Graded Regulation of the Kv2.1 Potassium Channel by Variable Phosphorylation

See allHide authors and affiliations

Science  18 Aug 2006:
Vol. 313, Issue 5789, pp. 976-979
DOI: 10.1126/science.1124254


Dynamic modulation of ion channels by phosphorylation underlies neuronal plasticity. The Kv2.1 potassium channel is highly phosphorylated in resting mammalian neurons. Activity-dependent Kv2.1 dephosphorylation by calcineurin induces graded hyperpolarizing shifts in voltage-dependent activation, causing suppression of neuronal excitability. Mass spectrometry–SILAC (stable isotope labeling with amino acids in cell culture) identified 16 Kv2.1 phosphorylation sites, of which 7 were dephosphorylated by calcineurin. Mutation of individual calcineurin-regulated sites to alanine produced incremental shifts mimicking dephosphorylation, whereas mutation to aspartate yielded equivalent resistance to calcineurin. Mutations at multiple sites were additive, showing that variable phosphorylation of Kv2.1 at a large number of sites allows graded activity-dependent regulation of channel gating and neuronal firing properties.

Ion channel phosphorylation is crucial to dynamic regulation of neuronal excitability and the integrated function of the brain (1). Numerous voltage-gated potassium (Kv) channels with distinct properties are expressed in mammalian neurons, where they exert diverse effects on membrane excitability (2). The voltage-gated potassium channel Kv2.1 constitutes the majority of delayed-rectifier potassium currents in most mammalian central neurons (35). However, due to its high threshold for voltage-dependent activation and slow activation kinetics, Kv2.1 does not appear to play a prominent role in the repolarization of single action potentials (4, 5). Rapid calcineurin-dependent dephosphorylation of neuronal Kv2.1 in response to increased intracellular Ca2+ concentration upon excitatory synaptic activity, epileptic seizures, neuromodulatory stimuli, and ischemia leads to graded enhancement of Kv2.1 activity by lowering the threshold for voltage-dependent activation and accelerating activation kinetics (68). As such, Kv2.1 acts a rheostat to homeostatically suppress neuronal firing (9, 10), especially during periods of high-frequency firing (4, 5). Recombinant Kv2.1 expressed in cultured human embryonic kidney 293 (HEK293) cells is similarly regulated by calcineurin (8). The mechanism whereby such graded changes in Kv2.1 function are achieved requires identification and functional analysis of Kv2.1 phosphorylation sites.

Treatment of native Kv2.1 protein isolated from rat brain (6, 11) or cultured neurons (68), or recombinant rat Kv2.1 from transfected HEK293 cells (8), with alkaline phosphatase (AP) results in large (≈30 kD) shifts in electrophoretic mobility, suggesting extensive phosphorylation. More than 20% (132 out of 653) of cytoplasmic amino acids in rat Kv2.1 are serine, threonine, or tyrosine residues, of which up to 60 are predicted as consensus phosphorylation sites. To conduct an unbiased analysis of Kv2.1 phosphorylation sites, we immunopurified recombinant Kv2.1 expressed in HEK293 cells and digested the proteins with trypsin for analysis by liquid chromatography tandem mass spectrometry (LC-MS/MS) (12). One representative example of the obtained LC-MS/MS data is shown in Fig. 1A, for a doubly charged, singly phosphorylated Kv2.1 peptide at a mass to charge ratio (m/z) of 1106.69 that was fragmented to produce a tandem mass spectrum with a nearly complete y-ion series and complementary b-ion series that described the sequence SGFFVEpSPR (amino acids 645 to 653). The phosphorylation site was unambiguously assigned to serine-651 due to mass assignments from β-eliminated y4, y5, and y6 fragment ions and the transition from y2 to y3. Similar LC-MS/MS analyses of other Kv2.1 peptides identified 15 serine residues and 1 threonine residue phosphorylated on recombinant Kv2.1 in intact cells (Fig. 1B and Table 1).

Fig. 1.

