cGMP Binding Domain D Mediates a Unique Activation Mechanism PDF

Article Cite This: ACS Infect. Dis. 2018, 4, 415−423...

Article Cite This: ACS Infect. Dis. 2018, 4, 415−423 cGMP Binding Domain D Mediates a Unique Activation Mechanism in Plasmodium falciparum PKG Eugen Franz, Matthias J. Knape, and Friedrich W. Herberg* Department of Biochemistry, University of Kassel, Heinrich-Plett-Strasse 40, 34132 Kassel, Germany * S Supporting Information ABSTRACT: cGMP-dependent protein kinase from Plasmodium falciparum (Pf PKG) plays a crucial role in the sexual as well as the See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles. asexual proliferation of this human malaria causing parasite. However, function and regulation of Pf PKG are largely unknown. Previous studies showed that the domain organization of Pf PKG Downloaded via UNIV MONTPELLIER on February 2, 2022 at 08:59:03 (UTC). significantly differs from human PKG (hPKG) and indicated a critical role of the cyclic nucleotide binding domain D (CNB-D). We identified a novel mechanism, where the CNB-D controls activation and regulation of the parasite specific protein kinase. Here, kinase activity is not dependent on a pseudosubstrate autoinhibitory sequence (IS), as reported for human PKG. A construct lacking the putative IS and containing only the CNB-D and the catalytic domain is inactive in the absence of cGMP and can efficiently be activated with cGMP. On the basis of structural evidence, we describe a regulatory mechanism, whereby cGMP binding to CNB-D induces a conformational change involving the αC-helix of the CNB-D. The inactive state is defined by a unique interaction between Asp597 of the catalytic domain and Arg528 of the αC-helix. The same arginine (R528), however, stabilizes cGMP binding by interacting with Tyr480 of the phosphate binding cassette (PBC). This represents the active state of Pf PKG. Our results unveil fundamental differences in the activation mechanism between Pf PKG and hPKG, building the basis for the development of strategies for targeted drug design in fighting malaria. KEYWORDS: malaria, Plasmodium falciparum, Pf PKG, cGMP-dependent protein kinase, cGMP, cyclic nucleotides W ith approximately 214 million cases per year, malaria is still the most pathogen-mediated infectious disease worldwide. In 2015, the number of malaria deaths reached differs significantly from their mammalian counterpart.5 cGMP signaling was shown to have a pivotal role in exflagellation and gametogenesis in the asexual blood stage and for schizogony in about 438 000.1 The disease is caused by protozoans of the liver stage development.9−11 In contrast, cAMP signaling genus Plasmodium, which are transmitted by the female controls apical exocytosis, sporozoite motility, anion transport Anopheles mosquito; thus, the distribution of malaria is directly through the infected red blood cell (RBC) membrane, and related to the living area of the mosquito. Malaria tropica is the invasion of liver cells and RBCs.12−15 most dangerous variant of malaria and is caused by Plasmodium The main effector of cGMP in Plasmodium is Pf PKG.16 The falciparum.2 For the treatment of malaria, various drugs such as P. falciparum genome encodes a single PKG isoform, which chloroquine, atovaquone, proguanil, or artemisin are currently consists of an N-terminal regulatory domain and a C-terminal used. Treatment of malaria has been hampered severely by catalytic domain. The regulatory domain of Pf PKG comprises drug resistance, making the development of next generation four cGMP binding domains (CNB-A/B/C/D) with a antimalarial drugs indispensable.3 degenerate CNB-C. With two cGMP binding domains P. falciparum exhibits a complex life cycle advancing through (CNB-A and -B), the regulatory domain of mammalian multiple stages in both its human and mosquito hosts. This PKGs differs significantly from that of Pf PKG. Moreover, involves tightly controlled intracellular processes including with 97.5 kDa, Pf PKG is larger than its mammalian orthologue cyclic nucleotide signaling and a variety of protein kinases.4 and is missing a dimerization domain.7,17,18 A recent The genome of P. falciparum encodes about 90 protein kinases phosphoproteomics study revealed 107 phosphorylation sites including the CMGC, NIMA, calcium-dependent, PIK, and on 69 proteins as direct or indirect cellular targets of Pf PKG, AGC kinase family.5,6 Among those, cAMP-dependent protein including proteins involved in cell signaling, gene regulation, kinase (PKA) and cGMP-dependent protein kinase (PKG) are proteolysis, and protein transport.19 the best characterized protein kinases biochemically and Studies on the activation mechanism of human PKGs (type I structurally. The P. falciparum homologue of PKA (Pf PKA) as well as type II isoforms) demonstrated that hPKG is was identified in the year 2000 and the PKG homologue (Pf PKG) only two years later.7,8 Due to early evolutionary Received: November 8, 2017 branching, cyclic nucleotide signaling in Plasmodium parasites Published: December 18, 2017 © 2017 American Chemical Society 415 DOI: 10.1021/acsinfecdis.7b00222 ACS Infect. Dis. 2018, 4, 415−423 ACS Infectious Diseases Article Figure 1. Domain organization and cGMP-dependent activation of Pf PKG. (A) Schematic representation of the domain organization in Pf PKG. Pf PKG contains four cyclic nucleotide binding domains (CNB-A, -B, -C, and -D) and a putative N-terminal autoinhibitory sequence (IS). The conserved arginine residues (R132 in CNB-A, R250 in CNB-B, and R492 in CNB-D) associated with high-affinity cGMP binding are highlighted. (B) Specific activity of Pf PKG deletion constructs (519-853: catalytic domain with αC-helix of CNB-D; 401-853: catalytic domain with CNB-D; 275-853: catalytic domain with CNB-C/D; 158-853: catalytic domain with CNB-B/C/D) as determined by the spectrophotometric assay. A final concentration of 200−300 nM of the respective Pf PKG construct was used. Bar diagram depicting basal activity (-cGMP) in black and specific activity (+cGMP, 100 μM) in white. To test the effect of the autoinhibitory sequence (IS), preactivated Pf PKG 401-853 and Pf PKG 158-853 (with 100 μM cGMP) were incubated with 500 μM IS peptide, respectively (gray). (C) cGMP activation by microfluidic mobility-shift assay (MSA) for truncated (401-853: catalytic domain with CNB-D; 275-853: catalytic domain with CNB-C/D; 158-853: