PHYSIOLOGICAL REVIEWS   Vol. 79 No. 1 January 1999, pp. 143-180
Copyright ©1999 by the American Physiological Society

Mitogen-Activated Protein Kinase: Conservation of a Three-Kinase Module From Yeast to Human

CHRISTIAN WIDMANN, SPENCER GIBSON, MATTHEW B. JARPE, AND GARY L. JOHNSON

Program in Molecular Signal Transduction, Division of Basic Sciences, National Jewish Medical and Research Center, Denver, Colorado

I. INTRODUCTION: DISCOVERY THAT MITOGEN-ACTIVATED PROTEIN KINASE^ERK IS REGULATED BY PHOSPHORYLATION ON THREONINE AND TYROSINE
II. MITOGEN-ACTIVATED PROTEIN KINASE PATHWAYS: ACTIVATION MODULES INVOLVING THREE KINASES
III. ACTIVATION OF MITOGEN-ACTIVATED PROTEIN KINASES: DUAL PHOSPHORYLATION OF THE ACTIVATION LOOP
IV. YEAST MITOGEN-ACTIVATED PROTEIN KINASE PATHWAYS
    A. Organization of MAPK Modules
    B. Role of MAPK Modules in Yeast
    C. Mating Pathways in Haploid Yeast
    D. Invasive Pathways: Pseudohyphal Pathway of Diploids and Invasive Growth of Haploids
    E. Cell Wall Remodeling Pathway
    F. Osmosensor and Stress Pathways
    G. Sporulation Pathway
V. MAMMALIAN MITOGEN-ACTIVATED PROTEIN KINASE PATHWAYS
    A. MAPK^erk Pathway
    B. MAPK^jnk Pathway
    C. MAPK^p38 Pathway
    D. MKK5/MAPK^erk Pathway
VI. MITOGEN-ACTIVATED PROTEIN KINASE PATHWAYS IN OTHER ORGANISMS
    A. Dictyostelium discoideum
    B. Caenorhabditis elegans
    C. Drosophila melanogaster
    D. Xenopus laevis
    E. MAPK Pathways in Plants
VII. CONCLUDING REMARKS
REFERENCES

	ABSTRACT

Top References

Widmann, Christian, Spencer Gibson, Matthew B. Jarpe, and Gary L. Johnson. Mitogen-Activated Protein Kinase: Conservation of a Three-Kinase Module From Yeast to Human. Physiol. Rev. 79: 143-180, 1999.  Mitogen-activated protein kinases (MAPK) are serine-threonine protein kinases that are activated by diverse stimuli ranging from cytokines, growth factors, neurotransmitters, hormones, cellular stress, and cell adherence. Mitogen-activated protein kinases are expressed in all eukaryotic cells. The basic assembly of MAPK pathways is a three-component module conserved from yeast to humans. The MAPK module includes three kinases that establish a sequential activation pathway comprising a MAPK kinase kinase (MKKK), MAPK kinase (MKK), and MAPK. Currently, there have been 14 MKKK, 7 MKK, and 12 MAPK identified in mammalian cells. The mammalian MAPK can be subdivided into five families: MAPK^erk1/2, MAPK^p38, MAPK^jnk, MAPK^erk3/4, and MAPK^erk5. Each MAPK family has distinct biological functions. In Saccharomyces cerevisiae, there are five MAPK pathways involved in mating, cell wall remodelling, nutrient deprivation, and responses to stress stimuli such as osmolarity changes. Component members of the yeast pathways have conserved counterparts in mammalian cells. The number of different MKKK in MAPK modules allows for the diversity of inputs capable of activating MAPK pathways. In this review, we define all known MAPK module kinases from yeast to humans, what is known about their regulation, defined MAPK substrates, and the function of MAPK in cell physiology.

	I. INTRODUCTION: DISCOVERY THAT MITOGEN-ACTIVATED PROTEIN KINASEERK IS REGULATED BY PHOSPHORYLATION ON THREONINE AND TYROSINE

Top Next References

In the early 1980s, it was realized in several different cell types stimulated with growth factors including platelet-derived growth factor (PDGF) and epidermal growth factor (EGF) that a predominant protein phosphorylated on tyrosine had a size of 42 kDa (59, 61). Proteins of the same size and behavior on two-dimensional gels were phosphorylated on tyrosine in response to phorbol esters (174), in virus-transformed cells (60), and in metaphase-arrested Xenopus laevis eggs (57). At the same time, serine-threonine protein kinases activated by the insulin receptor tyrosine kinase were being characterized (7, 287, 294). The hypothesis was considered that direct regulation by tyrosine phosphorylation catalyzed by the insulin receptor would regulate the activity of serine-threonine protein kinases (7, 287). Ray and Sturgill (287) demonstrated that a 42-kDa serine-threonine protein kinase, referred to as mitogen-activated protein kinase (MAPK), isolated from insulin-stimulated 3T3-L1 cells was phosphorylated on both threonine and tyrosine. It was quickly realized that the 42-kDa MAPK characterized to be threonine and tyrosine phosphorylated in response to insulin was the same protein shown to be tyrosine phosphorylated in response to other growth factors, phorbol esters, viral transformation, and metaphase arrest in Xenopus eggs (294). It was also demonstrated that phosphorylation of both the threonine and tyrosine was required for MAPK activation (294).

