Conformational control of protein kinases is an important way of modulating catalytic activity. Crystal structures of the C (catalytic) subunit of PKA (protein kinase A) in complex with physiological inhibitors and/or nucleotides suggest a highly dynamic process switching between open and more closed conformations. To investigate the underlying molecular mechanisms, SPR (surface plasmon resonance) was used for detailed binding analyses of two physiological PKA inhibitors, PKI (heat-stable protein kinase inhibitor) and a truncated form of the R (regulatory) subunit (RIα 92–260), in the presence of various concentrations of metals and nucleotides. Interestingly, it could be demonstrated that high-affinity binding of each pseudosubstrate inhibitor was dependent only on the concentration of divalent metal ions. At low micromolar concentrations of Mg2+ with PKI, transient interaction kinetics with fast on- and off-rates were observed, whereas at high Mg2+ concentrations the off-rate was slowed down by a factor of 200. This effect could be attributed to the second, low-affinity metal-binding site in the C subunit. In contrast, when investigating the interaction of RIα 92–260 with the C subunit under the same conditions, it was shown that the association rate rather than the dissociation rate was influenced by the presence of high concentrations of Mg2+. A model is presented, where the high-affinity interaction of the C subunit with pseudosubstrate inhibitors (RIα and PKI) is dependent on the closed, catalytically inactive conformation induced by the binding of a nucleotide complex where both of the metal-binding sites are occupied.

INTRODUCTION

Protein kinases share extensive sequence homology and a general structural fold in the catalytic core region [1] that was exemplified by the first X-ray crystal structure of a protein kinase, the C (catalytic) subunit of PKA (cAMP-dependent protein kinase; protein kinase A) [2]. PKA is a tightly regulated enzyme, functioning as an on/off switch in intracellular signal transduction. The inactive holoenzyme, consisting of two dimeric R (regulatory) subunits and two C subunits, is activated upon binding of the second messenger cAMP to the two highly conserved cyclic-nucleotide-binding domains of the R subunits. Another key player for the regulation of activity and for localizing the C subunit is PKI (heat-stable protein kinase inhibitor) [3], which also functions as a nuclear scavenger [4]. Moreover, PKI is also involved in neuronal signal transduction with effects on learning and memory by affecting LTP (long-term potentiation) and LTD (long-term depression) in the rat hippocampus [5].

Several high-resolution crystal structures of the C subunit have been solved [2,68], showing a bilobal structure with MgATP bound in the active-site cleft between the two lobes. It is currently well established by crystallographic and thermodynamic evidence that ligand binding (nucleotides, substrates and natural or synthetic inhibitors) as well as the interaction with the two classes of physiological inhibitors, the regulatory subunits (types I and II) and PKI, induce structural changes in the PKA C subunit that shifts the enzyme from an open to a more closed conformation [9]. The activity of the C subunit of PKA is controlled by the modulation between active and inactive states.

Generally, serine/threonine-specific protein kinases require MgATP, whereas MnATP seems to be the preferred substrate of the tyrosine-specific protein kinases [7,10]. At the same time many protein kinases are also inhibited by an excess of metal ions, so that active and inactive conformations can both exist in solution depending on the stoichiometry of metal binding [11]. The existence of two metal ion-binding sites in PKA and the role that the metal ions play in catalysis and inhibition has been investigated using a variety of methods [1115]. However, the importance of the metal- and nucleotide-binding sites for the structure and function of protein kinases is difficult to dissect.

In the absence of cAMP the C subunit is inhibited by forming a tight complex with the R subunits type I or type II. PKI and type I R subunit represent pseudosubstrate inhibitors, both containing an alanine residue instead of a serine residue (as in the RII subunit) in the autoinhibition site [16]. PKA does not only require Mg2+ and ATP for catalysis but also for inhibition by the two pseudosubstrate inhibitors. Type I R subunit and PKI have an absolute requirement for the presence of metal/ATP to maintain an inactive complex [17,18]. In contrast, high-affinity binding to the type II R subunit has no requirement for metal/ATP. All of these physiological inhibitors of the C subunit serve as competitive inhibitors with respect to protein substrates. Although MgATP binds with low affinity to the free C subunit with equivalent Km and KD values (10 μM, [13]), it binds with a high affinity (40–100 nM) to the inhibitor complexes (RI subunit and PKI [17,19]).

It was previously demonstrated that the metal–nucleotide complex in the active site cleft (i) increases the intrinsic stability of PKA C subunit, (ii) induces high-affinity binding to type I R subunit and PKI and, (iii) potentially induces the closing of the active site cleft by shifting PKA C subunit from an open to a closed conformation, the latter being thermodynamically more stable [12,17]. This effect was attributed to the second, low-affinity, metal-binding site chelating ADP and ATP in the active site.

The active site of PKA is deeply embedded within a pocket between the smaller ATP-binding domain and the larger substrate-binding domain. The catalytic cycle of the enzyme seems to be controlled by the binding of ATP, one or two metal ions and the release of ADP. The metal ions control the thermodynamic stability, as determined by thermal denaturation and CD [12], possibly by opening and closing of the active site cleft.

In the present study SPR (surface plasmon resonance) was employed to monitor directly the detailed association and dissociation kinetics of both PKI and a truncated form of R subunit type I (bovine RIα 92–260) with the C subunit in dependence of metal ions and nucleotides. Based on our investigations and the crystal structures of the C subunit in complex with PKI (5–24 peptide) and RIα 91–244 we present a model, where a high-affinity complex consisting of PKI or RIα and the C subunit could only be formed in the closed state, which is dependent on the presence of both metal ions and nucleotide.

MATERIALS AND METHODS

Reagents

The synthetic peptide substrate Kemptide (LRRASLG) was purchased from either Bachem Biochemicals or Biosyntan. Other reagents were purchased as follows: ATP, ADP, AMP, cAMP, adenine and p[NH]ppA (adenosine 5′-[β,γ-imido]triphosphate) from Sigma or from Biolog Life Science Institute. Fine chemicals and all other reagents were obtained in the purest grade available from Roth or from Sigma. CM5 sensor chips (research grade), goat anti-GST (glutathione transferase) antibody, NHS (N-hydroxysuccinimide), EDC [N-ethyl-N′-(dimethylaminopropyl)-carbodi-imide], ethanolamine-HCl and surfactant P20 were obtained from Biacore GE Healthcare. The DNA for the recombinant C and R subunits of PKA were a gift from Professor S.S. Taylor (Department of Chemistry and Biochemistry, University of California, San Diego, U.S.A.). The PKI cDNA corresponding to the rabbit skeletal muscle with a glycine residue after the start methionine residue, cloned into pT7-7, was a gift of Dr R. A. Maurer (Vollum Institute, Portland, Oregon, U.S.A.).

