Auto-phosphorylation of a voltage-gated K⁺ channel controls non-associative learning

Abstract

Here, we characterize a new K⁺ channel–kinase complex that operates in the metazoan Caenorhabditis elegans to control learning behaviour. This channel is composed of a pore-forming subunit, dubbed KHT-1 (73% homology to human Kv3.1), and the accessory subunit MPS-1, which shows kinase activity. Genetic, biochemical and electrophysiological evidence show that KHT-1 and MPS-1 form a complex in vitro and in native mechanosensory PLM neurons, and that KHT-1 is a substrate for the kinase activity of MPS-1. Behavioural analysis further shows that the kinase activity of MPS-1 is specifically required for habituation to repetitive mechanical stimulation. Thus, worms bearing an inactive MPS-1 variant (D178N) respond normally to touch on the body but do not habituate to repetitive mechanical stimulation such as tapping on the side of the Petri dish. Hence, the phosphorylation status of KHT-1–MPS-1 seems to be linked to distinct behavioural responses. In the non-phosphorylated state the channel is necessary for the normal function of the touch neurons. In the auto-phosphorylated state the channel acts to induce neuronal adaptation to mechanical stimulation. Taken together, these data establish a new mechanism of dynamic regulation of electrical signalling in the nervous system.

Introduction

Voltage-dependent potassium (K⁺) channels (Kv) regulate neuronal excitability by controlling the movement of K⁺ ions across the membrane in response to changes in cell voltage. Modulation of these channels by signalling molecules, such as protein kinases, represents a common mechanism by which neurons adapt their electrical responses to a variety of external events. In this classic picture, K⁺ channels provide the substrate for enzymes and as such they are the passive endpoint of signalling cascades. This wisdom has recently been challenged by the discovery that K⁺ channels might also possess their own enzymatic activity, which is conferred by non-conducting accessory subunits, including mammalian Kvβ2, Drosophila Slob and Caenorhabditis elegans MPS-1 (Zeng et al, 2004; Cai et al, 2005; Weng et al, 2006).

MPS-1, which shows kinase activity, phylogenetically belongs to the KCNE family (Bianchi et al, 2003; Cai et al, 2005). The kcne genes encode integral membrane proteins with a single transmembrane span, which act to modulate the basic attributes of K⁺ channels, including permeation, gating, trafficking, abundance at the plasma membrane and pharmacology, and when genetically mutated, cause congenital and acquired disease (reviewed by McCrossan and Abbott, 2004). There are four KCNE homologues in C. elegans, named mps genes (Park et al, 2005). The founder of the sub-family , MPS-1, shares significant homology with human KCNE2 and KCNE4, suggesting that MPS-1 might be a mosaic of both. MPS-1 is expressed in the nervous system of the worm, including the amphid sensory neurons, touch-sensing neurons (ALM/PLM) and vulva. In the amphid neurons, MPS-1 assembles with the voltage-gated K⁺ channel KVS-1 to form a complex that contributes to both the maintenance and the sensitivity of those cells (Bianchi et al, 2003). By contrast, the partners for MPS-1 in touch neurons and in the vulva are not known. In vitro studies have shown that MPS-1 is able to phosphorylate KVS-1 and that its kinase activity acts to decrease the open probability of the channel (Cai et al, 2005).

The discovery that K⁺ channels are capable of endogenous enzymatic activity prompts the intriguing question as to which are the physiological roles of this new mode of K⁺ channel regulation. To address this fundamental question we investigated the physiological role of the kinase activity of MPS-1. Here we show that MPS-1 assembles with a voltage-gated K⁺ channel, dubbed KHT-1 (for K⁺ channel for Habituation to Tap), in the touch neurons of the worm. Our data provide a model that predicts that the phosphorylation status of KHT-1–MPS-1 is linked to distinct behavioural responses. In the basal, non-phosphorylated state the channel is important for the normal function of the touch neurons. When mechanical stimulation is repeated, such as continuous tapping on the side of the Petri dish, the channel auto-phosphorylates and as a consequence, the animal learns that taps are not important (habituation). Thus, the fact that the kinase activity of MPS-1 is specifically required for habituation behaviour identifies MPS-1 as a true learning-susceptibility gene.

Results

KHT-1 is a candidate partner of MPS-1

MPS-1 does not form functional channels on its own. Therefore, to elucidate the role that MPS-1 plays in mechanosensation, we sought to identify potential pore-forming subunits with which MPS-1 might interact in the touch neurons. We focused on predicted R186.5, here christened KHT-1, as this gene shows high homology to KVS-1 (55%). We cloned KHT-1 using a standard RACE technology, and its cDNA sequence was as predicted. The KHT-1 channel is the homologue of the human Kv3.1 (73% homology, Figure 1), which is abundantly expressed in the brain, where it forms a complex with KCNE3 (Perney et al, 1992; McCrossan et al, 2003). Other likely candidates such as KQT-1, whose human homologue, KCNQ1, assembles with KCNE2 in the stomach and with KCNE1 in the heart and inner ear (Barhanin et al, 1996; Sanguinetti et al, 1996; Schulze-Bahr et al, 1997; Wei et al, 2005; Roepke et al, 2006) or EXP-2, which assembles with MPS-4 in the pharynx (Park and Sesti, 2007), failed to interact with MPS-1 in CHO cells (data not shown).

Figure 1.

KHT-1 is the C. elegans homologue of Kv3.1. KHT-1 and Kv3.1 alignment was calculated by ClustalW software (http://npsa-pbil.ibcp.fr/cgi-bin/npsa_automat.pl?page=/NPSA/npsa_clustalw.html). There is 73% homology. Red, green and blue colours indicate identity, strong similarity and low similarity, respectively.

To determine whether KHT-1 and MPS-1 might form a complex, we initially expressed the two proteins in mammalian (CHO) cells. Thus, KHT-1 yielded robust surface expression (Figure 2A) and co-immunoprecipitated with MPS-1 (Figure 2B). The fact that MPS-1 and KHT-1 can interact prompted the question to as whether MPS-1 can phosphorylate KHT-1. To evaluate KHT-1 phosphorylation, we used anti-phosphoserine/phosphothreonine (anti-phosphoS/T) antibodies as done before (Cai et al, 2005). Figure 2C shows that wild-type MPS-1, but not the D178N MPS-1 mutant, were able to increase anti-S/T staining of KHT-1, suggesting that MPS-1 phosphorylates KHT-1. We conclude that in CHO cells, KHT-1 forms a stable complex with MPS-1 and provides a substrate for the kinase activity of the latter.

