TOR complex 2 (TORC2) in Dictyostelium suppresses phagocytic nutrient capture independently of TORC1-mediated nutrient sensing

The protein kinase TOR functions within (at least) two distinct multiprotein complexes (Laplante and Sabatini, 2009; Lee et al., 2005; Loewith et al., 2002; Sarbassov et al., 2004; Soulard et al., 2009; Yang and Guan, 2007), TOR complexes 1 (TORC1) and 2 (TORC2), that integrate signals arising from nutrients, the cellular energy state, and other stimuli to regulate growth and cytoskeletal response.

TORC1 is comprised of TOR, Raptor and Lst8 (mammalian GβL), and is required for cellular growth. The phosphorylation targets of TORC1 include the protein translation regulators S6 kinase (S6K) and 4E-BP1 (Brunn et al., 1997; Price et al., 1992). The activated phosphorylated form of SK6 phosphorylates and activates ribosomal protein S6, which stimulates protein synthesis. Conversely, 4E-BP, an inhibitor of eukaryotic (protein translation) initiation factor eIF4E, is inactivated by phosphorylation by TOR. TOR–Raptor interactions are essential for TORC1 function, and genetic inactivation of either TOR or Raptor in yeast, Drosophila, Caenorhabditis elegans or mammals is lethal (Gangloff et al., 2004; Hara et al., 2002; Kim et al., 2002; Kunz et al., 1993; Long et al., 2002; Murakami et al., 2004; Oldham et al., 2000). TORC1 is sensitive to inactivation by the anti-fungal, anti-neoplasic and immunosuppressant agent rapamycin (Abraham and Wiederrecht, 1996; Laplante and Sabatini, 2009; Loewith et al., 2002; Sarbassov et al., 2006; Sarbassov et al., 2005; Soulard et al., 2009; Yang and Guan, 2007). Rapamycin in complex with the cellular protein FKBP12 (Heitman et al., 1991; Kim et al., 2002) disrupts TOR–Raptor interactions and suppresses TOR phosphorylation of S6K and 4E-BP1, protein synthesis and cell growth.

TOR and Lst8 are also constituents of a separate protein complex, TORC2. Unlike TORC1, TORC2 does not contain Raptor and does not exhibit direct sensitivity to inactivation by rapamycin (Loewith et al., 2002; Sarbassov et al., 2006; Sarbassov et al., 2005). In addition to TOR and Lst8, the TORC2 of mammals, Dictyostelium and Saccharomyces cerevisiae contain Rictor-like [(Sarbassov et al., 2004), Dictyostelium PIA (Chen et al., 1997), S. cerevisiae AVO3 (Loewith et al., 2002)] and SIN1-like [(Frias et al., 2006; Jacinto et al., 2006; Yang et al., 2006), Dictyostelium RIP3 (Lee et al., 1999), S. cerevisiae AVO1 (Loewith et al., 2002)] subunits. Although Lst8 is required for full activity of TORC1 (Kim et al., 2003; Roberg et al., 1997), it might primarily function within TORC2; Lst8-null mouse embryos more resemble rictor-null embryos than raptor-null embryos (Guertin et al., 2006).

Many TORC2 functions are mediated by effects on the AKT protein kinase (Huang and Manning, 2009; Laplante and Sabatini, 2009; Soulard et al., 2009; Yang and Guan, 2007). AKT is activated by phosphorylation within the kinase activation loop by phosphoinositide-dependent protein kinase 1 (PDK1) and within a C-terminal hydrophobic (HM) motif by TORC2. In mammalian cells, AKT activity does not require HM phosphorylation by TORC2; TORC2 phosphorylation increases the substrate target range of AKT. Also in mammalian cells, TORC1 can inhibit phosphoinositide 3-kinase (PI3K) activity and suppress phosphatidylinositol (3,4,5)-trisphosphate [PtdIns(3,4,5)P₃] accumulation (Shah et al., 2004). This action blocks recruitment of PDK1 and AKT to membranes and suppresses AKT phosphorylation by PDK1. Unregulated TORC1 can, thus, inhibit AKT activity, even in TORC2-activated cells, indicating antagonistic cross-talk between the TOR complexes. But, these regulatory interactions are not universal. In Dictyostelium, the activities of both PDK1 and TORC2 are completely independent of PI3K (Kamimura and Devreotes, 2010; Liao et al., 2010) and, thus, antagonistic cross-talk between TOR complexes 1 and 2 in the context of PI3K and AKT is not observed.

