InAKTivation of insulin/IGF-1 signaling by dephosphorylation

Abstract

Signal transduction pathways are tightly regulated by phosphorylation-dephosphorylation cycles and yet the mammalian genome contains far more genes that encode for protein kinases than protein phosphatases. Therefore, to target specific substrates, many phosphatases associate with distinct regulatory subunits and thereby modulate multiple cellular processes. One such example is the C. elegans PP2A regulatory subunit PPTR-1 that negatively regulates the insulin/insulin-like growth factor signaling pathway to modulate longevity, dauer diapause, fat metabolism and stress resistance. PPTR-1, as well as its mammalian homolog B56β, specifically target the PP2A enzyme to AKT and mediate the dephosphorylation of this important kinase at a conserved threonine residue. In C. elegans, the major consequence of this modulation is activation of the FOXO transcription factor homolog DAF-16, which in turn regulates transcription of its many target genes involved in longevity and stress resistance. Understanding the function of B56 subunits may have important consequences in diseases such as Type 2 diabetes and cancer where the balance of Akt phosphorylation is deregulated.

Introduction

Phosphorylation is an important post-translational modification that can have several pleiotropic effects on the fate of a protein, including its activation, inactivation or degradation. In particular, growth factor-induced signal transduction pathways are kinase cascades, where each phosphorylation step acts to amplify the signal, ultimately regulating processes such as cell growth, proliferation and survival.¹ To preserve cellular homeostasis and maintain the balance between aberrant growth and increased apoptosis, it is critical that signals from the kinases are counterbalanced and the phosphorylation events reversed. In this context, protein phosphatases have emerged as central regulators of cellular signaling processes.¹ Protein phosphatases are classified into three main groups: the phosphoprotein phosphatase (PPP) family, the divalent cation (Mg²⁺ or Mn²⁺)-dependent phosphatase (PPM family), both of which dephosphorylate serine/threonine residues, and the protein tyrosine phosphatase family (PTP).² Recent studies have also identified the Asp-based protein phosphatase family, which depend on an aspartate residue for their catalytic activity, as an additional group of serine/threonine phosphatases.³ Although, the majority of phosphorylation events in the cell occur on serine/threonine residues, the human genome encodes for a very small number of phosphatases (including the PPP and PPM families) that dephosphorylate these residues.^3–5 In fact, the majority of phosphatases encoded in the human genome belong to the PTP family.^6,7

How do these few serine/threonine phosphatases effectively counterbalance the activity of multiple kinases and substrates? Once thought of as ‘promiscuous’ regulators that exhibit little specificity for their targets, some phosphatases have been shown to be capable of dephosphorylating distinct residues even within a single protein.⁵ Multiple levels of regulation determine phosphatase substrate specificity. First, the sub-cellular localization of phosphatases may define a subset of its substrates; the presence of a nuclear, mitochondrial or membrane targeting signal leads to compartmentalization of the phosphatase and thereby may direct it to the local substrate(s).^3,8 Secondly, phosphatases may depend on additional co-factors for their activity. For example, PPM phosphatases dephosphorylate their targets without associating with additional structural components but depend upon divalent cations such as Mn²⁺ and Mg²⁺ for their function.⁹ Lastly, members of the PPP family such as PP1 and PP2A act as holoenzymes: in addition to the catalytic core that performs the actual dephosphorylation reaction, the enzyme complex often consists of additional structural and/or regulatory subunits that act as a scaffold and determine substrate specificity respectively.^5,9

Here, we highlight one such example by focusing on recent studies that identify the B56 regulatory subunit of PP2A as a critical modulator of insulin/IGF-1 signaling in the nematode Caenorhabditis elegans (C. elegans) and mammalian cells¹⁰ as well as in the fruitfly Drosophila melanogaster.¹¹ We discuss how this regulatory subunit directs the otherwise broadly expressed PP2A to Akt and regulates its dephosphorylation at its conserved Threonine residue.¹⁰ In C. elegans, this results in changes in longevity, fat metabolism, dauer diapause and stress resistance.¹⁰

