Inactivation of Two Dictyostelium discoideum Genes, DdPIK1 and DdPIK2, Encoding Proteins Related to Mammalian Phosphatidylinositide 3-kinases, Results in Defects in Endocytosis, Lysosome to Postlysosome Transport, and Actin Cytoskeleton Organization

The D. discoideum wild-type strain Ax3 and the double null mutant, Δddpik1/ddpik2, were grown axenically in HL5 medium attached to plastic T175 tissue culture flasks. The construction of the mutant strain is detailed elsewhere (Zhou et al., 1995).

Wild-type and Δddpik1/ddpik2 cells (3 × 10⁶) were solubilized in 10 μl of 0.1% Triton X-100 containing 50 μg/ml leupeptin, 1 mM PMSF, 2 mM benzamidine, and 10 μg/ml pepstatin. 1 μl (3 × 10⁵ cells) was used in each reaction. PI 3-kinase assays were performed basically as described (Whitman et al., 1988; Stack et al., 1993). The 50-μl reactions contained 20 mM Hepes, pH 7.5; 60 μm ATP, 10 mM MgC₂, 0.2 mg/ml sonicated PI, PI(4)P or PI(4,5)P₂ and 0.2 mCi/ml (γ ³²-P) ATP. The reactions were incubated at 25°C for 5 min and stopped by the addition of 80 μl of 1 M HCl. Lipids were extracted by the addition of 160 μl of CHCl₃:MeOH (1:1) and vortexing for 30 s. The organic phase was removed and 10-μl aliquots were dried under vacuum. The dried lipids were resuspended in 10 μl chloroform and spotted on a silica 60 TLC plate. Samples were separated using the borate buffer system as described (Walsh et al., 1991).

Fluid phase uptake, exocytosis, and flux was measured using FITC-dextran (FD) as previously described (Temesvari et al., 1996b). Measurement of pH was performed also as described (Padh et al., 1993) using FD.

Fluorescently labeled (crimson), carboxylate modified, latex beads (1 μm diameter; Molecular Probes, Eugene, OR) were added to cells in growth media at a concentration of 100 beads/cell. At the times indicated, 1 ml of cells were removed and processed by centrifugation as described (Temesvari et al., 1996b). Fluorescence was measured using a fluorometer set at an excitation wavelength of 625 nm and an emission wavelength of 645 nm. The number of beads internalized was calculated from a standard curve obtained by measuring the fluorescence of a known number of beads.

Radiolabel pulse-chase analysis was performed as follows. Exponentially growing cells were collected by centrifugation (1,000 g for 3 min) and resuspended to a titer of 10⁷ cells/ml in fresh growth media containing 650– 750 μCi/ml [³⁵S]methionine (New England Nuclear, Boston, MA). The cells were pulsed for 20–30 min while shaking at 150 rpm at 20°C, harvested, and resuspended in fresh media. At the times indicated, 5–7 × 10⁶ cells were removed and separated into cell and extracellular media fractions by centrifugation (1,000 g for 30 s). The pellets were solublized with 0.5% Triton X-100. α-Mannosidase was immunoprecipitated from cell pellet and media samples using monoclonal antibodies as described (Mierendorf et al., 1985), and immunoprecipitated proteins were separated on 7.5% polyacrylamide gels according to the procedure of Laemmli (1970). The gels were placed in 10% trichloroacetic acid (TCA) overnight to precipitate the proteins in the gels, treated with Enhance according to the manufacturer's instructions (New England Nuclear, Boston, MA), dried under vacuum, and exposed to Kodak XAR-5 film at −70°C.

Immunostaining of the contractile vacuole network was performed as described (Bush et al., 1994) using monoclonal antibodies to the 100-kD subunit of the vacuolar proton pump protein (a kind gift of Dr. Agnes Fok). The secondary antibodies used were goat anti–mouse IgG coupled with FITC. To visualize F-actin, cells grown overnight on coverslips in HL5 were washed twice with PBS (137 mM NaCl, 27 mM KCl, 10 mM Na₂HPO₄, 2 mM KH₂PO4, pH 7.4) and fixed in 3.7% formaldehyde in PBS for 10 min at room temperature. The coverslips were then washed twice with PBS and the cells were extracted in a solution of cold (−20°C) acetone for 3–5 minutes and air dried. FITC-labeled phalloidin solution (Molecular Probes) was evaporated under vacuum, resuspended in 200 μL of PBS and pipetted onto the coverslips which were incubated at room temperature for 20 min. The coverslips were then washed twice with PBS and mounted on glass slides using one drop of mounting media (10 ml PBS, 90 mL glycerol with 100 mg p-phenylenediamine) to prevent quenching of fluorescence. Photographs were taken with an Olympus BX50 fluorescence microscope using T-MAX 400 speed film.

