|Home | About | Journals | Submit | Contact Us | Français|
Mitogen-activated protein kinase (MAPK) cascades are ubiquitous signal transduction modules in eukaryotes that are of great interest and importance. Here, we summarize some useful methods for the analysis of MAPK signaling, including methods to (1) detect MAPK activation in cells, with an emphasis on using phosphorylation-state-specific antibodies raised against mammalian phosphopeptide sequences to detect the activation of MAPKs in other species; (2) estimate the cellular concentrations of MAPKs and other proteins of interest; (3) detect and quantify the stable physical association of MAPKs with their substrates and regulators, and estimate the relevant dissociation constants; (4) delineate the MAPK-binding regions or domains of MAPK-interacting proteins, with particular emphasis on the identification and verification of MAPK-docking sites. These procedures are broadly applicable to many organisms, including both yeast and mammalian cells.
Mitogen-activated protein kinase (MAPK) cascades are found in almost all eukaryotic organisms, and are expressed in virtually all tissues. MAPK cascades contribute to the regulation of diverse responses, including normal and pathological aspects of cell growth, division, differentiation, and death [1,2]. The MAPK cascade is a set of three sequentially acting protein kinases. Starting from the bottom and working back up, there is a MAPK (also termed extracellular-signal-regulated kinase, or ERK), which is phosphorylated and thereby activated by a MAPK/ERK kinase (MEK, MAPKK, MAP2K, or MKK). MEK activity is regulated, in turn, via phosphorylation by the top-most member of the module, a MEK kinase (MAP3K or MEKK) . Activated MAPKs phosphorylate targets including transcription factors, other kinases, and other enzymes [4,5]. Scaffolds proteins often assist in MAPK activation by binding to the MAPKs and other components of the cascade [6,7]. Finally, activated MAPKs are dephosphorylated and thereby deactivated by a battery of protein phosphatases [8–11].
Mammals contain four major MAPK cascades, designated the ERK1/2, ERK5, p38 and JNK cascades after their MAPK components. Most eukaryotes that have been examined to date appear to contain probable orthologs of p38 and ERK1/2. For example, in the yeast Saccharomyces cerevisiae, Hog1MAPK is orthologous to p38, and Kss1MAPK and Fus3MAPK are likely orthologs of ERK1/2 . The JNK pathway appears to have arisen after the fungal-animal split . The origin of the ERK5 pathway is uncertain at present.
MAP kinases are regulated by MEK-mediated phosphorylation of their activation loop, a region of the catalytic domain located between conserved kinase subdomains VII and VIII in the primary structure, just below the catalytic cleft in the tertiary structure . MEKs/MKKs phosphorylate both a tyrosine and a threonine in a TXY motif found in the MAPK activation loop; for this reason, MEKs are called “dual-specificity” kinases . Phosphorylation of the activation loop induces it to refold, causing subtle conformational changes which reverberate through the rest of the enzyme . For example, in the ERK2 MAPK, activation loop phosphorylation unblocks the active site and promotes a closure of the upper and lower lobes of the kinase domain that brings the catalytic residues into their correct orientation . As a result, the catalytic rate of ERK2 is increased by over 5000-fold .
MAPKs bind to many of their substrates and regulators in a relatively stable, relatively high-affinity fashion (‘relatively stable’ implies a half-life on the order of seconds to minutes, and ‘relatively high-affinity’ implies a Kd in the low micromolar range or lower). This physical association is often very important for efficient regulatory transactions such as phosphorylation, dephosphorylation or subcellular localization. Often, such interactions are mediated by MAPK-docking sites–short amino acid stretches that are located on MAPK-interacting proteins [17,18] and that bind to one or more of several characterized cognate binding regions on MAPKs [19,20].
In this paper, we summarize methods we have used successfully and repeatedly to detect MAPK activation in cells [21–24], estimate the cellular concentrations of MAPKs and other proteins of interest ; detect and quantify the stable physical association of MAPKs with their substrates and regulators [24–32], and delineate the MAPK-binding regions or domains of MAPK-interacting proteins [24–26,29–32].
