PerspectiveCell Biology

Phospholipase D and mTORC1: Nutrients Are What Bring Them Together

See allHide authors and affiliations

Science Signaling  27 Mar 2012:
Vol. 5, Issue 217, pp. pe13
DOI: 10.1126/scisignal.2003019


Mammalian target of rapamycin (mTOR) complex 1 (mTORC1) plays a central role in translating nutrient abundance into cell growth and proliferation. Although specific proteins have been described as mediators of this nutrient input, their mechanistic linkage remains incomplete. Two studies have added phospholipase D (PLD) as a mediator of nutrients to mTORC1. Furthermore, these studies link PLD and its product phosphatidic acid to previously identified activators of mTORC1 signaling, including the class III phosphoinositide-3 kinase, and provide evidence of the existence of two parallel nutrient-regulated pathways that converge on mTORC1 at late endosomes and/or lysosomes.

Sensing of nutrient abundance first occurs at the cell surface, where it acts to regulate cell growth and proliferation. In mammals and other higher eukaryotes, nutrient sensing in cells also triggers tissue-specific responses that regulate systemic metabolism to maintain metabolic homeostasis. These systems work together to promote survival of the organism during both nutrient excess and famine. Chronic nutrient overload, however, contributes to metabolic dysfunction, including cancer (1), driving a need to understand how nutrient abundance is sensed and communicated. A key nutrient-sensing system in the cell centers on the mammalian target of rapamycin (mTOR).

mTOR, a serine/threonine protein kinase, forms two complexes termed mTOR complex 1 (mTORC1) and mTORC2. Unlike mTORC2, mTORC1 is nutrient-responsive and phosphorylates its prototypical substrates, ribosomal protein S6 kinase (S6K) and eukaryotic initiation factor 4E-binding proteins (4E-BPs) (2, 3). The mTORC1 pathway responds to mitogens as well as to nutrients, and although these two inducers of signaling use parallel pathways, both converge at late endosomes and/or lysosomes to regulate mTORC1 (Fig. 1) (2, 414). In 2001, Chen and colleagues found that phosphatidic acid is required for mitogen-induced mTORC1 signaling (14). Although the exact mechanism is unclear, phosphatidic acid mediates activation of mTORC1 signaling by binding to the FKBP12-rapamycin binding (FRB) domain of mTOR (14). Furthermore, phospholipase D (PLD), the enzyme that generates phosphatidic acid, is also required for mitogen-induced mTORC1 signaling (8). Taken together, these studies suggest that PLD produces the phosphatidic acid required for mitogen-induced mTORC1 signaling.

Fig. 1

Model in which parallel amino acid–regulated pathways mediate mTORC1 signaling. Amino acids activate the class III PI-3 kinase Vps34 to mediate the recruitment of the GTPase RalA-PLD1 complex to late endosomes and lysosomes. This permits PLD1 activation through the small GTPases Arf6 and Rheb, producing phosphatidic acid to mediate mTORC1 signaling. Separately, amino acids activate the Rag GTPases to recruit mTORC1 to late endosomes and lysosomes in a manner dependent on the V-ATPase and Ragulator complex. This allows mTORC1 to interact with Rheb, activate mTORC1, and mediate signaling.


Chen and colleagues observed that mitogen-induced PLD1 activation required the presence of amino acids (8). This led them to hypothesize that PLD1 would also be activated in response to amino acids and, furthermore, would contribute to amino acid–induced mTORC1 signaling. Indeed, the work of Yoon et al. (15) and Xu et al. (16) found that amino acid deprivation attenuates PLD activity in cells. Yoon et al. also showed that PLD activity increased when cells were replete with amino acids, thus demonstrating the ability of amino acids to regulate PLD. Moreover, Xu et al. provide evidence that this is a general nutrient-dependent phenomenon because glucose deprivation also decreased PLD activity in cells.