Sterile, one centimeter square, glass coverslips were inserted into T25 tissue flasks and submerged in 10 ml of HL5 growth medium. Cells were seeded in the flask at an initial titer of 10⁵ cell/ml and were grown for 24 h to allow cell attachment. The coverslips with attached cells were removed and fixed in 1.25% glutaraldehyde buffered in 50 mM sodium cacodylate at pH 6.8. The cells were dehydrated in a graded acetone series and critical point dried from liquid carbon dioxide. The coverslips with adhering dried cells were mounted on specimen stubs, sputter coated with 10 nM of gold and viewed with a ISI DS 130 scanning electron microscope.

Wild-type and mutant cells were pulsed for 2 h with iron dextran (prepared as described; Rodriguez-Paris et al., 1993), washed, and fixed overnight at 4°C in 2.5% glutaraldehyde buffered with 0.1 M sodium cacodylate (pH 6.8). After fixation in glutaraldehyde, the cells were washed with 0.1 M cacodylate buffer, postfixed for 2 h on ice in 2% osmium tetroxide buffered with 0.1 M cacodylate (pH 6.8), washed again with 0.1 M cacodylate buffer, and dehydrated in a graded series of ethanol solutions. The dehydrated samples were infiltrated overnight in a 1:1 mixture of 100% ethanol and Polybed 812 (Polysciences, Warrington, PA) and then embedded in Polybed 812 by standard procedures. Thin sections were stained with uranyl acetate and lead citrate and viewed with a Philips CM-10 transmission electron microscope.

In vitro PI phosphorylation assays were performed to measure the relative intracellular concentrations of different PI kinases in wild-type Ax3 and Δddpik1/ddpik2 cells. Cells grown in tissue culture flasks were harvested by centrifugation, and lysed in 0.1% Triton X-100. Extracts prepared from the equivalent of 3 × 10⁵ cells were mixed with PI and γ labeled ATP. After incubation at 25°C for 5 min, the PI lipids were extracted and resolved by TLC. The only reaction products identified using protein from Ax3 extracts were PI(3)P and PI(4)P (Fig. 1; lane 1). Even with longer exposures times, no labeled PI(4,5)P₂ or PI(3,4,5)P₃ were detected. These later PI products were also not observed by modifying reaction times, by changing the concentration of protein added from the cell extracts, or by adding PI(4)P or PI(4,5)P₂ substrates. In addition, the level of PI 3-kinase activity was not altered in wild-type or mutant strains by generating extracts by freeze thaw in the absence of detergent; detergent has been reported to effect PI 3-kinase activity. Interestingly, the TLC profile of Dictyostelium phosphorylated PIs was nearly identical to the yeast phosphorylated PI generated in vitro; only PI(3)P and PI(4)P were observed using PI as a substrate (Fig. 1; lanes 2 and 3). Finally, no significant reduction in the levels of PI(3)P and PI(4)P was observed using extracts from Δddpik1/ddpik2 cells (lanes 4 and 5). We conclude that disrupting DdPIK1 and DdPIK2 does not alter the measurable levels of functional PI 3-kinase or PI 4-kinase using PI as a substrate. Because we cannot detect PI (3,4)P₂ or PI (3,4,5)₃ in our assays, we cannot address the effect of removing DdPIK1 and DdPIK2 on the synthesis of these lipids.

Δddpik1/ddpik2 cells do not grow axenically in shaken suspension culture, and only grow slowly in liquid medium while attached to solid surfaces such as plastic (Zhou et al., 1995). The slow growth rate of this strain indicates a possible defect in endocytosis; therefore, we examined the role of the DdPIK1 and DdPIK2 gene products in the endocytic pathway. To measure pinocytosis, logarithmically growing wild-type and mutant cells were exposed to FD for 60 min. At the times indicated in Fig. 2 A, cell samples were removed, washed, lysed with detergent, and the intracellular fluorescence measured. The level of internalized FD was standardized to total cell protein to correct for differences in cell size. Δddpik1/ddpik2 cells internalized FD at a greatly reduced rate compared to wild-type Ax3 cells. For instance, after 60 min (internalization was linear over this period), the control cells internalized 3–4-fold higher levels of FD than that internalized by mutant cells. The rate of uptake of FD was reduced over the first 5 min of internalization (results not shown). Taken together, these results suggest that the DdPIK1 and DdPIK2 gene products may regulate the earliest steps in pinocytosis.

