|Home | About | Journals | Submit | Contact Us | Français|
Extracellular signal regulated kinases (ERKs) are a class of MAP kinases that function in many signaling pathways in eukaryotic cells and in some cases, a single stimulus can activate more than one ERK suggesting functional redundancy or divergence from a common pathway. Dictyostelium discoideum encodes only two MAP kinases, ERK1 and ERK2, that both function during the developmental life cycle. To determine if ERK1 and ERK2 have overlapping functions, chemotactic and developmental phenotypes of erk1− and erk2− mutants were assessed with respect to G protein-mediated signal transduction pathways. ERK1 was specifically required for Gα5-mediated tip morphogenesis and inhibition of folate chemotaxis but not for cAMP-stimulated chemotaxis or cGMP accumulation. ERK2 was the primary MAPK phosphorylated in response to folate or cAMP stimulation. Cell growth was not altered in erk1−, erk2− or erk1−erk2− mutants but each mutant displayed a different pattern of cell sorting in chimeric aggregates. The distribution of GFP-ERK1 or GFP-ERK2 fusion proteins in the cytoplasm and nucleus was not grossly altered in cells stimulated with cAMP or folate. These results suggest ERK1 and ERK2 have different roles in G protein-mediated signaling during growth and development.
Extracellular signal response kinases (ERKs) are the prototypical MAP kinases (MAPKs) that function in many different signal transduction pathways including those regulated by G protein-coupled receptors and tyrosine kinase receptors [1–4]. Multiple ERKs are expressed in a wide range of eukaryotes suggesting that individual ERKs might have different functions in a single species but conserved functions among different species. Mammals, yeast and Dictyostelium genomes encode multiple ERKs with a highly conserved TEY sequence that can be phosphorylated (both T and Y residues) upon activation by MAP2Ks [5–7]. The simultaneous activation of multiple ERKs in response to a single stimulus opens the possibility that ERK paralogs might have overlapping or redundant functions [2, 3]. Genetic analysis in mice indicate the closely related ERK1 and ERK2 proteins have different roles in development . Loss of ERK2 results in an embryonic lethal phenotype whereas the loss of ERK1 has only subtle phenotypes such as defects in T cell maturation. Also, the down regulation of ERK2 but not ERK1 inhibits the rapid proliferation of tumor cells . The ERK orthologs in yeast, Fus3 and Kss1, are both activated in response to mating pheromone but genetic analysis indicates that only Fus3 is required for efficient mating . Therefore, the simultaneous activation of multiple ERKs might not represent redundancy in signaling but rather divergence of a signaling pathway to regulate multiple responses.
The Dictyostelium genome encodes only two MAPKs, ERK1 and ERK2, that share 37% sequence identity and both are expressed during vegetative growth and multicellular development [5–7, 9]. During the aggregation phase of development, external cAMP activates ERK2 allowing it to phosphorylate and inhibit the cAMP-specific phosphodiesterase, RegA, so that the cAMP signal can be relayed to other cells [10–12]. erk2− cells can chemotax to cAMP, although with low efficiently at high concentrations, but clonal populations of these cells do not form aggregates when starved [7, 13]. ERK2 is also important for the prespore cell specific gene expression and differentiation . ERK2 is also activated in response to folate and erk2− cells exhibit a slight reduction in folate chemotaxis [13, 14]. Less is known about the function of ERK1 but previous studies have reported erk1− cells to be defective in cAMP chemotaxis and to form small aggregates during development . ERK1 can be activated in response to cAMP and this activation is mediated by the MAP2K, MEK1 . While the erk1− and erk2− mutants have differences with respect to developmental phenotypes the specificity of ERK1 and ERK2 function in different G protein mediated signaling pathways has not been defined.