The cDNA encoding MAPK was isolated by Boulton et al. (32) who renamed it ERK1 for extracellular signal-regulated kinase 1, because of the variety of extracellular signals that could stimulate its activity. The isolation of cDNA for ERK2 and ERK3 quickly followed (31). Alignment of the ERK sequences showed they were closely related to the Saccharomyces cerevisiae protein kinases Fus3 and Kss1, demonstrating the close homology between mammalian and yeast MAPK.

Biochemical characterization and molecular cloning identified the upstream kinases in the MAPK^erk module in mammalian cells (67). The work defined a conserved three-kinase module in mammals. The MAPK regulatory system is a three-kinase module that establishes a sequential protein kinase activation pathway.

In parallel, geneticists were identifying the component genes in yeast mating. The sterile (ste) genes in S. cerevisiae involved in pheromone-induced mating were ordered and cloned (136). Quickly, the related genes in Schizosaccharomyces pombe were identified and their protein products characterized (260). Cumulatively, this work identified a conserved three-component protein kinase module that included the MAPK Fus3 and Kss1 in S. cerevisiae and Spk1 in S. pombe.

In this review, we detail the properties of the known MAPK modules. The expanding number of MAPK modules and their role in controlling complex cellular functions defines their importance in responsiveness of cells and organisms to their environment.

	II. MITOGEN-ACTIVATED PROTEIN KINASE PATHWAYS: ACTIVATION MODULES INVOLVING THREE KINASES

Top Previous Next References

Pathways involving MAPK are activated in response to an extraordinarily diverse array of stimuli. These stimuli vary from growth factors and cytokines to irradiation, osmolarity, and the shear stress of fluid flowing over a cell. The basic assembly of MAPK pathways is a three-component module conserved from yeast to humans. The minimal MAPK module is composed of three kinases that establish a sequential activation pathway (Fig. 1). The first kinase of the three-component activation module is a MAPK kinase kinase (MKKK) (92). Specific MKKK have been shown to be activated either by phosphorylation by a MAPK kinase kinase kinase (MKKKK) or by interaction with a small GTP-binding protein of the Ras or Rho family. Other potential modes of activation include oligomerization and subcellular relocalization. The MKKK are serine/threonine kinases that when activated phosphorylate and activate the next kinase in the module, a MAPK kinase (MKK) (315). The MKK are kinases that recognize and phosphorylate a Thr-X-Tyr motif in the activation loop of MAPK (109), defining MKK as dual-specificity kinases. Mitogen-activated protein kinases are the final kinase in the three-kinase module and phosphorylate substrates on serine and threonine residues. The vast majority of defined substrates for MAPK are transcription factors. However, MAPK have the ability to phosphorylate many other substrates including other protein kinases, phospholipases, and cytoskeleton-associated proteins.

View larger version (44K): [in this window] [in a new window]   FIG. 1.   Core module of a mitogen-activated protein kinase (MAPK) pathway is composed of 3 kinases [MAPK kinase kinase (MKKK), MAPK kinase (MKK), and MAPK] that are sequentially activated by phosphorylation.

Why have MAPK activation modules evolved having three kinases? The reason probably lies in the unique activation properties of MAPK. Mitogen-activated protein kinase must be phosphorylated on both a threonine and tyrosine for their activation, a dual phosphorylation catalyzed by a specific MKK. The different MKK recognize a tertiary structure of specific MAPK and not simply a linear sequence surrounding the Thr-X-Tyr activation motif of MAPK. As described in this review, very specific MKK and MAPK combinations are found in a MAPK module. Specific MKK appear to recognize the tertiary structure of different MAPK, effectively restricting their regulation of different MAPK subtypes. In contrast, MKKK are able to mix and match with different MKK-MAPK combinations. In mammalian cells, there are more known MKKK than MAPK. The MKK represent the fewest members of the three-component MAPK modules. The completion of the human genome project will be required to define the exact number of kinases in each of these groups. The large number of MKKK allows for diversity of inputs from numerous stimuli to feed into specific MAPK pathways. Some kinases that appear to be MKKK may regulate pathways not involving MAPK [such as regulation of the NFB pathway by MEKK1 (139, 196, 233, 393)]. Thus the regulation of MAPK pathways at the level of MKKK may represent branch points in regulation of signal pathways in some cases.

Table 1 lists the members of the known MAPK three-component modules defined to date. Perusal of Table 1 boggles the mind and clearly demonstrates that the current nomenclature, or more appropriately the lack of nomenclature, in the MAPK field poses a major problem of understanding the MAPK literature, especially for readers not familiar with the field. In the present review, we decided to name a given component of a MAPK module according to its position in the pathway with the current name of the protein in superscript. For example, the yeast Ste11 MAPK kinase kinase will be described as MKKK^ste11. This notation should greatly facilitate the reading of this review, since it is not required to know the identity of each kinase to determine its position in a given MAPK pathway.

