Distal Recognition Sites in Substrates Are Required for ... · Distal Recognition Sites in Substrates Are Required for Efﬁcient Phosphorylation by the cAMP-Dependent Protein Kinase

Copyright � 2009 by the Genetics Society of AmericaDOI: 10.1534/genetics.109.102178

Distal Recognition Sites in Substrates Are Required for EfficientPhosphorylation by the cAMP-Dependent Protein Kinase

Stephen J. Deminoff,* Vidhya Ramachandran*,† and Paul K. Herman*,†,1

*Department of Molecular Genetics, †Program in Molecular, Cellular and Developmental Biology,The Ohio State University, Columbus, Ohio 43210

Manuscript received February 24, 2009Accepted for publication April 10, 2009

ABSTRACT

Protein kinases are important mediators of signal transduction in eukaryotic cells, and identifying thesubstrates of these enzymes is essential for a complete understanding of most signaling networks. In thisreport, novel substrate-binding variants of the cAMP-dependent protein kinase (PKA) were used to identifysubstrate domains required for efficient phosphorylation in vivo. Most wild-type protein kinases, includingPKA, interact only transiently with their substrates. The substrate domains identified were distal to the sitesof phosphorylation and were found to interact with a C-terminal region of PKA that was itself removed fromthe active site. Only a small set of PKA alterations resulted in a stable association with substrates, and theidentified residues were clustered together within the hydrophobic core of this enzyme. Interestingly, theseresidues stretched from the active site of the enzyme to the C-terminal substrate-binding domain identifiedhere. This spatial organization is conserved among the entire eukaryotic protein kinase family, andalteration of these residues in a second, unrelated protein kinase also resulted in a stable association withsubstrates. In all, this study identified distal sites in PKA substrates that are important for recognition by thisenzyme and suggests that the interaction of these domains with PKA might influence specific aspects ofsubstrate binding and/or release.

PROTEIN kinases are key mediators of signal trans-duction in all eukaryotic cells. Each protein kinase

modifies a distinct set of substrates, and the biologicalconsequences of activating any kinase are the result ofthe collective actions of these target proteins (Hunter

2000; Manning et al. 2002). The ability to identifysubstrates is therefore essential for a complete un-derstanding of most signaling pathways. Unfortunately,this identification process tends to be difficult, and fewphysiologically relevant targets are known for mostprotein kinases (Manning and Cantley 2002; Johnson

and Hunter 2005). This situation may be changing as anumber of innovative approaches to this problem havebeen developed in recent years (reviewed in Ptacek andSnyder 2006; Deminoff and Herman 2007; Ubersax

and Ferrell 2007).This article is focused on the cAMP-dependent pro-

tein kinase (PKA) from the budding yeast, Saccharomycescerevisiae. The PKA enzyme is found in all eukaryotes andis one of the most intensely studied members of thisprotein family (Taylor et al. 2005). PKA was the firstprotein kinase structure to be described, and its struc-ture has provided essential insights into the generalorganization and catalytic mechanism of these enzymes

(Knighton et al. 1991; Smith et al. 1999). Subsequentwork has illustrated the conserved nature of the proteinkinase core and the different ways that the activity ofthese enzymes can be regulated (Hunter 2000; Huse

and Kuriyan 2002; Kannan and Neuwald 2005). InS. cerevisiae, PKA activity is a key regulator of cell growthand the response to environmental stress (Toda et al.1985; Thevelein and De Winde 1999; Herman 2002;Schneper et al. 2004). We are interested in understand-ing the role of PKA in these processes and have iden-tified a number of substrates for this enzyme (Howard

et al. 2003; Chang et al. 2004; Budovskaya et al. 2005;Deminoff et al. 2006). One of the approaches used forthis identification took advantage of PKA variants thatexhibit a stable binding to substrate proteins (Deminoff

et al. 2006). This binding is novel as most wild-typeprotein kinases, including PKA, interact only transientlywith their substrates (Manning and Cantley 2002).Interestingly, one of these PKA variants was altered at aresidue that is conserved in all protein kinases, suggest-ing that it might be possible to generate substrate-binding versions of other enzymes in this family.

These variants of PKA were used here to explore thenature of the protein kinase–substrate interaction.These studies identified substrate domains distal tothe sites of phosphorylation that were required forefficient recognition by the wild-type PKA, both in vitroand in vivo. These substrate domains were found to

1Corresponding author: Department of Molecular Genetics, Ohio StateUniversity, 484 W. 12th Ave., Room 984, Columbus, OH 43210.E-mail: [email protected]

Genetics 182: 529–539 ( June 2009)

interact with a C-terminal region of PKA that is itselfremoved from the active site of the enzyme. A systematicmutagenesis of PKA identified additional residues that,when altered, resulted in a stable association withsubstrates. These latter residues are in close proximityin the three-dimensional structure and may link theactive site with this C-terminal substrate-binding do-main of PKA. Finally, we show that similar alterationswithin a second protein kinase, the mammalian double-stranded RNA-dependent protein kinase (PKR), alsoled to an increased affinity for substrates. In all, the datasuggest that the interactions described here may begenerally important for protein kinase function andmodels that explain potential roles for these substratedomains are discussed.

MATERIALS AND METHODS

Yeast strains, plasmids, and growth conditions: The yeaststrains PJ69-4A, PHY1025, and PHY1220 have been described( James et al. 1996; Herman and Rine 1997; Chang et al. 2004).The plasmid vectors used for the copper-inducible expressionof the epitope-tagged versions of Tpk1 and the substrateproteins were described previously (Deminoff et al. 2006). Forthese TPK1 expression vectors, Tpk1 is tagged at its N terminuswith three copies of the HA epitope. Within the context of thisHA-tagged construct, site-directed mutagenesis was used tointroduce the T241A alteration for this study. The TPK1-C70construct, pPHY3062, is based on these HA-tagged expressionplasmids except that a GST tag was also introduced betweenthe HA and TPK1 sequences. This same vector was used forcopper-inducible expression of a negative control HA-GSTprotein. Unless otherwise noted, the Cdc25 and Yak1 substrateproteins comprised residues 51–338 and 193–362, respec-tively. These proteins, as well as the full-length Bcy1 protein,were tagged with six copies of the myc epitope at their Nterminus (Deminoff et al. 2006). Protein A-tagged Cdc25(codons 51–338; pPHY2017) was expressed in PHY1025 cellsas a TDH3-driven construct with two copies of the Protein Asequences fused at the N terminus. A variant of this CDC25construct, called Cdc25-R4A, encodes a product in which botharginine residues in each of the two PKA sites were changed toalanines (Cdc25 residues 148, 149, 171, and 172; pPHY2362).These proteins were extracted from log-phase yeast cellcultures, precipitated with IgG–Sepharose, and used as sub-strates for PKA phosphorylation in vitro. The quadruple R4Avariant was also expressed as a Gal4 DNA-binding domainfusion protein for two-hybrid analysis (pPHY1962).

