39. Ribosomal S6 kinase 1 Rskl (C-21) 0.7 SantaCruz sc-231

40. Ribosomal S6 kinase 2 Rsk2-PCT 0.8 StressGen KAP-ST007

41. S6 protein kinase S6K-PNT 1.8 Upstate 06-321

42. Stress-activated kinase (Jnk) SAPKβ 1.6 StressGen KAP-SA004

43. TGFβ -activated kinase Takl-CT 1.0 StressGen KAP-ST009E

Notes

(1) Upstate = Upstate Biotechnology Inc. (Lake Placid, New York, U.S.A.); StressGen = StressGen Biotechnologies Corp. (Victoria, British Columbia, Canada); Santa Cruz Biotechnology, Inc. (Santa Cruz, California, U.S.A.). Table 1. (continued)

(2) All of these antibodies are commercially available, except for Cot-PCT, Erk5-PNT, Ksrl-CT, PKC- β M7 and PKB2-PCT. For these antibodies, the amino acid sequence of the immunizing peptide is provided.

(3) All of the antibodies used were rabbit polyclonal antibodies with the exception of the PKC-β M7 mouse monoclonal antibody provided by Dr. Susan Jaken of the W. Alton Jones Cell Science Center in Lake Placid, New York.

Gel electrophoresis and immunoblotting '

The thawed cell lysates were measured for protein content using Bradford reagent (Bio-Rad) with bovine serum albumin as the reference standard. The protein concentration of the lysates was adjusted to 1 mg/ml in SDS-PAGE sample buffer (2% SDS, 5% glycerol, 50 mM Tris-HCl, pH 6.8, 0.1 M β-mercaptoethanol and 0.01% bromophenol blue) and boiled at 100 °C for 3 min. One mg of the cell lysate was loaded on to the stacking layer (2 mm x 4 cm x 20 cm; 4% acrylamide/0.11% bisacrylamide) of an SDS-PAGE gel. A comb was not used to create individual lanes, so that there was a single, wide lane over the width of the entire gel. The stacking gel was previously layered over a separating SDS-PAGE gel (2 mm x 12.5 cm x 20 cm; 13% aery lamide/0.086% bisacrylamide). Electrophoresis was performed at 30 mA (maximum voltage) and was continued until proteins of 25,000 Daltons had migrated to the bottom of the gel. The composition and concentrations of the other ingredients in the stacking and separate gels, and in the lower and upper chamber gel buffers were as described (Laemmli, U.K. (1970) Nature 227, 680-685). Proteins were then electrophoretically transferred from the gel at 300 mA (maximum voltage) for 3 h on to a nitrocellulose membrane, and the membrane was subsequently cut vertically into 1 cm wide strips. The strips were then blocked with 5% skim milk powder in Tris-buffered saline (20 mM Tris/HCl, pH 7.5, 0.5 M NaCl, 0.2 % Tween-20; TBST) and, after quickly rinsing the membrane with TBST, each strip was exposed to a unique mixture of different primary antibodies in TBST for 3 h with constant shaking at room temperature. The concentrations of the antibodies that were used are provided in Table 1. The strips were washed two times for 15 min with TBST and incubated with horse radish peroxidase-conjugated secondary antibody (goat anti-rabbit or anti-mouse IgG) in TBST for 30 min. After washing the strips three times for 10 min with TBST, the strips were reassembled, and subjected to the ECL Western blotting detection system (Amersham Pharmacia Biotech, Inc.). Exposure of the x-ray films was for 40 sec.

Results

Figure 1 shows an example of the application of the above-described multi-kinase immunoblotting technique of this invention to three different rat tissues. It is evident that the patterns of kinase expression differed markedly between the tissues. At least 45 known protein kinases were visualized on the immunoblots and clearly identified based on their predicted sizes and immunoreactivities. From the intensity of the signals for the immunoreactive kinase bands in Fig. 1, the relative expression levels of these specific protein kinases are provided in Table 2. As demonstrated by this example, this invention may be especially useful for tracking kinases as a function of development, long term mitogen, stress or drug stimulation, and disease progression.

Table 2. Relative expression of known protein kinases in rat tissues.

Kinase Relative Expression

Brain Heart Skeletal muscle

1. Calmodulin-dependent kinase kinase Moderate Low Low

2. Calmodulin-dependent kinase 4 Moderate None None

3. Cyclin-dependent kinase 1 (cdc2) Low None None

4. Cyclin-dependent kinase 2 None Low None

5. Cyclin-dependent kinase 5 Moderate None None

6. Cyclin-dependent kinase 6 Low Moderate Moderate

7. Cyclin-dependent kinase 7 Moderate Low Low

8. Cyclin-dependent kinase 8 Low Low Low

9. Casein kinase 1 o Low Low Low

10. Casein kinase 1 ε Moderate Low Low

11. Casein kinase 2« High High High

12. Cot (Tpl2) Low High Low

13. Erkl MAP kinase High High High

14. Erk2 MAP kinase High High High

15. Erk5 MAP kinase (Bmk) Moderate None None

16. Glycogen synthase kinase β High High Moderate

17. Integrin linked kinase 1 Moderate Moderate Moderate

18. Kkialre Cdk-like kinase None Moderate High

19. Kinase-suppressor ofRas Low None None

20. MAP kinase-activated kinase 2 High Low None

21. MAP kinase kinase 1 (Mekl) High Low Low

22. MAP kinase kinase 3 (Mek3) Low Low Low

23. MAP kinase kinase 4 (Mek4) High Low Low

24. MAP kinase kinase 5 (Mek5) High Moderate Low

25. MAP kinase kinase 6 (Mek6) High High High

26. MAP kinase kinase kinase 1 (Mekkl) Low Moderate Low

27. MAP kinase kinase kinase 3 (Mekk3) Low Low Low

28. Mos None Low Moderate

29. p38 c. MAP kinase Moderate High Moderate

30. p21 -activated kinase α High Low Low

31. Piml Moderate High Low Table 2. (continued)

32. Protein kinase A (cAMP-dep. kinase) Moderate High High

33. Protein kinase B α (Aktl) High Low Low

34. Protein kinase B β (Akt2) High Low Low

35. Protein kinase C β High Low Low

36. Protein kinase C ε High High High

37. Protein kinase C ζ High High High

38. Protein kinase G (cGMP-dep. kinase) Low High High

39. Raffl High Low None

40. Ribosomal S6 kinase 1 High High Moderate

41. Ribosomal S6 kinase 2 Low Low High

42. S6 protein kinase High Low High

43. Stress-activated kinase βp46 (Jnk) High Low Low

44. Stress-activated kinase βp54 (Jnk) High Moderate moderate

45. TGF β -activated kinase Low Low Low

Example 2

In this example, the method of the invention was used to probe for the presence and activation states of over 45 different protein kinases in soluble extracts prepared from the human Ramos B cell line that have been treated anti-IgM antibody in order to stimulate these cells through a B cell antigen receptor. The results demonstate the band shifting of several protein kinases as a consequence of their increased phosphorylation in response to B cell antigen receptor stimulation.

