A nonnatural peptide targeting the A-kinase anchoring function of PI3Kγ for therapeutic cAMP modulation in pulmonary cells

Abstract

A-kinase anchoring proteins (AKAPs) are key orchestrators of cAMP signaling that act by recruiting protein kinase A (PKA) in proximity of its substrates and regulators to specific subcellular compartments. Modulation of AKAPs function offers the opportunity to achieve compartment-restricted modulation of the cAMP/PKA axis, paving the way to new targeted treatments. For instance, blocking the AKAP activity of phosphoinositide 3-kinase γ (PI3Kγ) improves lung function by inducing cAMP-mediated bronchorelaxation, ion transport, and antiinflammatory responses. Here, we report the generation of a nonnatural peptide, D-retroinverso (DRI)-Pep #20, optimized to disrupt the AKAP function of PI3Kγ. DRI-Pep #20 mimicked the native interaction between the N-terminal domain of PI3Kγ and PKA, demonstrating nanomolar affinity for PKA, high resistance to protease degradation and high permeability to the pulmonary mucus barrier. DRI-Pep #20 triggered cAMP elevation both in vivo in the airway tract of mice upon intratracheal administration, and in vitro in bronchial epithelial cells of cystic fibrosis (CF) patients. In CF cells, DRI-Pep #20 rescued the defective function of the cAMP-operated channel cystic fibrosis transmembrane conductance regulator, by boosting the efficacy of approved cystic fibrosis transmembrane conductance regulator modulators. Overall, this study unveils DRI-Pep #20 as a potent PI3Kγ/PKA disruptor for achieving therapeutic cAMP elevation in chronic respiratory disorders.

Keywords: AKAP; PI3Kγ; PKA; cAMP; peptide; respiratory diseases.

Figure 1

DRI-Pep #20 is a potent PI3Kγ/PKA disruptor peptide.A, chemical structure of DRI-Pep #20. The amino acid sequence of DRI-Pep #20 comprises the nonnatural D-peptide RHQGK, the D-retroinverso (DRI)-isoform of the cell penetrating peptide Penetratin 1 (P1) and a glycine (G) linker. B, schematic representation of the fluorescence spectroscopy assays for the characterization of the interaction between DRI-Pep #20 (or PI3Kγ MP) and the recombinant fluorescein 5-maleimide–labeled PKA-RIIα (PKA-F5M). C, steady-state emission spectra of PKA-F5M in the presence of increasing concentrations of DRI-Pep #20 (0–20 μM). K_D: dissociation constant. Inset, nonlinear fitting of the fluorescence intensity maxima obtained at various concentrations of DRI-Pep #20 for the monitoring of bio-labeled PKA. K_A: association constant. D, for kinetic analysis, fluorescence spectra of PKA-F5M in the presence of increasing concentrations of DRI-Pep #20 or PI3Kγ MP (inset) were analyzed and fitted to a single exponential function to obtain the observed rate constant (k_obs). The binding of DRI-Pep #20 or PI3Kγ MP to biolabeled PKA was investigated under pseudo-first-order conditions, and the kinetic constants, k_on and k_off, were determined. E, schematic representation of the displacement assay between DRI-Pep #20 (or PI3Kγ MP) and the PI3Kγ/PKA-F5M complex. F, percentage displacement of the PI3Kγ/PKA-RIIα complex by DRI-Pep #20 or PI3Kγ MP, calculated from steady-state emission spectra of the PI3Kγ/PKA-F5M complex in the presence of increasing concentrations of the peptides (0–5 μM). The displacement efficiency was expressed as percentage of the binding between PI3Kγ and PKA-F5M relative to that in the absence of peptides. G, cAMP concentrations in peritoneal macrophages from WT (in green) and PI3Kγ^−/− mice (in gray) treated with DRI-Pep #20 (1–25 μM) for 30 min. The amount of cAMP was expressed as percentage of cAMP accumulation observed in untreated PI3Kγ^−/− cells. n ≥ 6 technical replicates from N > 3 independent experiments. ∗∗∗p < 0.001 WT versus PI3Kγ^−/− and #p < 0.05, ##p < 0.01, and ###p < 0.001 UT versus DRI-Pep #20 by one-way ANOVA, followed by Bonferroni’s post hoc test. Data are means ± SD. AU, arbitrary units; PKA, protein kinase A; PKA-RIIα, PKA regulatory subunit RIIα; PI3Kγ, phosphoinositide 3-kinase gamma; PKA-F5M, fluorescein 5-maleimide–labeled PKA-RIIα.

