Evidence for a role for the Dictyostelium Rap1 in cell viability and the response to osmotic stress

The mammalian Rap proteins are part of the Ras subfamily of proteins, which exhibit ∼50% identity to the Ha-Ras protein(Reuther and Der, 2000). The rap1 gene was discovered originally by its ability to suppress the malignant phenotype of ras-transformed cells(Kitayama et al., 1989), and Rap1 is capable of competitively inhibiting Ras signaling pathways by binding to downstream effectors, suggesting a mechanism for suppression(Cook et al., 1993;Boussiotis et al., 1997;Okada et al., 1998;Hu et al., 1999). However,Rap1 can itself act in growth-factor-induced signaling pathways(Yoshida et al., 1992;Altschuler and Rebeiro-Neto, 1998;Zwartkruis et al., 1998). Rap1 is also involved in B and T cell activation(McLeod et al., 1998;Reedquist and Bos, 1998),platelet activation (Franke et al.,1997), neutrophil activation(M'Rabet et al., 1998),neuronal cell differentiation (Vossler et al., 1997; York et al.,1998) and cyclic-AMP-mediated signaling(DeRooij et al., 1998;Kawasaki et al., 1998). More recently, evidence has been presented to indicate that Rap1 is involved in mammalian cell adhesion and in the inflammatory response(Tsukamoto et al., 1999;Reedquist et al., 2000;Katagiri et al., 2000;Caron et al., 2000). Thus, in mammals Rap1 is capable of many roles depending upon the cell type. Dissecting the various functions of Rap in mammalian cells is complicated by the fact that there are at least four rap genes(Reuther and Der, 2000),raising the possibility of redundant or overlapping functions.

In Drosophila, there is a single rap gene, and loss-of-function mutations are lethal at the larval stage(Hariharan et al., 1991). This lethality can be rescued by expressing rap under the control of a heat shock promoter, and cell proliferation was unimpaired but morphogenesis and cell movement were abnormal during development in the absence of Rap1 function (Asha et al.,1999).

In Dictyostelium, the available evidence suggests that there is also a single rap gene, rapA, previously designated rap1, encoding a protein, Rap1, 75% identical to mammalian Rap1A(Robbins et al., 1990). The overexpression of rapA results in cells with a variety of cytoskeletal defects, including a flattened cell morphology and failure to contract in response to contraction stimuli(Rebstein et al., 1993). In addition, cells overexpressing activated and dominant-negative forms of Rap1 exhibited alterations in phagocytosis and fluid phase endocytosis(Seastone et al., 1999). To provide a more definitive assessment of Rap1 function in Dictyostelium, we attempted to disrupt the rapA gene by standard procedures. However, these attempts failed, suggesting that rapA might be an essential gene in Dictyostelium. In the present study, we have expressed a rapA antisense construct under the control of the folate repressible discoidin promoter(Blusch et al., 1992) and have examined the effects of Rap1 depletion on cell function. We have also demonstrated that Rap1 is activated in response to hyperosmotic stress and low temperature.

Dictyostelium Ax-2 cells were grown in HL-5 broth culture, either in shake suspension (Watts and Ashworth,1970) or in 10 cm Petri dishes at 22°C. To determine growth rates, cell numbers were counted in a hemocytometer. To determine cell viability, Dictyostelium amoebae were diluted and plated in association with Klebsiella oxytoca on nutrient-rich agar plates(Sussman, 1987) and incubated at 22°C. Under these conditions only viable cells form plaques on the bacterial lawn.

To create the pVEII-AS5′construct, the 5′ portion of the rapA gene (nucleotides -15 to +218) was synthesized by PCR using the oligonucleotides 5′-TGCTCTAGAGCTCGAATTCATCATGCC-3′ and 5′-TGCTCTAGAGCAGTAAATTGTTCAGTACG-3′ as primers and rapAcDNA (Rebstein et al., 1993)as the template. Both primers contained KpnI sites, and the PCR product was inserted into the KpnI-digested pVEII vector. The resulting constructs contained the antisense rapA DNA linked to the folate repressible discoidin promoter. The orientation and promoter-gene fusions were confirmed by sequencing.

