Fig. 1. Phosphorylation-dependent binding of 14-3-3 to HERG. Anti-HERG immunoblots of GST–14-3-3 binding assays. (A) Wild-type HERG (enriched for the higher molecular weight glycosylated form) binds GST–14-3-3ε but not GST. Lysate Input indicates a sample of the total cellular lysate applied to the immobilized GST fusion proteins for the binding assay. (B) HERG binds ε and η isoforms of 14-3-3 but not mutant R56,60A 14-3-3η. (C) R18 peptide blocks 14-3-3 binding to HERG. (D) Interaction of the N-terminus of HERG with 14-3-3 in yeast depends on intact S283. The EGY48 yeast strain expressing full-length 14-3-3ε was mated with the YM4271 strain expressing the pGilda plasmid containing: 1, wild-type HERG 1–375; 2, S283A HERG 1–375; 3, wild-type HERG 1–215; 4, blank; 5, pLexA-laminin; or 6, wild-type HERG 690–1159 (C-terminus). (E) Binding of wild-type and PKA site HERG mutants to GST–14-3-3ε with (+) or without (–) CPT-cAMP stimulation prior to cell lysis. Δ indicates a serine or threonine to alanine mutation of a specified PKA phosphorylation site in HERG as indicated by the schematic below. (F) Pre-treatment of HERG-expressing cells with staurosporine (Staur.) decreases 14-3-3 interaction, and alkaline phosphatase (A.P.) treatment of HERG lysates blocks 14-3-3 interaction.

Fig. 2. Interaction between HERG and 14-3-3 in eukaryotic cells. (A) Double-labeled immunofluorescence of HEK 293 cells co-transfected with wild-type (upper panels) or Δ1234 HERG (lower panels) stained for 14-3-3 (red, middle panels), or HERG (green, left panels), with merged channels (right panels) demonstrating co-localization of the wild-type, but not Δ1234 HERG with 14-3-3. (B) Immunoblot analysis of HERG interaction with 14-3-3 in eukaryotic cells. HEK 293 cells were transfected with either HERG-myc or Δ1234 HERG-myc, with or without FLAG-14-3-3ε.

Fig. 3. Direct interaction of HERG and 14-3-3ε. Gels are Coomassie Blue stained for total protein. Recombinant HERG C-terminus (residues 979–1159, GST fusion) and 14-3-3ε (MBP fusion) were expressed separately in E.coli and examined for in vitro binding. (A) Purified and eluted MBP fusion proteins (MBP alone or MBP–14-3-3ε) applied to immobilized GST or GST–HERG C-terminus (after pre-phosphorylating with PKA). Expression of fusion proteins is confirmed in the purified bacterial lysates (Lysate Input). MBP–14-3-3 but not MBP alone co-precipitated with GST–HERG. GST alone failed to precipitate MBP–14-3-3. (B) Conversely, when purified GST–HERG (that had been in vitro PKA phosphorylated) was applied to immobilized MBP fusion proteins, it only co-precipitated with immobilized MBP–14-3-3ε but not with MBP alone.

Fig. 4. Functional consequences of HERG interaction with 14-3-3. Representative whole-cell current tracings from CHO cells expressing wild-type (WT) HERG alone (A) or co-expressing wild-type HERG and 14-3-3ε (WT+14-3-3ε) (B). Time and current scale bars apply to both (A) and (B). Insets show expanded views of the early development of outward current. (C) I–V curves for wild-type HERG with (open circles) or without (closed circles) overexpressed 14-3-3ε. (D) Normalized voltage-dependent activation curves with (open circles) or without (closed circles) 14-3-3ε overexpression. (E) Summary of shifts in V_h of wild-type HERG (n = 31), Δ23 HERG (n = 10), Δ1234 HERG (n = 20) and wild-type HERG + KCNE2 (n = 7) in the presence of 14-3-3ε. V_h measurements were obtained from the Boltzman fits of normalized voltage-dependent activation curves. Shifts in voltage dependence of activation were determined as the difference between V_h with 14-3-3 and V_h without 14-3-3 overexpression. (F) Time constants of onset of outward current shown for HERG with and without overexpressed 14-3-3ε.

Fig. 5. Functional effects of HERG association with 14-3-3 require dimerization of 14-3-3 and cross-bridging of HERG. (A) Summary of V_h shifts from CHO cells expressing different PKA mutants of HERG in the presence and absence of 14-3-3ε overexpression. Shifts are calculated as the difference in V_h between cells without and with 14-3-3. (B) V_h from CHO cells expressing HERG and 14-3-3ε with (+, n = 14) or without (–, n = 17) R56,60A 14-3-3η. Inset, heterodimerization of 14-3-3ε (FLAG) and R56,60A 14-3-3η (myc) confirming the mutant η isoform’s dominant-negative effect over the ε isoform (Lysate Input indicates a sample of the total cellular lysate applied to the gel without immunoprecipitation). (C and D) Anti-Myc immunoblots demonstrating the interaction of 14-3-3 with isolated C- (C) and N- (D) termini of HERG in GST–14-3-3 binding assays requiring intact PKA phosphorylation sites. (E) Anti-Myc immunoblot of simultaneous binding of HERG N- and C-termini to GST–HERG C-terminus mediated by 14-3-3(ε or η). A non-specific Myc-immunoreactive band was observed between the N- and C-terminal fragments of HERG whenever GST–HERG C-terminus-expressing bacterial lysates were probed. (Lysate Input indicates a sample of the total cellular lysate applied to the immobilized GST fusion proteins for the binding assay.)

Fig. 6. 14-3-3-mediated stabilization of PKA-dependent phosphates on HERG. (A) Overexpression of 14-3-3 with HERG reduces the capacity for in vitro back-phosphorylation with the catalytic subunit of PKA. Top panel: autoradiography signal from immunoprecipitated HERG (either Δ23 or Δ234 PKA mutants) phosphorylated with PKA and [³²P]ATP. Middle panel: anti-Myc detection of total HERG protein from the same gel. Bottom panel: 14-3-3 co-immunoprecipitated with HERG. (B) Recombinant 14-3-3 partially shields the removal of in vitro PKA-mediated ³²P labeling of HERG by 10 U of protein phosphatase 1-α (PP1) in vitro. Top panel: autoradiography signal from immunoprecipitated HERG phosphorylated with PKA and [³²P]ATP. Bottom panel: anti-Myc detection of total HERG protein. R18 was added at 1 µM. (C) PP1-mediated dephosphorylation of intracellularly ³²P- labeled HERG with (open circles) or without (closed circles) overexpressed 14-3-3. The specific activity of [³²P]HERG was obtained as the ratio of the densitometry signal from autoradiography and anti-HERG immunoblot. Data were normalized to display relative changes with time.