Expression and purification of recombinant proteins

Following overexpression in Escherichia coli, the PKA C subunit was purified by phosphocellulose chromatography (P11, Whatman) [20]. The purified C subunit was stored at 4 °C in buffer containing 20 mM potassium phosphate, 110 mM KCl and 5 mM 2-mercaptoethanol (pH 7.0). Bovine RIα 92–260 was overexpressed in E. coli BL21 (DE3) cells (17 h, 37 °C) and was purified as previously described [21]. Purified R subunits were stored at −20 °C in buffer A [20 mM Mops, 150 mM NaCl and 1 mM 2-mercaptoethanol (pH 7.0)] until further use. After subcloning PKIα cDNA into pGEX-KG in order to add a GST-fusion tag, PKIα was overexpressed in E. coli BL21 (DE3) cells and purified as a GST-fusion protein using glutathione–agarose (Sigma). Untagged protein was purified according to the procedure of Thomas et al. [22].

The purity of the recombinant proteins was confirmed by SDS/PAGE [23]. The biological activity of all proteins was verified using the spectrophotometric phosphotransferase assay according to Cook et al. [11] with the heptapeptide Kemptide (LRRASLG) as the substrate. Protein concentrations were measured by a colorimetric assay using BSA as a standard [24].

Calculations of the free and complexed metal ion concentrations

The concentrations of free and complexed metals were determined using the software BAD (Bound And Determined) by Brooks and Storey [25]. This software allows entering of multiple parameters which influence metal–ATP complex formation. Included are all ion species in the buffer and the resulting ionic strength, as well as the pH.

SPR measurements

SPR experiments were performed using Biacore 2000 and Biacore T100 instruments (Biacore GE Healthcare). In SPR, the interaction between an immobilized component referred to as the ligand and a molecule in the mobile phase, the analyte, is determined. Changes in surface concentration are proportional to changes in the refractive index on the surface resulting in changes in the SPR signal, plotted as RU (resonance units); 1000 RU corresponds to a surface concentration of 1 ng/mm2 [26].

For the interaction analysis of the Cα subunit with GST–PKI, an anti-GST antibody was immobilized to a level of 8000–10000 RU using standard NHS/EDC chemistry [27]. GST–PKI was then captured to a level of 800–1200 RU for high immobilization levels or less than 100 RU for low-density surfaces. Association and dissociation rates were determined by injecting a series of dilutions of the C subunit using a flow rate of 30 μl/min. All interaction studies were performed in buffer B [20 mM Mops, 150 mM NaCl (pH 7.0) and 0.005% surfactant P20] at 20 °C. After each interaction event, the entire GST–PKI–C subunit complex was removed and the antibody surface was regenerated using 10 mM glycine (pH 2.2). Less than 3 RU of residual binding was determined after the regeneration. This was a crucial prerequisite to obtain highly reproducible and accurate data by capturing exactly the same amount of GST–PKI in each cycle to probe against various concentrations of the C subunit. No reduction in the capacity of the anti-GST surface was detected during one set of experiments due to the regeneration procedure. Non-specific binding was removed by subtracting blank runs performed on surfaces with immobilized anti-GST antibody. Kinetic constants were calculated by non-linear regression of data using the Biaevaluation software version 3.1 and Biacore T100 Evaluation Software version 1.1 (Biacore GE Healthcare). A 1:1 binding model, assuming Langmuir conditions, was applied to the data as well as an equilibrium binding analysis under steady-state conditions in a plot analogous to a Scatchard-type analysis by plotting Req/C against Req to calculate Rmax and KD values, where Req is the analyte binding at equilibrium (steady state) in RU, C is the concentration of the analyte injected and Rmax is the maximum analyte binding capacity in RU. The binding constants were additionally calculated by non-linear regression analysis using the software GraphPad Prism, version 3.02 and 4.03 with the equation:

 
formula

where Bmax is the maximal binding, and Kd is the concentration of ligand required to reach half-maximal binding.

For the interaction analysis of RIα 92–260 with PKA Cα, the recombinant C subunit was immobilized on a CM5 sensor chip to a level of 500 RU using standard NHS/EDC chemistry (see above). Association and dissociation rates were determined by injecting a series of dilutions of the regulatory subunit using a flow rate of 25 μl/min. Both the association and dissociation phases were monitored for 240 s. The interaction studies were performed in buffer B (see above) at 20 °C. The respective concentrations of MgCl2 and ATP are indicated on the Figures and are mentioned in the Results section. Regeneration of the Cα subunit surface was performed by a washing step with 0.1 mM cAMP and 1 mM EDTA. Non-specific binding was removed by subtracting blank runs performed on a reference surface, which was activated and deactivated without coupling a protein. Data processing and evaluation was performed using Biaevaluation Software version 4.0.1 (Biacore GE Healthcare).

RESULTS

Nucleotides and metal ions play a pivotal role for cellular function and homoeostasis and furthermore seem to control stability and distinct conformations in a number of cellular proteins. Previous studies have established that the C subunit of PKA is more resistant against thermal denaturation in the presence of bivalent metal ions and nucleotides as determined by activity assays and CD [12]. Furthermore, it has been shown that the C subunit is stabilized when it is part of the type I holoenzyme complex or the PKI–C subunit complex [12]. In the present study we investigated the role of nucleotides and metal ions on the interaction with two important physiological inhibitors of PKA, PKI and the type I R subunit, using SPR. PKI, as well as the type I R subunit, are pseudosubstrate inhibitors of the C subunit and require ATP and Mg2+ for high-affinity binding.

Role of nucleotides for high-affinity binding of pseudosubstrate inhibitors

In a first approach, several nucleotides were tested in the presence and absence of a fixed concentration of MgCl2 (10 mM). Figure 1(A) shows qualitative plots of the association and dissociation patterns of the C subunit on GST–PKI using a fixed concentration of 16 nM of Cα subunit. In the presence of ATP and Mg2+, a fast association and a slow dissociation were monitored. However, when changing to ADP, both association and dissociation were affected, displaying slower association and faster dissociation kinetics. Since it is not clear whether ATP hydrolysis takes place once PKI has bound, a non-hydrolysable ATP analogue, p[NH]ppA, was tested. Surprisingly, the kinetics were strongly influenced with lower affinities indicating an improper steric setting probably owing to the exchange of an oxygen to a nitrogen between the β- and the γ-phosphate of ATP. In the absence of Mg2+ ions all tested nucleotides (ATP, ADP, AMP and p[NH]ppA) displayed very fast on- and off-rates causing transient interaction kinetics. With AMP, transient kinetics could be observed independent of the presence of metal ions, demonstrating the essential role of the β- and γ-phosphate of the nucleotide for complex formation.

Binding pattern of PKA C subunit to pseudosubstrate inhibitors based on SPR

Figure 1
Binding pattern of PKA C subunit to pseudosubstrate inhibitors based on SPR

(A) Interaction of 16 nM C subunit with GST–PKI captured to an anti-GST chip in the presence of adenine nucleotides {ATP, ADP, AMP and p[NH]ppA (AMP-PNP), 1 mM each, indicated on the plot} either in the presence or absence of 10 mM Mg2+. The experiments were carried out at 20 °C using a flow rate of 30 μl/min. Association and dissociation of C subunit to GST–PKI was monitored for 240 s and 360 s respectively. (B) Binding of 25 nM bovine RIα 92–260 to immobilized C subunit under the same conditions as described for (A). Association and dissociation were monitored for 240 s each. The experiment was performed at 20 °C at a flow rate of 25 μl/min.

