The phytotoxin fusicoccin differently regulates 14-3-3 proteins association to mode III targets

Authors

  • Alessandro Paiardini,

    1. Department of Biochemical Sciences, Sapienza University of Rome, Rome, Italy
    Search for more papers by this author
  • Patrizia Aducci,

    Corresponding author
    1. Department of Biology, University of Rome “Tor Vergata”, Rome, Italy
    • Address correspondence to: Patrizia Aducci, Department of Biology, University of Rome “Tor Vergata”, via della Ricerca Scientifica, 00133 Rome, Italy. or Lorenzo Camoni, Department of Biology, University of Rome “Tor Vergata”, via della Ricerca Scientifica, 00133 Rome, Italy; Tel., +39 06 72594343; Fax, +39 06 2023500.E-mail: aducci@uniroma2.it or E-mail: camoni@uniroma2.it.

    Search for more papers by this author
  • Laura Cervoni,

    1. Department of Biochemical Sciences, Sapienza University of Rome, Rome, Italy
    Search for more papers by this author
  • Francesca Cutruzzolà,

    1. Department of Biochemical Sciences, Sapienza University of Rome, Rome, Italy
    2. Istituto Pasteur-Fondazione Cenci Bolognetti, Rome, Italy
    Search for more papers by this author
  • Cristina Di Lucente,

    1. Department of Biology, University of Rome “Tor Vergata”, Rome, Italy
    Search for more papers by this author
  • Giacomo Janson,

    1. Department of Biochemical Sciences, Sapienza University of Rome, Rome, Italy
    Search for more papers by this author
  • Stefano Pascarella,

    1. Department of Biochemical Sciences, Sapienza University of Rome, Rome, Italy
    Search for more papers by this author
  • Serena Rinaldo,

    1. Department of Biochemical Sciences, Sapienza University of Rome, Rome, Italy
    2. Istituto Pasteur-Fondazione Cenci Bolognetti, Rome, Italy
    Search for more papers by this author
  • Sabina Visconti,

    1. Department of Biology, University of Rome “Tor Vergata”, Rome, Italy
    Search for more papers by this author
  • Lorenzo Camoni

    Corresponding author
    1. Department of Biology, University of Rome “Tor Vergata”, Rome, Italy
    • Address correspondence to: Patrizia Aducci, Department of Biology, University of Rome “Tor Vergata”, via della Ricerca Scientifica, 00133 Rome, Italy. or Lorenzo Camoni, Department of Biology, University of Rome “Tor Vergata”, via della Ricerca Scientifica, 00133 Rome, Italy; Tel., +39 06 72594343; Fax, +39 06 2023500.E-mail: aducci@uniroma2.it or E-mail: camoni@uniroma2.it.

    Search for more papers by this author

Abstract

Modulation of the interaction of regulatory 14-3-3 proteins to their physiological partners through small cell-permeant molecules is a promising strategy to control cellular processes where 14-3-3s are engaged. Here, we show that the fungal phytotoxin fusicoccin (FC), known to stabilize 14-3-3 association to the plant plasma membrane H+-ATPase, is able to stabilize 14-3-3 interaction to several client proteins with a mode III binding motif. Isothermal titration calorimetry analysis of the interaction between 14-3-3s and different peptides reproducing a mode III binding site demonstrated the FC ability to stimulate 14-3-3 the association. Moreover, molecular docking studies provided the structural rationale for the differential FC effect, which exclusively depends on the biochemical properties of the residue in peptide C-terminal position. Our study proposes FC as a promising tool to control cellular processes regulated by 14-3-3 proteins, opening new perspectives on its potential pharmacological applications. © 2014 IUBMB Life, 66(1):52–62, 2014

Introduction

Protein–protein interactions are highly dynamic processes that play pivotal roles in many biological systems. As their dysregulation can contribute to human disease, modulation of these binding events by small molecules represents a promising strategy in drug discovery [1].

A remarkable example of regulatory proteins involved in protein–protein interactions is represented by 14-3-3 proteins, a family of conserved dimeric proteins that regulate a wide variety of cell processes in eukaryotic organisms, including cell differentiation and proliferation, intracellular trafficking, and gene transcription [2, 3]. Given the 14-3-3 role in various biological pathways that contribute to human diseases, including cancer progression and neurodegenerative diseases, modulation of their interaction with different proteins is a main target in pharmacological research [4].

