The Interactive Fly
Zygotically transcribed genes
Studies in Drosophila have characterized Insulin receptor/Phosphoinositide 3-kinase (Inr/PI3K) signaling as a potent regulator of cell growth, but an understanding of its function during development has remained uncertain. Inhibiting Inr/PI3K signaling phenocopies the cellular and organismal effects of starvation, whereas activating this pathway bypasses the nutritional requirement for cell growth, causing starvation sensitivity at the organismal level. Consistent with these findings, studies using a pleckstrin homology domain-green fluorescent protein (PH-GFP) fusion as an indicator for PI3K activity show that PI3K is regulated by the availability of dietary protein in vivo. It is surmised that an essential function of insulin/PI3K signaling in Drosophila is to coordinate cellular metabolism with nutritional conditions (Britton, 2002).
To test whether PI3K is required for larval growth, its activity was inhibited by expressing p60, Deltap60, or Pten in large domains of the larva using the Gal4/UAS system. p60 is an adaptor that couples Inr to Pi3K92E (Dp110), and Deltap60 is a deletion variant lacking part of the Dp110 binding domain. These molecules and their mammalian homologs have dominant-negative effects on PI3K activity when overexpressed, presumably because they compete with endogenous Dp110/p60 complexes for binding sites on upstream activators such as insulin receptors and IRSs. When p60 is expressed under the control of Act-Gal4 (expressed ubiquitously) or Adh-Gal4 (expressed predominantly in fat body), larvae remain growth arrested in the first instar for as long as 2 weeks. Deltap60 and Pten had similar effects. Dissection of these animals revealed that all of their tissues and organs were proportionally reduced. This developmental arrest is indistinguishable from the effects of starvation or inhibition of protein synthesis (Britton, 2002).
To test whether PI3K is autonomously required for growth in the different larval cell types, the Flp/Gal4 technique was used to express p60, Deltap60, or Pten in scattered cells throughout the larva. This method employs the Act>CD2>Gal4 and hs-Flp transgenes to activate UAS-linked target genes, including the cell marker UAS-GFPnls, in cell clones. Heat shock-independent activation of Gal4 occurs prior to the onset of larval growth and DNA endoreplication in 1%-10% of cells (depending on organ) in the fat body, gut, salivary glands, renal (Malpigian) tubules, and epidermis. Cells overexpressing p60, Deltap60, or Pten in the salivary glands and fat body are greatly reduced in size and have much smaller nuclei with far less DNA than adjacent control cells. Similar effects are observed in other larval tissues. Despite their reduced growth, p60-, Deltap60-, and Pten-expressing cells are found at approximately the same frequencies as control GFP-marked cells. Apoptotic cells are not observed. Thus, reductions in PI3K activity are not incompatible with cell viability. Overt effects on cell morphology that might reflect changes in cell adhesion, motility, or identity are also not observed. It is concluded that reducing Inr/PI3K activity in the differentiated tissues of the larva has cell-autonomous effects that are limited to reducing cell growth and DNA replication (Britton, 2002).
These observations suggest that PI3K activity might be responsive to nutritional conditions. To directly test this possibility, a fusion protein was made for use as an in vivo reporter for PI3K activity. The pleckstrin homology (PH) domain of the Drosophila homolog of general receptor for phosphoinositides-1 (GRP1) was fused to green fluorescent protein (GFP), generating a protein called GPH (GFP-PH domain). PH domains from mammalian GRP1 genes bind specifically to phosphatidylinositol-3,4,5-P3 (PIP3), the second messenger generated by class I PI3-kinases. Since PIP3 generally resides in lipid membranes, particularly the plasma membrane, GRP1 is recruited to membranes when PI3-kinase activity raises cellular levels of PIP3. Fusion proteins containing the GRP1 PH domain are likewise recruited to plasma membranes by binding PIP3, and thus serve as in situ reporters for PI3K activity (Britton, 2002).