Figure 1
Binding pattern of PKA C subunit to pseudosubstrate inhibitors based on SPR

(A) Interaction of 16 nM C subunit with GST–PKI captured to an anti-GST chip in the presence of adenine nucleotides {ATP, ADP, AMP and p[NH]ppA (AMP-PNP), 1 mM each, indicated on the plot} either in the presence or absence of 10 mM Mg2+. The experiments were carried out at 20 °C using a flow rate of 30 μl/min. Association and dissociation of C subunit to GST–PKI was monitored for 240 s and 360 s respectively. (B) Binding of 25 nM bovine RIα 92–260 to immobilized C subunit under the same conditions as described for (A). Association and dissociation were monitored for 240 s each. The experiment was performed at 20 °C at a flow rate of 25 μl/min.

Next, the interaction of the R subunit type I (bovine RIα 92–260, [21]) with the C subunit was investigated under the same conditions (Figure 1B). Again, as for the C subunit–PKI complex, ATP, ADP and p[NH]ppA together with Mg2+ could promote complex formation. Interestingly, in the presence of ADP and Mg2+ the interaction showed a slower association rate of the holoenzyme complex compared with the ATP and p[NH]ppA complexes. In the absence of Mg2+ ions, no binding of the R subunit to PKA Cα subunit could be observed using all tested adenosine nucleotides (ATP, ADP, AMP and p[NH]ppA). This was in contrast with the C subunit–PKI complexes, where at least a transient binding could be observed. This indicates that, in the case of the type I R subunit, the holoenzyme formation is strictly dependent on the presence of metal and nucleotide.

Apparent rate constants of PKI binding to PKA C subunit in the presence of MgATP

In order to obtain association and dissociation rate constants, various concentrations of C subunit were injected over the GST–PKI surface, ranging from 1 nM to 32 nM and 8 nM to 1 μM in the presence and in the absence of ATP and Mg2+ respectively (Figure 2). From those sets of binding curves, apparent rate constants were calculated using global fit analysis as well as separate fits. In the presence of 1 mM ATP and 5 mM MgCl2 a fast association rate (1.5×106 M−1·s−1) and a relatively slow dissociation rate (7.7×10−4 s−1) was calculated (Table 1 and Figure 2A). From the respective association and dissociation rate constants a KD value of 0.5 nM was calculated, which was in excellent agreement with previous observations [18]. Global fit analysis based on a Langmuir 1:1 model yielded the fits with the lowest χ2 values, almost perfectly fitting the model when using a low-density surface. This indicates a 1:1 binding mechanism for PKI and the C subunit.

Interaction of C subunit with captured GST–PKI in either the presence (A) or absence (B) of 5 mM Mg2+ and 1 mM ATP

Figure 2
Interaction of C subunit with captured GST–PKI in either the presence (A) or absence (B) of 5 mM Mg2+ and 1 mM ATP

(A) In the presence of MgATP the C subunit was injected in concentrations ranging from 1 nM to 32 nM (as indicated). Association and dissociation were monitored for 240 s and 300 s respectively. The interaction analysis was performed at 20 °C at a flow rate of 30 μl/min. Original results are shown as a solid line. Association and dissociation rate constants were determined using a global fit analysis according to a Langmuir 1:1 model, employing the software Biaevaluation 3.1 (see also Table 1). Resulting fits are shown as broken lines. (B) In the absence of MgATP the C subunit was injected in concentrations ranging from 31.25 nM to 1 μM (as shown). The experimental set-up was performed as described for (A). Association and dissociation rate constants were determined using a global fit analysis employing the Biacore T100 Evaluation Software version 1.1, KD values were determined either from the rate constants or from equilibrium binding analysis by non-linear regression (hyperbolic fit) and Scatchard plot analysis. Results are shown in Table 1.

Figure 2
Interaction of C subunit with captured GST–PKI in either the presence (A) or absence (B) of 5 mM Mg2+ and 1 mM ATP

(A) In the presence of MgATP the C subunit was injected in concentrations ranging from 1 nM to 32 nM (as indicated). Association and dissociation were monitored for 240 s and 300 s respectively. The interaction analysis was performed at 20 °C at a flow rate of 30 μl/min. Original results are shown as a solid line. Association and dissociation rate constants were determined using a global fit analysis according to a Langmuir 1:1 model, employing the software Biaevaluation 3.1 (see also Table 1). Resulting fits are shown as broken lines. (B) In the absence of MgATP the C subunit was injected in concentrations ranging from 31.25 nM to 1 μM (as shown). The experimental set-up was performed as described for (A). Association and dissociation rate constants were determined using a global fit analysis employing the Biacore T100 Evaluation Software version 1.1, KD values were determined either from the rate constants or from equilibrium binding analysis by non-linear regression (hyperbolic fit) and Scatchard plot analysis. Results are shown in Table 1.

Table 1
Binding of the PKA C subunit to immobilized GST–PKI

Apparent association and dissociation rate constants determined with SPR and calculated KD values of the interaction of PKA Cα subunit with immobilized GST–PKI. The ATP and Mg2+ concentrations used are indicated in the Table (n.d., not determined).

 Interaction conditions 
Kinetic parameter 1 mM ATP/5 mM MgCl2 1 mM ATP/50 μM EDTA 50 μM EDTA 
kass (M−1·s−11.5×106 0.56×106 0.76×106 
kdiss (s−10.77×10−3 200×10−3 340×10−3 
KD (nM) 0.5 360 450 
KD (nM) hyperbolic fit n.d. 380 550 
KD (nM) Scatchard plot n.d. 396 352 
 Interaction conditions 
Kinetic parameter 1 mM ATP/5 mM MgCl2 1 mM ATP/50 μM EDTA 50 μM EDTA 
kass (M−1·s−11.5×106 0.56×106 0.76×106 
kdiss (s−10.77×10−3 200×10−3 340×10−3 
KD (nM) 0.5 360 450 
KD (nM) hyperbolic fit n.d. 380 550 
KD (nM) Scatchard plot n.d. 396 352 

Omitting MgATP from the buffer yielded an approx. 3-fold slower association rate constant and a dissociation rate constant more than two orders of magnitude faster (Figure 2B and Table 1). Again, global fit analysis suggested a Langmuir 1:1 model, and the calculated KD of approx. 400 nM was almost three orders of magnitude higher than the value in the presence of MgATP. Replacing Mg2+ with 50 μM EDTA leaving ATP in the buffer or measurements performed just in the presence of 50 μM EDTA yielded comparable kinetics (Table 1).

These findings clearly indicate a crucial role for metal binding in forming a stable inhibitor complex with the pseudosubstrate inhibitor PKI, where the dissociation of the complex rather than the association is influenced by metal binding.

Since the dissociation rate constant was very fast in the absence of MgATP, equilibrium was reached quite rapidly. Therefore only a few data points could be used for the calculation of the association and dissociation rate constants respectively. To test data for self-consistency as recommended by Schuck and Minton [28], equilibrium binding analysis was performed by plotting the response signal at the end of the injection phase against the concentration of the analyte or by plotting the respective values, similar to a Scatchard-type analysis. Non-linear regression analysis using a single hyperbolic fit or linear regression performed for the Scatchard-type analysis yielded equilibrium binding constants of 380 nM and 396 nM for the interactions in the absence of MgATP respectively. This was in excellent agreement with the data derived from the kinetic evaluation (Table 1).