The PKR open-reading frame was PCR amplified from ahuman cDNA (Open Biosystems) and cloned into thepGADT7 two-hybrid vector (Clontech). Site-directed muta-geneses were performed to generate the K296R, H412N, andR526A alleles. The mutagenized alleles were cloned into theexpression vector pPHY3170 for the immunoprecipitationexperiments. In this vector, the resulting products are underthe control of the CUP1 promoter and have GST at their Nterminus. The S. cerevisiae eIF2a fragment (codons 2–175) wasPCR amplified from yeast genomic DNA and cloned into thetwo-hybrid vector pGBKT7 (Clontech). This fragment is pre-cisely the one used to obtain the co-crystal with PKR (Dar et al.2005). This eIF2a clone was subsequently inserted down-stream of the CUP1 promoter and a myc6 epitope in pRS423.

Two-hybrid assays: All of the two-hybrid binding assays wereperformed using PJ69-4A cells. The Tpk1 or substrate proteins

were expressed as N-terminal fusion proteins using pGADT7or pGBKT7, encoding the Gal4 activation or DNA-bindingdomains, respectively. In either case, PCR amplification wasdone using primers that allowed for insertion of thesesequences in-frame into the respective two-hybrid vector.These primers also introduced termination codons at the 39-ends of each of these fragments. Cells expressing the appro-priate constructs were collected from mid-log phase culturesand diluted with water. Assays were performed by spotting 5 mlof cells at an initial concentration of 5 OD600 cell equivalents/ml,and serial 10-fold dilutions thereafter, onto the appropriateselective media and incubating at 30� for 2 days. For the Tpk1assays, binding was indicated by cell growth on minimal medialacking adenine ( James et al. 1996). For the PKR-bindingassays, growth was assessed on media lacking histidine. In allcases, the kinase and substrate constructs were tested againstthe appropriate vector controls. For the Tpk1 variants, theindicated alterations were all introduced into the Tpk1K116A

background to avoid the toxicity associated with elevated levelsof PKA activity in S. cerevisiae (Deminoff et al. 2006). Theintroduction of this K116A alteration into either Tpk1KH orTpk1R324A had no effect on the observed substrate binding ofthese variants (data not shown). Western blots were performedin all cases to ensure that similar levels of the substrates andenzyme were present.

Immunoprecipitation assays: Yeast cells expressing candi-date binding partners were grown to mid-log phase at 30�.CuSO4 was added to a concentration of 100 mm, and theincubation was continued for 2 hr. Typically, 100 OD600 cellequivalents were collected by centrifugation, resuspended in0.5 ml of IP buffer [0.02 m Tris, pH 7.5, 1 mm EDTA, 10 mm

MgCl2, 5% glycerol, 0.3 m (NH4)2SO4, 1 mm PMSF], and lysedby agitation with glass beads. The cell extract was dilutedtwofold into this buffer, and NP-40 (to 0.5%) and theappropriate antibody were added and incubated with theextract for 2 hr at 4�. Protein A–Sepharose was then added,and the slurry was incubated for another 2 hr at 4� with mixing.Where noted, an anti-HA affinity resin (Sigma) was used forprecipitation of HA-tagged proteins. The beads were washedsix times with 1 ml of IP-wash buffer (0.024 m Tris, pH 7.5,1 mm EDTA, 10 mm MgCl2, 5% glycerol, 0.3 m ammoniumsulfate, 15 mm NaCl, 1 mm PMSF, 0.5% NP-40, 0.0375%sodium deoxycholate, and 0.075% Triton X-100), and theprecipitated proteins were detected by Western blotting asdescribed (Deminoff et al. 2006).

Tpk1 was also isolated in vitro by eluting it with cAMP froman immune complex containing Bcy1 (see below). The abilityof this Tpk1 protein to associate with Yak1 was assessed asfollows. A sample of the Tpk1 eluate was added to 50 ml ofextract prepared from yeast cells that expressed the appropri-ate copper-inducible Yak1 constructs. These mixtures wereincubated at 30� for 2 hr and then diluted into 1 ml of IP bufferplus 0.5% NP-40. Anti-myc antibody was added and the Yak1proteins were immunoprecipitated. Washes, elution, andanalysis by Western blotting were then performed as describedabove.

For the PKR immunoprecipitation assays, the cells express-ing the PKR constructs were resuspended in lysis buffer(25 mm Tris–HCl, pH 7.4, 140 mm NaCl, 0.1% Tween-20,1 mm PMSF) and lysed by agitation with glass beads. The anti-GST antibody was added to the clarified cell extracts, and thePKR immunoprecipitates were collected on Protein A–Sepharosebeads. The beads were washed three times with wash buffer(25 mm Tris–HCl, pH 7.4, 500 mm NaCl, 0.1% Tween-20, 1 mm

PMSF) and then incubated for 2 hr at 4� with cell extracts ofthe yeast strain expressing eIF2a. The beads were then washedthree times with wash buffer and resuspended in SDS–ureasample buffer, and the eluted proteins were separated by SDS–

530 S. J. Deminoff, V. Ramachandran and P. K. Herman

polyacrylamide electrophoresis. The eIF2a present was de-tected by Western blotting with an anti-myc antibody. For theBcy1-binding experiments, the immunoprecipitations wereperformed as described above for PKR.

PKA phosphorylation assays: The in vitro kinase assays wereperformed as described with [g-32P]ATP and either Tpk1 or0.02 units of the bovine PKA catalytic subunit (bPKA; Sigma)(Budovskaya et al. 2005; Deminoff et al. 2006). The Tpk1protein was prepared as follows. PHY1025 cells expressinginducible Bcy1 and Tpk1 constructs were grown to mid-logphase at 30�, and CuSO4 was added to a final concentration of100 mm. The incubation was continued for 2 hr, and 100 OD600

cell equivalents were collected by centrifugation and brokenby agitation with glass beads. The myc-tagged Bcy1 protein wasimmunoprecipitated from the resulting extracts using an anti-myc antibody and Protein A–Sepharose. The washed beadswere then incubated in 50 ml of phosphorylation buffer plus1 mm cAMP to elute the catalytic subunit (Deminoff et al.2006). A 5-ml aliquot of this eluate was used in a 100-ml kinasereaction. The activity of this 5-ml aliquot toward the peptidesubstrate, Kemptide, was similar to that observed with 0.02units of bPKA (Sigma). The Cdc25 or Yak1 protein substrateswere added as immunoprecipitates bound to Protein A–Sepharose, and the reaction was allowed to proceed at roomtemperature for 30 min. The substrate proteins were elutedfrom the beads and separated by SDS–polyacrylamide gelelectrophoresis. The degree of phosphorylation was assessedeither by autoradiography or by a phosphorimager analysis.The amount of substrate protein was assessed by Westernblotting. For the in vitro analysis of Yak1 phosphorylationusing the anti-PKA substrate antibody, the reaction buffercontained 3 mm ATP and 0.02 units of the bPKA.

