We found novel development rescuing factors (DRFs) secreted from developing Dictyostelium cells, by using a mutant (erkB−) which is missing MAP-kinase ERK2 as a test strain for bioassay. The mutant erkB− fails to undergo multicellular morphogenesis due to impaired cAMP signaling. However, such developmental defect can be restored by the presence of low-molecular weight DRFs that are secreted from developing wild-type cells. We previously showed that DIF-1 (Differentiation-Inducing Factor 1 for stalk cells) possesses this activity, indicating a newly discovered role of DIF-1. Surprisingly, however, the mutant dmtA−, which is incapable of DIF-1 synthesis still exerts a strong inducing activity of the multicellular morphogenesis of erkB−. After analysis of HPLC fractions of conditioned media prepared from both wild type Ax2 and dmtA− strains revealed that both strains secrete at least two novel DRF activities with DIF-like mobility. However, these activities were not derived from other DIFs such as DIF-2 and DIF-3. Identification of these DRFs found in this study would provide insight into the mechanism by which the development of the erkB− mutant is restored and how these factors act in the normal development of Dictyostelium.
Intercellular communication mediated through secreted factors is essential for both the cell differentiation and morphogenesis of multicellular organisms. Dictyostelium, one of the simplest multicellular eukaryotes, provides a model system for studying the mechanisms underlying such intercellular communication. Upon starvation, free-living Dictyostelium amoebae initiate multicellular development via their aggregation to form a hemispherical mound where prestalk- and prespore-cell differentiation occurs. These cells are sorted eventually to form a fruiting body consisting of a round spore mass supported by a cellular stalk. Moreover, the proportion and sorted pattern of prestalk and prespore cells are highly regulated. Co-ordinate cell movement and regulated pattern of cell differentiation during morphogenesis suggest that intercellular signaling plays crucial roles in the morphogenesis of Dictyostelium.
Most studied extracellular signaling molecules in Dictyostelium are cAMP and DIF-1 (Differentiation Inducing Factor for stalk cells). cAMP acts as a chemoattractant for the aggregation and the morphogenetic movement of Dictyostelium cells and also regulates gene expressions for the cell differentiation (Konijn et al., 1969; Gross et al., 1976; Tomchik and Devreotes, 1981; Rietdorf et al., 1998). Several mutants with a defect in cAMP signaling have been isolated in Dictyostelium. They are all arrested at the preaggregation stage, however, some mutations lacking adenylyl cyclase, cytosolic regulator of adenylyl cyclase or MAP-kinase ERK2 are non-autonomous, because the mutants develop into fruiting bodies in the presence of wild-type cells (Pitt et al., 1992; Insall et al., 1994; Segall et al., 1995). In this connection, we further analyzed the effects of wild type cells on erkB− lacking functional ERK2 and found that the mutant phenotype was not restored by cAMP (Maeda and Kuwayama, 2000). However, certain factors secreted by the developing wild type cells restored developmental defect in the mutant (Maeda and Kuwayama, 2000; Kuwayama et al., 2000). This suggests that wild-type cells provide certain signaling molecules other than cAMP that are lacking in the mutant. These signaling molecules were named DRFs (Development Rescuing Factors) after their biological activity as mentioned above.
DIF-1 is another well-analyzed signaling molecule in Dictyostelium, which is a dichloro-alkylphenone, a hydro-phobic signaling molecule that induces differentiation of stalk cells in vitro (Town et al., 1976; Town and Stanford, 1979; Morris et al., 1987). Two DIF-1 related compounds, DIF-2 and DIF-3, were also identified (Morris et al., 1988). DIF-3 is the initial product in the breakdown process of DIF-1. DIF-2 has a quite similar structure to DIF-1, which could be a side-product of DIF-1 biosynthesis. DIF-1 accounts for more than 95% of the total bioactivity and DIF-2 for most of the rest (Kay et al., 1983; Brookman et al., 1987). Both DIF-1 and DIF-2 are also involved in the light scattering responses of early developing cells in suspension, suggesting that they play a distinct role from stalk cell inducers (Wurster and Kay, 1990). Furthermore, we revealed that DIF-1 acts as DRF which induces the multicellular development of erkB− mutant missing MAP kinase ERK2, a component of cAMP signaling cascades in Dictyostelium (Segall et al., 1995; Kuwayama et al., 2000). This is a newly discovered role of DIF-1.