14-3-3 proteins interact with several phosphorylated ligands through a conserved amphipathic groove present in each 14-3-3 monomer. Analysis of 14-3-3 binding sequence of the targets has allowed to identify common features and consequently to propose three main consensus binding motifs. The vast majority of targets bind 14-3-3s through the related motifs RSX(pS/pT)XP (mode I) and RXY/FX(pS/pT)XP (mode II), where X is any amino acid and pS/pT represents phosphorylated Ser or Thr [5].

In a restricted number of targets, the 14-3-3 binding site is located at their C-terminus. Although these binding sequences are not structurally related, they have been grouped in mode III consensus motifs, pS/pT(X0,1,2)-COOH [6].

The best characterized 14-3-3 target of the mode III subfamily is the plant plasma membrane H+-ATPase, a key enzyme for generation of the electrochemical gradient across the plasma membrane of plant cells, responsible for the control of fundamental processes [7]. 14-3-3 proteins associate to the C-terminal sequence YpTV-COOH, leading to enzyme activation. This interaction displays a remarkable trait, being strongly stabilized by fusicoccin (FC), a phytotoxic terpenoid produced by the fungus Phomopsis amygdali [8]. Despite the fungus being host-specific, FC is active in all higher plants, where it affects a number of biochemical and physiological processes, as a consequence of H+-ATPase activation [9].

FC mode of action has been clarified at the molecular level: the toxin binds to the preformed H+-ATPase/14-3-3 complex, thus greatly stabilizing the interaction and maintaining the enzyme in its activated state [10-12]. Crystal structure of the ternary complex between FC, 14-3-3, and a phosphopeptide reproducing the H+-ATPase binding sequence showed that the toxin inserts into the conserved amphipathic cavity of 14-3-3 and makes molecular contacts also with the C-terminal end of the peptide, thus promoting mutual stabilization of both ligands [13].

FC does not increase the affinity of 14-3-3 proteins to mode I or II targets [14, 15], because in these interactions the FC binding site of 14-3-3 is engaged in peptide binding, thus hampering FC binding in the groove. Conversely, structure similarity between the H+-ATPase and other mode III targets makes conceivable that 14-3-3 interaction with other mode III targets could be stabilized by FC. According to this hypothesis, we recently demonstrated the FC ability to stabilize 14-3-3 association to human GPIbα [15], a platelet glycoprotein which is part of GPI-IX-V, a protein complex that mediates the adhesion of circulating platelets to arteries and capillaries sub-endothelium [16]. As a consequence, FC promotes platelet adhesion and subsequent aggregation [15]. This finding proposes FC as a drug-like molecule potentially exploitable to control a number of physiological processes where 14-3-3 clients with mode III motifs take part [17].

To clarify the structural requisites of 14-3-3 targets to allow the FC regulatory function, the FC effect on the interaction between 14-3-3 proteins and phosphopeptides reproducing a mode III binding sequence has been investigated in this article.

Experimental Procedures

Chemicals

FC was prepared according to Ballio et al. [18]. The following peptides: p27Kip1 (VEQTPKKPGLRRRQpT), IL9-R (MLLPSVLSK ARSWpTF), KCNK3 (SLSTFRGLMKRRSpSV), GPR15 (HAEDFARRRKRSVpSL), HAP1 (GHPPASGTSYRSSpTL), and maize PLDδ (KILGASTSLPDSLpTM) were synthesized both in free and in biotin-bound form by JPT Peptide Technologies (Berlin, Germany). Maize H+-ATPase (MHA2 isoform) peptide (LKGLDIDTIQQNYTpV) was synthesized by Neosystem (Strasbourg, France).

Glutathione sepharose 4B was from GE Healthcare Biosciences (Pittsburgh, PA). Protein kinase A, streptavidin–agarose magnetic beads, and the other chemicals were from Sigma–Aldrich (St. Louis, MO).

Expression and Purification of 14-3-3ζ

Human 14-3-3ζ was expressed in Escherichia coli as fusion protein with the glutathione-S-transferase (GST) using pGEX-2TK vector, following the procedure described by Camoni et al. [15].