Kv2.1 is phosphorylated at multiple sites, a subset of which are regulated by calcineurin. (A) MS/MS spectrum of Kv2.1 phosphopeptide SGFFVEpSPR from control HEK293 cells. (B) Cartoon of approximate location of Kv2.1 phosphorylation sites identified in LC-MS/MS analyses; colored sites are those regulated by calcineurin. (C) The spectral intensity of the S800 and S537 phosphopeptides from one set of experiments reflecting the ratio of Arg0,Lys0-(black)– and Arg6,Lys8-(red)–containing peptides. The pS537 but not the pS800 phosphopeptide exhibits Ca2+-dependent dephosphorylation that is inhibited by pretreatment with FK520. (D) MS/MS spectrum of rat brain Kv2.1 phosphopeptide FSHpSPLASLSSK. Shown is the product ion spectrum of a doubly charged, singly phosphorylated tryptic peptide at m/z = 1340.64. Abbreviations for the amino acid residues are as follows: A, Ala; C, Cys; D, Asp; E, Glu; F, Phe; G, Gly; H, His; I, Ile; K, Lys; L, Leu; M, Met; N, Asn; P, Pro; Q, Gln; R, Arg; S, Ser; T, Thr; V, Val; W, Trp; and Y, Tyr.

Table 1.

Functional characterization of key Kv2.1 phosphorylation sites. G½ is the half-maximal conductance calculated from the conductance-voltage (G-V curve). Vi½ is the half-maximal steady-state inactivation potential calculated from the current-voltage (I-V) curve. ND, not determined. Values in bold are significantly different (P < 0.05) from the respective values for wild-type Kv2.1. Mutation of Ser/Thr to Ala at constitutive phosphorylation sites S492, S499, S516, S586, S795, and T832 did not alter functional phenotype under control, ionomycin-, or AP-treated conditions.

Phosphorylation sites identified by LC-MS/MSDephosphorylation by ionomycin (SILAC)Mutations studiedVoltage-dependent activation and steady-state inactivation parameters of channels
G ½ Vi½ G ½ Vi½ G ½ Vi½
Wild type +16.4 ± 0.6 -26.2 ± 0.4 -10.1 ± 0.6 -58.3 ± 0.5 -19.8 ± 0.7 -60.2 ± 0.8
S563 Yes S563A +0.2 ± 0.5 -43.5 ± 0.4 -9.6 ± 0.4 -59.5 ± 0.7 -18.5 ± 0.8 -56.8 ± 0.8
S563D +15.8 ± 0.8 -28.1 ± 0.7 +5.4 ± 0.3 -43.1 ± 0.4-5.7 ± 0.4-48.3 ± 0.5
S603 Yes S603A +4.3 ± 0.7 -34.8 ± 0.4 -9.3 ± 0.3 -55.4 ± 0.6 -18.9 ± 0.5 -58.4 ± 0.4
S603D +16.7 ± 0.6 -32.4 ± 0.4-4.3 ± 0.6-53.3 ± 0.8-15.3 ± 0.7-54.1 ± 0.5
S537 Yes S537A +5.7 ± 0.4 -36.2 ± 0.5 -9.3 ± 0.6 -58.2 ± 0.6 -19.7 ± 0.3 -59.4 ± 0.7
S537D +16.1 ± 0.7 -29.7 ± 1.0 -3.1 ± 0.5-51.9 ± 0.4-14.1 ± 0.8-53.2 ± 0.4
S715 Yes S715A +5.9 ± 0.4 -35.4 ± 0.5 -10.1 ± 0.5 -60.3 ± 0.8 -20.1 ± 0.7 -58.8 ± 0.6
S715D +15.8 ± 0.4 -26.3 ± 0.3 -2.7 ± 0.5-54.6 ± 0.7-12.1 ± 0.7-51.7 ± 0.6
S651 Yes S651A +15.9 ± 0.6 -26.2 ± 0.4 -9.7 ± 0.5 -59.3 ± 0.7 -20.4 ± 1.2 -59.7 ± 0.8
S453 Yes S453A +7.2 ± 0.7 -28.0 ± 0.3 -10.2 ± 0.3 -58.9 ± 0.4 -19.6 ± 0.4 -59.2 ± 0.5
S453D +15.9 ± 0.4 -26.2 ± 0.4 -3.1 ± 0.6-49.6 ± 0.8-9.2 ± 0.6-52.4 ± 0.5
S11 Yes S11A +16.7 ± 0.5 -42.1 ± 0.7 -9.7 ± 0.4 -59.1 ± 0.6 -20.3 ± 0.6 -61.1 ± 0.7
S11D +16.5 ± 0.6 -27.4 ± 0.5 -8.7 ± 0.9-47.3 ± 0.4 -19.2 ± 0.5 -48.2 ± 0.5
S11A + S453A +4.6 ± 0.5 -39.8 ± 0.7 -10.4 ± 0.6 -59.3 ± 0.5 -19.8 ± 0.5 -60.1 ± 0.4
S563A + S603A -4.3 ± 0.5-48.2 ± 0.6 -10.6 ± 0.6 -59.7 ± 0.6 -19.9 ± 0.5 -60.1 ± 0.7
S453A + S563A + S603A -6.9 ± 0.4-47.3 ± 0.4 -11.3 ± 0.6 -58.9 ± 0.5 -20.7 ± 0.6 -59.5 ± 0.6
S480 ND S480A -0.8 ± 0.5-41.1 ± 0.8-15.7 ± 0.9 -59.8 ± 0.4 -20.1 ± 0.8 -60.1 ± 0.5
S480D +16.2 ± 0.4 -25.9 ± 0.5 -2.4 ± 1.0-42.7 ± 0.5-4.1 ± 0.7-44.6 ± 0.6
S767 ND S767A +8.1 ± 0.7 -33.1 ± 0.8 -7.3 ± 1.1 -54.7 ± 0.8 -17.6 ± 0.5 -58.9 ± 0.8
S767D +15.4 ± 0.8 -26.8 ± 0.4 +0.7 ± 0.8 -48.4 ± 0.6-12.2 ± 0.8-52.1 ± 0.5
S800 No S800A 16.7 ± 0.8 -27.4 ± 0.5 -9.8 ± 0.5 -59.1 ± 0.7 -20.1 ± 0.6 -59.9 ± 0.4