The Tpk1 or Tpk1D264A phosphorylation of the Kemptidepeptide was assessed as described previously with the followingmodifications (Gibbs and Zoller 1991). Strains containingcopper-inducible Bcy1 and Tpk1 constructs were grown, andexpression was induced as described above. A small volume ofthe extract from this culture (0.27 OD600 equivalents) was usedin a 100-ml reaction that contained 200 mm Kemptide, 400 mm

ATP, 1 mm cAMP, 10 mm MgCl2, 4.5 mm DTT, and 10 mCi[g-32P]ATP (Gibbs and Zoller 1991; Deminoff et al. 2006).The reaction was allowed to proceed at room temperature for30 min. A 5-ml aliqot of the reaction was then removed andspotted onto Whatman P81 filters. The filters were washedfour times for 5 min each in 75 mm phosphoric acid, dried,and subjected to scintillation counting. A strain expressing thekinase-dead Tpk1K116A variant served as a negative control forthis experiment.

To monitor PKA phosphorylation in vivo, substrate proteinswere precipitated as described above, except that NP-40 wasomitted from the IP buffer and phosphatase inhibitors werepresent throughout. The level of PKA phosphorylation wasassessed by Western blotting with the anti-PKA phospho-substrate antibody (Cell Signaling) as described (Chang

et al. 2004; Deminoff et al. 2006). The signal observed withthis antibody was lost upon phosphatase treatment and wasrestored by the subsequent phosphorylation of the substrateby PKA in vitro (Chang et al. 2004; Searle et al. 2004;Deminoff et al. 2006). Recognition by this antibody was alsodependent upon the presence of the PKA phosphorylationsites in Cdc25 (data not shown).

Tpk1 phosphorylation at position T241 within the activationloop was assessed by Western blotting with an antibody specificfor the phosphorylated form of T197 in the mammalian PKAenzyme (Cell Signaling). This antibody cross-reacts specificallywith the phosphorylated form of T241 in Tpk1. For theseanalyses, the relevant Tpk1 proteins were immunoprecipitatedas described above and separated by SDS–polyacrylamide gel

electrophoresis, and the phosphorylation level within theactivation loop was assessed by Western blotting with thisphosphospecific antibody.

RESULTS

Identification of substrate sequences important forrecognition by PKA: Our previous work identified twovariants of the S. cerevisiae PKA isoform Tpk1 that exhibita greater affinity for substrates than the wild-type pro-tein (Deminoff et al. 2006). Tpk1R324A has an alanineresidue substituting for an invariant arginine at position324, and Tpk1KH has alanine residues replacing Lys-336and His-338. These positions correspond to R280, K292,and H294 in mammalian PKA enzymes, respectively. Atwo-hybrid assay was used here to map sequences in twosubstrates, Yak1 and Cdc25, that were required for theinteraction with Tpk1KH. Yak1 is a protein kinase impor-tant for the control of S. cerevisiae growth and Cdc25 isthe guanine nucleotide exchange factor for the Rasproteins (Broek et al. 1987; Garrett et al. 1991). Thisanalysis identified 51 and 27 amino acid segments ofYak1 and Cdc25, respectively, that were necessary andsufficient for the interaction with Tpk1KH (Figure 1, Aand B; Figure 2, A and B). Epitope-tagged versions ofthese substrate domains were also detected in immuno-precipitates with the Tpk1KH variant (Figure 1C; Figure2C). These Tpk1-recognition sites in Yak1 and Cdc25will be referred to as RS-Y and RS-C, respectively, in thisreport. These RS fragments did not contain any PKAconsensus sites and neither was phosphorylated byPKA in vitro (Figure 1A and Figure 2A). In addition, anonphosphorylatable form of Cdc25 interacted withTpk1KH as well as with the wild-type substrate (Figure2D). Therefore, the presence of a functional PKA sitewas not needed for the observed binding to the Tpk1variants described here.

To test whether the identified RS domains wereimportant for recognition by the wild-type Tpk1, weconstructed versions of Yak1 and Cdc25 that lackedthese sequence elements. We found that the in vitrophosphorylation signal for both substrates decreased bythree- to fourfold when the respective RS domain wasdeleted (Figure 1, D and E; Figure 2E). A similardecrease in phosphorylation was observed for reactionscontaining either Tpk1 or bPKA, indicating that theeffects were not specific for the S. cerevisiae enzyme.Removal of the RS domains also resulted in a significantdecrease in the in vivo phosphorylation level of bothYak1 and Cdc25 (Figure 1F; Figure 2F). For this latterassay, the substrate proteins were precipitated fromyeast cell extracts, and the level of PKA phosphorylationwas assessed by Western blotting with an anti-PKAsubstrate antibody. This antibody has been shown tospecifically recognize the phosphorylated forms of anumber of PKA substrates (Chang et al. 2004; Searle

et al. 2004; Deminoff et al. 2006). We found that the

Substrate Identification by PKA 531

recognition of Cdc25 and Yak1 by this antibody was lostupon phosphatase treatment and was restored by thesubsequent incubation of these dephosphorylated sub-strates with PKA and ATP (data not shown). In all, thedata suggested that PKA substrates might generallycontain sequences distal from the sites of phosphoryla-tion that are needed for efficient recognition by thewild-type enzyme.

A C-terminal fragment of Tpk1 was sufficient forbinding to the substrate RS domains: To identify kinasedomains that may mediate the above interactions withsubstrates, we tested whether various fragments of Tpk1would associate with Cdc25 and Yak1. We subdividedthis enzyme into non-overlapping 70-amino-acid frag-ments that covered the majority of the protein andassessed substrate binding with a two-hybrid assay. Thesestudies identified a C-terminal fragment of Tpk1 thatwas sufficient for interaction with both Cdc25 and Yak1(Figure 3A). This Tpk1-C70 fragment included residuesfrom 285 to 354. Substrate binding was specific to thisregion of Tpk1 because other domains of this enzymedid not interact with either Cdc25 or Yak1 (Figure 3A;data not shown). Moreover, this Tpk1-C70 fragment didnot exhibit an interaction with any protein examinedthat was not a substrate for PKA (Figure 3B). Finally, adeletion analysis identified a minimal 33-residue frag-ment from position 298 to 330, which was capable ofinteracting with Yak1 (Figure 3, C and D).

These substrate interactions with Tpk1-C70 werefound to require the presence of the RS domainsidentified here. For example, the loss of the RS-Cdomain resulted in a significant decrease in the amountof Cdc25 associated with Tpk1-C70 (Figure 4A). More-over, we found that the Tpk1-C70 fragment specificallyinteracted with the isolated RS-C and RS-Y domains inboth two-hybrid and co-immunoprecipitation assays(Figure 4, B–D). In all, the data suggested that the RSdomains within Cdc25 and Yak1 interact, directly orindirectly, with particular sequences near the C termi-nus of Tpk1.