During growth in the wild, Dictyostelium engulf bacteria and yeast for nutrient particle capture by phagocytosis, an actin-dependent process. TORC1 in Dictyostelium has an essential role in control of cellular growth, and although TORC2 will phosphorylate AKT (Kamimura and Devreotes, 2010; Lee et al., 2005; Liao et al., 2010), the mechanistic interconnections of PDK1 and TOR in control of AKT in Dictyostelium differ significantly from those in mammalian cells. In Dictyostelium, phosphorylation of AKT by PDK1 and TORC2 are very tightly coordinated (Liao et al., 2010). PDK1 is unable to phosphorylate AKT in the absence of phosphorylation at the HM site, yet PDK1 phosphorylation of AKT is absolutely required for activity in vivo. Furthermore, HM site phosphorylation is insufficient for AKT activity in vivo (Liao et al., 2010). Thus, in vivo, AKT requires phosphorylation at both the kinase and HM domains. In addition, PDK1 in Dictyostelium does not interact with PtdIns(3,4,5)P₃ and functions independently of PI3K (Liao et al., 2010). Finally, we have shown that phosphorylation of AKT at both the kinase and HM sites in growing Dictyostelium is unaffected by loss of TORC2 components Rictor/Pia, Lst8 or SIN1/RIP3 (Liao et al., 2010).

Despite the essential role of TORC1 in the control of cellular growth, we show that inhibition of TORC1 activity by depletion of nutrients from media, by genetic depletion of Raptor or by short-term treatment with rapamycin does not alter phagocytosis. Because TORC2 is implicated in the regulation of cell polarization and chemotaxis via its effects on actin function in Dictyostelium (Kamimura and Devreotes, 2010; Lee et al., 2005; Liao et al., 2010), we examined whether TORC2 is similarly required for regulation of actin assembly to promote nutrient particle capture. However, rather than suppressing yeast engulfment as might have been expected from previous hypotheses, we show that loss of TORC2 components promotes nutrient uptake via the upregulation of phagocytosis, supporting a negative role for TORC2 in nutrient particle capture.

We suggest that nutrient particle capture in Dictyostelium is negatively regulated by TORC2, but that particle capture is independent of TORC1-mediated nutrient sensing and growth regulation and of TORC1- or TORC2-dependent feedback loops that act on AKT. The parallel regulations of TORC1 and TORC2 might balance growth rates and regulate the growth-to-development transition. In this context, activation of TORC2 during early development might promote leading edge function of chemotaxing cells, but also inhibit phagocytosis, a potentially antagonistic pathway during Dictyostelium aggregation.

In the wild, Dictyostelium cells acquire nutrients through phagocytosis (Cardelli, 2001). Because TORC1 signaling promotes growth in response to nutrients, we examined whether TORC1 also regulates nutrient particle capture during Dictyostelium growth.

Dictyostelium grow logarithmically in nutrient-rich media provided that nutrients are not depleted and that cells are at moderate density (<5×10⁶ cells/ml). Upon transfer of logarithmically growing Dictyostelium from growth medium to starvation (e.g. into 12 mM phosphate buffer), cells undergo growth arrest and a growth-to-development transition (Kimmel and Firtel, 2004). To determine whether TORC1 in Dictyostelium impacted nutrient particle capture, we quantified phagocytic rates in cells with fully activated or suppressed TORC1. To assay relative in vivo TORC1 activity, we monitored phosphorylation levels of the TORC1 substrate 4E-BP.

Logarithmically growing Dictyostelium showed persistent phosphorylation of 4E-BP as monitored with antibodies against human 4E-BP1 phosphorylated at T37 and/or T46 (Fig. 1A), or at T70 (data not shown) by using immunoblots. However, within 10 minutes of transfer from growth medium to starvation buffer there was rapid dephosphorylation of 4E-BP. Levels of phosphorylated 4E-BP did not change further when cells were maintained in starvation buffer for 2 hours (Fig. 1A) or more (data not shown). These data match those of mammalian cells cultured in serum-replete or serum-deprived medium and indicate that in vivo TORC1 activity is rapidly suppressed if cells are removed from an exogenous nutrient source. Re-exposure of starved Dictyostelium to fresh nutrients promoted re-phosphorylation of 4E-BP and cell proliferation (data not shown).

To determine whether there is a functional correlation between in vivo TORC1 activity and phagocytic rates, growing Dictyostelium were mixed with yeast labeled with tetramethyl rhodamine isothiocyanate (TRITC), and phagocytic uptake was monitored over an extended time course by quantifying internalized fluorescent yeast (Khurana et al., 2005). We simultaneously assayed in vivo TORC1 activity (i.e. 4E-BP phosphorylation) and quantified phagocytosis over 2 hours using growing cells or cells that had been nutrient-starved for 2 hours and maintained in starvation buffer (Fig. 1B,C). As shown, phosphorylated 4E-BP levels from growing cells in nutrient-rich medium or from cells starved in buffer differed significantly in phosphorylated 4E-BP levels and remained so for the first hour of the phagocytosis time-course assay (Fig. 1B). Nonetheless, phagocytic rates were identical between the two cell populations (Fig. 1C), which is consistent with work of others (Katoh et al., 2007; Khurana et al., 2005). Eventually, incubation of Dictyostelium with yeast particles caused a moderate activation of TORC1 and 4E-BP phosphorylation. Nonetheless, phagocytic rates remained unaltered. 4E-BP re-phosphorylation could be similarly induced in starved cells by the addition of either glucose (80 mM) or the essential amino acid leucine (5 mM), but not glycine (5 mM), without an effect on phagocytosis (data not shown). Thus, nutrient particle capture efficiency does not correlate with in vivo TORC1 activity.

