Kelch-repeat proteins interacting with the Gα protein Gpa2 bypass adenylate cyclase for direct regulation of protein kinase A in yeast
- Laboratory of Molecular Cell Biology, Katholieke Universiteit Leuven, and Department of Molecular Microbiology, Flanders Interuniversity Institute of Biotechnology (VIB), Kasteelpark Arenberg 31, B-3001 Leuven-Heverlee, Belgium
-
Edited by Gregory A. Petsko, Brandeis University, Waltham, MA, and approved July 5, 2006 (received for review November 7, 2005)
Abstract
The cAMP–PKA pathway consists of an extracellular ligand-sensitive G protein-coupled receptor, a G protein signal transmitter, and the effector, adenylate cyclase, of which the product, cAMP, acts as an intracellular second messenger. cAMP activates PKA by dissociating the regulatory subunit from the catalytic subunit. Yeast cells (Saccharomyces cerevisiae) contain a glucose/sucrose-sensitive seven-transmembrane domain receptor, Gpr1, that was proposed to activate adenylate cyclase through the Gα protein Gpa2. Consistently, we show here that adenylate cyclase binds only to active, GTP-bound Gpa2. Two related kelch-repeat proteins, Krh1/Gpb2 and Krh2/Gpb1, are associated with Gpa2 and were suggested to act as Gβ mimics for Gpa2, based on their predicted seven-bladed β-propeller structure. However, we find that although Krh1 associates with both GDP and GTP-bound Gpa2, it displays a preference for GTP-Gpa2. The strong down-regulation of PKA targets by Krh1 and Krh2 does not require Gpa2 but is strictly dependent on both the catalytic and the regulatory subunits of PKA. Krh1 directly interacts with PKA by means of the catalytic subunits, and Krh1/2 stimulate the association between the catalytic and regulatory subunits in vivo. Indeed, both a constitutively active GPA2 allele and deletion of KRH1/2 lower the cAMP requirement of PKA for growth. We propose that active Gpa2 relieves the inhibition imposed by the kelch-repeat proteins on PKA, thereby bypassing adenylate cyclase for direct regulation of PKA. Importantly, we show that Krh1/2 also enhance the association between mouse R and C subunits, suggesting that Krh control of PKA has been evolutionarily conserved.
In budding yeast, PKA plays a major role in the control of growth, metabolism, stress resistance, and filamentous differentiation in response to nutrient availability. Nutrient-deprived quiescent cells or slow-growing cells on a nonfermentable carbon source accumulate high levels of storage carbohydrates (such as trehalose and glycogen), induce the expression of stress response element-controlled genes, and tolerate various stress conditions well. Cells cultured on complete medium with a fermentable carbon source, such as glucose, show the opposite phenotype (1, 2). Pseudohyphal differentiation occurs in response to nitrogen limitation in the presence of a rapidly fermentable sugar (3). All of these characteristics are in large part determined by the activity of PKA. Mutants with constitutively high PKA activity are hyperfilamentous, display the characteristics of fermenting cells, and are deficient in respiration and entry into stationary phase, whereas mutants with lowered PKA activity cannot switch to a filamentous form, grow slowly, and resemble stationary-phase cells in many aspects, even when supplied with a rapidly fermentable carbon source (4, 5). When glucose is added to glucose-deprived cells, cAMP rapidly and transiently accumulates and triggers the activation of PKA (6). PKA is an inactive tetramer composed of two regulatory subunits and two catalytic subunits in the absence of cAMP (7). Upon binding of cAMP to the regulatory subunits, encoded by the BCY1 gene, the partially redundant catalytic subunits encoded by TPK1, TPK2, and TPK3 dissociate and become active (8, 9).
In Saccharomyces cerevisiae, the synthesis of cAMP from ATP is catalyzed by a single adenylate cyclase enzyme, Cyr1 (10). The degradation of cAMP is catalyzed by low- and high-affinity phosphodiesterases, encoded by PDE1 and PDE2, respectively (11, 12). Adenylate cyclase activity in yeast is stimulated by the small GTPases Ras1 and Ras2 (13). By analogy to Gαs stimulation of adenylate cyclase in animal cells, it was proposed that the Gα protein Gpa2 activates Cyr1, based on the requirement of Gpa2 for glucose-induced cAMP signaling (14) and pseudohyphal differentiation (15). Gpa2 is coupled to the seven-transmembrane receptor Gpr1 (16, 17), which can be activated by extracellular sucrose (EC50 ≈ 0.5 mM) and glucose (EC50 ≈ 20 mM) (18). Gpa2 is turned off by the action of an RGS (regulator of G protein signaling) protein, called Rgs2, which stimulates the intrinsic GTPase activity on Gpa2 (19).