To date, 14 MKKK, 7 MKK, and 12 MAPK have been identified in mammalian cells (Table 1). Dendogram analysis indicates that these kinases belong to different subfamilies (Fig. 2). Four subfamilies among the MKKK can currently be defined. The Raf subfamily is the best characterized and comprises MKKK^B-raf, MKKK^A-raf, and MKKK^raf1. The MEK kinase (MEKK) subfamilly is made of the four MEKK (MKKK^mekk1-4). MKKK^ask1 and MKKK^tpl2 appear to form a third MKKK subfamily. The fourth group in the dendogram is more diverse and comprises MKKK^mst, MKKK^sprk, MKKK^muk, MKKK^tak1, and the most distantly related MKKK, MKKK^mos. For the MKK, MKK^mek1 and MKK^mek2 are closely related as are MKK^mkk3 and MKK^mkk6. The MAPK can be categorized into five subfamilies: the MAPK^erk1/2, the MAPK^p38, the MAPK^jnk, the MAPK^erk5, and the MAPK^erk3/4 subfamilies. These MAPK give the name to the MAPK pathways that employ them (i.e., the MAPK pathways using the MAPK^jnk are called the JNK MAPK pathways). This review focuses on the organization, regulation, and function of the different MAPK modules in eukaryotic cells.

View larger version (14K): [in this window] [in a new window]   FIG. 2.   Phylogenetic trees of mammalian MAPK, MKK, and MKKK family members. Dendrograms were created with CLUSTAL X program (334) using human sequences when available or mouse sequences otherwise (see Table 1).

	III. ACTIVATION OF MITOGEN-ACTIVATED PROTEIN KINASES: DUAL PHOSPHORYLATION OF THE ACTIVATION LOOP

Top Previous Next References

Mitogen-activated protein kinases are proline directed in that they only phosphorylate substrates that contain a proline in the P-1 site (219). A general consensus for MAPK^erk1/2 is Pro-X-Ser/Thr-Pro (6). The activity of MAPK is controlled by dual phosphorylation in an amino acid sequence known as the activation loop. The sequence Thr-X-Tyr in the activation loop, where X can be different amino acids among the MAPK, is the site for dual phosphorylation catalyzed by specific MKK (4). For MAPK^erk1/2, the phosphorylation sites correspond to Thr-183 and Tyr-185. Dual phosphorylation of these sites results in a >1,000-fold increase in specific activity of the MAPK.

The core three-dimensional structure of protein kinases, as resolved from the crystal structure of protein kinase A (PKA), is composed of two domains with the active site at the domain interface (177). Adenosine 5'-triphosphate binds in the active site cleft, and peptide substrate for some kinases has been shown to bind in a groove on the surface of the COOH-terminal domain of the active site (178). The phosphoreceptor amino acid (Ser, Thr, Tyr) of the substrate binds near the catalytic loop in this domain. A surface loop contiguous with this COOH-terminal domain is found in many protein kinases and encodes a phosphorylation site; this sequence is referred to as the activation loop or lip. Threonine-183 in MAPK^erk1/2 is homologous to the phosphorylation site in the activation loop of other protein kinases (165). The tyrosine phosphorylation site in the activation loop of MAPK is unique.

The recent crystal structure of active, dual-phosphorylated MAPK^erk2 has been recently solved (40). Phosphorylation of Thr-183 and Tyr-185 in the activation loop causes the loop to refold and interact with surface arginine-binding sites. The conformational changes in the activation loop and neighboring sequences result in activation of the kinase.

An interesting property of the requirement for dual phosphorylation of the Thr-X-Tyr sequence is that substitution of the Thr and Tyr by acidic amino acids like glutamate do not result in constitutive activation. The characteristics of the activation loop and the dual phosphorylation of the Thr and Tyr induce several conformational changes in the protein that cannot be mimicked by glutamates. For this reason, point mutations in MAPK do not activate their activity. The requirement for the P-Thr-X-P-Tyr regulation of MAPK activity has been proposed as why MAPK have never been identified as oncogenes (40), as could be expected from proteins regulating cell growth.

	IV. YEAST MITOGEN-ACTIVATED PROTEIN KINASE PATHWAYS

Top Previous Next References

Yeast probably represents the experimental model where the organization and regulation of MAPK pathways are best understood. Presently, five MAPK pathways have been well characterized in S. cerevisiae: the haploid mating pathway, invasive growth, cell wall remodeling, and two pathways involved in stress responses such as hyperosmolarity. These five MAPK pathways are discussed in detail in section IV. An important concept on MAPK pathways that has emerged from yeast studies is that the kinases employed in MAPK pathways are organized into modules. As discussed in section IVA, this is achieved by tethering to scaffold proteins as well as by direct interaction between the different kinases of the module. Organization into modules ensures segregation of the pathways from other signaling events in the cells and also allows the use of a given component kinase in more than one MAPK module without affecting the specificity of the response mediated by the MAPK pathways.

