Commentary Cellscience Reviews Vol 2 No.2 ISSN 1742-8130

AKAP-based scaffolds and insulin action

Craig C. Malbon¹ & Hsien-yu Wang²

Introduction

As our appreciation of cell signaling matures from understanding of signaling complexes in three dimensions to the addition of the fourth (i.e., time) and fifth dimension (i.e., space), lesions in understanding the role of scaffold proteins become acute. More and more, the detailed operation of cell signaling takes on the appearance of solid-state signaling devices of “meso”-scale dimensions. Although once approached as a system largely dependent upon diffusion-based models to explain protein-to-protein interactions and the generation of/movement of “second messengers” (Singer and Nicolson, 1972), such as intracellular cyclic nucleotides and free Ca²⁺, the time frame of many of such responses demand a more skeptical view of diffusion as a the major driving force over dimensions of the diameter of a cell. In addition, as our knowledge of various individual intracellular signaling pathways expands (Morris and Malbon, 1999) we are forced to find the physical basis of “integration” of signals from many signaling cascades operating in a seamless continuum required for cells to divide, progress, differentiate, and organize into higher level aggregates common in complex organisms, from Drosophila to humans. In a recent publication, Zhang and co-workers (Zhang et al., 2005) have probed an interesting set of earlier observations at the nexus of two of the two most well-studied paradigms of cell signaling, namely tyrosine kinase-mediated and G-protein-coupled receptor (GPCR)-mediated signaling. The results of their novel and insightful study of how insulin acts in disease (i.e., chronic elevation of insulin or hyperinsulinemia) and regulates a prominent GPCR-mediated pathway (i.e., catecholamine regulation of the breakdown of stored triglyceride, or lipolysis) focus attention on a well-known family of scaffold proteins and how scaffolds may integrate signaling both in health and disease, over time and space (Wong and Scott, 2004).

A-kinase anchoring proteins, a.k.a. AKAPs

AKAPs constitute a family of multivalent scaffolds that all share one signature feature, a highly conserved binding site for the RI/II regulatory subunits of protein kinase A (PKA), from whence the term “AKAP” arises (Langeberg and Scott, 2005). These AKAPs provide docking sites for the catalytic subunit of PKA as well as for other critical signaling elements, both protein kinases and phosphoprotein phosphatases (Malbon et al., 2004). In addition to protein kinase and phosphatase partners, these scaffolds have been shown capable of binding/docking adaptor molecules, cyclic nucleotide phosphodiesterases, and elements of the cytoskeleton (Michel and Scott, 2002). This is not to suggest that the AKAPs are merely passive templates upon which signaling occurs. Actually AKAPs may be best envisioned as “catalysts” for the fourth and fifth dimensions of cell signaling.

AKAP250 (gravin) and GPCRs

At least two AKAPs, AKAP250 (Shih et al., 1999) and AKAP79/150 (Cong et al., 2001), are known to interact directly with members of the superfamily of 1500+ GPCRs. The GPCR β2-adrenergic receptor (β2AR) is a heptihelical, plasma membrane-embedded receptor that can be activated by β-adrenergic agonists, such as epinephrine, norepinephrine, and the synthetic isoproterenol. AKAP250, like many AKAPs, can be found throughout the cytoplasm and decorating the inner leaflet of the plasma membrane. Docking of AKAP250 to the β2AR is dynamic and involves the dynamic phosphorylation of the Receptor-Binding Domain (RBD) of the scaffold associating with the C-terminal cytoplasmic tail of the receptor (Fig. 1). The dynamic AKAP250/β2AR interaction is observed in the basal state, but is dramatically increased upon activation of the β2AR with agonist, leading to receptor activation of the Gs protein, activation of adenylyl cyclase, the generation of cyclic AMP, and activation of the PKA that is docked to the AKAP scaffold (Tao et al., 2003). Subsequent to generation of cyclic AMP, PKA catalyzes the phosphorylation of both the AKAP250 (in cis) as well as the β2AR docked to the AKAP, events that appear to increase the dynamic association of this scaffold and the GPCR. AKAP79/150 also displays an RBD highly homologous to that of AKAP250, binds the β2AR, but does so in a more constitutive rather than dynamic manner (Baillie et al., 2005). The AKAP250 appears to function as a “toolbox”, carrying a number of enzymes essential for the regulation of the signaling events, especially the phosphoprotein phosphatase PP2B that appears to be responsible for the de-phosphorylation, resensitization, and the necessary trafficking of the receptor back to the surface of the cell following agonist-induced internalization. Agonist-induced internalization of GPCRs is a process that is facilitated by the adaptor molecule β-arrestin (Lefkowitz and Shenoy, 2005).