Although our data indicate that nutrient sensing via TORC1 does not regulate phagocytosis, it is possible that nutrient withdrawal might not fully suppress TORC1 function. Therefore, we inactivated TORC1 by depleting Raptor protein, an essential subunit of TORC1. We were unable to disrupt the Raptor gene in Dictyostelium using multiple targeting vectors and various selection conditions that optimized cell growth. Raptor is probably an essential gene in Dictyostelium, as it is in mammals, C. elegans, Drosophila and yeast (Hara et al., 2002; Kim et al., 2002). Therefore, we chose to deplete Raptor using a tetracycline repression (Tet-off) system that directs the conditional expression of an interfering RNA, based upon a long-hairpin (lhRNA) structure (Rosel and Kimmel, 2006).

For RNA interference (RNAi) experiments, cells were co-transformed with a raptor (Tet-off) expression vector and a vector that encodes the tetracycline-dependent transrepressor (Blaauw et al., 2000). Cells grown in the presence of tetracycline did not express the engineered raptor RNAi; however, upon removal of tetracycline, raptor RNAi was induced and, with time, endogenous raptor mRNA was depleted (Fig. 2A; to ~75%). Cell growth was also slowed and largely repressed by 24–30 hours (Fig. 2B). Similarly, phosphorylated 4E-BP levels in Raptor-depleted cells (raptor RNAi cells) were reduced to approximately the level observed in wild-type cells starved of nutrients (data not shown).

We next quantified phagocytosis in control and raptor RNAi-expressing cells after 34–40 hours of induction. Untransformed wild-type cells and mock (Blaauw et al., 2000) cells (wild-type cells containing the parental vectors that lacked the raptor RNAi sequence) grown with or without tetracycline served as additional standards. No differences in phagocytosis were observed among any of these various cell lines, regardless of exposure to tetracycline (Fig. 2C; data not shown). Thus, inactivation of TORC1 through depletion of Raptor does not regulate phagocytosis. These data suggest that despite the role of TORC1 as a nutrient sensor and growth regulator, phagocytosis functions independently of changes in TORC1 activity.

To further address the role of TORC1 in the control of phagocytosis, we examined the effects of the TORC1 inhibitor rapamycin. To control for potential non-specific actions of rapamycin, we identified and genetically inactivated the rapamycin sensor FKBP12. Growth of Dictyostelium cells was strongly inhibited by 500 nM rapamycin (Fig. 3A) (Lee et al., 2005), whereas cells mutated for FKBP12 were insensitive to rapamycin (Fig. 3A). To determine the rapidity of the rapamycin response, we examined its effect on the phosphorylation of 4E-BP. The addition of 500 nM rapamycin to enriched growth medium rapidly (<10 minutes) reduced phosphorylated 4E-BP levels to that of starved cells (Fig. 3B).

We next quantified the effects of 500 nM rapamycin on phagocytosis. Cells were pre-incubated with rapamycin for 30 minutes, and phagocytosis was monitored for another 120 minutes in the continued presence of rapamycin. There was no initial effect of rapamycin on phagocytosis, but with extended treatment times (>90 minutes) rapamycin exerted a mild, but reproducible stimulatory effect (Fig. 3C; supplementary material Table S1, Fig. S1). This stimulation was greatly enhanced when pretreatment with rapamycin was extended to 5 hours prior to the phagocytosis assay (Fig. 3C; supplementary material Table S1, Fig. S1). Phagocytic rates in fkbp12-null cells were unaffected by rapamycin regardless of treatment times or conditions (Fig. 3C; supplementary material Table S1, Fig. S1), confirming the specificity of the rapamycin response.

Because rapamycin and FKBP12 specifically alter the interaction of TOR and Raptor within TORC1, we utilized Raptor-depleted cells to determine whether the longer-term sensitivity of phagoctyosis to rapamycin required functional TORC1. Phagocytosis is stimulated by long-term rapamycin treatment in the various control cells, which express normal levels of Raptor, regardless of exposure to tetracycline (Fig. 3D; data not shown). However, Raptor-depleted cells (i.e. raptor RNAi cells grown in the absence of tetracycline; see Fig. 2A,B) were largely refractory to rapamycin-stimulated phagocytosis (Fig. 3D; supplementary material Fig. S1); the small differences were not statistically significant (P>0.1), and probably reflect residual Raptor levels. Because phagocytic rates did not correlate with TORC1 activity (Fig. 1B,C and Fig. 3B,C), the long-term effects of nutrient particle capture through TOR–Raptor destabilization by rapamycin suggest interplay with another pathway.

Long-term rapamycin treatment can reduce TORC2 activity through inhibition of TORC2 assembly in mammalian cells (Sarbassov et al., 2006) and can partially inhibit TORC2 phosphorylation of AGC kinases AKT and PKBR1 in Dictyostelium (supplementary material Fig. S2). This raised the possibility that TORC2 might be an inhibitory target for the action of rapamycin on phagocytosis in Dictyostelium. We, thus, compared phagocytosis by Dictyostelium cells mutated for various TORC2 component genes with that of wild-type cells. For consistency and clarity, we chose to unify Dictyostelium and mammalian nomenclatures (see Materials and Methods) and henceforth refer to TORC2 components in Dictyostelium as Rictor(Pia), Sin1(RIP3) and Lst8.

