Interaction of heterotrimeric G13 protein with an A-kinase-anchoring protein 110 (AKAP110) mediates cAMP-independent PKA activation

1686 Brief Communication

Interaction of heterotrimeric G13 protein with an A-kinaseanchoring protein 110 (AKAP110) mediates cAMP-independent PKA activation Jiaxin Niu*, Rita Vaiskunaite*, Nobuchika Suzuki*, Tohru Kozasa*, Daniel W. Carr†, Nickolai Dulin* and Tatyana A. Voyno-Yasenetskaya* Heterotrimeric G proteins [1] and protein kinase A (PKA) are two important transmitters that transfer signals from a wide variety of cell surface receptors to generate physiological responses. The established mechanism of PKA activation involves the activation of the Gs-cAMP pathway [2]. Binding of cAMP to the regulatory subunit of PKA (rPKA) leads to a release and subsequent activation of a catalytic subunit of PKA (cPKA). Here, we report a novel mechanism of PKA stimulation that does not require cAMP. Using yeast two-hybrid screening, we found that the ␣ subunit of G13 protein interacted with a member of the PKA-anchoring protein family, AKAP110. Using in vitro binding and coimmunoprecipitation assays, we have shown that only activated G␣13 binds to AKAP110, suggesting a potential role for AKAP110 as a G␣ subunit effector protein. Importantly, G␣13, AKAP110, rPKA, and cPKA can form a complex, as shown by coimmunoprecipitation. By characterizing the functional significance of the G␣13-AKAP110 interaction, we have found that G␣13 induced release of the cPKA from the AKAP110-rPKA complex, resulting in a cAMP-independent PKA activation. Finally, AKAP110 significantly potentiated G␣13-induced activation of PKA. Thus, AKAP110 provides a link between heterotrimeric G proteins and cAMP-independent activation of PKA. Addresses: *Department of Pharmacology, University of Illinois, Chicago, Illinois 60612. †Veteran Affairs Medical Center and Oregon Health Sciences University, R&D-8, 3710 SW Veterans Hospital Road, Portland, Oregon 97201. Correspondence: Tatyana Voyno-Yasenetskaya E-mail: [email protected] Received: 11 April 2001 Revised: 30 July 2001 Accepted: 4 September 2001 Published: 30 October 2001 Current Biology 2001, 11:1686–1690 0960-9822/01/$ – see front matter  2001 Elsevier Science Ltd. All rights reserved.

Results and discussion G␣13 interacts with A-kinase-anchoring protein 110 (AKAP110) in a yeast two-hybrid screen

To discover novel proteins that might participate in signaling via heterotrimeric G proteins, a two-hybrid screen was performed on a human testis library using G␣13 as bait. Because effector molecules interact with high affinity to the GTP-bound G␣ subunits, we have used the activated form of G␣13 (G␣13Q226L) as bait. One of the positive clones from this screen was purified, sequenced, and identified as a fragment of A-kinase-anchoring protein, AKAP110 (aa 683–853). AKAP110 was cloned recently and was initially found in sperm cells [3]. Since then, AKAP110 (and G␣13) was detected in the granular and Purkinje neurons of the cerebellum by Western blotting, indirect immunostaining, immunofluorescence, and RTPCR studies (D. Carr and T. Voyno-Yasenetskaya, unpublished data). To determine the specificity of the interaction of G␣13 with AKAP110, plasmids containing domain fragments of AKAP110 were transformed together with either G␣13 or G␣12 into the yeast reporter strain EGY48, and cotransformants were selected for ␤-galactosidase activity at 30⬚C for 4 days. ␤-galactosidase activity data indicated that G␣13, but not G␣12, interacted with the AKAP110 fragments (Figure 1a). The absence of interaction between AKAP110 and G␣12 was specific, because the same construct was used in the similar yeast twohybrid screen that yielded eight different positive clones coding for novel G␣12-interacting proteins (R. Vaiskunaite and T. Voyno-Yasenetskaya, unpublished data). In addition, both the C-terminal domain and full-length AKAP110, but not the N-terminal domain, interacted with G␣13. Thus, our data show that G␣13 specifically interacts with the C-terminal domain of AKAP110. G␣13 directly interacts with AKAP110 in an activationdependent manner in vitro

To confirm that the carboxyl-terminal domain of AKAP110 interacted directly with G␣13, we examined their binding using purified components in vitro. GST fusion proteins of both full-length AKAP110 and the putative G␣13 binding domain interacted with purified G␣13 (Figure 1b). Importantly, G␣13 bound to GST-AKAP110 (FL) and GST-AKAP110 (C-350) only in the presence of AlF4⫺, an activator of the G␣ subunit that promotes a conformation similar to that of the transition state for GTP hydrolysis [4]. These data suggest that AKAP110 may function as a G␣ subunit effector protein.