An alternative explanation for the lower levels of fluorescence in the mutant cells is that the FD was internalized at the same rate in all cultures but the fluid phase marker was exocytosed at a faster rate from mutant cells as compared to wild-type cells. To test this possibility, we measured the rate of release of FD from cells loaded with this fluid phase marker to steady state (influx rates equal efflux rates). Wild-type and mutant cells were exposed to FD for 3 h, washed, and resuspended in FITC-free media and at various times, cell samples were removed and processed for fluorimetry as described in the Materials and Methods section. After 90 min, ∼25% of the internalized FITC-dextran was retained in the control cells whereas the mutant cells retained 70% of the FITC-dextran, indicating a block in exocytosis (Fig. 2 B). Therefore, the reduced levels of fluorescence observed in the pinocytosis assays cannot be accounted for by an increased rate of exocytosis from the cells. Instead, the data indicate that DdPIK1 and DdPIK2 encoded products regulated an early step in pinocytosis and perhaps a later step in the endocytic pathway leading to the efflux of fluid phase.

The mammalian p85/p110 PI 3-kinase has been reported to play a role in phagocytosis in mammalian cells (Ninomiya et al., 1994). To determine if Δddpik1/ddpik2 cells were normal in phagocytosis in Dictyostelium, exponentially growing wild-type cells and mutant cells were exposed to crimson-labeled 1-μm latex beads for 90 min. At various times, cells were removed, washed, lysed, and the intracellular fluorescence was measured. Δddpik1/ddpik2 cells internalized beads (on a per protein basis) at a rate that was higher than wild-type cells throughout the time course; however, these rates were not significantly different (Fig. 2 C). A similar result was observed when FITClabeled bacteria were substituted for beads (results not shown). The phagocytosis results were somewhat surprising since this process, like pinocytosis, is a form of endocytosis, and pinocytosis was severely reduced in the Δddpik1/ ddpik2 strains. Therefore, these results indicate that the processes of phagocytosis and pinocytosis may be independently regulated processes in Dictyostelium.

The results presented in the previous section suggest that DdPIK1 and DdPIK2 gene products regulated an early step in pinocytosis and a later step in the endosomal pathway. In D. discoideum, fluid phase reaches acidic lysosomes 15–20 min after internalization, and begins to be exocytosed from nonacidic postlysosomal vacuoles after a 45– 60-min lag; no early recycling compartment has been identified in Dictyostelium. If DdPIK1 and DdPIK2 regulate fluid phase traffic at a stage like the lysosome to postlysosome step, then fluid phase should accumulate in acidic lysosomal compartments in the mutant cells. Therefore, in mutant cells, loaded with FD to steady state levels, significantly less FD-containing postlysosomal vacuoles and proportionally more lysosomes should be observed.

To test this, wild-type cells and Δddpik1/ddpik2 cells were loaded to steady state with FD for 3 h, washed, placed on glass coverslips, and examined using fluorescence microscopy. Essentially all control cells contained numerous small, FD-containing vesicles (less than 1 μm in diameter; closed arrow) as well as a few large (greater than 2 μm; open arrowhead) FD-containing postlysosomes (Fig. 3). Cells were also exposed to acridine orange (AO), a substance that accumulates to high concentrations in acidic compartments and fluoresces red-orange. As indicated in Fig. 3; upper right panel, AO accumulated in wildtype cells in greater than 90% of the small, FD-positive vesicles (closed arrow), indicating these vesicles were acidic in nature whereas the large postlysosomes did not accumulate this dye consistent with the lumens of these vacuoles being non-acidic. In contrast, the mutant cells (Fig. 3; bottom panels) contained numerous FD-dextran containing vesicles that accumulated AO but less than 10% of these cells contained FD positive vacuoles larger than 2.5 μm in diameter. Furthermore, greater than 90% of the FD positive vesicles in mutant cells were acidic suggesting that DdPIK1 and DdPIK2 do not play a major role in regulating traffic from less acidic prelysosomal vesicles to lysosomes.

