high affinity binding of the heat-stable protein kinase inhibitor to

Biochemistry and Molecular Biology I nc Vol. 269, No. 11, Issue of March 18, pp. 8423-8430. 1994

P n n t d 1 n U.S.A.

High Affinity Binding of the Heat-stable Protein Kinase Inhibitor to the Catalytic Subunit of CAMP-dependent Protein Kinase Is Selectively Abolished by Mutation of Arg 133*

(Received for publication, November 1, 1993)

Wei Wen and Susan S. Taylor+ From the University of California, San Diego, Department of Chemistry, La Jolla, California 92093-0654

The two classes of physiological inhibitors of the catalytic subunit of CAMP-dependent protein kinase are the regulatory subunits and the heat-stable protein kinase inhibitors (Pus) , and both share a common mechanism of inhibition. Each has a similar inhibitor site that resembles a peptide substrate, and this occupies the P-3 to P+l portion of the peptide recognition site. However, in addition to this consensus site, each inhibitor requires a peripheral binding site to achieve high affinity binding.

and Arg133 lie on the surface of the catalytic subunit with Arg'33 coming close to the amphipathic helix of PKI(5-24) (Knighton, D. R., Zheng, J., Ten Eyck, L. F., Xuong, N.-h., Taylor, S. S., and Sowadski, J. M. (1991) Science 253,414-420). Replacement of and with Ala selectively abolishes the high affinity binding of PKI. Replacement of alone is sufficient to give the same phenotype. In the presence of MgATP, the Kd,app, is increased from ~0.2 to 105 nM and, in the absence of ATP, the Kd is too large to be reliably measured. Based on the crystal structure, Arg133 hydrogen bonds to the P-7 backbone carbonyl of PKI(5-24). However, more importantly, it also contributes to the hydrophobicity of the P-11 binding site in the C.PKI(5-24) complex. We predict that it is the perturbation of this hydrophobic pocket that accounts for the effects of this mutation. In the absence of peptide, Arg133 may help to stabilize GluZ3O, a buried carboxylate that binds to the P-2 Arg in the crystal structure of C.PKI(5-24). Replacement of Arg133 and Arg134 with Ala has little effect on catalysis using a heptapeptide substrate and has no effect on the inhibition of the catalytic subunit by the regulatory subunit. The results thus demonstrate that these two inhibitor proteins that both bind to the catalytic subunit with a high affinity utilize different sites on the enzyme to achieve tight binding.

The y isoform of the catalytic subunit is insensitive to inhibition by PKI and in this isoform Arg133 is replaced with Gln. We predict that this change accounts for the altered inhibitor properties of Cy.

CAMP-dependent protein kinase (cAPK)' is unusual among the protein kinase family in that its activation is mediated by

* This work was supported in part by American Cancer Society Grant BE-48 (to S. S. T.1. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduerti.sernent" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Dept. of Chemistry, 0654, University of California, San Diego 9500 $ To whom correspondence and reprint requests should be addressed:

Gilman Dr., La Jolla, CA92093-0654. Tel.: 619-534-3677; Fax: 619-534- 8193; Email: staylorOucsd.edu.

C, catalytic subunit of cAPK FITC, fluorescein 5-isothiocyanate; ' The abbreviations used are: cAPK, CAMP-dependent protein kinase;

MOPS, 3-(N-morpholino1propanesulfonic acid; PKI, heat-stable protein kinase inhibitor; PKI(5-241, NH,-terminal peptide of PKI; R, regulatory

a mechanism that involves the dissociation of regulatory (R) and catalytic (C) subunits. The free catalytic subunit, therefore, has the potential to interact with a variety of different inhibitor proteins. There are actually two classes of protein inhibitors of the C-subunit. First are the R-subunits that bind tightly only in the absence of CAMP (1). Within this class are at least four unique gene products, R', (21, R', (3) R",,(4), and R"B (5). The other class of inhibitors are the heat-stable protein kinase inhibitors (PHs) (6, 7). These are small proteins that bind with high specificity to the C-subunit. Three unique gene products are known: PKIa (81, PKIP1, and PKIp2 (9, 10). PKIP can be expressed as two different proteins due to alternative splicing of the RNA at the NH2 terminus of the protein (10). So far no physiological mechanism is known for reversing this mechanism, although in vi tro CAMP-free R can successfully compete for C when it is complexed with PKI i l l ) .

In contrast to typical substrates, these inhibitor proteins bind to the C-subunit with a high affinity, typically in the subnanomolar range (12, 13). Both classes of inhibitor also use a common mechanism for inhibiting the catalytic activity of the C-subunit. Specifically, both PKI and the R-subunits contain an autoinhibitor site that resembles a peptide substrate (Fig. 1). The minimum consensus site for recognition of a peptide substrate is R-R-X-SPT-X', where X is a hydrophobic residue. This consensus sequence is common to most substrates and inhibitors of cAPK (14). In both the R-subunit and PKI this consensus region occupies the peptide binding site and renders the enzyme inactive by preventing the binding of other substrates. PKI and the R-subunits are thus competitive inhibitors of peptide substrates.

A crystal structure of the C-subunit of cAPK was obtained by co-crystallizing with a 20-residue inhibitor peptide, PKI(5-24), derived from PKI( 15-17). The structure of the ternary complex containing ATP as well a s PKI(5-24) describes the total inhibitor complex (18-21). These structures illustrate how the consensus sequence binds to the active site of the enzyme. They also explain in part the high affinity binding of PKI. An essen- tial feature for the high affinity binding of PKI is an amphipathic helix that precedes the consensus site and occupies a hydrophobic patch on the surface of the large lobe of the C- subunit. Like the R'-subunits, the high affinity binding of PKI requires the synergistic binding of MgATP (11, 13, 22).

The R-subunits will presumably occupy the consensus site of the C-subunit in a manner that is analogous to PKI. However, there is evidence to suggest that the R-subunits do not utilize the same mechanism as PKI to achieve high affinity binding. For example, based on sequence alone, it is unlikely that the

subunit of cAPK, rC, recombinant catalytic subunit; R', type I recombinant regulatory subunit; R", type I1 recombinant regulatory subunit; R133A, Argl:':+ in C-subunit substituted with Ala; R133,134A, Argl"" and Arg':j" in C substituted with Ala; R209K, Arg"Ig in R'-subunit substituted with Lys.

