INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Transient potassium currents, also known as A-currents (I_A), are active at subthreshold membrane potentials and help to control the excitability and firing properties of neurons and cardiac myocytes in vertebrate and invertebrate systems (Greenstein et al. 2000; Storm 1990; Tierney and Harris-Warrick 1992; Tsien and Carpenter 1978). I_A plays important roles in shaping spike frequency by regulating the interspike interval, in setting the bursting rate in oscillatory neurons, and in modulating postinhibitory rebound (Connor and Stevens 1971; Greenstein et al. 2000; Kloppenburg et al. 1999; Tierney and Harris-Warrick 1992;). The pyloric network in the stomatogastric ganglion (STG) of the spiny lobster, Panulirus interruptus, contains 14 neurons that organize a rhythmic pumping and filtering movement of the foregut. All the neurons are conditional oscillators. It has been suggested that I_A plays important roles in shaping the different firing patterns of the pyloric neurons (Golowasch et al. 1992; Guckenheimer et al. 1993; Harris-Warrick et al. 1998; Hartline 1979; Tierney and Harris-Warrick 1992). Even subtle changes in the conductance and kinetics of I_A by application of low concentrations of the antagonist 4-aminopyridine (4-AP) or neuromodulators such as dopamine dramatically alter the firing properties of neurons in the network (Harris-Warrick et al. 1995; Kloppenburg et al. 1999; Peck et al. 2001; Tierney and Harris-Warrick 1992).

In Drosophila, two different genes in the Shaker family of voltage-dependent potassium channel genes, shal and shaker, have been shown to encode A-type currents (Jan and Jan 1992, 1997; Pongs 1992). These are equivalent to the vertebrate Kv4 and Kv1 families of genes (Jan and Jan 1997; Salkoff et al. 1992; Wei et al. 1990). We have cloned these genes from P.interruptus and showed that both encode an A-type current in Xenopus oocytes (Baro and Harris-Warrick 1998; Baro et al. 1996a,b). However, several results argue that shal alone encodes I_A in the soma and neurites of pyloric neurons within the STG. First, only shal immunoreactivity is found in the soma and neuropil of pyloric neurons, while shaker immunoreactivity is selectively targeted to the axons of the neurons after they leave the ganglion (Baro et al. 2000). Second, there is a linear relationship between the number of shal transcripts within a neuron and the maximal conductance of its I_A measured from the soma (Baro et al. 1997), and there is a similar linear relationship between shal immunolabeling of the soma and maximal I_A conductance (Baro et al. 2000). Third, shal encodes currents in Xenopus oocytes that more closely resemble the endogenous I_A in pyloric neurons than shaker (Kim et al. 1997, 1998; Baro et al. 2001).

We have also expressed shal RNA by injection into pyloric neurons and obtained an increased current that resembles the endogenous I_A (MacLean et al. 1999, 2003). However, we show here that in a quantitative analysis, the increased I_A is not identical in its detailed biophysical properties to the endogenous I_A. In pyloric dilator (PD) neurons, the shal-expressed A-currents have a faster rate of inactivation, a slower time course of recovery from inactivation, and slightly hyperpolarized voltage dependence of inactivation.

These results raise the question of why the shal-encoded currents in pyloric neurons are different from the endogenous A-currents. In vertebrate neurons and cardiac myocytes, a similar discrepancy has been at least partly resolved by the finding of Kv channel interacting proteins (KChIPs) (An et al. 2000). The KChIPs are a group of EF-hand calcium binding proteins, belonging to the neuronal calcium sensor-recoverin family. Co-expression of KChIP with Kv4 proteins in Xenopus oocytes, or the CHO and HEK 293 cloned cell lines, increased the channel density, slowed the inactivation kinetics, and accelerated the rate of recovery from inactivation of Kv4 channels (An et al. 2000; Decher et al. 2001; Takimoto et al. 2002). Following these studies, Nakamura et al. (2001) showed that frequenin [also called neuronal calcium sensor 1 (NCS-1)], a distant-related protein in the same family, can also regulate Kv4 currents. Similar to KChIP, frequenin can specifically bind to Kv4 channel proteins to increase the surface expression, accelerate the recovery from inactivation, and slow the inactivation kinetics due to the increase in a slowly inactivating component of the current. However, the effect of frequenin on Kv4.2 is Ca²⁺-dependent.

