Reduced Delayed-Rectifier K⁺ Current in the Learning Mutant rutabaga

Abstract

In the Drosophila mutant rutabaga, short-term memory is deficient and intracellular cyclic adenosine monophosphate (cAMP) concentration is reduced. We characterized the delayed-rectifier potassium current (IK_DR) in rutabaga as compared with the wild-type. The conventional whole-cell patch-clamp technique was applied to cultured Drosophila neurons derived from embryonic neuroblasts. IK_DR was smaller in rutabaga (368 ± 11 pA) than in wild-type (541 ± 14 pA) neurons, measured in a Ca²⁺-free solution. IK_DR was clearly activated at ∼0 mV in the two genotypes. IK_DR typically reached its peak within 10–20 msec after the start of the pulse (60 mV). There was no difference in inactivation of IK_DR for wild-type (14 ± 3%) and rutabaga (19 ± 3%). After application of 10 mM TEA, in wild-type, IK_DR was reduced by 46 ± 5%, whereas in rutabaga, IK_DR was reduced by 28 ± 3%. Our results suggest that IK_DR is carried by two different types of channels, one which is TEA-sensitive, whereas the other is TEA-insensitive. Apparently, the TEA-sensitive channel is less expressed in rutabaga neurons than in wild-type neurons. Conceivably, altered neuronal excitability in the rutabaga mutant could disrupt the processing of neural signals necessary for learning and memory.

The Drosophila mutations rutabaga and dunce affect learning and memory due to defects in cAMP metabolism. The dunce mutant has a high intracellular cAMP concentration owing to cAMP-specific phosphodiesterase (PDE) disruption (Byers et al. 1981). The rutabaga mutant has a low intracellular cAMP concentration due to elimination of a calcium/cadmodulin-responsive adenylyl cyclase. The ability of the cyclase catalytic subunit to interact with calcium/cadmodulin may be affected in rutabaga (Livingstone et al. 1984; Levin et al. 1992). dunce and rutabaga, originally identified for affecting learning and memory (Dudai et al. 1976) have been shown to alter synaptic plasticity (Zhong and Wu 1991), disrupt habituation (Engel and Wu 1996), and reduce growth cone motility (Kim and Wu 1996).

cAMP-dependent modulation of ion channels has been shown to modify impulse activity of neurons (Kaczmarek and Kauer 1983). Modulation of neuronal electrical properties can change the operation of neural networks, it may be an important cellular mechanism for activity-dependent conditioning of behavior. Classical conditioning of the Aplysia siphon and tail-withdrawal reflex is believed to rely on presynaptic facilitation (Kandel et al. 1983). However, this form of simple learning, which depends on the synapse, cannot occur in complete isolation from the cell body, because changes in neuronal function can impact synaptic function. K⁺ current is one of the fundamental factors that regulate neuronal excitability (Klee et al. 1995) and neuronal function (Le Masson et al. 1993). Investigation of somal K⁺ current is required to elucidate possible changes in neuronal function of the Drosophila learning mutants. The learning deficit in rutabaga has been demonstrated in the adult fly, nonetheless, the embryonic cell culture system provides an excellent preparation that can be manipulated easily in electrophysiological studies. Furthermore, embryonic neurons can be used to show whether the K⁺ current is altered in early life of the fly.

It has been shown previously that short-term (10 min) treatment with dibutyryl (db) cAMP does not affect neuronal K⁺ current (Alshuaib and Byerly 1996), whereas long-term (2 d) treatment with db-cAMP enhances neuronal K⁺ current (Alshuaib and Mathew 1998). Moreover, neuronal K⁺ current was shown to be greater in dunce than in wild-type neurons (Alshuaib and Mathew 1998). dunce and rutabaga were shown to display altered firing patterns in giant (cleavage-arrested) cultured neurons (Zhao and Wu 1997). However, neuronal K⁺ current of normal embryonic cultures has not been investigated in the rutabaga mutant. In the present study, we compared the delayed-rectifier K⁺ current (IK_DR) in wild-type and rutabaga neurons from normal embryonic cultures. IK_DR was reduced in rutabaga neurons as compared with the wild-type, a defect that can impact neuronal excitability and, ultimately, learning and memory.