Wild-type cells exhibited linear unsaturated particle uptake during the time-course of the experiment. However, cells lacking Rictor(Pia), Sin1(RIP3) or Lst8, central members of the TORC2 complex, had significantly elevated rates of phagocytosis for both yeast particles (Fig. 4A, Table 1) and latex beads (supplementary material Fig. S3) in comparison with wild-type controls. The initial rates (i.e. at 30 minutes) of yeast particle uptake for the various mutant strains exceeded that of wild-type cells by about sixfold.

Because actin polymerization is an essential determinant for phagocytosis and because TORC2 is implicated in actin regulation, we examined filamentous (F)-actin cytoskeletal changes in real-time during phagocytosis by wild-type cells and by cells lacking TORC2 components. F-actin was monitored by expression of the F-actin binding domain (ABD) of ABP-120/filamin fused with GFP (GFP–ABD) (Pang et al., 1998). During phagocytosis there is an initial polymerization of actin at the site of particle attachment (see Fig. 4B, Table 2). Then, the phagocytic cup forms, with the yeast particle surrounded by F-actin-associated membrane (see Fig. 4B, Table 2). As the particle becomes fully internalized, F-actin is rapidly depolymerized. Actin polymerization and depolymerization at the phagocytic cups was not morphologically different in the various TORC2 mutant cell lines in comparison with wild-type (Fig. 4B). However, by carefully monitoring phagocytic progression by live-cell confocal microscopy (Khurana et al., 2005), we observed several differences in comparing wild-type cells with rictor(pia)-, sin1(rip3)- or lst8-null cells that described enhanced phagoctyosis in cells lacking TORC2.

For wild-type cells, ~50% of F-actin-mediated contacts with yeast particles convert to a phagocytic cup, and only 30% of attached yeast particles become fully internalized (Table 2). By contrast, 90% of TORC2-deficient cells were able to complete full particle engulfment (Table 2). Furthermore, the time from attachment of a TRITC–yeast particle to full internalization and dissociation from F-actin was decreased about twofold in the rictor(pia)-, sin1(rip3)- or lst8-null cells in comparison with wild-type cells (Fig. 4B, Table 2). Thus, TORC2-deficient cells were significantly more efficient than wild-type cells in the phagocytic process, ultimately defining the increased rates observed during the quantitative, time-course phagocytosis assay (Fig. 4A).

We next evaluated the potential role of other components of the TOR signaling pathway during phagocytosis. TSC2 is suggested to play a dual role in TOR signaling (Huang and Manning, 2009). As a Rheb–GAP, TSC2 serves to antagonize the TORC1 activator Rheb, but TSC2 has also been shown to activate TORC2 in mammalian cells (Huang et al., 2008).

We have identified TSC2 in the Dictyostelium genome by bioinformatic comparison (supplementary material Fig. S4A). Dictyostelium TSC2, of ~325 kDa, has the characteristic tuberin and GAP domains and a putative AKT target at Ser2428 within the interdomain region (supplementary material Fig. S4A). TSC1 is less well-defined by annotation and we have not identified a Dictyostelium ortholog. Currently, we do not know if a functionally equivalent Dictyostelium TSC1 is too diverged or absent. The TSC1-interacting region of TSC2 lies within in its N-terminus, a region in Dictyostelium TSC2 that is both highly diverged in sequence and interrupted by homopolymeric asparagine repeats (supplementary material Fig. S4A).

If TSC2 were a positive regulator of TORC2, we would anticipate that loss of TSC2 would activate phagocytosis. However, we generated null-cell lines for TSC2 and demonstrated that phagocytosis was suppressed in tsc2-null cells, not activated (Fig. 5, Table 3; supplementary material Fig. S5). Although TSC2 functions as a positive regulator of TORC2 in some networks (Huang et al., 2008; Huang and Manning, 2009), its role during phagocytosis in Dictyostelium must be distinct. To further study the TSC2 pathway, we quantified phagocytosis in Dictyostelium cells lacking AKT, a TSC2 inhibitor, and in Dictyostelium cells lacking Rheb (see supplementary material Fig. S4B), a direct target of TSC2 GAP activity (Avruch et al., 2008). Cells that lack AKT had elevated rates of phagocytosis (Fig. 5, Table 3) and, remarkably, cells lacking Rheb exhibited an increase in phagocytosis (Fig. 5, Table 3; supplementary material Fig. S5). The data suggest that AKT, TSC2 and Rheb function coordinately during phagocytosis, with the potential for AKT to function upstream during phagocytosis as a negative regulator of TSC2 signaling.