In the present study, we examined whether DIF-1 is a sole agent acting to induce the multicellular development of erkB− mutant. First we assessed whether DIF-2 and DIF-3 also act as DIF-1 did. Second we examined whether a DIFless mutant (dmtA−) that was generated by disrupting the gene encoding methyltransferase which catalyzes the final step of the pathway for the biosynthesis of DIF-1 (Kay, 1998; Thompson & Kay, 2000) secretes certain factors to induce the multicellular development of erkB− mutant. We show here that there are at least two novel hydrophobic factors which are able to induce the multicellular development of erkB− mutant and secreted from both wild type Ax2 and a dmtA− mutant. Identification and further analysis of these factors would provide insight into the mechanism by which the developmental defect of the erkB− mutant is corrected and how these DRFs act in the normal development of Dictyostelium.
MATERIALS AND METHODS
Strains and cultures
Dictyostelium discoideum Ax2 (a wild-type strain), HS175 (erkB− mutant),dmtA− (DIF-less mutant) and V12M2 were used. All strains except V12M2 were grown at 21°C in HL5 supplemented with 5 ng/ml of Vitamin B12 and 100 ng/ml of folic acid (Watts & Ashworth, 1970). V12M2 cells were grown on SM agar plates in association with Klebsiera aerogenes and used for DIF assay.
Spot test for rescue assay
DRF activities were assayed by the spot test. Cells were harvested at 2∼6×106 cells/ml and were washed twice with 12 mM Na2K phosphate buffer, pH 6.1 (PB). Then, the washed cells were resuspended in PB at a density of 5×107∼1×108 cells/ml. To examine the effects of DIF-1, DIF-2 and DIF-3, a drop of a 5 μl aliquot of the erkB− cell suspension was spotted on 1.6% hydrophobic agar containing various concentrations. Hydrophobic agar was prepared by extensively washing agar powder with deionized water. To examine the DRF activity of the secreted factors of dmtA−, a 5 μl aliquot of the dmtA− cell suspension was placed at the side of a drop of erkB− suspension. The distance between the two spots was about 2 mm. The DRF activities of HPLC fractions were tested in the same way. A 0.5 μl aliquot of each fraction and sometimes 5, 10, 50, 100, 200, and 300-folds dilutions were deposited instead of the dmtA− cell suspension.
Extraction of factors sereted from Dictyostelium cells
Secreted factors were collected and extracted by the same method used for DIF collection and extraction (Kay et al., 1983). Washed Ax2 or dmtA− cells were allowed to develop at a density of 1.5×107 cells/cm2 on a layer of cellophane (325P, British Cellophane Co, Bridgewater, UK) supported by a stainless steel mesh in a surgical tray. The underside of the cellophane was bathed with a medium containing 10 mM KCl, 2 mM NaCl, 1mM CaCl2 and 5–10 g/l washed Amberlite XAD-2 resin (BDH). The trays were shaken at about 16 rev./min on a New Brunswick G2 orbital shaker. After 3–5 days, XAD-2 beads whose color had changed to yellow from adsorbing material were harvested and washed with water. The adsorbed material was eluted with ethanol. The elute concentrated by rotary evaporation, was dispersed in water, and the process was repeated to take up the elute largely in water. Then, the elute was extracted with an equal volume of ethyl acetate five times and was dried down to dissolve with 70% ethanol.
The sample extracted as described above was eluted through a Hichrom reverse-phase 25 -cm column packed with 5-μm particles of Spherisorb ODS-2 and analyzed with a Beckman Gold solvent delivery module controlled by a NEC PC800 computer and fitted with a Beckman L160 UV detector. Elution was performed at 3 ml/min with a gradient from 60% to 90% solvent B in solvent A (solvent A=2% acetic acid, solvent B=2% acetic acid in methanol) in 120 min. Eluted fractions were collected every minute, dried in a Speed Vac vacuum concentrator (Savant), and redissolved in 500 μl of 70% ethanol.