For in vivo studies, the GPH gene was placed under control of the Drosophila ß-tubulin promotor, generating a gene called 'tGPH' (tubulin-GPH), and introduced into Drosophila by P element-mediated transformation. Membrane localization of tGPH was observed in the larval epidermis, fat body, salivary glands, malpighian tubules, and wing imaginal discs. Cytoplasmic and nuclear tGPH was also visible in these cell types. The degree of membrane localization depends upon the developmental stage. Epidermal cells show little membrane localization of tGPH in embryos or newly hatched first instar (L1) larvae, but have strong membrane localization in second (L2) and early third (L3) instar larvae. Later, in wandering stage L3 larvae, membrane-associated tGPH is again diminished. Similar trends are observed in the fat body. These variations might reflect changes in the levels of Inr, PI3K, Pten, or insulin-like peptides (dILPs) in the larva as it feeds and grows (Britton, 2002).
To test whether tGPH localization was responsive to PI3K activity in vivo, Inr or Dp110 was overexpressed using the Gal4 system. Either gene causes a striking redistribution of tGPH to plasma membranes in cells of the fat body, epidermis, Malpighian tubules, gut, and imaginal discs. To determine whether endogenous PI3K is responsible for the membrane localization of tGPH, PI3K activity was suppressed by expressing p60 with the heat-inducible driver hs-Gal4. In the larval epidermis, partial loss of membrane-bound tGPH is apparent 1.5 hr post-heat shock (phs), and by 4 hr, phs tGPH is nearly completely lost from cell membranes. Similar results were obtained in the fat body when either p60 or dPten were expressed mosaically using the Flp/Gal4 method (Britton, 2002).
To determine whether PI3K activity is nutritionally modulated, tGPH localization was monitored after starvation. Early L2 larvae (48-60 hr AED) were deprived of dietary protein by culture on either 20% sucrose or Sang's defined media lacking casein. These treatments arrest cell growth in all of the differentiated larval tissues. L2 larvae fed on either protein-free diet survive for up to 14 days, but a marked shrinkage of cells in the epidermis and fat body is observed. In the epidermis of early L2 larvae, culture on either protein-free diet causes membrane-bound tGPH to be diminished after 24 hr, and to be nearly undetectable after 48 hr. Levels of total tGPH also decrease after protein deprivation, but nuclear and cytoplasmic tGPH remain detectable for more than 6 days. Loss of membrane-associated tGPH also occurs in fat body cells of protein-deprived L2 larvae, with similar kinetics. Starvation causes shrinkage of the nuclei and nonlipid cytoplasm in fat body cells, leaving large tGPH-negative lipid droplets that occupy most of the cell volume. To test whether membrane association of tGPH is reversible, L2 larvae were cultured on 20% sucrose for 8 days and then returned to whole food. In this experiment, tGPH levels rose and the protein reassociated with plasma membranes in epidermal cells between 24 and 48 hr after feeding. These results indicate that cellular levels of PIP3 drop as a consequence of starvation for dietary protein (Britton, 2002).
Terminally differentiated endoreplicating tissues (ERTs) constitute most of a Drosophila larva, and the growth of these tissues accounts for virtually all of the ~200-fold mass increase sustained by the animal during the larval stages. During larval life, the ERTs provide a physiologically nurturing environment for undifferentiated imaginal cells and neuroblasts, which generate much of the reproductive adult stage. Most of the biomass accumulated in the ERTs is eventually recycled into these progenitor cells as they form the adult. This study provides evidence that Inr/PI3K signaling coordinates nutritional status with ERT cell metabolism and growth. To determine whether Inr/PI3K signaling can maintain cell growth in the face of starvation, Flp/Gal4 was used to express Dp110 or Inr in scattered ERT cells, and then changes in cell size and DNA replication were assessed at time points during a protein starvation regime. At larval hatching, prior to starvation, cells expressing Dp110 or Inr in the gut, fat body, Malpighian tubules, and epidermis are only slightly larger than nonexpressing cells. After several days of starvation on 20% sucrose, Inr- or Dp110-expressing cells in these organs are much larger than adjacent control cells, and have visibly increased DNA content. BrdU incorporation indicated that gut and fat body cells expressing Dp110 or Inr continue to replicate their DNA for at least 2 to 3 days under starvation conditions. Normally, DNA endoreplication in these cells ceases within 1 to 2 days of starvation (Britton, 1998). A catalytically inactive PI3K, Dp110D945A, does not promote cell growth or DNA endoreplication, indicating that lipid kinase activity was required. These experiments indicate that active Inr/PI3K signaling is sufficient to bypass the nutritional requirement for cellular growth and DNA replication in many larval cell types, and that this effect is cell autonomous (Britton, 2002).