To determine if Δddpik1/ddpik2 cells lacked postlysosomes or instead contained these vacuoles, but these cells were unable to accumulate FD, control and mutant cells were fed iron dextran for 2 h in growth medium. The cells were isolated and prepared for thin section transmission electron microscopy. Fig. 4, top panels, show an electron micrograph of a thin section of two representative wild-type cells that contained 2–3 large (greater than 2 μm in diameter) iron dextran containing postlysosomes (closed arrow) with other smaller iron dextran-containing lysosomes. Greater than 90% of the wild-type Ax3 cells contained at least one large postlysosomal vacuole and 20–30% of the cells contained 3–6 postlysosomes. In contrast, greater than 90% of the mutant cells completely lacked the large postlysosomal vacuoles found in the control cells and instead contained an increased number of smaller iron positive vesicles (Fig. 4, bottom panels). The absence of postlysosomes and the presence of numerous, smaller lysosomes in the mutant cells, supports the hypothesis that DdPIK1 and DdPIK2 proteins regulate the formation of postlysosomes from lysosomes.

α-Mannosidase is synthesized as a 140-kD precursor that is processed to an 80-kD intermediate polypeptide and a 58-kD mature subunit; the 80-kD form is processed in lysosomes to the other 60-kD subunit (Cardelli, 1993). Mature lysosomal enzymes are eventually secreted from postlysosomal vacuoles. To determine if trafficking from the secretory pathway to lysosomes (the biosynthetic targeting pathway) was normal in Δddpik1/ddpik2, radiolabel pulse-chase experiments were performed. Logarithmically growing cells were pulsed with [³⁵S]Met for 20 min, centrifuged, and resuspended in label free medium. At various times, samples were removed and separated into cell pellets and extracellular supernatants by centrifugation. α-Mannosidase was immunoprecipitated from both cells and supernatants using the monclonal antibody, 2H9, and the immunoprecipitates were separated by SDS-PAGE and analyzed by fluorography. As shown in Fig. 5 a and b, the newly synthesized 140-kD α-mannosidase precursor was processed to the 60- and 58-kD mature forms in wild-type and mutant cultures with a half time of ∼30–40 min. Since the final proteolytic processing event occurs in lysosomes, this result suggests that DdPIK1 and DdPIK2 do not play a role in regulating transport of hydrolases from the ER to lysosomes. Furthermore, in both control cells and the mutant cells <5% of the α-mannosidase precursors were secreted from either strain. This result suggests that DdPIK1 and DdPIK2 did not regulate hydrolase targeting to lysosomes. The newly synthesized precursors that were secreted, exited control and mutant cells by 10 min of chase further suggesting that DdPIK1 and DdPIK2 did not regulate membrane trafficking from the ER to the cell surface along the constitutive secretory pathway.

20% of the mature radiolabeled α-mannosidase was secreted from control cells by 120 min of chase, and by 300 min of chase, 70% of the enzyme had been exocytosed (Fig. 5,a). In contrast, by 300 min of chase <5% of the mature radiolabeled enzyme was secreted from the mutant cells (Fig. 5 b). This result supports the fluid phase exocytosis data presented above and suggests that the processed mature lysosomal enzymes may not reach the secretory postlysosomal vacuoles in mutant cells.

The vacuolar proton pump populates both the endocytic pathway and the contractile vacuole (CV) system of membranes in Dictyostelium (Temesvari et al., 1994; Bush et al., 1994; Fok et al., 1994; Heuser et al., 1993; Nolta and Steck, 1994). The CV complex consists of a reticular network of membranes and a few large associated bladderlike vacuoles (Bush et al., 1994; Nolta and Steck, 1994). Recent evidence also suggests that the vacuolar proton pump regulates the structure and function of the CV and endosomal pathway and that these two pathways may be functionally connected (Bush et al., 1996; Temesvari et al., 1996b). Conceivably, PI 3-kinases could regulate membrane flow between the CV system and endosomal pathway. Therefore, to determine if the organization of the CV network was normal in Δddpik1/ddpik2 cells, permeabilized wild-type and mutant cells were incubated with monoclonal antibodies directed against the 100-kD subunit of the V H⁺-ATPase followed by the addition of secondary antibodies coupled to FITC. Fluorescence microscopy indicated that all cells contained a reticular network enriched in the 100-kD proton pump subunit (Fig. 6, B and D). In addition to the reticular network, large proton pump positive vacuoles were seen in control and mutant cells (Fig. 6). We have also observed that mutant cells were normal in osmotic regulation (results not shown) suggesting that DdPIK1 and DdPIK2 do not regulate the function of the CV network. Together, these results suggest that DdPIK1 and DdPIK2 proteins do not play a major role in the regulation of the organization or function of the CV complex.