catalytic domain with CNB-B/C/D; 32- 853: without putative IS) and full length Pf PKG (1-853). A final concentration of 4−10 nM of the respective Pf PKG construct was used. Activation constant (Kact) is defined as the cGMP concentration where half-maximal kinase activity is observed. Kinase activity corresponds to substrate (PKStide, 1 mM) conversion. Normalized data were fitted to a sigmoidal dose−response plot. (D) Hill slopes as determined from the activation data using GraphPad Prism 6. Hill slope change from 0.7 to 1.6 with CNB-B and CNB-A showing positive cooperativity. Standard errors are specified as SD (standard deviation) of n ≥ 4 experiments. regulated by autoinhibition via an N-terminal pseudosubstrate inhibitory sequence (IS).20−22 According to this model, the N- RESULTS The detailed molecular basis of how cGMP binding controls terminal IS binds to the C-terminal catalytic domain, rendering catalytic activity of Pf PKG is not known so far. In mammals, the inactive state of hPKG. Binding of cGMP to both CNBs both PKA and PKG are regulated by a pseudosubstrate induces conformational changes, which displace the IS from inhibitory sequence (IS),20 either on a separate regulatory the catalytic domain.23−25 On the basis of sequence alignments subunit as in PKA or on the same polypeptide chain (PKG). However, for Pf PKG, biochemical evidence for the auto- of the apicomplexan PKGs and their mammalian orthologues, inhibitory function of an IS is missing so far. an N-terminal pseudosubstrate inhibitory sequence was Role of the Inhibitory Sequence in Regulation of predicted, suggesting a similar activation mechanism.26 Pf PKG. To investigate the role of an IS for the regulation of Previous studies demonstrated that the C-terminal cyclic Pf PKG, we generated a deletion construct of Pf PKG lacking nucleotide binding domain D (CNB-D) is important for the the N-terminal putative autoinhibitory sequence (Pf PKG 32- activation of Pf PKG.17,18 A crystal structure of the isolated 853, catalytic domain with CNB-A/B/C/D without IS, Figure cGMP-bound CNB-D revealed a capping triad formed upon 1A). The mutant protein was expressed in E. coli TP2000 cGMP binding. This capping triad is conserved in all Δcya, a bacterial strain lacking adenylyl cyclase activity, and purification was performed under cyclic nucleotide-free apicomplexan PKGs, consisting of an arginine (R484 in conditions allowing for affinity measurements using fluores- Pf PKG) in the phosphate binding cassette (PBC) and a cence polarization (FP) and activation studies based on glutamine (Q532) as well as an aspartate (D533) in the αC- microfluidic mobility-shift assays (MSA, Caliper).27 The helix. Yet so far, the exact activation mechanism of this parasite expression of Pf PKG 32-853 and full length Pf PKG in kinase is largely unexplored. TP2000 resulted in low yields, and consequently, it was not In this study, we analyzed the molecular mechanisms of possible to quantify the exact phosphotransferase activity for Pf PKG regulation, studying the details of how domain these constructs. However, it was shown previously that the full length protein including the IS is activated at nanomolar organization and cGMP binding mediate activation and concentrations of cGMP.18 Although lacking the IS, the inhibition. Since the regulatory domain of Pf PKG significantly activity of Pf PKG 32-853 was efficiently inhibited in the differs from hPKG, this Plasmodium kinase could act as a absence of cGMP. Moreover, this mutant protein could be potential drug target for the treatment of malaria. activated by cGMP (Kact(cGMP) = 315 nM, Figure 1C), 416 DOI: 10.1021/acsinfecdis.7b00222 ACS Infect. Dis. 2018, 4, 415−423 ACS Infectious Diseases Article comparable to the full length protein (Kact(cGMP) = 330 nM, peptide (500 μM) (Figure 1B). Moreover, on the basis of Figure 1C and Deng et al.17). We therefore conclude that fluorescence polarization (FP), we were not able to control of activity may not depend on the interaction of the N- demonstrate any interaction of a FITC-labeled IS peptide terminal IS motif with the catalytic domain as described for (FITC-ERNKKKAIFSND) with Pf PKG (S1 Figure). hPKG. Binding of cGMP to CNB-D Is a Prerequisite for CNB-D Is Sufficient for Autoinhibition of Pf PKG. Since Pf PKG Activation. In previous studies, we and others activation still depends on cGMP binding, we then analyzed demonstrated that the CNB-C of Pf PKG is degenerate and which part of the regulatory domain is responsible for binds neither cGMP nor cAMP. While the isolated CNB-A and activation. For this, we sequentially deleted the individual -B have been described as low affinity nucleotide binding sites CNBs and generated the following constructs: Pf PKG 158-853 with low selectivity between cAMP and cGMP, the isolated (catalytic domain with CNB-B/C/D), Pf PKG 275-853 CNB-D is a high-affinity, cGMP selective site.17,18 To further (catalytic domain with CNB-C/D), Pf PKG 401-853 (catalytic investigate the involvement of CNB-D in the activation domain with CNB-D), and Pf PKG 519-853 (catalytic domain) mechanism, we selectively changed the conserved Arg within (Figure 1A). All deletion constructs were expressed in E. coli the phosphate binding cassette (PBC). This residue (R492 in TP2000 Δcya. Pf PKG 158-853 (catalytic domain with CNB- CNB-D), interacting specifically with cGMP, was replaced by a B/C/D) and 275-853 (catalytic domain with CNB-C/D) Lys residue. The R492K mutant protein bound cGMP with showed low basal activity and again, both could be activated extremely low affinity (EC50 > 100 μM, Table 1) and showed a with cGMP (19-fold, Figure 1B). Surprisingly, even Pf PKG tremendously elevated activation constant (approximately 600- 401-853, just comprising CNB-D and the catalytic domain, fold) (Figure 2), yet only slightly decreased maximum showed a low basal activity in the absence of cGMP (about 0.2 U/mg) and was efficiently activated by cGMP (30-fold to 5.3 U/mg, Figure 1B, S1 Table). This corresponds to about 60% activity of the full length protein.18 The catalytic domain (Pf PKG 519-853) showed high level expression in TP2000 yielding a stable protein; however, no kinase activity could be detected with the peptide substrates PKStide or Kemptide in the presence or absence of cGMP. It should be noted that, for the calculation of specific activities, the active protein concentration was