We hypothesized that the difference between endogenous I_A and shal-evoked currents in PD neurons was due to the lack of sufficient regulatory proteins in vivo. To test this, we have co-expressed lobster shal with the human gene, hKChIP1, as well as the lobster frequenin gene, in Xenopus oocytes and in PD neurons in the lobster STG. Co-expression of hKChIP1 with shal altered the voltage inactivation properties as well as the kinetics of inactivation and rate of recovery from inactivation, to make the shal current identical to the endogenous I_A. Co-expression of shal with frequenin produced many of the same changes in shal currents in PD neurons, but did not modify the rate of recovery from inactivation.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

STG dissection and PD cell identification

Pacific spiny lobsters (P. interruptus) were purchased from Don Tomlinson Commercial Fishing (San Diego, CA) and maintained in artificial seawater at 16°C until use. Animals were anesthetized by cooling on ice for 30 min before dissection. The STG was dissected along with its motor nerves and associated commissural and esophageal ganglia (Mulloney and Selverston 1974), and pinned in a dish. The preparation was superfused continuously (3 ml/min) with saline (16°C) containing (in mM) 479 NaCl, 12.8 KCl, 13.7 CaCl₂, 3.9 Na₂SO₄, 10.0 MgSO₄, 2 glucose, and 11.1 Tris base, pH 7.35 (Mulloney and Selverston 1974). Extracellular recordings were made from identified motor nerves using glass suction electrodes. Individual somata were impaled with glass microelectrodes (10-25 M; 3 M KCl). The PD neurons were identified by the 1:1 correspondence of their intracellular action potentials with those recorded extracellularly on the PD motor nerve and by their typical shape of membrane potential oscillations and synaptic inputs (Kloppenburg et al. 1999).

Microinjection of neurons

Shal (L49135, Genbank), Frequenin (AF260780, Genbank), and hKChIP1 (NM_014592, Genbank) RNAs were transcribed in vitro and capped using a T3 mMessage mMachine kit (Ambion, Austin, TX). The transcripts were cleaned using the RNeasy Mini kit (Qiagen, Valencia, CA). Following neuronal identification, PD neurons were injected with an RNA solution containing 0.04% Fast Green using pressure pulses (40 psi, 0.2 Hz) driven by a home-made pressure injector and a pulse generator (Master-8, Jerusalem, Israel). The RNA solution contained 0.4 µg/µl Shal with or without 0.04 µg/µl hKChIP1 or 0.01 µg/µl frequenin. Various ratios of shal:hKChIP1 were tested, from 10:1 to 200:1 by weight, and all produced similar results. PD cells were injected with roughly equivalent amounts of RNA based on the color of the co-injected Fast Green. Fast Green alone was injected into control neurons, which were otherwise treated identically to the RNA-injected neurons. After injection, the ganglion, with attached nerves and the commissural and esophageal ganglia, was incubated in filter-sterilized recording saline without Tris-base, but with 5 mM HEPES, 2 g/l glucose, 50,000 unit/l penicillin, and 50 mg/l streptomycin at 16°C for 48-72 h to allow the expression of the proteins.

Xenopus oocyte expression

Xenopus oocytes were harvested and maintained as previously described (Baro et al. 1996b). Shal cRNA was diluted to a final concentration of 0.2 µg/µl and hKChIP1 to 0.02 µg/µl or frequenin at 0.005 µg/µl. Shal cRNA with or without hKChIP1 or frequenin was injected in 100 nl with a sterile glass microelectrode using a microinjector (NA-1, Sutter Instruments, San Rafael, CA). The oocytes were cultured in sterilized ND96 solution (in mM): 96 NaCl, 2 KCl, 1.8 CaCl₂, 5 HEPES, and 1 MgCl₂ with 1 mM NaPyruvate and 0.1 mg/ml gentamycin at 16°C. Recordings were typically made 24 h later.