RESULTS

Our purpose was to compare the delayed-rectifier K⁺ current between wild-type and rutabaga neurons using the conventional patch-clamp technique. Figure 1 shows neurons typical of those studied in the two genotypes. The diameters of the cells ranged from 4–7 μm, and each cell had one to three neurites. In general, neurons with isolated cell bodies (not in contact with other cells) were chosen for study, because they were easier to describe and to approach with the patch electrode.

Figure 1.

Photographs of cultured embryonic Drosophila neurons from wild-type and rutabaga. Cells were cultured (2 d) in a Drosophila medium, they represent typical examples of neurons that were studied. All photographs were taken using Carl Zeiss bright-field optics at the same magnification (X400). Scale bar (25 μm) applies to the two photographs.

Electrophysiological Properties

The electrophysiological properties of wild-type and rutabaga neurons are shown in Table 1. We calculated the total capacitance (C) for each cell by integrating the capacitive current flowing in response to a 50-mV hyperpolarizing step. The cell capacitance was essentially the same for wild-type (7.83 ± 1.26 pF, n = 60) and rutabaga (8.14 ± 1.12 pF, n = 60) neurons. This indicates that the membrane area of wild-type and rutabaga neurons is similar. There was no difference in resting membrane potential (RMP) between wild-type neurons (79.5 ± 0.1 mV, n = 20) and rutabaga neurons (79.3 ± 0.2 mV, n = 22). The whole-cell resistance (R_in) was measured by stepping the membrane potential from −60 mV to −110 mV and dividing this 50-mV step by the measured current amplitude between 90 and 100 msec. R_in was similar for wild-type (8.02 ± 0.89 GΩ, n = 17) and rutabaga (7.52 ± 0.76 GΩ, n = 22) neurons. This suggests that the specific resistances of wild-type and rutabaga neurons are similar.

Table 1.

Electrophysiological Properties of Wild-Type and rutabaga Neurons

Property	Wild-type	rutabaga
C (pF)	7.83 ± 1.26 (n = 60)	8.14 ± 1.12 (n = 60)
RMP (mV)	79.5 ± 0.1 (n = 20)	79.3 ± 0.2 (n = 22)
Rin (GΩ)	8.02 ± 0.89 (n = 17)	7.52 ± 0.76 (n = 22)

All values are means ± SEM; the means were compared using a two-tailed independent Student's t-test. There was no significant difference between wild-type and rutabaga neurons in any of these properties. 

Comparison of IK_DR in Wild-Type and rutabaga in 6K/0Ca Tris Solution

IK_DR was measured in a Ca-free Drosophila external solution, because inward calcium current can affect IK_DR (Alshuaib et al. 2001). Figure 2 shows typical examples of IK_DR recorded from neurons at potentials from −40 to +60 mV. IK_DR was calculated between 90 and 100 msec (steady state) of the pulse to exclude any possibility of A-current contribution to the measured amplitude. K⁺ currents recorded at +60 mV were smaller in rutabaga neurons (368 ± 11 pA, n = 80) than in wild-type neurons (541 ± 14 pA, n = 83) (P < 0.001) (Combined mean of stocks w23,w24,w25 versus combined mean of stocks r33, r34, r35; see Table 2). The IK_DR phenotype was consistent within wild-type (stocks w23,w24,w25) and within rutabaga (stocks r33, r34, r35) neurons.

Figure 2.

Current-voltage (I-V) relations of wild-type and rutabaga neurons in 6K/0Ca Tris Dros. saline (Ca²⁺-free) and in a Ca²⁺-containing Dros saline. K⁺ currents were recorded from wild-type and rutabaga neurons during six 100-msec clamp potentials between −40 mV (bottom trace) and +60 mV (top trace) in steps of 20 mV. Holding potential was −80 mV. Current traces are from individual neurons. Data points are mean ± SEM. The mean IK_DR was determined at steady-state (90–100 msec of the pulse). In the Ca²⁺-free solution, n = 83 for wild-type; n = 80 for rutabaga. In the Ca²⁺-containing solution, n = 16 for wild-type; n = 17 for rutabaga. The pipette contained potassium aspartate solution.

Table 2.