The Insulin/IGF-1 Signaling Pathway

The insulin/IGF-1 signaling pathway is structurally as well as functionally conserved between nematodes, flies and higher organisms such as rodents and humans (Fig. 1).¹² In C. elegans, the insulin/IGF-1 signaling pathway regulates longevity, dauer diapause, fat metabolism and stress resistance.^13–15 Dauer diapause is an alternative larval stage that worms enter in response to adverse environmental conditions (an integration of pheromone, food and temperature signals).¹⁶ Dauers are arrested from development and re-enter the life cycle to molt into fertile adults only when growth conditions become favorable. In addition to lifespan extension, loss-of-function kinase mutants in the insulin/IGF-1 signaling pathway form dauers even under favorable growth conditions.^14,16 These mutants also show changes in metabolism and stress resistance.^14,15

Figure 1.

the insulin/iGF-1 signaling pathway is conserved across phylogeny.

This pathway is typically termed insulin/IGF-1 signaling as the receptor in the pathway, encoded by the gene daf-2, is equally related to both, the mammalian insulin receptor and insulin-like growth factor receptor.¹⁷ Further, the ligand for this receptor has not been verified biochemically. Downstream of the daf-2 receptor is a phosphatidylinositol (PI) 3-kinase signaling pathway that ultimately regulates the major target of this cascade, the Forkhead transcription factor box O (FOXO) homolog daf-16.^18–20 In mammals and worms, PI 3-kinase activation results in the conversion of membrane phosphatidylinositol (3,4)-bisphosphate (PI(3,4)P2) to phosphatidylinositol (3,4,5)-trisphosphate (PI(3,4,5)P3).^21,22 In mammals and presumably in worms, PI(3,4,5)P3 phosphoinositides recruit phosphoinositide-dependent kinases such as phosphoinositol-dependent kinase 1 (PDK-1) and Akt to the plasma membrane through their pleckstrin homology (PH) domains, subsequently leading to their activation.^22,23

C. elegans has two different AKT proteins, AKT-1 and AKT-2.²⁴ In addition, the serum- and glucocorticoid-inducible kinase 1 (SGK1), which is closely related to Akt, has been shown to act at the same level as Akt in the pathway.^25,26 Similar to mammals, in C. elegans activated AKT-1/2 and SGK-1 can phosphorylate DAF-16/FOXO on distinct serine/threonine residues, and this leads to its inactivation as well as cytoplasmic sequestration by the 14-3-3 proteins^{18,19,26–28} (Fig. 1). Under low insulin-signaling conditions, DAF-16/FOXO is able to translocate into the nucleus and transactivate or repress its many target genes. Several genome-wide studies have revealed antioxidant genes, molecular chaperones, detoxification genes, antimicrobial genes and metabolic genes as direct or potential targets of DAF-16.^29–31 These data suggest that DAF-16 target genes may play a combinatorial role in regulating longevity, stress-resistance, dauer diapause and metabolism.

Given that insulin/IGF-1 signaling regulates multiple cellular processes, tight regulation of kinase activity at each step is important to maintain the threshold of signals to elicit appropriate cellular responses. While many of the kinases in insulin/IGF-1 signaling have been well characterized, little is known about the phosphatases that downregulate signals through the pathway. The most well-studied negative regulator of insulin/IGF-1 signaling is the lipid phosphatase and tensin homolog PTEN.^32–34 In C. elegans, the PTEN homolog DAF-18 negatively regulates PI 3-kinase signaling to ultimately promote DAF-16 nuclear localization and function. Loss-of-function mutations or knockdown by RNA interference of daf-18 results in stress resistance, dauer suppression and reduced fat storage.^10,35–38

PPTR-1/B56 Regulates Insulin/IGF-1 Signaling

We reasoned that there would be additional phosphatases in the pathway in addition to DAF-18/PTEN to counterbalance the effects of the kinases. We performed a directed RNAi screen in C. elegans that assessed the role of 60 putative serine/threonine phosphatases in modulating the insulin/IGF-1 signaling pathway by using dauer diapause as a readout.¹⁰ We also included six regulatory subunit genes of the PP2A holoenzyme (explained further below).