The pVEII-AS5′ vector was introduced into Ax-2 cells by electroporation, and single cell transformants were selected in 96-well plates in HL-5 medium supplemented with 10 μg/ml G418 and 1 mM folate. Six pVEII-AS5′ transformants were obtained, and each was plaque purified on a lawn of K. oxytoca. The isolated transformants were cultured on 10 cm Petri dishes in HL-5 growth media, supplemented with 100μg/ml G418 and 1 mM folate for three days and then maintained on HL-5 media containing 10 μg/ml G418 and 1 mM folate. The cultures were frequently divided into fresh growth media to maintain low cell density.

Exponentially growing cells were centrifuged at 600 g,washed three times with KK₂ and resuspended at 1×10⁶ cells/ml. The cell suspensions were left on ice for 15 minutes to produce isolated, spherical cells and then viewed through a microscope. The cells were photographed and their diameters determined from prints.

Exponentially growing cells were centrifuged at 600 g,washed three times with KK₂ and resuspended in the same buffer at 3.0×10⁷ cells/ml and shaken at 160 rpm for 1 hour at 22°C. To induce hyperosmotic shock, sorbitol was added to a final concentration of 0.4 M. Aliquots were removed periodically and analyzed for either cell viability, cGMP content or the level of activated Rap1.

To induce hypo-osmotic shock, cells that had been shaken at 22°C for 1 hour in KK₂ were centrifuged and resuspended in ddH₂O. Cells were subjected to temperature shock by placing cells at 8°C for cold shock or 30°C for heat shock. Aliquots were removed at various time points and analyzed for the level of activated Rap1.

Levels of cGMP were determined essentially as described previously(Oyama, 1996). 100 μl aliquots of cells that had been subjected to osmotic shock were added to 100μl of 3.5% perchloric acid and the mixture incubated on ice for 30-60 minutes with occasional vigorous shaking. The solution was neutralized by the addition of 50 μl of 50% saturated KHCO₃ and the mixture incubated for 60 minutes on ice with occasional vigorous shaking. The suspensions were centrifuged at 14,000 g for 10 minutes at 4°C and 100 μl of the supernatant was assayed for cGMP content using the Amersham Pharmacia Biotech radioisotope dilution assay.

Between 5×10⁶ and 1×10⁷ cells were washed twice in KK₂, resuspended in 1% SDS, and the protein concentration was determined using the Bio-Rad protein assay. A 20 μg protein aliquot of each sample was mixed with an equal volume of SDS sample buffer (0.5%β-mercaptoethanol; 0.5% SDS; 50 mM Tris-Cl, pH 6.8; 12.5% glycerol, and 0.04% bromophenol blue), boiled for 5 minutes and then fractionated by SDS-polyacrylamide gel electrophoresis(Laemmli, 1970) using 12%gels. After electrophoresis, the proteins were transferred to nitrocellulose membranes for 1 hour and probed with either polyclonal Rap1 antibody(Rebstein et al., 1997) or a monoclonal phosphotyrosine antibody(Ingham et al., 1998). The membranes were incubated for 1 hour at room temperature in TBS-Tween (25 mM Tris-Cl, pH 8.0, 1.0% NaCl, 1% Tween 20) containing 5% powdered milk(Carnation) for Rap1 detection or 4% BSA for phosphotyrosine detection. The Rap1 primary antibodies were diluted 1:2,000 in TBS-Tween containing 0.5%powdered milk and incubated with the nitrocellulose membranes overnight at room temperature. The phosphotyrosine antibodies were diluted 1:200 in TBS-Tween and incubated with the membranes overnight at 4°C. The membranes were washed three times for 5 minutes in TBS-Tween and exposed to a secondary antibody (donkey anti-rabbit IgG conjugated to horseradish peroxidase or donkey anti-mouse IgG conjugated to horseradish peroxidase) diluted 1:10,000 in TBS-Tween. The bound antibodies were detected by an enhanced chemiluminescence assay (Amersham). Blots were scanned using a ScanJet-II scanner (Hewlett-Packard, USA). and densitometry was performed using Image Quant (V.1.2) software for MacIntosh.