8423

at Biom

edical Library, UC

SD

on Septem

ber 26, 2006 w

ww

.jbc.orgD

ownloaded from

http://www.jbc.org

8424 cAPK, Importance o f Arg'33 for PKI Binding

FIG. 1. Physiological inhibitors of

high affinity binding site regions of PKI the catalytic subunit. Sequences in the

and the R'- and R"-subunits are compared with a commonly used heptapeptide substrate. The sequence NH, terminus to the conserved site is important for the high affinity binding of PKI (71, whereas the region COOH terminus to the consensus site appears to be important for the R-subunit binding (25, 26).2

Protein Kinase Inhibitor PKI Ki = 0.2 nM

/ \ T T Y A D F I A S G R T C ~ ~ ~ H D PKl(5-24) Ki I 2.3 nM

A E V Y T E E D A A S Y AI-Subunit Kd = 0.23 nM Kemptide Km = 16 pM

Dimerization Domain

segment of R' and R" that precedes the consensus site forms an a-helix (23, 24). Proteolysis also supports this conclusion since cleavage just before the consensus site of the RII-subunit does not eliminate high affinity binding of the R"-subunit to the C-subunit (25, 26). A mutant form of the R',-subunit where residues 1-91 are deleted further supports this conclusion.2

In an effort to identify specific regions beyond the consensus recognition site that bind to the R-subunit, Gibbs et al. (28) used a family of mutants of the yeast C-subunit, TPKI. These mutants were generated by selectively replacing all of the charged residues in TPKI with Ala. They then screened for mutants that were catalytically intact in terms of phosphorylating the heptapeptide, LRRASLG, but were not regulated properly by the R-subunit (29). Three mutants were isolated from this screen. One mutant, R133,134A, was located on the NH2-terminal side of the consensus site. The rest, K189A,K192H and K213A,K217A were located on the surface that lies just beyond the COOH-terminal side of the consensus sequence close to the autophosphorylation site, ThrIg7. In order to better understand these mutant proteins, the same mutations were introduced into the mammalian C-subunit. We describe here the consequences of replacing Arg133 and Arg134 with Ala in the mammalian C-subunit. The single mutant where onlyArg133 is replaced with Ala shows the same phenotype as the double mutant.

EXPERIMENTAL PROCEDURES Materials-MOPS was from Sigma; fluorescein 5-isothiocyanate

(FITC) was from Molecular Probes; Centricon 30 was from Amicon; and Mono Q HR5/5, Mono S HR 10/10, Superose 12 HR 10130, Superdex 75 HR 10/10, PD10, and NAP 10 columns were from Pharmacia LKB Biotechnology Inc. The peptide substrate, Leu-Arg-Arg-Ala-Ser-Leu- Gly (Kemptide), were synthesized at the Peptide and Oligonucleotide Facility at the University of California, San Diego. Oligonucleotides were synthesized with a DNA synthesizer (Applied Biosystem Inc. model 380B).

Site-directed Mutagenesis-The mutations in the C-subunit were introduced as described by Kunkel(30,31). The following oligonucleotides were used to introduce the double and single amino acid changes, respectively.

R133,134A: 5"GCTGAACCTTCCAATAGCTGCTAGGTGGG-3' SEQUENCE 1

R133A: 5"GCTGAACCTTCCAATCCGTGCTAGGTGGG-3' SEQUENCE 2

Bases differing from the wild-type cDNA are indicated in bold. All mutations were confirmed by DNA sequence analysis. The mutants were expressed inEscherichia coli BL21 (DE-3) using a vector described previously, pLWS-3 (32).

Preparation of Protein-The recombinant wild-type C-subunit (rC- subunit) and the R133,134A double mutant rC(R133A.Rl34A) were expressed in E. coli and purified by phosphocellulose chromatography as described previously (32). The purification of the R133A single mu-

W. R. G. Dostmann, F. W. Herberg, and S. S. Taylor, manuscript in preparation.

unit CAMP Binding Domain A CAMP Bindlng Domain B

tant r(R133A) was achieved by holoenzyme formation in vitro with mutant RI-subunit (R209K) according to the procedure described by Yonemoto et al. (33). The rC-subunits were further purified on a Mono S HR10/10 ion exchange column using an fast protein liquid chromatography instrument (Pharmacia LKJ3 Biotechnology Inc.) according to Herberg et al. (34). The major peak corresponding to isoform I1 was used for all experiments.

Two types of R-subunits were used. The type I R-subunit forms holoenzyme in a manner that is dependent on ATP. The wild-tye RI- subunit and a mutant form of the R1-subunit were expressed in E. coli 222 under conditions described previously (35). This mutant RI-subunit, rR(R209K1, contains a point mutation in CAMP binding site A so that CAMP binds poorly. This rR mutant forms holoenzyme instantaneously when mixed with the rC-subunit. The type I1 R-subunit was expressed in E. coli. BL2UDE-3) as described earlier (36).

The recombinant wild-type RI-subunit (rR) and the mutant R'-subunit, rR(R209K), were purified by DEAE-52 chromatography as described previously (37). The recombinant RII-subunit was purified first using a CAMP affinity column followed by Mono Q ion exchange chromatography. Briefly, the R"-subunit eluted from the CAMP affinity column (36) was dialyzed against buffer containing 50 mM Tris (pH 7.41, 5 mM p-mercaptoethanol, 2 mM EDTA, and 2 mM EGTA. The dialyzed protein was then applied to a Mono Q column equilibrated in the same buffer and eluted with a salt gradient from 0 to 500 mM KCl.

The recombinant skeletal muscle heat-stable protein kinase inhibitor (PKIa) was expressed in E. coli BL2UDE-3) and purified by DEAE-52 chromatography as described earlier (38). PKI was purified further using Superdex G-75 gel filtration chromatography in the following buffer (50 mM Tris (pH 7.41, 150 mM KCl, 2 mM EDTA).