Figure 1. AKAP250-based signaling complexes with the β2AR: regulation by β-adrenergic agonist-stimulated phosphorylation of the scaffold (AKAP250) on its Receptor-Binding Domain (RBD) and on the C-terminal, cytoplasmic domain of the β2AR. Shown in green in this cartoon are the additional proteins that constitute the essence of hormone-sensitive lipolysis found in fat cells and in mouse 3T3-L1 adipocytes. Activation by agonist leads to β2AR activation of Gs, Gs activation of adenylylcyclase (AC), generation of cyclic AMP (cAMP), and activation of protein kinase A (PKA). AKAP250 organizes the signaling complex and functions in the process of β2AR dephosphorylation, resensitization, and recycling in response to agonist-induced desensitization and internalization. The AKAP scaffold also docks protein kinase C (PKC), phosphoprotein phosphatase-2B (PP2B), as well as PKA. *, denotes activation of the molecule.

PKA is both sensing intracellular cyclic AMP concentrations and, upon activation, catalyzes the phosphorylation of its substrates. For the sake of the current discussion, if the cell is an adipocyte or fat cell, activation of some pool of PKA results in the activation of triglyceride lipase, a hormone-sensitive lipase that catalyzes the breakdown of stored triglyceride to fatty acids (Fain, 1973), a process termed “lipolysis”. It is essential to have this background information to appreciate the recent work of Zhang et al. in which they make use not of acutely prepared fat cells from animals, but rather of a cell line (mouse 3T3-L1 cells) in culture (Zhang et al., 2005). The 3T3-L1 cells are unique, upon differentiation from pre-adipocytes to adipocytes the cells displays many features of true adipocytes, e.g., regulation of lipolysis, including both catecholamine stimulation and insulin suppression of intracellular cyclic AMP and triglyceride breakdown (Green and Kehinde, 1975).

Hyperinsulinemia, β-arrestin suppression, and increased β-adrenergic stimulation of cyclic AMP

A most interesting paradox was discovered earlier in mouse 3T3-L1 adipocytes (Hupfeld et al., 2003), that chronic elevation of insulin levels in the culture media leads to enhanced, rather than suppressed stimulation of cyclic AMP accumulation in response to β-agonist. Typically, insulin is a potent antagonist of catecholamine-stimulated activation of adenylyl cyclase, of catecholamine-stimulated accumulation of intracellular cyclic AMP, and, for adipocytes, catecholamine-stimulated lipolysis (Fain and Shepherd, 1979). How then does chronic stimulation of 3T3-L1 cell cultures with insulin (~20 nM) for 8 hr stimulate enhanced β2AR-mediated accumulation of intracellular cyclic AMP? The answer to this paradox appears to involve the ability of chronic hyperinsulinemia to provoke a sharp decline in the levels of β-arrestin, an adaptor molecule that plays a role in the internalization of β2AR in response to stimulation by β-adrenergic agonist (Lefkowitz and Shenoy, 2005). The decline in β-arrestin1 would arrest aspects of agonist-stimulated internalization of β2ARs and thereby presumably enhance the cyclic AMP response in comparison with the normal response that includes the β-arrestin-mediated internalization (Hupfeld et al., 2003). An exciting test of this proposal would be to knock-down β-arrestin1 levels with RNAi in 3T3-L1 adipocytes and assay intracellular cyclic AMP, activation of PKA, and stimulation of lipolysis in response to β-adrenergic stimulation.