This expression vector produces a GST-fused protein containing a cAMP-dependent protein kinase phosphorylation site and a thrombin site between the two polypeptides. The fusion protein was immobilized onto glutathione sepharose beads and 14-3-3ζ purified by treating the immobilized protein with thrombin.

32P Labeling of 14-3-3 Proteins

14-3-3ζ was labeled with [32P]-ATP on the phosphorylation site present at junction between GST and the cloned protein using the catalytic subunit of PKA as already described [15]. Specific activity of 32P-14-3-3ζ was about 3 MBq/mg.

Binding of 14-3-3ζ to Resin-Bound Phosphopeptides

Biotinylated peptides (0.4 nmol) were immobilized onto 40 µL of streptavidin–agarose magnetic beads. Beads were incubated for 60 Min at RT in 50 µL of buffer H (20 mM Hepes-OH, 75 mM KCl, 5 mM MgCl2, 0.1 mM EDTA, pH 7.5) containing 0.04% Tween-20 and 10 µM of 32P-labelled 14-3-3ζ in the presence of 10 or 100 µM FC [19]. Resin-bound radioactivity was measured in a liquid scintillation β-counter.

Isothermal Titration Calorimetry

Isothermal titration calorimetry (ITC) experiments were performed using an iTC200 microcalorimeter (MicroCal, GE Healthcare Biosciences, Pittsburgh, PA), by titrating 14-3-3ζ with mode III peptides. Two microliters aliquots of 0.5 mM peptide in buffer H containing 1 mM benzamidine were injected into a 50 µM 14-3-3ζ in the same buffer at 25 °C. If indicated, 100 µM FC was added. Data were fitted using the “one-binding-site model” of the MicroCal version of ORIGIN, as previously described [20]. All measurements were done in duplicate.

Modeling

Structural superposition of the three-dimensional coordinates of 14-3-3 proteins was done using the Combinatorial Extension algorithm [21], as implemented in PyMod plugin [22]. 14-3-3ζ structure (PDB Code 1QJB) was chosen as archetypal of the ζ-isoform. PDB Codes 3IQV, 3P1O, and 1O9F were chosen as representatives of RSX, RRX, and QSX clusters, respectively, and subsequently used as initial structural templates to model mode III peptides. Each ternary complex was modeled using as structural template the 14-3-3c isoform from Nicotiana tabacum (PDB Code 2O98).

Subsequent in silico mutagenesis and energy minimization relied on the program Molecular Operating Environment (MOE, 2007). The energy minimization protocol has been applied according to Paiardini and Pascarella [23].

Results

FC Has Diverse Effects on 14-3-3 Binding to Phosphopeptides Reproducing Mode III Binding Sequences

14-3-3 targets with a mode III motif selected in this study (Fig. 1, top panel) are the cyclin-dependent kinase inhibitor p27Kip1 [24], the interleukin 9 receptor alpha chain (IL-9Rα, 25), the heart muscle system potassium channel KCNK3 [26], the G protein-coupled receptor GPR15 [27], and the neuronal Huntingtin-associated protein 1 (HAP1, 28). PLDδ, a maize phospholipase D with a putative mode III binding sequence (DSLpTM), has also been included. Phosphopeptides reproducing the mode III sequence of these targets and the reference H+-ATPase peptide were used in interaction experiments performed with the prototype 14-3-3ζ isoform.

Figure 1.

FC differentially regulates binding of 14-3-3ζ to mode III peptides. 14-3-3ζ binding assay to phosphopeptides reproducing the binding sites of p27Kip1, IL-9Rα, KCNK3, GPR15, HAP1, PLDδ, and H+-ATPase. Immobilized peptides were incubated with 32P-14-3-3ζ in the absence (white bars) or in the presence of 10 µM (gray bars) or 100 µM (black bars) FC. Resin-associated radioactivity was measured by scintillation counting. *P < 0.01 by Student's t-test (n = 9). [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]

The FC effect on 14-3-3/peptide association has been initially screened by measuring 14-3-3 binding to immobilized biotinyl-peptides under equilibrium conditions. As shown in Fig. 1, the FC effect is considerably different in the various peptides assayed. In fact, 14-3-3 binding to p27Kip1 peptide is unaffected by both 10 and 100 µM FC, while IL-9Rα association is inhibited by 100 µM FC. Conversely, 14-3-3 interaction with KCNK3, GPR15, HAP1, and PLDδ peptides is significantly stimulated both by 10 and 100 µM FC, resembling the well-known FC stabilizing effect on 14-3-3 interaction with the H+-ATPase peptide.