Materials Affinity-purified rabbit polyclonal or monoclonal antibodies and immunizing peptides used to raise these antibodies are listed in Table 1. Goat anti-rabbit IgG conjugated to alkaline phosphatase (AP) was obtained from Calbiochem. Enhanced chemiluminescence (ECL) detection reagents for immunoblotting were obtained from Amersham Pharmacia Biotech, Inc. Other reagents were obtained from Sigma- Aldrich, unless otherwise stated.

Preparation of cell extracts

The human Ramos B cell line (American Type Culture Collection, Rockville, MD) was cultured in Dulbecco's modified Eagle medium containing 10% heat inactivated fetal bovine serum and 2 mM glutamine at 37°C in a 5% CO2/air mixture. For each experimental analysis, 2 x 10⁷ cells were seeded in a 150 mm culture dish containing 20 ml of medium. Twelve hours prior to cell stimulation, the cells were cultured in the above media in the absence of serum, and then were incubated for 5 min with anti-IgM antibody. Subsequently the cells were lysed in 2 ml of ice-cold buffer that contained 20 mM MOPS, pH 7.2, 5 mM EGTA, 1 % (w/v) Nonidet P-40, 1 mM dithiothreitol, 75 mM β-glycerol phosphate, 1 mM Na3VO , and 1 mM phenylmethylsulfonyl fluoride by sonication with a Branson Probe Sonicator at 4°C with 3 x 30 s bursts. The homogenates were ultracentrifiiged at 100,000 rpm (240,000 x g) for 15 min in a Beckman TLA-100.2 ultracentrifuge at 4 °C. The supernatants were immediately frozen at -70 °C until subsequent analysis. Gel electrophoresis and immunoblotting

The thawed cell lysates were measured for protein content using Bradford reagent (Bio-Rad) with bovine serum albumin as the reference standard. The protein concentration of the lysates was adjusted to 1 mg/ml in SDS-PAGE sample buffer (2% SDS, 5% glycerol, 50 mM Tris-HCl, pH 6.8, 0.1 M β-mercaptoethanol and 0.01 % bromophenol blue) and boiled at 100 °C for 3 min. One mg of the cell lysate was loaded on to the stacking layer (2 mm x 4 cm x 20 cm; 4% acrylamide/0.11 % bisacrylamide) of an SDS-PAGE gel. A twenty lane comb was used, and the extracts from untreated (control) and anti-IgM-treated (experimental) cells were deposited into adjacent lanes. Molecular mass markers (glycogen phosphorylase, bovine serum albumin, ovalbumin, glyceraldehyde 3-phosphate dehydrogenase) were applied to the first and last lanes of the gel. The stacking gel was previously layered over a separating SDS-PAGE gel (2 mm x 12.5 cm x 20 cm; 13% acrylamide/0.086% bisacrylamide). Electrophoresis was performed at 30 mA (maximum voltage) and was continued until proteins of 25,000 Daltons had migrated to the bottom of the gel. The composition and concentrations of the other ingredients in the stacking and separate gels, and in the lower and upper chamber gel buffers were as described in Laemmli, U.K. (1970) Nature 227, 680-685. Proteins were then electrophoretically transferred from the gel at 300 mA (maximum voltage) for 3 h on to a nitrocellulose membrane, and the membrane was subsequently cut vertically into strips that contained one lane of the control and one lane of the experimental samples. The strips were then blocked with 5 % skim milk powder in Tris-buffered saline (20 mM Tris/HCl, pH 7.5, 0.5 M NaCl, 0.2 % Tween-20; TBST) and, after quickly rinsing the membrane with TBST, each strip was exposed to a unique mixture of different primary antibodies in TBST for 3 h with constant shaking at room temperature. The concentrations of the antibodies that were used are provided in Table 1. The strips were washed two times for 15 min with TBST and incubated with horse radish peroxidase-conjugated secondary antibody (goat anti-rabbit or anti-mouse IgG) in TBST for 30 min. After washing the strips three times for 10 min with TBST, the strips were reassembled, and subjected to the ECL Western blotting detection system (Amersham Pharmacia Biotech, Inc.). Exposure of the x-ray films was for 40 sec. Results

In the human Ramos B cell line, many of the tested protein kinases were not detected at measurable levels. Several of the protein kinases present in the Ramos cells were observed to undergo band shifts to high apparent molecular mass species following short term exposure for 5 min to the anti-IgM antibody. These kinases are evident in Fig. 2. This band shifting generally correlated with the phosphorylation and activation of these kinases. Table 3 summarizes the findings of this example with respect to the expression level and band shifting of the 45 known protein kinases that were examined. An umdentified 52-kDa immunoreactive protein (p52) in Lanes 1 and 2 that band shifted upon anti-IgM antibody treatment may correspond to a novel protein kinase within the MAP kinase family, based on immunoreactivity.

Table 3. Protein kinase expression and band shifting in anti-IgM antibody- treated Ramos cells.

Kinase Expression Leve :1 Band Shifted

1. Calmodulin-dependent kinase kinase None

2. Calmodulin-dependent kinase 4 None

3. Cyclm-dependent kinase 1 (cdc2) None

4. Cyclin-dependent kinase 2 None

5. Cyclin-dependent kinase 5 None

6. Cyclm-dependent kinase 6 None

7. Cyclin-dependent kinase 7 None

8. Cyclin-dependent kinase 8 None

9. Casein kinase 1 α High No

10. Casein kinase 1 ε None

11. Casem kinase 2« High No

13. Erkl MAP kinase High Yes

14. Erk2 MAP kinase High ?

15. Erk5 MAP kinase (Bmk) None

16. Glycogen synthase kinase β None

17. Integrm linked kinase 1 None

18. Kkialre Cdk-hke kinase None

19. Kmase-suppressor of Ras None

20. MAP kinase-activated kinase 2 High No

21. MAP kinase kinase 1 (Mekl) High No

22. MAP kinase kinase 3 (Mek3) None

23. MAP kinase kinase 4 (Mek4) None

24. MAP kinase kinase 5 (Mek5) Low No

25. MAP kinase kinase 6 (Mek6) Moderate No

26. MAP kinase kinase kinase 1 (Mekkl) None

27. MAP kinase kinase kinase 3 (Mekk3) None

28. Mos None

29 p38 α MAP kinase Moderate No

30. p21 -activated kinase c. None

32. Protein kinase A (cAMP-dep. kinase) None Table 3. (continued)

33. Protein kinase B α (Aktl) None

34. Protein kinase B β (Akt2) High Yes

35. Protein kinase C β High Yes

36. Protein kinase C ε High Yes

37. Protein kinase C ζ None

38. Protein kinase G (cGMP-dep. kinase) None

39. Raffl None

40. Ribosomal S6 kinase 1 High Yes

41. Ribosomal S6 kinase 2 None

42. S6 protein kinase Moderate Yes

43. Stress-activated kinase βp46 (Jnk) None

44. Stress-activated kinase βp54 (Jnk) Moderate No

45. TGF β -activated kinase High No

Example 3

In this example, the method of this invention was used to probe for the effect of various protein kinase inhibitors on the ability of hepatocyte growth factor (HGF) to activate protein kinases in the Erkl/Erk2, p38 MAP kinase and S6 kinase pathways. This example demonstrates that unique kinase band shift patterns may be produced by different drugs. Thus, this invention may be exploited to determine the mechanisms of action of known drugs and the identification of unknown targets of new drugs.