All proteins were a t least 95% pure as judged by SDS-polyacrylamide gel electrophoresis and analytical gel filtration. SDS-polyacrylamide gel electrophoresis was carried out according to Laemmli using 12.5% acrylamide for C-subunit and R-subunit and 154 acrylamide for PKIa (39).

Fluorescent Labeling of C-subunit and PKI-Both the rC-subunits and PKI were labeled with fluorescein 54sothiocyanate (FITC) using procedures that were described previously (40). The rC-subunit was labeled with a 35-fold molar excess of FITC for 30 min a t 22 "C in the labeling buffer containing 100 mM Hepes (pH 8.01, 8 mM MgCl,, 4 mM ATP. The labeled rC-subunit retained more than 97% of its initial phos- photransferase activity and was able to form holoenzyme with the rR1- subunit in a manner similar to unlabeled rC-subunit.

PKI was labeled under the same conditions as described for the rC-subunit except that no MgATP was included in the labeling buffer. Labeled and unlabeled PKI showed no difference in their capacity to inhibit the rC-subunit based on the spectrophotometric assay.

Protein Kinase Activity Assays-All activity assays were performed using the spectrophotometric method described by Cook et al. (41). This assay couples the production of MgADP to pyruvate kinase and lactate dehydrogenase, and the resulting decrease in absorbance a t 340 nM due to the oxidation of NADH is followed. Kemptide, LRRASLG, was used as substrate.

Holoenzyme Formation and Purification-In order to evaluate the role of MgATP in stabilizing the holoenzyme, holoenzyme was formed initially in the absence of ATP. FITC-labeled C-subunit (1 mg/ml) was mixed with wild-type R-subunit in a molar ratio of 1.2:l and dialyzed against buffer without ATP (25 mM potassium phosphate (pH 6.8),2 mM p-mercaptoethanol, 2 mM EDTA) for 3 days a t 22 "C. The resulting holoenzyme was filtered through a 0 .22 -p~ nylon filter (Rainin Instru- ment Co.) and loaded onto a Superose 12 HR10/30 column equilibrated with buffer (20 mM MOPS (pH 7.01, 150 mM KC1, 2 mM p-mercaptoeth-

at Biom

edical Library, UC

SD

on Septem

ber 26, 2006 w

ww

.jbc.orgD

ownloaded from

http://www.jbc.org

cAPK, Importance of Arg133 for PKI Binding

TABLE I Kinetic Droaerties of the R133.134A mutant

Wild type

R133,134A R133,134.4’ wild type

Knqxptide (p) 38 93 2.4 L . A T P (PM) 18 55 3.0 kcat (S-l) 20 33 1.6 Knt.pcpt idkcat 1.9 3.0 1.5 Km.ATPfkcat 0.9 1.6 2.0

anal, 0.2 m~ EDTA). The holoenzyme, eluting at a retention time of 13.28, was collected. The holoenzyme was concentrated to about 10 p~ by Centricon 30. The Concentration of holoenzyme was measured by titrating the concentration of free C in the presence of CAMP with PKI using the spectrophotometric assay to measure loss of enzyme activity.

Analytical Gel Filtration-In order to measure binding of R and PKI to C without depending on an activity assay, analytical gel filtration was performed according to the methods described by Herberg and Taylor (11). All the experiments for R-C interaction were carried out using a Superose 12 HR10/30 column, whereas all the experiments for PKI-C interaction were carried out using a Superdex G-75 column. Various concentrations of holoenzyme or PKI.C complex were preincubated for 30 minutes and then injected using a 25 p1 sample loop unless otherwise specified. Protein was detected in a double detector system connecting a UV M monitor (Pharmacia LKB) with a fluorescence detector (Hitachi fluorescence spectrophotometry) using an absorption wavelength of 493 nM and a n emission wavelength of 520 nM. The percentages of holoenzyme and PKI.C-subunit complex were calculated based on the peak areas corresponding to the complex and free protein. The values of the holoenzyme peak area were corrected by a quench factor of 0.77 in the presence of MgATP and 0.79 in the absence of MgATP, since the formation of holoenzyme quenched the fluorescent signal. The correction fac- tors for the C.PKI complex peak was 1.1 in the presence of MgATP.

RESULTS

Basic Kinetic Parameters of R133,134A--Before considering its capacity to be inhibited by PKI and the R-subunits, the kinetic properties of the mutant protein were characterized. These are summarized in Table I. The mutant did show a 2-3-fold increase in its K,,, for both peptide and ATP compared with the wild-type rC-subunit using the heptapeptide LR- RASLG. However, these changes were not major and the kcat was slightly increased. These mutations, therefore, did not have a significant impact on catalytic activity.

Interaction with Regulatory Subunits-Having established that the kinetic properties of mutant C-subunit were comparable with the wild-type catalytic subunit, inhibition by the two R-subunits was investigated. The interaction of the C-subunit with the R’-subunit was first determined by measuring the ability of the R1-subunit to inhibit the activity of the C-subunit using the spectrophotometric assay. A mutant form of the R1- subunit, R209K, that exhibits lower affinity for CAMP by ap- proximately an order of magnitude (351, was used since this mutant R’-subunit is able to bind C-subunit rapidly. A fixed amount of wild-type or mutant C-subunit was incubated with increasing amounts of R209K R’-subunit for 1 min at 22 “C. Kemptide was then added to initiate the kinase reaction. As shown in Fig. 2, the activity of the R133,134A C-subunit mutant (Fig. 2, panel A ) could be titrated readily by R209K R’- subunit. Linear titration line was obtained, when R133,134A was used at 43 nM, indicating that the affinity of double mutant C-subunit for R1-subunit is at least 10 times lower than the concentration of mutant C-subunit used in this assay condition, In other words, the dissociation constant (Kd) for R133,134A holoenzyme is less than 4.3 nM in the presence of ATP. This is quite comparable with the wild-type rC-subunit.