determined by titration with Compound 2 (specific inhibitor of Pf PKG).28 Depending on the protein preparation, 50−80% of the protein concentration as determined by the Bradford assay was catalytically active. We then analyzed cGMP binding to the deletion construct Figure 2. Binding of cGMP to CNB-D is critical for activation of Pf PKG. cGMP activation data determined with a microfluidic containing CNB-D and the catalytic domain (Pf PKG 401-853: mobility-shift assay (MSA) using PKStide as substrate (1 mM). EC50 = 42 ± 6 nM, Table 1), revealing similar affinities for Recombinant mutant proteins Pf PKG 401-853 R492K and Pf PKG 158-853 R492K show increased Kact values for cGMP of about 600- Table 1. EC50 Values for cGMP Binding Deficient Mutants fold. A final concentration of 4−10 nM of the respective Pf PKG of Pf PKGa construct was used. Standard errors are specified as SD (standard deviation) of n ≥ 4 experiments. His-Pf PKG EC50 ± SD (nM) 401-853FP 43 ± 7 401-853SPR 24 ± 2 phosphotransferase activity compared to Pf PKG 401-853 401-853 R492KSPR >100 000 (Figure 1B, S1 Table). Along this line, binding of cGMP to 158-853FP 78 ± 9 CNB-D may cause a conformational change in this domain, 158-853 R250KFP 58 ± 6 which unleashes catalytic activity potentially by stabilizing an 158-853 R492KFP 115 ± 30 a active conformation of the kinase. FP: mean values of the EC50 for cGMP determined using FP Interestingly, when comparing Pf PKG 401-853 and full solution competition assay with standard deviation (SD) of n ≥ 4 length Pf PKG, the Hill slopes changed from 0.7 to 1.6, experiments. SPR: mean values of the EC50 for cGMP determined using SPR solution competition assay with standard deviation (SD) of indicating that CNB-A and -B entail some positive n ≥ 4 experiments. cooperativity (Figure 1D). To investigate the role of CNB-B in the regulation of Pf PKG, we used a construct encompassing CNB-B/C/D and the catalytic domain (Pf PKG 158-853). cGMP as described for the isolated CNB-D.18 Strikingly, the We now disrupted cGMP binding in CNB-B and CNB-D in activation constant for cGMP (Kact = 270 ± 10 nM) was this construct by mutating the respective Arg residues to Lys almost identical to the full length Pf PKG (Kact = 330 ± 20 (R250K and R492K) (Figure 1A). Interestingly, both mutant nM) (Figure 1C). proteins displayed comparable affinities for cGMP (70 to 110 These data again suggest that control of activity may not nM, Table 1), while previously a micromolar affinity for cGMP depend on the interaction of an N-terminal IS with the was determined for the isolated CNB-B.18 The specific activity catalytic domain, as described for human PKG (hPKG). of both mutant proteins was reduced to 50% compared to Hence, we tested the putative IS26 as a short peptide Pf PKG 158-853 (S1 Table). While disruption of cGMP (ERNKKKAIFSNDDF) on Pf PKG kinase activity in a binding to CNB-B (R250K) had no influence on the activation spectrophotometric assay.29 Pf PKG 401-853, preactivated constant, mutation of CNB-D (R492K) strongly influenced with cGMP (100 μM), could not be inhibited by this IS Pf PKG activation (Figure 2). Again, the Kact value was about 417 DOI: 10.1021/acsinfecdis.7b00222 ACS Infect. Dis. 2018, 4, 415−423 ACS Infectious Diseases Article Figure 3. cGMP binding to a chimeric protein construct of Pf PKG. (A) Sequence alignment of the individual phosphate binding cassettes (PBCs) of cyclic nucleotide binding domains (CNB-A, -B, -C, and -D). Residues corresponding to the capping triad (R484, Q532, and D533) of CNB-D are marked in yellow. (B) Chimeric recombinant protein constructs of Pf PKG: top, 401-853; bottom, 158-294/540-853; color code as follows: catalytic domain (black), CNB-B (blue), αC-helix of CNB-B (violet), CNB-D (red), and αC-helix of CNB-D (green). (C) cGMP binding affinity of Pf PKG constructs (4−8 nM) as determined by fluorescence polarization (FP) competition. Bar diagrams depicting the EC50 values with standard deviation (at least 4 experiments) for cGMP. When replacing the CNB-D (401-542) by the CNB-B (158-294/540-853), a 10-fold higher cGMP affinity was observed compared to Pf PKG 401-853 (catalytic domain with CNB-D). Data for the isolated CNB-B (158-294) were taken from Kim et al.18 600-fold higher compared to Pf PKG 158-853, suggesting that nM) and R528K (Kact = 75 ± 10 nM) displayed 3-fold lower cGMP binding to CNB-D triggers activation. activation constants (Figure 4D). While the catalytic activity of CNB-B Cannot Substitute for the CNB-D. Amino acid the D597N and Y480F mutant proteins was only slightly alignments of the CNB-D between all CNBs revealed the reduced compared to Pf PKG 401-853, the R528K mutation highest identity between CNB-D and CNB-B (38%, Figure resulted in a very low activity (9-fold reduced). Furthermore, 3A). To test whether the CNB-B can structurally and although the basal activity of all constructs was similar, only functionally substitute for the CNB-D, we generated a D597N and Y480F could be activated in a similar manner (25- construct where the catalytic domain was fused to the CNB- fold and 15-fold compared to 30-fold for Pf PKG 401-853). B (Pf PKG 158-294/540-853). This chimera bound cGMP Notably, R528K could be activated only 3-fold (Figure 4C), with low affinity (EC50 = 1260 nM), which is in line with data again supporting the indispensable role of the αC-helix in for the isolated CNB-B as determined previously (EC50 = 1200 CNB-D. nM, Kim et al.18) (Figure 3C). However, the chimera was not active in the presence or absence of cGMP. Apparently, CNB- DISCUSSION B cannot substitute for the function of CNB-D in the activation of the parasite specific kinase. Malaria is expected to increase in the near future since vaccines Role of the αC-Helix of CNB-D in Pf PKG Activation. are not readily available, and the emergence of increasing drug We compared several available apo-structures of full length resistance requires alternative therapeutic strategies. Protein Pf PKG (PDB code: 5DYK) and P. vivax PKG (PvPKG) (PDB kinases represent one target for the development of code: 5DZC). Those structures reflect the inhibited state of antimalarial drugs. Recent phosphoproteomics studies revealed the protein, and in both, an α(C)-helix connects the C- that numerous protein kinases are involved in the life cycle of terminal CNB-D to the respective catalytic domain. Structural P. falciparum.19,30 In