Alteration of residues within the hydrophobic coreof PKA resulted in an increased binding to substrates:A systematic mutagenesis of TPK1 was carried out toidentify additional residues that, when altered, wouldresult in a stable interaction with Cdc25. For thisanalysis, the most conserved residues in each of the 11protein kinase subdomains in the catalytic core of Tpk1were altered (Hanks and Hunter 1995). We found thatonly particular changes in the C-terminal half of theenzyme resulted in an increased binding to Cdc25(Figure 5, A–C). The newly identified Tpk1 residueswere E252, W266, D264, and Y208. (These positions

Figure 1.—Yak1 sequences distant from the sites of phos-phorylation were required for efficient recognition by PKA.(A) Mapping the Tpk1-recognition site (RS-Y) in Yak1. Therelative binding affinity to Tpk1KH in a two-hybrid assay is in-dicated, along with the relative phosphorylation signal ob-served in an in vitro reaction with Tpk1. The positions ofthe four consensus PKA phosphorylation sites are indicatedwith asterisks, and the RS-Y domain by the solid area. (B)The RS-Y domain of Yak1 was sufficient for the interactionwith the Tpk1KH variant. A two-hybrid assay with Tpk1KH

and the indicated fragments of Yak1 are shown. The Yak1,Yak1-DRS, and RS-Y constructs included residues 193–362,193–311, and 312–362, respectively. (C) The RS-Y domain in-teracted with Tpk1KH in an immunoprecipitation assay. TheTpk1 proteins were precipitated with an anti-HA affinity resin,and the amount of associated RS-Y fragment was assessed byWestern blotting. (D) Deletion of the RS-Y domain resulted indecreased phosphorylation by Tpk1 in vitro. The indicatedYak1 fragments were precipitated from cell extracts and incu-bated with Tpk1 and [g-32P]ATP, as described in materials

and methods. The relative level of phosphorylation followingphosphorimager quantification of the data is shown. (E) Theabsence of the RS-Y domain resulted in decreased phosphor-ylation of Yak1 by bPKA in vitro. The Yak1 constructs were pre-cipitated from yeast cell extracts, treated with phosphatase,and incubated with 0.02 units of bPKA and 3 mm ATP. Thelevel of PKA phosphorylation was then assessed by Westernblotting with the anti-PKA substrate antibody (a-Subs). (F)The RS-Y domain was required for efficient Yak1 phosphory-lation by PKA in vivo. The indicated Yak1 constructs were pre-cipitated from yeast cell extracts, and the level of PKAphosphorylation was assessed by Western blotting with theanti-PKA substrate antibody (see materials and methods).

The band marked with the asterisk is the heavy chain ofthe antibody used for the immunoprecipitation, and the dot-ted oval indicates the position of the Yak1-DRS fragment.


correspond to E208, W222, D220, and H164 in mam-malian PKA enzymes, respectively.) These residues areall very close to each other in three-dimensional spaceand, together with R324, span the interior of the enzymefrom the active site (via Y208 in the catalytic loop) to thepotential RS domain-binding site at the base of the C-terminal lobe (the pink region in Figure 5D). Two ofthese residues, E252 and D264, correspond to positionsthat are invariant in protein kinases and the others,W266 and Y208, to positions that are conserved to alesser degree. The glutamate corresponding to position252 forms a buried ion pair with the invariant arginine(R324 in Tpk1) in all available protein kinase structures(Kannan et al. 2007a). These residues, in turn, pack upagainst W266 in the F-helix, and D264 in the F-helixinteracts with residues in the conserved H/Y208-R-Dmotif present in the catalytic loop (see Figure 5D)(Smith et al. 1999). Interestingly, a recent comparativeanalysis of protein kinase structures has indicated thatthis link between the invariant arginine (R324 in Tpk1)and the active site is conserved throughout this enzymefamily (Kannan et al. 2007b). Moreover, the authors ofthis study suggested that the residues identified heremight constitute an internal network that is generallyimportant for coupling the substrate- and ATP-bindingportions of the kinase molecule.

The absence of the regulatory subunit was notsufficient for substrate binding by the wild-type Tpk1:PKA enzymes are inhibited by a regulatory (R) subunitthat binds to the catalytic subunit within the active siteand at several distal positions (Kim et al. 2007; Wu et al.2007). One of these latter sites is near, or coincidentwith, the potential C-terminal substrate-binding domainidentified here (site 4 in Kim et al. 2007; Wu et al. 2007).In addition, we found that several of the Tpk1 variantsthat exhibited an elevated association with substrateshad a diminished capacity to bind to the S. cerevisiae Rsubunit, Bcy1 (Figure 6A). Therefore, it was formallypossible that a competition existed between substratesand the R subunit for binding to this C-terminal domainof Tpk1. In the simplest model, the absence of Bcy1would result in increased substrate binding to the wild-type Tpk1. To test this possibility, we assessed theinteraction between Tpk1 and Yak1 in a strain lackingBcy1 and found that there was no increase in substratebinding (Figure 6B). We also tested this model with aTpk1 variant, Tpk1T241A, that fails to interact normallywith Bcy1 (Levin et al. 1988; Levin and Zoller 1990).Thr-241 is in the activation loop of Tpk1, and phos-

Figure 2.—Identification of the RS-C domain in Cdc25.(A) Mapping the Tpk1-recognition site in Cdc25. The relativebinding affinity to Tpk1KH in a two-hybrid assay is indicatedalong with the relative phosphorylation signal observed inan in vitro reaction with Tpk1. The positions of the two pri-mary PKA phosphorylation sites in Cdc25 are indicated withasterisks, and the RS-C domain by the solid areas. (B) The RS-C domain was sufficient for the observed interaction withTpk1KH in a two-hybrid assay. The RS-C construct containedCdc25 residues 78–104. (C) A Cdc25 fragment containingthe RS-C domain (residues 78–135) was detected in immuno-precipitates with the Tpk1KH variant. The Tpk1 proteins wereprecipitated with an anti-HA affinity resin, and the amount ofassociated RS-C fragment was assessed by Western blotting.(D) The PKA phosphorylation sites were not required forCdc25 binding to Tpk1KH. The results of an in vitro kinase as-say with bPKA and [g-32P]ATP are shown for the wild-typeCdc25 and a variant that is altered at both of the PKA consen-sus sites (R4A). The typical PKA site has two arginine residueson the N-terminal side of the site of phosphorylation (R-R-x-S/T). The R4A variant has alanine residues replacing both ofthe arginines in each site (R148 R149 and R171 R172). The rel-ative binding to Tpk1KH observed for both Cdc25 fragments ina two-hybrid assay is indicated. (E) Deletion of the RS-C do-main resulted in decreased phosphorylation by Tpk1in vitro. Cdc25 fragments with and without the RS-C(Cdc25-DRS, residues 136–338) were precipitated from yeastcell extracts and incubated with Tpk1 and [g-32P]ATP, as de-scribed in materials and methods. (F) The RS-C domainwas required for the efficient in vivo phosphorylation ofCdc25 by PKA. The indicated Cdc25 constructs were precip-

itated from yeast cell extracts, and the level of PKA phosphor-ylation was assessed by Western blotting with the anti-PKAsubstrate antibody. The band marked with the asterisk isthe heavy chain of the antibody used for the immunoprecip-itation, and the dotted oval indicates the position of theCdc25-DRS fragment on the immunoblot.


phorylation at this position is needed for full enzymaticactivity (Levin and Zoller 1990). Consistent with theabove results with the bcy1D strain, we found that thisTpk1T241A variant did not exhibit any significant bindingto Yak1 (Figure 6C). In all, these data indicated that afailure to bind to the R subunit does not a priori result inan increased association with substrates. It is importantto note that these data are consistent with our previousimmunoprecipitation studies that were performed withstrains expressing relatively high levels of Tpk1 (Deminoff

et al. 2006). The endogenous levels of Bcy1 were muchlower in these cells, and most of the Tpk1 was pre-sumably free and not complexed with the R subunit.