Gpa2 appears to differ from other well characterized Gα proteins in that it does not associate with canonical Gβγ subunits (20, 21). Recently, two related interaction partners of Gpa2 have been described: Krh1/Gpb2 and Krh2/Gpb1 (22, 23). Henceforward, we refer to these proteins as Krh (kelch repeat homologue) 1 and Krh2, as argued in the first paragraph of Results. Mutants lacking Krh1/2 display a high-PKA phenotype, suggesting that these proteins function antagonistically to Gpa2. FLO11, a gene involved in pseudohyphal and invasive growth, is highly up-regulated in the double deletion mutant, resulting in precocious filamentation. Krh1 and Krh2 contain seven kelch repeats that are predicted to fold as a seven-bladed β-propeller, superficially similar to the three-dimensional structure of seven tandem WD-40 repeats in Gβ proteins, despite a lack of primary sequence identity. Based on this structural similarity, the preferential binding of Krh1 and Krh2 to the GDP-bound state of Gpa2 and the partial attenuation of the phenotype of the single deletion mutants by additional deletion of GPA2, Krh1, and Krh2 were proposed to function as Gβ mimics for Gpa2 (22). Recently, the same researchers reported that Krh1 impairs efficient coupling of Gpa2 to its cognate receptor, Gpr1, by binding to the extended N terminus of Gpa2 (24). However, absence of Gpa2 does not suppress the high expression of FLO11 or the hyperfilamentation in a double krh1/2Δ mutant, suggesting that these kelch-repeat proteins largely act downstream or independently of Gpa2 (23). On the other hand, deletion of TPK2, encoding the PKA catalytic subunit that is specifically required for filamentous growth (5, 25), largely abrogates the hyperinvasive phenotype of the krh1/2Δ mutant (22, 23), arguing that Krh1 and Krh2 function upstream of PKA.
In this study, we find that both adenylate cyclase and Krh1 preferentially bind to active, GTP-bound Gpa2, suggesting that both proteins are distinct Gpa2 effectors. Consistently, constitutively active Gpa2, or deletion of KRH1 and KRH2, reduces the cAMP requirement for growth of an adenylate cyclase mutant. The function of Krh1 and Krh2 strictly requires both the catalytic and regulatory subunits of PKA in vivo, and Krh1 directly interacts with PKA in vitro. We propose that, in addition to stimulating adenylate cyclase, active Gpa2 relieves the Krh1/2-stimulated association of the regulatory and catalytic subunits of PKA.
Results
Krh1 and Krh2 Potently Down-Regulate PKA Targets, Even in the Absence of Gpa2.
In a two-hybrid screen using myristoylation-deficient Gpa2(G2A) as the bait protein, we isolated a fragment of Krh1 (amino acids 531–740). The yeast genome encodes one protein that is closely related in sequence to Krh1, Krh2, which also interacts with Gpa2(G2A) in the two-hybrid system (data not shown). Both proteins are predicted to have seven tandem kelch repeats in the second half of the proteins (for a detailed protein domain structure, see refs. 22 and 23). The Gα protein Gpa2 is required for glucose-induced activation of cAMP synthesis and PKA activation (14). To investigate the functional importance of these proteins in Gpa2 signaling, we deleted KRH1 and KRH2 and determined typical PKA-associated parameters when cells approach stationary phase. Consistent with the hyperfilamentous phenotype of mutants lacking Krh1/2 (22, 23), indicative of high PKA activity, absence of the kelch-repeat proteins resulted in constitutively low levels of the reserve carbohydrates trehalose and glycogen and low expression of HSP12 (Fig. 1; glycogen data not shown). In addition, trehalase activity, a direct readout for PKA activity, was much higher in the krh1/2Δ mutant compared with WT (Fig. 8A, which is published as supporting information on the PNAS web site). Deletion of GPA2, however, results in a low-PKA phenotype. Krh1 and Krh2 have been proposed to act as Gβ mimics for the Gα protein Gpa2 (22, 24), which implies that Krh1/2 require the presence of Gpa2 to exert their function. However, absence of GPA2 did not prevent the decrease in trehalose and glycogen levels or the expression of HSP12 that is observed when KRH1/2 are deleted (Fig. 1). Similarly, absence of the Gpr1 receptor could not prevent the dramatic decrease in trehalose or glycogen levels in a krh1/2 deletion background (data not shown). The Gpa2Q300L mutant is deficient in GTP hydrolysis (this mutation is analogous to Giα Q204L), and, as expected, this mutant causes a dominant hyperactive PKA phenotype, as evidenced by constitutively low trehalose levels and low expression of HSP12, although trehalose and HSP12 levels are slightly higher than in the krh1/2Δ mutant (Fig. 1). The combined effect of the absence of Krh1/2 and constitutively active Gpa2 in these assays cannot be measured, because no detectable trehalose accumulation or HSP12 expression was observed in the krh1/2Δ mutant by itself. We conclude that Krh1 and Krh2 largely act in parallel or downstream of Gpa2. For this reason, we have adopted the neutral terminology “Krh” (kelch repeat homologue, as in ref. 23) throughout this article rather than the suggestive “Gpb” (G protein β; ref. 22).