The mammalian Rap-binding domain (RBD) of RalGDS was expressed in Escherichia coli as a GST fusion protein as described previously(McLeod et al., 1998). Exponentially growing cultures of bacteria were induced with 0.1 M IPTG for 16 hours at 22°C. The cells were lysed by sonication for two minutes in 50 mM Tris-Cl, pH 7.5; 150 mM NaCl; 1% TritonX-100; 1 mg/ml lysozyme and 0.1 mg/ml DNase I. The resulting cell lysate (10-50 μl) was incubated with 10 μl of glutathione-Sepharose beads (Pharmacia) at 4°C for one hour, and the beads were then washed three times with wash buffer (20 mM Tris-Cl, pH 7.6;150 mM NaCl, 0.1% Triton X-100; 10 μg/ml leupeptin; 1 μg/ml aprotinin and 1 mM PMSF).

Rap1 was also expressed as a GST fusion protein in E. coli. The original rapA cDNA was complete at the 3′ end, but the 5′end was truncated at an EcoRI site, 12 bp downstream of the ATG translation initiation site of the gene(Robbins et al., 1990). The 5′ end was completed through a single oligonucleotide mutagenesis, which eliminated the internal EcoRI site through a conservative change at nucleotide +15 and introduced an EcoRI site at the 5′ end of the completed cDNA. The rapA construct was treated with EcoRI, and the 700 bp fragment was ligated to an EcoRI-digested pGEX-1 vector. This gst-rapA construct was transformed into E. coli. Exponentially growing cultures of bacteria were induced with 0.1 M IPTG for 4 hours at 37°C and the bacteria lysed as described above. Cell lysate (0.5 ml) was mixed with 50 μl of packed glutathione-Sepharose beads at 4°C for 1 hour and the beads then washed three times with wash buffer. The bound GST-Rap1 was treated with 10 units/ml thrombin in 50 mM Tris-HCl (pH 7.6); 150 mM NaCl; 2 mM CaCl₂ to hydrolyze the peptide bond between Rap1 and GST, and the beads were removed by centrifugation. The supernatant containing the bacterically expressed Rap1 was concentrated using a Centricon 10 (Amicon).

The Rap1 protein (∼1 μg) was incubated with either 1 mM GDP or 1 mM GTP in 20 mM Tris-HCl (pH 7.6); 10 mM EDTA; 5 mM MgCl₂; 1 mM DTT(Dithiothreitol); 0.1 mM PMSF and 10% (v/v) glycerol in a total volume of 20μl for 30 minutes at 30°C, and MgCl₂ was then added to a final concentration of 20 mM to stabilize the binding. 210 ng Rap1 protein was then incubated with 100 ng of GST-RalGDS(RBD) bound to glutathione-Sepharose beads in 20 μl of binding buffer containing 100 μg BSA, 100 mM NaCl, 0.5 mM GTP or GDP, 6.5 mM EDTA, 12.5 mM MgCl₂. After 2 hours at 4°C, the beads were pelleted and the unbound material in the supernatant was removed. The beads were washed five times with 0.5 ml of ice-cold 10 mM Tris-HCl, pH 7.6; 5 mM MgCl₂; 1 mM DTT; 0.1 mM PMSF; 10% glycerol and 0.05% Triton X-100. An equal volume of SDS sample buffer was added to both the beads and the unbound fraction. Both fractions were subjected to SDS-PAGE and western blotting as described above.

Samples of Dictyostelium cells were pelleted and resuspended in 30 mM HEPES (pH 7.8), 10 mM KCl, 10% sucrose, 1% TritonX-100 and protease inhibitor (Roche). The protein concentration was determined using the BioRad protein assay. Cell lysate (5-100 μg protein) was incubated for 1 hour at 4°C with 20 μg GST-RalGDS(RBD) that had been pre-bound to glutathione-Sepharose beads. After three washes with 30 mM HEPES (pH 7.8), 200 mM KCl, the beads were incubated with an equal volume of SDS sample buffer and the solubilized material subjected to SDS-PAGE and western blotting.