Figure 2.

KHT-1 and MPS-1 form a complex in mammalian cells. (A) Immunolocalization of KHT-1 at the surface of CHO cells transfected with KHT-1 cDNA and phase-contrast light transmission image of the field. Images were taken with an Olympus BX61 microscope equipped with Nomarski optics. (B) KHT-1 and c-Myc–MPS-1 epitope tagged to the c-Myc tag in the N-terminus co-immunoprecipitate in CHO cells. Lysates were immunoprecipitated using c-Myc-conjugated beads and western visualization was with anti-KHT-1 antibody. (C) CHO cells transfected with KHT-1–HA alone or with the indicated MPS-1 variants were immunoprecipitated using HA-conjugated beads. Western visualization with rabbit anti-phosphoserine/threonine (S/T) antibodies is shown in the top panel. Western visualization with anti-KHT-1 antibody is shown in the bottom panel.

KHT-1 forms a complex with MPS-1 in the touch neurons

As our experiments indicated that KHT-1 and MPS-1 form a complex in CHO cells, we next investigated the possibility that this complex could exist in vivo. Thus, we sought to identify neurons in which these two proteins are co-expressed, using GFP fusions. In transgenic worms expressing GFP driven by the kht-1 promoter in a kht-1 KO background (P_kht∷gfp), intense GFP fluorescence was visible in the PLM touch neurons (Figure 3A), which also express MPS-1 (Bianchi et al, 2003), suggesting that KHT-1 and MPS-1 might form a complex in these cells. P_kht−1∷gfp also yielded signals in intestine and in several neurons in the head and tail (data not shown), which are currently unidentified. To further explore the notion that MPS-1 and KHT-1 might form a complex in the PLM neurons, we assessed their surface co-localization in primary cells prepared from embryos, as morphological, electrophysiological and GFP reporter studies have shown that the differentiation, morphology and functional properties of cultured cells are similar to those observed in vivo (Christensen et al, 2002; Zhang et al, 2002; Suzuki et al, 2003). We constructed transgenic worms expressing the wild-type mps-1 gene and a mutated gene encoding for the inactivated kinase fused to gfp in the C-terminus in a mps-1 KO background (we name these genotypes, respectively, P_mps-1∷wild type–mps-1∷gfp and P_mps-1∷d178n–mps-1∷gfp, where P_mps-1 is the mps-1 promoter). Both constructs gave strong GFP signals in the cells where MPS-1 is normally expressed, including the PLM neurons (Supplementary Figure S-1A-B), indicating that wild-type and D178N fold and traffic to the plasma membrane normally and that inactivation of the kinase activity of MPS-1 has no effect on its expression. In Figure 3B a representative confocal image of a PLM neuron obtained from P_mps-1∷wild type–mps-1∷gfp embryos is shown. (Nomarski images of the broad field are shown in Supplementary Figure S-2.) In Figure 3C the same cell stained with a custom-made polyclonal antibody raised against KHT-1 is shown. Both subunits were detected in the dendrite and along the perimeter of the cell, indicating surface expression as expected. Most importantly, their signals overlapped (Figure 3D), suggesting that KHT-1 and MPS-1 form a complex in the PLM neuron. (analogous results were obtained with P_mps-1∷d178n–mps-1∷gfp, Supplementary Figure S-3.) To confirm this notion we carried out native co-immunoprecipitations (n=3 experiments). Thus, lysates of P_mps-1∷wild type–mps-1∷gfp and P_mps-1∷d178n–mps-1∷gfp worms were immunoprecipitated with anti-KHT-1 antibody. Western blot analysis with anti-GFP antibody showed a specific band migrating with a molecular weight of ∼52 kDa, which corresponds to the molecular weight of the MPS-1–GFP fusion protein, which was not detected in control samples (Figure 3E). Together these data indicated that KHT-1 and MPS-1 form a stable complex in the PLM neuron.

Figure 3.

KHT-1 and MPS-1 form a complex in native PLM neurons. (A) kht-1 expression in PLM neurons was determined by a transcriptional gene reporter obtained by fusing ∼3.5 kb of the promoter sequence of kht-1 with GFP in a Fire vector pPD95.75. The figure shows side and top views using two distinct worms. The dendrites are indicated by arrows. (B) Confocal image of the plasma membrane of a PLM neuron cultured from embryos of P_mps-1∷wild type–mps-1∷gfp worms. (C) Confocal image of the same neuron in (B) stained with anti-KHT-1. (D) A merged image of the confocal images in (B) and (C). The yellow colour indicates co-localization of KHT-1 and MPS-1 at the plasma membrane. (E) Co-immunoprecipitations of wild type MPS-1–GFP and D178N MPS-1–GFP with KHT-1. Worm lysates were immunoprecipitated with anti-KHT-1 antibody and western visualization was done using anti-GFP antibody. Controls were obtained from precipitates obtained using non-conjugated protein A beads. Images were taken with a Zeiss LSM 510 META confocal microscope and merged by ImageJ software.