So is this nutrient-dependent PLD activity important for mediating mTORC1 signaling? Xu et al. relied exclusively on pharmacological approaches to inhibit PLD-dependent phosphatidic acid production, using either the PLD inhibitor 1-butanol or PLD isoform-specific inhibitors. The authors found that both 1-butanol and, to a lesser extent, a combination of PLD1- and PLD2-specific inhibitors repressed both amino acid– and glucose-induced mTORC1 signaling, implying that phosphatidic acid is required for nutrient-induced mTORC1 signaling. Foster and colleagues did not delineate the contribution of each PLD isoform in the human cancer cell lines tested; however, Yoon et al. demonstrated with RNA interference (RNAi) that knockdown of PLD1, but not PLD2, reduced amino acid–induced mTORC1 signaling in human embryonic kidney (HEK) 293 cells, arguing that PLD1 is the predominant PLD isoform mediating the nutrient signal. Yoon et al. also used several mutations that alter PLD1 activity and the ability of nutrients to enhance this activity. They found that if the ability of PLD1 to respond to nutrients is compromised, then nutrient-induced mTORC1 signaling is inhibited, even when the magnitude of PLD1 response is larger than that of wild-type PLD1.

The question remains as to the mechanism by which nutrients regulate PLD1 activity. PLD is regulated by the small guanosine triphosphatases (GTPases) Rheb and RalA (8, 1720), which are also implicated in mediating mTORC1 signaling (13, 2123). Rheb binds and activates PLD1 in response to mitogens, whereas a role for Rheb in nutrient-induced PLD1 activity has not been explored (8). Additionally, RalA constitutively interacts with PLD1 (19) and is required for the activation of PLD1 by the small GTPases ADP-ribosylation factor 1 (Arf1) and Arf6 at endosomes (17, 18, 24, 25). Xu et al. found that depletion of Rheb, RalA, Arf1, and Arf6 each reduced PLD activity; however, it is not clear under these conditions whether or not PLD can still respond to nutrients. Depletion also corresponded with decreased nutrient-induced mTORC1 signaling, corroborating previous studies (13, 2123), with one exception: Arf1 depletion decreased PLD activity, but not mTORC1 signaling. Although it is unclear why in this instance PLD activity did not correspond to mTORC1 signaling, it is noteworthy that Arf6, but not Arf1, interacts with the vacuolar H+–adenosine triphosphatase (V-ATPase) (26, 27), which is also required for mTORC1 signaling (4). Therefore, it is tempting to speculate that Arf1 regulates a specific pool of PLD1 not involved in mTORC1 signaling.

In addition to its regulation by small G proteins, PLD isoforms contain a Phox homology (PX) domain, which binds to phosphoinositol-3-phosphate (PI3P) (28). Of the phosphoinositide-3 kinases (PI3Ks) that function as phosphatidylinositol-3 kinases in cells (29, 30), only the class III PI3K can act as a mediator of amino acid–induced mTORC1 signaling (11, 12, 31). Therefore, does class III PI3K, through its product PI3P, mediate PLD1-dependent mTORC1 signaling? The class III PI3K comprises the catalytic subunit, Vps34, and its regulatory subunit, Vps15. Indeed, Yoon et al. found that genetic depletion or pharmacological inhibition of hVps34 or sequestration of PI3P decreased both PLD activity and amino acid–induced mTORC1 signaling, which is consistent with previous reports (11, 12, 31). Furthermore, Chen and colleagues also found that the PX domain is required for PI3P-mediated PLD1 activation and, through genetic manipulation of class III PI3K and PLD1, found that class III PI3K is upstream of PLD1-dependent amino acid–induced mTORC1 signaling. Although Xu et al. also found that depletion of either hVps34 or hVps15 potently decreased PLD activity, and although they observed that hVps15 depletion in parallel inhibited nutrient-induced mTORC1 signaling, they reported in data not presented in their paper that depletion of hVps34 had no effect on the mTOR response. This latter finding is a demonstration of the occasional disconnect between changes in PLD activity and mTORC1 signaling. Moreover, the reason behind the apparent inconsistent observations linking hVps34 and mTORC1 signaling is unknown, although they may be due to the differences in experimental conditions or cell lines.