particular, the cGMP-dependent protein analyses revealed an Arg residue (R528) in the αC-helix that kinase of P. falciparum offers potential for the rational design of forms a hydrogen bond with an Asp residue (D597) of the innovative antimalarial drugs. This protein kinase plays an catalytic domain. Interestingly, the same Arg (R528) was important role in gene regulation, proteolysis, and protein found to interact with the hydroxyl group of a Tyr residue transport within the malaria parasite life cycle in both sexual (Y480) in the PBC based on the crystal structure of the and asexual blood-stages.19 A recent study by Baker et al. isolated, cGMP-bound CNB-D (PDB code: 4OFG) (Figure demonstrated that targeting Pf PKG with the protein kinase 4A,B). To delineate the influence of these three amino acids on inhibitor ML10 blocks blood stage proliferation in vitro and activation, we introduced the following mutations into the hinders both merozoite egress and erythrocyte invasion in Pf PKG 401-853 construct: Y480F, R528K, and D597N. vivo.31 The primary structure of Pf PKG differs significantly Binding studies using FP demonstrated that none of these from its mammalian counterpart with regulatory domain of mutations had a significant effect on cGMP affinity (S2 Table); Pf PKG encompassing four instead of two CNBs.7,32 Align- however, activation studies with cGMP revealed an approx- ments of human and apicomplexan PKG sequences as well as imately 5-fold higher Kact value for Pf PKG 401-853 Y480F gel filtration chromatography experiments on Eimeria tenella compared to the wildtype construct (Kact = 1500 ± 170 nM PKG demonstrated that parasite PKGs lack the N-terminal versus 270 ± 10 nM). In contrast, D597N (Kact = 87 ± 10 leucine zipper motif and are monomeric.33,34 418 DOI: 10.1021/acsinfecdis.7b00222 ACS Infect. Dis. 2018, 4, 415−423 ACS Infectious Diseases Article Figure 4. Unique salt bridge forming residues (Y480, R528, and D597) are crucial for cGMP-dependent activation in Pf PKG. (A) Crystal structures of the isolated cyclic nucleotide binding domain D (CNB-D, Pf PKG 401-542) in complex with cGMP (PDB code: 4OFG) and (B) of the full length Pf PKG without cGMP (PDB code: 5DYK). The structure of the CNB-D with cGMP bound (A) shows an interaction between a critical Arg residue (R528) on the αC-helix (in red) with a Tyr residue (Y480) in the phosphate binding cassette (PBC in yellow). In contrast, in the full length structure without cGMP (B), R528 R528K forms a salt bridge with an Asp residue (D597) on the catalytic domain (in green). Recombinant protein constructs are shown in a cartoon representation with the secondary structure elements and cGMP shown as sticks. (C) Specific activity of recombinant proteins expressed by Pf PKG mutant constructs as determined by the spectrophotometric assay with 1 mM substrate (PKStide). Basal activity was determined without cGMP and maximum activity using 100 μM cGMP with a final protein concentration of 200−300 nM. R528K has a severe impact on the specific activity, and the respective construct can only be 3-fold activated with cGMP compared to wt (Pf PKG 401-853; 30-fold activation). (D) Kact values for Pf PKG mutant protein constructs (4−10 nM) using MSA. The mutant proteins Y480F show a 5-fold increased and D597N as well as R528K a 6-fold decreased Kact value compared to wt (Pf PKG 401-853). For all measurements, the standard errors are specified as SD of n ≥ 4 experiments. All structures were visualized using PyMOL v1.3 (Schrödinger LLC). For cyclic nucleotide-dependent protein kinases, two alanine. The pseudosubstrates tested, Ala-Glasstide37 and Ala- principle mechanisms of autoinhibition are described sharing Kemptide,38 showed no binding to the kinase based on FP the common theme of (pseudo)substrate inhibition. The first (data not shown). one, exemplified in mammalian PKG (mPKG), is based on Strikingly, all Pf PKG constructs with or without IS displayed kinase regulation by a short inhibitory sequence (IS), catalytic activity only in the presence of cGMP. This effect is mimicking the substrate consensus sequence.20−22 The second particularly prominent for the construct encompassing only the mechanism is employed by a close relative, cAMP-dependent catalytic domain and CNB-D (Pf PKG 401-853). This protein kinase (PKA), where autoinhibitory sequences are construct had low basal activity in the absence of cGMP but located on both types of physiological inhibitor proteins, the was efficiently activated by cGMP (30-fold). Thus, our results PKA regulatory subunits and the heat stable protein kinase indicate a completely novel regulatory mechanism, where the inhibitors (PKI). Peptides derived from these inhibitory control of kinase activity may not depend at all on sequences are efficient inhibitors of PKA kinase activity.35 autoinhibition by an N-terminal IS motif. As a consequence, Upon binding of cyclic nucleotides, the IS is displaced, other contacts between the regulatory and the catalytic allowing for catalytic activity.23−25 Interestingly, the sequences domains need to lock the kinase in a catalytically inactive of the respective IS in mammalian PKG isoforms are not highly state. This suggests a fundamentally different activation conserved.36 Yet, sequence alignments of apicomplexan PKGs mechanism for Pf PKG, where the parasite specific CNB-D with mammalian PKG isoforms indicated similarities in the IS, acts as a regulatory domain controlling catalytic activity in leading to the assumption that the activity of Pf PKG is also Pf PKG. These results are consistent with previous studies on controlled by autoinhibition.26 Our study, however, revealed Pf PKG, which showed that all three cGMP-binding sites are no influence of the putative IS on autoinhibition. This is based involved in kinase regulation, but the CNB-D has the strongest on two lines of evidence: (i) truncation of the putative N- influence on activation.17 In line with this, our chimeric terminal IS did not alter basal activity in the absence of cGMP; constructs, where CNB-D is replaced by CNB-B, still bound (ii) a peptide comprising the putative IS failed to inhibit cGMP but yielded an inactive protein. Pf PKG kinase activity in the presence of cGMP. We also To test in more detail how cGMP binding to CNB-D affects tested common substrate peptides of PKG, where in the activation, we generated a mutant enzyme defective