The activity of DIF for induction was measured as the ability to induce differentiation into stalk cells in isolated cells of strain V12M2 (Brookman et al., 1982). The cells were harvested from SM agar plates and washed free of bacteria. Then, 0.9×104 cells were plated in a tissue culture dish (ø=3 cm) under 1.5 ml of stalk salt solution (10 mM MES, 10 mM KCl, 2 mM NaCl, and 1 mM CaCl2, pH 6.2) containing 5 mM cAMP and each HPLC fraction. Stalk cells were scored by phase contrast microscopy after 2 days of incubation at 22°C. One unit (U) of the activity induces 1% stalk cell differentiation in this assay.
Comparison of DRF-activities among DIF-1, DIF-2 and DIF-3
The chemical structures of DIF-1, DIF-2 and DIF-3 are shown in Fig. 1. DIF-1 and DIF-3 have a C4 alkyl side chain, whereas DIF-2 has a C3 alkyl side chain. DIF-3 is the monochlorinated analogue that is the initial degradation product of DIF-1 (Nayler et al., 1992). Distinct from DIF-1, DIF-2 and DIF-3 show minor biological activity for stalk cell induction. Here, we compared the DRF activity of DIF-1, DIF-2 and DIF-3 by placing erkB− cells on agar plates containing various concentrations of each DIF. To quantify the activity, the number of fruiting bodies formed after two days of incubation was counted (Fig 2). The DRF activity of DIF-1 was highest at 5 μM though it nearly reached this peak at 2 μM, but it was greatly reduced at 10 μM. At 10 μM, almost all the erkB− cells differentiated into stalk cells (data not shown). On the other hand, the activity of DIF-2 was hardly detectable until 2 μM, but increased at 5 and 10 μM in a dose-dependent manner. The number of fruiting bodies on DIF-2 agar at 5 μM was almost the same as that on DIF-1 agar at 1 μM and was about 25% of that of 5 μM DIF-1 agar. DIF-3 showed weak activity only at 10 μM. These results demonstrate that DIF-1 is most effective not only for stalk cell induction but also for the restoration of the development of erkB−. The DRF activity of DIF-2 was 20∼30% of that of DIF-1. The activity of DIF-3 was 15% of that of DIF-2 at 10 μM, that is, 3∼4% of the activity of DIF-1. We could not see any striking difference among the phenotypes of erkB− mutants restored by these DIFs (data not shown).
Restoration of the development of erkB− by DIF-less mutant dmtA−
In an attempt to determine whether DIF-1 is the sole agent with strong DRF activity, we focused on the DIF-less mutant dmtA−. The gene dmtA encodes the methyltransferase that catalyzes the final step in the pathway of DIF-1 biosynthesis. AdmtA− mutant was created by homologous recombination (Kay, 1998: Thompson and Kay, 2000). We tested whether the mutant is able to rescue the development of erkB− and revealed that the mutant secreted a DRF with activity similar to the wild-type (Fig 3). We also found that precursor molecules such as THPH (2,4,6-trihydroxyphenyl-1-hexan-1-one) and dichloro-THPH at the consecutive steps of DIF-1 biosynthesis never showed any DRF activity (data not shown). These indicate the existence of novel DRFs other than DIF-1 and its precursors.
Profile of DRF activity in conditioned media prepared from wild-type and dmtA− after HPLC
In order to identify novel DRFs, we pooled the conditioned media prepared from developing wild-type and dmtA− cells, respectively, and analyzed DRF activity after HPLC (see MATERIALS AND METHODS). First, the activity for stalk cell induction (DIF activity) in the wild-type preparation was assayed by the DIF assay. A single major DIF-peak was resolved in the fractions at 87–89 min (termed DIF-1 peak) where the authentic DIF-1 was eluted (Fig 4a). A minor DIF peak was also resolved in the fractions at 73–75 min (termed DIF-2 peak) at which the authentic DIF-2 was eluted (Fig 4a). DRF activity was also assayed in each HPLC fraction by the spot test. Two peaks of DRF activity were resolved in the wild-type preparation. One peak corresponded to the DIF-1 peak and the other nearly matched the DIF-2 peak. Thus, we speculated that the DRF activities in these two peaks resulted from DIF-1 and DIF-2. However, we obtained some unexpected results regarding the DRF peak around the DIF-2 peak. The DRF activity of DIF-2 was about 20∼30% of that of DIF-1 as mentioned previously and the amount of DIF-2 was much less than that of DIF-1. Nevertheless, the DRF activity of the fractions at 73 and 74 min that should contain DIF-2 was as strong as that of the DIF-1 peak. Surprisingly, the fraction at 76 min demonstrated the strongest DRF activity (Fig 4a). As the fraction was supposed to lack any DIF (Fig 4a), these two DRF-peaks should be distinct from DIF-1 and DIF-2.