Delayed-Rectifier K⁺ Current (IK_DR) of Several Wild-Type and rutabaga Fly Stocks

Wild-type IKDR (pA)	rutabaga IKDR (pA)
533 ± 20 pA (stock w23, n = 43)	370 ± 17 pA (stock r33, n = 40)a
559 ± 28 pA (stock w24, n = 20)	353 ± 23 pA (stock r34, n = 20)a
541 ± 28 pA (stock w25, n = 20)	378 ± 22 pA (stock r35, n = 20)a
Combined mean	Combined mean
541 ± 14 pA (n = 83)	368 ± 11 pA (n = 80)a

IK_DR is presented as mean ± SEM. n is the number of neurons from 10–30 different cultures (1–5 neurons per culture). To control the possible accumulation of unidentified autosomal mutations (modifiers) that might contribute to the phenotypes examined, the following stocks were used: rutabaga phenotype; Stock r35, a homozygous rutabaga/rutabaga stock was reconstructed (see Materials and Methods). Stocks r33 and r34, these are copies of the parental rutabaga/rutabaga stock. IK_DR amplitude is essentially the same in the reconstructed rutabaga (r35) and parental rutabaga (r33, r34). Wild-type phenotype; Stock w25, a heterozygous wild-type/rutabaga stock was generated for examination (wild-type rutabaga allele is dominant over mutant rutabaga allele). Stocks w23 and w24, these are copies of the parental wild-type/wild-type stock. IK_DR amplitude is essentially the same in the wild-type/rutabaga stock (w25) and parental wild-type (w23, w24). The means were compared using a two-tailed independent Student's t-test and the difference is presented as follows: ^aSignificantly different from wild-type value in the same row (P <0.001). 

All wild-type and rutabaga neurons displayed the delayed-rectifier (non-inactivating) K⁺ current. However, in almost all neurons, the A-type (inactivating) K⁺ current was not observed even with the voltage protocol (−120 mV prepulse, +20 mV test pulse) that usually elicits this transient current. For this reason, we focused mainly on the delayed-rectifier K⁺ current. We applied a voltage protocol that maximizes the delayed-rectifier K⁺ current and diminishes the A-type K⁺ current (holding potential, −80 mV; test pulses, −40 to +60 mV) (Byerly and Leung 1988; Saito and Wu 1991; Alshuaib and Mathew 1998). IK_DR was clearly activated at 0 mV, but only weakly activated at −20 mV in wild-type and rutabaga neurons. IK_DR typically reached its peak within 10–20 msec after the start of the pulse (60 mV) in both wild-type and rutabaga. The time course of inactivation was quantified by calculating the percentage of the peak current that had inactivated at 100 msec. There was no statistically significant difference in values of IK_DR inactivation, which were 14 ± 3% (n = 18) for wild-type neurons, and 19 ± 3% (n = 19) for rutabaga neurons (60 mV pulse; Fig. 2). The current-voltage relations are shown in Figure 2 for wild-type and rutabaga neurons. The regression coefficient (the slope between −20 mV and +60 mV) is smaller for rutabaga (B = 3.90 pA/mV) than that for wild-type neurons (B = 5.66 pA/mV). At each of the clamp voltages between 0 and +60 mV, the current is significantly smaller in rutabaga than in wild-type neurons.

Effect of External Ca²⁺ on IK_DR

To investigate the effect of Ca²⁺ on potassium current, IK_DR was measured in a Ca²⁺-containing Drosophila solution and the current amplitude was compared with that measured in the 6K/0Ca Tris solution. For wild-type neurons, IK_DR measured in the Ca²⁺-containing solution (399 ± 36 pA, n = 16) was smaller than that measured in the 6K/0Ca Tris solution (541 ± 14 pA, n = 83) (P < 0.001). Similarly, for rutabaga neurons, IK_DR measured in the Ca²⁺-containing solution (236 ± 30 pA, n = 17) was smaller than that measured in the 6K/0Ca Tris solution (368 ± 11 pA, n = 80) (P < 0.001). IK_DR typically reached its peak within 10–20 msec in both wild-type and rutabaga neurons. There was no significant difference in values of IK_DR inactivation, which were 11 ± 5% (n = 4) for wild-type neurons and 20 ± 8% (n = 4 ) for rutabaga neurons (60 mV pulse). Figure 2 shows the I-V relations of IK_DR measured in the Ca²⁺-containing solution for both wild-type and rutabaga neurons. IK_DR is clearly activated at 0 mV, but only weakly activated at −20 mV in both wild-type and rutabaga neurons. At each of the clamp voltages between 0 and +60 mV, the current amplitude was significantly smaller in rutabaga than in wild-type neurons. The regression coefficient (the slope between −20 mV and −60 mV) was smaller for rutabaga neurons (B = 2.35 pA/mV) than that for wild-type neurons (B = 4.08 pA/mV).