The kinase activity of MPS-1 acts to inhibit the KHT-1 current

Next, we characterized the functional properties of the KHT-1 current using the whole-cell configuration of the patch clamp. CHO cells transfected with KHT-1 alone expressed robust, voltage-dependent, rapidly activating K⁺ currents (Figure 4A and B). By progressive substitution of bath potassium with sodium, we found the KHT-1 channel to be selective for potassium, which shifted the reversal potential by kT/e=23.1±1.1 mV, n=11 (Figure 4C). The half-maximal voltage for activation, V_1/2, was 67.2±5.4 mV and the slope, V_s=16.3±1.4 mV, n=27 (Figure 4D). Thus, C. elegans KHT-1 encodes a K⁺-selective channel with properties similar to mammalian channels of the Kv3 family, indicating functional conservation suggested by sequence homology. Next, we characterized the functional effects of MPS-1 on KHT-1. Cells co-transfected with wild-type MPS-1 and KHT-1 expressed currents, which were roughly 5-fold smaller than KHT-1 currents alone (Figure 4E and G). It should be noted that when we replaced wild-type MPS-1 with the constitutively inactive D178N mutant, the magnitude of the current returned to KHT-1-alone levels (Figure 4F and G). This was specifically due to the lack of enzymatic activity in the MPS-1 variant because the D to N mutation does not affect the expression of the protein nor its interactions with pore-forming subunits (Cai et al, 2005), which are mainly controlled by its transmembrane domain (Wang and Sesti, 2007). Neither wild-type MPS-1 nor D178N modified other gating attributes of the complex, such as activation. (Figure 4H, The theoretical line is the Boltzmann fit of KHT-1 alone.) We conclude that MPS-1 is an inhibitory subunit of KHT-1 and that the inhibition of the KHT-1 current is expressly mediated by the kinase activity of MPS-1. It is interesting to observe that the kinase activity of MPS-1 acts similarly to inhibit the current of KVS-1 (Cai et al, 2005), underscoring not only a sequence but also functional homology in the two channels.

Figure 4.

MPS-1 in an inhibitory subunit of KHT-1. (A) Representative whole-cell currents recorded in CHO cells transfected with KHT-1 cDNA. Voltage protocol (inset) was 1 s voltage steps from −80 mV to +120 mV (20 mV increments). (B) Current density–voltage (σ–V) relationships for KHT-1 channels alone expressed in CHO cells. n=27. (C) Reversal potential of KHT-1 currents alone on isotonic replacement of sodium with various external potassium levels. Data fit to a Nernst function (equation (2)) gave kT/e=23.1±1.1. n=11 cells. (D) Normalized macroscopic conductances, G/G_Max (equation (3)) for KHT-1 channels alone. The theoretical line was constructed by fitting the data to the Boltzmann function (equation (4)) with 65.4 mV and V_s=16.1 mV. n=27 cells. (E) Representative whole-cell currents expressed in CHO cells transfected with KHT-1 and wild-type MPS-1 cDNA. (F) Representative whole-cell currents expressed in CHO cells transfected with KHT-1 and D178N MPS-1 cDNA. (G) σ–V relationships for KHT-1-MPS-1 channels (triangles) and KHT-1–D178N channels (circles). n=17 and 18 cells, respectively. The dotted line is the (σ–V) relationship for KHT-1 channels alone. (H) G/G_Max relationships for KHT-1–MPS-1 channels (triangles) and KHT-1–D178N channels (circles). The theoretical line is the Boltzmann fit of the G/G_Max relationship for KHT-1 alone (panel D). Data are presented as mean±standard error of the mean (s.e.m.). Statistically significant differences from control are indicated with ^**P<0.01.

The KHT1–MPS-1 complex is required for mechanosensation

C. elegans. senses gentle touch on the body mainly through two neuron types, the ALM and PLM neurons—located in the middle of the body and in the tail, respectively—both of which express MPS-1 (Driscoll and Kaplan, 1997; Bianchi et al, 2003). Therefore, to probe the physiological role of the kinase activity of MPS-1 in mechanosensation, we constructed transgenic worms expressing the wild-type gene and a mutated gene encoding for the inactivated kinase in an mps-1 null background (we name these genotypes, respectively, P_mps-1∷wild type–mps-1 and P_mps-1∷d178n–mps-1) and assessed how these worms responded to mechanical challenges. By a gentle touch on the head or tail the worm responds by moving backward or forward, respectively. However, when the mechanical stimulus is repeated, for example by rhythmically tapping on the Petri dish, the worm habituates (Rankin et al, 1990; Rankin and Broster, 1992). Thus, we evaluated both the response of the transgenic animals to gentle touch (body-touch phenotype) and their ability to habituate to repetitive mechanical stimulation (tap phenotype). Figure 5A shows that parental worms (N2 genotype) responded nearly 100% of the times to gentle touches on the tail by moving forward. By contrast, the response of mps-1 null worms was significantly diminished in agreement with earlier data, in which mps-1 was inactivated by RNA interference (Bianchi et al, 2003). Also, genetic ablation of kht-1 caused defective forward response (Figure 5A), underscoring the active role played by this channel in mechanosensation. Furthermore, the defective phenotype of the null worms could not be ascribed to abnormal development of the touch neurons because they did not show developmental or morphological abnormalities (Supplementary Figure S-1B-D). The rescue of the wild-type mps-1 gene in the null genotype fully restored mechanosensation, as expected (Figure 5A). It should be noted that inactivation of the kinase activity of MPS-1 (P_mps-1∷d178n-mps-1 genotype) did not affect the ability of the animal to respond to touch on the tail (Figure 5A). Nearly identical responses were recorded when the animals were touched on the head (reverse movement, Figure 5B). We conclude that the kinase activity of MPS-1 is not required for the basal function of the touch neurons. Interestingly, in both kht-1–null and mps-1–null worms forward and reverse movement were equally impaired. This is expected for MPS-1, which is expressed in the ALM and PLM neurons and might suggest that KHT-1 is also expressed and is functional in both cells. The lack of P_kht-1-driven GFP signals in the ALM neurons could be because of several factors—most importantly, the fact that GFP requires high levels of protein expression to yield detectable signals.

Figure 5.

The KHT1–MPS-1 complex is required for mechanosensation. (A) Responses to touches in the tail in parental (N2), mps-1 null, kht-1 null, P_mps-1∷wild type–mps-1 (wild type), P_mps-1∷d178n–mps-1 (d178n) genotypes. The mec-4(d) mutant, which induces necrotic touch-neuron death and therefore is touch insensitive, was used as control strain. n⩾50 worms/genotype. (B) Response to touch on the head in the indicated phenotypes. n⩾50 worms/genotype. Behavioural tests were carried out without knowledge of the worms' genotype. Data are presented as mean±s.e.m. Statistically significant differences from control are indicated with ^*P<0.0 and ^**P<0.01.