Materials

The protein kinase inhibitors PD98059 (for Mekl), LY294002 (for phosphatidylinositol 3-kinase), SB203580 (for p38 MAP kinase) and rapamycin (for mTor/FRAP) were obtained from ProMega Corporation (Madison, WI). Affinity- purified rabbit polyclonal or monoclonal antibodies and immunizing peptides used to raise these antibodies are listed in Table 1. Goat anti-rabbit IgG conjugated to alkaline phosphatase (AP) was obtained from Calbiochem. Enhanced chemiluminescence (ECL) detection reagents for immunoblotting were obtained from Amersham- Pharmacia Biotech, Inc. Other reagents were obtained from Sigma-Aldrich, unless otherwise stated.

Preparation of cell extracts

Primary cultures of human ovarian surface epithelial cells were cultured in Dulbecco's modified Eagle medium containing 10% heat inactivated fetal bovine serum and 2 mM glutamine at 37°C in a 5 % CO2/air mixture. For each experimental analysis, 2 x 10⁷ cells were seeded in a 150 mm culture dish containing 20 ml of medium. Twelve hours prior to cell stimulation, the cells were cultured in the above media in the absence of serum, and then were incubated at 37°C in the absence or presence of either 50 μM PD98059, 50 μM LY294002, 10 μM SB203580 or 20 nM rapamycin for 30 min. The cells were subsequently incubated for another 10 min with 20 ng/ml of HGF prior to their lysis at 4°C in 2 ml of buffer that contained 20 mM MOPS, pH 7.2, 5 mM EGTA, 1 % (w/v) Nonidet P-40, 1 mM dithiothreitol, 75 mM β-glycerol phosphate, 1 mM Na3VO , and 1 mM phenylmethylsulfonyl fluoride. The cells were sonicated with a Branson Probe Sonicator at 4°C with 3 x 30 s bursts. The homogenates were ultracentrifiiged at 100,000 rpm (240,000 x g) for 15 min in a Beckman TLA-100.2 ultracentrifuge at 4 °C. The supernatants were immediately frozen at -70 °C until subsequent analysis.

Gel electrophoresis and immunoblotting

The thawed cell lysates were measured for protein content using Bradford reagent (Bio-Rad) with bovine serum albumin as the reference standard. The protein concentration of the lysates was adjusted to 1 mg/ml in SDS-PAGE sample buffer (2% SDS, 5% glycerol, 50 mM Tris-HCl, pH 6.8, 0.1 M β-mercaptoethanol and 0.01 % bromophenol blue) and boiled at 100 °C for 3 min. One mg of the cell lysate was loaded on to the stacking layer (2 mm x 4 cm x 20 cm; 4% acrylamide/0.11 % bisacrylamide) of an SDS-PAGE gel. A twenty lane comb was used, and the extracts from untreated (control; Lane 1), HGF-treated (Lane 2), PD98059- and HGF-treated (Lane 3), LY294002- and HGF-treated (Lane 4), SB203580- and HGF-treated (Lane 5), and rapamycin- and HGF-treated (Lane 6) cells were deposited into adjacent lanes. Molecular mass markers (glycogen phosphorylase, bovine serum albumin, ovalbumin, glyceraldehyde 3-phosphate dehydrogenase, trypsinogen) were applied to the first and last lanes of the gel. The stacking gel was previously layered over a separating SDS-PAGE gel (2 mm x 12.5 cm x 20 cm; 13% acrylamide/0.086% bisacrylamide). Electrophoresis was performed at 30 mA (maximum voltage) and was continued until proteins of 25,000 Daltons had migrated to the bottom of the gel. The composition and concentrations of the other ingredients in the stacking and separate gels, and in the lower and upper chamber gel buffers were as described in Laemmli, U.K. (1970) Nature 227, 680-685. Proteins were then electrophoretically transferred from the gel at 300 mA (maximum voltage) for 3 h on to a nitrocellulose membrane, and the membrane was subsequently cut vertically into strips that contained six lanes of the control and various HGF and inhibitor-treated cells. The strips were then blocked with 5% skim milk powder in Tris-buffered saline (20 mM Tris/HCl, pH 7.5, 0.5 M NaCl, 0.2 % Tween-20; TBST) and, after quickly rinsing the membrane with TBST, each strip was exposed to a unique mixture of different primary antibodies in TBST for 3 h with constant shaking at room temperature. The concentrations of the antibodies that were used are provided in Table 1. The strips were washed two times for 15 min with TBST and incubated with horse radish peroxidase-conjugated secondary antibody (goat anti-rabbit or anti-mouse IgG) in TBST for 30 min. After washing the strips three times for 10 min with TBST, they were subjected to the ECL Western blotting detection system (Amersham Pharmacia Biotech, Inc.). Exposure of the x-ray films was for 40 sec.

Results

Many different protein kinases were clearly detected in the human ovarian surface epithelial (OSE) cells. For the purposes of this example, only the band shifted states of six protein kinases are presented in Figure 3. These kinases are the extracellular regulated kinases Erkl (Panel A) and Erk2 (Panel A) isoforms, p38 Hog MAP kinase (Panel B), protein kinase B alpha (PKBl, Panel C) and beta (PKB2, Panel D) isoforms, and p70 S6 kinase (S6K). A schematic of the signalling pathways within which these kinases participate is present on the right in Figure 3. On the one hand, HGF is known to stimulate the enzyme activities of the Erkl and Erk2 MAP kinases as well as p70 S6K. On the other hand, there are no reports of activation of p38 MAP kinase by HGF. In concordance, HGF treatment of the OSE cells produced reduced mobilities of Erkl, PKBl, PKB2 and S6K, consistent with their phosphorylation and activation. No band shift of Erk2 was evident due to comigration of activated Erk2 with the inactive form of Erkl. There was no change in the mobility in p38 MAP kinase. Consistent with the known action of PD98059, the Mekl inhibitor prevented the HGF-induced band shift in Erkl, but not of PKBl and PKB2. There was some reduction of the HGF-induced S6K band shift, possibly because some of the phosphorylation of S6K was catalyzed by Erkl and Erk2. The phosphatidylinositol 3-kinase (PI3K) inhibitor LY294002 caused a slight reduction in HGF-induced Erkl band shifting, in part because the protein kinase C zeta (PKCξ) isoform is normally activated by the lipid products of the PI3K reaction, and PKCξ causes the activation of Mekl. The inhibition of PI3K by LY294002 completely blocked the HGF-induced band shifting of PKBl, PK2 and S6K, consistent with their actions distal to PI3K. Furthermore, the LY294002 produced band shifts in these kinases to forms of lower molecular mass than were detected in the untreated cells. These results indicate that there was an intermediate state of activation of these kinases in the control cells in the absence of HGF. The p38 MAP kinase inhibitor SB203580 had no discernable effect on any of the kinases in the absence or presence of HGF. The mTOR/Frap inhibitor rapamycin only reduced the band shift in S6K, which is expected since this protein kinase appears to lie upstream of only the S6K and none of the other protein kinases that were tracked.