AKT is an essential regulator of nutrient uptake through macropinocytosis (Rupper et al., 2001), a fluid phase uptake process that differs from phagocytosis (Cardelli, 2001). Because activated TORC2 suppresses nutrient particle uptake and because TORC2 can also target AKT, we wished to determine whether TOR regulation was also involved in nutrient uptake through macropinocytosis. First, utilizing a previously characterized akt-null strain (Meili et al., 1999), we showed that, as expected, macropinocytosis was suppressed in akt-null cells (Fig. 6A). Nonetheless, we also showed that macropinocytosis was unaffected by the loss of TORC2 components Rictor(Pia), Sin1(RIP3) or Lst8 (Fig. 6A; data not shown). We further defined the specificity of TORC2 and TORC1 action by the inability of Raptor depletion or of long- (or short-) term 500 nM rapamycin treatment to affect macropinocytosis (Fig. 6A; data not shown). Thus, neither TORC2 nor TORC1 appear to regulate macropinocytosis.

To test whether the effect of TORC2 suppression of phagocytosis resulted from altered cellular adhesion, we quantified cell association of the various strains with surface matrices. We found that inactivation of TORC2 or TORC1, or treatment of the cells with rapamycin, did not alter cellular adhesive properties (Fig. 6B). We also showed that neither TSC2 nor Rheb regulate macropinocytosis or cell adhesion (Fig. 6A,B).

Because macropinocytosis is not regulated by TORC1 or TORC2, but is regulated by AKT, the defined inhibitions of phagocytosis via TORC2 or rapamycin must not involve suppression of AKT function. These observations are consistent with our previous studies that show that loss of TORC2 components during growth does not alter the basal phosphorylation state or in vivo activity of AKT (Liao et al., 2010). These data also indicate that TORC2 must regulate phagocytosis by a pathway that differs from that of macropinocytosis and cell adhesion. The differences between TORC2 regulation of macropinocytosis and phagocytosis also cannot be dismissed as differences in AKT substrate specificity; in Dictyostelium, PDK1 is unable to phosphorylate AKT if phosphorylation at the TORC2 target site is prevented. Collectively, these data indicate that the regulation of phagocytosis by TORC2 occurs independently of its action on AKT as a downstream target.

Data obtained using sin1(rip3)-, rictor(pia)- and lst8-null cells indicate that TORC2 is a major regulator of phagocytosis (see Fig. 4). Although nutrient regulation of TORC1 is minimally involved in phagocytosis (see Figs 1, 2), AKT, TSC2 and Rheb also have a significant function (Fig. 5). Because AKT is an upstream inhibitor of TSC2, as well as a potential downstream positive target of TORC2, its genetic position in the regulation of phagocytosis is not clear.

To investigate the genetic connections among the various components further, we examined strains carrying additional mutations in the pathway. First, we expressed mutated forms of Rheb that exist in preferentially GTP-bound (Rheb^Q64L) and non-GTP-bound (Rheb^D60I) states (Long et al., 2005; Tabancay et al., 2003) and that function, respectively, as dominant-active and dominant-negative variants in other systems (supplementary material Fig. S4B). The Rheb constructs were N-terminally tagged with GFP, and expression of each variant was confirmed in cell populations by fluorescence and immunoblotting.

Expression of (non-GTP-bound) Rheb^D60I in wild-type cells enhanced phagocytosis, mimicking the rheb-null phenotype (Fig. 7A, Table 3), similar to the repression in tsc2-null cells (Fig. 5). We then examined genetically the epistatic relationship between Rheb and AKT and between Rheb and TORC2 by comparing rates of phagocytosis among wild-type and akt-, rip3- and lst8-null cells that expressed or did not express the dominant-active GTP-bound Rheb variant Rheb^Q64L. Whereas wild-type and akt-null cells expressing Rheb^Q64L were suppressed for phagocytosis (Fig. 7B), expression of Rheb^Q64L in either the rip3- or lst8-null background had no effect on phagocytosis (Fig. 7C). Yeast particle uptake remained elevated in these cells compared with akt-null or wild-type cells that express Rheb^Q64L.

TORC2 in Dictyostelium also phosphorylates PKBR1, an AGC kinase related to, but distinct from, AKT (Meili et al., 1999). Although AKT is the more active kinase during growth phase, TORC2 regulation of PKBR1 could still mediate regulation of phagocytosis. Nonetheless, we showed that pkbr1-null cells expressing Rheb^Q64L were suppressed for phagocytosis to the same degree as wild-type cells expressing Rheb^Q64L (supplementary material Fig. S6). These results also indicate that akt/pkbr1-null cells expressing Rheb^Q64L would have defects in both macropinocytosis and phagocytosis and, thus, would grow extremely poorly under all conditions. Consistent with this prediction, we were unable to isolate for study akt/pkbr1-null cells expressing Rheb^Q64L.

Collectively, the data indicate that Rheb-mediated suppression of phagocytosis requires TORC2, but that Rheb–GTP functions independently of AKT (or of PKBR1) as a downstream target, but potentially within an AKT–TSC2–Rheb circuit.

In its ecological niche, Dictyostelium phagocytose bacteria and yeast to capture nutrients. We have shown that cells deficient in TORC2 components are more effective throughout the phagocytic pathway during quantitative, time-course measurements (see Fig. 4A). The initial rates (i.e. at 30 minutes) for the various TORC2-deficient strains exceeded wild-type rates by six- to tenfold, positioning TORC2 as a major negative regulator of phagocytosis in Dictyostelium.

