Neurochondrin is an atypical RIIα-specific A-kinase anchoring protein

Abstract

Protein kinase activity is regulated not only by direct strategies affecting activity but also by spatial and temporal regulatory mechanisms. Kinase signaling pathways are coordinated by scaffolding proteins that orchestrate the assembly of multi-protein complexes. One family of such scaffolding proteins are the A-kinase anchoring proteins (AKAPs). AKAPs share the commonality of binding cAMP-dependent protein kinase (PKA). In addition, they bind further signaling proteins and kinase substrates and tether such multi-protein complexes to subcellular locations. The A-kinase binding (AKB) domain of AKAPs typically contains a conserved helical motif that interacts directly with the dimerization/docking (D/D) domain of the regulatory subunits of PKA. Based on a pull-down proteomics approach, we identified neurochondrin (neurite-outgrowth promoting protein) as a previously unidentified AKAP. Here, we show that neurochondrin interacts directly with PKA through a novel mechanism that involves two distinct binding regions. In addition, we demonstrate that neurochondrin has strong isoform selectivity towards the RIIα subunit of PKA with nanomolar affinity. This article is part of a Special Issue entitled: Inhibitors of Protein Kinases.

Keywords: A-kinase anchoring proteins; Neurochondrin; Protein kinase A; RII-specific.

Figure 1. Neurochondrin selectively binds RIIα through the dimerization/docking (D/D) domain

(a) Surface plasmon resonance measurements. Human RIIα (~160 RU), RIIβ (~300 RU), RIα (~600 RU) and RIβ (600 RU) subunits were captured in separate flow cells of 8-AHA-cAMP sensor chips. Binding of neurochondrin (500 nM) and dissociation was monitored for 300 s, respectively. As an additional control, binding of neurochondrin to 8-AHA-cAMP without immobilized R-subunit was subtracted. Shown is a representative plot from 5 independent experiments. RU: response units. (b) Far Western blot analysis. Equimolar amounts of RIα, RIIα and GST-1-44 RIIα as well as 125 ng neurochondrin (used as control) were separated by SDS-PAGE and transferred onto PVDF membranes. Membranes were incubated with (+) and without (−) neurochondrin. Binding of neurochondrin to the R-subunits was detected using Strep-Tactin® HRP conjugates (F: hyperfilm). As loading controls proteins on the membranes were stained with amido black (M: membrane). Neurochondrin interacts with RIIα as well as with the 1-44 RIIα deletion construct, which only contains the D/D domain (aa 1-44).

Figure 2. Inhibition of the interaction between neurochondrin and RIIα subunits by blocking the D/D domain with peptides Ht31 or AKAP18δL314E

Human RIIα (330 RU) was captured on 8-AHA-cAMP surfaces. Neurochondrin (300 nM) was injected alone (black) or in the presence of the inhibitor peptides Ht31 or AKAP18δL314E (5 μM each). Association and dissociation phases were monitored for 300 s each. Ht31 (red) and AKAP18δL314E (blue) efficiently prevent association of neurochondrin with RIIα. The inactive control peptide Ht31P had no inhibitory effect (5 μM; gray). Representative plot (n=2). RU: response units.

Figure 3. Neurochondrin binds RIIα with high affinity

SPR studies were performed with 70 RU of immobilized RIIα subunit (8-AHA-cAMP surface). Neurochondrin was injected at concentrations ranging from 7.8 nM to 250 nM for 300 s. After the end of the injection, the dissociation phase lacking neurochondrin was monitored for 300 s. A blank surface was subtracted and a Langmuir 1:1 binding model was applied (red lines). Shown is a representative sensorgram including fits (n=2). Association and dissociation rate constants were determined to 5.9 × 10⁴ M⁻¹s⁻¹ and 4.5 × 10⁻⁴ s⁻¹, respectively, yielding an equilibrium binding constant of 8 nM.

Figure 4. RII-binding domains of neurochondrin

Mapping the RII-binding domain of neurochondrin. Panels (a) and (b): The entire sequence of neurochondrin (729 amino acid residues) was spot-synthesized as overlapping peptides (25-mers, 20 amino acid residues overlap) and probed for RII binding using RII overlay assays. Signals were detected by autoradiography. RII overlay assays were performed in the presence of the inactive control peptide AKAP18δ-PP (left) or the peptide AKAP18δL314E, which blocks AKAP-PKA interactions [34]. The binding was not affected by the presence of either peptide. Panels (c) and (d): Two potential RII-binding domains were identified. Peptide sequences binding RII: position A11–A15 and D6–D11. (e) As control, binding of RII to the peptides Ht31 and AKAP18δL314E and the corresponding inactive control peptides Ht31PP and AKAP18δ-PP were monitored.

Figure 5. Determination of association and dissociation rate constants for the binding of the RIIα subunit to designed neurochondrin peptide I and II

Surface plasmon resonance studies were performed using covalently immobilized neurochondrin peptides I (a) and II (b). RIIα was injected in the indicated concentrations for 300 s. After the injection, the dissociation phase lacking RIIα was monitored for additional 300 s. A blank surface was subtracted and a Langmuir 1:1 binding model was applied (red lines). RIIα shows high affinity binding with both neurochondrin peptides (see Table 2). Representative plot (n=2). RU: response units.

Figure 6. The neurochondrin-derived peptides individually interact with RIIα in a non-canonical manner

Neurochondrin peptides I (a) and II (b) were immobilized covalently on a Biacore sensor surface. RIIα (500 nM) was either injected alone (black) or in the presence of the peptide Ht31 (5 μM), which inhibits canonical AKAP-PKA interactions. Association and dissociation phases were monitored for 300 s. Ht31P is the corresponding negative control peptide for Ht31. The interactions between RIIα and the neurochondrin peptides were not affected by Ht31 (red) or Ht31P (dark gray).

Figure 7. Surface charges of the neurochondrin RII-binding regions and the D/D domain of RIIα

(a) Model of neurochondrin. Both RII-binding regions of neurochondrin exhibit pI values of approximately 9. Therefore, they are positively charged under physiological conditions (blue label). (b) D/D domain of RIIα complexed with the D-AKAP2 peptide (PDB ID: 2HWN). Shown is the hydrophobic groove, formed by the D/D domain, where AKAPs usually bind (demonstrated by the D-AKAP2 peptide in yellow). The D/D domain of RIIα contains charged regions (negative charges are labeled in red whereas positive charges are labeled in blue). The arrows label negatively charged regions that may be important for the neurochondrin binding interface.

Figure 8. Phyre2-structure prediction of neurochondrin

The full-length amino acid sequence of neurochondrin was analyzed using Phyre2 [58]. (a) Neurochondrin shows a primarily helical architecture. Neurochondrin contains an N-terminal armadillo (ARM)-domain (aa: 36–468) with characteristic repeats of three helices (blue: helix I, magenta: helix II, yellow: helix III). (b) Different views of Neurochondrin (magenta: loop, cyan: helix). The N terminus is labeled with black dots.