One application of this invention is for drug profiling. It is evident from Example 3, that the different protein kinase inhibitors generated distinct changes in the 6 protein kinases that were tracked. Potentially several hundred protein kinases can be monitored for the specific effects of selected drugs. The short term actions of the drug on the basal, mitogen-stimulated and stress-stimulated phosphorylation states of the various protein kinases can be assessed. Distinct sets of band shifts should be produced by different drugs. These patterns can be interpreted to deduce the mechanisms of action of these drugs. By matching the patterns of kinase alterations induced by known drugs, it is possible to determine targets of unknown drugs. If two drugs generate exactly the same patterns of kinase changes, then they should have the same cellular target.

Example 4

In this example, the kinase multi-blot analysis is used to probe for changes in the expression of protein kinases in extracts from human breast tumours compared to patient-matched "normal" breast tissue. The method allows for the detection of known protein kinases and other proteins that are increased in samples from diseased tissues or cells. These proteins can serve as markers of disease progression and possibly targets for therapeutic intervention.

Materials Affinity-purified rabbit polyclonal or monoclonal antibodies and immunizing peptides used to raise these antibodies are listed in Table 1. Goat anti-rabbit IgG conjugated to alkaline phosphatase (AP) was obtained from Calbiochem. Enhanced chemiluminescence (ECL) detection reagents for immunoblotting were obtained from Amersham Pharmacia Biotech, Inc. Other reagents were from Sigma-Aldrich, unless otherwise stated.

Tissue procurement and homogenization

Breast tumours from patients and their adjacent control samples were obtained through the Pathology Department at Vancouver General Hospital. The samples were immediately placed in liquid nitrogen until analysis. Homogenization was carried out by placing the tissue in 5 ml of buffer containing 20 mM MOPS, 50 mM β-glycerophosphate, 1 % NP40, 50 mM sodium fluoride, 1 mM Na3VO , 5 mM EGTA, 2 mM EDTA, 1 mM dithiothreitol, 1 mM benzamidine and 1 mM phenylmethanesulphonyl fluoride, 10 μg/ml of leupeptin and aprotinin, and applying three, 20 second bursts of a Brinkman Polytron at a setting of 10,000 rpm whilst on ice. The samples were then subjected to centrifugation at 150,000 x g for 30 min and the supernatant fractions were stored at -70°C until used.

Gel electrophoresis and immunoblotting

The thawed cell lysates were measured for protein content using Bradford reagent (Bio-Rad) with bovine serum albumin as the reference standard. The protein concentration of the lysates was adjusted to 1 mg/ml in SDS-PAGE sample buffer (2% SDS, 5% glycerol, 50 mM Tris-HCl, pH 6.8, 0.1 M β-mercaptoethanol and 0.01 % bromophenol blue) and boiled at 100 °C for 3 min. One mg of the cell lysate was loaded on to the stacking layer (2 mm x 4 cm x 20 cm; 4% acrylamide/0.11% bisacrylamide) of an SDS-PAGE gel. A twenty lane comb was used, and the extracts from patient-matched normal (control) and tumour breast tissue biopsy extracts were deposited into adjacent lanes. The stacking gel was previously layered over a separating SDS-PAGE gel (2 mm x 12.5 cm x 20 cm; 13% acrylamide/0.086% bisacrylamide). Electrophoresis was performed at 30 mA (maximum voltage) and was continued until proteins of 25,000 Daltons had migrated to the bottom of the gel. The composition and concentrations of the other ingredients in the stacking and separate gels, and in the lower and upper chamber gel buffers were as described (Laemmli, U.K. (1970) Nature 227, 680-685). Proteins were then electrophoretically transferred from the gel at 300 mA (maximum voltage) for 3 h on to a nitrocellulose membrane, and the membrane was subsequently cut vertically into strips that contained one lane of the control and one lane of the experimental samples. The strips were then blocked with 5% skim milk powder in Tris-buffered saline (20 mM Tris/HCl, pH 7.5, 0.5 M NaCl, 0.2 % Tween-20; TBST) and, after quickly rinsing the membrane with TBST, each strip was exposed to a unique mixture of different primary antibodies in TBST for 3 h with constant shaking at room temperature. The concentrations of the antibodies that were used are provided in Table 1. The strips were washed two times for 15 min with TBST and incubated with horse radish peroxidase-conjugated secondary antibody (goat anti-rabbit or anti- mouse IgG) in TBST for 30 min. After washing the strips three times for 10 min with TBST, they were subjected to the ECL Western blotting detection system (Amersham Pharmacia Biotech, Inc.). Exposure of the x-ray films was for 2 min.

Results

Many different known protein kinases were clearly detected in the human breast tumours and patient matched "normal" breast biopsy samples. Most of these kinases were unaltered in their relative expression levels. Five protein kinases that consistently exhibited increased amounts in the tumour samples are presented in set of left panels in Figure 4. These kinases were p38 Hog MAP kinase, protein kinase B- alpha (PKBα), the various isoforms of the catalytic subunit of casein kinase 2 (CK2), cGMP-dependent protein kinase (PKG) and cyclin-dependent kinase 8 (Cd8). The elevated amounts of one or more of these kinases may contribute to the neoplastic transformation of breast tissue. Alternatively, they may represent feedback responses to counteract the loss of growth control in the breast tumours. In either event, they may serve as useful markers of cancer progression.

Many antibodies are able to cross-react with related proteins that share the epitopes that are recognized for binding by these antibodies. Since most protein kinases are evolutionarily related, there is a high probability of cross-reactivity of kinase-directed antibodies with homologous kinases. In the experiment described in this example, there were at least 12 unidentified immunoreactive proteins with kinase antibodies that were selectively detected in the tumour samples, but poorly if at all the control samples. Five of these proteins are shown in the immunoblots in the set of panels on the right side of Figure 4. It is possible to identify these proteins following their enrichment by standard purification techniques and protein microsequence analysis. For example, the immunoreactive protein p38 in Figure 4B has been identified as the MAP kinase kinase, Mek6. Three major advantages of kinase purification with antibodies are: (1) it is unnecessary to know what will serve as a selective substrate to momtor the presence of the kinase during its purification; (2) it is not critical to preserve the enzyme activity of the kinase during the purification procedures; and (3) the position of the kinase can be clearly detected on an SDS-PAGE gel by immunoblotting for its sequencing, even if the protein is not completely pure.