Figure 2

Structural prediction of the binding between DRI-Pep #20 and PKA-RIIα.A, DRI-Pep #20 structure prediction by PEP-FOLD3.5. P1-G and RHQGK domains are shown as cartoons in gray and red, respectively. R-1, H-2, Q-3, and K-5 residues are indicated and shown as sticks. B, circular dichroism spectra of DRI-Pep #20 showing a peak at 190–240 nm. The percentage of α-helical and β-sheet secondary structures calculated by the K2D3 software are indicated. C, molecular docking simulation of the interaction between DRI-Pep #20 and the PKA-RIIα dimer by HADDOCK 2.4. The docked pose of DRI-Pep #20 in complex with residues 2 to 44 of PKA-RIIα (cartoon in green) is shown. The key residues involved in the binding are indicated and shown as sticks, with DRI-Pep #20 residues in bold. Hydrogen bonds between DRI-Pep #20 and PKA-RIIα are indicated by yellow dashed lines. In (A and C), the structural models were developed using PyMOL. DRI, D-retroinverso; HADDOCK, high ambiguity driven biomolecular DOCKing; PI3Kγ, phosphoinositide 3-kinase gamma; PKA, protein kinase A; PKA-RIIα, PKA regulatory subunit RIIα.

Figure 3

Structural prediction of the native binding between the N-terminal domain of PI3Kγ and PKA-RIIα.A, molecular docking simulation of the interaction between PI3Kγ and the PKA-RIIα dimer by HADDOCK 2.4. The docked pose of residues 109 to 159 of PI3Kγ in complex with residues 2 to 44 of the PKA-RIIα dimer (green cartoon) is shown. The amino acids critical for the binding between the two proteins are shown and indicated as sticks, with the residues of PI3Kγ in bold. The putative PKA-binding motif of PI3Kγ (126–150) is shown in orange and blue. The sequence in orange indicates the region of PI3Kγ that was identified as being at the core of the interaction (KATHR). Hydrogen bonds between PI3Kγ and PKA-RIIα are indicated by yellow dashed lines. B, structural prediction of the KATHR sequence by PEP-FOLD3.5. KATHR and P1-G domains are shown as cartoons in orange and gray, respectively. K-18, H-21 and R-22 residues of the KATHR sequence (corresponding to K-126, H-129 and R-130 of native PI3Kγ) are indicated and shown as sticks. C, molecular docking simulation of the interaction between KATHR and the PKA-RIIα dimer by HADDOCK 2.4. The docked pose of KATHR in complex with residues 2 to 44 of PKA-RIIα (cartoon in green) is shown. Yellow dashed lines indicate hydrogen bonds between KATHR and 2 to 44 PKA-RIIα. The amino acids critical for the binding are indicated and shown as sticks, with KATHR residues in bold. Throughout, the structural models were developed using PyMOL. HADDOCK, high ambiguity driven biomolecular DOCKing; PI3Kγ, phosphoinositide 3-kinase gamma; PKA, protein kinase A; PKA-RIIα, PKA regulatory subunit RIIα.

Figure 4

DRI-Pep #20 increases cAMP levels locally in vivo in the airway tract of mice.A, schematic representation of the treatment schedule. Mice received DRI-Pep #20 through intratracheal (i.t.) instillation. B–D, cAMP concentrations in tracheas (B), lungs (C) and hearts (D) from BALB/c mice 24 h after i.t. instillation of different doses of DRI-Pep #20 (0–750 mg/kg). Values in brackets indicate the dose of DRI-Pep #20 expressed as mg/kg. The number of mice (n) ranged from three to six per group. EC₅₀, median effective concentration. E–G, cAMP concentrations in tracheas (E), lungs (F) and hearts (G) from WT and PI3Kγ^−/− mice 24 h after i.t. instillation of 10 μg/Kg DRI-Pep #20 (in green) or PBS (in gray). The number of mice (n) ranged from three to four per group. In (A and B), ∗p < 0.05, ∗∗p < 0.01 and ∗∗∗p < 0.001 by one-way ANOVA, followed by Bonferroni’s post hoc test. In (E and F) ∗p < 0.05 and ∗∗p < 0.01 PBS versus DRI-Pep #20 by two-way ANOVA test, followed by Bonferroni’s post hoc analysis. Throughout, data are means ± SD. DRI, D-retroinverso; PI3Kγ, phosphoinositide 3-kinase γ.