A number of recent reports have implicated the PI 3-kinase p85/p110 in controlling membrane ruffling and cell shape (Wennstrom et al., 1994), perhaps through activation of Rac GTPases (Hawkins et al., 1995) which in turn regulate changes to the actin cytoskeleton. The following experiments provide evidence that in Dictyostelium, the putative PI 3-kinases DdPIK1 and DdPIK2 may also regulate these processes. First, control cells grown overnight on plastic appeared amoeboid in shape when examined by phase contrast microscopy (Fig. 7,A). In contrast, the mutant cells appeared as small, irregularly shaped cells with jagged edges around their periphery (Fig. 7,B). Interestingly, 1–3% of the mutant cells contained long cellular bridges (inset Fig. 7 B) that connected cells. These structures were stable for at least 1 h and probably represent cellular material shared by cells unable to efficiently complete fission. It remains to be demonstrated that these mutant cells are actually defective in cytokinesis.

Secondly, scanning electron micrographs of control cells revealed amoeboid-like cells with a smooth ruffled exterior (Fig. 8,A) similar to what has been observed previously (Novak et al., 1995). Some filopodia were seen but these structures were usually short (<2 μm in length) and mainly found in areas where the cell was attached to the surface. In contrast, scanning EM of mutant cells demonstrated that these cells were smaller than wild-type and more irregularly shaped (assuming a “puzzle piece” morphology), and the cells contained numerous filopodia (Fig. 8 B) extending along both the plane of attachment of cells to the surface (closed arrowhead) and outward from the surface of cells (open arrowhead).

To determine the distribution of F-actin in control and mutant cells, we stained detergent permeabilized cells with FITC-labeled phalloidin, a compound that tightly binds F-actin. As seen in Fig. 9, A and B, F-actin in the control cells distributed primarily in a cortical pattern (open arrowhead) with only a few actin rich filopodia being observed (Fig. 9,B). In mutant cells (Fig. 9, C and D), there was a decrease in the overall staining of cortical actin and a striking increase in the amount of actin-rich filopodia. Most of the remaining F-actin was found in patches around the cells and in the long cellular bridges described earlier (results not shown). In summary, compared to wild-type cells, mutant cells were smaller, more irregularly shaped and contained long cellular bridges that connected cells in the process of division. In addition, these cells contained numerous, actin-rich filopodia coupled with a decrease in cortical actin staining. Taken together, these results suggest DdPIK1 and DdPIK2 regulate cell morphology and organization of the actin cytoskeleton in Dictyostelium.

During the aggregation phase of development, Dictyostelium cells chemotax toward cAMP, resulting in the formation of a multicellular organism containing up to 10⁵ cells. This process is regulated through G protein–coupled, cell surface receptors that activate downstream effector pathways, including the activation of guanylyl cyclase, adenylyl cyclase, and aggregation-stage gene expression (Firtel, 1995; Parent and Devreotes, 1996). To determine if chemotactic movement was altered in Δddpik1/ddpik2 cells, we examined the ability of aggregation-competent cells to chemotax toward a micropipette containing the chemoattractant cAMP (Gerisch et al., 1995) (see legend to Fig. 10 for details). As can be observed by comparing the movement of Δddpik1/ddpik2 null to wild-type cells in Fig. 10,A, Δddpik1/ddpik2 cells chemotax significantly more rapidly toward a pipette tip emitting cAMP than do wild-type cells. To examine this in more detail, we followed the ability of individual cells to undergo chemotaxis toward cAMP and then to reverse direction as the position of the micropipette was changed. Fig. 10,B shows this analysis for Δddpik1/ddpik2 cells. As is observed, Δddpik1/ ddpik2 cells rapidly respond both to the initial insertion of the micropipette (Fig. 10,B, a–e) and when the position of the tip was placed on the other side of the cells (Fig. 10,B, f–i). Comparison of these images to the movement of wildtype cells toward the pipette (Fig. 10,C, a–f) and as the position of the needle is moved (Fig. 10,C, g–i) shows that the wild-type cells respond and move more slowly. (Note Δddpik1/ddpik2 cell images are every 1 min while those of the wild-type cells are every 2 min on the initial movement. See legend to Fig. 10 for details.) Δddpik1/ddpik2 cells show rapid reorientation and movement when the tip is moved, while, within the similar period of time, wildtype cells had just started to reorient themselves. These results indicate that both the response to the chemoattractant and the rate of chemotaxis are faster in the double knockout strain.