Insulin stimulated phosphorylation of the β2AR: GPCRs as substrates for receptor tyrosine kinases

The ability of insulin to suppress β-catecholamine action goes far beyond its well-known ability to increase cyclic AMP phosphodiesterase activity (Houslay, 1998), acting rather at additional sites more proximal to the insulin and β2-adrenergic receptors (Hadcock et al., 1992; Baltensperger et al., 1996). The β2AR has been shown to be a substrate for the insulin-stimulated insulin receptor (IR, Baltensperger et al., 1996), a receptor with intrinsic tyrosine kinase activity. In response to physiological levels of insulin, both in vivo and in vitro, the IR directly phosphorylates the β2AR on several distinct sites in the C-terminal cytoplasmic domains of this GPCR. Phophorylation of the β2AR at Tyr350 creates a bona fide SH2 binding site (Baltensperger et al., 1996), a site capable of docking the β2AR with a variety of proteins, including the non-receptor tyrosine kinase Src, the adaptor molecule Grb2, and the GTPase dynamin (Shih and Malbon, 1998). The phospho-Tyr350 β2AR with docked proteins is impaired in its functional interactions with its cognate G protein Gs and this effectively leads to “uncoupling” of the β2AR from the downstream signaling to Gs, to adenylyl cyclase, and to generation of intracellular cyclic AMP (Karoor et al., 1995). The phosphorylation of Tyr350 is also obligate for insulin-stimulated internalization of the β2AR (Karoor et al., 1998), a process that like β-agonist-induced internalization, uncouples the β2AR from the signaling cascade. Insulin-stimulated internalization of β2AR requires not only the direct phosphorylation of the β2AR by the IR, but also the insulin-stimulated activation of protein kinase B/Akt (Doronin et al., 2002), that then also phosphorylates the cytoplasmic, C-terminal tail of the GPCR. Of interest, PKA and Akt share the same serine residue targets in the cytoplasmic C-terminal tail of the β2AR (Fig. 2). What additional effects chronic stimulation by insulin might have on the pathway, which may explain the enhanced rather than attenuated cyclic AMP response observed in the mouse 3T3-L1 adipocytes, was the target of the work by Zhang et al.(2005).

Figure 2. Schematic of possible regulation of AKAP-based signaling of β2AR to PKA in 3T3-L1 adipocytes in culture exposed to chronic stimulation by insulin. Effects of hyperinsulinemia on intracellular levels of the adaptor molecule, β-arrestin1, are suspected to account for increased agonist-stimulated cyclic AMP accumulation. β-arrestin1 is an adaptor molecule participating in the post-receptor sequellae including receptor desensitization, sequestration, and activation of the mitogen-activated protein kinase cascade. The paradoxical loss of PKA activation suggests that the PKA responsible for activation of hormone-sensitive lipase is compartmentalized away from the bulk cyclic AMP. Since AKAP250 (as well as AKAP79) bind PKA and β2AR, these scaffolds may be involved in ability of insulin to suppress PKA, in the face of elevated bulk cyclic AMP levels. Insulin acts at several points in the signaling, catalyzing the direct phosphorylation of the β2AR and catalyzing the activation of Akt, which then also phosphorylates the β2AR on specific residues located in the cytoplasmic C-terminal tail of the receptor. Although speculative, it may be that the insulin receptor or some downstream serine/threonine protein kinase(s) regulated by insulin may make use of the AKAP as a substrate and thereby regulate its interaction with the PKA necessary for hormone-sensitive lipolysis. Since PKA also can be found elsewhere in the cell and the 3T3-L1 cells display elevated cyclic AMP, it appears that the non-AKAP250/AKAP79/150 PKA pool is not capable of activating the lipase and stimulating lipolysis.