Many different G protein-mediated signaling pathways exist in Dictyostelium and several of them play important roles in growth and development [16–20]. Responses to cAMP are mediated through cAMP receptors and the G protein containing the Gα2 subunit [21–23]. cAMP stimulation is responsible for the aggregation phase of development and aids in the establishment of prespore and prestalk cell zones in the aggregate [24, 25]. Response to folate or related pterin compounds is mediated through a pathway using the Gα4 subunit and this pathway allows cells to chemotax to bacterial food sources . This Gα4 subunit-mediate pathway is also important for the localization and development of prespore cells in multicellular aggregates and the morphogenesis associated with fruiting body formation [18, 27, 28]. Responses to folate are inhibited by another G protein pathway using the Gα5 subunit and the signal activating this pathway is unknown . The Gα5 subunit helps to regulate cell size, growth, and the rate of morphogenesis after aggregate formation . All three of these Gα subunits presumably couple to a common Gβγ dimer and function in pathways that affect ERK function [30, 31]. All three Gα subunits also contain known or putative MAPK docking sites (D-motifs) . The Gα4 subunit is required for folate stimulated ERK2 activation and recently the Gα4 subunit has been shown to associate with ERK2 [10, 14, 28]. The lethality associated with Gα5 subunit over-expression requires a MAPK docking motif and ERK1 function and the Gα2 subunit is not required for ERK2 activation in response to cAMP [10, 32, 33].
In this report we describe an analysis of ERK1 and ERK2 function with respect to different G protein-mediated signaling pathways. Strains defective in ERK1 or ERK2 or both ERKs were analyzed for complementation or suppression with ERK expression vectors. The ERK mutants were also assessed for chemotaxis, development, and sensitivity to Gα subunit over-expression. The phosphorylation of ERKs in response to folate stimulation was examined in ERK mutants and the distribution of ERK1 and ERK2 in the cell was monitored using GFP fusions. These analyses indicate very different functional roles for ERK1 and ERK2 with respect to signaling pathways mediated by G proteins.
All Dictyostelium strains used in the study were derived from the wild-type strain KAx3. The gα4−, gα5−, erk2− strains were previously described [18, 19, 28]. The ERK1(erkA) locus was disrupted in KAx3 and erk2− strains using an ERK1 gene construct with a blasticidin S resistance (Bsr) gene inserted into the coding region and erk1− mutants were identified using primers specific to the ERK1 and Bsr sequences for PCR analysis and sequencing. Strains were labeled with GFP using the expression vector pTX-GFP2. Cells were grown in axenic HL5 medium or on bacterial lawns of Klebsiella aerogenes . DNA vectors were electroporated into cells as previously described . Transformed cells were selected and maintained in medium containing 3–10 µg/ml of the drug G418 or 3 µg/ml of the drug blasticidin S and drug selection was removed several hours prior to analysis. Folate solutions were prepared by neutralizing folic acid with NaHCO3.