All six independently isolated transformants that contained the folate repressible pVEII-AS5′ construct could be maintained in HL-5 media supplemented with 1 mM folate and 10 μg/ml G418 for as long as two months without loss in viability. However, the tranformants grew progressively more slowly upon removal of folate. Fig. 1A shows the growth of one of these transformants, pVEII-AS5′-1, following folate removal. Growth was initially normal, but after cells were passaged into fresh media their growth rate declined and growth terminated after ∼10 days of incubation. Cell viability started to decline after 4 days of incubation in the absence of folate and by 12 days only 30% of the cells were still viable(Fig. 1B). After 15 days in the absence of folate, no viable cells were detected (data not shown). The growth of cells that had been transformed with the pVEII vector in HL-5 media containing 10 μg/ml G418 was identical to that of the parental Ax-2.

Rap1 protein levels in the pVEII-AS5′-1 transformant were found to gradually decline during growth without folate, reaching ∼20% of their initial value by 10 days and being barely detectable beyond 12 days(Fig. 2). These results indicate that the reduction in Rap1 protein level correlated with reduced cell growth. A similar decrease in viability and Rap1 protein level was observed in the two other pVEII-AS5′ transformants that were examined in detail (data not shown). As a control, samples were also assayed for RasG by western blot analysis, as this protein is another member of the Dictyostelium Ras subfamily and any non-specific effect of the rapA antisense construct would be expected to be manifested as a drop in RasG level. Only a slight decrease in RasG was observed during the course of the experiment.

During the period of declining growth in axenic media, the cells were tested for growth on bacterial lawns. The size of the plaques formed on the bacteria were considerably smaller as the time of depression increased,suggesting that growth on bacteria was also impaired. Both pinocytosis and phagocytosis were considerably reduced in the pVEII-AS5′-1transformant (data not shown), which may account for the slow growth rate in both axenic media and on bacteria. In addition to the defects in cell growth,the majority of pVEII-AS5′-1 cells were smaller than the parental Ax-2 cells, although about 5% of the cells were considerably larger(data not shown).

In view of the effect of Rap1 depletion on cell viability and cell size, we determined the resistance of the Rap1-depleted cells to hyperosmotic conditions. To ensure that the cells used in these tests were viable, we used cells that contained ∼40% of the normal level of Rap1. When these partially Rap1-depleted cells were exposed to 0.4 M sorbitol for 120 minutes,only ∼15% of the transformant cells survived. By contrast, under identical conditions all wild-type Ax-2 cells remained viable(Fig. 3A).

Dictyostelium cells respond to hyperosmolarity by reducing their cell volume by 50% within 5 minutes(Zischka et al., 1999). This rapid reduction in cell volume correlates with myosin phosphorylation and the activation of guanylyl cyclase (Kuwayama et al., 1996). The importance of the guanylyl cyclase activation in osmoregulation is indicated by the observation that some of the osmosensitive mutants that have been identified are deficient in cGMP production. Furthermore, these mutants become less osmosensitive in the presence of the cell-permeable cGMP analog, 8-Br-cGMP(Kuwayama et al., 1996). We found that cGMP levels were reduced in the pVEII-AS5′-1transformant relative to the wildtype following treatment with 0.4 M sorbitol(Fig. 3B), suggesting a possible requirement for Rap1 in the accumulation of cGMP. Furthermore,addition of 8-Br-cGMP to Rap1 depleted cells enhanced their survival following sorbitol treatment (Fig. 3A). These results suggest that Rap1 might be important in regulating the accumulation of cGMP during the response to osmotic shock. Consistent with this idea was the finding that cells overexpressing the constitutively activated Rap1, Rap1(G12V), produced more cGMP in response to sorbitol addition than did wild-type cells (Fig. 3B).

During osmotic shock, there is also a dramatic increase in actin tyrosine phosphorylation (Zischka et al.,1999). The data in Fig. 4 shows that the expected increase in actin tyrosine phosphorylation occurred in wild-type cells in response to osmotic shock. Under these conditions, actin is by far the major protein to be tyrosine phosphorylated (Zischka et al.,1999). However, in pVEII-AS5′-1 cells the basal pre-shock level of actin tyrosine phosphorylation was considerably higher than the pre-shock level in wild-type cells, and actin phosphorylation increased only slightly after osmotic shock, suggesting actin tyrosine phosphorylation was deregulated in Rap1-depleted cells.