Linking changes in cyclic AMP to activation of PKA: scaffolds as an explanation for apparent “compartmentalization”

Using a clever and improved, genetically-encoded fluorescent PKA activity reporter, Zhang et al. (2005) revealed that the paradoxical enhanced elevation of intracellular cyclic AMP levels in response to activation of the β2AR are not reflected in the activation of the PKA. Whilst other indirect (e.g., plant diterpene adenylylcyclase activator forskolin) or direct (e.g., photolysis of caged cyclic AMP) activation of PKA proceeded normally in the insulin-treated cells, the activation of PKA by β-adrenergic agonist was impaired. Thus, the pool of PKA seemed normal in the insulin-treated cells, but the PKA activated in response to β-adrenergic agonist was seemingly “compartmentalized” and not accessed by the sharp increase in intracellular cyclic AMP. This interesting study is by no means the first to report a set of observations that provoked the concept of “compartmentalization” of cyclic AMP/ PKA action. Although such observations are decades old, we now know that the spatial constraints on the operation of solid-state-like signaling devices such as the β2AR-based signaling complex are far greater than first imagined. Studies of AKAP biology opened our eyes to the fourth and fifth dimensions of cell signaling, enabled by scaffolds with spatial trafficking.

Some keen insights to a possible explanation of the basis for compartmentalization of signaling is found in recent analysis of PKA activation by β2AR, operating via the AKAP250 scaffold. Investigators found that when expressed in essentially equivalent amounts as wild-type AKAP250, a mutant form of the scaffold devoid of the RII binding site for PKA was no longer able to catalyze the phosphorylation of either the scaffold’s own RBD or of the β2AR to which the AKAP was docked (Tao et al., 2003). The conventional wisdom would suggest that the availability of catalytic subunits of PKA released from the wild-type AKAP upon stimulation of intracellular cyclic AMP accumulation by β-agonist would be sufficient for the phosphorylation of both the wild-type AKAP as well as the mutant AKAP molecules (and its docked β2AR) in the same cell, but this was not the case. Only those AKAP with PKA RI/II subunit docking sites were phosphorylated and only those β2ARs docked to these wild-type scaffolds were phosphorylated in response to stimulation by β-agonist. Thus, the content of the AKAP tool box may well only function in "cis" with the scaffold, temporally and spatially restricted in their sphere of catalytic function. The results of Zhang et al. (2005) may reflect this notion. Even in the presence of β-agonist-stimulated cyclic AMP accumulation and wild-type levels of PKA, the PKA that drives triglyceride lipase seems to be inaccessible to normal activation by cyclic AMP in these 3T3-L1 adipocytes subjected to chronic stimulation by insulin (Fig. 2).

AKAP250 and AKAP79/150: scaffolds in insulin action?

Are AKAP scaffold central to these effects of chronic insulin on the ability of PKA to be activated in the mouse 3T3-L1 adipocytes? We can only speculate as to the molecular aspects of chronic insulin treatment on the inability of PKA to be activated by elevated intracellular cyclic AMP levels in response to catecholamine stimulation in 3T3-L1 adipocytes. The data presented by Zhang et al. are consistent with a potential role of a scaffold that docks both PKA and the β2AR. AKAP250 and AKAP79/150 fulfill this requirement, as both AKAPs bind PKA RI/II subunits (by definition) and both display the capacity to bind the β2AR. The situation for catecholamine action in the mouse 3T3-L1 adipocytes is, however, complicated by the fact that the cells express both β1AR as well as β2AR (Guest et al., 1990). Do both β-adrenergic receptor sub-types dock AKAPs? This is a central, unanswered question. Catecholamine stimulation of lipolysis in response to isoproterenol can be and is mediated by both sub-types of βAR. The association between β2AR and PKA in the mouse cells seems to be constitutive, which could reflect action by either AKAP250 or AKAP79/150. The GPCR-PKA interactions of AKAP79/150 appears to be solely constitutive (Baillie et al., 2005), whereas that of β2AR-AKAP250 is observed also in the basal, unstimulated state, but is enhanced markedly in response to catecholamine stimulation (Fan et al., 2001; Tao et al., 2003; Baillie et al., 2005).