An ERK1(erkA) gene with flanking restriction enzymes sites was created from an ERK1 cDNA, kindly provided by the Dictyostelium cDNA project in Japan and National BioResource Project [36, 37]. The open reading frame was amplified using PCR with oligonucleotides (sense strand: 5’-CGCGGATCCCTCGAGAATTAATGCCACCACCACCAACAAGTG) and (antisense strand: 5’-GCGGTCGACTCTAGATTAATTTTTAATTGATTGTTGATTTACTTGTTG) to produce an ERK1 gene with BamHI and XhoI sites 5' to the start codon and XbaI and SalI sites 3' to the stop codon. The HindIII/SalI digested PCR product was inserted into the same sites of the vector pBluescriptIISK+ (Stratagene) and verified by sequence analysis. To create an ERK1 expression vector, the ERK1 gene construct was transferred as SalI/XbaI fragment into the Ddp2-based vector pDXA-3H vector . This act15 promoter driven ERK1 gene construct was then inserted into the Ddp1-based Dictyostelium expression vector pTX-GFP2, replacing the SalI/XbaI fragment that contained a GFP gene . Similar results were obtained using either the Ddp1- and Ddp2-based vectors but only data from Dpd1 vectors is shown. The ERK1 gene was also transferred as XhoI/XbaI fragment into the vector pTX-GFP to create the GFP-ERK1 expression vector. To create an ERK1 gene disruption, the 3' HindIII fragment was deleted from the ERK1 gene and a BamHI fragment containing the Bsr gene was inserted into the unique BclI site in the ERK1 coding region. This construct was excised from the vector as a BamHI-HindIII fragment and electroporated into cells. PCR verification of the ERK1 gene disruption used oligonucleotides specific to the Bsr gene (antisense strand: 5’-CTGCAATACCAATCGCAATGGCTTCTGCAC) and a region of the ERK1 gene outside the sequence used in the disruption construct (antisense strand: 5’-GTATGGTGCCTGTGGATCTTCAGGATG). The PCR product was sequenced to confirm disruption of the ERK1 locus. Multiple erk1− and erk1−erk2− clones with similar phenotypes were identified but the data represents the analysis of a single representative clone of each mutant. ERK2 and GFP-ERK2 expression vectors were constructed as described for the ERK1 expression vectors except that a HindIII/XbaI ERK2(erkB) gene fragment was inserted into the same sites of the pDXA-GFP2 vector and then the act15 promoter driven ERK2 gene was tranferred into the pTX-GFP2 vector. The Ddp1-based Gα4 and Gα5 subunit expression vectors were previously described .
Cells were grown to mid-log phase (approximately 2–3 × 106 cells/ml), washed twice in phosphate buffer (12mM NaH2PO4 adjusted to pH 6.1 with KOH), and suspended in phosphate buffer (1 × 108 cells/ml), before spotting on nonnutrient plates (phosphate buffer, 1.5% agar) for development or chemotaxis as described . Cell development was analyzed using a dissecting microscope or fluorescence microscopy. Above-agar assays were used to measure chemotaxis as previously described . Briefly, folate chemotaxis assay was performed by spotting droplets of cell suspension (107 to 108 cells/ml) on nonnutrient plates followed by the spotting of 1µl of folate solutions (10−2 to 10−4 M) approximately 2–3 mm away from the cell droplet. For cAMP chemotaxis assay, cells were shaken for 4 hours with addition of 100 nM cAMP every 15 min and then spotted on nonnutrient plates followed by the spotting of 1µl of cAMP (10−4 M). Cell movement was monitored with a dissecting microscope. All strains were treated identically for each experiment.
cGMP concentration in Dictyostelium was determined using a radioimmunoassay as previously described . Cells were grown to mid-log phase, washed twice in phosphate buffer (12 mM NaH2PO4 adjusted to pH 6.1 with KOH) and deposited on nonnutrient plates for starvation. After 6 hours of starvation, cells were collected and bubbled with air for 10 min prior to treatment with 1 mM cAMP. At times indicated, the cellular responses were terminated by addition of perchloric acid and then the samples were neutralized with ammonium sulfate. The concentration of cGMP in each sample was determined using a radioimmunoassay kit (Amersham).
Strains were electroporated with the same amount of each vector and each electroporation was performed in duplicate. The number of transformants was determined on each plate section after 7–10 days of drug selection. Electroporations of strains with no DNA were used as controls. After transformants were identified at a low drug selection (3 ul/ml G418), culture plates were treated with increasing selection of drug selection and the number of surviving clones was determined.