Electrophysiology

PD NEURONS. After 2-3 days in organ culture, PD neurons were voltage-clamped using an Axoclamp 2B amplifier driven by pClamp8 software (Axon, Instruments, Foster City, CA). Microelectrodes were filled with 3 M KCl and had a tip resistance 8 M. To isolate PD neurons from most synaptic input and to isolate I_A from most other currents, we superfused the ganglion with saline containing 10⁷ M tetrodotoxin (TTX), 5 × 10⁶ M picrotoxin (PTX), 2 × 10⁴ M CdCl₂, 5 × 10³ M CsCl, and 2 × 10² M tetraethylammonium chloride (TEA). The cells were held at 50 mV. The voltage dependence of activation was measured following a deinactivating prepulse to -120 mV for 400 ms with a series of 400-ms voltage steps from 60 to +30 mV in 10-mV increments. The data were leak subtracted using a P/6 protocol with steps opposite to the sign of activation. A control protocol for activation of non-I_A currents was the same as the activation, but without the deinactivating step to 120 mV. The control protocol currents were digitally subtracted from the activation protocol currents to produce the isolated I_A. To measure the voltage dependence of inactivation, the cell was held at 50 mV, where most of the I_A was inactivated; a series of 1-s prepulses from 130 to 30 mV in 10-mV increments was given to remove the inactivation, followed by a 400-ms test pulse to 20 mV to determine the peak I_A after each prepulse. The data were leak subtracted as described above. To measure the time course of recovery from inactivation, the cell was held at 50 mV and hyperpolarized to -120 mV for increasing step lengths before a test step to +20 mV to measure recovery of the current amplitude.

XENOPUS OOCYTES. The oocytes were voltage clamped using a Geneclamp amplifier driven by Clampex 8.0 software (Axon Instruments). All recordings were made in standard ND96 solutions (Baro et al. 1996b). The A-current was recorded using similar protocols as in PD neurons, except that when recording the voltage dependence of activation and inactivation, the membrane potential was held at 70 mV and the prepulses lasted 1 s to 90 mV for activation and 2 s for inactivation.

CURRENT ANALYSIS. The voltage dependence of activation of I_A was determined by converting the peak current to a peak conductance, g, assuming E_K = 94 mV for oocytes and E_K = 86 mV for PD neurons (Hartline and Graubard 1992). The resulting g/V curve was fitted to a third-order (n = 3 for activation) and first-order (n = 1 for inactivation) Boltzmann equation of the form

(1)

where g_max is the maximal conductance and s is a slope factor. For the activation curve, V_1/2 is the voltage at which half-maximal activation of the individual gating steps occurs, assuming a third-order activation relation (Hodgkin and Huxley 1952). For the inactivation curve, V_1/2 is the voltage at which one-half the channels are inactivated.

(2)

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Injection of shal RNA enhances I_A

In each STG, there are two PDs with almost identical electrophysiological properties. Thus one PD neuron, injected only with Fast Green, can serve as an internal control for the other, RNA-injected PD neuron. After injection, the ganglion, with the appropriate motor nerves and associated commissural and esophageal ganglia attached, was cultured in sterilized HEPES-buffered saline at 16°C for 48-72 h to allow expression of the proteins.

The I_A in control PD neurons after a 48- to 72-h culture had almost identical voltage and kinetic properties as the currents recorded in PD neurons from acutely dissected STG preparations (Baro et al. 1996, 1997) (Table 1; Fig. 1). The amplitude of endogenous I_A was around 300 nA at +20 mV (Fig. 1Ai). Activation of I_A was voltage dependent. The normalized peak conductance/voltage relation was fitted with a third-order Boltzmann relation (Eq.1). The voltage of half-maximal activation of the individual gating particles was around 41 mV (n = 10; Table 1), corresponding to a half-maximal activation of the current itself at -13.6 mV; the slope factor was 20 mV (Fig. 1Bi). The current inactivated with time in a voltage-dependent manner. To determine the voltage dependence of inactivation, prepulses from -130 to -30 mV were given, followed by a step to +20 mV; the peak currents were plotted against the prepulse voltages and fit by a first-order Boltzmann equation (Fig. 1Bii). The V_1/2 for inactivation was -63 mV with a slope factor of 5.1 mV (n = 10; Fig. 1Bii). The kinetics of inactivation for this current was rapid and biphasic, and was fit by Eq. 2. At +20 mV, approximately 52% of the peak current inactivated with a slow time constant of 81 ms, while the remaining current inactivated with fast time constant of 20 ms (Table 1; Fig. 1Aii). The A-currents were mostly inactivated when the membrane potential was held at 50 mV; when the PD neurons were stepped to -120 mV, inactivation was removed with a single exponential time course and the time constant was about 18 ms (Fig. 2).