The following reagents are used in the protocols described herein:
Lysis/binding buffer is 20 mM Tris–HCl, pH 7.5, 0.5 mM EDTA, 125 mM potassium acetate, 12.5% (v/v) glycerol, 0.1% (v/v) Tween 20, and 1 mM DTT (added fresh from a 1 M stock solution stored at –20 °C). This buffer can be used for extract preparation, and for both the incubation and wash steps of binding assays. A reasonable alternative buffer is phosphate-buffered saline that has been raised to 12.5% (v/v) glycerol, 0.1% (v/v) Tween 20, and 1 mM DTT.
The following additions to the above lysis/binding buffers are recommended depending on the application:
3× SDS sample buffer is 60 mM Tris–HCl, pH 6.8, 6% (w/v) SDS, 1% (w/v) DTT (Note. 1 M DTT = 15%), 20% (v/v) glycerol, and 0.001% (w/v) bromphenol blue. This buffer is stored frozen in 1 ml aliquots.
1× High pH SDS sample buffer is 60 mM Tris–HCl, pH 8.8, 2% (w/v) SDS, 0.3% (w/v) DTT, 7% (v/v) glycerol, and 0.001% (w/v) bromphenol blue. This buffer is stored frozen in 1 ml aliquots.
For detecting cross-species-reactivity with phospho-specific antibodies, we have found that polyclonal antibodies generally work better than monoclonal. An anti-phospho-ERK antibody we have had success with is “Phospho-p44/p42 MAP kinase (Thr202/Tyr204) antibody” (Catolog No. 9101) from Cell Signaling Technology. An anti-phospho-p38 antibody we have had success with is “Phospho-p38 MAP kinase (Thr180/Tyr182) antibody” (Catolog No. 9211), also from Cell Signaling Technology.
To detect MAPK activation by immunoblotting or immune-complex kinase assay, or to study MAPK protein–protein interactions by co-immunoprecipitation or co-sedimentation from cells, it is first necessary make a protein extract by lysing the cells of interest in an appropriate buffer (such as the lysis/binding buffer described in Section 2). How the cells are lysed will depend upon the specific cell type and organism, but some general principles apply:
Activation loop phosphorylation can be conveniently detected by phosphorylation state-specific antibodies . These antibodies are typically produced by inoculating rabbits with a synthetic phospho-peptide corresponding to residues around the TXY motif. Affinity purification techniques are then used to enrich for the antibodies that bind to the phosphopeptide but not to the corresponding unphosphorylated sequence. Ideally, the result is a reagent that will only bind to the phospho-epitope, and can be used to detect whether a given protein is phosphorylated on a particular residue (or residues, assuming they are closely spaced like the Tyr and Thr in the MAPK activation loop). Several anti-phospho-MAPK antibodies, raised against mammalian phosphopeptide sequences, are available from commercial suppliers. Here we describe techniques that may facilitate getting these antibodies to cross-react with MAPKs from non-vertebrate species, e.g. insects  and fungi .
As shown in Fig. 1, for both ERK1/2-like MAPKs and p38-like MAPKs, the TXY motif itself, as well as the residues C-terminal to the activation loop, are well conserved in many distantly related eukaryotes such as invertebrates and fungi. In contrast, residues N-terminal to the TXY motif are not very well conserved. Typically, the documentation provided by commercial suppliers does not state exactly which residues around the TXY were included in their antigenic peptide, which makes it difficult for those wishing to evaluate the likelihood of cross-species-reactivity. For example, Khokhlatchev et al.  state that a 16 residue peptide was used as the immunogen but do not give the sequence.
Nevertheless, we have had success using anti-phospho-ERK1/2 antibodies to detect the phosphorylation of the yeast Kss1, Fus3 and Mpk1 MAP kinases in immunoblots  (see Fig. 2), and have also been able to detect phosphorylation of the yeast Hog1 MAP kinase using anti-phospho-p38 antibodies. Based on our experience, we believe the following modifications to the “standard” immunoblot protocol are likely to maximize the chance of success of obtaining cross-species-reactivity:
Trichloroacetic acid (TCA) precipitation is useful for concentrating proteins. It can also in some cases result in a cleaner sample, as certain cellular components that degrade the quality of SDS–PAGE and immunoblot analysis will remain in the supernatant. TCA works by acid denaturation of proteins, so this method should only be used to concentrate and clean up protein samples prior to immunoblotting. Note. Protect eyes and avoid contact with skin when preparing and handling TCA solutions.