Example 5

In this example, standard immunoblotting of 2D isoelectric focusing/SDS- PAGE gels of rat brain extracts was undertaken to demonstrate the limitations of the traditional approach for proteomic analysis as applied to protein kinases.

Materials

The Immobiline DryStrip Kit with immobilized IPG ampholytes was obtained from Amersham Pharmacia Biotech, Inc. The rabbit polyclonal antibodies Erkl-CT (Catalogue No. 06-182) and PKC-III (Catalogue No. KAP-PK003) were obtained from Upstate Biotechnology Inc. and StressGen Biotechnologies Corp., respectively. Goat anti-rabbit IgG conjugated to alkaline phosphatase (AP) were obtained from Calbiochem. Enhanced chemiluminescence (ECL) detection reagents for immunoblotting were obtained Amersham Pharmacia Biotech, Inc. Other reagents were from Sigma-Aldrich, unless otherwise stated.

Preparation of rat brain extract

The brain from a 50-day old male Sprague-Dawley rat was rapidly excised, after induction of anesthesia by intraperitoneal injection of pentobarbital (60 mg/kg). The brain was cut, rinsed with phosphate buffered saline at 4 °C, and re-suspended in 10 volumes of ice-cold homogenization buffer containing: 10 mM Tris, pH 7.2,

0.1 mM EGTA, 0.1 mM Na2EDTA, 1 mM Na3VO4, 50 mM dithiothreitol, 10 mM β-glycerophosphate, 1 mM phenylmethanesulfonyl fluoride, 10% glycerol and 1 % Triton X-100. This was then homogenized with a Brinkman Polytron homogenizer. The homogenates were ultracentrifiiged at 100,000 rpm (240,000 x g) for 15 min in a Beckman TLA-100.2 ultracentrifuge at 4 °C. The supernatant was immediately frozen at -70 °C until subsequent analysis.

Gel electrophoresis and immunoblotting

The thawed cell lysate was measured for protein content using Bradford reagent (Bio-Rad) with bovine serum albumin as the reference standard. One mg of the soluble protein was applied to an Immobiline isoelectric focusing (IEF) Dry Strip and electrophoresis was performed on Hoefer SE600 standard vertical gel system as recommended by the manufacturer (Pharmacia Biotech). The 18 cm, pH 3-10 DryStrip was previously rehydrated in 6 M urea, 2 M fhiourea, 4% CHAPS, 0.02% bromophenol blue, 2% Nonidet P-40, 0.7% dithiothreitol, 10 mM Tris, 2% IPG ampholyte, 10% glycerol, and 4 mM tributyl phosphine. Following the isoelectric focusing, the IPG strip was overlaid lengthwise on to an 11 % SDS- polyacrylamide gel, and electrophoresis was continued into the second dimension. In a separate lane at the left side of the same 11 % SDS-PAGE gel, 200 μg of the cytosolic protein was directly loaded on to the gel prior to the electrophoresis into the second dimension. Subsequently, the resolved proteins were then electrophoretically transferred from the 2D gel at 300 mA (maximum voltage) for 3 h on to a nitrocellulose membrane. The membrane was then blocked with 5% skim milk powder in Tris-buffered saline (20 mM Tris/HCl, pH 7.5, 0.5 M NaCl, 0.2 % Tween-20; TBST) and, after quickly rinsing the membrane with TBST, it was exposed to a mixture of 0.2 μg/ml of Erkl-CT and 1 mg/ml of the PKC-III primary antibodies in TBST for 3 h with constant shaking at room temperature. The membrane was washed two times for 15 min with TBST and incubated with horse radish peroxidase-conjugated secondary antibody (goat anti-rabbit IgG) in TBST for 30 min. After washing the membrane three times for 10 min with TBST, it was subjected to the ECL Western blotting detection system (Amersham Pharmacia Biotech, Inc.). Exposure of the x-ray film was for 5 min. Results

The migration positions of PKCβ, Erkl and Erk2 are shown in Figure 5. The right most of the series of spots for Erkl and Erk2 correspond to their fully dephosphorylated states, and these proteins shifted progressively to the left with the acquisition of each phosphate group. Only one large spot was evident for PKCβ. The large black smear on the left of the 2D gel corresponds to five-fold less brain extract applied directly to the same SDS-PAGE gel than was also loaded on to the IPG Drystrip, which was then transferred into the SDS-PAGE gel. These results were consistently observed in multiple experiments, even when diverse agents were used to try to release the proteins from the IPG Drystrip. The adsorption of protein to the IPG Drystrip was especially problematic for less abundant cellular proteins such as protein kinases. Evidently, less than 5% of the Erkl and Erk2 that loaded on to the IPG Drystrip actually entered the SDS-PAGE gel when electrophoresis was performed in the second dimension. The unpredictable retention of protein kinases and their substrates by the IPG

Dry strips calls into question the useful of standard 2D gel electrophoresis (e.g. according to O'Farrell) for resolution of proteins to perform quantitative studies to monitor changes in protein expression and post-translational covalent modification. Many proteins are least soluble at the pH that corresponds to their isoelectric points when they are uncharged in solution. It is also possible that the phosphorylation state of proteins may also influence their binding to the IPG Drystrips. This would compromise on the analysis of the relative amounts of the dephosphorylated to phosphorylated species of some proteins on 2D gels. It is possible that some of the problems associated with protein retention on the IPG Drystrips with immobilized ampholytes might be avoided through the use of tube gels and soluble ampholytes. However, the protein loading capacity of such tube gels is markedly lower. This could still result in insufficient protein on the 2D gel for detection of protein kinases by silver staining or immunostaining.

An attempt was made to detect more than twenty different protein kinases by immunoblotting 2D gels of rat brain lysates with the antibodies described in Table 1. The immunoblots were performed with mixes composed of two to four different antibodies. With few exceptions, it was very difficult to detect any spots that corresponded to the appropriate sizes of the target proteins. The data shown in Figure 5 represents the best finding using particularity sensitive antibodies. 2D gel electrophoresis is thus insufficiently sensitive and inconsistently reproducible for the simultaneous detection of rare proteins such as protein kinases.

Example 6

In this example, the kinase multi-blot analysis of this invention is used to probe for the presence of over 53 different protein kinases in detergent solubized extracts prepared from the whole brain of 1 and 50 day old male Sprague-Dawley rats. These results demonstrate many differences in the expression patterns associated with development. Materials

Affinity-purified rabbit/goat polyclonal or mouse monoclonal antibodies used to detect protein kinases are listed in Table 4. Most of these antibodies were purchased from commercial sources as indicated in Table 4. Goat anti-rabbit IgG conjugated to alkaline phosphatase (AP) was bought from Calbiochem (San Diego, CA). Enhanced chemiluminescence (ECL) detection reagents for immunoblotting were purchased from Amersham Pharmacia Biotech, Inc. (Baie dUrfe, Quebec). Other reagents were purchased from Sigma- Aldrich (St. Louis, MI), unless otherwise stated.