As it is known that peptide immobilization to a solid matrix may greatly hamper the interaction with a macromolecule such as the dimeric 14-3-3 protein, the affinity of these interactions has been quantitatively investigated by means of ITC, an innovative approach that determines the thermodynamic parameters of interactions in solution.

In these experiments, 50 µM 14-3-3ζ has been titrated with 0.5 mM peptide solutions in the presence or absence of 100 µM FC. As expected for specific binding, integration of the titration peaks produces sigmoidal enthalpy curves (Fig. 2), whose fit led to determine the stoichiometry of binding, the thermodynamic parameters, and the derived KD for each interaction (Table 1).

Figure 2.

Binding of mode III peptides to 14-3-3ζ by ITC measurements. 14-3-3ζ interaction with p27Kip1 (A), IL9-Rα (B), KCNK3 (C), GPR15 (D), HAP1 (E), PLDδ (F), and H+-ATPase (G). Analysis was performed by titrating 14-3-3ζ with each mode III peptide both in the absence (■) or in the presence of 100 µM FC (□). Experiments were performed in 20 mM Hepes-OH buffer (pH 7.5), 75 mM KCl, 5 mM MgCl2, 0.1 mM EDTA, and 1 mM benzamidine. Graphs show the integrated energy values normalized for injected protein. Binding isotherms were fitted using a single binding site model. All measurements were done in duplicate.

Table 1. Thermodynamic constant parameters of mode III peptides/14-3-3ζ interaction obtained by ITC
 Binding stoichiometry (n)Ka × 106 (M−1)KD (µM)ΔH (kcal/mol)−TΔS (kcal/mol)ΔG (kcal/mol)
  1. Dissociation constants (KD = 1/Ka), Gibbs free energy changes (ΔG = −RTlnKa), and entropy changes (TΔS = ΔH − ΔG) were calculated from ΔH and Ka.

P27Kip1
−FC0.82 ± 0.030.16 ± 0.036.25−7.10 ± 0.640.54 ± 0.03−6.56
+FC0.86 ± 0.060.23 ± 0.054.35−8.30 ± 1.200.96 ± 1.30−7.34
IL-R9α
−FC0.73 ± 0.0522.20 ± 3.200.045−4.00 ± 0.14−6.03 ± 0.23−10.03
+FC0.79 ± 0.066.06 ± 0.230.16−3.30 ± 0.10−5.99 ± 0.08−9.29
KCNK3
−FC1.05 ± 0.064.51 ± 0.390.22−6.40 ± 0.01−2.62 ± 0.05−9.02
+FC1.08 ± 0.0123.70 ± 2.00.042−8.50 ± 0.22−1.55 ± 0.17−10.05
GPR15
−FC0.96 ± 0.066.00 ± 0.270.17−5.90 ± 0.24−3.37 ± 0.21−9.27
+FC0.95 ± 0.0337.00 ± 9.400.03−4.40 ± 0.40−5.87 ± 0.25−10.27
HAP1
−FC0.82 ± 0.081.84 ± 0.170.54−5.2 ± 0.17−3.31 ± 0.14−8.51
+FC0.75 ± 0.1325.00 ± 11.000.04−5.9 ± 0.07−4.10 ± 0.23−10.00
PLDδ
−FC0.83 ± 0.030.27 ± 0.053.70−2.00 ± 0.46−5.44 ± 0.57−7.44
+FC0.84 ± 0.023.07 ± 1.300.33−3.80 ± 0.27−5.03 ± 0.54−8.83
H+-ATPase
−FC0.81 ± 0.0211.11 ± 1.870.09−5.43 ± 0.64−3.70 ± 0.03−9.13
+FC0.73 ± 0.0699.65 ± 29.800.01−8.58 ± 1.20−1.87 ± 1.30−10.45

ITC data are in total agreement with the peptide binding assay performed with immobilized biotinyl-peptides. Similarly to the H+-ATPase peptide, FC stabilizes 14-3-3 association to KCNK3, GPR15, HAP1, and PLDδ peptides, while the interaction with p27Kip1 peptide is unaffected and that with IL-9Rα is inhibited by the toxin.