Gpa2 and Krh1/2 act antagonistically and independently on several PKA targets. (A) The strains BY4742, Y10152, TP261, TP675, TP676, and TP677 were grown to stationary phase on YPD; samples were taken after 1, 2, and 3 days (white, gray, and dark gray bars, respectively); and trehalose content was determined (OD600 and glucose consumption were followed to verify that all strains reached saturation at approximately the same time). (B) The strains were grown and samples were taken as in A. Expression of HSP12 and ACT1 (as an internal standard) were measured by quantitative PCR. (C and D) The strains S18-1D, TP626, and TP701 were grown to stationary phase on yeast extract/peptone glycerol. cAMP (C) and trehalase (D) activity were measured after addition of 100 mM glucose.
The cAMP signal induced by addition of glucose to glucose-deprived cells was reported to be longer lasting when either Krh1 or Krh2 is absent, supporting a negative regulatory role of Krh1 and Krh2 on Gpa2 (22). However, as illustrated in Fig. 8B, we did not detect significant differences in glucose-induced cAMP accumulation between the WT and the krh1Δ or krh2Δ mutant. In agreement with ref. 22, glucose-induced cAMP signaling in a krh1/2Δ double mutant was somewhat attenuated compared with the WT, consistent with increased negative feedback regulation on cAMP levels mediated by the high PKA activity in the krh1/2Δ mutant (26, 27). Even in a strain with attenuated PKA activity (and, hence, reduced feedback inhibition), absence of Krh1/2 did not increase cAMP levels, in sharp contrast to expression of an overactive mutant in Ras, a well established activator of adenylate cyclase, which caused a dramatic increase in cAMP levels (Fig. 1 C). Although deletion of Krh1/2 did not increase cAMP levels, it did cause a similar increase in trehalase activity (Fig. 1 D) and, consequently, reduced trehalose (see Fig. 4 B), as the overactive Ras mutant in the Tpkw mutant, indicative of increased PKA activity. We conclude that Krh1/2 down-regulate PKA without affecting cAMP levels.
We also tested whether the GTPase activity of Gpa2 is affected by Krh1 binding. The steady-state GTPase rate on Gα proteins in vitro, in the absence of regulatory factors such as guanine nucleotide exchange factors or GTPase-activating proteins, is limited by nucleotide exchange rather than GTP hydrolysis (28). Canonical Gβγ subunits stabilize GDP-Gα and prevent GDP-for-GTP exchange, thereby reducing overall GTP turnover (29). However, Gpa2 and a stoichiometric complex of Gpa2 and the kelch repeats of Krh1 displayed undistinguishable steady-state GTP turnover (Fig. 8C).
Both Adenylate Cyclase and Krh1 Preferentially Bind to GTP-Gpa2.