Preparation of rat tissue extracts

Whole brains from 1 and 50-day old male Sprague-Dawley rats were rapidly excised, after induction of anesthesia by intraperitoneal injection of pentobarbital (60 mg/kg). The brains were cut, rinsed with phosphate buffered saline at 4 °C, frozen in liquid nitrogen, and stored at -70 °C until use. The brains were pulverized with 5 strokes of a liquid nitrogen-cooled hand French press and re-suspended in 10 volumes of ice-cold homogenization buffer containing: 20 mM MOPS, 15 mM EGTA, 2 mM Na₂EDTA, 1 mM Na₃VO₄, 1 mM dithiothreitol, 75 mM β-glycerophosphate, 0.1 mM phenylmethanesulfonyl fluoride, 1 μg/ml aprotinin, 0.7 μg/ml pepstatin, 1 μg/ml leupeptin, and 1% Triton X-100. This was then sonicated with a Branson Probe

Sonicator at 4°C with 3 x 30 s bursts. The homogenates were ultracentrifiiged at 100,000 rpm (240,000 x g) for 15 min in a Beckman TLA-100.2 ultracentrifuge at O ° C. The supernatants were immediately frozen at -70 C until subsequent analysis.

Table 4. Antibody sources, concentrations and mixes.

Lane Final Antibody Commercial Type Catalog No. Target Number Cone. Name Source Protein

Protein Size (kDa)

(μg/ml)

16 3 A t2 (D-17) Santa Cruz gpAb¹ sc-7127 62

19 2.45 CaMPK4-NT rpAb² SG KAP-CA003 65

2 1.125 Cdkl Bio-Source mmAb³ AHZ0122 34

3 2.0 Cdk2 Bio-Source mmAb AHZ0142 34

4 2.0 Cdk4 Bio-Source mmAb AHZ0202 36

5 2.5 Cdk5 (H-291) Santa Cruz rpAb sc-750 34

6 2.0 Cdk6 Bio-Source mmAb AHZ0232 39

8 2.0 Cdk7 (C4) Santa Cruz mmAb sc-7344 39

8 2.0 Cdk7-PCT rpAb M015-PCT 39

17 1.25 CK2α-III Ab SG KAP-ST010E 38-43

12 2.5 DNAPK StressGen rpAb KAP-PI001E 460

2 0.175 Erkl-CT φAb SG KAP-MA001 42-44

3 0.9 Erk2 SG StressGen φAb KAP-MAω7E 42

1 2.0 Erk3 (1-15) Santa Cruz φAb sc-156 62

6 2.5 Erk6 (N-19) Santa Cruz gpAb sc-6023 41

18 0.9 GSK3α/β StressGen mmAb KAM-ST002E 48

10 3.75 Krs-1 (N-19) Santa Cruz gpAb sc-6212 63

11 0.6 Mekl Transd.Lab mmAb M 17020 45

12 0.9 Mek2 Transd.Lab mmAb M24520 44

13 1.0 Mek4 (K-18) Santa Cruz φAb sc-964 44

14 2.2 Mek6 StressGen φAb KAP-MA014 38

15 3.0 Mek7 (T-19) Santa Cruz gpAb sc-7104 43

10 1.05 Mos-III φAb SG KAP-MA004 38

15 1.6 Nek2 (N-20) Santa Cruz gpAb sc-7440 50

0.3 p38 Hog-CT StressGen φAb KAP-MA009 40 Table 4. (continued)

12 0.5 PAKα (C19) Santa Cruz φAb sc-881 67-72

13 2.0 PAKβ (N19) Santa Cruz gpAb sc-1871 67-72

14 3.0 PAKγ (N19) Sant Cruz gpAb sc-1872 67-72

3 3.0 PDK1 (C-19) Santa Cruz gpAb sc-7686 65

19 2.5 Piml (C20) Santa Cruz gpAb sc-7856 36

16 1.0 PKA Transd. Lab mmAb P73420 41

2 3.0 PKB-CT φAb 60

4 0.3 PKC-α Transd.Lab mmAb P16520 80

5 0.6 PKC-βl (C16) Santa Cruz φAb sc-209 80

6 0.4 PKC-γ (C19) Santa Cruz φAb sc-211 80

7 1.5 PKC-δ (C17) Sant Cruz φAb sc-213 78

8 0.8 nPKC-ε (C15) Santa Cruz φAb sc-214 95

9 2.0 PKC-λ(N-17) Santa Cruz φAb sc-1091 75

10 2.0 PKC-Θ Transd.Lab mmAb 15120 79

11 1.0 PKC-ζ (C20) Santa Cruz φAb sc-216 70

4 1.2 PKC-μ (C20) Santa Cruz φAb sc-639 100

16 4.0 PKG1-CT φAb 75

14 4.0 PKN (C19) Santa Cruz gpAb sc-1842 120

15 2.5 PKR (D-20) Santa Cruz φAb sc-708 75

1 0.8 RafB (H145) Santa Cruz φAb sc-9ω2 94

20 2.5 Rockl (K18) Santa Cruz gpAb sc-6056 160

13 1.5 ROKα Transd.Lab mmAb R54520 180

19 1.8 Rskl (C-21) Santa Cruz φAb sc-231 85-90

20 1.5 Rsk2 StressGen φAb KAP-ST007E 75-90

18 3.5 S6K- (C-18) Santa Cruz φAb sc-230 65-70

9 1.5 SAPKβ φAb SG KAP-SA004E 44-54

17 1.5 ZAP70 kinase Transd.Lab mmAb Z24820 70 Table 4. (continued)

Footnotes

¹ gpAb= goat polyclonal antibody

² φAb= rabbit polyclonal antibody

³ mmAb = mouse monoclonal antibody

Gel electrophoresis and immunoblotting

The thawed brain lysates were measured for protein content using Bradford reagent (Bio-Rad) with bovine serum albumin as the reference standard. The protein concentration of the lysates was adjusted to 1 mg/ml in SDS-PAGE sample buffer (2% SDS, 25% glycerol, 50 mM Tris-HCl, pH 6.8, 0.1 M β-mercaptoethanol and 0.01% bromophenol blue) and boiled at 100 °C for 3 min. One mg of the cell lysate was loaded on to the stacking layer (0.75 mm x 0.8 cm x 9 cm; 4% acrylamide/0.11% bisacrylamide) of an SDS-PAGE mini gel designed for a Bio-Rad mini gel system. A comb was not used to create individual lanes, so that there was a single, wide lane over the width of the entire gel. The stacking gel was previously layered over a separating SDS-PAGE gel (0.75 mm x 5.2 cm x 9 cm; 13% acrylamide/0.086% bisacrylamide). Electrophoresis was performed at 30 mA (maximum voltage) and was continued for 45 minutes until proteins of 25,000 Daltons had migrated to the bottom of the gel. The composition and concentrations of the other ingredients in the stacking and separate gels, and in the lower and upper chamber gel buffers were as described in Laemmli, U.K. (1970) Nature 227, 680-685. Proteins were then electrophoretically transferred from the gel at 300 mA (maximum voltage) for 30 min on to a nitrocellulose membrane. The membrane was blocked with 1.5% skim milk powder and 2% bovine serum albumin in Tris-buffered saline (20 mM Tris/HCl, pH 7.5, 0.5 M NaCl, 0.2 % Tween-20; TBST) and, after quickly rinsing the membrane with TBST, the membrane positioned in an Immunetics 20 lane multi-blotter and exposed to a unique mixture of different primary antibodies in TBST for 3 h with constant shaking at room temperature. The final concentrations of the antibodies that were used are provided in Table 4. The membrane was washed two times for 15 min with TBST and incubated with horse radish peroxidase-conjugated secondary antibodies (sheep anti-mouse IgG (Amersham), donkey anti-rabbit IgG (Santa Cruz Biotechnology) or bovine anti-goat IgG (Santa Cruz Biotechnology) in TBST for 30 min. After washing the strips three times for 10 min with TBST, the membranes were subjected to the ECL Western blotting detection system (Amersham Pharmacia Biotech, Inc.) and scanned for chemoluminescence with a Bio-Rad FluorS MAX imager.