in cGMP respective phosphorylation site a serine was replaced with an binding (Pf PKG 401-853 R492K). Introducing R492K within 419 DOI: 10.1021/acsinfecdis.7b00222 ACS Infect. Dis. 2018, 4, 415−423 ACS Infectious Diseases Article the PBC of CNB-D reduces cGMP binding affinity by more the R528K mutant protein could be activated only 3-fold with than 3 orders of magnitude, on the basis of surface plasmon cGMP and showed 10-fold reduced activity (0.6 U/mg). resonance (SPR) data. This Arg is a conserved feature of all Introducing a Phe at position 480 (Y480F), abolishing the functional PBCs and interacts with the phosphodiester of the interaction with R528, increased the Kact value about 5-fold. cyclic nucleotide.39 Interestingly, mutagenesis of a correspond- Here, the αC-helix cannot be stabilized in the active, cGMP- ing residue in bovine PKA RIα, R209K in CNB-A and R333K bound state anymore. Taken together, our model suggests that in CNB-B, reduced the cAMP affinity for CNB-A and CNB-B the catalytic activity of Pf PKG is mainly controlled by the αC- only by a factor of 10.40 While the Pf PKG cGMP deficient helix of CNB-D. Apparently, there are two different construct (Pf PKG 401-853 R492K) had only slightly conformations of the αC-helix. Without cGMP, the αC-helix decreased specific activity compared to Pf PKG 401-853, the stabilizes the catalytic domain in the inactive state via activation constant for cGMP was highly increased (600-fold interaction between R528 and D597. Upon cGMP binding, higher), in accordance with the strongly reduced affinity. Even the αC-helix moves toward the PBC and the interaction adding back CNB-B and -C (158-853 R492K) did not change between Y480 and R528 stabilizes the αC-helix in a the respective activation constants. In contrast, mutating solely conformation corresponding to the active state of Pf PKG. CNB-B (R250K) did not influence the activation constant. It Recent crystal structures reveal this specific interaction should be noted that the specific activity for these cGMP between R528 and D597 also in Plasmodium vivax PKG.31 deficient mutants was reduced by 50% compared to Pf PKG Having shown that the interaction between Y480 and R528 158-853, indicating that all functional CNBs seem to be is a prerequisite for Pf PKG activation, disruption of this required for maximal kinase activity. Mutation of the respective interaction could be a potential target for allosteric inhibitors. CNB-A and -B in Eimeria tenella PKG and in Toxoplasma This novel strategy may include specific small molecule gondii PKG resulted in a small effect on activity; however, competitors, restrained scaffolds like stapled peptides, or cyclic when CNB-D was disrupted, activity was decreased by >70%.32 nucleotide analogues (antagonists) selective for CNB-D. Already early studies on mPKG Iα demonstrated that occupation of CNB-A partially activates the kinase; yet, saturation of all CNBs is required for full activation.23,25,41 MATERIAL AND METHODS Construct Design and Mutagenesis. All protein Truncated constructs, where the CNBs were sequentially constructs of Pf PKG were ligated into the pQTEV-vector deleted, revealed similar activation constants comparable to full (N-terminal His-Tag) 27 applying BamHI and HindIII length Pf PKG (Figure 1C). However, the Hill slopes changed restriction sites. Site-directed mutagenesis was performed from 1.6 (Pf PKG 1-853, full length) to 0.7 (Pf PKG 401-853), using the KAPA HiFi Polymerase with the respective site- suggesting that N-terminal CNB-A and CNB-B induce positive specific primer pairs (KAPA Biosystems, Wilmington, MA, cooperativity. Our results are consistent with a study by Salowe USA): et al. reporting positive cooperativity (Hill slope of 1.7) for R250K for: GATGAACCGAAATCAGCCACAATT; another parasite PKG, EtPKG.32 R250K rev: AATTGTGGCTGATTTCGGTTCATC; R492K With the crystal structure of the full length Pf PKG (PDB for: CGACGAGCCTAAAACCGCAAGC; R492K rev: code: 5DYK) available, first insight into the domain GCTTGCGGTTTTAGGCTCGTCG; Y480F for: GCA- organization of PKG catalytic domains is provided. This AAAATGATTTCTTTGGCGAGCG; Y480F rev: CGCT- crystal structure reveals a single helix (αC-helix (AA 520-542)) CGCCAAAGAAATCATTTTTGC; R528K for: CATCT- between CNB-D and catalytic domain, originally assigned to GGAAGAAAAGATTAAAATGC; R528K rev: GCATTTT- the CNB-D.18 This αC-helix most likely corresponds to a AATCTTTTCTTCCAGATG; D597N for: highly conserved A-helix of the catalytic domain also found in CCGCAGAAAACAATCATCCGTTT; D597N rev: AAAC- PKA of various organisms.42,43 Here, this N-terminal GGATGATTGTTTTCTGCGG. Template DNA was di- amphipathic helix (A-helix, AA 15-40 in murine PKA) spans gested using DpnI (2 h, 37 °C). All mutant constructs were the surface between the small and the large kinase lobe verified by Sanger sequencing (GATC Biotech Konstanz, covering a hydrophobic surface. The A-helix is important for Germany). thermostability and for activation of PKA Type II holoen- Protein Expression and Purification. All Pf PKG protein zyme.42,43 For Pf PKG, we conclude that both the catalytic constructs were expressed in E. coli TP2000. Cells were grown domain and CNB-D share this helix. at 37 °C until an OD600 nm of 0.6−0.8. Expression was then Our previous structural analyses showed that binding of induced with 500 μM IPTG, and the protein constructs were cGMP to CNB-D induces a drastic conformational change in expressed overnight at 18 °C. The cells were harvested by the C-terminal αC-helix.18 An arginine (R528) located on the centrifugation at 7000g for 20 min and 4 °C. Next, the cells αC-helix directly interacts with an aspartate (D597) of the were resuspended using lysis buffer containing 50 mM sodium catalytic domain, stabilizing the αC-helix in a cGMP-free form, phosphate (pH 8.0), 300 mM NaCl, 10 mM imidazole, 5 mM corresponding to the inactive state of Pf PKG. Binding of 2-mercaptoethanol, and EDTA-free protease inhibitor (cOm- cGMP to CNB-D induces conformational changes, reorienting pleteTM, Roche, Basel, Switzerland) and lysed using a French R528 toward the PBC. In this cGMP-bound form, an press (Thermo Scientific, Waltham, MA, USA). The lysate was interaction between Y480 and R528 stabilizes the active centrifuged at 45 000g for 20 min at 4 °C and incubated with state. To investigate this proposed activation mechanism in Protino Ni-NTA agarose (Macherey-Nagel, Dueren, Ger- more detail, we introduced three specific mutations (Y480F, many). After incubation, the agarose was washed four times R528K, and D597N, Figure 4A,B). Weakening the interaction with washing buffer containing 50 mM sodium phosphate (pH between R528 and D597 by introducing either a Lys at 8.0), 300 mM NaCl, 60 mM imidazole, and 5 mM 2- position 528 (R528K) or an Asn at position 597 (D597N) mercaptoethanol. The His-tagged protein constructs were resulted in a 3-fold lower Kact value, suggesting that the αC- eluted with elution buffer containing 50 mM sodium helix cannot be stabilized longer in the inactive state. Notably, phosphate (pH 8.0), 300 mM NaCl, 250 mM imidazole, and 420 DOI: 10.1021/acsinfecdis.7b00222 ACS Infect. Dis. 2018, 4, 415−423 ACS Infectious Diseases Article 5 mM 2-mercaptoethanol. Finally, the samples were loaded screening pressure, −1.3 psi. Data analysis was done using onto a HiLoad 16/60 Superdex 75 gel filtration column (GE GraphPad Prism 6 by plotting the measured substrate Healthcare, Chalfont St Giles, United Kingdom) equilibrated conversion against the logarithm of the cGMP concentration. with a buffer containing 20 mM MOPS (pH 7.0), 300 mM The Kact values were calculated from sigmoidal dose−response NaCl, and 1 mM DTT. Sample purity was verified using SDS- (variable slope) curves. PAGE.44 Protein concentration was determined using the Determination of Specific Activity. The specific activity Bradford assay (Bio-Rad, Hercules, CA, USA). The protein was measured using a coupled spectrophotometric assay constructs were stored at 4 °C until use. At least two separate according to Cook et al.29 All measurements were performed protein preparations were used for all measurements. at room temperature in reaction buffer (100 mM MOPS (pH Fluorescence Polarization (FP). The fluorescence polar- 7.0), 10 mM MgCl2, 1 mM phosphoenolpyruvate, 1 mM ATP, ization (FP) assay was performed as described before.45 All 220 μM NADH, 5 mM 2-mercaptoethanol, 15.1 U/mL lactate measurements were performed in 20 mM MOPS (pH 7.0), dehydrogenase, and 8.4 U/mL pyruvate kinase). Pf PKG 150 mM NaCl, and 0.005% CHAPS. For the competition samples were mixed with reaction buffer and measured with assays, the cGMP (Biolog Life Science Institute, Bremen, cGMP (100 μM) and without cGMP in a quartz cuvette. Since Germany) dilutions were mixed with 0.5 nM (final some of the mutant proteins required higher cGMP concentration) 8-Fluo-cGMP (Biolog Life Science Institute, concentration for full activation (in the μM range) compared Bremen, Germany) and a fixed protein concentration (4−8 to wildtype (about 300 nM), we used 100 μM cGMP for all nM) in a black 384-well microtiter plate (Perking Elmer, spectrophotometric measurements (specific activity under full Waltham, MA, USA). For the IS peptide studies, fixed activation) to ensure comparable conditions. The kinase concentrations of the FITC-IS peptide (FITC-ERNKKKAIF- reaction was started by adding 1 mM PKStide (GRTGRRNSI; SNDDF) and the protein samples were mixed. Data GeneCust, Luxembourg). The absorption at 340 nm was acquisition was done using a Fusion α-FP microtiter plate monitored photometrically (Specord 205, Analytik Jena) for at reader at room temperature for two seconds at 485 nm (Ex) least 60 s, and the slope was determined. The specific activity and 535 nm (Em) with a PMT voltage of 1100 V. Binding data in U/mg was calculated on the basis of the Beer−Lambert law was analyzed using GraphPad Prism 6 by plotting the using GraphPad Prism 6. measured FP signal in mPol against the logarithm of the Active protein concentration for all protein constructs cGMP concentration. The EC50 values were calculated from including the catalytic domain were determined by titrating sigmoidal dose−response (variable slope) curves. with the inhibitor Compound 2. For this, 200−300 nM of the Surface Plasmon Resonance (SPR). Surface plasmon respective preactivated protein (with 100 μM cGMP) was resonance (SPR) assays were performed at 25 °C with running measured in the presence of various concentrations of buffer (20 mM MOPS (pH 7.0), 150 mM NaCl, and 0.005% Compound 2 (3−600 nM) in reaction buffer containing 1 (v/v) surfactant polysorbate P20) using a Biacore T200 system mM PKS-tide. The determined slopes were analyzed with GraphPad Prism 6. (GE Healthcare, Chalfont, UK). The cyclic nucleotide analog (8-AET-cGMP) was coupled to a CM5 sensor chip surface (GE Healthcare, Chalfont, UK) as described previously.46 For ASSOCIATED CONTENT solution competition assays, the cGMP (Biolog Life Science * S Supporting Information Institute, Bremen, Germany) dilutions were mixed with a fixed The Supporting Information is available free of charge on the protein concentration (10−20 nM) before injection. All ACS Publications website at DOI: 10.1021/acsinfec- measurements were performed at a flow rate of 30 μL/min. dis.7b00222. The association and dissociation phases were monitored for at S1 Figure, interaction between the putative IS peptide least 150 s, respectively. After each measurement cycle, the and Pf PKG recombinant proteins; S1 Table, specific surface was regenerated by two injections of SDS (0.5% w/v) activities of Pf PKG constructs; S2 Table, EC50 values for for 30 s and 1 M NaCl for 60 s. Data was analyzed using cGMP binding of Pf PKG constructs (PDF) GraphPad Prism 6 by plotting the SPR signal (RU) 10 s before the end of the association phase against the logarithm of the cGMP concentration. The EC50 values were calculated from AUTHOR INFORMATION sigmoidal dose−response (variable slope) curves. Corresponding Author Microfluidic Mobility-Shift Assay. The respective *E-mail: [email protected]. activation constants (Kact) were determined using a micro- ORCID fluidic mobility-shift assay on a Caliper DeskTop Profiler Eugen Franz: 0000-0002-6160-7477 (PerkinElmer, Waltham, MA, USA). cGMP dilution series Matthias J. Knape: 0000-0002-0474-9997 were prepared in 20 μL of buffer (20 mM MOPS (pH 7.0), Friedrich W. Herberg: 0000-0001-7117-7653 300 mM NaCl, 1 mM DTT, 0.05% L-31, 990 μM PKStide (GRTGRRNSI; GeneCust, Luxembourg), 10 μM FITC- Author Contributions PKStide (FITC-GRTGRRNSI; GeneCust, Luxembourg), 1 E.F. performed the biochemical experiments. E.F. and M.J.K. mM ATP, and 10 mM MgCl2) with a fixed protein analyzed the data. E.F., M.J.K., and F.W.H. wrote the concentration. Samples without cGMP were used as a control. manuscript. The samples were incubated for 2 h at room temperature in a Notes The authors declare no competing financial interest. 