The thermodynamic profiles of the aforementioned binding events indicate that, for the most part, protein/peptide interactions present both favorable enthalpic and entropic components, with the exception of p27Kip1 peptide, which displays a typical enthalpy driven interaction.

In more detail, binding affinity of the H+-ATPase peptide to 14-3-3 is increased 9.0-folds in the presence of FC and involves a strong favorable enthalpic variation. Binding affinity of KCNK3 to 14-3-3 increases 5.2-folds in the presence of FC and, under these experimental conditions, the ternary complex formation contributes favorably to the enthalpic component. Similarly, PLDδ/14-3-3 interaction in the presence of FC involves a predominant favorable enthalpic variation (as compared to the entropic one), with respect to the corresponding FC-free experiment, and the affinity increases 11.2-folds.

The presence of FC also increases the affinity of binding of GPR15 and HAP1 peptides to 14-3-3, 6.3- and 13.5-folds, respectively. The first interaction mainly involves a favorable entropic change, while the latter presents a similar favorable change of both the enthalpic and the entropic factors.

Conversely, p27Kip1 peptide binds 14-3-3 with low affinity, and the interaction is not significantly affected by FC. In agreement with binding experiments with immobilized biotinyl-peptides, the presence of 100 µM FC negatively affects (3.5-folds) the affinity of 14-3-3 for peptide IL-9Rα, thus confirming the inhibiting activity of the toxin in this interaction.

A Structural Rationale by Molecular Modeling

The analysis of available crystal structures revealed that several structural features are invariant in target peptides. These positions are represented by: i) the phosphate moiety (position 0), interacting via ion-pairs with Lys49, Arg56, Arg127, and a hydrogen-bond with Tyr128 (Fig. 3B, left); ii) the carboxy-terminal moiety (position +1), engaged in electrostatic and hydrogen-bonding interactions with the evolutionarily conserved Lys120 and Asn173 (Fig. 3B, center); iii) the main-chain at position +1, invariably interacting with Asn224 via hydrogen-bonds (Fig. 3, right).

Figure 3.

(A) Superposition of known peptide structures (green ribbons) into the binding site of 14-3-3ζ (PDB Code 1QJB, gray cartoons). (B) Conserved polar interactions of target peptides (light gray) in the 14-3-3 binding groove. Residues are numbered according to 14-3-3ζ. Left, the phosphate moiety (position 0), interacting via ion-pairs with Lys49, Arg56, Arg127, and a hydrogen-bond with Tyr128; center, the carboxy-terminal moiety (position +1), engaged in electrostatic and hydrogen-bonding interactions with Lys120 and Asn173 of 14-3-3ζ; right, the main-chain at position +1, interacting with Asn224 via hydrogen-bonds. (C) superposition of the α-chain of peptides used as structural templates for modeling. RRX (PDB Code 31PO), RSX (PDB Code 31QV), and QSX (PDB Code 1O9F) are represented as cyan, pink, and green scaffolds, respectively. [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]

Consequently, different affinities displayed by tested mode III peptides are mainly due to residues in position −1, −2, and −3. Position −1 can accommodate a great variety of residues with different physicochemical properties, with a slight preference for aromatic residues [5]. The presence of a Trp residue in −1 in IL9-Rα peptide can thus explain the high affinity of the IL9-Rα/14-3-3 complex (Table 1).

The residue in position −2 is mainly responsible for the N-terminal conformation of the backbone of bound peptides. Ser and Arg are the most favored residues in this position [29]. The small side chain of Ser allows peptide accommodation at the bottom of the binding cleft, while a bulky residue like Arg projects the main chain away from the floor of the groove. Position −3 is almost invariably occupied by an Arg residue, which is involved in an ion-pair interaction with the phosphate moiety of the phosphopeptide and in a close stacking interaction with the guanidinium ring of Arg60 [30]. In the PLDδ peptide, Arg in −3 is replaced by the acidic residue Asp. This substitution is detrimental for peptide affinity, as supported by the observed low affinity of PLDδ/14-3-3 complex (Table 1).