The kinase activity of MPS-1 is necessary for habituation to tap

On tapping the Petri dish every 5 s, N2 worms and P_mps-1∷wild type–mps-1 worms quickly habituated (Figure 6A and movie in Supplementary data). Similar results were obtained at ISI 2, 10 and 60 s (data not shown), in agreement with earlier reports (Rankin et al, 1990; Rankin and Broster, 1992). Surprisingly, when P_mps-1∷d178n–mps-1 worms were subjected to the same training, they habituated considerably slower (Figure 6A and movie in Supplementary data. 2 s, 10 s and 60 ISI, data not shown) and retained a significant (∼30%) residual response to tap. For quantification, we fitted the habituation coefficient to a single exponential function and used the time constant of the exponential, τ_H, as a measure of the ability of the worms to habituate. Thus, parental and P_mps-1∷wild type–mps-1 genotypes habituated quickly (τ_H values ranging between 2–4 taps) in an ISI-dependent fashion (Figure 6B). In contrast, the number of taps necessary to habituate P_mps-1∷d178n–mps-1 worms was 10-fold higher and fairly independent of the ISI used during training. N2 and P_mps-1∷wild type–mps-1 worms could retain habituation for nearly 1 h (Figure 6C). Recovery was more rapid in worms trained at short (ISI=5 s, Figure 6C) than at long ISI (ISI=60 s, data not shown), ruling out possible effects of fatigue. In contrast, P_mps-1∷d178n–mps-1 worms recovered almost instantaneously (ISI=5 s, Figure 6C; ISI=60 s, data not shown). In reality, this behaviour did not reflect defective retention of habituation training, but rather was a consequence of the inability of these worms to habituate to tap. It is important to emphasize that the inability of the P_mps-1∷d178n–mps-1 genotype to habituate to tap was not caused by defective function of the touch neurons. In fact, we found that, as in other species, habituation to tap in C. elegans arises as a consequence of progressive desensitization to the stimulus (Figure 6D). Thus, the animals harbouring D178N MPS-1 could not habituate because they retained touch sensitivity during tapping. Consistent with this mechanism, when we tried to habituate kht-1 null worms we found that the animals were less receptive to taps than normal worms, as though they were already desensitized (Figure 6E). Notably, mps-1 KO worms showed a similar behaviour (Figure 6F), further suggesting that the two proteins form a complex in the touch neurons. It is likely that the weak response in the null genotypes was a consequence of the fact that both KHT-1 and MPS-1 are required for the normal function of the touch neurons, which therefore do not respond to tap when this current is suppressed.

Figure 6.

The kinase activity of MPS-1 is necessary for habituation to tap. (A) Habituation to tap in N2 (hollow squares), P_mps-1::wild type–mps-1 (filled squares) and P_mps-1::d178n–mps-1 (hollow triangles) animals at ISI=5 s. The theoretical lines were constructed by fitting the data to a single exponential function: where T is the number of taps, A and B are constants and τ_H is the time constant of the exponential. n=7 experiments/genotype. (B) τ_H coefficients as a function of the ISI for N2, P_mps-1::wild type–mps-1 and P_mps-1::d178n–mps-1 animals. Data represent averages of 5–7 experiments/genotype. (C) Recovery from habituation to tap in N2, P_mps-1::wild type–mps-1 and P_mps-1::d178n–mps-1 animals trained at 5 s ISI. Data represent averages of 3–4 experiments/genotype. (D) Body-touch phenotype in N2, P_mps-1::wild type–mps-1 and P_mps-1::d178n–mps-1 animals earlier trained with the indicated number of taps. n=3–5 experiments/genotype. (E) Habituation to tap in the kht-1-null genotype. Dotted line represents N2 worms. ISI=5 s. n=5 experiments. (F) Habituation to tap in the mps-1-null genotype. Dotted line represents N2 worms. ISI=5 s. n=7 experiments. Behavioural tests were carried out without knowledge of the worms' genotype. Data are presented as mean±s.e.m. Statistically significant differences from control are indicated with ^*P<0.0 and ^**P<0.01.

Touch neurons express rapidly inactivating voltage-gated K⁺ currents

To gain insight into the molecular mechanisms underlying the effect of the KHT-1–MPS-1 complex on mechanosensation, we characterized the electrophysiology of native cells cultured from embryos. In agreement with earlier reports (Suzuki et al, 2003), primary touch neurons expressed robust, inactivating currents, which were abolished on substitution of N-methyl-D-glucamine for K⁺ in the pipette solution (data not shown), suggesting that they were mainly carried by K⁺ ions. Representative currents, along with peak and steady-state current density–voltage relationships, are shown in Figure 7A and B. Thus, N2 currents showed half-maximal activation, V_1/2=65.5±5.0 mV, with a slope factor, V_s=19.3±1.8 mV (Figure 7C). Inactivation was moderately voltage-dependent (∼−0.064 ms/mV) and fast (τ∼7 ms at +80 mV, Figure 7D). As expected, rescue of wild-type or d178n mps-1 in the null background gave rise to currents that fully recapitulated the parental currents (Figure 7A–D). No substantial overexpression of MPS-1 occurred in the transgenic worms, consistent with the observation that the body-touch phenotype of P_mps-1∷wild type–mps-1 and P_mps-1∷d178n–mps-1 worms is normal. We also occasionally recorded non-inactivating, slowly activating currents (data not shown) in neurons of all genotypes. These currents had been earlier reported by others (Suzuki et al, 2003), and are probably the result of some degenerative process in the cell.

Figure 7.