In order to confirm that the interaction was not influenced by steric hindrance due to the immobilization of GST–PKI on the sensor chip surface or by the GST-fusion tag itself a vice versa experiment was performed. Therefore recombinant Cα subunit was coupled on to a biosensor surface via primary amines as described in the Materials and methods section, and GST–PKI, as well as untagged PKI, were injected over the immobilized C subunit and the interaction kinetics were monitored. Again, the major effect of 5 mM MgCl2 and 1 mM ATP was to slow down the dissociation rate constant from 200×10−3 s−1 to 0.77×10−3 s−1. The association rate constant was accelerated from 0.56×106 M−1·s−1 to 1.5×106 M−1·s−1 (see also Table 1). No influence of the 26 kDa GST-fusion part on the kinetics was detected when using PKI as a ligand (results not shown). Although no difference in the association and in the initial dissociation phase could be seen, a second dissociation phase was detected when using GST–PKI in the flow phase. This could be due to the fact that GST can form a dimer, whereas PKI alone is running as a monomer based on gel-filtration experiments (F. W. Herberg, unpublished work). However, this effect was not detected on low-density surfaces of immobilized Cα subunit, indicating that at higher surface densities dimeric fusion proteins might cause more complex binding kinetics due to avidity effects.

Metal dependency of pseudosubstrate inhibitor binding

It has been shown previously that metal ions stabilize the C subunit against heat denaturation only at higher concentrations, probably owing to the occupancy of the second, low-affinity, metal-binding site [12]. Therefore the concentration dependency of MgCl2 in complex with 1 mM ATP on the Cα subunit–PKI interaction was investigated. Cα subunit, at a fixed concentration of 32 nM, was pre-incubated with various concentrations of MgCl2 in the presence of 1 mM ATP and was then injected over a GST–PKI surface as described above. Figure 3(A) shows the association and dissociation kinetics for MgCl2 concentrations varying from 10 μM to 10 mM (total concentration). Although variation of the Mg2+ ion concentration had almost no effect on the association rate constant, the dissociation rate constant was strongly affected. With increasing Mg2+ the fast dissociation phase transiently modulates into a slow dissociation phase. To investigate this phenomenon in more detail, the dissociation pattern of the C subunit from GST–PKI at numerous Mg2+ concentrations was measured (results not shown) and the dissociation rate constants for each Mg2+ concentration were determined directly from the kinetic data. When plotting the apparent kdiss determined at different Mg2+ concentrations against the total Mg2+ concentration (Figure 3B), a transition from fast dissociation kinetics (kdiss ranging from 8×10−2 to 2×10−2 s−1) to slower dissociation kinetics (kdiss ranging from 1.5×10−3 to 5×10−4 s−1) could be observed. Fast dissociation kinetics could be attributed to a total Mg2+ concentration in the lower micromolar range (EC50=26 μM), whereas slow dissociation kinetics could be correlated with a total Mg2+ concentration in the higher micromolar range (EC50=322 μM). Calculating the concentration of the free metal ions (Mg2+free) and the MgATP complex yielded an apparent EC50 of 2 μM and 5.3 μM for the fast dissociation kinetics and 36 μM and 150 μM for the slow dissociation kinetics respectively. This influence of the Mg2+ ion concentration on the dissociation rate constant of PKI could potentially be correlated with the occupation of the second metal-binding site in the C subunit.

Influence of the metal ion concentration on binding kinetics of the PKA C subunit to PKI

Figure 3
Influence of the metal ion concentration on binding kinetics of the PKA C subunit to PKI

(A) Interaction of 32 nM C subunit with captured GST–PKI in the presence of 1 mM ATP and various concentrations of MgCl2 (total concentrations, as shown). Association and dissociation were monitored for 240 s and 360 s respectively. The experiment was performed at 20 °C and at a flow rate of 25 μl/min. (B) Apparent dissociation rate constants for the dissociation of C subunit from GST–PKI were calculated at different MgCl2 concentrations using Biaevaluation 3.1 and plotted against the total MgCl2 concentration. The inset shows an enlargement of the plot for the higher MgCl2 concentrations. Non-linear regression of the data was performed with the software GraphPad Prism 4.0 using a sigmoidal dose–response equation.

Figure 3
Influence of the metal ion concentration on binding kinetics of the PKA C subunit to PKI

(A) Interaction of 32 nM C subunit with captured GST–PKI in the presence of 1 mM ATP and various concentrations of MgCl2 (total concentrations, as shown). Association and dissociation were monitored for 240 s and 360 s respectively. The experiment was performed at 20 °C and at a flow rate of 25 μl/min. (B) Apparent dissociation rate constants for the dissociation of C subunit from GST–PKI were calculated at different MgCl2 concentrations using Biaevaluation 3.1 and plotted against the total MgCl2 concentration. The inset shows an enlargement of the plot for the higher MgCl2 concentrations. Non-linear regression of the data was performed with the software GraphPad Prism 4.0 using a sigmoidal dose–response equation.

To investigate whether ATP alone, Mg2+ ions alone or the metal–nucleotide complex were responsible for the effect on the dissociation phase of the Cα subunit from GST–PKI, different dissociation conditions were tested. The association in the absence of Mg2+ ions but in the presence of 1 mM ATP was followed by dissociation in the presence of EDTA, ATP alone, Mg2+ alone or Mg2+–ATP. The same experiment was performed but this time with an association phase in the presence of 1 mM ATP and 10 mM Mg2+ ions, depicted in Figure 4. As shown on both plots, only the Mg2+–ATP complex, but not ATP alone, is responsible for the slow dissociation rate. No difference could be detected between ATP alone and EDTA; therefore the nucleotide alone seemed to have a negligible effect on the dissociation rate. However, adding Mg2+ ions during the dissociation phase dramatically decreased the dissociation rate, suggesting that Mg2+ is stabilizing the C subunit–PKI complex.

Effects of Mg2+ and ATP on association and dissociation patterns of C subunit binding to PKI

Figure 4
Effects of Mg2+ and ATP on association and dissociation patterns of C subunit binding to PKI

Interaction of 125 nM of PKA C subunit with immobilized GST–PKI under different conditions for association and dissociation. (A) shows the association of Cα subunit to GST–PKI in the presence of 1 mM ATP and the dissociation in the presence of Mg2+, Mg2+–ATP, ATP or EDTA respectively (as indicated in the graph). (B) shows the association of Cα subunit to GST–PKI in the presence of 1 mM ATP and 10 mM Mg2+ and the dissociation in the presence of Mg2+, MgATP, ATP and EDTA respectively (as indicated in the graph).

Figure 4
Effects of Mg2+ and ATP on association and dissociation patterns of C subunit binding to PKI

Interaction of 125 nM of PKA C subunit with immobilized GST–PKI under different conditions for association and dissociation. (A) shows the association of Cα subunit to GST–PKI in the presence of 1 mM ATP and the dissociation in the presence of Mg2+, Mg2+–ATP, ATP or EDTA respectively (as indicated in the graph). (B) shows the association of Cα subunit to GST–PKI in the presence of 1 mM ATP and 10 mM Mg2+ and the dissociation in the presence of Mg2+, MgATP, ATP and EDTA respectively (as indicated in the graph).