MAPK activation was determined as previously described . Cells were grown to mid-log phase (approximately 2–3 × 106 cells/ml), washed twice in phosphate buffer (12mM NaH2PO4 adjusted to pH 6.1 with KOH), and suspended in phosphate buffer (5 × 107 cells/ml). For folate responses, cell suspensions were shaken for 1 hr and then stimulated with 50 mM folate. For cAMP responses, cell suspensions were shaken for 6 hrs and pulsed with 100 nM cAMP every 15 min after the first 3 h. After this preconditioning to cAMP, cell suspensions were stimulated with 100 nM cAMP. Stimulated cells were treated with SDS-PAGE loading buffer at times indicated and boiled. Cell lysates were subjected to SDS-PAGE (8 × 107 cells/lane) and immunoblot analysis was conducted using a rabbit α-phospho-p44/p42 MAPK antibody and secondary goat HRP-conjugated α-rabbit IgG antibody for chemiluminescence detection (Cell Signaling Technology). Densitometry of immunoblot bands was determined using Image J software (NIH).
To compare ERK1 and ERK2 functions in growth and development, ERK1 and ERK2 gene disruptions were created in strains with the same genetic background to reduce strain-specific phenotypes. An erk2− strain previously created using a REMI allele (PYR5-6 selection) insertion into the JH8 strain (KAx3 pyr5-6− strain) was used in all analyses of ERK2 function . The ERK1 gene was disrupted in the KAx3 strain using a blasticidin resistance gene insertion into a unique BclI site within the central part of the coding region and starvation of erk1− clones resulted in small aggregates as previously described by others (Fig. 1) . However, a more detailed inspection of the erk1− cell aggregation revealed relatively short aggregation streams compared to that of wild-type cells. In addition, the timing of mound formation was accelerated with mounds forming as early as 5 hrs and most of these aggregates were either delayed or completely blocked in tip formation (movement of prestalk cells to the top of the aggregate). Aggregates that produced tips were capable of completing development but the size of the mature fruiting bodies was small due to the limited number of cells in the aggregate. At higher cell densities the aggregates of erk1− cells were nearly that of wild-type cells. The introduction of an ERK1 expression vector rescued of the wild-type aggregate size and development indicating the phenotypes are ERK1 specific.
The aggregation phenotype of erk1− cells suggests that cell-cell signaling might be compromised possibly due to defects in responses to the external cAMP signal. Therefore, starved erk1− cells were examined for cAMP chemotaxis using an above-agar assay and erk1− cells were found to travel nearly the same distance as wild-type cells in most experiments suggesting that erk1− and wild-type cells have similar capabilities for cAMP chemotaxis (Fig. 2A). However, many erk1− cells moved in directions other than toward the source of cAMP suggesting an altered chemotactic response compared to wild-type cells. Tracking the movement of individual cells in an above agar assay revealed that erk1− cells typically travel faster than wild-type cells but that erk1− cells were also less directed toward the cAMP source compared to wild-type cells (Fig. 2B). An earlier report had indicated that another erk1− strain was defective in directional movement and also speed but this previous study had examined the movement of individual cells toward a micropipet filled with cAMP . The above agar chemotaxis assay assesses the movement of cells in a large population and therefore a subpopulation of cells with a defect in speed might not be detected. However, the above-agar assay does indicate that many erk1− cells are capable of rapid cell movement in response to cAMP. erk1− cells were also labeled with a GFP expression vector and mixed with wild-type cells for an above-agar cAMP chemotaxis assay and the ratio of erk1− cells to wild-type cells in the leading edge of chemotaxing cells remained constant indicating both wild-type and erk1− cells have similar chemotaxis rates (Fig. 2C). In addition, erk1− cell responsiveness to external cAMP was also assayed with respect to cGMP accumulation because the loss of MEK1 reduces cGMP production. erk1− cells were found to produce similar levels of of cGMP compared to that of wild-type cells indicating that ERK1 function is not necessary for cGMP accumulation (Fig. 2D). This result also suggests that MEK1 regulates cGMP levels through a pathway that does not require ERK1 function.