To provide a structural rationale for the differential FC effect observed in peptides/14-3-3 interaction experiments, structures of mode III peptide/14-3-3/FC ternary complexes have been inferred by molecular docking studies. The initial step of this method is the selection of known peptide/14-3-3 structures that will be used as templates for modeling. At the beginning of this study, different crystal structures of 14-3-3 in complex with mode I, II, and III targets have been solved (Table 2). The three-dimensional coordinates of peptides present in those complexes were transferred to the binding cavity of a representative structure of 14-3-3ζ (Fig. 3A) upon superposition.

Table 2. Dataset used in molecular modeling
PDB code14-3-3 isoform (organism)Amino acid sequence of peptides crystallized in complex with 14-3-3 proteins
Amino acid position respect to pS/pT
 −4−3−2−10+1+2
1A3714-3-3ζ (Bos taurus) QRSTpSTP
1IB114-3-3ζ (Homo sapiens) QRRHpTLP
1O9D14-3-3c (Nicotiana tabacum)  QSYpTV 
1O9F14-3-3c (N. tabacum)  QSYpTV 
1QJA14-3-3ζ (H. sapiens) RLYHpSLP
1QJB14-3-3ζ (H. sapiens) ARSHpSYP
1YWT14-3-3σ (H. sapiens) ARSHpSYP
2B0514-3-3ϒ (H. sapiens)  RAIpSLP
2BR914-3-3ε (H. sapiens) RRQRpSAP
2BTP14-3-3θ (H. sapiens)  RQRpSAP
2C1N14-3-3ζ (H. sapiens)  ARKpSTG
2C6314-3-3η (H. sapiens)  RAIpSLP
2C7414-3-3η (H. sapiens) RRQRpSAP
2NPMCF14ε (Cryptosporidium parvum)  RAIpSLP
2O9814-3-3c (N. tabacum) QQSYDI 
2V7D14-3-3ζ (B. taurus)  KSApTTT
2WH014-3-3ζ (H. sapiens) DRSKpSAP
3AXYGF14-c (Oryza sativa) QRVLpSAP
3CU814-3-3ζ (H. sapiens)  RSTpSTP
3E6Y14-3-3c (N. tabacum)  QSYpTV 
3IQJ14-3-3σ (H. sapiens) QRSTpSTP
3IQU14-3-3σ (H. sapiens) QRSTpST 
3IQV14-3-3σ (H. sapiens) QRSTpST 
3LW114-3-3σ (H. sapiens)   FKpTEG
3M5014-3-3c (N. tabacum) QQSYDI 
3M5114-3-3c (N. tabacum) QQSYDI 
3MHR14-3-3σ (H. sapiens)  RAHpSSP
3NKX14-3-3σ (H. sapiens) QRSTpSTP
3O8I14-3-3σ (H. sapiens) QRSTpSTP
3P1N14-3-3σ (H. sapiens) KRRKpSV 
3P1O14-3-3σ (H. sapiens) KRRKpSV 
3P1P14-3-3σ (H. sapiens) KRRKpSV 
3P1Q14-3-3σ (H. sapiens) KRRKpSV 
3P1R14-3-3σ (H. sapiens) KRRKpSV 
3P1S14-3-3σ (H. sapiens) KRRKpSV 
3UAL14-3-3ε (H. sapiens) IRSFpSEP
3UBW14-3-3ε (H. sapiens) IRSFpSEP
3UZD14-3-3ϒ (H. sapiens) IRSPpSLP

Comparison of peptide structures with their position in the 14-3-3 binding groove reveals three main structural clusters (Fig. 3C). In the first cluster (hereinafter “RSX”), positions −2 and −3 are almost invariantly occupied by Ser and Arg. The small side-chain of Ser in −2 keeps the peptide's main-chain lying on the cleft of 14-3-3. The second cluster, “RRX,” is characterized by two Arg residues in position −2 and −3. In this case, the presence of a bulky residue like Arg in position −2 projects the main-chain away from the floor of the binding cleft. Finally, the “QSX” cluster is characterized by the presence of a Ser residue at position −2 and mainly a Gln residue (not Arg) at −3.