The KHT-1–MPS-1 complex contributes to the K⁺ current in the touch neurons. (A) Representative whole-cell currents elicited by single voltage jumps from −80 to +120 mV (inset) in N2, P_mps-1::wild type–mps-1 (wild-type) and P_mps-1::d178n–mps-1 (D178N) touch neurons. Currents were recorded 4 days after seeding. (B) Peak and steady-state σ–V relationships in touch neurons of N2 (hollow squares), P_mps-1::wild type–mps-1 (filled squares) and P_mps-1::d178n–mps-1 (triangles). n=22, 15 and 17 cells, respectively. (C) G/G_Max relationships for N2, P_mps-1::wild type–mps-1 and P_mps-1::d178n–mps-1. Data are from n=22, 15 and 17 cells, respectively. The theoretical line was calculated by fitting the N2 curve to the Boltzmann function (equation (4)) with 65.5±5.0 mV and V_s=19.3±1.8 mV. (D) Inactivation time constants for N2, P_mps-1::wild type–mps-1 and P_mps-1::d178–mps-1 currents. Time constants were calculated by fitting macroscopic currents to a single exponential function (equation (1)). n=22, 15 and 17 cells respectively. Linear fit of the data gave a slope of −0.064±0.002 ms/mV. (E) Representative whole-cell currents elicited by single voltage jumps from −80 to +120 mV in kht-1-null and mps-1-null touch neurons. Currents were recorded 4 days after seeding. (F) Representative KHT-1 current obtained by algebraically subtracting kht-1-null currents from N2 currents. Left, normalized current–voltage (I–V) relationships of native (hollow triangles) and heterologously expressed (filled triangles) KHT-1 currents. n=7 and n=27 cells, respectively. (G) Macroscopic peak and steady-state (hollow) current densities, σ, at +120 mV in N2, kht-1-null and mps-1-null touch neurons. n=22, 18 and 25 cells respectively. (H) Window currents in N2, kht-1-null and mps-1-null touch neurons. Fractional currents, I_{steady state}/I_peak were normalized and fitted to a modified Boltzmann function (equation (5)) with V_1/2=12.5±1.0 mV and V_s=23.2±1.8 mV for N2, V_1/2=6.1±0.5 mV and V_s=19.6±2.0 mV for kht-1-null and −2.6±0.4 mV and V_s=17.6±1.1 mV for mps-1-null. n=22, 18 and 25 cells, respectively. Activation curves were calculated by fitting normalized conductances (G/G_Max) to the Boltzmann function (equation 4) with respectively, V_1/2=65.5±5.0 mV and V_s=19.3±1.8 mV for N2, V_1/2=60.8±5.7 mV and V_s=17.7±1.2 mV for kht-1-null and V_1/2=64.1±3.2 mV and V_s=17.9±1.1 mV for mps-1-null. n=22, 18 and 25 cells, respectively. The total areas of the window currents are, respectively, 2.3, 0.7 and 0.9 mV. Touch neurons (ALM and PLM) were marked by the Pmec-4∷gfp reported, which specifically expresses in these neurons. Data are presented as mean±s.e.m. Statistically significant differences from control are indicated with ^*P<0.0 and ^**P<0.01.

KHT-1 and MPS-1 contribute to the K⁺ current in the touch neurons

Currents expressed in the touch neurons of the kht-1 null genotype are shown in Figure 7E. These currents inactivated, with peak values roughly 20% smaller than those of the N2 current (Figure 7G). However, although a substantial steady-state component remained in native N2 currents after inactivation, in the kht-1 null genotype, this steady-state component was markedly diminished (Figure 7G), suggesting that it was conducted by KHT-1. Therefore, we algebraically isolated the native KHT-1 current, I_KHT-1, by subtracting kht-1 null currents from N2 currents (a representative example is shown in Figure 7F). Native I_KHT-1 seemed qualitatively similar to cloned I_KHT-1, an impression that was corroborated by quantitative analysis. Thus, normalized I–V relationships of native and cloned I_KHT-1 overlapped (Figure 7F). The V_1/2 was 64.4±7.7 mV, n=7 in the native channel (data not shown), a value very close to the V_1/2=67.2±5.1 mV shown by the cloned channel. In addition the slope factors of native and cloned channels, V_s=19.7±1.4 mV, n=7 and V_s=16.5±0.7 mV, respectively, were comparable, indicating that the native and the cloned channel show similar voltage dependence. Thus, we conclude that KHT-1 conducts a steady-state K⁺ current in the touch neurons. The electrophysiological analysis further strengthened the notion that KHT-1 is expressed in both ALM and PLM neurons because we did not record N2-like currents in cells of this genotype. If KHT-1 was expressed only in the PLM neurons, roughly 50% of the currents should have shown N2 characteristics. Next, we analysed the effect of ablating mps-1 on the electrophysiological properties of the touch-neuron current. Representative mps-1 null currents are shown in Figure 7E. Like in the kht-1 null genotype, genetic ablation of mps-1 suppressed the steady-state component of the current and moderately reduced the amplitude of the peak current (Figure 7G). Taken together, these data indicate that genetic ablation of kht-1 or mps-1 modifies the touch-neuron current in a similar manner. This suggests that MPS-1 is an essential component of the complex, which therefore cannot operate properly without MPS-1. This explains why these genotypes show similar mechanosensory defects.

The lack of KHT-1–MPS-1 decreases K⁺ flux

A decrease in the steady-state current is expected to reduce the net flux of K⁺ ions during depolarization and therefore to affect the neuronal output of the touch neuron. To quantify this effect we calculated the overlap in the relationships for channel activation and inactivation, referred to as the window current. The inactivation curve was given by the fractional current, (I_{steady state}/I_peak). The activation curve was given by the normalized peak conductance, G/G_Max. The window currents for N2, kht-1-null and mps-1-null genotypes are shown in Figure 7H; the area under the intersection of the activation/inactivation curves gives an estimate of the flux of potassium across the membrane during depolarization. The area was significantly reduced in both the kht-1-null and mps-1-null genotypes compared with the N2 genotype (0.7 and 0.9 mV versus 2.3 mV). Together, these data suggest that genetic ablation of kht-1 or mps-1 results in decreased K⁺ flux and therefore, acts to hinder touch-neuron repolarization. This mechanism provides a likely explanation to as why kht-1-null and mps-1-null worms are touch defective.

Discussion

In this study we find that the bifunctional accessory subunit MPS-1 assembles with the pore-forming subunit KHT-1 in the touch neurons of C. elegans to form a complex that plays a crucial role in the functioning of these cells. KHT-1 and MPS-1 co-immunoprecipitated in vivo and in vitro and co-localized in vivo. Behavioural and electrophysiological analyses showed that genetic ablation of these genes caused similar mechanosensory defects and affected the current in the touch neurons in a similar manner. This suggests that the KHT-1–MPS-1 complex is functional in vivo and further, that it is assembled in the early stages of the secretory pathway, so that the lack of any subunit is sufficient to destabilize the entire complex. This seems to be a general property of the MPS and the KCNE proteins (Bianchi et al, 2003; McCrossan et al, 2003; Chandrasekhar et al, 2006; Wang and Sesti, 2007).