Figure 5

DRI-Pep #20 can penetrate pathological mucus and resist protease degradation.A, schematic representation of the parallel artificial membrane permeability assay (PAMPA) with and without cystic fibrosis (CF) sputum deposition on top of the artificial lipid membrane (PM). B, apparent permeability (P_app) measurements of DRI-Pep #20 (2 mg/ml), in the absence (green box) and in the presence (blue box) of CF sputum. The green dashed line indicates the P_app typical of high-medium permeable compounds (4 × 10⁻⁶ cm·s⁻¹), while the red dashed line denotes the P_app of medium-low permeable molecules (1 × 10⁻⁶ cm·s⁻¹). ns: nonsignificant by Student’s t test. C, representative transmission electron microscopy (TEM) images of DRI-Pep #20 (0.1 mg/ml in water). The scale bar represents 20 nm. D, size distribution profile of DRI-Pep #20 (4 mg/ml in 2 mM PBS) obtained by dynamic light scattering (DLS) analysis. E and F, cAMP concentrations in 16HBE14o- cells treated with the DRI-Pep #20 and Pep #20 (25 μM for 30 min) in the absence (green bars) and in the presence (gray bars) of either 3 μg/ml (E) or 20 μg/ml (F) human neutrophil elastase (HNE). The amount of cAMP was expressed as percentage of cAMP accumulation elicited by Pep #20 without HNE. Dashed lines indicate the amount of cAMP (%) induced by Pep #20 with HNE as a reference. n ≥ 6 technical replicates from N > 3 independent experiments. ∗p < 0.05 and ∗∗p < 0.01 by one-way ANOVA, followed by Bonferroni’s post hoc test. G, cAMP elevation in 16HBE14o- cells covered with a layer of CF sputum and then treated with 25 μM DRI-Pep #20 for 30 min and 1 h. The amount of cAMP was expressed as percentage of cAMP accumulation elicited by DRI-Pep #20 in the absence of sputum at 30 min ∗∗p < 0.01 and ∗∗∗p < 0.001 versus DRI-Pep #20 without sputum by two-way ANOVA test, followed by Bonferroni’s post hoc analysis. n ≥ 6 technical replicates from N > 3 independent experiments. Throughout, data are means ± SD. DRI, D-retroinverso; ns, non-significant.

Figure 6

DRI-Pep #20 modulates WT and F508del-CFTR activity.A, schematic representation of CFTR activity measurement through the Premo Halide Sensor assay. B, average fluorescence quenching traces of 16HBE14o- cells expressing the halide-sensitive yellow fluorescent protein (HS-YFP) and treated with either 25 μM DRI-Pep #20 or equimolar amount of the control peptide P1 for 30 min before addition of Premo Halide stimulus buffer. Fluorescence was continuously read (1 point per second) starting at 1 s before addition of the buffer and up to 120 s. The CFTR inhibitor CFTR_inh-172 (10 μM for 5 min) was used to evaluate the selective activation of the CFTR channel. C, CFTR activity (expressed as the change in fluorescence ΔF/F0) in response to 30-min stimulation with increasing concentrations of DRI-Pep #20 (31.6 nM–316 mM) in 16HBE14o- cells expressing HS-YFP. To determine the EC₅₀, nonlinear regression analysis was used. D, CFTR activity (expressed as the change in fluorescence ΔF/F0) in 16HBE14o- cells expressing HS-YFP and treated with 10 to 25 μM DRI-Pep #20 for 30 min in the absence or in the presence of the CFTR inhibitor CFTR_inh-172 (10 μM for 5 min). The adenylyl cyclase activator, forskolin (FSK), was used as a positive control (100 nM for 5 min), while P1 was used as a negative control (25 μM for 30 min). E, CFTR activity in F508del-CFTR-CFBE41o- cells expressing HS-YFP and treated with elexacaftor/tezacaftor/ivacaftor alone (ETI) or together with DRI-Pep #20. Cells were corrected with elexacaftor (3 μM) and tezacaftor (10 μM) for 24 h and then exposed acutely to ivacaftor (1 μM) for 30 min, alone (ETI) or together with 25 μM DRI-Pep #20. The CFTR inhibitor CFTR_inh-172 was used as in (B). UT: untreated cells. In (B and E), n ≥ 3 technical replicates from N > 3 independent experiments. ∗∗p < 0.01 and ∗∗∗p < 0.001 versus UT and ###p < 0.001 ETI versus ETI plus DRI-Pep #20 by one-way ANOVA, followed by Bonferroni’s post hoc test. Throughout, data are means ± SD. CFBE, cystic fibrosis bronchial epithelial; CFTR, cystic fibrosis transmembrane conductance regulator; DRI, D-retroinverso; HS-YFP, halide-sensitive yellow fluorescent protein; P1, penetratin 1; UT, untreated cells.