In this report, we used cell biological and biochemical approaches to demonstrate that the DdPIK1 and DdPIK2 PI 3-kinase–related gene products regulated multiple processes in Dictyostelium discoideum. Δddpik1/ddpik2 cells internalized fluid phase at 20–25% of the rate of wild-type cells; the efflux rate of internalized fluid phase was also greatly reduced in the mutant cells. The lysosome to postlysosome transport step appeared to be delayed in Δddpik1/ddpik2, accounting for the reduction in exocytosis of fluid phase in mutant cells. Phagocytosis, lysosomal enzyme targeting, and CV structure and function were normal in mutant cells demonstrating the specificity of action of the enzymes. Finally, DdPIK1 and DdPIK2 also appeared to play a role in regulating cell shape and cell movement perhaps by regulating the actin cytoskeleton.

Five genes encoding proteins related to PI kinases on the basis of amino acid sequence homology have been cloned from Dictyostelium. Three of these genes, DdPIK1, DdPIK2, and DdPIK3, encode proteins most related to the mammalian 110-kD PI 3-kinase subunit (Zhou et al., 1995). The strain characterized in this report Δddpik1/ddpik2 is disrupted in two of these genes, DdPIK1 and DdPIK2, and grows slowly in axenic medium when attached to plastic; the strain does not grow at all in suspension in axenic medium (Zhou et al., 1995). Single knockout strains grow normally with no observable defects. This suggests that the DdPIK1 and DdPIK2 gene products have overlapping or redundant functions, therefore, requiring that both genes be disrupted before an effect is observed. The rate of fluid phase uptake was much less in the mutant cells as compared to wild-type cells. This might account for the growth defects observed for the mutant since nutrients are internalized from growth medium by pinocytosis in Dictyostelium.

Disruption of the DdPIK1 and DdPIK2 genes did not significantly reduce the levels of functional enzymes capable of phosphorylating PI in vitro to generate PI(3)P. The levels of mRNA encoding DdPIK1 and DdPIK2 are extremely low, and it is possible that the loss of these enzymes in the mutant would not be detected even if these enzymes use PI as a preferred substrate. Furthermore, using PI(4)P and PI(4,5)P₂ as substrates we were still only able to identify PI(3)P and PI(4)P products following in vitro kinase assays, therefore, we cannot determine if the levels of PI 3-kinase enzymes recognizing PI(4)P and PI(4,5)P were reduced in Δddpik1/ddpik2. Based on amino acid sequence homology, however, DdPIK1 and DdPIK2 are more closely related to mammalian p85/p110 PI 3-kinase than to yeast Vps34 (only recognizing PI as a substrate). Therefore, we speculate that disruption of DdPIK1 and DdPIK2 reduce the levels of PI 3-kinase enzymes capable of recognizing substrates like PI(4,5)P₂ and that reduced levels of PI(3,4,5)P₃ may account for the phenotypic changes presented here.

The reduction in the rate of fluid phase internalization was evident at the earliest time points examined suggesting that a very early step in pinocytosis (perhaps the internalization step) was regulated by DdPIK1 and DdPIK2. Consistent with our results, two recent reports have indicated that a PI 3-kinase sensitive to low levels of wortmannin (hypothesized to be p85/p110) also played a role in pinocytosis in mammalian cells (Li et al., 1995; Clague et al., 1995). Although it has been reported that the clathrin heavy chain is involved in pinocytosis in Dictyostelium (O'Halloran and Anderson, 1992; Ruscetti et al., 1994), recent studies examining endocytosis in yeast (Kubler and Riezman, 1993) and in Dictyostelium (Temesvari, L., and J. Cardelli, unpublished results) indicate that actin may also be critical in regulating pinocytosis. In support of this, influx of fluid phase is regulated by select myosin I proteins that are known to interact with and regulate the actin cytoskeleton (Luna and Condeelis, 1990; Novak et al., 1995; Temesvari et al., 1996a).

It has been reported that the p85/p110 PI 3-kinase regulates phagocytosis in mammalian cells (Ninomiya et al., 1994); however, phagocytosis of beads and bacteria were normal in the mutant cells. It is possible though that other Dictyostelium PI 3-kinases, like DdPIK3 or DdPIK5, regulate phagocytosis. As discussed below, Δddpik1/ddpik2 cells contained an increased number of filopodia, structures that may play an important role in binding beads and bacteria as the first step of phagocytosis (Maniak et al., 1995). Conceivably, increased filopodia may partially compensate for the reduction of cortical actin in the mutant cells; the organization of cortical actin has been suggested to be important in phagocytosis (Maniak et al., 1995). Regardless, these observations suggest that the mechanisms regulating pinocytosis and phagocytosis may be overlapping but they are not identical, a possibility further supported by recent studies that indicate cells lacking functional heterotrimeric G proteins (Wu et al., 1995) are normal in pinocytosis but defective in phagocytosis (Buczynski, G., and J. Cardelli, unpublished results).