The observation that transgenic worms bearing a inactivated kinase variant (D178N) retain normal touch sensitivity but habituate to taps only partially and slowly underscores the specific role of the kinase activity of MPS-1 in controlling this form of non-associative learning. Consistent with the fact that P_mps-1∷d178n–mps-1 worms, as well as P_mps-1∷wild type–mps-1 worms, show normal sensitivity to touch, the K⁺ currents expressed in the touch neurons of these genotypes recapitulated parental currents. Unfortunately, the small neurons and hydrostatic skeleton of C. elegans make physiological studies technically demanding, especially in live animals. For this reason we could not directly measure the KHT-1–MPS-1 current in habituated touch neurons. However, the evidence presented here provides a model that predicts that MPS-1 is inactive in the basal state and that repetitive stimulation initiates a cascade of events that eventually leads to auto-phosphorylation of KHT-1–MPS-1. This might imply that other, currently unidentified proteins might act to regulate the timing by which the kinase activity of MPS-1 is switched on, to phosphorylate KHT-1.

Inhibition of the KHT-1–MPS-1 current through auto-phosphorylation induced by repetitive mechanical stimulation or through genetic ablation of kht-1 or mps-1 seems to be a central step to induce touch insensitivity and, therefore, habituation. Suzuki and collaborators showed that repetitive mechanical stimulation caused a progressive decrease in the electrical response of the touch neurons by a reduced Ca²⁺ influx conducted by the EGL-19 channel (Suzuki et al, 2003). In human heart, longer action-potential duration and slower [Ca²⁺]_i decline slow the recovery from inactivation of the L-type Cav1.2 calcium channel, thereby decreasing calcium current availability (Altamirano and Bers, 2007). It is plausible that the KHT-1–MPS-1 complex might similarly induce cumulative inactivation in the EGL-19 calcium channel—which, incidentally is the homologue of the human cardiac channel (90% homology to Cav1.2)—by prolonging the duration of the mechanoreceptor potential. Thus, auto-phosphorylation of the KHT-1–MPS-1 complex would diminish K⁺ flux, delay touch-neuron repolarization, induce cumulative inactivation in the EGL-19 calcium channel and as a consequence, dampen touch-neuron excitability.

Modification of K⁺ channels by protein kinases is a mechanism contributing to learning and memory formation in vertebrate and invertebrate organisms (Nestler and Greengard, 1983; Cohen, 1989; Krebs, 1994; Levitan, 1999; Biron et al, 2006; Tomioka et al, 2006). For example, in mammals, phosphorylation of Kv4.2 by protein kinase A, protein kinase C and of mitogen-activated protein kinase can modulate action-potential back propagation—a mechanism that acts to enhance the long-lasting synaptic plasticity (Magee and Johnston, 1997; Markram et al, 1997; Hoffman and Johnston, 1998; Johnston et al, 1999)—by decreasing channel's activity. In invertebrates, conditioning of the phototactic response in Hermissenda crassicornis has been linked to phosphorylation-mediated changes in K⁺ channel activity by protein kinase C and mitogen-activated protein kinase (Crow and Alkon, 1978; Farley and Alkon, 1980; Alkon et al, 1983; Farley and Auerbach, 1986; Neary and Alkon, 1986; Farley and Schuman, 1991; Etcheberrigaray et al, 1992; Crow and Forrester, 1993; Crow et al, 1998). There is evidence that in Drosophila neural and behavioural plasticity might be mediated by modulation of EAG by calcium/calmodulin-dependent protein kinase II kinase (Griffith et al, 1994; Yao and Wu, 2001). These physiological processes share a common molecular mechanism: enhanced neuronal excitability through kinase-mediated decrease of a K⁺ current. Hence, significant similarity might exist between the role of K⁺ channels in determining non-associative learning in C. elegans and in other organisms, including mammals.

In conclusion, these data underscore a marked level of conservation between mammals and invertebrates in the mechanisms causing habituation. Considering that both KHT-1 and MPS-1 have human homologues that interact in vivo (Kv3.1–KCNE3) and in vitro (Kv3.1–KCNE2/3 McCrossan et al, 2003; Lewis et al, 2004), we speculate that the KHT-1–MPS-1 complex might represent a precursor of mammalian systems in which the K⁺ channel and the protein kinase have evolved to become separate entities.

Materials and methods

Strains

Strain used were Bristol (N2), mps-1(ok1376) (mps-1 null; outcrossed four times), shw-3(ok1884) (kht-1 null, outcrossed 4 times) and mec-4(u231) (mec-4(d)). We constructed mps-1(ok1376)(wild type)(myo-2∷gfp), termed P_mps-1∷wild type–mps-1, P_mps-1∷mps-1(ok1376)(d178n)(myo-2∷gfp), termed P_mps-1∷d178n–mps-1, mps-1(ok1376)(wild type∷gfp), termed P_mps-1∷wild type–mps1∷gfp, mps-1(ok1376)(d178n∷gfp), termed P_mps-1∷d178n–mps-1∷gfp, mps-1(ok1376)(mec-4∷gfp) and shw-3(ok1884)(mec-4∷gfp).

Molecular biology

The wild-type mps-1 gene plus its promoter (P_mps-1∷wild type–mps-1, 3836 bp) were amplified from genomic DNA by PCR using the following primers: 5′-CCCAAGCTTGAATGTGGCTGCTCAATCGAAGGTACCC-3′ and 5′-CGCGGATCCGTCACGTCTAAGCTAAATGATTTATCGTCA-3′ and were inserted in the Fire vector pPD95.75 using Hind III and BamHI restriction sites. The P_mps-1∷d178n–mps-1 construct was constructed by PCR using the wild-type construct as template. To obtain the P_mps-1∷wild type–mps-1∷gfp and P_mps-1∷d178n–mps-1∷gfp constructs, the STOP codon in the wild-type and d178n constructs was mutated to glutamine by PCR, in-frame with the start codon of gfp.