To identify Kv2.1 phosphorylation sites dephosphorylated by calcineurin, we compared phosphorylation of Kv2.1 in control cells with Kv2.1 in cells treated with ionomycin, a Ca2+ ionophore that activates calcineurin (8). Stable isotope labeling with amino acids in cell culture (SILAC) (1315) was used to metabolically label proteins with mass tags, allowing for subsequent analyses of two chemically identical but isotopically distinct samples in a single LC-MS/MS analysis. Transfected HEK293 cells expressing Kv2.1 were metabolically labeled with either normal 12C6-Arg, 12C6, 14N2-Lys (Arg0,Lys0), or isotopic variant 13C6-Arg, 13C6, 15N2-Lys (Arg6,Lys8) amino acids to ensure mass labeling of every tryptic peptide. Arg0,Lys0-labeled cells were left unstimulated, whereas Arg6,Lys8-labeled cells were stimulated with ionomycin. Equal amounts of cell lysates were mixed, and Kv2.1 was immunopurified from the mixed lysates and subjected to in-gel trypsin digestion and LC-MS/MS. The relative intensities of the two isotopically distinct MS peaks directly reflects the amount of the corresponding phosphopeptide in the two different samples.

Using this assay, we found seven phosphorylation sites in Kv2.1 that were dephosphorylated upon calcineurin activation (Table 1), six in the cytoplasmic C terminus and one in the cytoplasmic N terminus (Fig. 1B). Representative mass spectra for two phosphopeptides, NHFESSPLpTPSPK (pS800) and TQpSQPILNTK (pS537), showed Ca2+-dependent dephosphorylation at S537 but not at S800 (i.e., the Lys8-labeled MS peak height changes relative to the Lys0 peak for pS537 but not pS800; Fig. 1C). None of the seven sites exhibiting Ca2+-dependent dephosphorylation were modified in ionophore-stimulated cells pretreated with the calcineurin inhibitor FK520 (Fig. 1C).