To further investigate the R-C interaction, analytical gel filtration was employed. This assay was described in a previous paper (11). Using this method, the direct measurement of the extent of holoenzyme formation in the presence and absence of

0 10 20 30 40 50 60 R209K Mutant R (nH)

8425

15j x , \ R133A

0 0 10 20 30 40 50 60 70 80

R209K Mutant R (nH)

FIG. 2. Inhibition of the activity of mutan t and wild-type catalytic subunits b y the mutan t (R209K) regulatory subunit. In the panel A, either the wild-type rC-subunit (0) or mutant rC(R133A,R134A)-subunit (0) (43 n~ each) was preincubated with increasing amounts of the mutant RI-subunit, rR(R209K), for 1 min at 22 “C in the assay buffer as described under “Experimental Proce- dures.” Kemptide was added to initiate the kinase reaction. The percent activity represents the ratio of remaining C-subunit activity assayed in the presence of rR(R209K) compared with the initial activity assayed in the absence of rR(R209K). Panel B shows the inhibition of activity of the &subunit (0) and rC(R133A) (0) (78 nM each) by rR(R209K).

MgATP can be achieved. For these studies, wild-type rR-subunit was used, not the mutant R. Holoenzyme was prepared by dialyzing the mixture of FITC-labeled rC-subunit and rR-subunit against the buffer in the absence of MgATP. The holoenzyme was then purified by gel filtration and was subsequently concentrated. Serial dilutions of the concentrated holoenzyme were applied to a Superose 12 column. This column readily separates holoenzyme from the free C-subunit. The percentage of holoenzyme present at various concentrations initially applied to the column was determined based on the peak areas of holoenzyme and free C-subunit. The apparent dissociation constant for wild-type and R133,134Amutant type I holoenzyme in the absence of MgATP are 390 and 580 nM, respectively, as shown in Fig. 3. In the presence of 100 UM ATP and 1 mM MgC12, the R’-C interaction of both wild-type holoenzyme and R133,134A holoenzyme are much stronger (Fig. 3, panel A). The apparent Kd values could not be determined accurately because they exceeded the sensitivity of our detection system. However, the data indicate that the apparent Kd values for both wild-type holoenzyme and for holoenzyme formed with R133,134A are less than 1 nM. Therefore, the R133,134A mutant C-subunit can still recognize the R’-subunit in a similar way to wild-type C-subunit both in the presence and absence of MgATP.

Since the binding of ATP to the type I holoenzyme synergistically enhances the association of R- and C-subunits, the effect of MgATP on the holoenzyme formation was also measured using analytical gel filtration. Holoenzyme was diluted to a concentration of 1 nM into buffer (20 mM MOPS (pH 7.0) 150 mM KCI, 2 mM P-mercaptoethanol, 1 mM MgC12) containing various concentrations of ATP and then applied to the Superose 12 column pre-equilibrated with the same buffer. As shown in Fig. 4, MgATP facilitated the formation of both wild-type and mutant type I holoenzyme. The concentration of ATP required for 50% holoformation under these conditions was 267 nM for wild-

at Biom

edical Library, UC

SD

on Septem

ber 26, 2006 w

ww

.jbc.orgD

ownloaded from

http://www.jbc.org

8426 M K , Importance of Arg'33 for PKI Binding

Mutant +ATP

0

L N C

0 X x

Mutant -ATP

- 100 10' 102 103 104 105

Type I Holoenzyme (nM)

.1 1 10 100 1000 Type I1 Holoenzyme (nM)

FIG. 3. Concentration dependence of dissociation of wild-type and mutant holoenzymes. Inpanel A, holoenzymes containing rC (0, U) and rC(R133A,R134A) (0, 0) were prepared as described under "Experimental Procedures." Serial dilutions of each holoenzyme in the presence (U, 0) and the absence (0, 0) of MgATP were applied to the Superose 12 column. When MgATP was present, 100 PM ATP and 1 mM MgClz were used. The percentage of holoenzyme at various concentrations was calculated based on the peak areas of holoenzyme and free C-subunit. Panel B showed the concentration dependence of dissociation for the type I1 holoenzymes formed with the wild type (0) and

MgATP. R133,134A (0). The type I1 holoenzymes were run in the absence of

100

u E

I"

c 4 50

x

0

loo 10' 10' lo3 lo4 10" lo6 10' ATP (nM)

FIG. 4. MgATP dependence of the dissociation of wild-type and mutant holoenzymes. Wild-type holoenzyme (0) or holoenzyme formed with rC(R133A,R134A) (0) were prepared as described (1 nM). Each holoenzyme sample was incubated in the buffer with various concentrations of ATP for 30 min a t 22 "C and then loaded onto Su- perose 12 equilibrated with same buffer using a 2 0 0 4 sample loop. The percentage of holoenzyme at various concentrations of ATP was calculated based on the peak areas of holoenzyme and C-subunit.

type holoenzyme and 363 nM for R133,134A mutant holoenzyme. Therefore, the high affinity binding of ATP to the type I holoenzyme was not altered significantly in this mutant.

A 100

X

2 6 50 <

0 0 500 1000 1500 2000 2:

PKI (nM)

0 ~ " ' " " ' " ' 5000 10000

PKI(5-24) (nM)

L

k \ HlJJA

-0 1000 2000 3000 PKI [nM]

lytic subunits by PKI and PKI(5-24). rC (01, rC(R133A,R134A) (O), FIG. 5. Inhibition of the activity of wild-type and mutant cata-

and rC(R133A) (0) were preincubated with varying amounts of PKI (panels A and C ) or PKI(5-24) (panel B) for 1 min at 22 "C. Kemptide was added to initiate the reaction. The percent activity represents the ratio of remaining C-subunit activity assayed in the presence of inhibitor compared with the initial activity in the absence of inhibitor. Dixon plots are shown in the insets.

To test whether the R133,134A rC-subunit can still interact with R"-subunit, analytical gel filtration was performed under the same conditions as those used for measuring the Kd of the type I holoenzyme. The apparent Kd values for wild-type and R133J34Amutant type I1 holoenzyme are close to 0.5 n~ in the absence of MgATP (Fig. 3, panel B ) , indicating that the R133,134A mutant also interacts normally with R"-subunit.

Taken together, the R133,134Amutant is able to bind to both type I and type I1 R-subunits with an affinity similar to that of wild-type C-subunit.