Cloning of KHT-1 was carried out with a Smart Race kit (Clontech) using poly(A)⁺ mRNA extracted from total C. elegans RNA with the Oligotex kit (Qiagen). The primer for KHT-1 5′-RACE was TCTGTACTGCGTGCCCACTTCTAAACATTCT and that for 3′-RACE was AGTCATCCTCCCAACCAAATCGTTTGTACAGT. cDNA was amplified by PCR and inserted in pCI–neo vector (Promega) for functional expression in CHO cells. The full sequence of KHT-1 was communicated to GenBank, which issued the accession number DQ185514. KHT-1–HA was epitope tagged by replacing the terminal stop codon with nucleotides encoding hemagglutinin (HA) residues (YPYDVPDYA-STOP). MPS-1 was tagged by inserting the c-Myc sequence (ISMEQKLISEEDLN) (Wang and Sesti, 2007). The constructs were subcloned into pcI–neo vector (Promega) for expression in CHO cells. All sequences were confirmed by automated DNA sequencing. Transcripts were quantified with spectroscopy and compared with control samples separated by agarose gel electrophoresis stained with ethidium bromide.

Construction of transgenic animals

For transgenic expression of P_mps-1∷wild type–mps-1 and P_mps-1∷d178n–mps-1, the constructs were injected into the syncitial gonads of adult mps-1(ok1376) hermaphrodites. Transformant lines each for P_mps-1∷wild–type-mps-1 and P_mps-1∷d178n–mps-1 were stabilized by a mutagenesis-induced integration into a chromosome, by irradiating 40 animals with 4000 rads of γ-ray for 40 min. The progeny were checked for 100% transmission of the marker (myo-2∷gfp) and also for the presence of the transgene by PCR amplification. Two lines for mps-1(ok1376)(wild type)(myo-2∷gfp) (#8 and #9) and two lines for mps-1(ok1376)(d178n)(myo-2∷gfp) (#4 and #8) were outcrossed four times. The data presented in this study were obtained from lines #8 and #4, respectively (all four lines gave similar results in gentle body touch assays). For electrophysiology lines #8 and #4, mps-1(ok1376) and shw-3(ok1884) were outcrossed with Pmec-4∷gfp transgenic worms that express GFP in the touch neurons under the mec-4 promoter, and the progenies were selected for homozygous transmission of the transgenes. For construction of P_mps-1∷wild-type–mps-1∷gfp and P_mps-1∷d178n–mps-1∷gfp, we proceeded as described above, except that lines were not integrated.

Immunoprecipitations and co-immunoprecipitations

CHO cells. For immunoprecipitations, CHO cells were plated/transfected with cDNA as described before (Bianchi et al, 2003) and used 24–36 h post transfection. CHO cells were washed with 10 ml ice-cold PBS and lysed with ∼2 ml ice-cold RIPA buffer (50 mM TRIS pH 7.4, 150 mM NaCl, 1 mM EDTA, 1% IGEPAL CA-630, 0.5% (w/v) deoxycholate, 0.1% (w/v) SDS, freshly-added 10 mM iodoacetamide, phosphatase and protease inhibitors) for 30 min at 4 °C. Cell lysates were centrifuged for 60 min at 4 °C and the supernatant was mixed with HA-conjugated beads (Roche) and rocked at 4 °C for 3 h. Beads were washed thrice with ice-cold TBST and incubated in SDS sample buffer at ∼90–95 °C for 15 min.

For co-immunoprecipitation, cells were lysed with 1% NP-40 buffer, 50 mM TRIS pH 7.4, 150 mM NaCl, 1 mM EDTA, 1% IGEPAL CA-630, phosphatase and protease inhibitors (Calbiochem). Cell lysates were centrifuged for 60 min at 4 °C and the supernatant mixed with c-Myc-conjugated beads and rocked at 4 °C for 3 h. Beads were washed thrice with ice-cold 1% NP-40 buffer and then incubated in SDS sample buffer at ∼90–95 °C for 15 min.

Native. P_mps-1∷wild type–mps-1∷gfp and P_mps-1∷d178n–mps-1∷gfp worms were harvested and collected in 30% sucrose at 1000 rpm for 5 min. The pelleted worms were washed 2 times in cold water in a 15 ml conical Falcon tube. Worms were lysed in 3 ml lysis buffer (50 mM Tris, 150-mM NaCl, protease inhibitor, 0.1 mM PMSF, pH 7.6) by sonication (2-s pulses with an interval of 30 s chill between two pulses). The lysate was centrifuged at 18 000 rpm for 30 min. The supernatant was removed and the pellet was resuspended in 1% NP-40 buffer (50 mM Tris, 150 mM NaCl, protease inhibitor, 0.1 mM PMSF pH 7.6), rocked at 4 °C for 1 h and then centrifuged at 18 000 rpm for 30 min at 4 °C. The supernatant was incubated with 50 μl protein A beads for 2 h. The beads were then centrifuged at 18 000 rpm at 4 °C and the supernatant was transferred to a fresh tube. In all, 4 μl of anti-KHT-1 antibody (from rabbit) was added to the supernatant and incubated for 3 h. Then 50 μl protein A beads were added, and rocked at 4 °C for 2 h. The beads were collected by centrifugation at 18 000 rpm at 4 °C, washed thrice with NP-40 buffer and then incubated in SDS sample buffer at 90–95 °C for 15 min. Western visualization was carried out with anti-GFP antibody from mouse (Invitrogen).

Behavioural assays

Age synchronization. Nematodes were grown in standard 10 cm NGM plates+OP50 Escherichia coli until a large population of gravid adults was reached (3–5 days). The animals were collected in 50 ml Falcon tubes washed in M9 buffer (22 mM KH₂PO₄, 22 mM NaH₂PO₄, 85 mM NaCl, 1 mM MgSO₄) and treated with 10 volumes of basic hypochlorite solution (0.25 M NaOH, 1% hypochlorite freshly mixed). Worms were incubated at room temperature for 10 min, the eggs (and carcasses) collected by centrifugation at 400 g for 5 min at 4 °C, incubated overnight in M9 buffer and seeded on standard NMG plates.

Behavioural tests were carried out ‘blind' that is the experimenter did not know the worms' genotype:

Gentle body touch. Experiments were carried out as described earlier (Bianchi et al, 2003). Briefly, single worms were picked to individual plates and tested six times for response to light touch to the head and tail with an eyelash. Responses to head/tail touch were recorded as backward/forward movement. The overall response of each group of worms to touch was expressed as the average percentage of times the worms responded.