Brief Communication 1687

Figure 1

G␣13 interacts with AKAP110. (a) A specific interaction of AKAP110 with G␣13 in yeast two-hybrid system. pB42AD constructs containing full-length (AKAP-FL), 1–683 aa (AKAP-N) regions, and 685–853 aa (AKAP-C) regions of AKAP110 as indicated were transformed into yeast reporter strain EGY48 either with or without pLexA-G␣13Q226L or pLexA-G␣12Q229L. Cotransformants were selected and underwent a ␤-galactosidase assay. The blue color indicates a protein-protein interaction. (b) The domain structure of AKAP110. The domain that binds the regulatory subunit II of PKA, the RII binding domain, is located between amino acids 124 and 141 [3]. The putative G␣13 binding domain is located at 685–853 aa. (c) Direct interaction of purified G␣13 protein with AKAP110 GST fusion proteins in vitro. 1 ␮M purified GST or GST-AKAP110 fusion proteins bound to glutathione-sepharose beads were mixed with 0.25 ␮M purified G␣13-GDP in 100 ␮l binding buffer to perform an in vitro binding assay as described in the Materials and methods. Where indicated, AlF4⫺ was added. Glutathione-sepharose

beads were pulled down by centrifugation, and bound G␣13 was separated by SDS-PAGE and immunoblotted with antibody against G␣13. C-350, C-terminal 350 amino acids; FL, full-length AKAP10. (d) PKA, AKAP110, and G␣13 form a protein complex in vivo. Coimmunoprecipitation of AKAP110 with G␣13. HEK-293 cells grown in 10-cm2 dishes were transiently transfected with vector alone or with various Myc-tagged constructs of AKAP110 together with wildtype HA-tagged G␣13. A total of 48 hr after transfection, cells were lysed, an equal amount of cell lysates was immunoprecipitated with agarose-conjugated anti-Myc antibodies in the presence or absence of AlF4⫺. Immunoprecipitates and cell lysates were analyzed by Western blotting using antibodies against Myc or HA. (e) Coimmunoprecipitation of AKAP110 with endogenous rPKA. The same immunoprecipitates were probed with Myc antibodies or antibodies against RII. All experiments (b–e) were performed 3–4 times with similar results.

G␣13 and AKAP110 form a protein complex in intact cells

and C-350 were ⵑ1:4. Thus, it is likely that apparently weak interactions could be a result of lower expression or stability of C-terminal domains of AKAP110. Both short (685–853 aa) and long (503–853 aa) C-terminal fragments, but not N-terminal (1–500 aa and 1–350 aa) fragments, bind to G␣13. The ability of the N-terminal domain to interact with the RII subunit was performed as a positive control for the N-terminal domain activity (see the Supplementary material available with this article online). Thus, we have mapped the G␣13 binding site of AKAP110 to a region between residues 685 and 853 (Figure 1e). Consistent with results from the two-hybrid screen and in vitro binding, full-length AKAP110 bound only to an

To assess the in vivo interaction, we assayed the physical association between G␣13 and AKAP110 by coimmunoprecipitation from human kidney embryonic 293 (HEK293) cells transfected with Myc-tagged AKAP110 constructs and HA-tagged G␣13. Western blotting showed that the expression of full-length AKAP110 is higher than the expression of any AKAP110 construct (Figure 1e). As it is not unusual for protein domains to be less stable than full-length protein, we have determined relative intensities of the immunoprecipitated G␣13 and AKAP110 constructs’ bands by densitometry using ImageQuant software. The ratios of G␣13 to AKAP110 full-length, C-170,

1688 Current Biology Vol 11 No 21

AlF4⫺-induced active state of G␣13. Additionally, AKAP110 interacted with the endogenous RII subunit. Interestingly, interaction of the G␣13 with AKAP110 did not affect the interaction of RII with AKAP110 (Figure 1f). These data suggest that AKAP110 can interact simultaneously with RII and G␣13 proteins and that G␣13 binding does not affect RII binding to AKAP110.