Interaction with PKI-The interaction of R133,134A mutant C-subunit with PKI was first studied using the spectrophotometric method. Unlike the R-subunit, the heat-stable protein kinase inhibitor, PKI, as well as the NH2-terminal peptide of the inhibitor, PKI(5-24), do not inhibit the phosphotransfer- ence activity of R133,134A double mutant with a high affinity (Fig. 5, panels A and B ) . The inhibitory constant (Ki) values were determined by a Dixon plot shown in the insets. The Ki of PKI and PKI(5-24) for R133,134Amutant increased by a factor of 430 and 173, respectively, compared with that for wild-type C-subunit in the presence of MgATP.

To determine whether one of the two arginines was primarily responsible for this increased Ki, the single mutant, R133A, was studied. Like the double mutant, R133A behaved like wild- type C-subunit in terms of its inhibition by the R-subunit (Fig. 2, panel B ). However, when the Ki of PKI for this single mutant

at Biom

edical Library, UC

SD

on Septem

ber 26, 2006 w

ww

.jbc.orgD

ownloaded from

http://www.jbc.org

M K , Importance of Arg133 for P K l Binding

TABLE I1 Znhibition constant (KJ of wild-type and mutant C-subunit

by PKI and PH(5-24)

Wild typea R133,134A R133A

PKI 0.2 nM 86 nM 105 IIM PKI(5-24) 2.4 nM 435 nM ND

a K, of PKI and PKI(5-24) for wild-type C is from Ref. 27. ND, not determined.

C-subunit was measured using the spectrophotometric method, an increase of 525-fold was observed compared with that for wild-type C-subunit (Fig. 5, panel C ) . This mutant thus dem- onstrates that and not Arg134, is primarily responsible for the decreased affinity of the mutant protein for PKI. All the Ki values in the presence of MgATP are summarized in Table 11.

Since, like the R1-subunit, the interaction of PKI and C is synergistically dependent on MgATP, the increased Ki in the presence of MgATP may have two explanations. One possibility is that the mutant C-subunit can still recognize PKI normally but that MgATP can no longer facilitate the tight complex formation. Alternatively, this mutant C-subunit may be defective in recognizing PKI both in the presence and absence of ATP. To resolve these questions, analytical gel filtration was used to directly measure the complex formation in the presence and absence of MgATP. The same conditions to those described for holoenzyme were used, except that the complex of FITC-labeled PKIC-subunit and free FITC-labeled PKI were separated using Superdex G-75 chromatography. No complex was formed in the absence of MgATP at concentrations up to 60 p~ of both C-subunit and PKI (Fig. 6, panel A). This suggested that the R133,134A mutant C-subunit is defective in its recognition of PKI, not just in its capacity to bind ATP with a high affinity. In the presence of 1 mM ATP and 2 mM MgC12, labeled PKI forms a complex very poorly with the R133,134A mutant, indicating that the presence of MgATP can stabilize the complex formation (Fig. 6, panel A). To investigate the effect of MgATP on the formation of the PKIC complex, 60 VM of the rC(R133A,R134A) was incubated with 1 p~ of PKI in the presence of various concentrations of MgATP. Samples were then applied onto the Superdex G-75 column to separate the complex and free FITC- labeled PKI. As shown in Fig. 6, panel B , an increased propor- tion of complex was observed with increasing amount of MgATP. 50% PKI was associated with R133,134A in the presence of 7 p~ ATP and 1 mM MgC1,. Using same amounts of wild-type C-subunit and PKI, all of the wild-type C-subunit binds to PKI even in the absence of MgATP. When 100 n~ each of the wild-type C-subunit and PKI were incubated with increasing amounts of ATP, 220 nM of ATP was required to achieve half-maximal complex formation (11); however, under these conditions, no complex formation was observed for the R133,134A mutant C-subunit, even in the presence of 1 mM ATP. Therefore, R133,134A mutant C-subunit also needs higher ATP to stabilize the PKI.C complex. The R133,134A mutant thus is defective in its recognition of PKI as well as PKI(5-24) both in the presence and absence of MgATP.

DISCUSSION

Both classes of physiological inhibitors of the C-subunit, the R-subunits, and the PKIs contain an auto-inhibitory site that resembles peptide substrates. This region, extending from the P-3 through the P+l site, occupies the peptide binding site and renders the enzyme inactive by preventing the binding of other substrates. However, in addition to this region, these two inhibitors require other sites to achieve high affinity binding. To

A 100-

8427

PKkC-Subunit (nM)

B 1ooF k 2 - I

:" I I

ATP (nM)

FIG. 6. MgATP dependence of the PKI.C complex. Panel A, serial dilutions of the rC.PKI complex (0) (11) or the mutant

the presence (H, 0) and absence (0, 0) of MgATP. When MgATP was rC(R133A,R134A).PKI complex (0) were applied to Superdex G-75 in

present, 100 PM ATP and 1 m~ MgC1, were used for the rC-subunit complex, whereas 1 mM ATP and 2 m~ MgClz were used for the rC(R133A,R134A) double mutant complex. E , percentage of the rC- subunit.PKI complex formed at various concentrations of MgATP. For these experiments, 120 p~ C-subunit and 2 PM FITC-labeled PKI were used.

identify those additional sites, Gibbs et al . (29), using charge- to-alanine scanning mutagenesis, identified three mutants in the yeast C-subunit that recognized the consensus sequence but were not regulated properly by the R-subunit. One of these mutants, equivalent to R133,134Ain the mouse C-subunit, dis- played an ICso value that was increased 4.2-fold. The catalytic efficiency of this mutant in yeast also was reduced slightly due to a 2.3-fold increase in the K,,, for Kemptide. In order to char- acterize this mutation in vitro, Arg133 and Argi34 in the recombinant mouse C-subunit were both replaced with Ala. In addition, a single mutant was made where only Arg133 was replaced with Ala. These mutations did not have a major effect on catalytic activity similar to what was observed for the yeast enzyme (29).

To investigate interactions between the C-subunit and its two physiological inhibitors, two methods were used. The interactions were first determined by measuring the capacity of these proteins to inhibit the catalytic activity of the C-subunit in the presence of MgATP using a spectrophotometric activity assay. However, in order to compare the Kd,spp values for subunit interactions in the presence and absence of MgATP, it was necessary to use a method that was independent of the activity assay. To accomplish this, analytical gel filtration was used (11).