In the screen, we assayed for the contribution of the different serine/threonine phosphatases on daf-2 dauer formation. Our positive control was daf-18 RNAi, as we were primarily interested in genes that functioned as negative regulators of the pathway. We also identified several genes that when knocked down by RNAi, resulted in enhanced dauer formation. These may represent phosphatases that act as activators of the pathway or even positive regulators of kinase activity. Among the negative regulators identified in the screen, our top candidate was the gene protein phosphatase two A (2A) regulatory subunit-1 (pptr-1).¹⁰ Knockdown of pptr-1 by RNAi resulted in robust suppression of dauer formation similar to daf-18 RNAi.

In addition to dauer diapause, the insulin/IGF-1 pathway also regulates longevity, stress resistance and fat metabolism.¹³ Phenotypic analyses revealed that pptr-1 was a robust modulator of these additional outputs. Indeed, while pptr-1 RNAi resulted in a reduction in the lifespan of long-lived insulin/IGF-1 receptor (daf-2) mutants, overexpression of pptr-1 conferred almost 30% extension in lifespan.¹⁰ Further, pptr-1 RNAi resulted in decreased thermotolerance as well as decreased fat storage in daf-2 mutants, but not in wild type animals.¹⁰ Similarly, a recent screen performed in the fruitfly Drosophila melanogaster to identify phosphatases that function in the insulin/IGF-1 signaling pathway identified the pptr-1 homolog widerborst as negative regulator of the pathway and consistent with our findings, widerborst was found to be an important regulator of fat metabolism.¹¹

As mentioned earlier, pptr-1 encodes a regulatory subunit of the PP2A holoenzyme. PP2A is an abundant serine/threonine protein phosphatase that functions as a holoenzyme consisting of a 36 kDa invariant catalytic core, a 65 kDa structural scaffold and a variable regulatory subunit.³⁹ PP2A has been implicated in several cellular processes such as insulin signaling, cell cycle progression and translation.^5,40,41 In our screen, we found that RNAi knockdown of the PP2A catalytic and structural subunits in C. elegans resulted in lethality,¹⁰ substantiating the fact that knockdown of such a broadly required phosphatase produces detrimental effects on multiple cellular functions. Importantly, it is the association of the catalytic and structural core with distinct regulatory subunits that enables the PP2A holoenzyme to modulate multiple processes and yet retain substrate specificity.^39,42,43

There are three predominant families of mammalian PP2A regulatory subunits: the B/B55/PR55, B′/B56/PR61 and B″/PR72 families^39,44,45 and currently at least 15 human genes have been identified that encode distinct regulatory subunits.⁴⁵ Further, these genes may have additional splice forms that vastly increase the total number of regulatory subunits that associate with the catalytic and structural subunits, ultimately providing substrate specificity as well as distinct spatio-temporal localization within the cell. Post-translational modifications such as methylation and phosphorylation of the C-terminal tail of the PP2A catalytic subunit can affect the binding of members of the B subunit to the holoenzyme.^44,45 This may constitute additional level of regulation to achieve substrate specificity.

C. elegans has at least seven genes that encode for regulatory subunits and they fall into each of the three PP2A regulatory subunit families (www.wormbase.org).¹⁰ pptr-1 belongs to the B56 family of PP2A regulatory subunits. In our studies, we examined six out of the seven regulatory subunits and found that only pptr-1, and none of the other five genes, showed significant effects on dauer formation, thus highlighting a specific role for this regulatory subunit.¹⁰

Tissue-expression patterns of these regulatory subunits indicate that they are not as broadly expressed, spatially or temporally, as the catalytic subunit of PP2A. Indeed, we found that PPTR-1/B56β was only expressed in discrete sets of tissues, including the spermatheca, vulva and several neurons in the worm (Fig. 2).¹⁰ In Drosophila as well as mammals, members of the B56 family show distinct sub-cellular localization patterns as well, with B56α, B56β and B56ε showing cytosolic expression while B56γ shows both, nuclear as well as cytosolic expression.^3,11 In agreement with this finding, DAPI staining and confocal microscopy showed that C. elegans PPTR-1/B56β was predominantly cytosolic.¹⁰

Figure 2.

Overlapping and distinct expression patterns of insulin/iGF-1 signaling components suggest a possible neuroendocrine regulation of the pathway.