Figure 2

AKAP110 does not regulate GTPase activity or rate or nucleotide exchange activity

Some effector proteins, such as phospholipase C and cGMP phosphodiesterase ␥ subunit, also act as GTPaseactivating proteins (GAPs) that accelerate deactivation of heterotrimeric G proteins [5, 6]. Therefore, we tested whether AKAP110 affects the ability of G␣13 to hydrolyze GTP. AKAP110 did not stimulate GTPase activity of G␣13 over the untreated control (Figure 2a). The positive control, p115RhoGEF, significantly stimulated GTPase activity at all time points. We also tested whether AKAP110 can regulate the rate of nucleotide exchange of G␣13 protein. In the absence of a receptor, the ratelimiting step for binding of the nonhydrolyzable GTP analog GTP␥S to the G␣ protein is the release of GDP. The full-length AKAP110 did not affect GTP␥S binding by G␣13 (Figure 2b). Therefore, AKAP110 does not regulate GTPase activity or rate or nucleotide exchange activity. Rho guanine nucleotide exchange factor, p115RhoGEF, is a recently identified G␣13 effector. G␣13 directly stimulates p115RhoGEF, leading to Rho activation [7, 8]. We next tested whether AKAP110 could affect the guanine nucleotide exchange of Rho protein. Consistent with previous reports, p115RhoGEF induced GTP␥S binding to Rho. This effect increased in the presence of AlF4⫺-activated G␣13. However, AKAP110 did not affect GTP␥S binding by Rho (data not shown). Although many G␣13induced signaling events are Rho dependent, it appears unlikely that AKAP110 is involved in G␣13-induced Rho activation. G␣13 induces release of the PKA catalytic subunit from the AKAP110-RII complex

To investigate the functional consequences of the interaction between AKAP110 and G␣13 in a heterologous expression system, HEK-293 cells were transfected with cDNAs for AKAP110 (tagged with MYC) and were G␣13 tagged with an HA epitope. Expression of AKAP110 and G␣13 did not change the amount of endogenous cPKA or rPKA, as determined by Western blotting of total cell lysates (data not shown). Both endogenous rPKA and cPKA coprecipitated with AKAP110 (Figure 3a). PKA catalytic activity in immunoprecipitated material was determined in the absence or presence of 0.5 mM cAMP. PKA activity was defined as the activity inhibited by 4 ␮M PKA inhibitor PKI (Figure 3a) and was enriched by 58.4 ⫾ 5-fold in the presence of cAMP (n ⫽ 7). We then cotrans-

AKAP does not affect either guanine nucleotide binding to G␣13 or the GTPase activity of G␣13. (a) The GTPase activity of G␣13 was assayed in the presence of 400 nM full-length AKAP110 (closed circles) or 20 nM p115 RhoGEF (open square) or control, GST protein (open circles). (b) Binding of GTP␥S to G␣13 (100 nM) was performed by incubating the proteins at 30⬚C for 60 min in the presence or absence of 500 nM full-length AKAP110. All experiments (a,b) were performed 2–3 times with similar results.

fected the increasing concentrations of constitutively activated G␣13. Interestingly, both PKA activity and the amount of cPKA associated with AKAP110 was decreased, while the amount of RII detected by Western blotting did not change in the presence of G␣13Q226L (Figure 3a). Importantly, wild-type G␣13 did not affect PKA activity associated with AKAP110 (Figure 3a). Only G␣13Q226L, but not G␣iQ205L or G␣12Q229L, reduced PKA activity associated with AKAP110, indicating that this effect by G␣13 is specific (Figure 3b). To confirm the observation that G␣13 induces the release of cPKA from the AKAP110 complex, AKAP110 was immunoprecipitated from HEK-293 cells transiently transfected with Myc-tagged AKAP110. Preincubation of the

Brief Communication 1689

Figure 3

G␣13 induces the release of the PKA catalytic subunit from the AKAP110-RII complex. (a) Coprecipitation of PKA with AKAP110. HEK-293 cells were cotransfected with Myc-tagged AKAP110 and different amounts of G␣13Q226L or G␣13-wt as indicated, and were then immunoprecipitated with anti-Myc antibody. Immunoprecipitated material was incubated in the absence (basal) or presence (cAMP) of 0.5 mM cAMP and was assayed for kinase activity. As a control, the plus cAMP samples were also assayed in the presence of 4 ␮M PKA inhibitor PKI (cAMP ⫹ PKI). Values are expressed as means ⫾ SEM of seven separate experiments. The presence of cPKA and RII in immunoprecipitates was detected by Western blot. (b) G␣13Q226L specifically decreased PKA activity associated with AKAP110. HEK-293 cells were transiently transfected with Myc-tagged AKAP110 together with either vector alone, G␣iQ205L, G␣12Q229L, or G␣13Q226L as indicated. They were then lysed and immunoprecipitated with Myc antibodies, followed by measurement of the PKA activity associated with AKAP110. (c) AlF4⫺stimulated G␣13 induces the release of cPKA from the PKAAKAP110 complex in vitro. AKAP110 was immunoprecipitated from transiently transfected HEK-293 cells as described above. Immunoprecipitates were incubated in 40 ␮l kinase buffer in the