Genetic and biochemical studies in yeast have demonstrated that specific MKKK-MKK-MAPK signaling modules are functionally segregated from one another (98). In the budding yeast, S. cerevisiae, the three component MAPK modules are assembled either by tethering to scaffolding proteins such as Ste5 or by binding domains encoded in the kinases themselves. Ste5 is a 110-kDa dimeric protein that is capable of binding the three-kinase components of the pheromone-regulated mating MAPK pathway, MKKK^ste11, MKK^ste7, and MAPK^fus3. This ensures that this particular module is coordinately regulated and segregated from other signaling pathways. Tethering and segregation of MAPK modules may in fact be vital for MAPK pathways to function properly (217). Mitogen-activated protein kinase module kinases may have high affinity for each other in the absence of a tethering protein like Ste5. For example, MKKK^fus3 and MKK^ste7 have an ~5 nM dissociation constant binding affinity, indicating they could have significant interaction in the the absence of Ste5 (13). Similarly, MKK^pbs2 has the ability to bind several different MKKK involved in osmoregulation (see sect. IVF). The complexity of regulation of MAPK pathways is further exemplified by the fact that some kinases participate in more than one MAPK module. For example, MKKK^ste11 functions in the MAPK pathways for mating, pseudohyphal development, and osmoregulation. However, because of differential upstream inputs from MKKKK, GTP-binding proteins, and the assembly of three kinases into modules, there is generally little if any cross talk between the different MAPK pathways in S. cerevisiae.

Yeast may exist either as haploid or diploid cells. The haploid cells have two sexual phenotypes characterized by the expression of a set of genes involved in mating that are not expressed in diploids. The mating response to generate diploids is controlled by the a and -pheromones that bind to their respective receptors that are coupled to a heterotrimeric G protein. Pheromone binding to its receptor leads to G protein activation and the dissociation of the -subunit complex from -GTP. In budding yeast, the -complex binds to Ste5 and by a process still poorly understood stimulates activation of the mating MAPK pathway. In addition to mating, yeast respond to their environment with metabolic changes that involve MAPK pathways. For example, in response to starvation, yeast undergo dramatic morphological changes involving pseudohyphal formation and invasiveness in an attempt to find nutrients that are controlled in part by a MAPK pathway. Yeast adapt to additional adverse environmental conditions such as high or low osmolarity by activation of specific MAPK pathways. Thus many aspects of yeast physiology controlling their response to environmental cues are, at least in part, regulated by MAPK pathways (136). Knowledge of the entire genome sequence for the budding yeast S. cerevisiae has defined the number of protein kinases in the organism (147). Relative to MAPK modules, it appears that the S. cerevisiae genome encodes four MKKK, four MKK, and six MAPK (Table 1). Among the six different MAPK present in S. cerevisiae (147), only two cannot be attributed to one of the five well-characterized MAPK pathways: MAPK^smk1, which is involved in spore wall assembly, and the putative MAPK^YKL161C identified by genome sequencing. MAPK^smk1 and MAPK^YKL161C may be involved in as yet undefined MAPK pathways. MAPK^YKL161C has a K-X-Y motif in its activation loop distinguishing it from other MAPK.

The best-defined yeast MAPK pathway in S. cerevisae is involved in the mating of haploid cells (21, 58, 89, 305) (Fig. 3). The two haploid cell types (a and ) of S. cerevisiae, upon binding the sexual pheromone secreted by the opposite cell type (a-factor and -factor), stop growing and differentiate into mating-competent cells by inducing transcription of mating genes. The S. cerevisiae genes whose disruption inhibited mating and caused sterility were designated as sterile (ste) genes. The seven transmembrane receptors for the - and a-factors are designated as Ste2 and Ste3, respectively, and are coupled to a heterotrimeric G protein. Pheromone activation of the G protein induces the dissociation of the heterotrimeric G protein subunits designated Gpa1 (-subunit), Ste4 (-subunit), and Ste18 (-subunit). The released -subunit complex (Ste4/Ste18) activates Ste20 and interacts with the scaffolding protein Ste5, resulting in the stimulation of the MAPK module MKKK^ste11/MKK^ste7/MAPK^fus3. It should be noted that involvement of Ste20 in the activation of the mating MAPK pathway is still debated, since it has not yet been unequivocally proven that Ste20 phosphorylates and activates MKKK^ste11. Activated MAPK^fus3 regulates the activity of transcription factors required for the expression of components of the mating pathway itself and genes necessary for cell cycle arrest and cell fusion (136).

View larger version (43K): [in this window] [in a new window]   FIG. 3.   Saccharomyces cerevisiae MAPK pathway used for mating. Note that Ste5 is a scaffolding protein that is able to bind to Ste4, MKKKste11, MKKste7, and MAPKfus3.