View larger version (23K): [in this window] [in a new window]   Fig. 1. shal produces large A-currents in PD neurons. Ai: IA measured at +20 mV after a prepulse to -120 mV from control and shal-injected PD neurons. Aii: the currents are scaled to show their difference in inactivation rate. B: plots of g/gmax vs. voltage for activation (Bi) and inactivation (Bii) of A-currents in control () and shal ().

View larger version (22K): [in this window] [in a new window]   Fig. 2. The shal-evoked A-currents in PD neurons have slower recovery from inactivation than endogenous currents. A: current traces from vehicle (Ai) and shal- (Aii) injected PD neurons. The voltage protocol is shown at the top. B: plots of the time course of recovery from inactivation in control () and shal ().

In the PDs that were injected with shal RNA, the amplitude of I_A more than doubled 48-72 h following injection (7.2 µS in the shal-injected PDs vs. 3.4 µS in control PDs; P < 0.01; Fig. 1Ai). However, the newly expressed currents had somewhat different biophysical parameters from the endogenous I_A. The voltage dependence of activation of the exogenous I_A was similar to the endogenous current (Fig. 1Bi), but the voltage of half-inactivation showed a trend to hyperpolarize by 5 mV, and its slope factor was increased by 1.7 mV (P < 0.05; Fig. 1Bii). Although the fast and slow time constants of inactivation did not change, the percentage of current inactivating with _slow decreased from 52% to 36% (P < 0.01; Table 1). As a consequence, the I_A in shal-injected PD neurons inactivated more quickly than control neurons, which is clearly seen when the amplitude-normalized currents are superimposed (Fig. 1Aii). Finally, the rate of recovery from inactivation during a prepulse to 120 mV was significantly slowed by 56% (P < 0.05; Table 1; Fig. 2)

hKChIP1 alters the shal currents expressed in Xenopus oocytes

As described previously by Baro et al. (1996), 24 h after injection into Xenopus oocytes, shal RNA produced a transient A-type current resembling the I_A in pyloric neurons (Fig. 3A; n = 7). The voltage of half-maximal activation of the individual gating particles was around 41 mV (n = 5; Table 1), corresponding to a half-maximal activation of the current itself at -13.6 mV (n = 5); the slope factor was -18 mV. The V_1/2 for inactivation was -73 mV with a slope factor of 7 mV (n = 7; Fig. 3B). The kinetics of inactivation for this current was rapid and biphasic. At +20 mV, approximately two-thirds of the peak current inactivated with a fast time constant of 35 ms, while the remaining current inactivated with a slow time constant of 139 ms (Table 1; Figs. 1A and 3A). The shal currents were mostly inactivated when the membrane potential was held at 50 mV; when the oocytes were stepped to -90 mV, removal of inactivation occurred with a time constant of about 470 ms (Fig. 3C).

View larger version (27K): [in this window] [in a new window]   Fig. 3. hKChIP1 alters the shal currents expressed in Xenopus oocytes. Ai: oocytes were injected with shal or shal + hKChIP1 RNA. Currents were triggered by a 400-ms pulse to +20 mV after a prepulse to -90 mV. Aii: currents in Ai are scaled to show the difference in inactivation rates. B: plots of g/gmax vs. voltage for activation and inactivation of A-currents in shal- () or shal + hKChIP1- () expressing cells. C: plots of the time course of recovery from inactivation at -90 mV from shal- () or shal + hKChIP1- () expressing cells.