Perhaps the most reliable way to measure MAPK activation is to monitor the kinase activity by immunoprecipitating the MAPK of interest from cell extracts and assessing its ability to phosphorylate a suitable substrate in an in vitro reaction. This is a standard procedure that will not be further discussed here .
Some MAPKs exhibit a detectable shift in their electrophoretic mobility upon phosphorylation, which can be used in lieu of anti-phospho-antibodies to monitor their activation state. An excellent example is Xenopus p42MAPK .
This procedure is described elsewhere .
To estimate the cellular concentration of a given protein X requires (1) a standard containing a known concentration of protein X, and (2) a whole cell extract for which the total protein concentration has been accurately measured. A dilution series of the standard and a dilution series of the sample are then run on the same SDS–PAGE and subjected to immunoblotting, or perhaps, given a clean enough antibody, compared by ELISA. This exercise will determine the weight of protein X present in a given weight of total cell protein.
The standard is purified from bacteria or some other source, or produced by in vitro transcription and translation. This latter option requires the ability to accurately determine the yield of a protein translated in vitro. A method for this is given below.
Recently, a similar method has been used to measure cellular protein concentration on a genomic scale in yeast . This was accomplished by semi-quantitative immunoblotting of extracts prepared from over 4000 different yeast strains, each expressing a different tandem-affinity-purification (TAP)-tagged-yeast protein from the endogenous promoter. A control TAP-tagged protein was used as the standard. This method may be prone to some bias for fusion proteins that do not transfer to the immunoblot membrane as efficiently as the standard; this would lead to an underestimation of the cellular concentrations of poorly transferring proteins. Nevertheless, for Ste7MEK, Fus3MAPK and Kss1MAPK, this method gave estimates in good agreement with those obtained using native test proteins and standards produced by in vitro translation .
Suppose we desire to calculate the cellular abundance of a protein called X. We run a dilution series of our standard of X on the same gel as a dilution series of extract from the cells of interest. Suppose the lane containing 10 ng of protein X standard gives the same intensity staining pattern as the amount of protein X in 10 μg of total cell protein. This indicates that protein X constitutes 0.1% of total cell protein. It has been estimated that the average yeast cell contains 6 pg of total protein. Hence, there are 6 fg of protein X per cell. If protein X has a mass of 50 kDa, then protein X weighs ~10–19 g (50,000 g/mole/6.023 × 1023 molecules/mole = ~10–19 g/molecule). Hence the total number of molecules of protein X per cell is 60,000 (6 × 10–15 g/cell/10–19g/molecule = 6 × 104 molecules/cell = 100 zmoles/cell). For mammalian cells, perform the same calculations assuming 50 pg protein/cell. To convert to concentrations, assume a volume of 0.1 pL for a yeast cell and 0.5–1 pL for a typical mammalian cell. Thus, in yeast, 60,000 molecules/cell corresponds to a cellular concentration of approximately 1 μM.
In vitro translation is a useful method to produce protein standards for estimating cellular concentrations, and to produce radiolabeled protein for use in quantitative binding assays. Kits that enable the coupled in vitro transcription and translation of plasmid- or PCR-product-encoded proteins are available from manufacturers such as Novagen or Promega. Some of the advantages and disadvantages of this method are as follows:
In our experience, using kits based on rabbit reticulocyte lysate, typical protein yields are 20–100 fmoles/μL, that is, 20–100 nM. For an average size protein of 50 kDa, this converts into a weight per volume of 1–5 ng/μL. We typically use 1 pmole of in vitro-translated protein in a 200 μL binding reaction. This typically requires the addition of between 10 and 50 μL of translated lysate.
There are a wide variety of methods for detecting protein–protein interactions , most of which are suitable for analyzing the interactions of MAPKs with their substrates and regulators. In our work, we find ourselves often using one of the following methods:
All three methods are well established, with many published protocols. Here, we present a procedure for estimating dissociation constants from the results of binding assays, and give some advice for using co-immunoprecipitation from cell extracts to examine MAPK-mediated protein–protein interactions. Information on the yeast two-hybrid assay is available elsewhere [45,46].