384-well plate (Corning LV, nonbinding surface) before measurement. A ProfilerPro LabChip (4-sipper mode, PerkinElmer, Waltham, MA, USA) was used for the separation ACKNOWLEDGMENTS of product and substrate using the following conditions: We thank Michaela Hansch (University of Kassel) for expert upstream voltage, 1900 V; downstream voltage, 500 V; technical assistance. We also thank David Baker (London 421 DOI: 10.1021/acsinfecdis.7b00222 ACS Infect. Dis. 2018, 4, 415−423 ACS Infectious Diseases Article School of Hygiene & Tropical Medicine, London) for cGMP-dependent protein kinase from Plasmodium falciparum. providing Compound 2 and the Pf PKG 1-853 DNA clone. Biochem. J. 374 (Pt 2), 559−565. We acknowledge Jascha Manschwetus (University of Kassel) (18) Kim, J. J., Flueck, C., Franz, E., Sanabria-Figueroa, E., and Eileen J. Kennedy (University of Georgia) for the peptide Thompson, E., Lorenz, R., Bertinetti, D., Baker, D. A., Herberg, F. W., and Kim, C. (2015) Crystal structures of the carboxyl cGMP synthesis of FITC-IS. E.F. is supported by a Ph.D. fellowship of binding domain of the Plasmodium falciparum cGMP-dependent the Kassel University as a member of the graduate program protein kinase reveal a novel capping triad crucial for merozoite Functionomics. F.W.H. is funded by the Deutsche For- egress. PLoS Pathog. 11 (2), e1004639. schungsgemeinschaft [Grant ID: He1818/10]. F.W.H. and (19) Alam, M. M., Solyakov, L., Bottrill, A. R., Flueck, C., Siddiqui, M.J.K. are supported in the funding line Future (Phosmorg) F. A., Singh, S., Mistry, S., Viskaduraki, M., Lee, K., Hopp, C. S., and by research training group (Clocks) of Kassel University. Chitnis, C. E., Doerig, C., Moon, R. W., Green, J. L., Holder, A. A., Baker, D. A., and Tobin, A. B. (2015) Phosphoproteomics reveals malaria parasite Protein Kinase G as a signalling hub regulating egress REFERENCES and invasion. Nat. Commun. 6, 7285. (1) Phillips, M. A., Burrows, J. N., Manyando, C., van Huijsduijnen, (20) Francis, S. H., Poteet-Smith, C., Busch, J. L., Richie-Jannetta, R. H., Van Voorhis, W. C., and Wells, T. N. C. (2017) Malaria. Nat. R., and Corbin, J. D. (2002) Mechanisms of autoinhibition in cyclic Rev. Dis Primers 3, 17050. nucleotide-dependent protein kinases. Front. Biosci., Landmark Ed. 7, (2) Cowman, A. F., Healer, J., Marapana, D., and Marsh, K. (2016) d580−d592. Malaria: Biology and Disease. Cell 167 (3), 610−624. (21) Francis, S. H., Smith, J. A., Colbran, J. L., Grimes, K., Walsh, K. (3) Miller, L. H., Ackerman, H. C., Su, X. Z., and Wellems, T. E. A., Kumar, S., and Corbin, J. D. (1996) Arginine 75 in the (2013) Malaria biology and disease pathogenesis: insights for new pseudosubstrate sequence of type Ibeta cGMP-dependent protein treatments. Nat. Med. 19 (2), 156−67. kinase is critical for autoinhibition, although autophosphorylated (4) Brochet, M., and Billker, O. (2016) Calcium signalling in malaria serine 63 is outside this sequence. J. Biol. Chem. 271 (34), 20748−55. parasites. Mol. Microbiol. 100 (3), 397−408. (22) Heil, W. G., Landgraf, W., and Hofmann, F. (1987) A (5) Baker, D. A. (2011) Cyclic nucleotide signalling in malaria catalytically active fragment of cGMP-dependent protein kinase. parasites. Cell. Microbiol. 13 (3), 331−9. Occupation of its cGMP-binding sites does not affect its (6) Doerig, C., Billker, O., Haystead, T., Sharma, P., Tobin, A. B., phosphotransferase activity. Eur. J. Biochem. 168 (1), 117−21. and Waters, N. C. (2008) Protein kinases of malaria parasites: an (23) Zhao, J., Trewhella, J., Corbin, J., Francis, S., Mitchell, R., update. Trends Parasitol. 24 (12), 570−7. Brushia, R., and Walsh, D. (1997) Progressive cyclic nucleotide- (7) Deng, W., and Baker, D. A. (2002) A novel cyclic GMP- induced conformational changes in the cGMP-dependent protein dependent protein kinase is expressed in the ring stage of the kinase studied by small angle X-ray scattering in solution. J. Biol. Plasmodium falciparum life cycle. Mol. Microbiol. 44 (5), 1141−51. Chem. 272 (50), 31929−36. (8) Li, J., and Cox, L. S. (2000) Isolation and characterisation of a (24) Wall, M. E., Francis, S. H., Corbin, J. D., Grimes, K., Richie- cAMP-dependent protein kinase catalytic subunit gene from Jannetta, R., Kotera, J., Macdonald, B. A., Gibson, R. R., and Plasmodium falciparum. Mol. Biochem. Parasitol. 109 (2), 157−63. Trewhella, J. (2003) Mechanisms associated with cGMP binding and (9) Taylor, C. J., McRobert, L., and Baker, D. A. (2008) Disruption activation of cGMP-dependent protein kinase. Proc. Natl. Acad. Sci. U. of a Plasmodium falciparum cyclic nucleotide phosphodiesterase gene S. A. 100 (5), 2380−5. causes aberrant gametogenesis. Mol. Microbiol. 69 (1), 110−8. (25) Alverdi, V., Mazon, H., Versluis, C., Hemrika, W., Esposito, G., (10) Taylor, H. M., McRobert, L., Grainger, M., Sicard, A., van den Heuvel, R., Scholten, A., and Heck, A. J. (2008) cGMP- Dluzewski, A. R., Hopp, C. S., Holder, A. A., and Baker, D. A. binding prepares PKG for substrate binding by disclosing the C- (2010) The malaria parasite cyclic GMP-dependent protein kinase terminal domain. J. Mol. Biol. 375 (5), 1380−93. plays a central role in blood-stage schizogony. Eukaryotic Cell 9 (1), (26) Baker, D. A., and Deng, W. (2005) Cyclic GMP-dependent 37−45. protein kinases in protozoa. Front. Biosci., Landmark Ed. 10, 1229− (11) Falae, A., Combe, A., Amaladoss, A., Carvalho, T., Menard, R., 1238. and Bhanot, P. (2010) Role of Plasmodium berghei cGMP-dependent (27) Bussow, K., Scheich, C., Sievert, V., Harttig, U., Schultz, J., protein kinase in late liver stage development. J. Biol. Chem. 285 (5), Simon, B., Bork, P., Lehrach, H., and Heinemann, U. (2005) 3282−8. Structural genomics of human proteins–target selection and (12) Mota, M. M., Hafalla, J. C., and Rodriguez, A. (2002) Migration generation of a public catalogue of expression clones. Microb. Cell through host cells activates Plasmodium sporozoites for infection. Fact. 4, 21. Nat. Med. 8 (11), 1318−22. (28) Donald, R. G., Zhong, T., Wiersma, H., Nare, B., Yao, D., Lee, (13) Ono, T., Cabrita-Santos, L., Leitao, R., Bettiol, E., Purcell, L. A., A., Allocco, J., and Liberator, P. A. (2006) Anticoccidial kinase Diaz-Pulido, O., Andrews, L. B., Tadakuma, T., Bhanot, P., Mota, M. inhibitors: identification of protein kinase targets secondary to cGMP- M., and Rodriguez, A. (2008) Adenylyl cyclase alpha and cAMP dependent protein kinase. Mol. Biochem. Parasitol. 