Endocytosis in Dictyostelium differs from endocytosis in mammalian cells in at least three fundamental ways. First, no early fluid phase endosomal recycling compartment has been identified in Dictyostelium, second, nearly 100% of internalized fluid phase is secreted from Dictyostelium cells (Padh et al., 1993; Aubry et al., 1993) and third, large vacuoles exist located downstream from lysosomes along the endosomal pathway. These vacuoles termed postlysosomes are greater than 2.5 μm in diameter, contain less vacuolar proton pumps and have lumens that are much less acidic than lysosomes (Nolta et al., 1995). Postlysosomes may arise from fusion of lysosomes (Rodriguez-Paris et al., 1993). Mammalian cells do not contain a compartment analogous to postlysosomes although this type of vacuole may exist in other free living amoebae including the human pathogen Entamoeba histolytica (Aley et al., 1984). Our results demonstrate that DdPIK1 and DdPIK2 played an important role in regulating traffic between lysosomes and postlysosomes. A combination of biochemical and microscopic approaches revealed that FD accumulated in mutant cells in small, acidic, protease positive vesicles (0.5–1.0 μm in diameter) that have all the characteristics of Dictyostelium lysosomes (Temesvari et al., 1994; Rodriguez-Paris et al., 1993). The block in transport to postlysosomes may account for the reduction in rates of fluid phase exocytosis. Electron microscopy also indicated a nearly complete absence of postlysosomal vacuoles in mutant cells supporting the model proposed by others (Nolta et al., 1995) that postlysosomes are not stable compartments but instead form from the fusion of smaller lysosomes and are later consumed during exocytosis. It has been reported recently that a late step in transcytosis is regulated by PI 3-kinase sensitive to low concentrations of wortmannin (Hansen et al., 1995), consistent with our data supporting a role for DdPIK1 and DdPIK2 later in the endosomal pathway.

It has been demonstrated that low concentrations of wortmannin inhibit endosomal fusion in mammalian cells (Jones and Clague, 1995; Li et al., 1995), and Stahl and coworkers have interpreted this to mean that a PI 3-kinase might regulate this event (Li et al., 1995). They further propose that this PI 3-kinase regulates the fusion of endosomal vesicles by activation of Rab5 (Li et al., 1995), a small M_r GTPase previously demonstrated to play an important role in fusion of early endosomes (Bucci et al., 1992). The Rab GTPases are members of the Ras superfamily of G proteins, and Rab proteins have been demonstrated to regulate vesicle trafficking and fusion in a variety of systems (Pfeffer, 1994). Although a number of Dictyostelium Rab proteins have been characterized (Bush et al., 1993, 1994; Bush and Cardelli, 1995), no Rab5 encoding gene has been isolated from this organism. However, we have recently determined that a Rab4-like GTPase (named RabD) regulated lysosome to postlysosome transport (Bush et al., 1996), and conceivably PI 3-kinase may activate or regulate this Rab protein, or other Rab proteins such as a Rab7-like GTPase and a novel GTPase (RabB), also localized to the endosomal pathway (Bush and Cardelli, 1995; Buczynski, G., and J. Cardelli, unpublished results).

Δddpik1/ddpik2 cells were normal in sorting and processing of lysosomal α-mannosidase, suggesting that DdPIK1 and DdPIK2 were not critical in the regulation of the biosynthetic targeting pathway to lysosomes. In contrast another PI 3-kinase, Vps34p, plays a role in the sorting of hydrolases to the yeast vacuole (Stack et al., 1995) but this enzyme is more related to the Dictyostelium PI 3-kinase, DdPIK5, than to DdPIK1, DdPIK2, or DdPIK3 (Zhou et al., 1995). Also, DdPIK5 can partially compensate for the absence of Vps34 in S. pombe suggesting that DdPIK5 may be a Dictyostelium Vps34 homologue. Sorting of hydrolases may be normal in Δddpik1/ddpik2 because another PI 3-kinase (perhaps DdPIK5) is responsible for this process. Therefore, it remains to be tested whether lysosomal α-mannosidase is targeted by a PI 3-kinase independent pathway. Others, however, studying the targeting of certain plant proteins have concluded that PI 3-kinase independent lysosomal targeting pathways do exist.