The Yak1 RS domain was required for interactionswith the wild-type Tpk1: Although the binding isweaker than that observed with specific variants, thereis a measurable interaction between substrates and thewild-type Tpk1 (Figure 7A) (Deminoff et al. 2006). Inaddition, one of the substrate-binding variants de-scribed above, Tpk1D264A, has been shown previously toexhibit near wild-type kinase activity in vitro (Gibbs and

Zoller 1991). Here, we tested whether the RS domainswere also required for substrate interactions with thesetwo enzymatically active forms of Tpk1. We first con-firmed that Tpk1D264A possessed near-normal catalyticactivity and further showed that this variant exhibited anormal interaction with Bcy1 and wild-type levels ofphosphorylation within the activation loop (Figure 7, Band C; data not shown). Subsequent work showed thatthe interactions between Yak1 and both Tpk1 andTpk1D264A were dependent upon the presence of theRS-Y domain (Figure 7, A and D). Therefore, the RSdomains in substrates were required for binding to thewild-type Tpk1, as well as to the inactive variantsdiscussed above.

Similar alterations in a second protein kinase, PKR,resulted in an increased binding to substrates: Most ofthe alterations that resulted in the increased substratebinding observed here occurred at positions that arevery highly conserved among protein kinases. Forexample, an arginine is found in all protein kinases ata position that corresponds to R324 in Tpk1. This

Figure 3.—A C-terminal fragment of Tpk1 in-teracted specifically with substrates. (A) Cdc25and Yak1 interacted specifically with the C70 frag-ment of Tpk1. Substrate binding to the indicatedfragments of Tpk1 was assessed with a two-hybridassay. The constructs contain the indicated resi-dues of Tpk1; Tpk1-C70 includes residues 285–354. ‘‘2x’’ indicates that twice as many cells wereused in the initial inoculum. (B) The Tpk1-C70fragment did not exhibit an interaction with pro-teins that were not substrates of PKA. The relativeinteractions between Tpk1-C70 and the indicatedproteins were assessed with a two-hybrid assay.Bur1, Ctk1, and Kin28 were not substrates forPKA in an in vitro assay (data not shown). Yak1was used as a positive control, and the empty vec-tor (�) was the negative control. (C) Mappingthe Tpk1 domain sufficient for the interactionwith Yak1. The relative binding to Yak1 withthe indicated fragments of Tpk1 was assessedin a two-hybrid assay. The endpoints of each frag-ment along with the positions of residues R324,K336, and H338 are provided. The relative posi-tions of the G, H, I, and J a-helices are indicatedby the lavender shading. Note that it is not clearwhether the helical nature of these regions is re-tained within the isolated fragments. (D) A sche-matic of Tpk1 is shown. The positions of the F, G,and H a-helices are indicated. The image wasgenerated with the ProteinWorkshop programusing the PDB file, 1FOT, for Tpk1 (Mashhoon

et al. 2001; Moreland et al. 2005).


conservation suggests that this residue performs animportant function for protein kinases and that alter-ations at this position might have a similar effect inother enzymes of this family. Therefore, we testedwhether altering this position in an unrelated proteinkinase would also influence substrate interactions. Forthese experiments, we worked with the mammalianRNA-dependent protein kinase, PKR, an enzyme thatplays an important role in translational control inresponse to viral infection (Sadler and Williams

2007). We chose PKR because this enzyme was pre-viously crystallized in a complex with its primary sub-strate, the translation initiation factor eIF2a (Dar et al.2005). These previous studies showed that eIF2a con-tacted a C-terminal region of PKR that is near the Tpk1domain described here.

We found that an alanine substitution for the arginineat position R526 in PKR resulted in a stable interactionwith eIF2a (Figure 8, A and B). In contrast, alterationof a critical lysine in the PKR active site, K296 (K116 in

Tpk1), did not lead to increased binding to eIF2a

(Figure 8). These results therefore mirrored thoseobtained with Tpk1 and showed that the substrate-binding phenotype described here was not unique toPKA enzymes. The final point concerns the complexformed between PKR and eIF2a in the previous crystal-lization studies. To obtain this complex, the authorsused a PKR enzyme that had several alterations, in-cluding H412N (Dar et al. 2005). This residue corre-sponds to Y208 in Tpk1 and is one of the residues thatresulted in an increased affinity for substrates whenaltered (see Figure 5). We therefore examined thePKRH412N variant and found that it also interacted stablywith eIF2a. Thus, two different protein kinases havebeen altered in similar ways to give rise to an elevatedaffinity for substrates.

DISCUSSION

Protein kinases are able to recognize their appropri-ate targets in a complex milieu of cellular protein. Thisprocess must be carried out with high fidelity to ensureproper signal transduction in eukaryotic cells (Hunter

2000). In this study, we attempted to obtain insight intothis recognition by examining PKA variants that exhibita stable association with substrates. This binding pro-vided a facile assay that allowed us to identify domains inboth enzyme and substrates that were important forPKA phosphorylation. The substrate domains identifiedwere physically removed from the sites of phosphoryla-tion and were required for efficient recognition by PKAboth in vivo and in vitro. To the best of our knowledge,these studies are the first to show that such distalsequence elements in substrates are required for phos-phorylation by PKA. These observations may helpexplain why only a fraction of proteins that contain aPKA consensus site are phosphorylated by this enzymein vivo (Budovskaya et al. 2005). The proteins that failto be recognized may lack these secondary recognitionsites, a possibility that will need to be tested by furtherexperimentation.