In mammals, ERK function has been associated with cell growth and development because down regulation of ERK2 function reduces the rapid proliferation of tumor cells and loss of ERK2 results in an embryonic lethal phenotype [8, 40, 41]. Dictyostelium has only two MAPKs and so an erk1−erk2− double mutant strain was created to examine phenotypes in growth and development and to identify possible redundant or oppositional relationships that might exist between the two ERKs. The resulting erk1−erk2− mutant was viable suggesting that ERK function is not required for vegetative growth. The erk1−, erk2−, and erk1−erk2− mutants were tested for growth in axenic and bacterial cultures. Growth rates of all the ERK mutants and wild-type cells were similar in both the axenic and bacterial cultures suggesting that ERK function is not required for vegetative growth of cells (Fig. 3A & B). The starvation of the erk1−erk2− cells resulted in an aggregation defect identical to erk2− cells indicating the loss of ERK1 does not suppress the erk2− phenotype (Fig. 3C). The expression of ERK2 but not ERK1 in the erk1−erk2− cells rescued developmental aggregation but the aggregates displayed the small aggregate phenotype associated erk1− cells. Under high drug selection (5 µg/ml G418), ERK1 expression was not sufficient to rescue the aggregation defect of erk2− cells but ERK2 expression in erk1− cells did allow for larger aggregates to form at low cell densities. The ERK1 expression vector in wild-type cells delayed development starting at the aggregation phase but only when the vector was selected at higher drug concentrations (10 µg/ml). The ERK2 expression vector blocked the aggregation of wild-type cells when drug selection was relatively high (10 µg/ml). The impaired development of wild-type cells expressing ERK1 or ERK2 vectors suggests the over-expression of ERK1 or ERK2 interferes signaling pathways involved with development.
The developmental phenotypes of erk1−, erk2−, and erk1−erk2− cells suggest ERK1 and ERK2 might be required for different processes in cell differentiation or cell sorting. To examine the role of the ERKs in these processes, ERK mutants were labeled using GFP expression vectors and then monitored during development in the presence of wild-type cells. GFP-labeled erk1− cells were found primarily in the anterior half of chimeric slugs and these cells later contributed to both the stalk and spore mass of the mature fruiting body (Fig. 4). The erk1− cells were very sparse in the posterior of slugs, a region typically composed of anterior-like cells (ALCs) and late developing prespore cells. This pattern of cell sorting suggests that erk1− cells have enhanced capability to migrate to the anterior of developing aggregates and that the wild-type cells are providing extracellular signaling necessary to overcome the delay or block in anterior tip formation observed with erk1− clonal aggregates. GFP-labeled erk2− cells aggregated with wild-type cells but many of the erk2− cells in these aggregates were left behind as the mounds developed into slugs. The erk2− cells retained in the slug were found primarily in the prestalk O region near the anterior but not in the extreme anterior, designated as the prestalk A region. The low level of erk2− cells in the central prespore region of the slugs was consistent with earlier reports of erk2− sorting and deficiencies in spore development [7, 9]. The erk2− cells in the prestalk O region contributed to the upper and lower cups around the spore mass of the fruiting body but these cells were relatively scarce in the spore mass or stalk. Similar to erk2− cells, many of the GFP-labeled erk2−erk2− cells found in chimeric aggregates were left behind during slug formation. However, the erk1−erk2− cells retained in slugs were found primarily in the prestalk A region rather than the prestalk O region and these cells contributed to both the stalk and the cups surrounding the spore mass.