In the case of the type I R subunit binding to PKA Cα subunit, the effect of various concentrations of Mg2+ on the association and dissociation rate constants was tested by injecting a concentration series ranging from 6.25 nM to 50 nM of RIα 92–260 over immobilized C subunit in the presence of 1 mM ATP and various concentrations of Mg2+ (ranging from 10 μM to 10 mM). In the absence of Mg2+, no binding of RIα 92–260 to Cα subunit could be observed (see Figure 1). Apparent rate constants as well as equilibrium binding constants were calculated using a Langmuir 1:1 binding model. The resulting values are shown in Table 2. Figure 5 shows the interaction of 25 nM RIα 92–260 with Cα subunit under the influence of various concentrations of Mg2+ in the presence of 1 mM ATP. Interestingly, as seen from the plot and Table 2, compared with the C subunit–PKI interaction, the dissociation rate constant is less affected by variation of metal ion concentration. In contrast, the association rate constant is increased by two orders of magnitude with increasing metal ion concentration. This results in a shift in apparent KD of two orders of magnitude from 130 nM at 10 μM Mg2+ to 1.9 nM at 10 mM Mg2+. Thus occupation of the second metal binding site in PKA Cα subunit has only minor effects on the dissociation of the holoenzyme complex, but strongly influences the association of the type I holoenzyme complex.

Table 2
Influence of Mg2+ concentration on association and dissociation rate constants of the interaction of the PKA C subunit with RIα 92–260

Apparent association and dissociation rate constants and calculated KD values of the interaction of bovine regulatory subunit RIα 92–260 with the immobilized C subunit of PKA (No bdg., could not be determined because of no significant binding). The interaction was measured in the presence of 1 mM ATP and MgCl2 concentrations as indicated in the Table.

 [MgCl2] (mM) 
Kinetic parameter 0.01 0.05 0.1 0.2 0.5 10 
kass (M−1·s−1No bdg. 2.1×104 1.1×105 2.3×105 7.3×105 1.3×106 2.3×106 3.3×106 
kdiss (s−1No bdg. 2.7×10−3 3.3×10−3 3.6×10−3 4.8×10−3 4.8×10−3 5.5×10−3 6.2×10−3 
KD (nM) No bdg. 130 30 16 6.5 3.7 2.4 1.9 
 [MgCl2] (mM) 
Kinetic parameter 0.01 0.05 0.1 0.2 0.5 10 
kass (M−1·s−1No bdg. 2.1×104 1.1×105 2.3×105 7.3×105 1.3×106 2.3×106 3.3×106 
kdiss (s−1No bdg. 2.7×10−3 3.3×10−3 3.6×10−3 4.8×10−3 4.8×10−3 5.5×10−3 6.2×10−3 
KD (nM) No bdg. 130 30 16 6.5 3.7 2.4 1.9 

Influence of metal ion concentration on PKA type I holoenzyme formation

Figure 5
Influence of metal ion concentration on PKA type I holoenzyme formation

Interaction of 25 nM bovine RIα 92–260 with immobilized C subunit in the presence of 1 mM ATP and various (total) concentrations of MgCl2 as indicated on the plot. Association and dissociation were monitored for 240 s each at 20 °C with a flow rate of 25 μl/min. The resulting equilibrium binding constants and association and dissociation rate constants are shown in Table 2.

Figure 5
Influence of metal ion concentration on PKA type I holoenzyme formation

Interaction of 25 nM bovine RIα 92–260 with immobilized C subunit in the presence of 1 mM ATP and various (total) concentrations of MgCl2 as indicated on the plot. Association and dissociation were monitored for 240 s each at 20 °C with a flow rate of 25 μl/min. The resulting equilibrium binding constants and association and dissociation rate constants are shown in Table 2.

DISCUSSION

SPR measurements

The kinetic properties and dynamics of the PKA C subunit have been studied intensively using various activity-based or direct binding assays [11,1315,17]. However, activity assays, either radioactive, spectrophotometric or fluorescence-based, always require Mg2+ and ATP as a co-substrate. Therefore the direct interaction of the C subunit with its physiological inhibitors cannot be determined in the absence of MgATP following the loss of catalytic activity. Complex formation can be studied independently of activity, for example by analytical gel filtration. However, the method is somewhat inaccurate, since in small zone experiments dilution effects influence the apparent equilibrium binding constants [17]. Biosensors based on SPR allow a direct observation of detailed association and dissociation kinetics in real time, yielding not only equilibrium binding constants but also apparent association and dissociation rate constants. SPR has been previously used as a method of choice to investigate PKA interactions with its inhibitors [16,27,29]. We used a Biacore system to study in depth the binding kinetics of the PKA C subunit to its physiological pseudosubstrate inhibitors, namely PKI and the type I R subunit. The influence of both metal ions and a variety of nucleotides on rate and equilibrium binding constants was investigated, which in turn allowed us to study the effect of metal binding and nucleotide interaction separately for C subunit–inhibitor complex formation and dissociation (Tables 1 and 2).

PKI is a monomer (molecular mass of 7.8 kDa), binding with sub-nanomolar affinity to the PKA C subunit [18]. The interaction with the C subunit [27] can easily be analysed using SPR with PKI immobilized to the sensor chip, whereas the analysis of the R subunit dimer is quite difficult, since multiple binding events overlay during analysis. Thus we used a truncated construct of bovine regulatory subunit type I (RIα 92–260) which consists of the autoinhibitor site (pseudosubstrate sequence) and one cAMP-binding domain (A site) only [21]. The construct cannot form a dimer, because the dimerization and docking domain (first 91 amino acids) is deleted, but it is still capable of forming a stable holoenzyme complex with the C subunit. The crystal structure of a similar deletion construct (RIα 91–244) in complex with the C subunit also demonstrates a stable complex formation [30]. The binding kinetics of the truncated construct are much less complex than those of the full-length wild-type protein with its two binding sites for the C subunit. All measurements performed in the present study displayed an extremely high accuracy with the S.E.M. within the same experiment being less than 1%. The interaction with both inhibitors could be fitted perfectly to a Langmuir 1:1 model, suggesting a true 1:1 interaction with a single binding site. In contrast, the full-length R subunit protein shows various effects such as biphasic binding, not reaching equilibrium, which do not fit well to any binding model (C. Hahnefeld and F. W. Herberg, unpublished work).

Metal and nucleotide binding in PKA

High-resolution crystal structures of the catalytic subunit of PKA can provide a model for understanding the metal-binding sites of protein kinases. Originally, crystals of the mouse recombinant enzyme containing ATP and the inhibitor peptide PKI(5–24), forming the ternary complex, were grown under conditions where nucleotide was in excess of Mg2+ so that only a single metal-binding site was occupied. By soaking the crystals in Mn2+, two metal-binding sites could be unambiguously identified [8]. A similar approach was used to define the metal-binding sites in crystals of the porcine C subunit crystallized in the presence of inhibitor peptide and p[NH]ppA [7].