In contrast to phagocytosis, macropinocytosis, the process of cellular uptake of fluid phase nutrients, is unaffected by the loss of TORC2 components or by the inhibition of TORC1 by rapamycin. We recognize that the laboratory strains used were selected for growth in liquid media (Maniak, 2001), and that the independence of macropinocytosis from TOR regulation might reflect second-site optimization for macropinocytosis in these strains.

Nonetheless, we have established phagocytosis in Dictyostelium as a downstream target of TORC2 signaling (see Fig. 8). This suggests that in addition to TORC1, TORC2 in Dictyostelium is also involved in nutrient-related signaling. In contrast to TORC1, which activates growth in response to high nutrients, TORC2 inhibits nutrient uptake. Consistent with these conclusions, we observed an inverse relationship (supplementary material Fig. S7) between in vivo TORC2 activation during development (Lee et al., 2005; Liao et al. 2010; Meili et al., 1999) and the relative ability of cells to undergo phagocytosis (Katoh et al., 2007). The local activation of TORC2 during chemotaxis might possibly serve to both activate the leading edge but also suppress phagocytosis, a potentially antagonistic pathway.

It should also be noted that phagocytic processes in Dictyostelium and mammalian macrophages are mechanistically conserved (Cardelli, 2001). Conceivably, TORC2 plays a crucial role during innate immune response and microbial defense (Sinclair et al., 2008), as for nutrient particle capture in Dictyostelium. The interrelationships of the TOR complexes might define new modes for balancing cell proliferation with growth arrest and other cellular processes.

The defined role of TORC2 in phagocytosis in Dictyostelium has permitted development of a novel and quantifiable assay for analyzing additional components in TOR signaling. This unique approach has allowed us to explore genetic interactions that had not been previously appreciated in Dictyostelium. Expression of GTP-bound Rheb suppresses phagocytosis in wild-type and akt-null cells, but it is unable to do so in either lst8- or rip3-null cells. Thus, activated Rheb regulates phagocytosis in a pathway that is independent of AKT, but integrated with TORC2.

A mechanism consistent with the genetic and known biochemical data would place AKT–TSC2–Rheb in a single functional regulatory pathway, with activation of Rheb antagonistic to nutrient particle capture in Dictyostelium (Fig. 8). Thus, akt- and rheb-null cells exhibit elevated rates of phagocytosis, whereas tsc2-null cells are suppressed. In addition, loss of PTEN, an antagonist of PI3K–AKT signaling, represses phagocytosis in Dictyostelium (Dormann et al., 2004).

An axis involving AKT–TSC2–Rheb would seem to implicate TORC1 in the regulation of phagocytosis in Dictyostelium. However, other aspects of TORC1 regulation are not consistent with TORC1 as the primary regulator. Manipulation of in vivo TORC1 activity by withdrawal of nutrients or short-term treatment with rapamycin does not alter phagocytosis, but reduces 4E-BP phosphorylation. In addition, depletion of Raptor causes growth arrest without affecting phagocytosis. Although long-term treatment with rapamycin will induce growth arrest and stimulate phagocytosis, the latter event will not occur in Raptor-depleted cells. Unlike mammalian cells, Rheb is not required in Dictyostelium for growth or for TORC1 activity in growth media. A Rheb–TORC1 connection with phagocytosis would be separate from a TORC1 dependency for cell growth and phosphorylation of 4E-BP.

Long-term effects of rapamcycin on phagocytosis could result from secondary transcriptional changes. Although we cannot exclude this completely, several observations are not simply consistent with this as the major response. Rates of phagocytosis in cells deficient for Raptor, Rictor(Pia) or SIN1(RIP3) were not altered upon treatment with rapamycin (supplementary material Table S1, Fig. S1). If the primary response to rapamycin were transcriptional, we might have detected some changes in phagocytosis. We also suggest that if a major effect on phagocytosis were transcriptional, we would have observed differences in the rates of phagocytosis between growing and 2-hour starved cells, in which TORC1 activity was also rapidly repressed (Fig. 1) and major transcriptional changes ensued.

Loss of TORC2 components in Dictyostelium clearly activates phagocytosis, and genetic data indicate that Rheb suppression of phagocytosis cannot occur in the absence of productive TORC2. Although these might be unlinked events, data from Schizosaccharomyces pombe and mammalian cells indicate a potential function for Rheb that is independent of TORC1 (Aspuria et al., 2007; Karbowniczek et al., 2004; Otsubo and Yamamato, 2008; Urano et al., 2007; Weisman et al., 2005).

During growth of Dictyostelium, PDK1 and HM site phosphorylations of AKT and AKT in vivo enzymatic activity are unaffected by loss of TORC2 components (Liao et al., 2010). The cross-dependencies of PDK1 and TORC2 on AKT phosphorylation (Liao et al., 2010) and the independence of PDK1 from PI3K and TORC1 regulation (Kamimura and Devreotes, 2010; Liao et al., 2010) further demonstrate that AKT, or the related kinase PKBR1, do not function as downstream targets of either TORC2 or TORC1 in control of phagocytosis. Although during development there is a dependency of actin polarization on a TORC2–AKT pathway (Lee et al., 2005), during growth these components are unlinked for phagocytic uptake.