PPTR-1/PP2A and Akt

To investigate how a single regulatory subunit, PPTR-1, was able to modulate multiple outputs of insulin/IGF-1 signaling, we performed genetic epistasis analyses with mutants in the insulin/IGF-1 signaling pathway. Epistasis analyses using dauer formation as a readout of the insulin/IGF-1 signaling pathway revealed that pptr-1 acts downstream of pdk-1, at the level of akt-1.¹⁰ Knockdown of pptr-1 by RNAi could strongly suppress dauer formation of daf-2, and pdk-1 single mutants as well as daf-2; akt-2 and daf-2; sgk-1 double mutants. However, pptr-1 RNAi had no effect on the dauer formation of daf-2; akt-1 double mutants, and therefore, we concluded that there was a genetic interaction between pptr-1 and akt-1.

Akt belongs to the AGC family of protein kinases that also include Protein Kinase A, C, ribosomal S6 kinase (S6k) and SGK1.⁴⁶ Akt has been shown to be at the crossroads of several signaling cascades such that active Akt is a regulator of cell cycle progression, cell survival, glucose metabolism as well as protein synthesis.^23,47 Mammalian studies have shown that activation of Akt is achieved through the phosphorylation of two main residues, Threonine 308 and Serine 473.^23,47,48 While the Thr 308 residue is phosphorylated by the PDK-1 kinase, the mammalian target of rapamycin (mTOR) complex 2 (TORC2) phosphorylates Serine 473.⁴⁸ At the protein level, C. elegans AKT-1 and AKT-2 share nearly 60% sequence homology.²⁴ Interestingly, AKT-1 contains both the Thr 350 (mammalian 308) as well as the Serine 517 (mammalian 473) residues, whereas AKT-2 lacks the C-terminal serine residue.^10,24 Tissue expression analyses show that AKT-1 and AKT-2 share overlapping expression in multiple tissues.^10,24

In C. elegans, AKT-1, AKT-2 and SGK-1 can form a complex to negatively regulate DAF-16 by direct phosphorylation.²⁶ However these kinases show phenotypic differences as well. Reduction of function mutations or RNAi of akt-1 and/or akt-2 results in enhanced dauer formation as well as lifespan extension.^24,49,50 Mutation or RNAi of sgk-1 has been shown to either increase or decrease lifespan and have a minor effect on dauer formation.^26,51 As shown in Figure 2, there are also differences in the expression patterns in C. elegans for AKT-1, AKT-2 and SGK-1.

Humans have three Akt proteins, Akt1, Akt2 and Akt3 that are encoded by distinct genes and share nearly 80% sequence homology.⁵² Based on homology, C. elegans AKT-1 is more related to mammalian Akt2 while C. elegans AKT-2 shows homology to mammalian Akt3. Studies using gene knockouts in mice have revealed more specific roles for each Akt isoform: Akt1 null mice are small and show defects in placental development. In contrast, Akt2 null mice show severe defects in glucose metabolism including insulin resistance and age-dependent loss of adipose tissue while Akt3 null mice show a reduced brain size.⁵² Consistent with the role of Akt in growth and cell survival, all three Akt mutants show greatly reduced cell size as well as mass.^52,53

Our genetic epistasis studies showed that pptr-1 acted at the level of akt-1 but not on the closely related akt-2 or sgk-1. Tissue expression analyses revealed a remarkable overlap in the expression of PPTR-1 with AKT-1, partial overlap with AKT-2 and little or no overlap with SGK-1, pointing at the specificity of the pptr-1/akt-1 interaction (Fig. 2).¹⁰ Consistent with this, in Drosophila, genetic epistasis analysis placed widerborst within the PI 3-kinase pathway at the level of Akt1.¹¹

From these results, we hypothesized that PPTR-1 was modulating insulin/IGF-1 signaling and DAF-16 activity by regulating AKT-1 phosphorylation. Using affinity-purified phospho-specific antibodies raised against each of the two AKT phosphorylation sites in C. elegans, we showed that PPTR-1 modulated Thr 350 dephosphorylation, and to a lesser extent, Ser 517. This interaction was then verified in mammalian 3T3-L1 adipocytes, where mammalian B56β but not the other B56 regulatory subunits robustly regulated Akt dephosphorylation at Thr 308,¹⁰ and the Ser 473 site was unaffected. In Drosophila, Widerborst interacts with Akt1 and regulates its dephosphorylation in a PP2A-dependent manner.¹¹ These phosphorylation experiments highlight the remarkable conservation of the insulin/IGF-1 signaling pathway in terms of regulation and functionality between worms, flies and mammals. Importantly, regulation of AKT phosphorylation is critical in humans as well. Reduced AKT phosphorylation has been associated with insulin resistance in patients with type 2 diabetes and hyperphosphorylated AKT is common in cancers where PTEN is mutated.^46,52,54