Our data suggest that these RS domains interact,perhaps directly, with particular sequences in the C-terminal lobe of the catalytic subunit of PKA. Thisdomain is removed from the active site and is similar tothe region in PKR that has been shown to mediate theassociation with its primary substrate, eIF2a (Dar et al.2005). The relevant question here is how do these RSdomains influence substrate phosphorylation by PKA?One possibility is that these domains might act astraditional docking sites, such as those described forsubstrates of particular MAP kinases (Bardwell et al.1996; Kallunki et al. 1996). These latter sites aregenerally thought to function by tethering substratesnear the appropriate enzyme (Bardwell 2006; Remenyi

et al. 2006; Shi et al. 2006). However, it should be pointedout that the RS domains described here appear to bind

Figure 4.—The Tpk1-C70 fragment interacted with the iso-lated RS domains of substrates. (A) The interaction betweenthe Tpk1-C70 fragment and Cdc25 required the presence ofthe RS-C domain. The GST-C70 fusion protein was immuno-precipitated from yeast cell extracts, and the relative amountof associated Cdc25 was then assessed by Western blottingwith an anti-myc antibody. (B) The Tpk1-C70 fragment wassufficient for the observed interaction with Cdc25 and Yak1.Two-hybrid assay results for Tpk1-C70 with the indicated sub-strates are shown. The bottom three rows show the two-hybridinteractions for Tpk1-C70 with the indicated fragments ofCdc25. (C and D) The RS-C and RS-Y domains interacted withthe Tpk1-C70 fragment in an immunoprecipitation assay. Therelative level of the respective RS domain associated with theprecipitated Tpk1-C70 construct was assessed by Western blot-ting with an anti-myc antibody.


rather weakly to the wild-type PKA. Nonetheless, thepresence of one or more of these weaker interactionmotifs in substrates might be enough to facilitate theirrecognition by PKA in vivo.

An alternative possibility may be suggested by theidentity of the Tpk1 residues that resulted in an in-creased binding to substrates when altered. Theseresidues spanned the interior of the enzyme and couldlink the active site to the potential RS-binding domain atthe C terminus. Therefore, it is possible that interac-tions at one of these sites could influence eventsoccurring at the other. For example, substrate bindingat the C terminus might stimulate catalysis within theactive site, perhaps by stabilizing the formation of apreviously identified structure known as the regulatory,

or R, spine (Kornev et al. 2006). This structure appearsto be needed for protein kinase activation and includesthe Y208 position identified here (Azam et al. 2008;Kornev et al. 2008). Evidence for such long-rangecommunication between the active site and C-terminalregions has been reported for a number of proteinkinases (Hoofnagle et al. 2001; Hamuro et al. 2002; Lee

et al. 2004; Anand et al. 2007). In addition, an allostericrole has been proposed for docking sites found insubstrates for other protein kinases (Biondi et al. 2000;Frodin et al. 2000; Bhattacharyya et al. 2006; Shi et al.2006; Zhou et al. 2006; Goldsmith et al. 2007). De-termining which of these possibilities might be relevantto PKA will clearly require additional work. In particular,it will be important to interrogate the dynamics of the

Figure 5.—Identification of Tpk1 residues that, when altered, resulted in a stable binding to substrates. (A) The substratebinding associated with various Tpk1 alterations. The top schematic shows the 11 subdomains of the catalytic core of the PKAenzyme and the major secondary structural features of these regions (adapted with permission from Kannan et al. 2007b).The relative binding to Cdc25 for representative alterations from the different regions of PKA is shown. Each of the indicatedresidues was changed to an alanine, and the binding was assessed with a two-hybrid assay. (B and C) Representative two-hybridbinding data for a subset of the Tpk1 alterations. In B, equal numbers of cells were spotted in the left-most column, and 10-foldserial dilutions were spotted thereafter. The growth on the minimal medium lacking adenine (‘‘Selection’’) is an indicator of therelative binding between the indicated Tpk1 variants and Cdc25. In C, the two-hybrid assay was performed with the indicated Tpk1variants and either Cdc25 or an empty vector control. (D) The residues identified by the elevated substrate-binding phenotypewere clustered within the hydrophobic core of Tpk1. A three-dimensional surface view of the Tpk1 enzyme is shown on the left.The five residues that gave rise to an increased affinity for Cdc25 when altered are shown as space-filling spheres. An expandedview of these latter residues is shown in the inset. The pink color indicates the surface-exposed regions of the minimal Tpk1fragment that was capable of binding to Cdc25 (residues 298–330; see Figure 3C). The images were generated with PyMOL usingthe PDB file for Tpk1 (1FOT) (Mashhoon et al. 2001).


enzyme and to determine how, for example, the in-teraction with the RS domains might influence otherregions of the enzyme. These studies might thereforeprovide insight into why the PKA alterations describedhere influence interactions occurring at the surface ofthe enzyme.

Several observations suggest that the results reportedhere for PKA are likely to be relevant for other enzymesin this family. First, the Tpk1 residues that gave rise tothe increased substrate binding when altered are highlyconserved, if not invariant, in eukaryotic protein kinases(Hanks and Hunter 1995). Moreover, these residuesmay provide a means to generally link the active site tomore C-terminal domains of the kinase molecule

(Kannan et al. 2007b; Kornev et al. 2008). Second, wefound that the alteration of these same residues in anunrelated protein kinase, the mammalian PKR, also ledto a stable association with substrates. Therefore, it ispossible that changes at these positions could provide ageneral approach to protein kinase substrate identifica-tion. Finally, a recent study identified the positionscorresponding to E252, W266, and R324 in Tpk1 asresidues in protein kinases that are hotspots for single-nucleotide polymorphisms associated with human dis-eases (Torkamani et al. 2008). In all, we feel that acontinued analysis of these substrate-binding variants ofTpk1 is likely to provide insights into the mechanismsgoverning substrate recognition in this enzyme family.

Figure 6.—A failure to interactwith the regulatory subunit Bcy1did not necessarily result in in-creased substrate binding. (A)Several of the substrate-bindingvariants of Tpk1 exhibited a de-creased interaction with the Bcy1regulatory subunit. The indicatedTpk1 proteins were immunopre-cipitated with an anti-HA anti-

body and collected on protein A–Sepharose beads. These beads were washed and then added to equivalent amounts of a cellextract that contained a GST-Bcy1 fusion protein. The beads were subsequently washed, and the amount of associated Bcy1was assessed by Western blotting with an anti-GST antibody. Vec, vector control. (B) The wild-type Tpk1 did not exhibit an in-creased interaction with Yak1 in cells lacking Bcy1. The wild-type Tpk1 (WT) and Tpk1KH (KH) proteins were precipitated fromeither wild-type (BCY1) or bcy1D cell extracts, and the amount of associated Yak1 was assessed by Western blotting. (C) TheTpk1T241A variant did not exhibit significant binding to the Yak1 substrate. The indicated Tpk1 enzymes were precipitated fromyeast cell extracts, and the amount of associated Yak1 was assessed by Western blotting.

Figure 7.—Substrate interactions with thewild-type Tpk1 were dependent upon the pres-ence of the RS domain. (A) The RS-Y domainwas required for Yak1 binding to the wild-typeTpk1 and Tpk1D264A. The indicated Yak1 con-structs were precipitated from yeast cell extracts,and the amount of associated Tpk1 was assessedby Western blotting. The band marked by an as-terisk is the heavy chain of the antibody usedin the immunoprecipitation. A longer exposureof the wild-type Tpk1 lanes is shown below theInput controls. (B) The Tpk1D264A variant ex-hibited a normal interaction with Bcy1. TheBcy1 protein was immunoprecipitated from yeastcell extracts, and the amount of associated Tpk1was detected by Western blotting. The right lanesfor both the wild-type and Tpk1D264A strains showthe immunoprecipitation results for extractslacking Bcy1. (C) Tpk1D264A was phosphorylatednormally at position T241 within the activationloop. The indicated Tpk1 proteins were precipi-tated, and the extent of phosphorylation in theactivation loop was assessed by Western blottingwith an antibody specific for the phosphorylatedform of T241 as described in materials and

methods. (D) The RS domain of Yak1 was re-quired for interaction with the wild-type Tpk1. A myc-tagged Bcy1 was immunoprecipitated from a cell extract, and the associatedTpk1 was released by the addition of 1 mm cAMP. Equal aliquots of this partially purified Tpk1 were then added to extracts con-taining the indicated Yak1 proteins. The Yak1 was subsequently precipitated, and the amount of associated Tpk1 was assessed byWestern blotting.