Results

Figure 6 shows an example of the application of the multi-kinase immunobotting technique of this invention using mini SDS-PAGE and a commercial 20-lane immunoblotter from Immunetics. At least 40 protein kinases could be visualized on the immunoblots and clearly identified based on their predicted sizes and immunoreactivities. Changes in expression and/or band shifting were evident for 23 kinases between the 1 and 50 day old rat brain samples. The levels of chemoluminescence were determined for each kinase band using a Bio-Rad FluorS Max Imager and Bio-Rad Quantity One software, and these are provided in Table 5. In particular, during development: the protein levels of PKC-α, PKC-γ, PKC-λ, PKC-Θ, PAK-α, ERK6, ERK3, p54 SAPK, MEK1 and MEK7 were increased; the protein levels of PKB-α, PKC-μ, PAK-γ, PKA, CDK5, CDK7, p38-α, MEK4, MEK6, p70 S6 kinase and RSK2 were decreased; the isoform distribution of CK2 changed; and MEK2 was band-shifted up to its phosphorylated and activated form. As demonstrated by this example, this invention may be especially useful for tracking kinases as a function of development, short or long term mitogen, stress or drug stimulation, and disease progression.

Table 5 - Quantitation of intensity of protein kinase bands on multi-kinase immunoblots of brain lysates from 1 and 50 day old rats.

1 day 1 day 1 day 1 day 1 day

Lane Band Protein Identity Relative Trace

Number Number Front ⁽ 3NT x

1 τιm

1 1 RafB PSTK 0.362 18533

1 2 Extracellular regulated 0.469 11773 kinase 3 PSTK P2

1 3 Extracellular regulated 0.481 14753 kinase 3 PSTK P 1

1 4 ? p43 MAP kinase 0.623 8417

2 1 PKB alpha (Akt1) PSTK 0.496 15365

2 2 Extracellular regulated 0.615 19609 kinase 1 PSTK P0

2 3 Extracellular regulated 0.654 23393 kinase 2 PSTK PO

3 1 Extracellular regulated 0.665 29531 kinase 2 PSTK PO

4 1 Protein kinase C mu 0.319 9471 PSTK P1

4 2 Protein kinase C mu 0.342 6753 PSTK P0

4 3 Protein kinase C alpha 0.415 15944 PSTK P0

4 4 ? 0.519 11255

4 5 ? 0.65 4934

5 1 Protein kinase C beta 0.423 39408 PSTK P0

5 2 ? 0.6 14063

5 3 p38 Hog-alpha MAPK 0.665 16494 PSTK P0

5 4 Cyclin-dep. kinase 5 0.812 34832 PSTK P0

6 1 Protein kinase C gamma 0.415 23250 PSTK P0

7 1 Protein kinase C delta 0.442 35548 PSTK P0

8 1 Protein kinase C epsilon 0.365 50521 PSTK P0 Table 5. (continued)

Lane Band Protein Identity Relative Trace

Number Number Front i CNT x

I mm

8 2 Protein kinase C epsilon 0.377 32565 PSTK P1

8 3 ? 0.562 8161

8 4 ? 0.608 8909

8 5 Cyclin-dep. kinase 7 0.685 25945 PSTK

8 6 ? 0.819 15583

9 1 ? 0.362 11983

9 2 ? 0.373 11450

9 3 ? 0.373 12327

9 4 Protein kinase C lambda 0.512 9138 PSTK P0

9 5 SAPK beta (JNK2) p54 0.577 11346 PSTK P0

9 6 SAPK beta (JNK2) p46 0.65 28285 PSTK P0

11 1 Protein kinase C zeta 0.419 35543 PSTK P0

11 2 Mek1 PTYK P0 0.615 7873

12 1 ? 0.331 5045

12 2 ? 0.431 7708

12 3 Pak alpha (Pak1 ) PSTK 0.469 10473 P1

12 4 Pak alpha (Pak1) PSTK 0.477 15507 P0

12 5 Mek2 PTYK P1 0.477 25725

12 6 Mek2 PTYK P0 0.608 11679

12 7 ? 0.662 7819

13 1 Mek4 PTYK P0 0.658 48686

13 2 ? 0.704 32140

14 1 ? 0.5 8571

14 2 ? 0.662 10897

14 3 Mek6 PTYK PO 0.727 21441

15 1 ? 0.5 4948

15 2 Nek2 PSTK PO 0.531 8364

15 3 Mek7 PTYK PO 0.665 8288

16 1 PKA (cAMP-dep. kinase) 0.669 24691 PSTK PO Table 5. (continued)

Lane Band Protein Identity Relative Trace

Number Number Front i CNT x

I mm

17 1 ? 0.492 11641

17 2 Casein kinase 2 alpha 0.658 5882 PSTK PO

17 3 Casein kinase 2 alpha' 0.685 17224 PSTK PO

17 4 Casein kinase 2 alpha" 0.704 8012 PSTK P0

18 1 ? 0.292 17873

18 2 S6K p90 PSTK P1 0.431 21997

18 3 ? 0.435 0

18 4 S6K p70 PSTK P1 0.496 38666

18 5 S6K p70 PSTK PO 0.508 41938

18 6 GSK3 alpha PSTK PO 0.596 16984

18 7 ? 0.623 8753

18 8 GSK3 beta PSTK 0.646 11845

18 9 ? 0.646 14558

18 10 ? 0.677 9570

18 11 ? 0.7 12872

18 12 ? 0.85 17230

19 1 Rsk1 PSTK PO 0.438 55218

19 2 Calmodulin-dep. kinase 4 0.492 23021 PSTK PO

20 1 Rsk2 PSTK PO 0.427 19383

20 2 Mnk2 PO PSTK 0.596 10033

20 3 ? 0.781 21622

Table 5. (continued)