Dictyostelium AX3 cells were grown axenically in D3T medium (KD Medical) at 21°C (Kim et al., 1999). For transgene expression, Dictyostelium cells were transformed with 5–10 μg of circular plasmid DNA by electroporation. Linearized plasmids were used for targeted gene disruptions. Depending upon the antibiotic resistance cassette used, the medium was supplemented with 20 μg/ml G418 (Invitrogen) and/or 10 μg/ml blasticidin S (Fisher).

For Tet-off regulated expression (Blaauw et al., 2000), parental AX3 cells were co-transformed with the transactivator plasmid (MB35) and response plasmid (MB38) with or without the RNAi insert that was generated by linking a 760-nucleotide (nt) hairpin Raptor sequence separated by 372 nt of unpaired sequence to the tetracycline-repressible promoter (see below). Transformed cells were selected in the presence of tetracycline. To silence gene expression, media were supplemented with 30 μg/ml of tetracycline. To induce expression, exponentially growing cultures were washed twice with D3T medium and resuspended at 2×10⁵ cells/ml in fresh D3T lacking tetracycline. Cells in media with or without tetracycline were grown in shaking culture and counted twice daily to assess growth. Samples were also withdrawn for RNA analysis.

For rapamycin growth curve analysis, log phase cells were diluted to ~2×10⁵ cells/ml. Rapamycin was added to a concentration of 500 nM in DMSO, and DMSO was added to control cells. Shaking cultures were counted twice daily to assess growth.

Null mutants of FKBP12, Lst8, TSC2 and Rheb were generated by homologous recombination using targeting constructs disrupted by insertion of a blasticidin resistance cassette. All the transformants were screened by PCR and further confirmed by DNA and RNA blot hybridization. All of the constructs were verified by sequence analyses. rip3-, akt-, crac- and pia-null cells were generously provided by Rick Firtel (University of California, San Diego, CA), Carol Parent (Centre for Cancer Research, Bethesda, MD) and Peter Devreotes (John Hopkins University School of Medicine, Baltimore, MD).

To generate regulated RNAi expression of Raptor, we constructed a long-hairpin RNA construct directed to a central region of Dictyostelium Raptor as part of the TET-off system (Blaauw et al., 2000; Rosel and Kimmel, 2006). The region corresponding to 1860–2999 nt of Raptor complementary DNA (cDNA) was PCR-amplified from genomic DNA using the forward primer 5′-AGATCTATGAAGTTAAAATTGGTGATCTTTG-3′ and the reverse primer 5′-GACGTCGGATAAATCTTGTCAATTGGATTTAG-3′ and ligated in the antisense orientation into the BglII/AatII sites of the TET-off MB38 response plasmid to express antisense Raptor. Sense Raptor sequence 2239–2999 nt was PCR amplified from genomic DNA using the forward primer 5′-GACGTCGGTGGTGCAGAAGGAAGTG-3′ and reverse primer 5′-ACGCGTGTTGATGAAGTTAAAATTGGTGATC-3′ and ligated in the sense orientation at the AatII/MluI sites of MB38 immediately 3′ to the antisense insert. In both cases, genomic DNA was amplified from an intronless region of Raptor. All of the constructs were verified by sequence analyses.

Full-length Rheb cDNA was amplified by RT-PCR from total Dictyostelium vegetative RNA using the forward primer 5′-GGATCCATGGCACCACAAAAACATAG-3′ and the reverse primer 5′-TCTAGATTACATTAAAATACAACCTTCTTTTTG-3′ and the Qiagen OneStep RT-PCR Kit. To generate GTP-bound (Q64L) and non-GTP-bound (D60I) forms of Rheb, we mutagenized the Rheb cDNA using QuikChange II Kit (Stratagene) and fused them to the C-terminus of GFP in the pTX–GFP expression vector (Knetsch et al., 2002). All of the constructs were verified by sequence analyses.

Relevant dictyBase ID numbers are TOR (DDB0214908), Raptor (DDB0191024), FKBP12 (DDB0203129), Lst8 (DDB0233210), Pia (DDB0185055), RIP3 (DDB0201626), TSC2 (DDB0218242) and Rheb (DDB0229441). TORC2 components Pia and RIP3 have functional equivalents in the yeasts, Drosophila and mammals, and although the identification of the Dictyostelium genes (Chen et al., 1997; Lee et al., 1999) predates that of other systems (Frias et al., 2006; Jacinto et al., 2006; Loewith et al., 2002; Sarbassov et al., 2004; Yang et al., 2006), there has not been universal agreement on a common nomenclature. As we believe consistency and clarity are essential for communication among workers in diverse systems, we have chosen to modify nomenclature of the Dictyostelium genes by additionally incorporating mammalian terminology. Hence, we refer to Pia as Rictor(Pia) and RIP3 as SIN1(RIP3).