Population Studies of Blockage of IK_DR With TEA

IK_DR was measured in a 10 mM TEA-6K/ 0 Ca Tris Drosophila saline (10 min), for both wild-type and rutabaga neurons. Comparing the two genotypes after blockage with TEA, IK_DR amplitude was not significantly different in wild-type (356 ± 44 pA, n = 18) and rutabaga (338 ± 38 pA, n = 27) (Fig. 3). This indicates that the blockage of IK_DR was greater in wild-type than in rutabaga neurons, making IK_DR amplitude similar in wild-type and rutabaga neurons. Nonetheless, the contribution of variability due to sampling cannot be completely excluded in such comparisons of limited samples from the population.

Figure 3.

Comparison of IK_DR in wild-type and rutabaga neurons, population studies. Neurons in the control saline (6K/0Ca Tris) and the 10-mM TEA saline were not the same. The K⁺ currents were elicited by stepping to +60 mV (holding potential was −80 mV). Currents were measured between 90 and 100 msec after the beginning of the pulse. The internal solution was potassium aspartate. Values are means ± SEM. In the control solution, n = 83 for wild-type; n = 80 for rutabaga. In the TEA solution, n = 18 for wild-type; n = 27 for rutabaga. The means were compared using a two-tailed independent Student's t-test and the difference is presented as follows: (*) significant difference between wild-type and rutabaga (P < 0.001); (†) significant difference within the same genotype (P <0.001).

This TEA-blocked IK_DR was compared with the control IK_DR (without TEA) obtained from other cells in 6K/0Ca Tris saline (see above). Within the wild-type genotype, the IK_DR reduction with TEA was statistically significant (356 ± 44 pA Vs. 541 ± 14 pA, P < 0.001). Within the rutabaga genotype, there was no statistically significant difference with TEA (338 ± 38 pA) and without TEA (368 ± 11 pA) (P = 0.303) (Fig. 3).

Single-Cell Studies of Blockage of IK_DR With TEA

The best experimental treatment is obtained when comparing the subject with itself, rather than comparing independent samples. Therefore, to exclude variability due to sampling, we measured IK_DR from the same cell before and after application of 10 mM TEA-6K/0Ca Tris Drosophila saline (10 min). In wild-type, IK_DR was reduced by 46 ± 5% (before TEA, 575 ± 38 pA, n = 14; after TEA, 310 ± 52 pA, n = 14, P < 0.001), whereas in rutabaga, IK_DR was reduced by 28 ± 3% (before TEA, 417 ± 36 pA, n = 14; after TEA, 302 ± 42 pA, n = 14, P = 0.048, marginally significant) (see Fig. 4).

Figure 4.

Comparison of TEA sensitivity in wild-type and rutabaga neurons, single-cell studies. The K⁺ current was measured from the same cell, first in the control saline (6K/0 Ca Tris) and then in the 10-mM TEA saline. For wild-type, n = 14; for rutabaga n = 14. The K⁺ currents were elicited by stepping to +60 mV (holding potential was −80 mV). The internal solution was potassium aspartate. In the wild-type neuron, 55% of the K⁺ current was blocked; In the rutabaga neuron, only 30% of the K⁺ current was blocked (values were measured between 90 and 100 msec after the beginning of the pulse).

In these single-cell studies, after IK_DR blockage with TEA, IK_DR amplitude was not significantly different in wild-type (310 ± 52 pA, n = 14) and rutabaga (302 ± 42 pA, n = 14) neurons. This experimental approach confirmed that the blockage of IK_DR was greater in wild-type than in rutabaga neurons, due to a difference in delayed-rectifier K⁺ channels in the two genotypes.

Effect of 4-AP on IK_DR

IK_DR was measured in a 5 mM 4-aminopyridine (4-AP)-6K/0Ca Tris Drosophila saline (10 min), for both wild-type and rutabaga neurons. In the 4-AP saline, IK_DR was smaller in rutabaga (371 ± 56 pA, n = 11) than in wild-type (516 ± 22 pA, n = 13, P < 0.02). These IK_DR amplitudes were not significantly different from those measured in the control 6K/0Ca Tris saline, which were 541 ± 14 pA for wild-type and 368 ± 11 pA for rutabaga (as already mentioned above). Thus, 4-AP did not have any significant effect on IK_DR in both genotypes.