This prediction was supported by similar analysis of the conditioned media prepared from dmtA−. Stalk-cell inducing activity by DIF-1 and DIF-2 was not detected at all in the mutant preparation, which is consistent with the report of Thompson and Kay (2000). Consistently, one of the major peaks of DRF activity in the DIF-1 fractions disappeared in the mutant preparation (Fig 4b). However, two other peaks of DRF-activity were left at around the DIF-2 peak of the wild-type (Fig 4b), indicating the existence of two novel DRFs other than DIF-1 and DIF-2. The development of the erkB− mutant rescued by the DRFs was phenotypically indistinguishable from that rescued by DIFs (data not shown).
We previously reported that diffusible factors released from developing wild-type cells rescue the development of the erkB− mutant and one of these factors is DIF-1 (Maeda and Kuwayama, 2000; Kuwayama et al., 2000). In this study, we demonstrated that DIF-2 and DIF-3 also have DRF activity. Much importantly, we revealed that at least two novel DRFs were secreted by the developing cells.
The bioactivity for stalk cell induction of DIF-2 is much weaker than that of DIF-1 and DIF-2 evokes the light scattering response in early developing Dictyostelium cells (Wurster and Kay, 1990). Therefore, we expected DIF-2 to possess particular functions in vivo related to DRF activity rather than stalk cell induction. This, however, is very unlikely since DIF-2 is less effective than DIF-1 in restoring the development of the erkB− mutant. At 5 μM, DIF-1 showed the highest level of DRF activity among the DIFs. DIF-2 and DIF-3 were respectively 20∼30% and 3∼4% as active as DIF-1, which is roughly comparable to the activity for stalk cell induction of DIF-2 and DIF-3 being 40% and 16% of that of DIF-1 (Masento et al., 1988). In this regard, it is notable that the DRF activity of DIF-1 was highest at 5 μM but declined to be less than that of DIF-2 at 10 μM. At high concentrations such as 10 μM, DIF-1 stimulated erkB−cells to differentiate into stalk cells without multicellular development. These results suggest that the DRF activity of DIF is related to the pathway of prestalk cell differentiation. As DIF-1 is known to increase intracellular proton and calcium levels (Inouye, 1988a; Schaap et al., 1996; Azhar et al., 1997), we expected such an increase to be involved in how DIF-1 rescues the development of erkB−. Then, we tested whether a rise in the cytoplasmic calcium or proton concentration exerts a DRF-effect using calcium pump or proton ATPase inhibitors (Tsujioka et al., unpublished; Inouye, 1988b; Kubohara and Okamoto, 1994). Our preliminary results showed that these inhibitors induced stalk cell differentiation, but not the restoration of the mutant development (Tsujioka et al., unpublished data). However, we also found that 5 mM NH4Cl eliminated the DRF activity of DIF-1 (Tsujioka et al., unpublished data), suggesting the involvement of a rise in the cytoplasmic proton concentration in the mechanism by which DIF-1 functions as a DRF.
Our study clearly showed the existence of at least two novel DRFs, though they remain to be identified. Both are diffusible in agar and have a similar hydrophobicity to DIF-2 as revealed by the fact that they eluted in nearly the same fractions as DIF-2 on HPLC. However, these DRFs never show stalk-inducing activity. Also, Maeda and Kuwayama (2000) have demonstrated that DRF activity becomes detectable after 8h of starvation and that the molecular masses of all DRFs are less than 3000. It is important to purify these DRFs and determine their structure. The analysis of these factors should provide insight into one more intercellular communication system of Dictyostelium during early development.
This study was supported by grant-in-aids for Scientific Research on Priority Areas from the Ministry of Education, Science and Culture of Japan to MM (12640633 and 13024248).