Using both of these methods, the R133,134A mutant was shown to function like the wild-type C-subunit in terms of forming holoenzyme with either the R'-subunit or the R"-subunit. Furthermore, no significant differences were found for the R'-subunit either in the presence or absence of ATP. This is in contrast to the yeast C-subunit, where the comparable mutation showed a 4.2-fold increase in the IC5,,. We attribute this to subtle differences between BCYl and the mammalian R-subunit. BCYl is a type I1 R-subunit, and it does not bind to the yeast C-subunit nearly as tightly as the mammalian R-subunit binds to the mammalian C-subunit. The difference in R-C interaction in the presence and absence of CAMP is only a little more than an order of magnitude (42) in contrast to the mam-

at Biom

edical Library, UC

SD

on Septem

ber 26, 2006 w

ww

.jbc.orgD

ownloaded from

http://www.jbc.org

8428 cAPK Importance of Arg133 for PKI Binding

malian R- and C-subunits, where the difference is 4-5 orders of magnitude (43). The yeast cAPK may, therefore, be more sen- sitive to small perturbations. Although the precise explanation for the differences between yeast and mammalian C A P & is not totally clear, it is unambiguous that the R133,134A mutant in the mammalian C-subunit does not interfere with holoenzyme formation or holoenzyme stability.

These findings with regard to holoenzyme formation are in stark contrast to complex formation with PKI. This mutation severely impairs the interaction of C with PKI, both in the

FIG. 7. A stereo view showing the immediate environment surrounding R133. This diagram was taken from the coordinates of the binary complex of PKI(5-24) and the C-subunit (17).

r

presence and absence of MgATP. The ICd values are increased by over 400-fold, so that under physiological conditions, this C-subunit would no longer be inhibited by PKI. The yeast C- subunit does not have a high affinity binding site for PKI. This can be attributed to alterations in the hydrophobic pocket that binds the critical P-11 Phe in PKI (44, 45). This hydrophobic pocket is comprised of T~P~~-Pro-Pro-Phe-Phe with w35, Phe239, and Pro236 interacting predominantly with the P-11 Phe as seen in Fig. 7. In the yeast TPKI, the equivalent of 11.0236 is replaced with a Thr and Phe239 is replaced with QT.

A

I.

L

2s 133

FIG. 8. Space filling model showing the location of and in the crystal structure of the catalytic subunit. Residues

PKI(5-24) is shown in red with P-11 Phe highlighted as turquoise. andArg134 on the surface of the large lobe are shown with their side chain 15-127, predominantly located in the small lobe, are shown in purple, whereas residues 128-350 are shown in pink. The inhibitor peptide

orange. The hydrophobic P-11 binding pocket is comprised of Tyr2", Pro236, and Phe239, all shown in yellow. Panel A shows the position ofArglS3 carbons in green and nitrogens in blue. G1uZ3" that comes close to and ion pairs with the P-2 Arg is shown in yellow with its oxygens in

relative to the overall structure of the ternary complex containing MgATP and PKI(5-24). Panel B shows the immediate environment surrounding with PKI(5-24) removed. Panel C shows the immediate environment surrounding where Arg133 is replaced with Ala.

at Biom

edical Library, UC

SD

on Septem

ber 26, 2006 w

ww

.jbc.orgD

ownloaded from

http://www.jbc.org

cAPK, Importance of Arg133 for PKI Binding 8429

TABLE 111 Distances between Arg133 and other residues in the catalytic

subunit.PKI(5-24) binary complex

C::iEt Residue Distance (A) ' b e of interactlon

Vanderwaals NE CE2 P - 11, Phe 3.90 NE CZ P - 11, Phe 3.91 CZ CZ P - 11, Phe 3.71 CZ CE2 P - 11, Phe 3.85 NH2 CE2 P - 11, Phe 3.98 NH2 CZ P - 11, Phe 3.38

NE 0 P - 7, Thr 2.98 CD 0 P - 7, Thr 3.62

Hydrogen bond

CZ C P - 7, Thr 3.93 CZ 0 P - 7, Thr 3.99 cz OEI ~ 1 ~ 2 3 0 3.41 Hydrogen bond and

NH, OEI G1uz3O 3.11 NH2 OEI G1u2"o 2.84

electrostatic interaction

These differences most likely account for its inability to bind PKI (46, 47).

In order to determine whether Arg133 or Arg'34 contributed more prominently to the altered properties of this mutant, a single mutant was engineered by replacing Arg'33 with Ala. Since this mutant behaves in all regards like the double mutant, we conclude that the phenotype of the double mutant is due predominantly to the Arg133 replacement.

Does the crystal structure of the C-subunit provide an explanation for the observed phenotype? A space filling model of this region of the C-subunit- is shown in Fig. 8 while the specific contacts of less than 4 A that Arg'33 makes with other neigh- boring atoms are summarized in Table 111. The proximity of Arg133 to the P-11 hydrophobic pocket is particularly clear. Arg'33 also hydrogen bonds to the backbone a-carbonyl of the P-7 residue. This backbone interaction is not important for the Kemptide, which only occupies the P-4 to P+2 sites, or for the R-subunits; however, it is important for PKI. Arg133 also is close enough to ion pair with GIuz3O. Based on its reactivity with the hydrophobic carbodiimide, dicyclohexyl carbodiimide, Gluz3' was predicted to be in a hydrophobic environment (48,49), and this is consistent with the crystal structure. In the binary and ternary complexes Glu"' ion pairs with the P-2 Arg and thus neutralizes the buried charge (16). How this negative charge is neutralized in the apoenzyme is unclear, but its interactions with Arg133 could become more prominent. Perturbation of this interaction may very well account for the 2-fold increase in K,,, for the peptide. However, the greatest effect of this mutation with respect to PKI may be to reduce the hydrophobicity of the P-11 pocket. The methylene carbons ~ f A r g ' ~ ~ contribute to the hydrophobicity of this site and come within 3.5-4.0 of the P-11 Phe. When the P-11 Phe is replaced with dihydro-Phe, an artificial amino acid that is even larger and more hydrophobic than Phe, the afinity of PKI(5-24) for the C-subunit is increased by an order of magnitude. This finding supports the conclusion that this is quite an extended hydrophobic pocket (7, 47). As seen in Fig. 8, replacing Argi33 with Ala leaves a deep hole, and we conclude that the C-subunit compensates by re- structuring the hydrophobic P-11 binding pocket.