Recent mammalian studies identified the PH domain leucine rich repeat protein phosphatases (PHLPP), members of the PP2C family as important regulators of Ser 473, but not Thr 308 of Akt.^55,56 Specifically, PHLPP1 can dephosphorylate Akt1 and Akt3 at Ser 473, while PHLPP2 can dephosphorylate Akt2 and Akt3 at Ser 473.⁵⁶ As a consequence, these two phosphatases elicit different outputs of Akt signaling such as cell cycle control and glycogen metabolism respectively. The PHLPP homolog in C. elegans did not affect dauer formation in our screen (Padmanabhan D and Tissenbaum HA, unpublished observation). Consistent with this, our mammalian data showed that siRNA of B56β or PP2A in insulin-stimulated 3T3-L1 adipocytes resulted in enhanced Akt Thr 308 phosphorylation but had no effect on Ser 473 phosphorylation.¹⁰ Together, these studies reveal how distinct phosphatases can dephosphorylate two distinct residues within a single protein, thereby achieving a remarkable level of complexity in the modulation of signal transduction pathways.

PPTR-1/B56 Modulates DAF-16/FOXO Activity

What are the functional consequences of PPTR-1-dependent modulation of AKT-1 in C. elegans? The major output of C. elegans insulin/IGF-1 signaling is the negative regulation of DAF-16/FOXO.¹³ Dosage modulation of PPTR-1 had opposite effects on DAF-16 nuclear localization: while pptr-1 RNAi resulted in more cytosolic and inactive DAF-16, pptr-1 overexpression resulted in enhanced DAF-16 nuclear localization and lifespan extension.¹⁰ Similarly, overexpression of widerborst in flies results in a reduction in the adult eye, a phenotype similar to dFoxo overexpression and co-overexpression of both genes results in the enhancement of the dFoxo overexpression phenotype.¹¹ Therefore, both of these studies show that modulation of the PPTR-1/B56β dosage can affect FOXO-dependent phenotypes.

Phosphorylation of FOXO by Akt at three serine/threonine residues is an important determinant of its sub-cellular localization in mammals.⁵⁷ In addition, DAF-16/FOXO is also positively regulated by JNK, MST-1 and AMP-dependent kinase (AMPK) through phosphorylations at separate residues.^14,15,58 Therefore, there may be phosphatases that directly dephosphorylate and activate/inhibit DAF-16/FOXO itself. In a genome-wide screen for kinases and phosphatases that modulated dFOXO subcellular localization, activity and protein stability in Drosophila S2 cells, several kinases were identified, including many well-known regulators such as JNK and AKT1, but few phosphatases were identified.⁵⁹ Since kinases can have both stimulatory and inhibitory functions and multiple phosphorylation sites exist, the identification of a DAF-16/FOXO phosphatase(s) may require a more detailed approach. For example, the sensitization of the insulin/IGF-1 pathway, the type of stress or metabolic state (fed versus starved) may result in the association of DAF-16/FOXO proteins with distinct phosphatases. Given that DAF-16/FOXO nuclear-cytosolic localization is so dynamic, the interaction with the phosphatase(s) may be transient and difficult to capture. Identification of phosphatases that directly modulate DAF-16/FOXO function will not only provide a better perspective on the hundreds of genes that FOXO proteins transcriptionally activate/repress, but also have implications in our understanding of FOXOs in disease.^60,61

Future Directions

Several questions regarding PPTR-1/B56 and its regulation of the insulin/IGF-1 signaling pathway via AKT dephosphorylation stem from these findings.