Coexpression of shal with hKChIP1 significantly altered the properties of the resulting A-currents in Xenopus oocytes. Compared with the current evoked by shal alone, the maximum conductance of the shal + hKChIP1 current more than doubled (106.9 µS in shal + hKChIP-injected cells vs. 50.8 µS in shal alone; P < 0.01; Fig. 3A). When co-injected with hKChIP1, the voltage of half-maximal activation of shal currents was significantly (P < 0.01) shifted by 7mV in the hyperpolarizing direction, while the voltage of half-inactivation was significantly (P < 0.01) shifted by 6 mV in the depolarizing direction, and its slope factor became slightly steeper (Fig. 3B; Table 1). With regard to the kinetics of inactivation, the slow and fast time constants of inactivation for shal were not statistically different from the shal + hKhIP1 current (P > 0.1). However, the relative contribution of the slowly inactivating component was dramatically increased from 30% to 94% (P < 0.01; Table 1). As a consequence, the overall time course of inactivation slowed down, as can be seen when the currents are normalized to the same amplitude and superimposed (Fig. 3Aii). In addition, co-injection with hKChIP1 almost doubled the rate of recovery from inactivation of the shal current at -90 mV (Fig. 3C; Table 1). All these biophysical changes generated by co-injection of hKChIP1 with shal would increase the effective I_A in the subthreshold voltage range (below -45 mV), which in turn would hyperpolarize the membrane potential and reduce the excitability of the cell. The increased peak conductance, slowed inactivation kinetics, faster recovery from inactivation, and more negative activation voltage induced by co-expression of hKChIP1 with shal are similar to those observed when KChIPs were co-expressed with mammalian Kv4 channels in Chinese hamster ovary (CHO) cells and Xenopus oocytes (An et al. 2000).

Co-expression of hKChIP1 with shal in PD neurons produces an A-current that closely resembles the endogenous A-current

We then co-injected shal RNA with hKChIP1 into lobster PD neurons under the same culture conditions as described above. After 48-72 h, co-injection produced a large I_A compared with the endogenous A-current (Fig. 4Ai). The maximum conductance of the I_A produced by shal + hKChIP injection was significantly different from control values (7.6 ± 2.0 vs. 3.4 ± 0.6 µS; P < 0.01), but was not significantly different from shal alone injection (7.6 vs. 7.2 µS, respectively). However, all of the biophysical parameters of the current were closer to those of the endogenous I_A in PD neurons compared with shal injection only (Fig. 4; Table 1). Co-injection of hKChIP1 with shal did not alter the V_Act significantly (Fig. 4Bi), but it shifted the voltage for half-inactivation back to a more depolarized value with a steeper slope, which was almost identical to the native current values (P > 0.4; Fig. 4Bii; Table 1). Once again, the inactivation rate constants did not change, but the percentage of current inactivating with _s increased from 34% to 45%, not significantly different (P > 0.15) from the native current value of 52%. This caused the enhanced I_A to inactivate at a rate very similar to I_A in control PD neurons when amplitude-normalized and superimposed (Fig. 4Aii). Finally, compared with shal over-expressing neurons, the co-injection significantly reduced the time constant for recovery from inactivation to a value not significantly different from that seen in native currents (Fig. 5). As seen in Table 1, none of the voltage and kinetic parameters of I_A in shal + hKChIP1-injected neurons are significantly different from I_A parameters in the paired control PD neurons. These data show that co-injection of Shal + hKChIP1 RNA can reproduce the endogenous A-currents in PD neurons. Injection of hKChIP1 RNA alone into PD neurons did not significantly alter the amplitude or biophysical parameters of the endogenous I_A (n = 4; data not shown).

View larger version (23K): [in this window] [in a new window]   Fig. 4. Coexpression of shal with hKChIP1 produces an A-current similar to the endogenous A-current in PD neurons. Ai: IA measured at +20 mV after a prepulse to -120 mV from control and shal + hKChIP-injected PD neurons. Aii: currents are scaled to compare their inactivation rate. Bi and Bii: plots of g/gmax vs. voltage for activation (Bi) and inactivation (Bii) of A-currents in control () and shal + hKChIP1 (). The control and shal + hKChIP1 curves are almost completely overlapping.

View larger version (22K): [in this window] [in a new window]   Fig. 5. hKChIP1 accelerates the recovery from inactivation of the shal-encoded A-currents in PD neurons. A: current traces from vehicle (Ai) and shal + hKChIP1- (Aii) injected PD neurons. The voltage protocol is the same as shown in Fig. 2. B: plots of the time course of recovery from inactivation in control () and shal + hKChIP1 (). The control and shal + hKChIP1 curves are almost completely overlapping

Coexpression of shal with lobster frequenin partially restores the properties of A-currents in PD neurons