To ascertain if Rap1 is activated in response to hyperosmolarity, we adopted the GST-RalGDS(RBD) pull-down assay that had been used to determine Rap1 activation in mammalian cells (Franke et al., 1997). The rationale for this approach was the fact that Dictyostelium Rap1 and mammalian Rap1 have identical effector domains(Robbins et al., 1990). To demonstrate that RalGDS(RBD) binds specifically to the activated Dictyostelium Rap1, we compared the binding of bacterially expressed GST-RalGDS(RBD) fusion protein to Rap1 in lysates of wild type cells with Rap1 in lysates of cells overexpressing the constitutively activated Rap1 (G12V). As shown in Fig. 5A, there was appreciably more RalGDS(RBD)-bound Rap1 in the lysates from the rap1(G12V) transformant than in the lysate from the wild-type cells relative to the total amounts of Rap1 present in these extracts. These results indicated that RalGDS(RBD) preferentially bound to activated Rap1.

To determine if there was any binding of RalGDS(RBD) to GDP-bound Rap1,bacterially expressed Rap1 was equilibrated with 1 mM GTP or 1 mM GDP prior to binding to RalGDS(RBD). The proportion of Rap1 that binds to RalGDS(RBD) is high in the presence of GTP but low in the presence of GDP(Fig. 5B). Rap1 did not bind to the control GST protein. Hence, interaction with mammalian RalGDS (RBD) is clearly a good measure of the amount of activated Rap1 in the Dictyostelium cell.

Vegetative Ax-2 cells were exposed to 0.4 M sorbitol, and cell lysates were incubated with GST-RalGDS(RBD) bound to glutathione-Sepharose beads. As shown in Fig. 6, the amount of Rap1 bound to RalGDS(RBD) increased within 5 minutes of exposure to 0.4 M sorbitol. The extent of Rap1 activation was reproducibly three-to-four-fold, and there was a slight but consistent decrease in bound Rap1 during the next 5 minutes(Figs 6 and7). Levels of bound Rap1 increased again as the sorbitol shock continued (Figs6 and7). The initial kinetics of Rap1 activation correlated reasonably well with the increase in the cGMP level(Fig. 3B), a result consistent with the possibility that Rap1 regulates the pathway that leads to the activation of guanylyl cyclase. Treatment of cells with 2 mM EDTA enhances the cGMP response to osmotic shock (Oyama,1996), and we found that treatment of wild-type cells with 2 mM EDTA prior to osmotic shock also produced an enhanced activation of Rap1(Fig. 7). The addition of EDTA alone had no effect on Rap1 activation (data not shown).

To determine whether Rap1 activation in response to hyper-osmotic conditions was a general or specific response to shock, vegetative Ax-2 cells were subjected to three additional stress conditions: low temperature, high temperature and hypo-osmotic conditions. When cells were switched from 22°C to 8°C, Rap1 activation increased after approximately 5 minutes,and the level of activation remained high for 20 minutes(Fig. 8). However, neither a switch to 30°C or resuspension in H₂O had a noticeable effect on the level of activated Rap1 (Fig. 8), indicating that activation of Rap1 is not part of a general stress response.

Since previous attempts at isolating Dictyostelium cells with a disrupted rapA gene were unsuccessful (P. J. Rebstein, PhD thesis,University of British Columbia, 1996) (X. Insall, personal communication), we expressed the antisense RNA from the rapA gene to reduce the cell content of Rap1. The data presented here are consistent with the idea that Rap1 is essential. As levels of Rap1 in a pVEII-AS5′transformant were gradually reduced, there was a decrease in growth rate and cell viability. It is difficult to pinpoint the actual cause of cell death,but two distinct possibilities are consistent with our data. It had been shown that Rap1 is involved in endocytosis(Seastone et al., 1999), and since the rates of endocytosis in the pVEII-AS5′ transformant are low (data not shown), the reduced uptake of nutrients could account for the reduced rates of growth and might ultimately lead to cell death. Alternatively, since the overexpression of Rap1 or Rap1-G12V generates alterations in cell shape and contractile responses(Rebstein et al., 1993;Rebstein et al., 1997;Seastone et al., 1999), the lack of Rap1 could impair cytoskeletal function.