The results presented above suggest that Krh1 is an effector of Gpa2, which controls PKA targets independently of cAMP synthesis, implying that Krh1 binds preferentially to activated Gpa2. Contradictory data have been reported with regard to the binding preference of Krh1 for GDP-Gpa2 versus GTP-Gpa2. In an in vitro binding assay, Krh1 and Krh2 were found to bind to both GDP- and GTP-loaded Gpa2, but they bound with the highest affinity to GDP-Gpa2 (22). On the other hand, in the yeast two-hybrid system, a constitutively active Gpa2 mutant (Gpa2R273A) was reported to associate better with the C-terminal half of Krh1 than WT Gpa2 (23). Furthermore, adenylate cyclase (Cyr1) has been suggested to be an effector of Gpa2, based on genetic data (14). Nonetheless, in contrast to the well established control of Cyr1 by Ras in yeast, stimulation of adenylate cyclase activity by Gpa2 or even physical interaction between Gpa2 and Cyr1 has never been demonstrated. Two groups have recently reported that Schizosaccharomyces pombe Gpa2 physically associates with adenylate cyclase, although the precise interaction site in adenylate cyclase and the nucleotide dependence of this interaction is controversial (30, 31). We decided to reinvestigate the guanine nucleotide dependence of the association between Gpa2 and Krh1 and to include Cyr1 in the binding assays. GST-Gpa2 was purified from Escherichia coli, loaded with either GDP or the nonhydrolyzable GTP analog GTPγS, and incubated with yeast extract containing either Krh1-HA3 or HA3-Cyr1. Cyr1 bound to GTP-Gpa2 and not detectably to GDP-Gpa2 (Fig. 2 A), providing evidence of the biochemical basis for the role of Gpa2 in glucose-induced cAMP signaling. Moreover, this clear differential interaction confirmed that nucleotide loading onto GST-Gpa2 had occurred efficiently. Krh1 bound to both GDP and GTP-loaded Gpa2, with a modest but reproducible preference for GTP-Gpa2 (Fig. 2 A). The same results were obtained by using His-6-Krh1 purified from bacteria (data not shown). Similar to GTP-Gpa2, loading of GST-Gpa2 with GDP-AlF4 − resulted in the same preference of HA3-Cyr1 and Krh-HA3 for this GTP-hydrolysis transition state of Gpa2 over GDP-Gpa2. To resolve the discrepancy between these findings and a previous report (22), we decided to assess the interaction between Gpa2 and Krh1 in a second manner. HA3-Cyr1 or Krh1-HA3 was coexpressed with either GST-Gpa2 or GST-Gpa2Q300L in yeast, and the amount of HA-tagged protein that bound to the GST fusions was determined. A strict preference of HA3-Cyr1 for Gpa2Q300L over WT Gpa2 was observed (Fig. 2 B). In the case of Krh1, a modest preference for the constitutive GTP-bound Gpa2Q300L mutant was observed, in agreement with the previous assay. We conclude that adenylate cyclase is a bona fide effector of Gpa2, which binds only to active GTP-Gpa2. Krh1 binds directly to both nucleotide-bound states of Gpa2 but also displays a preference for the GTP-bound form.
Both adenylate cyclase and Krh1 preferentially bind to active Gpa2. (A) GST and GST-Gpa2 were purified from bacteria, preloaded with the indicated nucleotide, and incubated with yeast extracts from the strains MV601 and MV600, expressing HA3-Cyr1 and Krh1-HA3, respectively. Quantification indicates that the intensity of the GDP-bound fraction of Krh1-HA is 61% of the GTP-bound fraction. The input fraction represents 10% of the total amount added to each binding reaction. (B) (Left) Extracts from yeast strains expressing GST-Gpa2 and GST-Gpa2Q300L were mixed with extract from strain MV601 containing HA3-Cyr1 and incubated with glutathione-agarose. (Right) GST-Gpa2 and GST-Gpa2Q300L were coexpressed with Krh1-HA3 (using strains MV605 and MV606) and incubated with glutathione-agarose. Quantification indicates that the intensity of the Gpa2-bound fraction of Krh1-HA is 68% of the Gpa2Q300L-bound fraction. The input fraction represents 10% of the total amount added to each binding reaction.
Absence of Kelch Proteins or Overactive Gpa2 Reduces the Requirement of cAMP for Growth in the Absence of Adenylate Cyclase.
To find out how Krh1 and Krh2 down-regulate PKA targets, we determined whether Krh1 and Krh2 still function in the absence of adenylate cyclase. Krh1/2 were deleted in a diploid background that lacked Cyr1 and the high-affinity phosphodiesterase Pde2, and, after sporulation, the progeny was recovered on media supplemented with 3 mM cAMP. We selected cyr1Δ pde2Δ spores with or without additional deletions in KRH1/2 and spotted dilution series of these strains on nutrient plates with decreasing concentrations of cAMP (Fig. 3 A). Strikingly, deletion of Krh1/2 clearly suppressed the growth deficiency of the cyr1Δ pde2Δ mutant at 2 mM and 1 mM exogenously added cAMP, demonstrating that Krh1/2 can function in the absence of adenylate cyclase. However, in the complete absence of cAMP, none of the strains was able to grow, indicating that a low amount of cAMP is required for the growth suppression by deletion of KRH1/2. If Gpa2 controls the activity of Krh1/2, overactive Gpa2 (Gpa2Q300L) might also reduce the cAMP requirement for growth in the absence of adenylate cyclase. Similar to deletion of Krh1 and Krh2, overactive Gpa2 suppresses the growth defect of an adenylate cyclase deletion mutant at low cAMP concentration but not in the complete absence of exogenous cAMP (Fig. 3 B). Finally, in contrast to constitutively active Gpa2, overactive Ras2G19V was unable to suppress the growth deficiency of an adenylate cyclase deletion mutant at low cAMP concentrations (Fig. 3 B), confirming that Ras activates PKA only through stimulation of adenylate cyclase (13).