Habituation to tap. The experiments were carried out by two investigators: one analysed the worm's behaviour under a stereomicroscope (provided with a grid that allowed to distinguish 1/8 worm's length) and delivered the taps, and the other set the ISI and recorded the scores. Individual worms were transferred to the test plate (bacteria free) and allowed to rest for 15–20 min to recover from habituation caused by the transfer. Taps (impulse ∼10 mN s) were delivered manually by raising the Petri dish (2.5 cm) and dropping it. The elevation of the dish was controlled by an L-shaped glass rod, placed vertically at the base of the microscope. The habituation coefficient, HC, was defined as the backward distance travelled by the worm in response to a tap. Large responses (distance⩾1/2 of the animal's length) scored HC=1. Medium responses (1/4<distance <1/2 and/or stops) scored HC=0.5. Distances < 1/4 of the animal's length scored HC=0. N2 and P_mps-1∷d178n–mps-1 worms were also filmed (representative movies are shown in Supplementary data) and analysed using ImageJ software. Taps were delivered by hitting the Petri dish on the side at ISI=5 s. τ_H values were τ_H=3.1±0.7 (n=11) and τ_H=23.1±6.1 (n=13, P<0.021) for N2 and D178N, respectively (data not shown). These values are in good agreement with τ_H values obtained manually.

Recovery from habituation. To evaluate recovery from habituation, the worms were habituated by a train of 30 taps (block 1) and then 2, 5, 10, and 30 min after habituation training the same treatment was repeated (block 2). Training was carried at 5s and 60 s ISI. Recovery was expressed as the percent increase of the average initial HC (calculated by averaging the HC in response to the first and second tap after training).

A standard experiment consisted of three independent tests with groups of 10 or more animals in each group. (This number was generally sufficient to achieve statistical significance because the homogeneous genetic background of the nematodes limits individual behavioural differences.) Individual responses were averaged and s.e.m. calculated.

Electrophysiology

Heterologous expression system. CHO cells were plated/transfected as described before (Bianchi et al, 2003) and used for 24–36 h post transfection. The bath solution consisted of 4 mM KCl, 100 mM NaCl, 10 mM Hepes (pH=7.5 with NaOH), 1.8 mM CaCl₂ and 1.0 mM MgCl₂. Pipette solution consisted of 100 mM KCl, 10 mM Hepes (pH=7.5 with KOH), 1.0 mM MgCl₂, 1.0 mM CaCl₂, and 10 mM EGTA (pH=7.5 with KOH).

Primary cultures. Cultured cells were prepared as described before (Park et al, 2005). Briefly, gravid adult worms were lysed using 0.5 M NaOH and 1% NaOCl. Released eggs were washed thrice with sterile egg buffer containing 118 mM NaCl, 48 mM KCl, 2 mM CaCl₂, 2 mM MgCl₂ and 25 mM Hepes (pH 7.3, 340 mosM), and adult carcasses were separated from washed eggs by centrifugation in sterile 30% sucrose. Eggshells were removed by resuspending pelleted eggs in a sterile egg buffer containing 1 unit/ml chitinase at room temperature for 1.5 h. Embryos were resuspended in L-15 cell culture medium containing 10% foetal bovine serum, 50 units/ml penicillin and 50 μg/ml streptomycin (Sigma) and dissociated by gentle pipetting. Intact embryos, clumps of cells and larvae were removed from the cell suspension by filtration. Dissociated cells were plated on glass cover slips pre-coated with peanut lectin (0.1 mg/ml) dissolved in water. For electrophysiological recordings, bath solution consisted of 145 mM NaCl, 5 mM KCl, 1 mM CaCl₂, 5 mM MgCl₂, 10 mM Hepes/NaOH (pH7.50) and 20 mM D-glucose. The pipette solution contained 125 mM potassium gluconate, 18 mM KCl, 0.7 mM CaCl₂, 2 mM MgCl₂, 2 mM MgATP, 10 mM EGTA/KOH and 10 mM Hepes/KOH (pH 7.5). Currents were repeatedly elicited 3–5 times (with the same voltage stimulus) and digitally averaged online. Leak currents were recorded in the cell-attached configuration before establishing the whole-cell configuration and were digitally subtracted during analysis. Data were recorded with an amplifier Axopatch 200B (Axon) a PC (Dell) and Clampex software (Axon). Data were filtered at f_c=1 kHz and sampled at 2.5 kHz.

Immunolabelling of cultured touch neurons

Cells were washed thrice with PBS and fixed with paraformaldehyde (4% in PBS) for 15 min at room temperature. After fixing, cells were washed thrice for 5 min with PBS. Cells were incubated with 0.1% Triton X-100 in PBS for 5 min, then washed thrice with PBS and blocked for 1 h at room temperature with 5% non-fat dry milk in PBS plus 0.1% Tween 20. Cells were incubated in fresh M9 media with anti-KHT-1 antibody (1:100 dilution) at room temperature for 1.5 h. Cells were then incubated with a Cy3-conjugated anti-rabbit secondary antibody (1:2000, in 5% non-fat dry milk in PBS plus 0.1% Tween 20), for 1.5 h at room temperature, and subsequently washed thrice for 5 min with PBS+0.1% Tween 20. The samples were analysed and photographed by our departmental Zeiss LSM 510 META confocal microscope.

Data analysis

Mathematical functions. The Nernst equation was given by where k is the Boltzmann constant, T the temperature in Kelvin, e is electronic charge and [K⁺]_o and [K⁺]_i are, respectively, the concentration of K⁺ in the bath solution and in the pipette solution.

Macroscopic conductances were calculated as where V and I are, respectively, the test voltage and the corresponding current. V_rev is the theoretical reversal potential calculated by the Nernst function (equation (2)).

The Boltzmann function was given by where V is the test voltage and V_1/2 and V_s are constants. Fractional currents I_steady/I_peak were fitted to a modified Boltzmann function: where A and B are two constants.

Statistical analysis

Quantitative data are presented as mean±s.e.m. The level of significance was calculated using Student's t-test for single comparisons and ANOVA for multiple comparisons, and corrections carried out using a Tukey means comparison test with 95% confidence interval level using OriginPro 7.5 software. Statistical significance was accepted when the level of significance, P, was lesser than 5% (P<0.05).

Supplementary Material

Supplementary Movie S1

Supplementary Movie S2

Supplementary Figure S1

Supplementary Figure S2

Supplementary Figure S3