The properties of the pVEII-AS5′-1 Rap1-depleted cells resemble those of a strain that lacks both α-actinin and gelation factor, two F-actin crosslinking proteins(Rivero et al., 1996). This double mutant grows slowly, exhibits reduced phagocytosis and is more sensitive to osmotic shock. In addition, most of the double mutant cells are smaller and more rounded than wild-type cells, although approximately 5% of the population are larger (Rivero et al.,1996). All of these characteristics are shared by the Rap1-depleted cells. In view of this similarity, we determined the amounts ofα-actinin and gelation factor in depleted cells. Gelation factor levels were normal, but α-actinin levels were reduced (data not shown). However, cells disrupted only in the gene encoding α-actinin exhibit normal properties (Rivero et al.,1996). Thus, either essential regulatory pathways controlled by Rap1 and by α-actinin and gelation factor intersect at some point or there are multiple means of generating the phenotype seen in Rap1-depleted cells. In any case, it is unlikely the deleterious effects resulting from Rap1 depletion are solely caused by the regulation of α-actinin.

The activation of guanylyl cyclase is important for the response to hyperosmolarity (Kuwayama et al.,1996; Oyama,1996), and the data we have provided indicate that Rap1 plays a role in the activation of guanylyl cyclase. In particular, there was a reduced accumulation of cGMP in cells expressing rapA antisense RNA in response to osmotic shock, and the osmosensitivity of these cells was partially reduced by the addition of the cGMP analog 8Br-cGMP. In addition, we showed that Rap1 was activated in response to osmotic shock during cGMP accumulation. These results suggest that Rap1 acts upstream in the pathway that transmits the signal responsible for guanylyl cyclase activation. Consistent with this interpretation, EDTA treatment, which stimulates cGMP production in response to osmotic stress(Oyama, 1996), also stimulated the activation of Rap1. The activation of guanylyl cyclase in response to osmotic shock does not appear to involve the heterotrimeric G protein complex,as gβ-null cells have no deficiency in cGMP accumulation(Kuwayama and van Haastert,1998). However, GTPγS stimulates guanylyl cyclase in vitro(Janssens et al., 1988;Janssens et al., 1989) and in electro-permeabilized cells (Schoen et al., 1996), suggesting an interaction between guanylyl cyclase and a GTP-binding protein. Rap1 could be this presumptive GTP-binding protein.

A putative intracellular histidine kinase, encoded by the dokAgene is important in osmoregulation, demonstrated by the fact that a dokA-null strain exhibits increased sensitivity to osmotic shock(Schuster et al., 1996). However, the dokA-null strain exhibits normal cGMP accumulation in response to osmotic shock (Schuster et al., 1996), and there is now evidence that DokA acts in a signaling pathway parallel to the cGMP pathway(Ott et al., 2000). We found that Rap1-activation in the dokA-null mutant was similar to that in the parental Ax-2 strain during the osmotic shock (data not shown), consistent with the idea that Rap1 acts upstream of guanylyl cyclase. Although Rap1 depletion clearly affected actin tyrosine phosphorylation, there is no evidence as yet to indicate that this phosphorylation is dependent on guanylyl cyclase activation.

Rap1 activation does not occur under all stress conditions, as shown by the fact that there was no activation in response to low osmolarity or to high temperature (30°C). However, Rap1 was activated in response to low temperatures (8°C). These results indicate that Rap1 is not activated as part of a general stress response, signal transduction pathway.

An intriguing question is whether Rap1 has a function during Dictyostelium differentiation. It had been shown previously that the overexpression of Rap1 during development was capable of partially reversing the developmental defects produced by activated RasD, suggesting a possible role for the protein during development(Louis et al., 1997). However,other than the fact that the Rap1 protein levels remain constant during development (S. M. Robbins, PhD thesis, University of British Columbia, 1991),nothing more is known about a possible developmental function. Rap1-depleted cells do exhibit delayed differentiation (data not shown), but this may simply be caused by the reduced viability of these cells.

This work was supported by grants from the Medical Research Council of Canada to G.B.S. and CIHR to G.W. We thank Richard Pachal for constructing the gst-rap1 plasmid.