Results from studies of the regulation of many AKAPs demonstrates that AKAP themselves are substrates for post-translational modifications, especially phosphorylation (Wong and Scott, 2004). There are limited published data on the ability of AKAPs to be phosphorylated directly by tyrosine kinases, either receptor tyrosine kinases (such as the IR of IGFIR) or non-receptor kinases (such as Src). Clearly it is possible that in response to chronic hyperinsulinemia, AKAPs are phosphorylated directly by tyrosine kinases or indirectly by serine/threonine protein kinases (e.g., Akt) that are activated downstream of such tyrosine kinases. Ideally, β2AR-based signaling complexes from the naïve adipocytes and those chronically stimulated with insulin must be isolated and components in the complexes catalogued. Once the composition of the signaling complexes are known (e.g., AKAP, GPCR, β-arrestin, etc.), detailed proteomic analysis may reveal the post-translational modifications (e.g., phosphorylation) of components of the complex that might explain the lesion in PKA activation. Finally, the mouse 3T3-L1 adipocytes, at best, are a useful model of fat cells and the analysis needs to be conducted using white adipocytes isolated acutely from normal and hyperinsulinemic mice to see if the phenomenon noted in the 3T3-L1 cells mimics the pathological state in real adipocytes.

Concluding remarks

Based upon the provocative results of Zhang et al. (2005), much further analysis will be needed to reveal the molecular features of hyperinsulinema on PKA signaling in adipocytes. In addition, it is of interest to ascertain how the short-term effects of insulin on catecholamine action that result in β2AR uncoupling from Gs, internalization of β2AR, and activation of cyclic AMP PDE lead to the effects of chronic insulin treatment observed by these investigators. The apparent compartmentalization of cyclic nucleotide signaling may well reflect the role of AKAPs (or other scaffolds) that bring the fourth (time) and fifth (spatial localization) dimensions into focus for both insulin and catecholamine action. The composition of scaffold/GPCR-based signaling complexes, their structural modifications, and their spatial trafficking are new challenges for investigators in cell signaling. Although proteomics and fluorescent microscopy may provide valuable new leads, the study of such low-abundance signaling “devices” that can demonstrate such high “gain” will tax our ingenuity and resolve.

Acknowledgements

The authors thank the National Institutes of Health for the United States Public Health Services grants that support the research in their laboratories.

REFERENCES

Baillie, G. S., Scott, J. D., and Houslay, M. D. (2005) Compartmentalisation of phosphodiesterases and protein kinase A: opposites attract. FEBS Lett. 579, 3264-3270.

Baltensperger, K., Karoor, V., Paul, H., Ruoho, A., Czech, M. P., and Malbon, C. C. (1996) The beta-adrenergic receptor is a substrate for the insulin receptor tyrosine kinase. J.Biol.Chem. 271, 1061-1064.

Cong, M., Perry, S. J., Lin, F. T., Fraser, I. D., Hu, L. A., Chen, W., Pitcher, J. A., Scott, J. D., and Lefkowitz, R. J. (2001) Regulation of membrane targeting of the G protein-coupled receptor kinase 2 by protein kinase A and its anchoring protein AKAP79. J.Biol.Chem. 276, 15192-15199.

Doronin, S., Shumay, E., Wang, H. H., and Malbon, C. C. (2002) Akt Mediates Sequestration of the beta 2-Adrenergic Receptor in Response to Insulin. J.Biol.Chem. 277, 15124-15131.

Fain, J. N. (1973) Biochemical aspects of drug and hormone action on adipose tissue. Pharmacol.Rev. 25, 67-118.

Fain, J. N. and Shepherd, R. E. (1979) Hormonal regulation of lipolysis: role of cyclic nucleotides, adenosine, and free fatty acids. Adv.Exp.Med.Biol. 111, 43-77.

Fan, G., Shumay, E., Wang, H. H., and Malbon, C. C. (2001) The scaffold protein gravin (AKAP250) binds the ?2-adrenergic receptor via the receptor cytoplasmic R329 to L413 domain and provides a mobile scaffold during desensitization. J.Biol.Chem. 276, 24005-24014.