Absence of Krh1 and Krh2 or activation of Gpa2 reduces the cAMP requirement for growth of an adenylate cyclase mutant. The strains with indicated genotypes (TP433 and TP390 in A; TP433, TP506, and TP519 in B) were grown on YPD supplemented with 3 mM cAMP to midexponential phase, spotted in serial 2-fold dilutions on YPD plates with the indicated cAMP concentrations, and incubated at 30°C for 3 days.
Both the Catalytic and the Regulatory Subunits of PKA Are Required for the Function of Krh1 and Krh2.
To determine whether PKA itself is required for the phenotypes caused by absence of Krh1 and Krh2, or whether Krh1/2 may act on components downstream of PKA, we deleted each of the genes encoding the catalytic subunits of PKA (TPK1, TPK2, and TPK3) in a WT or krh1/2Δ background. Deletion of any TPK gene alone did not alter the effect of deletion of KRH1 and KRH2 on trehalose and glycogen levels (data not shown). Also, combined deletion of two TPKs, leaving a single functional TPK gene, did not abrogate the krh1/2 deletion defect (data not shown). Deletion of all three TPKs is lethal, but this lethality can be suppressed by loss of the YAK1 gene (32). We created a quadruple tpk1Δ tpk2Δ tpk3Δ yak1Δ mutant containing either WT or deleted versions of the KRH1 and KRH2 genes, as well as a yak1Δ mutant with or without KRH1/2. Absence of Yak1 alone did not prevent down-regulation of trehalose and glycogen levels by deletion of Krh1 and Krh2 (Fig. 4 A; glycogen data not shown). However, absence of all three TPK genes completely prevented the reduction of trehalose and glycogen by deletion of Krh1/2 (Fig. 4 A; glycogen data not shown). We also found that the decreased expression of HSP12 in a krh1/2Δ mutant could be reversed by deletion of all Tpks (Fig. 9A, which is published as supporting information on the PNAS web site).
Both the catalytic and the regulatory subunits of PKA are required for Krh control. (A) Trehalose determination in strains of the indicated genotypes (TP573, TP575, TP572, and TP571) grown to stationary phase on YPD. Samples were taken during growth on YPD after 1, 2, and 3 days (white, gray, and dark gray bars, respectively; OD600 and glucose consumption were followed to verify that all strains reached saturation at approximately the same time). The data shown in the two graphs were measured in a single experiment but are illustrated separately because of the large difference in trehalose levels in the yak1Δ tpk1–3Δ background compared with the yak1Δ mutant. (B) Trehalose determination in strains of the indicated genotypes (S18-1D, TP408, TP626, RS13-58A-1, TP415, and TP627).
The catalytic subunits of PKA are strongly inhibited by association with their regulatory subunits (encoded by a single BCY1 gene). To determine whether down-regulation of PKA by Krh1 and Krh2 involves the regulatory subunits, we compared the phenotypic consequences of the presence of Krh1/2 in a strain with or without Bcy1. However, a strain lacking BCY1 has such high PKA activity that no trehalose or glycogen can be detected. Therefore, we used strains that have attenuated PKA activity (4). Absence of Krh1/2 still increased trehalase activity (Fig. 1 D) and lowered trehalose levels in a tpk1w mutant containing WT BCY1, indicating that the wimp allele (Tpk1L217S) is still sensitive to Krh regulation. In contrast, deletion of BCY1 in this background completely abrogated the reduction of trehalose levels normally caused by deletion of Krh1/2 (Fig. 4 B). We also tested the effect of deletion of KRH1/2 in a temperature-sensitive PKA mutant lacking Bcy1 (33). KRH1/2 deletion could not suppress the growth defect of this tpk2ts mutant (neither at the semipermissive temperature of 32°C nor at the restrictive temperature of 35°C; data not shown). Moreover, absence of Krh1/2 did not decrease trehalose levels in such a strain either at the permissive temperature (24°C) (not even in stationary-phase cells) or after a 1-h shift to 35°C (Fig. 9B), again indicating that the presence of Bcy1 is essential for the function of Krh1/2.