Regulation of PPTR-1

We found that PPTR-1 function was more important under low signaling conditions. This was evident by the fact that pptr-1 RNAi reduced the lifespan, fat storage and thermotolerance of daf-2 mutant worms but did not affect wild type worms. It is possible that under low insulin/IGF-1 signaling, PPTR-1 functions to sensitize the pathway even further to ultimately promote survival. In the context of mammals, it will be intriguing to further study the role of PPTR-1/B56, and determine if levels of blood glucose (or more broadly, nutritional status) modulate the activity of this protein. Studies have shown that PP2A itself is downregulated under normal insulin signaling conditions.⁶² PPTR-1 itself, may also be regulated either at the transcriptional level or posttranscriptionally. Indeed in cardiomyocytes, an increase in wild type or constitutively active FOXO resulted in a corresponding decrease in PP2A activity and subsequent activation of Akt, thereby indicating a feedback loop.⁶³ Therefore, further studies will help to identify the upstream cues that activate or repress PPTR-1/B56 activity.

The role of other subunits

Our studies in C. elegans show that PPTR-1 modulates multiple processes associated with insulin/IGF-1 signaling. Importantly, given that AKT is at the focal point of several signaling pathways, the question that arises is whether PPTR-1 is a broad regulator of AKT activity, or does it specifically play a role in insulin-dependent activation of AKT? The other regulatory subunits did not significantly affect dauer diapause, but are likely to modulate other PP2A-dependent processes, even possibly other outputs of the pathway such as lifespan, stress resistance and fat metabolism. Biochemical approaches such as immunoprecipitation followed by mass spectrometry may help to identify the C. elegans substrates for these other regulatory subunits and PP2A.

Tissue-specificity

A third aspect that would be interesting to investigate further is the tissue-specific regulation of PPTR-1/B56. Although PPTR-1 is expressed in a subset of tissues, modulation of PPTR-1 dosage resulted in changes in organismal longevity suggesting that cell non-autonomous regulation and neuroendocrine signaling is important for this function (Fig. 2). In addition to the head neurons, PPTR-1 shares remarkable overlap with AKT-1 in the spermatheca and vulva. Interestingly, AKT-1::GFP^10,24 and PPTR-1::GFP¹⁰ did not show any expression in the intestine, the major tissue for fat storage in the worm. Moreover, tissue-specific studies have shown that the intestine is the most important tissue for DAF-16-dependent regulation of lifespan.⁶⁴ Akt-dependent phosphorylation is the major mechanism by which the activity of DAF-16/FOXO is regulated. Therefore, how and when does the direct regulation of AKT and DAF-16 occur? Further studies are necessary in worms as well as mammals to determine how PPTR-1/B56 regulates insulin/IGF-1 signaling in the context of the whole organism.

All of our expression studies and biochemistry experiments in C. elegans used overexpression strains, as we were unable to pull-down endogenous AKT-1. It is possible that there are low levels of expression that we could not detect in the intestine or that the AKT-1::GFP strain for some reason was not expressed in the intestine. We found that SGK-1::GFP and PPTR-1::mCherry-FLAG showed no overlap in their expression, and yet AKT-1, AKT-2 and SGK-1 have been shown biochemically to form a complex to regulate DAF-16 activity.²⁶ It is unclear if the tagged versions of these proteins entirely phenocopy the roles of native proteins and further experiments such as immunostaining with antibodies to target endogenous AKT, SGK-1 and PPTR-1 may provide a better understanding of their expression patterns, interaction and regulation. Finally, we identified a number of additional candidates in our RNAi screen that could potentially be important regulators of insulin/IGF-1 signaling. The phosphatases that negatively regulate PDK-1, AKT-2, SGK-1 and DAF-16 itself are currently unknown, and identification of these would provide us with a much better understanding of the regulation of this important pathway.

Taken together, the PPTR-1/B56 regulatory subunit of PP2A is a novel and robust modulator of the insulin/IGF-1 signaling pathway. By regulating the dephosphorylation of a conserved threonine residue on Akt, PPTR-1/B56 can activate DAF-16/FOXO and positively regulate its transcriptional activity. The genes that are upregulated/downregulated as a consequence are likely to play a combinatorial role in regulating longevity, stress resistance, dauer diapause and fat metabolism. Given its extensive conservation and the key role AKT-1 plays in mammals, further studies on PPTR-1/B56 could be of critical importance for diseases, such as cancer and diabetes.