To explore the means by which Inr/PI3K signaling induces cell growth, an examination was carried out of the morphology of fat body cells in which PI3K activity had been manipulated. Fat body cells accumulate large stores of protein, carbohydrate, and lipids during larval life, and also produce growth factors. During the third larval instar, these accumulations of nutrients cause fat body cells to become opaque. These nutrients are normally utilized during metamorphosis, but if a larva is starved, they are precociously mobilized into the haemolymph to support the animal during the ensuing dietary crisis. This causes the fat body cells to shrink and become clear as they lose organelles by autophagy and deplete stored metabolites (Britton, 2002).
Expression of Inr or PI3K in fat body cells increases the opacity of the cytoplasm, and thus promotes nutrient storage. A similar cytoplasmic effect is observed in intestinal cells from L1 animals. Close inspection has revealed that in both cell types, ectopic Inr or PI3K decreases the size of prominent vesicles in the cytoplasm. Induction of p60 in early L3 larvae has opposite effects, causing fat body cells to become more translucent. A loss of opacity was observed in fat bodies from Dp110 mutants after their growth arrest at L3. In further tests, fat body cells were stained for lipids with Nile red or for protein with Texas red X succinimidyl ester. This revealed that the cytoplasm of Inr-expressing cells contains many more, but much smaller, lipid droplets than neighboring control cells. In summary, activation of the Inr/PI3K pathway has profound effects on cytoplasmic composition. These effects mimic changes in the fat body that normally take place late in the L3 stage when nutrient storage by these cells is maximal. Suppression of Inr/PI3K activity has opposite effects on cytoplasmic composition, and these appear to mimic the mobilization of nutrients that normally accompanies starvation (Britton, 2002).
Considering the above results, it should be advantageous to larvae to downregulate insulin/PI3K signaling when nutrients are limited, since this would suppress nutrient storage and cell growth and allow nutrient mobilization by tissues such as the fat body. This idea was tested by hyperactivating Inr/PI3K signaling and then tracking development under different nutritional conditions. Several Gal4 drivers were used to induce expression in large numbers of cells, including Adh-Gal4 (expressed in the fat body, trachea, and a few cells in the gut), en-Gal4 (expressed in posterior epidermal cells, the hindgut, and some neural cells), Act-Gal4 (expressed ubiquitously), and hs-Flp/Act>Cd2>Gal4 (induced by heat shock in all cells). In several cases, overexpressed Dp110 and Inr were tolerated in feeding animals. For instance, animals expressing Dp110 under Adh-Gal4 or en-Gal4 control developed without delay and eclosed at the same frequency as controls, giving viable fertile adults. Inr was more deleterious, but some animals expressing Inr under Adh-Gal4 control developed to the L3 stage and a few viable adults eclosed. Ubiquitous expression of Dp110 or Inr using the Act-Gal4 driver, however, was 100% lethal at prelarval stages (Britton, 2002).
In contrast, hyperactivating Inr/PI3K signaling under starvation conditions was catastrophic. When Adh-Gal4 was used to drive Dp110 or Inr expression, for instance, L1 larvae raised on the sucrose/PBS diet all perished within 3 to 4 days of hatching, whereas control animals survived 8 to 9 days. Animals expressing Dp110 under en-Gal4 control also perished within 2 to 3 days, 4 to 5 days before controls, when deprived of dietary protein. Suppressing PI3K activity by expressing p60, Deltap60, Pten, or Dp110D945A using Adh-Gal4, en-Gal4, or even Act-Gal4 had no effect on viability under starvation conditions. These results suggest that the starvation sensitivity caused by high Inr/PI3K activity is specifically related to nutrient uptake and storage, functions that appear to be unique to Inr and PI3K (Britton, 2002).