Krh1 Directly Binds to the PKA Holoenzyme by Means of the Catalytic Subunits.
To investigate the possibility that Krh1 directly regulates PKA, we set up a binding assay between each of the catalytic subunits, or the regulatory subunit of PKA, and Krh1. N-terminal GST fusions of Tpk1, Tpk2, Tpk3, Bcy1, and Gpa2 (as a positive control) were expressed in bacteria, purified, and incubated with yeast extract containing Krh1-HA3. Krh1-HA3 associates with Gpa2, as expected, but Krh1-HA3 was also recovered when either of the Tpks was pulled down (Fig. 5 Left). By contrast, no or very weak interaction was observed with the regulatory subunit, Bcy1. To determine whether Krh1 binds to the PKA holoenzyme or only to free C subunits, we reconstituted PKA by coexpression of His-6-Tpk1 with untagged Bcy1. Krh1 binds to free His-6-tagged Tpk1, confirming the interaction observed in the GST pull-down assay, but a clear interaction was also observed with the Tpk1–Bcy1 complex (Fig. 5 Right). We confirmed that the Krh1–Tpk1 interaction is direct by purifying both proteins from bacteria (Fig. 10, which is published as supporting information on the PNAS web site). In addition, His-6-Krh1 alone or in complex with Gpa2 (prepared by bacterial coexpression; see Fig. 8B) interacted equally well with Tpk1 (data not shown), suggesting that interaction of Krh1 with Gpa2 and Tpk1 is not mutually exclusive. Indeed, we found that Tpk1 interacts with the N terminus of Krh1 (amino acids 1–275), as opposed to the Gpa2–Krh1 interaction, which requires the C-terminal kelch repeats in Krh1 (Fig. 10).
Krh1 binds to PKA by means of the catalytic subunits. (Left) The indicated GST fusions were purified from bacteria and incubated with cell extracts from strain MV600 expressing Krh1-HA3. (Right) His-6-Tpk1 expressed alone in bacteria or coexpressed with untagged Bcy1 was purified and incubated with cell extracts from strain MV600 expressing Krh1-HA3. The input fraction represents 10% of the total amount added to each binding reaction.
Importantly, the mouse PKA Cα subunit interacts with Krh1, either expressed in yeast or purified from E. coli (Fig. 10B). Moreover, mouse Cα was down-regulated in vivo by Krh1/2, as evidenced by a decrease in trehalose levels when KRH1/2 were deleted in a tpk1–3Δ mutant expressing mouse Cα as the sole source of the PKA catalytic subunit (Fig. 10C).
Krh1/2 Enhance the Association Between the Catalytic and Regulatory Subunits of PKA in Vivo.
Because we found that Krh1 can bind to the PKA holoenzyme in vitro and that both the regulatory and the catalytic subunits of PKA are required for the down-regulation of PKA targets by Krh1/2 in vivo, we tested whether Krh1/2 affect the association between Bcy1 and Tpk1. We transformed BCY1 and TPK1, fused to either the Gal4 transactivation domain (AD) or the Gal4 DNA-binding domain (BD), to a KRH WT, a krh1Δ, and a krh1/2Δ two-hybrid reporter strain. Absence of Krh1 and Krh2 strongly reduced the apparent interaction between Tpk1 and Bcy1, as judged by growth on media lacking histidine (data not shown) and a quantitative β-gal assay (Fig. 6). This reduction was clear in either orientation of the two-hybrid system: AD-Bcy1 and BD-Tpk1 or vice versa. Protein levels of the Bcy1 and Tpk1 BD fusion proteins were not affected by deletion of Krh1 alone or Krh1 and Krh2 (Fig. 11, which is published as supporting information on the PNAS web site). Overexpression of BCY1 stabilized the TPK1 fusion protein, as reported in ref. 7. We did not observe significant differences in nucleocytoplasmic distribution of GFP-Tpk1 and GFP-Bcy1 in WT versus a krh1/2Δ mutant (data not shown). Hence, absence of Krh1/2 reduces the interaction between the catalytic and regulatory subunits of PKA in vivo. Moreover, the two-hybrid interaction between mouse Cα and PKA regulatory subunit I is drastically attenuated in the absence of Krh1/2, confirming our finding that yeast Krh1/2 can also regulate mammalian PKA.