It was initially thought that MAPK^fus3 and MAPK^kss1 were redundant kinases in the mating response, because genetic experiments indicated that expression of one protein could complement the mating defect induced by the absence of the other. This notion was widely accepted in the MAPK field even though a number of observations indicated that MAPK^fus3 and MAPK^kss1 were differently regulated and were in fact components of different signaling pathways. First, MAPK^fus3 and MAPK^kss1 are not activated by the same stimuli. MAPK^fus3 activity is very low, whereas MAPK^kss1 exhibits high activity, in the absence of mating pheromones. Conversely, in presence of pheromones, MAPK^fus3 is strongly activated but MAPK^kss1 only weakly if at all (13, 85). Second, MAPK^fus3 is expressed only in haploid cells, whereas MAPK^kss1 is expressed both in haploid and diploid cells, indicating that these two MAPK have different functions. Third, MAPK^fus3 and MAPK^kss1 have different substrate targets (259). With regard to substrate recognition, it has been shown that MAPK^fus3, but not MAPK^kss1, phosphorylates Far1 (85, 275) that represses the transcription of the G₁ cyclins Cln1 and Cln2 (83), leading to cell cycle arrest that is a feature associated with mating. Cumulatively, these studies suggested that MAPK^fus3 is the MAPK involved in mating, whereas MAPK^kss1 is not. This was confirmed recently with experiments showing that the expression of a kinase inactive mutant of MAPK^fus3 in a MAPK^fus3 null background hampered the ability of MAPK^kss1 to fulfill mating MAPK functions (217). The findings indicated that MAPK^kss1 cannot activate the mating MAPK pathway in wild-type cells because it is excluded from this pathway by MAPK^fus3. Only in null mutants lacking MAPK^fus3 can MAPK^kss1 substitute for MAPK^fus3 in the mating pathway, explaining the earlier results that suggested redundancy between the two MAPK. Thus MAPK^fus3 has an additional function that is to segregate the mating MAPK pathway from imposter proteins, in this case MAPK^kss1, that are normally involved in other signaling pathways. Segregation of the MAPK module involved in mating is further achieved by Ste5, the scaffold molecule which has no apparent enzymatic activity itself. Different domains of Ste5 bind, in the two-hybrid system, to MKKK^ste11 and MAPK^fus3 (53, 151, 184, 226, 280) as well as to Ste4 (G protein -subunit) (365) and is predicted to couple the -complex (Ste4/Ste18) to activation of the MAPK mating pathway (Fig. 3).

The NH₂-terminal portion of Ste5 (residues 177-229) contains a cysteine-rich region that is the prototype for the RING-H2 motif. Proteins of the zinc RING family possess two fingerlike domains connected by a linking region and require zinc for folding (18, 29). Crystal structures indicate that RING domains are globular pseudosymmetric folds that coordinate two zinc atoms through a cross-bridging element (14, 28). When this structure is mutated, the corresponding Ste5 mutants are unable to complement the mating defect of yeast strains lacking Ste5 (ste5 cells) (152). The RING-H2 mutants, while being able to bind MKKK^ste11, MKK^ste7, and MAPK^fus3 as efficiently as the wild-type protein, were not capable of binding Ste4. Moreover, the RING-H2 mutants could not dimerize (152). When the Schistosoma japonicum glutathione S-transferase protein, known to form stable dimers (230, 349), was fused to the COOH terminus of the RING-H2 Ste5 mutants, the resulting protein could complement the mating defect of ste5 cells but could not associate with Ste4. Oligomerization of Ste5 is, however, not sufficient to transduce the mating signal, since wild-type Ste5 fused to the S. japonicum glutathione S-transferase was able to complement ste5 cells but not ste4 ste5 cells (152). The conclusion of these studies is that the RING-H2 domain has inhibitory functions that can be alleviated by Ste4. Once the inhibitory role of the RING-H2 domain is suppressed, it oligomerizes, leading to the activation of the MAPK mating module, possibly by allowing Ste20 to phosphorylate MKKK^ste11. It is presently unclear if MKKK^ste11 phosphorylation is the primary regulatory event that controls its activity, even though it is a substrate for Ste20.

Stimulation of the MAPK mating pathway leads to the activation or repression of the activity of a variety of proteins. For example, Far1 phosphorylated by MAPK^fus3 binds to and inactivates the cell division control (Cdc)28/Cln kinase complex and thus inhibits cell growth. Phosphorylation of Far1 is critical, since a mutant allele of Far1 that is not fully phosphorylated is unable to mediate pheromone-induced cell cycle arrest (275). Ste12, a transcription factor mediating the induction of pheromone-response genes, is also phosphorylated by MAPK^fus3 (85). One of the functions of Ste12 is to induce Far1 transcription (266). Thus MAPK^fus3 regulates Far1 in two complementary ways: a direct phosphorylation-dependent activation and an increase in transcription through activation of Ste12.

Yeast cells that are ready to undergo mating must make sure that they do not grow, and this is achieved in part by the MAPK^fus3 phosphorylation and activation of Far1. Similarly, cells undergoing cell division should not be responsive to pheromones. Interestingly, the G1 cyclins that are inhibited by Far1 in mating competent cells are themselves inhibitors of the mating pathways in dividing cells. It has indeed been shown that Cln2 represses the mating MAPK pathway, and this inhibition takes place at the level of MKKK^ste11 (362). There are other examples demonstrating the reciprocal actions of the mating MAPK pathway and cyclin-dependent kinases. For example, the rat glucocorticoid receptor ectopically expressed in S. cerevisiae is phosphorylated by cyclin-dependent kinases and MAPK at distinct sites. Glucocorticoid receptor-dependent transcriptional enhancement is reduced in a yeast strain deficient in cyclin-dependent kinases but is increased in a strain devoid of MAPK^fus3 (186).