Does this mutation have any physiological relevance? Al- though it could obviously be an important tool for probing the physiological role of PKI, the behavior of this mutant in vivo remains to be determined. However, there is another obvious physiological implication of this mutation. Three gene products of the C-subunit have been identified, Cu, Cp, and Cy (50-53). Cu is expressed constitutively in all cells, whereas the expres- sion of CP is much more tissue-specific. These two C-subunits are indistinguishable in their catalytic properties, and both can

be inhibited readily by both R and PKI. The Cy-subunit was first identified in human testis and is expressed only at low levels. The only property that has been identified so far that distinguishes Cy from Cu and Cp is that it can no longer be inhibited by PKI (54). In Cy, Arg133 is conspicuously replaced by Gln, and we predict that this single amino acid difference accounts for the inability of Cy to be inhibited by PKI. The physiological significance of this observation also remains to be established.

What implications does this mutation have for protein kinase function in general and, in particular, for understanding how protein kinases recognize their substrates and inhibitors? It reflects a rather remarkable flexibility when one compares how this enzyme recognizes its various classes of physiological inhibitors. To recognize all of these inhibitors, as well as its nor- mal substrates, the enzyme utilizes a common consensus recognition site that flanks the site of phosphotransfer. This extends, in general, from the P-3 Arg to the P+l site and, in some substrates such as PLCy, may extend as far as the P-6 Arg subsite. However, to achieve high affinity binding requires further interactions. The findings here demonstrate that the C-subunit can use one surface to achieve high affinity binding of PKI. This is the surface that complements the amphipathic helix of PKI that lies NHz-terminal to the consensus site. This surface, however, is not required to achieve high affinity binding of the R-subunits. Instead, the high affinity binding of the R-subunit apparently requires the surface of the C-subunit that lies COOH-terminal to the consensus site (29, 55). Whether the C-subunit uses similar flexibility in recognizing different substrates remains to be determined. Whether other protein kinases also display this type of versatility also remains to be established.

Acknowledgments-We thank the following individuals and re- sources for contributions: Drs. Joseph Adams for advice and kinetic analysis, Friedrich W. Herberg for help with fast protein liquid chromatography and thoughtful discussion; Maria L. McGlone, Siv Garrod, and Karen Prather for excellent technical assistance; and Gene Hase- gawa and Jill Baumgartner for assistance in preparation of this manuscript.

REFERENCES 1. Beebe. S. J.. and Corbin, J. D. (19861 in The Enzymes: Control by Phospho-

Academic Press, Inc., New York rylation Part A (Krebs, E. G., and Boyer, P. D., eds) Vol. XVII, pp. 43-111,

2. Lee, D. C., Carmichael, D. F., Krebs, E. G., and McKnight, G. S. (19831 Proc.

3. Clegg, C. H., Codd, G. G., and McKnight, G. S. (1988) Proc. Natl. Acad. Sci. Natl. Acad. Sci. U. S. A. 80, 3608-3612

4. Scott, J. D., Glaccum, M. B., Zoller, M. J., Uhler, M. D., Helfman. D. M., U. S. A. 85, 3703-3707

McKnight, G. S. , and Krebs, E. G. (19871 Proc. Natl. Acad. Sci. U. S. A. 84, 5192-5196

5. Jahnsen, T., Hedin, L., Kidd, V. J., Beattie, W. G., Lohmann, S. M., Walter, U., Purica, J., Schulz, T. Z., Schiltz, E., and Browner, M. (1986) 261, 12352-

6. Walsh, D. A., Ashby, C. D., Gonzalez, C . , Calkins, D., Fischer, E. H., and Krebs, 12361

7. Walsh, D. A., Angelos, K. L., Van Patten, S. M., Glass, D. B., and Garetto. L. P. E. G. (1971) J. Biol. Chem. 246, 1977-1985

(1990) in Peptides and Protein Phosphorylation (Kemp, B. E., ed) pp. 43-84, CRC Press, Inc., Boca Raton. FL

8. Olsen, S. R., and Uhler, M. D. (1991) J. Biol. Chem. 266, 11158-11162 9. Van Patten, S. M., Ng, D. C., Thng, J. P., Angelos, K. L., Smith, A. J., and

10. Scarpetta, M. A., and Uhler, M. D. (1993) J. Biol. Chem. 268, 10927-10931 Walsh. D. A. (19911 Proc. Natl. Acad. Sci. U. S. A. 88, 5383-5387

11. Herberg, F. W., and Taylor, S. S. (19931 Biochemistry 32, 14015-14022 12. Hofmann, F. (19801 J. Biol. Chem. 255, 1559-1564 13. Whitehouse, S., and Walsh, D. A. (1983) J. Biol. Chem. 258, 3682-3692 14. Zetterqvist, 0. Z., Ragnarsson, U., and Engstrom, L. (1990) in Peptides and

Protein Phosphorylation (Kemp, B. E., ed) pp. 171-187, CRC Press, Inc., Boca Raton, FL

15. Knighton, D. R., Zheng, J., Ten Eyck. L. F., Ashford, V A., Xuong, N.-h., Taylor, S. S., and Sowadski, J. M. (1991) Science 253,407414

16. Knighton, D. R., Zheng, J., Ten Eyck, L. F., Xuong, N.-h., Taylor, S. S., and Sowadski, J. M. (1991) Science 253.414420

17. Knighton, D. R., Bell, S. M., Zheng, J., Ten Eyck, L. F., Xuong, N.-h., Taylor, S. S., and Sowadski, J. M. (19931 Acta Crystallogr. D 49, 357-361

18. Zheng, J., Knighton, D. R., Ten Eyck, L. F., Karlsson, R., Xuong, N.-h., Taylor, S. S., and Sowadski, J. M. (1993) Biochemistry 32,2154-2161