4E-BP phosphorylation was monitored in whole cell lysates by immunoblot (NuPage, Invitrogen) using antibodies against human 4E-BP1 phosphorylated at T37 and/or T46, or at T70 (Cell Signaling) that was normalized to actin. Antibody specificity to 4E-BP was verified by cross-reactivity to Dictyostelium 4E-BP that had been C-terminally tagged with the hemagglutinin (HA) epitope (data not shown). When expressed in Dictyostelium, phosphorylation of HA-tagged, Dictyostelium 4E-BP shows comparable sensitivity to nutrient removal and rapamycin treatment as does endogenous 4E-BP (data not shown).

Quantitative phagocytosis using heat-killed yeast cells was performed as described (Khurana et al., 2005). Dictyostelium were washed and resuspended in 17 mM phosphate buffer, pH 6.2 (PB) at 2×10⁶ cells/ml. After recovery for 15 minutes in shaking suspension at room temperature, TRITC-labeled, heat-killed yeast cells were added to the cell suspension. Samples of 1 ml were withdrawn at indicated times and added to Trypan Blue solution (20 mg/ml dissolved in 20 mM sodium citrate containing 150 mM NaCl) to quench the fluorescence of non-internalized yeast. Cell pellets were collected and resuspended in PB, and fluorescence measured in a luminescence spectrometer LS50B (Perkin-Elmer) using 544 nm light for excitation and 574 nm for emission. Arbitrary fluorescence units were normalized to the maximal value obtained for wild-type cells within the same experiment. Experiments were performed three to eight times. Identical data were obtained when phagocytosis was assayed in standard nutrient broth rather than in PB. Rapamycin (500 nM) treatments are described in the text; controls for rapamycin included DMSO. The cell lines for each data set within a particular figure were measured in parallel under the same conditions and using an identical aliquot of TRITC-labeled, heat-killed yeast. The average phagocytosis rates between data sets varied, but were statistically insignificant and primarily reflected differences between individual preparations of TRITC-labeled, heat-killed yeast.

Quantitative phagocytosis using latex beads has been described (Rivero and Maniak, 2006). Dictyostelium were washed and resuspended in 20 mM phosphate buffer, pH 6.2 (PB) at 2.5×10⁶ cells/ml. After recovery for 15 minutes in shaking suspension at room temperature, 1 μm, FITC-labeled latex microspheres (Polysciences) were added to the cell suspension (200 beads/cell). Samples of 1 ml were withdrawn at indicated times and added to 2 ml PB at 4°C, which quenches further bead internalization. Cell pellets were collected by centrifugation through 20% PEG 6000. Cells were resuspended and lysed in 50 mM phosphate buffer, pH 9.2, containing 0.2% Triton X-100. Fluorescence was measured in the Spectromax Gemini EM (Molecular Devices) using 470 nm light for excitation and 520 nm for emission. Arbitrary fluorescence units were normalized to the maximal value obtained for wild-type cells within the same experiment. Experiments were performed three times.

Statistical significances were determined using an unpaired, two-tailed, Student's t-test.

Polymerized (filamentous) F-actin was monitored by expression of GFP–ABD, the F-actin binding domain (ABD) of ABP-120/filamin fused with GFP (Pang et al., 1998). Cells expressing GFP–ABD fusion proteins were harvested in log phase, washed and resuspended in PB. Cells were plated on chambered cover glasses (Nalge Nunc International) and allowed to settle for 15 minutes. In vivo phagocytosis, using TRITC-labeled yeast, was performed as described previously (Khurana et al., 2005). Live cells were observed using an inverted microscope (Axiovert 100M) equipped with a confocal laser system (LSM 510, Carl Zeiss) and an oil immersion 63× NA 1.4 objective lens. An argon laser (488 nm) was used for excitation of GFP, and HeNe1 (535 nm) was used for excitation of TRITC.

Log-phase cells were incubated with TRITC–dextran at room temperature, with shaking at 160 r.p.m. (Khurana et al., 2005). At selected times, 1-ml samples were withdrawn and quenched with Trypan Blue. Cell pellets were collected and resuspended in PB, and intracellular fluorescence measured; excitation was at 544 nm and emission at 574 nm. Arbitrary fluorescence units were normalized to the maximal value obtained for wild-type cells within the same experiment. All experiments were performed three to four times.

Adhesion of cells to a solid matrix was performed as described (Khurana et al., 2005). Briefly, log-phase cells were plated on tissue culture dishes and allowed to settle for 2 hours. The dishes were then shaken for 60 minutes at varying speeds, from 0 to 90 r.p.m. Unattached cells were removed and counted using a hemocytometer.

We are very much in debt to Xin-Hua Liao, Vanessa McMains, Pamela Schwartzberg and Julie Wu for their continuous and helpful insights. We are grateful to Elizabeth Geras-Raaka for assistance with fluorescence detection and Xin-Hua Liao for data on rapamycin inhibition of TORC2 activity. Finally, we thank the Firtel, Devreotes and Parent Laboratories for their generous access to mutant strains.