The dissociation constant, or Kd, is a number that represents how strong the interaction between two ligands is: the lower the Kd, the stronger the interaction (Kd and free energy are directly interconvertible). Knowing the Kd allows an informed comparison of the relative strength of binding of different ligands. We present this method using the example of an in vitro-translated fish protein co-sedimenting with a GST fusion bait protein, but it is readily applicable to other types of binding assays, e.g. co-immunoprecipitation. The essence of the method is that dissociation constants can be calculated from the known input concentrations of both the bait and fish, and a determination of the amount of bait-fish complex formed.
For a simple bimolecular binding reaction, the relevant chemical equation is A + B AB. Here, A is the 35S-labeled protein, and B is the GST fusion protein. Then
and [A]0 and [B]0 are the input concentrations of A and B. [AB]eq is the percentage of A that co-sediments with B (readily determined if A is radiolabelled) multiplied by [A]0. This number must be corrected, if necessary, for the efficiency of sedimentation/precipitation of the bait. The bait sedimentation efficiency can be assumed to be 100% if bead-bound GST fusion protein was added to the binding reactions, but is likely to be substantially less than 100% if eluted GST fusion proteins were added to the binding reaction, and then and glutathione sepharose or an antibody/bead combination was added to promote sedimentation/precipitation.
Kd estimates should be calculated by applying the above formula to the results of multiple replicate experiments and averaging the results [25,30]. An example is shown in Table 2. As is true of most activity calculations, the resulting Kd estimate is based on the assumption was that 100% of input molecules are active (i.e. capable of binding).
Clearly, accurate determination of protein concentrations is necessary to achieve reliable estimates. GST-fusion proteins bound to gluthathione–Sepharose beads can be quantified by comparison to known amounts of BSA on a Coomassie blue-stained protein gel. Eluted GST fusions can be quantified by a standard Bradford assay. As indicated above, it is quite convenient to have the fish protein be radiolabelled (e.g. as a result of in vitro translation), in which case its concentration can be determined by the method given in Section 5.5.
Examples of GST-pull down assays involving MAPKs that were quantified using the above technique have been described [24,28–30,32]. In these studies, GST–MAPKs were used to pull down in vitro-translated MEKs or substrates, or GST-fusions to portions of MEKs, substrates or scaffold proteins were used to pull down in vitro-translated MAPKs.
Typical dissociation constants for the interaction of MAPKs with substrates and regulators is in the range of 5–200 μM [28–30]. Practically, this means that these interactions are somewhat difficult to detect by co-immunoprecipitation. Success is critically dependent on the use of a sufficiently high concentration of extract protein in the reaction. For both yeast and mammalian cell applications, a protein concentration of 5 mg/ml is recommended as a starting point (i.e. 1 mg total cell protein in a reaction volume of 200 μL); if this works, the use of less protein can be attempted during optimization. In addition, overexpression of both the MAPK and its binding partner greatly increases the chances of success. An example of a successful co-immunoprecipitation experiment can be found in Molina et al. . In this study, the ERK2 substrate micropthalmia-associated transcription factor (MITF; bait) and ERK2 (fish) were co-immunopreciptated from human kidney epithelial (HEK293) cells.
The ability of MAPKs to effectively and specifically recognize their substrates is enhanced by their capacity to bind to those substrates with relatively high affinity. Often, this interaction occurs via MAPK-docking sites on substrate proteins [17,18]. These docking sites are small regions within substrates that bind with relatively high affinity to MAPKs, in some cases showing specificity for particular MAPKs. Docking sites are also found on MAPK regulators such as MEKs, phosphatases and scaffold proteins.
One class of MAPK-docking site, designated the ‘D-site’, has the consensus (K/R)1–3–X1–5–ϕXϕ, where ϕ is a hydrophobic residue, typically Leu, Ile, Val or Met. D-sites (also called ‘D-domains’, ‘DEJL motifs’, or ‘kinase interaction motifs’) were first identified in the yeast MEK Ste7 and the transcription factors c-Jun and Elk-1 [25,47–49], and were subsequently found in numerous MAPK-interacting proteins. A second docking motif (consensus L–X1–2–(R/K)2–5), related to the D-site, is found in MAPK-activated kinases such as RSK1 . Another docking motif for MAP kinases (consensus FXFP) has been named the ‘DEF motif’ . Fig. 3 shows examples of selected MAPK docking sites culled from human and yeast proteins.