These results demonstrate that downregulation of Inr/PI3K activity is critical to maintaining metabolic homeostasis under starvation conditions. The remarkable ability of Inr/PI3K-expressing cells to continue stockpiling nutrients and grow, even in starved animals, may account for the starvation sensitivity observed at the organismal level. This idea was supported by observations made in starved tGPH larvae that overexpressed Dp110 in posterior compartment epidermal cells (genotype, en-Gal4 UAS-Dp110 tGPH). In these animals, high levels of membrane-bound tGPH persisted in Dp110-expressing epidermal cells until 2 days after nutrient withdrawal, at which point the animals died. In anterior (A) cells, which did not overexpress Dp110, tGPH was completely lost from plasma membranes within 18 hr after nutrient deprivation, and tGPH protein became undetectable by 36 hr. This is a much more rapid starvation response than observed in animals that did not contain Dp110-expressing cells. Anterior epidermal cells also shrank rapidly during starvation, whereas posterior, Dp110-expressing cells maintained their large size. These effects might result from the rapid depletion of nutrients by the PI3K-expressing cells, and a consequent drop in levels of hemolymph insulins (Britton, 2002).
In performing these experiments, it was noticed that larvae that overexpress Inr or Dp110 wander away from their food. To more carefully analyze this phenotype, animals expressing various PI3K signaling components under Adh-Gal4 control were cultured on agar plates with red-colored food (yeast paste) in the center for ~24 hr after hatching. These animals were then scored for the presence of red food in the gut as well as their proximity to the food source. Animals overexpressing Inr, Dp110, or Dp110CAAX feed poorly (i.e., often have no food in the gut) and frequently wander away from their food. Similar aberrant behaviors were observed when Dp110 was expressed ubiquitously using Flp/Gal4, in which case nearly all animals wandered out of the food and pupated precociously. Thus, elevated levels of Inr/PI3K signaling alter larval feeding behavior, perhaps by affecting the animal's perceived level of hunger (Britton, 2002).
Is nutrition sensed in Drosophila at the cell level or by the animal as a whole? Although animal cells can sense amino acids directly, these studies suggest that cells in Drosophila larvae sense and respond to changes in dietary protein indirectly, using secondary humoral signals -- most likely insulins -- long before they become acutely starved for amino acids. In support of this idea, some cells in the larva can continue to grow and replicate their DNA long after the animal is deprived of dietary protein. Imaginal cells and neuroblasts do this, as do ERT cells in which Inr or PI3K have been artificially switched on. This attests to the fact that starvation for dietary protein does not completely deplete the larval hemolymph of amino acids or cause a global shutdown of protein synthesis. This is probably possible because nutrients stored in ERTs such as the fat body are mobilized during starvation to maintain levels of hemolymph nutrients. Nevertheless, starvation does cause a rapid, global shutdown of ERT cell growth, and thus some essential signal is lost (Britton, 2002).
One factor that all animal cells use for nutritional sensing is TOR (target of rapamycin), a protein kinase that mediates diverse effects on cell metabolism including protein synthesis, amino acid import, ribosome biogenesis, and autophagy (Raught, 2001; Schmelzle, 2000). What role might Drosophila TOR (dTOR) play in the nutrition response system addressed here? The mechanism by which TOR 'senses' nutrition remains uncertain (Kleijn and Proud, 2000), but cellular levels of amino acids, aminoacylated tRNAs, and ATP have been suggested as direct inputs (Dennis, 2001; Iiboshi, 1999). Drosophila dTOR mutations or the TOR-specific inhibitor rapamycin inactivate the TOR target S6-kinase and phenocopy starvation in fed larvae (Zhang, 2000; Oldham, 2000). Starvation, however, does not completely inactivate S6K, suggesting that dTOR retains some activity under starvation conditions (Oldham, 2000). Consistent with this interpretation, overexpressed PI3K is a potent promotor of cell growth in starved larvae, but PI3K cannot drive cell growth in dTOR mutant larvae (J. Lande and T. Neufeld, personal communication to Britton, 2002). This suggests that although dTOR may act as a cell-autonomous nutrient-dependent checkpoint for metabolism, the larva's physiology is so effective in buffering cells against absolute starvation that this checkpoint is rarely if ever fully engaged. TOR is found in fungi and plants and so seems to be a metabolic regulator that was used prior to the advent of multicellularity. Insulin signaling, which is absent from fungi and plants, probably evolved later when multicellular animals required a system to coordinate and fine-tune metabolism in communities of cells. The insulin system is clearly advantageous, since animals in which it is 'short-circuited' by hyperactivation of Inr or PI3K are unable to tolerate even brief periods of starvation (Britton, 2002).