Mode III peptides have been therefore modeled starting from a structural representative of RSX, RRX, or QSX clusters. In particular, RSX cluster has been selected for GPR15, IL9-R, and HAP1, RRX for KCNK3 and p27Kip1, and QSX for PLDδ peptide. To obtain models of 14-3-3ζ/peptide/FC complexes, in silico peptide mutagenesis followed by energy minimization has been then applied.

Structures obtained by modeling (Fig. 4) can account for the different FC effect observed in peptides/14-3-3 interactions experiments (Table 1). In fact, side-chains of amino acids occupying the +1 position of mode III peptides are predicted to be located in a hydrophobic cleft facing the FC-binding site of 14-3-3. In absence of FC, part of this position is exposed to the solvent; upon toxin binding, however, the cleft becomes occupied by FC hydrophobic rings. p27Kip1 peptide, lacking the residue in position +1, is not able to establish molecular contacts with FC (panel A). This predicted structure consequently confirms our experimental data, which point out the FC inability to modulate p27Kip1 peptide/14-3-3 association. Conversely, bulky hydrophobic residues could be disfavored in the presence of FC, due to steric hindrance of the latter inside the cleft. This observation could explain the behavior of IL-9Rα, bearing a Phe residue at position +1. In fact, the benzyl group of Phe sterically interferes with FC (panel B), which therefore inhibits peptide association to 14-3-3. Finally, peptides with small, non-polar side-chains at position +1, like KCNK3 (panel C), GPR15 (panel E), HAP1 (panel F), and PLDδ (panel D) are sterically compatible and able to establish hydrophobic interactions, thus explaining the FC-mediated energy stabilization experimentally observed.

Figure 4.

Modeling of mode III peptides bound to 14-3-3ζ. (A) Close-up view of p27Kip1 (purple) modeled with FC (cyan) in the binding groove of 14-3-3ζ (gray α-helices). (B) IL-9Rα/14-3-3ζ complex. FC was also modeled to show the predicted steric clash between Phe in position +1 and the toxin. (C–F) Predicted ternary complexes with KCNK3 (C), GPR15 (D), HAP1 (E), and PLDδ (F). [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]

The FC Effect Depends on the C Terminal Residue of Mode III Targets

As previously pointed out, pSer/pThr position and consequently peptide positioning are almost invariant in the inferred structures. Therefore, FC stimulatory/inhibitory action uniquely depends on the biochemical features of amino acid in position +1. To evaluate how each single amino acid in C-terminal position can accommodate in the binding cleft and establish molecular interactions with FC, in silico site-directed mutagenesis of the C-terminal residue of mode III peptide has been performed using the H+-ATPase/14-3-3/FC ternary structure as template. For each analysis, the more stable FC rotamer has been chosen. As shown in Table 3, positive potential energies of C-terminal amino acids with cyclic rings, like His, Pro, Tyr, Phe, and Trp, indicate that these groups are sterically incompatible with FC. Therefore, binding of peptides with these bulky residues at the C terminus is strongly disfavored by the toxin (Fig. 5, upper panel). Slight positive energy variations associated to Gln and Asn indicate that these residues at the C terminus weakly hamper peptide binding in the presence of FC, while null energy variations associated to Gly and Ala indicate that binding of peptides ending with these amino acids is unaffected by the toxin. Conversely, negative potential energies of remaining residues indicate that they are sterically compatible and able to establish molecular interactions with FC (Fig. 5, lower panel). As a consequence, 14-3-3 interaction of peptides ending with these amino acids is stimulated by FC.

Table 3. In silico site-directed mutagenesis of residue in position +1
ResiduePotential energy (kcal/mol)
  1. Internal strain energy of the best rotamer plus hydrogen bonding and electrostatic energy is shown.

His27.03
Pro25.22
Tyr23.97
Phe20.05
Trp15.76
Gln0.77
Asn0.68
Gly0.00
Ala0.00
Leu−1.81
Ile−2.95
Cys−3.85
Lys−3.86
Thr−3.88
Glu−4.15
Met−4.24
Arg−4.35
Ser−4.38
Asp−4.38
Val−5.37
Figure 5.