The metal requirements for the PKA C subunit were first described in detail by Armstrong et al. [31], who showed that from two Mg2+-binding sites, one is required for catalysis and one is associated with inhibition. Based on crystal structures of the C subunit [7,8] and NMR studies [14] with inert complexes of ATP it was concluded that the activating metal bridges the β- and γ-phosphates and involves an enzyme–ATP–metal bridge [14]. On the other hand, the second, inhibitory, metal site was predicted to be an enzyme–metal–ATP bridge between the α- and γ-phosphate [15]. Furthermore, it could be shown that the occupation of the second metal-binding site is crucial for stabilizing the C subunit against thermal denaturation [12].

From the results of the present study, we conclude that a stable complex formation of the PKA C subunit with the pseudosubstrate inhibitors PKI and type I R subunit can only occur in the presence of both ATP and Mg2+. Thus for high-affinity binding of this type of inhibitors the second, low-affinity, metal-binding site in the C subunit has to be occupied. For the interaction with PKI it could be demonstrated that, particularly the dissociation rate, was strongly influenced by the presence of Mg2+ ions, which allowed dissection of the two metal-binding sites, when plotting the dissociation rate constants against the metal concentration (Figure 3). Interestingly, in contrast with the C–PKI complex the concentration of Mg2+ has almost no effect on the dissociation of the type I holoenzyme complex. However, it has a strong effect on the association of RIα 92–260 to the PKA Cα subunit (Table 2). From previous studies, it is known that the Mg2+–ATP complex itself has a faster off-rate from the C–PKI complex than from type I holoenzyme, i.e. 17 min and 11.7 h respectively [17]. This is reflected in the fact that full-length RIα forms a more stable holoenzyme complex (KD 0.2 nM) than the C–PKI complex [17,32]. The truncated construct of the R subunit used in the present study with an apparent KD of 1.9 nM (Table 2) lacks additional interaction sites with the C subunit present in the full-length RI subunit and therefore displays significantly reduced affinity. For the full-length R subunit the dissociation rate constant from the C subunit in the presence of high concentrations of Mg2+ and ATP is three times slower compared with the dissociation of PKI [27]. This may underline the physiological role of PKI, acting as a nuclear scavenger for the C subunit [4]. Free PKA C subunit can enter the nucleus to phosphorylate its substrate proteins. The activity of PKA C subunit in the nucleus cannot be modulated through the R subunits because either these inhibitors are kept in the cytoplasm or are anchored via AKAPs (a kinase anchoring proteins) [33]. However, PKI is able to enter the nucleus and inactivates the catalytic subunit upon binding. After complex formation the nuclear export signal of PKI ensures the transport of the C–PKI complex to the cytoplasm [4]. In the cytoplasm, the R subunits may effectively compete for the catalytic subunit when it is bound to PKI, as suggested by in vitro competition experiments [17]. At a given cellular Mg2+ concentration the association rate of the type I R subunit is very fast, which in turn promotes the holoenzyme formation, thereby recapturing the C subunit released from PKI. The present study and previous work [4,16,17] established that the spatial and temporal regulation of PKA catalytic activity is governed by a variety of binding events in dynamic equlibria. Minute differences in apparent rate constants of the interaction with regulatory proteins and cofactors define the composition of functional complexes in subcellular compartments. Although the total cellular Mg2+ is well in the millimolar range (14–20 mM, [34]), the free intracellular Mg2+, determined in a variety of tissues or cells, is between 0.4 and 3 mM according to several reports reviewed in [35,36]. However, concentrations as low as 0.2 mM free Mg2+ have to be considered pathological [37]. Total ATP concentrations are generally >3 mM and free ATP concentrations have been reported at approx. 1 mM, whereas kinases usually have a Km for ATP which is below 0.5 mM [36]. The values for free Mg2+ and free ATP are well above the critical concentration for the PKA C subunit forming a high-affinity complex with a pseudosubstrate inhibitor. The intracellular levels of metal ions favour the occupation of the second metal-binding site in the PKA C subunit, and besides lead to fast holoenzyme formation, presumably locking the enzyme in a closed conformation.

Open and closed conformation

Recently solved crystal structures of a complex of the PKA C subunit with type I R subunits (RIα 91–244 and RIα 91–379:R333K) suggest that the C subunit is adopting a closed conformation in the type I holoenzyme complex [30,38]. Both crystal structures clearly show the occupation of the nucleotide-binding site of the C subunit by p[NH]ppA and two metal ions (Mn2+). Compared with the crystal structure of the ternary complex of the C subunit with PKI, also containing p[NH]ppA and two metal ions (Mn2+) [7], one partly overlapping the contact surface on the large lobe of the C subunit can be identified for binding the RI subunit and PKI to the active site cleft. However, additional, but distinct, extended binding surfaces exist for each inhibitor. This is depicted in Figure 6(A), a structural alignment of the two crystal structures of the C subunit with its pseudosubstrate inhibitors. In both complexes the C subunit adopts an identical, fully closed conformation with both metal ion-binding sites occupied. The autoinhibitor sites of the R subunit and PKI perfectly align in both crystal structures. The opening and closing of the active site cleft in the C subunit, depending on complex formation with inhibitor and/or cofactor, can be observed in several crystal structures (reviewed in [9]). This is reflected by the positioning and distance of the glycine-rich loop relative to the catalytic loop [16] and can be followed from a fully closed conformation over intermediate states to an open conformation in the apoenzyme [9]. In Figure 6(B) the difference between the open conformation in the apoenzyme and the fully closed conformation in the C–PKI–Mn2+–p[NH]ppA complex is clearly visible, when comparing the position of the glycine-rich loop in the open conformation (yellow) with the closed conformation (red). No crystal structure of a C subunit inhibitor complex is available to date where only one metal-binding site is occupied, from which an intermediate state could be assumed. Interestingly, another recently solved crystal structure of the type II holoenzyme (Cα subunit with RIIα 90–400) shows the C subunit in an open conformation when crystallized without metal ions and nucleotide [39], thus supporting our model in the present study.

Crystal structures of the C subunit pseudosubstrate inhibitor complexes

Figure 6
Crystal structures of the C subunit pseudosubstrate inhibitor complexes

(A) Structural alignment of the C subunit pseudosubstrate inhibitor complexes with PKI (PDB ID: 1CDK, orange) and type Iα regulatory subunit (91–244, PDB ID: 1U7E, red). The C subunit assumes a fully closed conformation in both crystal structures (tan and blue). p[NH]ppA and two Mn2+ (rendered as green spheres) are bound in the active-site cleft. The Figure was created using VMD 1.8.5 standard settings. (B) Representation of the open and closed conformation of the PKA C subunit reflected by the distance between the catalytic loop (tan) and the glycine-rich loop in the apoenzyme (PDB ID: 1J3H, yellow) and the PKI inhibitor complex (PDB ID: 1CDK, red). The nucleotide is shown as black sticks and the two metal ions (Me) are rendered as grey spheres. The Figure was created using VMD 1.8.5 standard settings.