In the fission yeast S. pombe, mating involves a MAPK module containing homologs of the S. cerevisiae MAPK mating pathway (MKKK^byr2 is homologous to MKKK^ste11, MKK^byr1 to MKK^ste7, and MAPK^spk1 to MAPK^fus3) (89) (Fig. 4). The S. pombe and S. cerevisiae mating pathways are controlled by different upstream regulators. First, in the budding yeast, the G complex transmits the signal, whereas in S. pombe, it is the G subunit that transmits the signal. Second, although haploid budding yeast express the genes required for the pheromone response in all nutritional conditions, differentiation of fission yeast into mating competent cells is strictly dependent on nutritional starvation. The S. pombe Ras homolog Ras1 plays a role in the starvation-dependent control of the mating pathway, possibly in a G-independent manner (384). In response to mating pheromones, mutants defective in Ras1 or in Ste6, the fission yeast homolog of the Ras GDP/GTP exchange factor, are unable to induce transcription of the mat1-Pm gene that controls entry into meiosis (262). Ras1 may be directly linked to the S. pombe MAPK module involved in mating as suggested by the observation that Ras1 can bind MKKK^byr2 in a two-hybrid assay (48, 344). Interestingly, a Ste5 equivalent has not been defined in S. pombe, nor is there evidence that phosphorylation of MKKK^byr2 by a Ste20 homolog is required. Current evidence argues that GTP-binding proteins regulate MKKK^byr2 activity.

View larger version (38K): [in this window] [in a new window]   FIG. 4.   Schizosaccharomyces pombe MAPK pathway used for mating.

In response to nitrogen starvation, diploid S. cerevisiae activate an intracellular pathway that will eventually lead to profound morphological changes. When starved for nitrogen, the elliptical diploid yeast undergo an asymmetric cell division to produce a long thin daughter cell that will keep producing long daughter cells. Because the mother and daughter cells remain attached, reiteration of this unipolar division pattern produces filaments composed of a linear chain of elongated cells called a pseudohypha. The mother yeast produce colonies on the surface of an agar plate, whereas the pseudohypha invades the agar (112, 113). Haploid cells can also invade the agar and grow beneath the surface. The response of haploids is not as dependent on nutritional starvation as the pseudohyphal development of diploid cells. In particular, nitrogen limitation seems to play no role in the invasive growth response of haploids (293). A MAPK module consisting of MKKK^ste11, MKK^ste7, and MAPK^kss1 can transduce the signals leading to invasive growth (Fig. 5). Mitogen-activated protein kinase^kss1 was initially thought to play no role in invasiveness because diploid mutants lacking MAPK^kss1 had a normal pseudohyphal response (136). However, with the use of strains bearing hypermorphic alleles of MKKK^ste11 and MKK^ste7 to increase filamentous growth, it was shown that the absence of MAPK^kss1 blocked hyperfilamentation, as well as the enhancement of the activity of promoters containing filamentation and invasion response elements. The kinase activity of MAPK^kss1 is required for filamentous growth, both in haploids and diploids, because removal of MAPKK^ste7, the upstream regulator of MAPK^kss1, or point mutations impairing the catalytic activity of MAPK^kss1 result in severe filamentation defects (121, 217). If MAPK^kss1 is required for a normal invasive response, how is it that MAPK^kss1 knockout mutants have no pseudohyphal defect? The postulated answer is that MAPK^kss1 also has an inhibitory function in the invasive response and that there are MAPK-independent signaling pathways mediating invasiveness (121, 217). In the absence of MAPK^kss1, the MAPK-independent pathways can stimulate the invasive growth because they are not hampered by the inhibitory function of MAPK^kss1. It has been possible to genetically separate the inhibitory and stimulatory functions of MAPK^kss1 on pseudohyphal development. There are mutants of MAPK^kss1 that have lost their inhibitory function that when expressed lead to hyperfilamentation in a MAPK^kss1 null background (217). In contrast, in the situation where MAPK^kss1 cannot be activated (i.e., when MAPKK^ste7 is absent), filamentation is drastically reduced because MAPK^kss1, while unable to activate filamentation since this requires the stimulation of its catalytic activity, retains its negative function (121, 217). This also shows that the inhibitory function of MAPK^kss1 does not require its catalytic activity. When not activated by MKK^ste7, MAPK^kss1 is thus an inhibitor or silencer of the invasive growth response. In contrast, when MAPK^kss1 is activated by MKK^ste7, it stimulates invasive growth, and this response depends on an intact kinase activity. The negative function of MAPK^kss1 is mediated by its interaction with its target Ste12. MAPK^kss1 interacts with Ste12 predominantly in its inactive state. This interaction inhibits the activity of Ste12. Once MAPK^kss1 is activated, it no longer binds and inhibits Ste12; rather, active MAPK^kss1 phosphorylates and activates Ste12, leading to activation of genes under the control of promoters containing filamentation and invasion response elements (217).

View larger version (52K): [in this window] [in a new window]   FIG. 5.   Saccharomyces cerevisiae MAPK pathway used during invasive responses.