To determine how Ca2+/calcineurin-dependent modulation of phosphorylation at these sites regulates voltage-dependent gating of Kv2.1, we expressed isoforms carrying mutated phosphorylation sites and wild-type Kv2.1 in HEK293 cells. Whole-cell patch-clamp was used to analyze Kv2.1 in untreated cells, and cells treated with ionomycin, or with AP dialyzed through the patch pipette. Maximal ionomycin treatment led to FK520-sensitive hyperpolarizing shifts in voltage-dependent activation (≈26 mV) and steady-state inactivation (≈32 mV) gating of wild-type Kv2.1 (Fig. 2 and Table 1). Maximal Ca2+/calcineurin-dependent dephosphorylation of Kv2.1 did not lead to complete dephosphorylation, because AP treatment induced a further shift in electrophoretic mobility (8) and in voltage-dependent gating (by an additional 9 mV; Fig. 2 and Table 1). Serine-to-alanine point mutations in four of the seven calcineurin-regulated Kv2.1 phosphorylation sites resulted in significant hyperpolarizing shifts in both voltage-dependent activation and inactivation in control cells (Fig. 2 and Table 1). Mutations at S11 and S453 affected only inactivation and activation, respectively; the corresponding double mutations exhibited additive effects (Table 1). Mutations at two identified phosphorylation sites (S480, S767) that were not detected in SILAC experiments also yielded shifts in both activation and inactivation gating (Table 1). Treatment with ionomycin or AP resulted in hyperpolarizing shifts in voltage-dependent activation and inactivation of each of these mutants to the same endpoints as wild-type Kv2.1, demonstrating that the effects of the mutations were solely through altered phosphorylation state, and not through secondary effects on overall structure (Table 1). Substitution of phosphoacceptor serine with phosphomimetic and phosphatase-resistant aspartate at these sites yielded mutant channels with wild-type gating characteristics in untreated cells, but with decreased sensitivity to ionomycin- and AP-induced effects (Table 1). The magnitude of the effect of each serine-to-alanine mutation on gating in untreated cells, and of serine-to-aspartate mutations in calcineurin- or AP-treated cells, was similar, suggesting that each phosphorylation site makes a characteristic contribution to gating (Table 1). Point mutations mimicking dephosphorylation (e.g., S563A), as well as calcineurin- and AP-mediated dephosphorylation of wild-type Kv2.1, also led to faster kinetics of channel activation (Fig. 2B, inset) and inactivation (Fig. 2C, inset). Ionomycin treatment of cells expressing S480A caused a hyperpolarizing shift in voltage-dependent gating ≈6 mV greater than that obtained upon ionomycin stimulation of wild-type Kv2.1 and all other mutants, to an endpoint similar to that achieved with AP treatment. This suggests that the S480 phosphorylation site singularly distinguishes dephosphorylation catalyzed by endogenous calcineurin and exogenous AP. Serine-to-alanine mutations at the seven phosphorylation sites refractory to Ca2+/calcineurin-dependent modulation yielded, in each case, channels that exhibited wild-type activation and inactivation gating. That mutations at calcineurin-sensitive sites yielded functional effects, whereas those at calcineurin-refractory sites did not, suggests that calcineurin dephosphorylation selectively targets sites affecting voltage-dependent gating of Kv2.1. These data point to complex in vivo modulation of Kv2.1 function by changes in phosphorylation at multiple functionally distinct sites, each with a characteristic contribution to gating. Analyses of one double mutant (S563A/S603A) and one triple mutant (S453A/S563A/S603A) showed a graded response to dephosphorylation at multiple sites, yielding channels that represent a larger shift than obtained with any single point mutant (Table 1), but smaller than predicted from the simple additive effects of the individual mutations.

Fig. 2.