19. Zheng, J., Trafny, E. A., Knighton, D. R., Xuong, N.-h., Taylor, S. S., Ten Eyck,

at Biom

edical Library, UC

SD

on Septem

ber 26, 2006 w

ww

.jbc.orgD

ownloaded from

http://www.jbc.org

8430 cAPK, Importance of Arg133 for PKI Binding

20. Taylor, S. S., Knighton, D. R., Zheng, J . , Sowadski, J. M., Gibbs, C. S., and

21. Bossemeyer, D., Engh, R. A,, Kinzel, V., Ponstingl, H., and Huber, R. (1993)

22. Van Patten, S. M., Fletcher, W. H., and Walsh. D. A. (1986) J. Biol. Chem. 261,

23. Titani, K., Sasagawa, T., Ericsson, L. H., Kumar, S., Smith, S. B., Krebs, E. G.,

24. Takio, K., Smith, S. B., Krebs, E. G . , Walsh, K. A,, and Titani, K. (1984)

25. Weldon, S. L., and Taylor, S. S. (1985) J. Bid. Chem. 260, 42034209 26. Weber, W., and Hilz, H. (1979iBiochem. Biuphys. Res. Commun. 90,1073-1081 27. Cheng, H:C., van Patten, S. M., Smith, A. J., and Walsh, D. A. (1986) Biochem.

28. Gibbs, C. S., and Zoller, M. J . (1991) Biochemistry 30, 5329-5334 29. Gibbs, C. S., Knighton, D. R., Sowadski, J . M., Taylor, S. S., and Zoller, M. J .

30. Kunkel, T. A. (1985) Proc. Natl. Acad. Sei. U. S. A. 82, 488492 31. Tasken, K., Skllhegg, B. S., Soldherg, R., Andersson, K. B., Taylor, S. S., Lea,

T., Blomhoff, H. K., Janson, T., and Hansson, V. (1993) J. Bid. Chem. 268, 21276-21283

L. F., and Sowadski, J. M. (1993) Acta Crystallogr: D 49, 362-365

Zoller, M. J . (1993) Trends Biochem. Sci. 18, 84-89

EMBO J. 12, 849-859

5514-5523

and Walsh, K. A. (1984) Biochemistry 23, 41934199

Biochemistry 23, 42004206

J. 231,655-661

(19921 J. Bid. Chem. 267, 4806-4814

32. Slice, L. W., and Taylor, S. S. (1989) J. Bid. Chem. 264, 20940-20946 33. Yonemoto, W., McGlone, M. L., Slice, L. W., and Taylor, S. S. (1991) in Protein

Phosphorylation (Part A ) (Hunter, T., and Sefton, B. M., eds) Vol. 200, pp. 581-596, Academic Press, Inc., San Diego, CA

34. Herherg, F. W., Bell, S. M., and Taylor, S. S. (1993) Protein Eng. 6, 771-777 35. Bubis, J., Neitzel, J . J., Saraswat, L. D., and Taylor, S. S. (195815. Bid . Chem.

36. Scott, J . D., Stofko, R. E., McDonald, J. R., Comer, J. D., Vitalis, E. A., and

37. Buechler, Y. J., and Taylor, S. S. 11991) J. Biol. Chem. 266, 3491-3497

263,9668-9673

Mangili, J . A. (1990) J. Biol. Chem. 265, 21561-21566

38. Richardson, J . M., Howard, F'., Massa, J. S., and Maurer, R. A. (1990) J. B i d .

39. Laemmli, U. K. (1970) Nature 227, 680-685 40. Buechler, Y. J., Herberg, E W., and Taylor, S. S. (1993) J. Bid. Chem. 268,

41. Cook, P. F., Neville, M. E., Vrana, K. E., Hartl, F. T., and Roskoski, J., R. (1982)

42. Kuret, J., Johnson, K. E., Nicolette, C., and Zoller, M. J. (1988) J. Biol. Chem.

43. Granot, J., Mildvan, A. S., Hiyama, K., Kondo, H., and Kaiser, E. T. (1980) J.

44. Glass, D. B., Cheng, H.-C., Mende-Mueller, L., Reed, J., and Walsh, D. A.

45. Glass, D. B., Lundquist, L. J., Katz, B. M., and Walsh, D. A. (19891 J. Biol.

46. Toda, T., Cameron, S., Sass, P., Zoller, M. J., and Wigler, M. (1987) Cell 50,

47. Glass, D. B., Feller, M. J. , Levin, L. R., and Walsh, D. A. (1992) Biochemistry

48. Buechler, J . A., and Taylor, S. S. (1988) Biochemistry 27, 7356-7361 49. Buechler, J. A,, and Taylor, S. S. (19901 Biochemistry 29, 1937-1943 50. Uhler, M. D., Carmichael, D. F., Lee, D. C., Chrivia, J. C., Krebs, E. G., and

51. Uhler, M. D., Chrivia, J . C., and McKnight, G. S. (1986) J. Biol. Chem. 261, McKnight, G . S. (1986) Proc. Natl. Acad. Sci. U. S. A. 83, 1300-1304

52. Showers, M. 0.. and Maurer, R. A. (1986) J. Bid. Chem. 261, 16288-16291 15360-15363

53. Beehe, S., Oyen, 0.. Sandberg, M., Froysa, A., Hansson, V., and Jahnsen, T.

54. Beebe, S. J. , Salomonsky, P., Jahnsen, T., and Li, Y. (1992) J. E d . Chem. 267,

55. Orellana, S. A,, and McKnight, G. S. (1992) Proc. Nutl. Acad. Sci. U. S. A. 89,

Chem. 265, 13635-13640

16495-16503

Biochemistry 21,5794-5799

263, 9149-9154

Bid. Chem. 255, 45694573

(1989) J. Biol. Chem. 264,8802-8810

Chem. 264, 14579-14584

277-287

31, 1728-1734

(1990) Mol. Endocrinol. 4, 465475

25505-25512

47264730

at Biom

edical Library, UC

SD

on Septem

ber 26, 2006 w

ww

.jbc.orgD

ownloaded from

http://www.jbc.org

high affinity binding of the heat-stable protein kinase inhibitor to

Documents