We thank Susan Taylor for helpful discussions and for sharingresults prior to publication, Susie Howard for assisting with the analysisof Cdc25 phosphorylation, and members of the Herman lab forcomments on the manuscript. This work was supported by NationalInstitutes of Health grant R01 GM65227 (to P.K.H.). S.J.D. wassupported in part by a T32 CA106196 fellowship in Cancer Geneticsfrom the National Institutes of Health.

LITERATURE CITED

Anand, G. S., M. Hotchko, S. H. Brown, L. F. Ten Eyck, E. A.Komives et al., 2007 R-subunit isoform specificity in protein ki-nase A: distinct features of protein interfaces in PKA types I andII by amide H/2H exchange mass spectrometry. J. Mol. Biol. 374:487–499.

Azam, M., M. A. Seeliger, N. S. Gray, J. Kuriyan and G. Q. Daley,2008 Activation of tyrosine kinases by mutation of the gate-keeper threonine. Nat. Struct. Mol. Biol. 15: 1109–1118.

Bardwell, L., 2006 Mechanisms of MAPK signalling specificity.Biochem. Soc. Trans. 34: 837–841.

Bardwell, L., J. G. Cook, E. C. Chang, B. R. Cairns and J. Thorner,1996 Signaling in the yeast pheromone response pathway: spe-cific and high-affinity interaction of the mitogen-activated pro-tein (MAP) kinases Kss1 and Fus3 with the upstream MAPkinase kinase Ste7. Mol. Cell. Biol. 16: 3637–3650.

Bhattacharyya, R. P., A. Remenyi, M. C. Good, C. J. Bashor, A. M.Falick et al., 2006 The Ste5 scaffold allosterically modulates sig-naling output of the yeast mating pathway. Science 311: 822–826.

Biondi, R. M., P. C. Cheung, A. Casamayor, M. Deak, R. A. Currie

et al., 2000 Identification of a pocket in the PDK1 kinase do-main that interacts with PIF and the C-terminal residues ofPKA. EMBO J. 19: 979–988.

Broek, D., T. Toda, T. Michaeli, L. Levin, C. Birchmeier et al.,1987 The S. cerevisiae CDC25 gene product regulates theRAS/adenylate cyclase pathway. Cell 48: 789–799.

Budovskaya, Y. V., J. S. Stephan, S. J. Deminoff and P. K. Herman,2005 An evolutionary proteomics approach identifies sub-strates of the cAMP-dependent protein kinase. Proc. Natl. Acad.Sci. USA 102: 13933–13938.

Chang, Y. W., S. C. Howard and P. K. Herman, 2004 The Ras/PKAsignaling pathway directly targets the Srb9 protein, a componentof the general RNA polymerase II transcription apparatus. Mol.Cell 15: 107–116.

Dar, A. C., T. E. Dever and F. Sicheri, 2005 Higher-order substraterecognition of eIF2alpha by the RNA-dependent protein kinasePKR. Cell 122: 887–900.

Deminoff, S. J., and P. K. Herman, 2007 Identifying atg1 substrates:four means to an end. Autophagy 3: 667–673.

Deminoff, S. J., S. C. Howard, A. Hester, S. Warner and P. K.Herman, 2006 Using substrate-binding variants of the cAMP-dependent protein kinase to identify novel targets and a kinasedomain important for substrate interactions in Saccharomyces cer-evisiae. Genetics 173: 1909–1917.

Frodin, M., C. J. Jensen, K. Merienne and S. Gammeltoft, 2000 Aphosphoserine-regulated docking site in the protein kinase RSK2that recruits and activates PDK1. EMBO J 19: 2924–2934.

Garrett, S., M. M. Menold and J. R. Broach, 1991 The Saccharomy-ces cerevisiae YAK1 gene encodes a protein kinase that is inducedby arrest early in the cell cycle. Mol. Cell. Biol. 11: 4045–4052.

Gibbs, C. S., and M. J. Zoller, 1991 Rational scanning mutagenesisof a protein kinase identifies functional regions involved in catal-ysis and substrate interactions. J. Biol. Chem. 266: 8923–8931.

Goldsmith, E. J., R. Akella, X. Min, T. Zhou and J. M. Humphreys,2007 Substrate and docking interactions in serine/threonineprotein kinases. Chem. Rev. 107: 5065–5081.

Hamuro, Y., L. Wong, J. Shaffer, J. S. Kim, D. D. Stranz et al.,2002 Phosphorylation driven motions in the COOH-terminalSrc kinase, CSK, revealed through enhanced hydrogen-deuteriumexchange and mass spectrometry (DXMS). J. Mol. Biol. 323: 871–881.

Hanks, S. K., and T. Hunter, 1995 Protein kinases 6. The eukary-otic protein kinase superfamily: kinase (catalytic) domain struc-ture and classification. FASEB J. 9: 576–596.

Herman, P. K., 2002 Stationary phase in yeast. Curr. Opin. Micro-biol. 5: 602–607.

Herman, P. K., and J. Rine, 1997 Yeast spore germination: a require-ment for Ras protein activity during re-entry into the cell cycle.EMBO J. 16: 6171–6181.

Hoofnagle, A. N., K. A. Resing, E. J. Goldsmith and N. G. Ahn,2001 Changes in protein conformational mobility upon activa-tion of extracellular regulated protein kinase-2 as detected by hy-drogen exchange. Proc. Natl. Acad. Sci. USA 98: 956–961.

Howard, S. C., A. Hester and P. K. Herman, 2003 The Ras/PKAsignaling pathway may control RNA polymerase II elongationvia the Spt4p/Spt5p complex in Saccharomyces cerevisiae. Genetics165: 1059–1070.

Hunter, T., 2000 Signaling–2000 and beyond. Cell 100: 113–127.Huse, M., and J. Kuriyan, 2002 The conformational plasticity of

protein kinases. Cell 109: 275–282.James, P., J. Halladay and E. A. Craig, 1996 Genomic libraries and

a host strain designed for highly efficient two-hybrid selection inyeast. Genetics 144: 1425–1436.

Johnson, S. A., and T. Hunter, 2005 Kinomics: methods for deci-phering the kinome. Nat. Methods 2: 17–25.

Kallunki, T., T. Deng, M. Hibi and M. Karin, 1996 c-Jun can re-cruit JNK to phosphorylate dimerization partners via specificdocking interactions. Cell 87: 929–939.