Analysis of Kv2.1 phosphorylation sites. (A) Whole-cell voltage-clamp traces of Kv2.1 currents in HEK293 cells expressing wild-type Kv2.1 (Kv2.1-WT) without or with ionomycin treatment (1 μM for 15 min) or intracellular dialysis of alkaline phosphatase (+AP; 100 U/ml for 30 min), and of the S563A phosphorylation site point mutant under control conditions. From a holding potential of –100 mV, the cells were depolarized for 200 ms in 10-mV increments to +80 mV. (B) Conductance-voltage (G-V) relationships as a function of voltage-dependent activation and (C) voltage-dependent steady-state inactivation relationships of wild-type (WT) and S563A mutant Kv2.1 in HEK cells without or with the drug treatments as described in (A) (18). The half-maximal conductance (G½) and steady-state inactivation (Vi½) potentials are detailed in Table 1. Current traces for the comparison of activation and steady-state inactivation kinetics, respectively, of WT and S563A mutant Kv2.1 in HEK cells without or with the drug treatments as given in (A) are shown in the insets in (B) and (C). (D) Immunoblot analyses performed against extracts from HEK293 cells expressing WT Kv2.1, and the respective phosphorylation-site mutants. We normalized extracts by comparing total Kv2.1 immunoreactivity of the immunoblot samples using the general Kv2.1-specific antibody KC. (E) Immunoblot analysis of Kv2.1 levels in input and products of immunoprecipitation reactions performed on control and AP-treated rat brain membranes with the indicated phosphospecific antibodies and the general Kv2.1-specific antibody KC.

To determine whether the sites identified as being phosphorylated on Kv2.1 expressed in HEK293 cells were also phosphorylated on native Kv2.1 expressed in rat brain, we generated phosphospecific antibodies against four of the calcineurin-regulated phosphorylation sites (S453, S563, S603, S715) identified above. Phosphospecific antibodies were affinity purified and validated by immunoblot analyses of extracts from HEK293 cells expressing wild-type Kv2.1 and phosphorylation site mutants (Fig. 2D). Each of the phosphospecific antibodies efficiently immunoprecipitated Kv2.1 protein from control rat brain membranes but not AP-treated brain membranes (Fig. 2E), demonstrating that these sites are phosphorylated on rat brain Kv2.1 in vivo. LC-MS/MS analysis (Fig. 1D) of Kv2.1 immunopurified from rat brain revealed an additional seven in vivo sites (S11, S480, S499, S516, S537, S651, S800) originally identified on Kv2.1 expressed in HEK293 cells, and it independently confirmed two of the sites already identified with phosphospecific antibodies (S453, S603).

Kv2.1 is extraordinary in (i) its extensive phosphorylation in untreated HEK293 cells and resting neurons, (ii) its graded modulation through stimulus-induced calcineurin-dependent dephosphorylation, and (iii) the large magnitude (≈35 mV) of the effects of maximal dephosphorylation on voltage-dependent activation. Our unbiased LC-MS/MS analyses, combined with SILAC labeling, reveal that the molecular basis for such regulation is phosphorylation at 16 sites under control conditions and targeted dephosphorylation of at least seven of these sites by calcineurin. Mutagenesis studies reveal that each of the seven calcineurin-modulated sites imparts a unique and incremental change in voltage-dependent gating. That the overall effect of AP treatment on voltage-dependent activation (≈36 mV) is smaller than that predicted (≈96 mV) from simply summing the incremental shifts obtained with each mutant, that certain sites (e.g., S453, S11) yield changes in only activation or inactivation gating, and that mutation of multiple sites increases the magnitude of the effects together suggest that phosphorylation-dependent regulation of Kv2.1 function is based on both the number and context of sites, analogous to phosphorylation-dependent regulation of Ets-1 DNA binding (16). The large number of phosphorylation sites, and the unique individual contribution from sites that can be dephosphorylated in different combinations, provide an effective mechanism for achieving graded regulation of Kv2.1 gating as observed in neurons and HEK293 cells (68). Variable dephosphorylation of Kv2.1 generates a wide array of function that acts as a sensitive rheostat to couple intracellular Ca2+ levels to neuronal excitability, analogous to direct Ca2+ binding to BK-type potassium channels, and distinct from binary switchlike mechanisms seen in phosphorylation-dependent regulation of other ion channels. That calcineurin is key to this modulation is intriguing given its prominent role in activity-dependent synaptic depression (17), such that calcineurin is a homeostatic suppressor of both evoked and intrinsic neuronal activity.

Supporting Online Material

Materials and Methods


References and Notes

View Abstract

Stay Connected to Science

Navigate This Article