Krh1 and Krh2 stimulate the interaction between Tpk1 and Bcy1 in vivo. Two-hybrid reporter strains PJ69-4A, TP636, and TP637 transformed with the indicated GAL4 fusion constructs (R, Bcy1; C, Tpk1), were grown overnight in liquid selective medium (SD-Leu-Trp) and analyzed for β-gal activity and total protein content (A) or spotted on a selective SD-Leu-Trp-Ade plate (B). The values in A are an average of two independent cultures; the error bars represent the range between those values.
We have also determined whether recombinant His-6-Krh1 can inhibit PKA activity at various cAMP concentrations, but we did not detect a significant effect of His-6-Krh1 on PKA activity at any sample point (Fig. 12A, which is published as supporting information on the PNAS web site). This finding may indicate either that an essential factor required for the regulation of PKA by Krh1/2 is missing or that bacterially purified Krh1 must be posttranslationally modified to efficiently inhibit PKA.
Discussion
A Bypass of Adenylate Cyclase from Gpa2 to PKA.
We have detected a previously unidentified mechanism by which the Gα protein Gpa2 activates PKA through two kelch-repeat proteins, bypassing adenylate cyclase stimulation (Fig. 7). Hence, Gpa2 regulates PKA activity via two distinct pathways: through stimulation of adenylate cyclase, as suggested previously based on genetic data (14, 15, 34) and confirmed biochemically in this study, and through inhibition of the Krh proteins. Conceivably, Krh1/2-mediated regulation of PKA is confined to different conditions compared with cAMP activation of PKA. Alternatively, the activity of Krh1/2 may not be controlled by Gpa2 alone. The presence of extracellular glucose/sucrose may be merely one of several conditions needed to effect full inhibition of Krh1/2 and hence fully active PKA. For instance, activation of PKA is sustained only when not only glucose but also other essential nutrients are available (35). Thus, the kelch-repeat proteins could serve as integrators of different signals that impinge on PKA. The predicted seven-bladed β-propeller structure made up of the kelch repeats in Krh1 and Krh2 seems particularly well suited to mediate multiple protein–protein interactions for such an integration function (36).
A model for the role of Gpa2 and Krh1/2 in PKA signaling. Upon activation of Gpa2, PKA is stimulated by means of two distinct mechanisms: direct activation of adenylate cyclase and inhibition of Krh1/2-mediated down-regulation of PKA. Krh1/2 regulate PKA activity by stimulating the association between Tpk1 and Bcy1. As explained in Discussion, other (nutrient) signals besides the availability of sugars may affect Krh regulation of PKA.
A Dual Effect of Krh1/2 on the cAMP–PKA Pathway?
It has been proposed that Krh1 negatively regulates the coupling of Gpa2 to its cognate receptor, Gpr1, by binding to the extended N terminus of Gpa2 (24). This proposal was based on the inhibition by overexpression of Krh1/2 of the plasma membrane recruitment of the C-terminal tail of Gpr1 by a mutant Gpa2 allele that is predicted to be deficient in GTP-induced conformational changes. Here, we find that the effect Krh1/2 exert on PKA targets is quantitatively equivalent in the absence and presence of Gpa2 (see Fig. 1). Negative regulation of Gpa2 by Krh1/2 could be limited to transient Gpr1-Gpa2 signaling. However, the rapid glucose-induced cAMP signal was not affected by deletion of KRH1 or KRH2 in our hands. Our finding that Krh1, despite having a preference for the GTP-bound form of Gpa2, also has affinity for GDP-Gpa2 is consistent with an additional role in GDP-Gpa2, but whether binding of Krh1 to Gpa2 mimics formation of a Gαβ dimer, as suggested previously (22), remains unclear.
Krh1/2 Stimulate the Association Between the Catalytic and Regulatory Subunits of PKA.
A previous report showed that the hyperfilamentous phenotype of a krh1/2Δ mutant can be largely suppressed by deletion of TPK2 (22, 23). In this study, we have extended this finding by demonstrating that, in a krh1/2Δ mutant, the up-regulation of other PKA targets that are controlled by all three Tpks, such as reserve carbohydrate mobilization and STRE (stress response element) gene repression, can be completely suppressed by deletion of all three TPKs. Moreover, we found that the regulatory subunit of PKA, Bcy1, is required for the krh1/2Δ phenotype. In addition, we have demonstrated that PKA itself is the direct target of Krh1 by showing a physical interaction between Krh1 and Tpk1. Finally, we have shown that Krh1/2 stimulate the association between Tpk1 and Bcy1 in vivo. Hence, Krh1/2 antagonize cAMP-mediated dissociation of the PKA holoenzyme. At least a small amount of cAMP is required to enable Krh1/2 control on PKA, presumably reflecting that, in the absence of cAMP, PKA is completely inactive, regardless of Krh1/2. Moreover, the requirement of a minimal amount of cAMP highlights the difference between Krh1/2 and downstream suppressors of PKA, such as the protein kinases Yak1 and Rim15, the deletion of which completely suppresses cyr1 lethality, and reinforces the idea that Krh1/2 regulate PKA itself.