Other studies have indicated that G protein mediated pathways activate ERK signaling in Dictyostelium and in some cases the Gα subunits of these pathways contain D-motifs for MAPK interactions [10, 14, 28, 42]. A D-motif in the Gα5 subunit has been determined to be important for the lethality associated with over-expression or constitutive activation of the Gα5 subunit suggesting that MAPK interactions are necessary for this phenotype . To compare the sensitivity to Gα subunit over-expression, ERK mutants were transformed with a Gα5 or Gα4 subunit expression vector and the number of viable transformants was determined. Many viable transformants were found when erk1− and erk1−erk2− cells were transformed with the Gα5 subunit vector but very few wild-type and erk2− cell transformants survived suggesting the Gα5 subunit lethality is specifically mediated through ERK1 and not ERK2 (Fig. 5). In contrast, transformation of a Gα4 subunit expression vector resulted in a similar number of transformants regardless of the ERK mutations. Increasing the drug selection had little impact on the survival of erk1− cells transformed with the Gα5 subunit vector but substantially reduced the number of erk1− cells transformed with the Gα4 subunit vector. In contrast, most erk2− cells transformed with the Gα4 subunit vector retained viability at the higher drug selection. Some of the few viable erk2− transformants with the Gα5 subunit expression vector were not affected by increased drug selection but this might be attributed to mutations or rearrangements in the Gα5 subunit expression allowing the G418 resistance gene to amplify without increasing Gα5 subunit function. Earlier studies have shown a strong selection against Gα5 function, often resulting in Gα5 subunit vector rearrangements . Transformation of an ERK1 or ERK2 expression vector into wild-type or ERK mutants consistently resulted in fewer transformants with the ERK1 expression vector suggesting ERK1, like the Gα5 subunit, might be detrimental to cells when over-expressed.
The possibility that Gα5-mediated signals regulate ERK1 function suggests ERK1 might be necessary for Gα5 specific phenotypes such as the inhibition of folate chemotaxis. Like gα5− cells, erk1− cells have a slightly enhanced chemotactic response to folate (Fig. 6A). Over-expression of the Gα5 subunit inhibits folate chemotaxis but not in erk1− cells suggesting ERK1 is required for this inhibition. Likewise, ERK1 over-expression inhibits folate chemotaxis but not in gα5− cells further supporting that the Gα5 subunit and ERK1 function in the same pathway. The over-expression of the Gα4 subunit in wild-type and erk1− cells enhances folate chemotaxis suggesting the Gα4 subunit-mediated responses to folate do not require ERK1 function.
Both ERK1 and the Gα5 subunit affect tip morphogenesis suggesting these proteins might function in a common signaling pathway to regulate this process. To examine this possibility, gα5− and erk1− cells carrying expression vectors for the Gα5 subunit and ERK1 were monitored for tip formation. The Gα5 subunit expression vector promotes small aggregate formation and accelerated tip formation in gα5− cells even under low drug selections of 2 µg/ml G418 (Fig. 6B). This Gα5 subunit-mediated precocious tip formation is not observed in erk1− cells expressing the Gα5 subunit vector even under high drug selection (10 µg/ml) indicating ERK1 is required for this phenotype. However, over-expression of the Gα4 subunit delayed aggregation in erk1− cells, as previously described for wild-type cells, indicating this Gα4 subunit-mediated phenotype does not require ERK1 function . Over-expression of ERK1 resulted in a delay in aggregation in wild-type and erk1− cells but not in gα5− cells suggesting this ERK1-mediated phenotype is dependent on Gα5 subunit function.
Previous reports have indicated that cAMP stimulation of Dictyostelium results in the activation of ERK1 and ERK2 and some of these studies have used antibodies generated against the conserved activation site (human phospho-p44/42 MAP kinase - TEY region) to detect phosphorylated MAPKs [10, 11, 42]. Using these antibodies a band approximately 42 kDa is detected from wild-type cells and this band is greatly reduced in erk2− or erk1− erk2− cells but not in erk1− cells, suggesting this band corresponds to phosphorylated ERK2 and not ERK1 (Fig. 7A). The detection of a weak band in erk2− and erk1− erk2− cells possibly represents leakiness of the erk2− allele or another phosphorylated protein. ERK1 is predicted to be approximately 48 kDa but these antibodies do not detect a band of this size after cAMP stimulation even though ERK1 is slightly more conserved than ERK2 to the human ERKs in this region. A protein blot of cells expressing His6-tagged ERK1 or ERK2 confirms that ERK1 migrates slightly slower than ERK2, as expected for the difference in protein size (Fig 7B). Therefore the inability to detect ERK1 after cAMP stimulation is likely due to either a very weak phosphorylation of ERK1 or another mechanism for ERK1 activation.