50 day 50 day 50 day 50 day i 50 day

Lane Band Protein Identity Relative Trace

Number Number Front i CNT x

I mm

1 1 RafB PSTK 0.357 14704

1 2 Extracellular regulated 0.469 15896 kinase 3 PSTK P2

1 3 Extracellular regulated 0.469 18221 kinase 3 PSTK P1

1 4 Extracellular regulated 0.481 6879 kinase 3 PSTK PO

1 5 ? p45 MAP kinase 0.504 12389

1 6 ? p43 MAP kinase 0.593 11531

1 7 ? 0.651 7723

2 1 PKB alpha (Akt1) PSTK 0.484 9932

2 2 Extracellular regulated 0.628 13846 kinase 1 PSTK PO

2 3 Extracellular regulated 0.659 14467 kinase 2 PSTK PO

3 1 Extracellular regulated 0.663 13301 kinase 2 PSTK PO

4 1 Protein kinase C alpha 0.415 25125 PSTK PO

4 2 ? 0.512 9384

4 3 ? 0.616 7471

5 1 Protein kinase C betal 0.419 37187 PSTK PO

5 2 ? 0.605 11492

5 3 ? 0.841 17086

6 1 Protein kinase C gamma 0.419 31227 PSTK PO

6 2 Extracellular regulated 0.593 8803 kinase 6 PSTK P 1

6 3 Extracellular regulated 0.616 9806 kinase 6 PSTK PO

7 1 ? 0.353 7382

7 2 ? 0.364 5018

7 3 Protein kinase C delta 0.438 32009 PSTK PO

7 4 ? 0.5 6150

7 5 ? 0.616 8910 Table 5. (continued)

Lane Band Protein Identity Relative Trace

Number Number Front < CNT x

I mm

7 6 ? 0.632 12062

8 1 ? 0.26 4222

8 2 Protein kinase C epsilon 0.353 35558 PSTK PO

8 3 Protein kinase C epsilon 0.372 19836 PSTK P1

8 4 ? 0.508 7898

8 5 ? 0.55 7711

8 6 ? 0.585 14390

8 7 0.609 8916

8 8 Cyclin-dep. kinase 7 0.682 7995 PSTK

9 1 ? 0.357 9267

9 2 ? 0.368 8738

9 3 Protein kinase C lambda 0.504 13755 PSTK PO

9 4 SAPK beta (JNK2) p54 0.581 17209 PSTK PO

9 5 SAPK beta (JNK2) p46 0.647 21514 PSTK PO

10 1 Protein kinase C theta 0.419 12320 PSTK PO

10 2 Mos PSTK PO 0.713 9756

11 1 Protein kinase C zeta 0.415 24956 PSTK PO

11 2 ? 0.527 3801

11 3 Mek1 PTYK PO 0.632 15379

12 1 ? 0.306 3976

12 2 ? 0.415 9167

12 3 Pak alpha (Pak1 ) PSTK 0.461 18990 P1

12 4 Pak alpha (Pak1 ) PSTK 0.473 15272 PO

12 5 Mek2 PTYK P1 0.616 7714

12 6 ? 0.628 7532

13 1 ROK alpha PYK PO 0.256 15200

13 2 Pak beta (Pak1) PSTK PO 0.469 16452

13 3 Mek4 PTYK PO 0.663 21963 Table 5. (continued)

Lane Band Protein Identity Relative Trace

Number Number Front < 3NT x

I mm

13 4 ? 0.705 25512

14 1 Mek6 PTYK PO 0.725 10467

15 1 ? 0.357 6804

15 2 ? 0.469 7507

15 3 Nek2 PSTK PO 0.512 7422

15 4 Mek7 PTYK PO 0.651 5217

16 1 PKG1 (cGMP-dep. 0.438 4686 kinase1) PSTK PO

16 2 PKA (cAMP-dep. kinase) 0.659 9564 PSTK PO

17 1 ? 0.399 4803

17 2 Casein kinase 2 alpha 0.643 9746 PSTK PO

17 3 Casein kinase 2 alpha' 0.667 10639 PSTK PO

18 1 ? 0.318 8933

18 2 S6K p90 PSTK P1 0.407 16763

18 3 S6K p70 PSTK P1 0.473 14861

18 4 S6K p70 PSTK PO 0.484 10086

18 5 GSK3 alpha PSTK PO 0.574 22202

18 6 ? 0.678 14292

19 1 ? 0.372 9226

19 2 Rsk1 PSTK P1 0.411 19417

19 3 Rsk1 PSTK PO 0.419 20134

19 4 Calmodulin-dep. kinase 4 0.469 10954 PSTK PO

Although the foregoing invention has been described in detail by way of illustration and example for purposes of clarity of understanding, it will be readily apparent to those of skill in the art in light of the teachings of this invention that changes and modification may be made thereto without departing from the spirit or scope of the appended claims. All patents, patent applications and publications referred to herein are hereby incorporated by reference.

Claims

CLAIMS:

1. A method for detection of multiple moieties selected from kinases and kinase substrates in a test sample, comprising: i) electrophoretically separating proteins in the sample to produce a one- dimensional array of proteins so separated; ii) contacting the array with two or more antibodies selected from anti-kinase and anti-kinase substrate antibodies; and, iii) detecting the presence of antibodies bound to kinase or kinase substrate moieties in the array.

2. The method of claim 1 further comprising recording one or more values representative of location of antibodies bound on the array.

3. The method of claim 2 comprising measuring intensity of a signal representative of an amount of antibody at a location on the array and recording a value or values representative of said intensity.

4. The method of claim 1 wherein at least one protein phosphatase is added to the sample prior to separating to dephosphorylate proteins in the sample.

5. The method of claim 4 wherein said at least one protein phosphatase is inactivated prior to separating.

6. The method of claim 1 wherein at least one kinase and ATP is added to the sample to phosphorylate substrates of the added kinase prior to separating.

7. The method of claim 6 wherein the added kinase is inactivated prior to separating.

8. The method of claim 1 wherein said separating is in a SDS-PAGE gel in a single dimension.

9. The method of claim 8 wherein the proportions of acrylamide and bisacrylamide in the gel are selected to permit separation of phosphorylated and non-phosphorylated states of a single kinase or kinase substrate moiety.

10. The method of claim 9 wherein proteins electrophoretically separated exclude proteins from the sample of less than about 25 KDa.

11. The method of claim 1 wherein at least one of the antibodies comprises a polyclonal antibody mixture.

12. The method of claim 1 wherein the array is transferred to a membrane prior to contacting with antibodies.

13. The method of claim 1 wherein said detecting is by monitoring the presence of an antibody selectively bound to anti-kinase or anti-kinase substrate antibodies.

14. The method of claim 13 wherein said monitoring is of an enzymatic reaction catalyzed by a moiety conjugated to the antibody selectively bound.

15. The method of claim 2 wherein said recording comprises exposing photographic material to the array and developing said photographic material to produce a record.

16. The method of claim 2 wherein said recording comprises scanning the array to produce a record comprising a signal, said signal comprising a data structure consisting of said values.

17. The method of claim 16 further comprising storing said signal in a machine readable medium.