Kannan, N., and A. F. Neuwald, 2005 Did protein kinase regulatorymechanisms evolve through elaboration of a simple structuralcomponent? J. Mol. Biol. 351: 956–972.

Kannan, N., N. Haste, S. S. Taylor and A. F. Neuwald, 2007a Thehallmark of AGC kinase functional divergence is its C-terminaltail, a cis-acting regulatory module. Proc. Natl. Acad. Sci. USA104: 1272–1277.

Kannan, N., S. S. Taylor, Y. Zhai, J. C. Venter and G. Manning,2007b Structural and functional diversity of the microbial ki-nome. PLoS Biol. 5: e17.

Kim, C., C. Y. Cheng, S. A. Saldanha and S. S. Taylor, 2007 PKA-Iholoenzyme structure reveals a mechanism for cAMP-dependentactivation. Cell 130: 1032–1043.

Knighton, D. R., J. H. Zheng, L. F. Ten Eyck, V. A. Ashford, N. H.Xuong et al., 1991 Crystal structure of the catalytic subunit ofcyclic adenosine monophosphate-dependent protein kinase. Sci-ence 253: 407–414.

Kornev, A. P., N. M. Haste, S. S. Taylor and L. F. Eyck,2006 Surface comparison of active and inactive protein kinasesidentifies a conserved activation mechanism. Proc. Natl. Acad.Sci. USA 103: 17783–17788.

Figure 8.—Similar alterations in the human PKR enzymealso resulted in a stable interaction with the substrate eIF2a.(A) A two-hybrid assay showing the relative binding of the in-dicated PKR variants to eIF2a. (B) The PKRH412N and PKRR526A

variants interacted with eIF2a in an immunoprecipitation as-say. Anti-GST antibody was added to cell extracts containingthe indicated GST–PKR constructs or GST alone. The im-mune complexes were collected on Protein A–Sepharose,and the washed beads were then incubated with equalamounts of a cell extract that contained eIF2a. The relativeamount of eIF2a associated with the PKR constructs was as-sessed by Western blotting.


Kornev, A. P., S. S. Taylor and L. F. Ten Eyck, 2008 A helix scaf-fold for the assembly of active protein kinases. Proc. Natl. Acad.Sci. USA 105: 14377–14382.

Lee, T., A. N. Hoofnagle, Y. Kabuyama, J. Stroud, X. Min et al.,2004 Docking motif interactions in MAP kinases revealed by hy-drogen exchange mass spectrometry. Mol. Cell 14: 43–55.

Levin, L. R., and M. J. Zoller, 1990 Association of catalytic and reg-ulatory subunits of cyclic AMP-dependent protein kinase re-quires a negatively charged side group at a conservedthreonine. Mol. Cell. Biol. 10: 1066–1075.

Levin, L. R., J. Kuret, K. E. Johnson, S. Powers, S. Cameron et al.,1988 A mutation in the catalytic subunit of cAMP-dependentprotein kinase that disrupts regulation. Science 240: 68–70.

Manning, B. D., and L. C. Cantley, 2002 Hitting the target: emerg-ing technologies in the search for kinase substrates. Sci. STKE162: 1–4.

Manning, G., G. D. Plowman, T. Hunter and S. Sudarsanam,2002 Evolution of protein kinase signaling from yeast toman. Trends Biochem. Sci. 27: 514–520.

Mashhoon, N., G. Carmel, J. W. Pflugrath and J. Kuret,2001 Structure of the unliganded cAMP-dependent protein ki-nase catalytic subunit from Saccharomyces cerevisiae. Arch. Bio-chem. Biophys. 387: 11–19.

Moreland, J. L., A. Gramada, O. V. Buzko, Q. Zhang and P. E.Bourne, 2005 The Molecular Biology Toolkit (MBT): a modu-lar platform for developing molecular visualization applications.BMC Bioinformatics 6: 21.

Ptacek, J., and M. Snyder, 2006 Charging it up: global analysis ofprotein phosphorylation. Trends Genet. 22: 545–554.

Remenyi, A., M. C. Good and W. A. Lim, 2006 Docking interactionsin protein kinase and phosphatase networks. Curr. Opin. Struct.Biol. 16: 676–685.

Sadler, A. J., and B. R. Williams, 2007 Structure and function of theprotein kinase R. Curr. Top. Microbiol. Immunol. 316: 253–292.

Schneper, L., K. Duvel and J. R. Broach, 2004 Sense and sensibil-ity: nutritional response and signal integration in yeast. Curr.Opin. Microbiol. 7: 624–630.

Searle, J. S., K. L. Schollaert, B. J. Wilkins and Y. Sanchez,2004 The DNA damage checkpoint and PKA pathways con-verge on APC substrates and Cdc20 to regulate mitotic progres-sion. Nat. Cell Biol. 6: 138–145.

Shi, Z., K. A. Resing and N. G. Ahn, 2006 Networks for the allo-steric control of protein kinases. Curr. Opin. Struct. Biol. 16:686–692.

Smith, C. M., E. Radzio-Andzelm, Madhusudan, P. Akamine andS. S. Taylor, 1999 The catalytic subunit of cAMP-dependent pro-tein kinase: prototype for an extended network of communica-tion. Prog. Biophys. Mol. Biol. 71: 313–341.

Taylor, S. S., C. Kim, D. Vigil, N. M. Haste, J. Yang et al.,2005 Dynamics of signaling by PKA. Biochim. Biophys. Acta1754: 25–37.

Thevelein, J. M., and J. H. de Winde, 1999 Novel sensing mecha-nisms and targets for the cAMP-protein kinase A pathway inthe yeast Saccharomyces cerevisiae. Mol. Microbiol. 33: 904–918.

Toda, T., I. Uno, T. Ishikawa, S. Powers, T. Kataoka et al., 1985 Inyeast, RAS proteins are controlling elements of adenylate cyclase.Cell 40: 27–36.

Torkamani, A., N. Kannan, S. S. Taylor and N. J. Schork,2008 Congenital disease SNPs target lineage specific structuralelements in protein kinases. Proc. Natl. Acad. Sci. USA 105:9011–9016.

Ubersax, J. A., and J. E. Ferrell, Jr., 2007 Mechanisms of specific-ity in protein phosphorylation. Nat. Rev. Mol. Cell Biol. 8: 530–541.

Wu, J., S. H. Brown, S. von Daake and S. S. Taylor, 2007 PKA typeIIalpha holoenzyme reveals a combinatorial strategy for isoformdiversity. Science 318: 274–279.

Zhou, T., L. Sun, J. Humphreys and E. J. Goldsmith,2006 Docking interactions induce exposure of activation loopin the MAP kinase ERK2. Structure 14: 1011–1019.

Communicating editor: M. Nonet


Documents

Distal Recognition Sites in Substrates Are Required for ... · Distal Recognition Sites in Substrates Are Required for Efﬁcient Phosphorylation by the cAMP-Dependent Protein Kinase