Stimulation of Dictyostelium with folate results in rapid phosphorylation of ERK2 [28, 42]. Consistent with earlier studies this phosphorylation requires Gα4 subunit function and ERK2 expression as indicated by the large reduction in the detected protein from gα4− and erk2− cells (Fig. 8). The phosphorylation of ERK2 in erk1− cells was similar to that of wild-type cells and the phosphorylation level of erk1− erk2− cells was similar to that of erk2− cells. These results suggest ERK1 has very little impact on ERK2 activation in response to folate. No bands were detected in the range of 48 kDa implying ERK1 is not activated in response to folate (data not shown). ERK2 phosphorylation in gα5− cells was similar to that of wild-type, consistent with the Gα5 subunit not being required for the folate responses.
In many organisms, MAPKs can move from the cytoplasm into the nucleus after cell stimulation [43, 44]. A previous study of ERK2 using protein blots of Dictyostelium lysates indicated that only a very small percentage of ERK2 was associated with nuclear fractions after cAMP stimulation . As an alternative method to assess ERK distribution in the cell, GFP-ERK1 and GFP-ERK2 fusions were created and expressed in cells so that ERK distribution could be monitored in live cells. Both of these fusion proteins were capable of rescuing developmental defects of the respective ERK mutant (Fig. 9A). Both GFP-ERK1 and GFP-ERK2 were found in the cytoplasm and nucleus of cells (Fig. 9B). Upon stimulation of these cells with cAMP or folate no major changes were observed with the distribution of GFP-fusion proteins inside of cells (data for folate stimulation not shown). Although phosphorylated MAPKs can be detected as early as 20 sec after cAMP or folate stimulation, no major changes were observed even up to 3 min after stimulation (data not shown). No major changes in GFP-ERK1 and GFP-ERK2 distribution were observed during chemotaxis to folate on agar plates (data not shown). However, the detection of GFP-ERK1 or GFP-ERK2 does not differentiate between phosphorylated and unphosphorylated forms of the MAPK and so a large pool of inactive GFP-MAPK might possibly mask changes in a smaller pool of activated protein. The analysis of MAPK distribution was also conducted on cells expressing high levels of GFP-ERK1 or GFP-ERK2 as the result of increased drug selection. At the higher expression levels, GFP-MAPK fusions were more concentrated in the nucleus, an observation also found for GFP-MAPKs in mammalian cells . The stimulation of these cells with cAMP or folate also did not result in any major changes in GFP-ERK1 or GFP-ERK2 distribution (data for folate stimulation not shown).
The analysis of ERK1 and ERK2 function in G protein-mediated signaling pathways suggests these ERKs are not redundant or overlapping in function but rather have different roles in different signaling pathways (Fig. 10). The requirement of ERK1 but not ERK2 for the precocious tip development and lethality associated with Gα5 subunit over-expression support a role for ERK1 in Gα5 subunit mediated signal transduction. The requirement for both ERK1 and the Gα5 subunit to inhibit folate chemotaxis is another indication these signaling components function in a common pathway. Previous studies have indicated the importance of ERK2 for Gα4 subunit mediated responses such as chemotaxis to folate even though the erk2- cells are only partially defective in this response [13, 26]. Although not tested, the Gα4 subunit and ERK2 are also likely to function together in folate-stimulated cAMP accumulation and prespore cell localization and development based on the phenotypes associated with gα4− and erk2− cells [7, 9, 18, 26]. ERK2 rather than ERK1 appears to the primary MAPK phosphorylated in response to cAMP and the importance of ERK2 in cAMP-mediated aggregation is exemplified by the aggregation defect of erk2− cells. Unlike the responses to folate, cAMP-stimulated ERK2 activation does not require the Gα2 or Gβ subunits [10, 11, 32]. The inability of erk1− cells to form large aggregation streams suggests a possible limitation on the aggregation territory that might be attributed to deficiencies in intercellular signaling. However most erk1− cells appear to be acutely responsive to cAMP-stimulated chemotaxis or cGMP accumulation.