If PLD1 is downstream of class III PI3K, then exogenous phosphatidic acid should be sufficient to induce mTORC1 signaling even in the absence of hVps34. This is indeed what Yoon et al. observed, but only when amino acids were present. Furthermore, whereas the addition of exogenous PI3P recapitulated amino acid–induced PLD1 activity in the absence of amino acids, this did not result in induction of mTORC1 signaling. Taken together, this suggests that the Vps34-PLD1 pathway is necessary but not sufficient to drive mTORC1 signaling, which still requires a parallel amino acid input. mTORC1 translocates to late endosomes and/or lysosomes in response to amino acids in a manner dependent on the heterodimeric Rag GTPases and Ragulator complex, which is composed of p18, p14, and MP1 (6, 9). In agreement with these observations, Yoon et al. found that depletion of Rags or p18 prevented mTORC1 translocation, whereas hVps34 depletion or inhibition had no effect. On the other hand, the authors found that PLD1 also translocated to late endosomes and/or lysosomes in response to amino acids, which was dependent on hVps34 and the PX domain of PLD1. Moreover, this response was not dependent on Rags, p18, or raptor, nor did these factors influence PLD activity, indicating that the amino acid–hVps34-PLD1 pathway operates in parallel to the amino acid–Rag pathway, with both pathways converging at late endosomes and/or lysosomes to mediate mTORC1 signaling (Fig. 1).

mTORC1 signaling is critical for the promotion of cell growth and proliferation (2). Ablation of mTORC1 signaling in all developing organisms is lethal (2). If PLD1 and class III PI3K are indeed critical components of mTORC1 signaling, then ablation of either factor should phenocopy mTORC1 disruption. As anticipated, Yoon et al. found that PLD1 and hVps34 depletion inhibited cell growth (15). Consistent with this observation, the phenotypes of Vps34-null mouse embryos are similar to the phenotypes of mTOR-null embryos (32): They are unviable, have major defects in inner cell mass differentiation, and demonstrate loss of phosphorylation of ribosomal protein S6 (33), which is a marker of mTORC1 signaling. On the other hand, mutations in the gene encoding dVps34 do not influence cell growth, proliferation, or steady-state dTORC1 signaling in Drosophila, whereas deletion of dVps34 results in hemizygous lethality at the third instar larval stage (34). Likewise, the critical nature of phosphatidic acid and PLD in mediating mTORC1 signaling has been questioned because the residue Arg2109 in the mTOR FRB domain critical for phosphatidic acid binding is not conserved in Drosophila (35). Furthermore, Drosophila cells depleted of dPLD do not manifest a growth phenotype (36). This may suggest that the necessity of the Vps34-PLD pathway in mTORC1 signaling has only recently evolved in mammals. However, this view has been recently challenged during the characterization of liver- and heart-specific Vps34 knockout mice (37). Despite major defects in autophagy, Vps34 deletion in these tissues does not affect mTORC1 signaling, which is similar to the findings in dVps34 mutant flies (34). Vps34-null mouse embryonic fibroblasts also do not have defects in mTORC1 signaling during serum deprivation (37). Xu et al. found that Vps15 knockdown results in normal serum-induced mTORC1 signaling (16). However, similar to Xu et al., Zong and colleagues observed attenuated mTORC1 signaling upon acute amino acid stimulation after amino acid deprivation (37). Taken together, this suggests that Vps34 does not contribute to steady-state mTORC1 signaling but rather alters the acute response of mTORC1 to changes in nutrients. This is also reflected in the requirement for amino acid–induced mTORC1 signaling in trophectoderm differentiation during early embryonic development (38, 39); however, this paradigm still remains to be directly tested in vivo.

What does this then imply about PLD1? Does PLD1 play a role similar to that of Vps34 in mTORC1 signaling in vivo? Although this question has not been explored, PLD1-deficient mice are reported to be viable and indistinguishable from wild-type littermates (40). Furthermore, PLD1 depletion in HEK 293 cells does not result in decreased mTORC1 signaling after serum deprivation (41, 42), despite the clear effect of PLD1 depletion on acute amino acid–induced signaling (15). It is tempting to speculate that PLD1, like Vps34, mediates only the acute response to changes in nutrients. However, caution needs to be applied when extrapolating these observations to the requirement of phosphatidic acid. Although mitogens and nutrients do not appear to regulate PLD2 (15), it still permits production of phosphatidic acid in the absence of PLD1. In fact, complete disruption of PLD-generated phosphatidic acid by using 1-butanol or dominant negative PLD expression disrupts formation of mTORC1, which is associated with an inhibition of mTORC1 signaling (43). Taken together, although nutrient-induced changes in phosphatidic acid abundance may only acutely regulate mTORC1, some basal amount of phosphatidic acid in mammals appears permissive for the stability of mTORC1 complex. Further mechanistic studies are required to sort out these critical details, including generation of PLD1 and PLD2 double knockout animals, which would shed more light onto the complexity of nutrient-regulated mTORC1 signaling.


View Abstract

Navigate This Article