Figure 6
Crystal structures of the C subunit pseudosubstrate inhibitor complexes

(A) Structural alignment of the C subunit pseudosubstrate inhibitor complexes with PKI (PDB ID: 1CDK, orange) and type Iα regulatory subunit (91–244, PDB ID: 1U7E, red). The C subunit assumes a fully closed conformation in both crystal structures (tan and blue). p[NH]ppA and two Mn2+ (rendered as green spheres) are bound in the active-site cleft. The Figure was created using VMD 1.8.5 standard settings. (B) Representation of the open and closed conformation of the PKA C subunit reflected by the distance between the catalytic loop (tan) and the glycine-rich loop in the apoenzyme (PDB ID: 1J3H, yellow) and the PKI inhibitor complex (PDB ID: 1CDK, red). The nucleotide is shown as black sticks and the two metal ions (Me) are rendered as grey spheres. The Figure was created using VMD 1.8.5 standard settings.

Based on the results of the present study we propose a model where complex formation of the C subunit with either the RI subunit or PKI is dependent on the presence of high metal concentrations and therefore the occupation of both metal-ion-binding sites. Only with both metal ion-binding sites occupied, the C subunit can adopt a fully closed conformation, allowing high-affinity complex formation with pseudosubstrate inhibitors. The results of the present study and early work by Armstrong et al. [14] suggest that the first metal ion is associated with catalysis, with the opening of the active-site cleft and probably with complex formation of the C subunit with R subunit type II. Occupation of the second metal-binding site induces the closed conformation which provides the scaffold for binding the RI subunit and PKI (Figure 6A). Binding of those physiological pseudosubstrate inhibitors locks the metal–nucleotide complex in the active site cleft. Taken together the second metal-binding site represents a gatekeeper which, when occupied, promotes the C subunit in a closed, inactive conformation and thereby allows stable high-affinity binding to its physiological pseudosubstrate inhibitors type I R subunit and PKI. The model in the present study describes an extremely efficient mechanism to tightly regulate catalytic activity of this ubiquitously expressed protein kinase.

We thank Susan Taylor and Dirk Bossemeyer for helpful discussion and Frank Gesellchen and Laura De Francesco for carefully reading the manuscript. This work was supported by grants from the DFG (Deutsche Forschungsgemeinschaft, He 1818/4), the Nationales Genomforschungsnetz (NGFN2, 01GR0441) to F. W. H. and the EU (thera-cAMP, 037189) to F. W. H. and B. Z. S. S. was supported by the graduate program of the University of Kassel. The group of F. W. H. is a member in the EU FP6 ProteomeBinders consortium.