Invasiveness is more pronounced in the absence of MAPK^fus3 (either in mutant haploids cells or in diploids that do not express MAPK^fus3) (121, 293), indicating that MAPK^fus3 functions as a repressor of the invasive growth response. Thus, when activated, MAPK^fus3 and MAPK^kss1 have opposite functions in invasive growth. Possibly, one role of the MAPK pathway employing MAPK^fus3 is to suppress invasive growth in haploid cells to facilitate their entry into the mating program where the cells must stop growing. In such a scenario, in the presence of pheromones, the inhibitory activity of MAPK^fus3 on invasive growth would be greater than the stimulatory activity of MAPK^kss1. This is consistent with the fact that pheromones activate MAPK^fus3 much more strongly than MAPK^kss1 (13, 85). Clearly, the balance of these kinase activities, both regulated by MKK^ste7, will play a role in determining whether a cell will exhibit an invasive response.

The exact role of the MKKK^ste11/MKK^ste7/MAPK^kss1 module in the control of invasive growth is still incompletely understood. The need for a better understanding of invasive growth becomes obvious when it is considered that removal of all the components of the MKKK^ste11/MKK^ste7/MAPK^kss1 module does not markedly affect the invasive growth response (121). Thus, even though genetic studies show invasive growth can be strongly affected by mutants in the MKKK^ste11/MKK^ste7/MAPK^kss1 module, the pathway itself is not required. This shows that filamentation and invasion response elements are under the control of alternative regulatory pathways in the absence of a functional MKKK^ste11/MKK^ste7/MAPK^kss1 pathway. Such pathways may involve Ras, since it has been shown that activated forms of Ras potentitate the invasive growth response (113, 121, 250).

In yeast, growth depends on efficient cell wall remodeling, and this response also uses a MAPK pathway (Fig. 6). In the budding yeast, the MAPK module is composed of MKKK^bck1, MKK^mkk1 or MKK^mkk2, and MAPK^mpk1. The biological relevance of having two seemingly redundant MKK in the cell wall remodeling pathway is unclear. The upstream regulator of the MAPK module of the cell wall remodeling pathway is the yeast homolog of mammalian protein kinase C (PKC), PKC1, which could function as a MKKKK for this pathway although it has not been proven biochemically that PKC1 can directly phosphorylate and activate MKKK^bck1. Mutants lacking PKC1, MKKK^bck1, or MAPK^mpk1 lyse under isotonic conditions because of a deficiency in cell wall construction (154, 199, 206). However, the phenotypes of the mutants defective in PKC1 are more severe than in the mutants defective in the downstream components of the pathway; PKC1 mutants have a lysis defect at all temperatures (200). This suggests that PKC1 controls two pathways required for an optimal cell-wall remodeling response, one of which contains the MKKK^bck1/MKK^mkk1 or MKK^mkk2/MAPK^mpk1 module.

View larger version (41K): [in this window] [in a new window]   FIG. 6.   Saccharomyces cerevisiae MAPK pathway used for cell wall construction. PKC, protein kinase C.

In the fission yeast S. pombe, MAPK^pmk1 may be part of a MAPK pathway involved in maintenance of cell wall integrity (337, 395). However, MAPK^pmk1 may not be a downstream target of PKC in S. pombe but may function in coordination with PKC to regulate cell integrity (337). The upstream regulators of MAPK^pmk1 remain to be characterized in fission yeast.

The budding yeast S. cerevisiae activates two pathways in response to hyperosmolarity (220) (Figs. 7 and 8). The osmosensors that stimulate these pathways function in a very different manner. One pathway is regulated by a set of three proteins functioning like the prokaryotic two-component transduction system, and the other osmosensor is an integral membrane protein. These two sensors regulate two different MAPK pathways utilizing common kinase elements that will lead to the transcription of genes necessary for survival in hyperosmotic conditions such as those required for the synthesis of glycerol to increase the internal osmolarity (136). Although external high osmolarity will activate both of these MAPK pathways, the corresponding osmosensors do not detect the same osmolarity changes. The integral membrane osmosensor is activated by high osmolarity and stimulates its corresponding MAPK pathway. In contrast, the two-component osmosensor is active in low-osmolarity conditions, and this results in the repression of the associated MAPK module. In high-osmolarity conditions, however, this osmosensor is inactivated and no longer inhibits the activation of the MAPK. Apparently, fission yeast also employ two related MAPK pathways to sense changes in osmolarity. Two different MAPK pathways are thus activated in response to hyperosmotic conditions in both S. cerevisiae and S. pombe. The reason for this is not yet understood, but it is likely that each of the osmosensor MAPK pathways do not fulfill exactly the same function in response to increased osmolarity and that they may be differentially involved in sensing and responding to other types of stress.

View larger version (51K): [in this window] [in a new window]   FIG. 7.   "Two-component" osmosensor MAPK pathway of Saccharomyces cerevisiae. Solid and dotted boxes correspond to active and inactive state of proteins, respectively.

View larger version (50K): [in this window] [in a new window]   FIG. 8.   Sho1-dependent osmosensor MAPK pathway in Saccharomyces cerevisiae. MKKpbs2 functions both as a MKK and as a scaffold protein that binds to Sho1, MKKste11, and MAPKhog1.

View larger version (54K): [in this window] [in a new window]   FIG. 9.   Stress-induced MAPK pathways in Schizosaccharomyces pombe.