We used several sets of degenerate primers, based on the sequences of the vertebrate KChIP genes, in reverse transcription-polymerase chain reactions (RT-PCRs) with lobster RNA to try to clone the lobster KChIP homologue, but without success. However, one primer set did yield a fragment of the lobster frequenin gene. This is another protein in the same family of EF-hand calcium-binding proteins as KChIP, and the P.interruptus homologue had already been previously cloned by Jeromin et al. (1999). We co-injected shal and frequenin RNAs into Xenopus oocytes, but frequenin significantly suppressed the expression of shal A-current in these cells (data not shown). However, when we co-expressed shal and frequenin in lobster PD neurons, the resulting increase in I_A was only slightly (and not significantly) suppressed (5.5 µS in Frequenin + shal vs. 6 µS in shal alone injected cells). Frequenin significantly modified the shal-evoked currents to more closely resemble the endogenous I_A (Fig. 6A; Table 1). Compared with the currents evoked by shal alone, the shal + frequenin-evoked current had similar voltage dependence of activation (with a slightly steeper slope for activation) but a more depolarized voltage for half-inactivation (62.1 mV in Frequenin + shal vs. 68 mV shal alone injected cells; P < 0.05; Table 1; Fig. 6Bi). In addition, the currents produced by co-expression of shal and frequenin had slower inactivation kinetics compared with the shal-evoked currents (Fig. 6Ai), due to the larger percentage of slowly inactivating current (56% vs. 36%; Table 1). All these parameters were significantly different from those obtained with shal injection alone (P < 0.05), and not significantly different from the parameters of the endogenous I_A (Fig. 6, Aii and Bii; Table 1). However, unlike hKChIP1, frequenin co-injection failed to accelerate the time constant for release from inactivation of the shal-evoked currents (Fig. 6C). The value of this time constant was around 24 ms, which was somewhat faster but statistically similar to shal-evoked currents (28 ms; Fig. 6Ci; Table 1) and significantly slower than endogenous currents (18 ms; P < 0.05; Fig. 6Cii; Table 1).

View larger version (26K): [in this window] [in a new window]   Fig. 6. Frequenin modifies the shal-evoked currents resemble endogenous A-curents in PD neurons. A: scaled IA currents measured at +20 mV after a prepulse to -120 mV from frequenin + shal or shal (Ai) and frequenin + shal or vehicle (Aii) injected PD neurons to compare their inactivation kinetics. Bi and Bii: plots of g/gmax vs. voltage for activation and inactivation of A-currents in control (), shal () and shal + frequenin (). C: plots of the time course of recovery from inactivation in shal () or shal + frequenin (; Ci) and shal + frequenin () or vehicle (; Cii) injected cells.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Electrophysiological, pharmacological, and modeling studies all show that I_A plays an important role in setting the activity states and firing pattern of the six cell types in the pyloric network (Golowasch 1999; Goldman 2001; Harris-Warrick et al. 1995; Kloppenburg et al. 1999; Tierney and Harris-Warrick 1992). Our earlier studies suggested that lobster shal encodes I_A in pyloric neurons (Baro and Harris-Warrick 1998; Baro et al. 1997, 2000; MacLean et al. 1999). However, when injected by itself into PD neurons, shal RNA produces an A-type current which, although similar to the endogenous I_A, differs in several parameters. The currents generated by shal in PD neurons inactivate significantly more rapidly, recover from inactivation significantly more slowly and have a slightly hyperpolarized V_1/2 for inactivation than the native A-currents.

One possible explanation for the failure of injected shal RNA to reproduce the endogenous I_A in PD neurons is that there are a rate-limiting number of some auxiliary proteins that normally interact with shal proteins in the neurons. A number of auxiliary proteins have been reported that interact with different Shaker family potassium channels and modify their gating, kinetics, and localization in vivo and in vitro (Heinemann et al. 1994; Pongs et al. 1999; Rettig et al. 1994). The recently cloned KChIPs are a group of calcium binding proteins that can interact with vertebrate shal-related Kv4 channels and modify their properties to become more similar to native currents in different cell types (An et al. 2000; Decher et al. 2001; Rosati et al. 2001). For example, hKv4.3 is known to generate the I_to1 current in human ventricular cells, but hKv4.3 alone failed to reproduce this current in heterologous cell lines; in particular, the artificial current had a much slower rate of recovery from inactivation. Co-expression of hKChIP2 with hKv4.3 in Xenopus oocytes produced A-currents much more similar to the ventricular I_to1 (Decher et al. 2001). Thus we hypothesized that co-injection of shal with a KChIP protein into PD neurons would restore normal properties to the shal-evoked I_A.