149 (1), 86−98. signaling mediate Plasmodium sporozoite apical regulated exocytosis (29) Cook, P. F., Neville, M. E., Vrana, K. E., Hartl, F. T., and and hepatocyte infection. PLoS Pathog. 4 (2), e1000008. Roskoski, R. (1982) Adenosine cyclic 3′,5′-monophosphate depend- (14) Merckx, A., Nivez, M. P., Bouyer, G., Alano, P., Langsley, G., ent protein kinase: kinetic mechanism for the bovine skeletal muscle Deitsch, K., Thomas, S., Doerig, C., and Egee, S. (2008) Plasmodium catalytic subunit. Biochemistry 21 (23), 5794−5799. falciparum regulatory subunit of cAMP-dependent PKA and anion (30) Solyakov, L., Halbert, J., Alam, M. M., Semblat, J. P., Dorin- channel conductance. PLoS Pathog. 4 (2), e19. Semblat, D., Reininger, L., Bottrill, A. R., Mistry, S., Abdi, A., Fennell, (15) Leykauf, K., Treeck, M., Gilson, P. R., Nebl, T., Braulke, T., C., Holland, Z., Demarta, C., Bouza, Y., Sicard, A., Nivez, M. P., Cowman, A. F., Gilberger, T. W., and Crabb, B. S. (2010) Protein Eschenlauer, S., Lama, T., Thomas, D. C., Sharma, P., Agarwal, S., kinase a dependent phosphorylation of apical membrane antigen 1 Kern, S., Pradel, G., Graciotti, M., Tobin, A. B., and Doerig, C. (2011) plays an important role in erythrocyte invasion by the malaria parasite. Global kinomic and phospho-proteomic analyses of the human PLoS Pathog. 6 (6), e1000941. malaria parasite Plasmodium falciparum. Nat. Commun. 2, 565. (16) Hopp, C. S., Bowyer, P. W., and Baker, D. A. (2012) The role (31) Baker, D. A., Stewart, L. B., Large, J. M., Bowyer, P. W., Ansell, of cGMP signalling in regulating life cycle progression of Plasmodium. K. H., Jimenez-Diaz, M. B., El Bakkouri, M., Birchall, K., Dechering, Microbes Infect. 14 (10), 831−7. K. J., Bouloc, N. S., Coombs, P. J., Whalley, D., Harding, D. J., (17) Deng, W., Parbhu-Patel, A., Meyer, D. J., and Baker, D. A. Smiljanic-Hurley, E., Wheldon, M. C., Walker, E. M., Dessens, J. T., (2003) The role of two novel regulatory sites in the activation of the Lafuente, M. J., Sanz, L. M., Gamo, F. J., Ferrer, S. B., Hui, R., 422 DOI: 10.1021/acsinfecdis.7b00222 ACS Infect. Dis. 2018, 4, 415−423 ACS Infectious Diseases Article Bousema, T., Angulo-Barturen, I., Merritt, A. T., Croft, S. L., Gutteridge, W. E., Kettleborough, C. A., and Osborne, S. A. (2017) A potent series targeting the malarial cGMP-dependent protein kinase clears infection and blocks transmission. Nat. Commun. 8 (1), 430. (32) Salowe, S. P., Wiltsie, J., Liberator, P. A., and Donald, R. G. (2002) The role of a parasite-specific allosteric site in the distinctive activation behavior of Eimeria tenella cGMP-dependent protein kinase. Biochemistry 41 (13), 4385−91. (33) Diaz, C. A., Allocco, J., Powles, M. A., Yeung, L., Donald, R. G., Anderson, J. W., and Liberator, P. A. (2006) Characterization of Plasmodium falciparum cGMP-dependent protein kinase (PfPKG): antiparasitic activity of a PKG inhibitor. Mol. Biochem. Parasitol. 146 (1), 78−88. (34) Gurnett, A. M., Liberator, P. A., Dulski, P. M., Salowe, S. P., Donald, R. G., Anderson, J. W., Wiltsie, J., Diaz, C. A., Harris, G., Chang, B., Darkin-Rattray, S. J., Nare, B., Crumley, T., Blum, P. S., Misura, A. S., Tamas, T., Sardana, M. K., Yuan, J., Biftu, T., and Schmatz, D. M. (2002) Purification and molecular characterization of cGMP-dependent protein kinase from Apicomplexan parasites. A novel chemotherapeutic target. J. Biol. Chem. 277 (18), 15913−22. (35) Glass, D. B., Cheng, H. C., Kemp, B. E., and Walsh, D. A. (1986) Differential and common recognition of the catalytic sites of the cGMP-dependent and cAMP-dependent protein kinases by inhibitory peptides derived from the heat-stable inhibitor protein. J. Biol. Chem. 261 (26), 12166−12171. (36) Hofmann, F., Dostmann, W., Keilbach, A., Landgraf, W., and Ruth, P. (1992) Structure and physiological role of cGMP-dependent protein kinase. Biochim. Biophys. Acta, Mol. Cell Res. 1135 (1), 51−60. (37) Hall, K. U., Collins, S. P., Gamm, D. M., Massa, E., DePaoli- Roach, A. A., and Uhler, M. D. (1999) Phosphorylation-dependent inhibition of protein phosphatase-1 by G-substrate. A Purkinje cell substrate of the cyclic GMP-dependent protein kinase. J. Biol. Chem. 274 (6), 3485−95. (38) Whitehouse, S., Feramisco, J. R., Casnellie, J. E., Krebs, E. G., and Walsh, D. A. (1983) Studies on the kinetic mechanism of the catalytic subunit of the cAMP-dependent protein kinase. J. Biol. Chem. 258 (6), 3693−3701. (39) Mohanty, S., Kennedy, E. J., Herberg, F. W., Hui, R., Taylor, S. S., Langsley, G., and Kannan, N. (2015) Structural and evolutionary divergence of cyclic nucleotide binding domains in eukaryotic pathogens: Implications for drug design. Biochim. Biophys. Acta, Proteins Proteomics 1854 (10 Pt B), 1575−1585. (40) Herberg, F. W., Taylor, S. S., and Dostmann, W. R. (1996) Active site mutations define the pathway for the cooperative activation of cAMP-dependent protein kinase. Biochemistry 35 (9), 2934−42. (41) Wolfe, L., Francis, S. H., Landiss, L. R., and Corbin, J. D. (1987) Interconvertible cGMP-free and cGMP-bound forms of cGMP-dependent protein kinase in mammalian tissues. J. Biol. Chem. 262 (35), 16906−16913. (42) Veron, M., Radzio-Andzelm, E., Tsigelny, I., Ten Eyck, L. F., and Taylor, S. S. (1993) A conserved helix motif complements the protein kinase core. Proc. Natl. Acad. Sci. U. S. A. 90 (22), 10618−22. (43) Leon, D. A., Herberg, F. W., Banky, P., and Taylor, S. S. (1997) A stable alpha-helical domain at the N terminus of the RIalpha subunits of cAMP-dependent protein kinase is a novel dimerization/ docking motif. J. Biol. Chem. 272 (45), 28431−7. (44) Laemmli, U. K. (1970) Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227 (5259), 680−5. (45) Moll, D., Prinz, A., Gesellchen, F., Drewianka, S., Zimmermann, B., and Herberg, F. W. (2006) Biomolecular interaction analysis in functional proteomics. J. Neural Transm (Vienna) 113 (8), 1015−32. (46) Hanke, S. E., Bertinetti, D., Badel, A., Schweinsberg, S., Genieser, H. G., and Herberg, F. W. (2011) Cyclic nucleotides as affinity tools: phosphorothioate cAMP analogues address specific PKA subproteomes. New Biotechnol. 28 (4), 294−301. 423 DOI: 10.1021/acsinfecdis.7b00222 ACS Infect. Dis. 2018, 4, 415−423