Yeast sporulation is the process, involving meiosis, that leads to the packaging of haploid nuclei into spores. This reponse occurs only in diploids and is elicited by nutritional starvation (241). Upon completion of meiosis, the four-haploid nuclei, which still remain within a single nuclear membrane, are enveloped by the double membraneous prospore wall. The spore wall is then deposited from the space between the layers of the prospore wall. The final differentiated spore wall consists of four layers. The two inner layers appear indistinguishable from the vegetative cell wall, whereas the third layer is a spore-specific structure composed primarily of chitosan and chitin. The outermost electron-dense layer is a dityrosine coat (36, 37). One gene encoding a MAPK, MAPK^smk1, has been shown to be involved in the sporulation pathway (MAPK^smk1 mutants are defective in spore wall assembly) (185). A putative MAPK module, which employs MAPK^smk1 and an uncharacterized MKKK and MKK, may thus be used to regulate the sporulation response. A MKKKK homolog (Sps1) has also been shown to be involved in the sporulation pathway. Both Sps1 and MAPK^smk1 mutants display similar phenotypes: they both proceed normally through meiosis but then are defective in spore wall assembly (136, 185). This suggests that Sps1 controls the putative MAPK module mediating the sporulation response.

	V. MAMMALIAN MITOGEN-ACTIVATED PROTEIN KINASE PATHWAYS

Top Previous Next References

One important point that emerges from the the studies of the yeast MAPK pathways is that MAPK pathways form modules that are held together through protein-protein interactions. Activation of one MAPK pathway does not normally lead to the activation of other MAPK pathways, even if a given component is found within multiple pathways. Mitogen-activated protein kinase pathways are thus spatially regulated in yeast. It is thus predicted that this also applies to the mammalian MAPK pathways. In mammalian cells, however, multiple MAPK pathways can be activated by a single receptor type. For example, the high-affinity FcR1 receptor for IgE activates the MAPK^erk, MAPK^jnk, and MAPK^p38 pathways in mast cells. Mammalian cells also do not lend themselves to genetic studies as performed in yeast. For this reason, many studies in mammalian cells involve overexpression of activated and inhibitory mutant components of MAPK pathways. A potential pitfall, however, is that this methodology may perturb the segregation mechanisms holding individual MAPK modules. The challenge in the forthcoming years will be to improve existing methods or develop new technologies in mammalian systems to overcome these limitations. Obviously, gene knock-out strategies may be particularly usefull in this context.

View larger version (49K): [in this window] [in a new window]   FIG. 10.   MKKK, MKK, and MAPK that can be components of MAPKerk pathway. GPCR, G protein-coupled receptor; RTK, related tyrosine kinase; ERK, extracellular-regulated kinase; MEK, MAPK or ERK kinase.

In yeast, it appears that transcription factors comprise the majority of the known substrates regulated by MAPK pathways. In mammalian cells, many MAPK substrates defined to date are also transcription factors. In addition, several cytoskeletal proteins, protein kinases, and phopholipases are also substrates for specific MAPK. As our understanding of the yeast and mammalian MAPK pathways increases, it is likely that additional classes of MAPK substrates will be defined.

Four types of MAPK pathways have been defined to date in mammalian cells. However, within a given pathway, several MKKK, MKK, and MAPK can generally be found to be interchangeable. For example, there are 3 isoforms and 10 different splice variants of MAPK in the c-Jun kinase (MAPK^jnk) pathway (123). The relevance of this complexity is poorly understood and is difficult to assess, because it is not straightforward to dissect the role of individual components of the MAPK pathways in mammalian cells. Mammalian MAPK pathways are involved in a diverse set of responses affecting cell fate, including cell proliferation and differentiation, adaptation to environmental stress, and apoptosis.

The best described MAPK signaling pathway in mammalian cells is the extracellular signal-regulated kinase (MAPK^erk) pathway. This pathway can apparently include a number of different MKKK and MKK (Fig. 10). There are five MAPK defined as ERK (MAPK^erk1-5) (52, 274, 308). However, amino acid sequence comparisons indicate that these proteins belong to different subfamilies of MAPK (see Fig. 2). There is actually more divergence between the MAPK^erk1/MAPK^erk2 subfamily and the MAPK^erk3/MAPK^erk4 subfamily than between the MAPK^erk1/MAPK^erk2 subfamily and any other MAPK. Of the collective group of MAPK referred to as ERK, MAPK^erk1 and MAPK^erk2 are the most extensively studied. MAPK^erk1 and MAPK^erk2 are 44- and 42-kDa isoforms, respectively. MAPK^erk1/2 is activated by many different inputs to the cell that lead to the activation of several transcription factors and other serine threonine kinases contributing to cellular proliferation, differentiation, cell cycle regulation, and cell survival. The other ERK are less well characterized. MAPK^erk3 is localized in the nucleus and is activated by PKC isoforms (52, 302). MAPK^erk4 has been shown to be activated in response to nerve growth factor (NGF) and EGF via a Ras-dependent pathway (274). The MAPK pathway employing MAPK^erk3 and MAPK^erk4 remains to be characterized. MAPK^erk5 is discussed in section VD.