Abbreviations

     
  • C subunit

    catalytic subunit

  •  
  • EDC

    N-ethyl-N′-(dimethylaminopropyl)-carbodi-imide

  •  
  • GST

    glutathione transferase

  •  
  • NHS

    N-hydroxysuccinimide

  •  
  • SPR

    surface plasmon resonance

  •  
  • PKA

    protein kinase A

  •  
  • PKI

    heat-stable protein kinase inhibitor

  •  
  • p[NH]ppA

    adenosine 5′-[β,γ-imido]triphosphate

  •  
  • R subunit

    regulatory subunit

  •  
  • RU

    resonance unit

References

References
1
Hanks
S. K.
Quinn
A. M.
Hunter
T.
The protein kinase family: conserved features and deduced phylogeny of the catalytic domains
Science
1988
, vol. 
241
 (pg. 
42
-
52
)
2
Knighton
D. R.
Zheng
J. H.
Ten Eyck
L. F.
Ashford
V. A.
Xuong
N. H.
Taylor
S. S.
Sowadski
J. M.
Crystal structure of the catalytic subunit of cyclic adenosine monophosphate-dependent protein kinase
Science
1991
, vol. 
253
 (pg. 
407
-
414
)
3
Walsh
D. A.
Ashby
C. D.
Gonzalez
C.
Calkins
D.
Fischer
E. H.
Krebs
E. G.
Purification and characterization of a protein inhibitor of adenosine 3′,5′-monophosphate-dependent protein kinases
J. Biol. Chem.
1971
, vol. 
246
 (pg. 
1977
-
1985
)
4
Wen
W.
Meinkoth
J. L.
Tsien
R. Y.
Taylor
S. S.
Identification of a signal for rapid export of proteins from the nucleus
Cell
1995
, vol. 
82
 (pg. 
463
-
473
)
5
de Lecea
L.
Criado
J. R.
Rivera
S.
Wen
W.
Soriano
E.
Henriksen
S. J.
Taylor
S. S.
Gall
C. M.
Sutcliffe
J. G.
Endogenous protein kinase A inhibitor (PKIα) modulates synaptic activity
J. Neurosci. Res.
1998
, vol. 
53
 (pg. 
269
-
278
)
6
Narayana
N.
Cox
S.
Nguyen-huu
X.
Ten Eyck
L. F.
Taylor
S. S.
A binary complex of the catalytic subunit of cAMP-dependent protein kinase and adenosine further defines conformational flexibility
Structure
1997
, vol. 
5
 (pg. 
921
-
935
)
7
Bossemeyer
D.
Engh
R. A.
Kinzel
V.
Ponstingl
H.
Huber
R.
Phosphotransferase and substrate binding mechanism of the cAMP-dependent protein kinase catalytic subunit from porcine heart as deduced from the 2.0 Å structure of the complex with Mn2+ adenylyl imidodiphosphate and inhibitor peptide PKI(5–24)
EMBO J.
1993
, vol. 
12
 (pg. 
849
-
859
)
8
Zheng
J.
Knighton
D. R.
Ten Eyck
L. F.
Karlsson
R.
Xuong
N.
Taylor
S. S.
Sowadski
J. M.
Crystal structure of the catalytic subunit of cAMP-dependent protein kinase complexed with MgATP and peptide inhibitor
Biochemistry
1993
, vol. 
32
 (pg. 
2154
-
2161
)
9
Taylor
S. S.
Yang
J.
Wu
J.
Haste
N. M.
Radzio-Andzelm
E.
Anand
G.
PKA: a portrait of protein kinase dynamics
Biochim. Biophys. Acta
2004
, vol. 
1697
 (pg. 
259
-
269
)
10
Swarup
G.
Dasgupta
J. D.
Garbers
D. L.
Tyrosine-specific protein kinases of normal tissues
Adv. Enzyme Regul.
1984
, vol. 
22
 (pg. 
267
-
288
)
11
Cook
P. F.
Neville
M. E.
Jr
Vrana
K. E.
Hartl
F. T.
Roskoski
R.
Jr
Adenosine cyclic 3′,5′-monophosphate dependent protein kinase: kinetic mechanism for the bovine skeletal muscle catalytic subunit
Biochemistry
1982
, vol. 
21
 (pg. 
5794
-
5799
)
12
Herberg
F. W.
Doyle
M. L.
Cox
S.
Taylor
S. S.
Dissection of the nucleotide and metal-phosphate binding sites in cAMP-dependent protein kinase
Biochemistry
1999
, vol. 
38
 (pg. 
6352
-
6360
)
13
Bhatnagar
D.
Roskoski
R.
Jr
Rosendahl
M. S.
Leonard
N. J.
Adenosine cyclic 3′,5′-monophosphate dependent protein kinase: a new fluorescence displacement titration technique for characterizing the nucleotide binding site on the catalytic subunit
Biochemistry
1983
, vol. 
22
 (pg. 
6310
-
6317
)
14
Armstrong
R. N.
Kondo
H.
Granot
J.
Kaiser
E. T.
Mildvan
A. S.
Magnetic resonance and kinetic studies of the manganese(II) ion and substrate complexes of the catalytic subunit of adenosine 3′,5′-monophosphate dependent protein kinase from bovine heart
Biochemistry
1979
, vol. 
18
 (pg. 
1230
-
1238
)
15
Granot
J.
Kondo
H.
Armstrong
R. N.
Mildvan
A. S.
Kaiser
E. T.
Nuclear magnetic resonance studies of the conformation of tetraamminecobalt (III)-ATP bound at the active site of bovine heart protein kinase
Biochemistry
1979
, vol. 
18
 (pg. 
2339
-
2345
)
16
Diskar
M.
Zenn
H. M.
Kaupisch
A.
Prinz
A.
Herberg
F. W.
Molecular basis for isoform-specific autoregulation of protein kinase A
Cell. Signalling
2007
, vol. 
19
 (pg. 
2024
-
2034
)
17
Herberg
F. W.
Taylor
S. S.
Physiological inhibitors of the catalytic subunit of cAMP-dependent protein kinase: effect of MgATP on protein–protein interactions
Biochemistry
1993
, vol. 
32
 (pg. 
14015
-
14022
)
18
Whitehouse
S.
Walsh
D. A.
Mg X ATP2-dependent interaction of the inhibitor protein of the cAMP-dependent protein kinase with the catalytic subunit
J. Biol. Chem.
1983
, vol. 
258
 (pg. 
3682
-
3692
)
19
Lew
J.
Coruh
N.
Tsigelny
I.
Garrod
S.
Taylor
S. S.
Synergistic binding of nucleotides and inhibitors to cAMP-dependent protein kinase examined by acrylodan fluorescence spectroscopy
J. Biol. Chem.
1997
, vol. 
272
 (pg. 
1507
-
1513
)
20
Slice
L. W.
Taylor
S. S.
Expression of the catalytic subunit of cAMP-dependent protein kinase in Escherichia coli
J. Biol. Chem.
1989
, vol. 
264
 (pg. 
20940
-
20946
)
21
Hahnefeld
C.
Moll
D.
Goette
M.
Herberg
F. W.
Rearrangements in a hydrophobic core region mediate cAMP action in the regulatory subunit of PKA
Biol. Chem.
2005
, vol. 
386
 (pg. 
623
-
631
)
22
Thomas
J.
Van Patten
S. M.
Howard
P.
Day
K. H.
Mitchell
R. D.
Sosnick
T.
Trewhella
J.
Walsh
D. A.
Maurer
R. A.
Expression in Escherichia coli and characterization of the heat-stable inhibitor of the cAMP-dependent protein kinase
J. Biol. Chem.
1991
, vol. 
266
 (pg. 
10906
-
10911
)
23
Laemmli
U. K.
Cleavage of structural proteins during the assembly of the head of bacteriophage T4
Nature
1970
, vol. 
227
 (pg. 
680
-
685
)
24
Bradford
M. M.
A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding
Anal. Biochem.
1976
, vol. 
72
 (pg. 
248
-
254
)
25
Brooks
S. P.
Storey
K. B.
Bound and determined: a computer program for making buffers of defined ion concentrations
Anal. Biochem.
1992
, vol. 
201
 (pg. 
119
-
126
)
26
Stenberg
E.
Persson
B.
Roos
H.
Urbaniczky
C.
Quantitative determination of surface concentration of protein with surface plasmon resonance using radiolabeled proteins
J. Colloid Interface Sci.
1991
, vol. 
143
 (pg. 
513
-
526
)
27
Zimmermann
B.
Chiorini
J. A.
Ma
Y.
Kotin
R. M.
Herberg
F. W.
PrKX is a novel catalytic subunit of the cAMP-dependent protein kinase regulated by the regulatory subunit type I
J. Biol. Chem.
1999
, vol. 
274
 (pg. 
5370
-
5378
)
28
Schuck
P.
Minton
A. P.
Kinetic analysis of biosensor data: elementary tests for self-consistency
Trends Biochem. Sci.
1996
, vol. 
21
 (pg. 
458
-
460
)
29
Viht
K.
Schweinsberg
S.
Lust
M.
Vaasa
A.
Raidaru
G.
Lavogina
D.
Uri
A.
Herberg
F. W.
Surface-plasmon-resonance-based biosensor with immobilized bisubstrate analog inhibitor for the determination of affinities of ATP- and protein-competitive ligands of cAMP-dependent protein kinase
Anal. Biochem.
2007
, vol. 
362
 (pg. 
268
-
277
)
30
Kim
C.
Xuong
N. H.
Taylor
S. S.
Crystal structure of a complex between the catalytic and regulatory (RIα) subunits of PKA
Science
2005
, vol. 
307
 (pg. 
690
-
696
)
31
Armstrong
R. N.
Kondo
H.
Kaiser
E. T.
Cyclic AMP-dependent ATPase activity of bovine heart protein kinase
Proc. Natl. Acad. Sci. U.S.A.
1979
, vol. 
76
 (pg. 
722
-
725
)
32
Hofmann
F.
Apparent constants for the interaction of regulatory and catalytic subunit of cAMP-dependent protein kinase I and II
J. Biol. Chem.
1980
, vol. 
255
 (pg. 
1559
-
1564
)
33
Tasken
K.
Aandahl
E. M.
Localized effects of cAMP mediated by distinct routes of protein kinase A
Physiol. Rev.
2004
, vol. 
84
 (pg. 
137
-
167
)
34
Schmitz
C.
Deason
F.
Perraud
A. L.
Molecular components of vertebrate Mg2+-homeostasis regulation
Magnes. Res.
2007
, vol. 
20
 (pg. 
6
-
18
)
35
Romani
A.
Scarpa
A.
Regulation of cell magnesium
Arch. Biochem. Biophys.
1992
, vol. 
298
 (pg. 
1
-
12
)
36
Traut
T. W.
Physiological concentrations of purines and pyrimidines
Mol. Cell. Biochem.
1994
, vol. 
140
 (pg. 
1
-
22
)
37
Schlingmann
K. P.
Gudermann
T.
A critical role of TRPM channel-kinase for human magnesium transport
J. Physiol.
2005
, vol. 
566
 (pg. 
301
-
308
)
38
Kim
C.
Cheng
C. Y.
Saldanha
S. A.
Taylor
S. S.
PKA-I holoenzyme structure reveals a mechanism for cAMP-dependent activation
Cell
2007
, vol. 
130
 (pg. 
1032
-
1043
)
39
Wu
J.
Brown
S. H.
von Daake
S.
Taylor
S. S.
PKA type IIα holoenzyme reveals a combinatorial strategy for isoform diversity
Science
2007
, vol. 
318
 (pg. 
274
-
279
)

Author notes

1

These authors contributed equally to this work.