presence of 0.5 mM cAMP or AlF4⫺-activated G␣13 at 30⬚C for 30 min, and the reactions were terminated by centrifugation. Both beads and supernatant were analyzed by Western blot using antibodies against Myc, RII, or cPKA. (d) G␣13 specifically stimulated PKA activity in vitro. AKAP110 was immunoprecipitated from transiently transfected HEK-293 cells as described above. Immunoprecipitates were incubated in 40 ␮l kinase buffer in the presence of AlF4⫺-activated G␣13 or AlF4⫺-activated G␣i at 30⬚C for 30 min and were assayed for kinase activity as described in the Materials and methods. (e) G␣13Q226QL does not affect cAMP levels. HEK-293 cells were cotransfected with either empty vector, activated G␣sR145C, activated G␣13Q226L, or full-length AKAP110. The cells were then labeled with [3H]adenine and were subjected to cAMP assay as described in the Materials and methods. (f) AKAP110 potentiates G␣13-induced VASP phosphorylation. Flag-tagged VASP was cotransfected with G␣sR145C, G␣13Q226L, or AKAP110 as indicated. Cells were lysed, and the phosphorylation-dependent electrophoretic mobility shift of VASP was determined by immunoblotting cell lysates with FLAG antibodies. All experiments (b–f) were performed 3–4 times with similar results.

immunocomplexes with 0.5 mM cAMP induced the release of cPKA bound to rPKA into the supernatant (Figure 3c). This release was accompanied by a dramatic increase in PKA activity (Figure 3a). Preincubation of the immunocomplexes with AlF4⫺-activated G␣13 also induced the release of cPKA, as determined by Western blotting and increased kinase activity (Figure 3c,d). In a control experiment, G␣i2 did not affect PKA associated with AKAP110 immunocomplexes (Figure 3d). We have quantitated relative PKA-specific activity as pmol of phosphate incorporated into the PKA substrate malantide per minute per unit of immunoprecipitated cPKA. An arbitrary unit of immunoprecipitated cPKA was determined by Western blotting, followed by densitometry. PKA activity was determined as activity inhibited by 4 ␮M PKI. The halfmaximal PKA-specific activity stimulated with 0.5 mM cAMP was ⵑ60 pmol/min/U cPKA, and the half-maximal

PKA-specific activity stimulated with 0.2 ␮M G␣13 was ⵑ6 pmol/min/U cPKA. These data suggest that purified recombinant G␣13 can induce the release of catalytic subunits from the PKA-AKAP110 complex. This G␣13induced release of the catalytic subunit of PKA from the regulatory subunit of PKA did not require cAMP. Based on this data, we postulated that the G␣13-AKAP110 complex would enhance the cAMP-independent phosphorylation of PKA substrates. The established mechanism of PKA activation in response to various hormones involves stimulatory G proteins, Gs, which activate adenylyl cyclase, resulting in the production of cAMP. Binding of cAMP to rPKA leads to the release and activation of cPKA [2]. However, we have previously reported that G␣13 does not induce the production of intracellular cAMP in HEK-293 cells [9]. This observation was con-

1690 Current Biology Vol 11 No 21

Figure 4

A proposed model for G protein activation of PKA. Activated G␣13 binds to AKAP110 and induces the release of the catalytic subunit from the regulatory II subunit of PKA.

firmed by determining cAMP levels in HEK-293 cells transfected with either constitutively activated G␣s, G␣13, or AKAP110. Consistent with previously reported data, G␣13 did not induce cAMP production (Figure 3e). AKAP110 alone or in combination with G␣13Q226L also did not affect the cAMP accumulation, these data are consistent with our model of a cAMP-independent activation of PKA. As vasodilator-stimulated phosphoprotein (VASP) is a PKA substrate that is involved in the reorganization and/or polymerization of actin filaments [10], we then investigated the effect of G␣13 and AKAP110 on VASP phosphorylation. As expected, constitutively active G␣s induced VASP phosphorylation (Figure 3f). Surprisingly, G␣13Q226L alone also induced VASP phosphorylation, although it was less pronounced. Importantly, AKAP110 alone did not induce VASP phosphorylation, but it dramatically potentiated G␣13-induced VASP phosphorylation. In conclusion, these data suggest that G␣13 protein directly interacts with AKAP110, leading to cyclic AMPindependent activation of protein kinase A (see model, Figure 4), revealing a new mechanism for the action of G proteins. It is possible that the G protein-coupled receptors that are not coupled to the Gs-cAMP pathway could potentially stimulate PKA, which in turn regulates a vast variety of cell functions. Our recent report [11], showing that endothelin I and angiotensin II receptors that are not coupled to G␣s could stimulate PKA, supports this hypothesis.