MAPK-docking sites are typically found near (within 50–100 residues of) the target phosphorylation site(s) on substrate proteins. For example, the JNK2 MAPK binds to a D-site on c-Jun located within residues 30–45, and phosphorylates c-Jun on Ser63 and Ser73 . A single docking site can enhance the phosphorylation of multiple target sites, although some substrates contain multiple MAPK docking sites, each of which may direct the MAPK to phosphorylate a specific subset of residues . Fig. 4 is a cartoon that shows one way to visualize how docking sites work.
MAPKs are proline-directed serine/threonine kinases, thus they phosphorylate the Ser or Thr in the dipeptide motif S/T-P. Leu is a preferred residue at the –1 position, and Pro is preferred at –2 (or, less so, at –3); thus, the optimal site is P-L-S/T-P . Many physiological target sites, however, do not match the ‘optimal’ consensus.
Structural and mutagenesis studies indicate that the basic submotif of the D-site contacts two closely spaced acidic patches on the MAPK surface—the ‘CD/sevenmaker’ region and the ‘ED region’ [24,54–58]. The hydrophobic submotif of the D-site binds to a ‘hydrophobic docking groove’ located close the CD/ED regions [54,56,57,59].
In contrast, the DEF motif binds to a different hydrophobic pocket (i.e. distinct from the one that the D-site binds to). This pocket is located below the MAPK active site, and is occluded in unphosphorylated ERK2 ; thus, the DEF motif preferentially binds to the dually phosphorylated, activated isoform of ERK2 .
Although many MAPK-interacting proteins contain one or more of the MAPK dockings sites described above, there are also numerous MAPK interacting proteins that apparently do not contain one of these well-recognized, commonly used MAPK docking sites. Examples include MITF , yeast Ste12, which binds preferentially to unphosphorylated Kss1MAPK , and Ets-1, which binds to ERK2 via its pointed domain . Of course, it is possible that some of these proteins will turn out to bind to the same regions of MAPKs that the more widely utilized docking sites interact with.
Presuming that a novel MAPK-interacting protein has been identified, a standard technique for localizing the particular domain or region that is responsible for MAPK binding is to produce a series of protein fragments and test them for MAPK binding. This is facilitated by making the fragments by in vitro translation. An example of such a procedure can be found in Bardwell et al.  and Molina et al. . In the former example, the use of progressively smaller N-terminal fragments of Ste7MEK eventually led the authors to identify a high-affinity D-site at the very N-terminus of Ste7 . In the latter case, it was concluded that the MAPK-binding domain of MITF constitutes approximately 100 residues and does not contain any obvious MAPK-docking motifs .
At present, the best way to identify putative MAPK docking sites in a sequence of interest is to scan the sequence by eye or with a simple pattern matching algorithm. The program Scansite  also contains probability weight matrixes for the identification of D-sites and DEF motifs.
How can a putative MAPK-docking site that has been identified in a MAPK-binding protein be verified? Some of the experimental results obtained with bona fide MAPK docking sites are:
Potentially complicating such analysis is the possibility that the MAPK-binding protein of interest may contain multiple docking sites. For example, Elk-1  and MKP-1  contain both a D-site and an FXFP/DEF site, Net contains two D-sites (one for JNK, one for ERK) as well as a putative FXFP/DEF site , and MKK7 contains three low-affinity D-sites that interact to create a relatively high-affinity MAPK-docking platform .
The binding affinity of isolated D-site peptides for their cognate MAPKs, as well the affinity of D-site-dependent binding interactions involving larger polypeptides, have been measured in several publications by several methods. The D-site of yeast Ste7 has an exceptionally high-affinity for the MAP kinases Fus3 and Kss1, with a Kd of 5–100 nM [25,29,59]. D-sites found in mammalian MAPK-binding partners typically have Kds in the low micromolar range [28,30,31], e.g. ~5 μM for MEK2-ERK2 , ~30 μM for MKK4-JNK3  (see Table 2). A particularly strong interaction is that between the D-site of the JIP-1 scaffold protein and JNK MAP kinases (apparent Kd <1 μM) . To our knowledge, the binding affinities of other classes of MAPK-docking sites have not been extensively studied. In our experience, FXFP sites do not bind nearly as well as D-sites, but we have not studied this rigorously.
Work in the authors’ laboratory is supported by research Grants GM60366 and GM69013 from the US National Institute of General Medical Science.