Inferred position of each amino acid substituting the residue in position +1 in the H+-ATPase phosphopeptide. Panel A, residues inhibiting the formation of the H+-ATPase/FC/14-3-3 ternary complex (red) and neutral residues (orange). Panel B, residue stimulating the interaction (green). [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]

Discussion

The identification of small cell-permeable molecules able to selectively modulate 14-3-3 binding with client protein represents a promising strategy to control physiological processes where 14-3-3 proteins play a role. In particular, the use of protein−protein stabilizers could allow to restore physiological processes impaired in pathologies correlated to 14-3-3 malfunctioning [4].

In this study, we show that FC, a cell-permeant fungal toxin whose activity has been so far deeply described only in plants, can represent a useful tool to selectively regulate 14-3-3 interactions with any target containing a mode III motif.

Mode III peptides used in this study bind 14-3-3ζ with different affinity, ranging from 45 nM of IL9-Rα/14-3-3 to 6.25 µM of p27Kip1/14-3-3 interactions. Analysis of crystallographic structures and modeling studies underline that some interactions are totally conserved, such as those established by the phosphate moiety, the main chain, and the carboxyl group in position +1. Therefore, the affinity of a peptide basically arises from the biochemical feature of residues surrounding the pSer/pThr. In particular, positions −3 and −2 mainly contribute to establish molecular connections with the 14-3-3 binding groove [5]. The optimal motifs Arg-Arg (KCNK3) and Arg-Ser (IL9-Rα and GPR15) can thus explain their high binding affinity. Conversely, in PLDδ peptide Arg in −3 is replaced by an Asp residue. This could explain the low binding affinity (KD = 3.7 µM) of the PLDδ/14-3-3 complex.

p27Kip1, although sharing the Arg–Arg motif, binds 14-3-3 at very low affinity (KD = 6.25 µM). This could result from the lack of any residue in position +1, which is essential to anchor the peptide to the 14-3-3 groove. Despite the affinity of each mode III peptide/14-3-3 complex, the FC effect is solely dependent on the physicochemical features of the residue in position +1. In fact, FC hampers the association of IL-9Rα (Phe in +1), while it greatly stimulates 14-3-3 interactions with KCNK3, GPR15, HAP1, and PLDδ, all displaying an hydrophobic side chain in +1 position. More generally, steric hindrance and limited plasticity of residues with cyclic rings obstruct FC insertion in the 14-3-3 binding groove, thus hampering ternary complex formation. In addition, we suggest that different residues, although biochemically diverging, are able to form favorable connections (Table 3).

Our studies open new perspectives in the control of cellular processes where mode III targets take part. A recurring outcome of 14-3-3 binding is the promotion of target delivery to the plasma membrane. A selective stabilizer of 14-3-3/target interaction could significantly affect levels of protein delivered to the plasma membrane, bringing about a strong perturbation of the physiological/pathological processes controlled by the proteins. Notably, it has been recently shown that a FC derivative enhances the surface expression of the mode III target TASK-3 (KCNK6) by stabilizing its interaction with 14-3-3 proteins [31].

In eukaryotes, elevated 14-3-3 expression has been associated to cellular transformation and proliferation [32].The development of inhibitors of 14-3-3 interactions is therefore a major target in cancer research [33]. However, in the treatment of some diseases, the alternative approach of drug-mediated stabilization of 14-3-3 protein/target interactions could be preferable [34]. In this respect, FC action on mode III targets could allow to selectively modulate specific pathways without influencing the complex network between 14-3-3 and their targets in the cell.

Very recently, a novel regulatory mechanism of estrogen receptor α (ERα), a master regulator of tumor cell proliferation expressed in the majority of breast tumors, has been discovered [35]. Binding of 14-3-3 proteins to the extreme ER C terminus (sequence ApTV-COOH) reduces dimerization and receptor activation. Intriguingly, FC stabilizes 14-3-3/ERα complex and reduces the estradiol-stimulated ERα dimerization, thus acting as an antiestrogenic compound with antiproliferative activity. Future work will address the potential pharmacological application of FC as a drug directed to 14-3-3/target complexes.

Acknowledgements

This work has been supported by a grant from Regione Lazio, Fondo per lo Sviluppo Economico, la Ricerca e l'Innovazione (P.A.), by the Italian Ministry of Education, University and Research (20094BJ9R7 to F.C., RBFR10LHD1 to S.R.), and by Sapienza, University of Rome (F.C. and A.P.). The authors thank Prof. Haian Fu for providing the 14-3-3ζ cDNA.

Ancillary