Green, H. and Kehinde, O. (1975) An established preadipose cell line and its differentiation in culture. II. Factors affecting the adipose conversion. Cell 5, 19-27.

Guest, S. J., Hadcock, J. R., Watkins, D. C., and Malbon, C. C. (1990) Beta 1- and beta 2-adrenergic receptor expression in differentiating 3T3-L1 cells. Independent regulation at the level of mRNA. J.Biol.Chem. 265, 5370-5375.

Hadcock, J. R., Port, J. D., Gelman, M. S., and Malbon, C. C. (1992) Cross-talk between tyrosine kinase and G-protein-linked receptors. Phosphorylation of beta 2-adrenergic receptors in response to insulin. J.Biol.Chem. 267, 26017-26022.

Houslay, M. D. (1998) Adaptation in cyclic AMP signalling processes: a central role for cyclic AMP phosphodiesterases. Semin.Cell Dev.Biol. 9, 161-167.

Hupfeld, C. J., Dalle, S., and Olefsky, J. M. (2003) Beta -Arrestin 1 down-regulation after insulin treatment is associated with supersensitization of beta-2 adrenergic receptor Galpha s signaling in 3T3-L1 adipocytes. Proc.Natl.Acad.Sci U.S.A 100, 161-166.

Karoor, V., Baltensperger, K., Paul, H., Czech, M. P., and Malbon, C. C. (1995) Phosphorylation of tyrosyl residues 350/354 of the beta-adrenergic receptor is obligatory for counterregulatory effects of insulin. J.Biol.Chem. 270, 25305-25308.

Karoor, V., Wang, L., Wang, H. Y., and Malbon, C. C. (1998) Insulin stimulates sequestration of beta-adrenergic receptors and enhanced association of beta-adrenergic receptors with Grb2 via tyrosine 350. J.Biol.Chem. 273, 33035-33041.

Langeberg, L. K. and Scott, J. D. (2005) A-kinase-anchoring proteins. J.Cell Sci. 118, 3217-3220.

Lefkowitz, R. J. and Shenoy, S. K. (2005) Transduction of receptor signals by beta-arrestins. Science 308, 512-517.

Malbon, C. C., Tao, J., and Wang, H. Y. (2004) AKAPs (A-kinase anchoring proteins) and molecules that compose their G-protein-coupled receptor signalling complexes. Biochem J 379, 1-9.

Michel, J. J. and Scott, J. D. (2002) AKAP mediated signal transduction. Annu.Rev.Pharmacol.Toxicol. 42, 235-257.

Morris, A. J. and Malbon, C. C. (1999) Physiological regulation of G protein-linked signaling Physiol.Rev. 79, 1373-1430.

Shih, M., Lin, F., Scott, J. D., Wang, H. Y., and Malbon, C. C. (1999) Dynamic complexes of beta2-adrenergic receptors with protein kinases and phosphatases and the role of gravin. J.Biol.Chem. 274 , 1588-1595.

Shih, M. and Malbon, C. C. (1998) Serum and insulin induce a Grb2-dependent shift in agonist affinity of beta-adrenergic receptors. Cellular Signalling 10, 575-582.

Singer, S. J. and Nicolson, G. L. (1972) The fluid mosaic model of the structure of cell membranes. Science 175, 720-731.

Tao, J., Wang, H. Y., and Malbon, C. C. (2003) Protein kinase A regulates AKAP250 (gravin) scaffold binding to the beta2-adrenergic receptor. EMBO J 22, 6419-6429.

Wong, W. and Scott, J. D. (2004) AKAP signalling complexes: focal points in space and time. Nat.Rev.Mol.Cell Biol. 5, 959-970.

Zhang, J., Hupfeld, C. J., Taylor, S. S., Olefsky, J. M., and Tsien, R. Y. (2005) Insulin disrupts beta-adrenergic signalling to protein kinase A in adipocytes. Nature 437, 569-573.