The distribution of erk1− and erk2− cells in chimeric organisms implies cell differentiation and/or sorting preferences of these cells is significantly different. The presence of erk1− in the prestalk and prespore regions near the anterior of the migratory slug is consistent with the ability of erk1− cells to form both spore and stalk cells but inconsistent with the block or delay of tip morphogenesis in clonal populations of erk1− cells. While capable of migrating to the tip regions, erk1− cells must have defects in intercellular signaling that can be diminished by the presence of wild-type cells. Therefore, ERK1 might possibly mediate the production of signals necessary for tip morphogenesis rather than responding to chemotactic signals such as cAMP that promote localization to the tip . The spatial pattern of erk2− cells in chimeric organisms reflects the partial deficiencies of these cells to chemotax to cAMP and to become prespore or prestalk A cells as previously described [5, 7]. The erk1− erk2− cells have these same deficiencies as erk2− cells except the erk1− erk2− cells can migrate to the prestalk A region. This difference is likely due to the absence of ERK1 as indicated by the strong anterior positioning of erk1− cells.
The inability to detect phosphorylated ERK1 in response to cAMP or folate suggests very limited if any ERK1 is activated by the typical phosphorylation mechanism in these responses. Possible explanations for the absence of phosphorylated ERK1 might be a limitation on ERK1 expression as suggested by the reduced viability of cells transformed with ERK1 compared to ERK2. Alternatively, ERK1 activation might occur with very different kinetics compared to ERK2. A previous report of ERK1 activation in response to cAMP indicated a relatively rapid peak in ERK1 kinase activity at 15 s. Interestingly, the phosphorylation of mammalian ERKs has been reported as early as 1 min after stimulation by a chemoattractant and the activated state can be maintained for over 15 min after stimulation with a growth factor. Differences in Dictyostelium ERK activation might also reflect the specificity of ERK1 or ERK2 involvement in a particular pathway or a subpopulation of developing cells. In mice, the loss of ERK1 can result in increased ERK2 activation suggesting an oppositional relationship between related MAPKs . The similarity of ERK2 activation in wild-type and erk1− cells suggests there is no such oppositional relationship between ERK1 and ERK2 function, at least with respect to cAMP- and folate-stimulated MAPK phosphorylation in early development.
Neither ERK1 nor ERK2 is required for cell growth suggesting Dictyostelium MAPK function is not an important factor for the proliferation of unicellular eukaryotes. This conclusion was implied by earlier studies showing the viability of yeast fus3−kss1− mutants but yeast have other MAPKs that might compensate for the loss of these ERKs . The requirement of mammalian ERK2 for tumor cell growth and embryonic development suggests that ERK function might be necessary for the proliferation of a specific differentiated cell type or to overcome growth inhibition mechanisms in multicellular organisms [8, 47]. This concept is consistent with the regulation of cell growth through growth factors activate MAPKs.
Our results indicate that ERK1 and ERK2 in Dictyostelium function in different G protein-mediated signaling pathways rather than providing redundant functions in a common pathway. ERK1 is necessary for signaling in Gα5 subunit-mediated pathways and ERK2 is necessary for signaling in Gα4 subunit-mediated pathways. While ERK1 and ERK2 could possibly perform similar functions in different pathways, these pathways clearly regulate different developmental processes as indicated by the phenotypic analysis of erk1− and erk2− mutants.
The authors thank Meggie Finley and Nicole Clarkson for technical assistance. This work was supported by a grant R15 GM073698-01 from the National Institute of General Medical Sciences provided to JAH.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.