The most direct evidence that insulin signaling is nutritionally controlled was obtained using a cellular indicator of PI3K activity, tGPH, which is recruited to plasma membranes by the second messenger product of PI3K, PIP3. Subcellular tGPH distributions indicate that PIP3 levels are high in many cell types in fed larvae, but low in larvae that have been starved for protein. Although these changes in PIP3 levels might be due to altered expression of Inr, PI3K, or Pten, expression profiling experiments using cDNA microarrays indicate that levels of p60 and Dp110 mRNA are not depressed in L2 larvae that had been deprived of protein for 4 days. Perhaps the most attractive explanation for the apparent loss of PIP3 upon starvation is that some of the seven Drosophila insulin-like peptides (dILPs) are produced in a nutrition-dependent fashion. Several of the dilp genes are expressed in the larval gut (Brogiolo, 2001) which, as the conduit of nutritional influx, might be expected to mediate metabolic responses to feeding throughout the animal. Other dilps are expressed in the salivary glands, imaginal discs, and small numbers of cells in the central nervous system (Brogiolo, 2001; Britton, 2002).
In mammals, insulin promotes the cellular uptake and storage of carbohydrates, proteins, and lipids, and is the strongest anabolic inducer known. Insulin-mediated responses are indirectly antagonized by the hormone glucagon, which stimulates catabolic reactions and the mobilization of stored nutrients. Insulin and glucagon are produced in the pancreas, and their relative levels are constantly adjusted to maintain proper blood sugar levels. The mammalian liver is also a key player in the regulation of metabolic homeostasis. In humans, most of the accessible glycogen, the principle form of stored carbohydrate, is found in the liver. This glycogen can be mobilized in response to exercise or starvation. In Drosophila larvae, hyperactivation of the Inr/PI3K pathway leads to increased accumulation of nutrients in the fat body, an organ that resembles the mammalian liver as the principal site of stored glycogen. Conversely, inhibition of PI3K activity depletes stored nutrients from the fat body, as does starvation. This suggests that like mammals, insects regulate storage of metabolites in response to changes in levels of Inr/PI3K signaling. Direct assays of the levels of carbohydrates, storage proteins, and lipids in the fat body after starvation or manipulation of Inr/PI3K activity should prove informative. While there are as yet no known Drosophila homologs of glucagons, there must be some mechanism by which stored resources can be mobilized during starvation or at the transition from feeding to metamorphosis (Britton, 2002).
Britton, J. S. and Edgar, B. A. (1998). Environmental control of the cell cycle in Drosophila: nutrition activates mitotic and endoreplicative cells by distinct mechanisms. Development 125(11): 2149-58
Britton, J. S., et al. (2002). Drosophila's insulin/pi3-kinase pathway coordinates cellular metabolism with nutritional conditions. Dev. Cell 2: 239-249. 11832249
Brogiolo, W., et al. (2001). An evolutionarily conserved function of the Drosophila insulin receptor and insulin-like peptides in growth control. Curr. Biol. 11: 213-221. 11250149
Dennis, P., et al. (2001). Mammalian TOR: a homeostatic ATP sensor. Science 294: 1102-1105. 11691993
Iiboshi, Y., et al. (1999). Amino acid-dependent control of p70(s6k). Involvement of tRNA aminoacylation in the regulation. J. Biol. Chem. 274: 1092-1099
Kleijn, M. and Proud, C. G. (2000). Glucose and amino acids modulate translation factor activation by growth factors in PC12 cells. Biochem. J. 347: 399-406. 10749669
Oldham, S., et al. (2000). Genetic and biochemical characterization of dTOR, the Drosophila homolog of the target of rapamycin. Genes Dev. 14: 2689-2694. 11069885
Raught, B., Gingras, A. C. and Sonenberg, N. (2001). The target of rapamycin (TOR) proteins. Proc. Natl. Acad. Sci. 98: 7037-7044. 11416184
Schmelzle, T. and Hall, M. N. (2000). TOR, a central controller of cell growth. Cell 103: 253-262. 11057898
Zhang, H., et al. (2000). Regulation of cellular growth by the Drosophila target of rapamycin dTOR. Genes Dev. 14: 2712-2724. 11069888
Home page: The Interactive Fly © 1995, 1996 Thomas B. Brody, Ph.D.
The Interactive Fly resides on the
Society for Developmental Biology's Web server.