Possible Conservation of Krh Control of PKA.
We describe a mechanism that allows a Gα protein to bypass adenylate cyclase and activate PKA by means of kelch-repeat proteins (Fig. 7). If this mechanism turns out to be conserved, it constitutes a unique concept in PKA signaling. Clear full-length sequence orthologs of Krh1 and Krh2 can be found in fungal genomes but cannot easily be recognized in the genomes of animals or plants. However, a large family of proteins has been found in the human genome that contains six or seven tandem kelch repeats preceded by an N-terminal extension (36). A subset of these kelch proteins is involved in targeting proteins for degradation (41), but many others have not been characterized in any detail. Our finding that Krh1 directly binds to mouse PKA Cα and can down-regulate it in yeast suggests that a Krh interaction motif has been conserved in mammalian PKA C subunits.
Materials and Methods
Plasmids, Strains, and Growth Conditions.
Yeast strains and plasmids used in this study are listed in Tables 1 and 2, which are published as supporting information on the PNAS web site. Standard growth media containing 2% glucose, 2% raffinose, 2% glycerol, or 2% galactose as the carbon source are described in Supporting Methods, which is published as supporting information on the PNAS web site. Yeast two-hybrid strains were transformed with specific combinations of pGAD424 and pGBT9 constructs (see Table 2). Determination of β-gal activity in extracts of these strains was performed according to standard procedures.
Biochemical Analyses.
Quantitative determination of glycogen, trehalose, trehalase activity, and cAMP levels were performed as reported in ref. 14. Steady-state GTPase assays of Gpa2 are described in ref. 19. The PKA assay and a list of antibodies is described in Supporting Methods.
Quantitative PCR and RT-PCR.
Quantitative PCR and RT-PCR are described in Supporting Methods.
Protein Purifications.
GST- and His-6-fusion proteins and protein complexes were purified from E. coli strain BL21(DE3)-STAR (Invitrogen) with glutathione-Sepharose and Ni2+-agarose, respectively, essentially as described in ref. 42. In the case of Ni2+ purification, imidazole eluates were desalted by using Sephadex G-25 (Amersham Pharmacia, Uppsala, Sweden) in a buffer consisting of 50 mM NaCl/50 mM Tris·HCl, pH 8.0.
Acknowledgments
We thank J. Hirsch (Mount Sinai School of Medicine, New York, NY) for sharing results before publication; S. McKnight (University of Washington, Seattle, WA), T. Den Abt (Katholieke Universiteit Leuven), and G. Griffioen (Katholieke Universiteit Leuven) for providing plasmids; F. Rolland and K. Voordeckers for critically reading the manuscript; and W. Verheyden, V. Maes, R. Wicik, and M. De Jonge for excellent technical assistance. This work was supported by predoctoral fellowships from the Institute for the Promotion of Innovation by Science and Technology in Flanders (to T.P., D.N., and R.G.), a Return Grant from the Belgian Federal Science Policy Office (to M.V.), grants from the Fund for Scientific Research–Flanders, the Research Fund of Katholieke Universiteit Leuven (Concerted Research Actions), and Interuniversity Attraction Poles Network P5/30 (to J.M.T.).
Footnotes
- *To whom correspondence should be addressed. E-mail: matthias.versele{at}bio.kuleuven.be
-
Author contributions: T.P., W.L., R.G., D.N., J.M.T., and M.V. designed research; T.P., W.L., R.G., D.N., and M.V. performed research; T.P., W.L., R.G., D.N., J.M.T., and M.V. analyzed data; and J.M.T. and M.V. wrote the paper.
-
Conflict of interest statement: No conflicts declared.
-
This paper was submitted directly (Track II) to the PNAS office.
- Abbreviations:
- AD,
- Gal4 transactivation domain;
- BD,
- Gal4 DNA-binding domain;
- YPD,
- yeast extract/peptone/dextrose
-
Freely available online through the PNAS open access option.
- © 2006 by The National Academy of Sciences of the USA