Our results show that lobster shal can indeed interact with KChIP proteins to restore normal properties of the resulting A-currentin in PD neurons. In fact, all the significant differences between the endogenous I_A and the additional current generated by injection of shal alone are corrected by co-injection of hKChIP1 with shal RNA. Most significantly, the voltage for half-inactivation is shifted in the depolarized direction, the ratio of slowly to rapidly inactivating current is increased, and the rate constant for recovery from inactivation is accelerated compared with the shal-induced current. As a consequence, none of the biophysical parameters of I_A in the shal + hKChIP1-injected neurons are significantly different from the endogenous I_A in the control PD neurons. This demonstrates that the shal current can be modified by auxiliary proteins.

However, despite repeated efforts, we have not yet successfully cloned the lobster KChIP homologue. Using KChIP- specific degenerate primers, we did clone a fragment of the lobster frequenin gene, which encodes another small calcium sensor protein in the same superfamily as KChIPs (Jeromin et al. 1999). Compared with the KChIP proteins, frequenin is involved in a number of different biological processes. It can interact with many proteins, from phosphatidylinositol 4-kinase to Ca²⁺ channels, resulting in modification of exocytosis and Golgi transport in different systems (Hendricks et al. 1999; Koizumi et al. 2002; Pongs et al. 1994; Walch-Solimena and Novick 1999; Wang et al. 2001). The mechanisms underlying many of its functions are still unclear. Recent research showed that frequenin could also modify the properties of the vertebrate Kv4 channel in a similar way as KChIPs (Nakamura et al. 2001).

We first looked at the possible interaction between shal and frequenin by co-injection into Xenopus oocytes. However, frequenin significantly depressed the current evoked by co-injection with shal. This failure might be due to species-specific differences in posttranslational modification of the proteins. For example, Nakamura et al. (2001) showed that the effect of frequenin on Kv4 channels was Ca²⁺-dependent. Therefore it would not be surprising that some posttranslational modification is involved in the interaction between lobster frequenin and shal, which is not efficiently mimicked in Xenopus oocytes.

To avoid inter-species incompatibilities, we decided to co-inject lobster frequenin with lobster shal in the lobster PD neurons. When this was done, the resulting I_A was significantly different from that seen with shal alone, and most biophysical parameters were similar to the control I_A. Specifically, the voltage for half-inactivation and kinetics of inactivation were not significantly different from the control I_A parameters. The only parameter that was not corrected by frequenin was the rate of recovery from inactivation, which remained slower than the endogenous current. We are not sure why this one parameter remained unchanged. One possibility is that multiple auxiliary proteins interact with lobster shal, including frequenin and additional unknown proteins. One of these might be the lobster homologue of KChIP. Nakamura et al. (2001) also showed that when co-expressed with Kv4.2 in oocytes, frequenin had a greater effect on the kinetics of inactivation, whereas KChIP1 had a stronger effect to enhance surface expression and accelerate recovery from inactivation. Patel et al. (2002) also suggested that different protein motifs in KChIP2 affect inactivation and recovery from inactivation of Kv4.3 differently. Therefore these two functions may be separate. In PD neurons, frequenin may only be involved in the modification of inactivation kinetics and voltage dependence of the shal current, while other proteins may shape the recovery from inactivation.

Although we still do not completely understand all of the factors that shape the shal potassium currents in pyloric neurons, our results support our earlier hypothesis that the shal gene encodes I_A in the soma and neuropil of lobster pyloric neurons. We previously used molecular and immunocytochemical methods to show that shal mRNA and protein levels are linearly correlated to the amplitude of I_A and the shal protein is present in appropriate locations in the STG neurons and neuropil (Baro and Harris-Warrick 1998; Baro et al. 1996a,b). Our ability to correctly mimic I_A by co-injection of shal with hKChIP1, and to a significant extent lobster frequenin, provides an important additional step in demonstrating that shal encodes lobster I_A. Our results suggest that shal proteins may interact with frequenin and/or other KChIP-like proteins to produce the natural transient potassium current in pyloric neurons. Our results are also the first to show that frequenin/KChIP proteins can modify shal channels in functional neurons.