Another interesting possibility is that signal specificity is achieved (in addition to other mechanisms) via the interaction of ubiquitous signaling molecules with tissuespecific signaling molecules. Such signal specificity is probably the case in the G␣13-AKAP110 interaction, in which G␣13 is ubiquitously expressed, and AKAP110 expression is restricted to a few cell types. Thus, based on data showing that G␣13 regulates the actin cytoskeleton and that disruption of the AKAP110-PKA interaction blocks the sperm motility, we predict that the G␣13AKAP110 interaction could be important in the regulation of sperm motility. An additional important physiological question about the role of the G␣13-AKAP110 interaction in the cerebellar neurons is based on data showing that G␣13 induces growth cone collapse in neuronal cell lines and AKAP110 is located in cerebellum neurons. Finally, a major challenge for future studies is to determine the molecular mechanism of G␣13-dependent release of cPKA from the AKAP110-RII complex. Supplementary material Supplementary material including the Materials and methods, which includes details concerning yeast two-hybrid screening, protein purification, in vitro binding assay, immunoprecipitation, and Western blotting, [35S]GTP␥S binding assay, gap assay, cyclic AMP assay, in vitro kinase assay, and PKA activity in intact cells, is available at http:// images.cellpress.com/supmat/supmatin.htm.

Acknowledgements We thank Sarah Andrews for editing this manuscript. This work was supported by grants from the National Institutes of Health: GM56159 (T.V.-Y.), GM59427 (T.K.), and HD36408 (D.W.C.). J.N. was supported by predoctoral fellowship 0110205Z from the American Heart Association.

References 1. Gilman AG: G proteins: transducers of receptor-generated signals. Ann Rev Biochem 1987, 56:615-649. 2. Taylor SS, Buechler JA, Yonemoto W: cAMP-dependent protein kinase: framework for a diverse family of regulatory enzymes. Annu Rev Biochem 1990, 59:971-1005. 3. Vijayaraghavan S, Liberty GA, Mohan J, Winfrey VP, Olson GE, Carr DW: Isolation and molecular characterization of AKAP110, a novel, sperm-specific protein kinase A-anchoring protein. Mol Endocrinol 1999, 13:705-717. 4. Berman DM, Kozasa T, Gilman AG: The GTPase-activating protein RGS4 stabilizes the transition state for nucleotide hydrolysis. J Biol Chem 1996, 271:27209-27212. 5. Berstein G, Blank JL, Jhon DY, Exton JH, Rhee SG, Ross EM: Phospholipase C-␤1 is a GTPase activating protein (GAP) for Gq/11, its physiologic regulator. Cell 1992, 70:411-418. 6. Arshavsky VY, Bownds MD: Regulation of deactivation of photoreceptor G protein by its target enzyme and cGMP. Nature 1992, 357:416-417. 7. Kozasa T, Jiang X, Hart MJ, Sternweis PM, Singer WD, Gilman AG, et al.: p115 RhoGEF, a GTPase activating protein for G␣12 and G␣13. Science 1998, 280:2109-2111. 8. Hart MJ, Jiang X, Kozasa T, Roscoe W, Singer WD, Gilman AG, et al.: Direct stimulation of the guanine nucleotide exchange activity of p115 RhoGEF by G␣13. Science 1998, 280:21122114. 9. Voyno-Yasenetskaya TA, Conklin BR, Gilbert RL, Hooley R, Bourne HR, Barber DL: G␣13 stimulates Na/H exchange. J Biol Chem 1994, 269:4721-4724. 10. Vasioukhin V, Fuchs E: Actin dynamics and cell-cell adhesion in epithelia. Curr Opin Cell Biol 2001, 13:76-84. 11. Dulin N, Niu J, Browning D, Ye R, Voyno-Yasenetskaya TA: Cyclic AMP-independent activation of protein kinase A by vasoactive peptides. J Biol Chem 2001, 276:20827-20830.