Two proteins, RabD (Bush et al., 1996) and the vacuolar proton pump (Temesvari et al., 1996b), play important roles in regulating endocytosis and the structure and function of the CV in Dictyostelium. Based in part on this evidence, it has been proposed that the CV and the endosomal pathway are physically connected by membrane trafficking (Bush et al., 1996; Padh and Tanjore, 1995). In contrast, the Dictyostelium DdPIK1 and DdPIK2 gene products, although important in endocytosis, did not play a role in regulating the structure and function of the CV network, suggesting that only a subset of the proteins regulating endocytosis also regulate the structure and function of the CV network.

1–3% of the Δddpik1/ddpik2 cells growing attached to plastic were connected by thin cytoplasmic bridges that lasted for 1–2 h suggesting defects in cytokinesis. Furthermore, defects in cytokinesis could account for the inability of Δddpik1/ddpik2 cells to grow in shaking suspense culture. Cytokinesis is regulated by a variety of proteins including myosin II (Knecht and Loomis, 1987; De Lozanne and Spudich, 1987), myosin essential light chain (Chen et al., 1995), and calmodulin (Liu et al., 1992). All of these proteins regulate the function of actin and/or myosin, and acto-myosin complexes are important in controlling the retraction of the contractile ring during cytokinesis leading to the fission of newly replicated cells. Therefore, given the proposed role for PI 3-kinases in regulating the actin cytoskeleton, it is possible but remains to be formally tested that DdPIK1 and DdPIK2 regulate cytokinesis.

Δddpik1/ddpik2 cells were also morphologically different from wild-type cells. Mutant cells were smaller than wild-type cells and were more irregular in shape. Light microscopy and scanning electron microscopy also indicated that these cells contained significantly more actin rich filopodia than wild-type cells and less cortical actin. Membrane ruffles were almost never seen in mutant cells consistent with the demonstrated role of p85/p110 PI 3-kinase in regulating membrane ruffling in mammalian cells. The altered distribution of actin may be responsible for the decrease observed in pinocytosis although phagocytosis was normal.

It is not yet clear how the putative PI 3-kinases, DdPIK1 and DdPIK2, regulate the distribution of F-actin or why reductions in the level of these enzymes would lead to an increase in filopodia formation. It is conceivable that these putative PI 3-kinase(s) normally function to repress the formation of filopodia and the loss of these proteins results in an increased number of filopodia. However, in mammalian cells, PI 3-kinase has been shown to activate Rac GTPases that, in turn, stimulate actin redistribution and induce membrane ruffling (Wennstrom et al., 1994; Ridley et al., 1992). Therefore, we consider it more likely that formation of filopodia may be part of a signal transduction pathway that triggers the activation of DdPIK1 and DdPIK2 which in turn stimulates changes in F-actin distribution leading to the formation of plasma membrane lamellipodia “webs” that fill the space between the filopodial “fingers” (Nobes and Hall, 1995a).

Our results also indicate that there is a significant alteration in the ability of aggregation-competent cells to move toward the chemoattractant cAMP. Unexpectedly, Δddpik1/ddpik2 cells respond and move more rapidly to the cAMP than do wild-type cells. It is possible that if DdPIK1 and DdPIK2 encode PI-3 kinases that the phospholipid products may be involved in modulating the effect of chemoattractant on the ability of the cells to move, possibly by regulating pathways controlled by members of the Rho/Rac family that affect membrane ruffling, filopod extension, etc. (Nobes and Hall, 1995). Alternatively, the ability of the cells to respond and move more rapidly may be the result of our observed changes in the actin cytoskeleton in resting cells.

In summary, we have found that DdPIK1 and DdPIK2, genes that encode proteins homologous to p110-like PI 3-kinases, play diverse but select roles in regulating endocytosis, cell shape, and cell motility in Dictyostelium. Future studies will be directed towards determining the substrate specificity of these proteins as well as the nature of the effector molecules they interact with.

We thank members of the Cardelli laboratory for careful reading of the manuscript. We thank Dr. Agnes Fok for antibodies recognizing the 100kD subunit of the vacuolar proton pump.

CHC

clathrin heavy chain

CV

contractile vacuole

FD

FITC-dextran

PI

phosphatidylinositide