DISCUSSION

Neuronal potassium current has not been characterized completely in rutabaga, a mutation with a low intracellular cAMP concentration due to complete elimination of a calcium/cadmodulin-responsive adenylyl cyclase. The second messenger cAMP has been found to reduce K⁺ current in Aplysia sensory neurons (Siegelbaum et al. 1982) and increase both the Ca²⁺ current (Alshuaib and Byerly 1996) and the K⁺ current (Alshuaib and Mathew 1998) in Drosophila neurons. A larger number of rutabaga larval neurons showed reduction of K⁺ current by 8-bromo-cAMP than wild-type neurons (Yu et al. 1999). In rutabaga larval muscle, pituitary adenylyl cyclase-activating polypeptide (PACAP)-induced enhancement of K⁺ current was abolished (Zhong 1995), whereas IK_DR was unchanged compared with that of the wild-type (Zhong and Wu 1993). Therefore, it was necessary to characterize the K⁺ current in rutabaga neurons.

Reduced IK_DR in rutabaga

The present study shows for the first time that the delayed rectifier K⁺ current is reduced in rutabaga (368 ± 11 pA) as compared with wild-type (541 ± 14 pA) normal culture neurons. This is in agreement with the reduction of K⁺ current reported previously for cleavage-arrested giant rutabaga neurons (Zhao and Wu 1997). The reduced current mainly included a non-inactivating component that was sustained throughout the 100 msec. IK_DR amplitude was the same in the wild-type phenotype (wild-type/rutabaga, stock w25) and parental wild-type (stock w23, stock w24). IK_DR amplitude was the same in the reconstructed rutabaga (stock r35) and parental rutabaga (stock r33, stock r34).

The steady-state inactivation that was determined here (range 6%–20%) is typical of the delayed-rectifier K⁺ current (Saito and Wu 1991; Alshuaib and Mathew 1998). Although the majority of cells did not display the transient A-type current, pharmacological experiments were carried out to test whether 4-AP affects the difference in K⁺ current between the two genotypes. Because IK_DR remained smaller in rutabaga (371 ± 56 pA) than in wild-type (516 ± 22 pA) neurons after 4-AP application, this confirms that the difference between the two genotypes is independent of the transient A-type current. For both wild-type and rutabaga neurons, IK_DR was smaller in the Ca²⁺-containing solution than in the Ca²⁺-free solution. This is probably due to contamination of the outward IK_DR by the inward Ca²⁺ current, and this justifies the measurement of IK_DR in a Ca²⁺-free external solution, which was the approach used in our study. This result is also in agreement with a previous report (Alshuaib and Byerly 1996) that the neuronal K⁺ current in Drosophila does not have a Ca²⁺-dependent component. At any rate, IK_DR reduction in rutabaga is independent of Ca²⁺ effects on the K⁺ current.

IK_DR Reduction is Due to rutabaga Mutation

Our results confirmed that the observed IK_DR was not derived from unidentified second-site mutations other than rutabaga. First, the w25 stock was a result of blending genetic backgrounds from the parental wild-type and rutabaga stocks. Because the w25 embryos were the F1 generation (wild-type/rut¹f ) of a cross of wild-type males with rut¹f females, they are heterozygous at all autosomal loci (bearing one copy from each parental stock). Female w25 embryos are also heterozygous at all X-linked loci. This provides evidence for attributing the physiological phenotype to rutabaga, in spite of the fact that one copy of each autosomal gene was from the rut¹f parental stock, the phenotype was perfectly wild-type (Table 2). In a large sample of embryos, female w25 embryos would be rut¹f/rut⁺f⁺, but male w25 embryos would be rut¹f /Y (mutant physiological phenotype). With the small sample of embryos used (10–30 cultures), it appears that only female embryos were used in our cell culture.

Second, the r35 stock was a result of partial replacement of the genetic background of the parental rutabaga stock. Female rut¹f were crossed with male wild-type (this produced the F1 generation), and then a homozygous rut¹f/rut¹f stock was reconstituted. The perfect accord of the r35 result with the r33 and r34 phenotypes (Table 2), in spite of the mixing of wild-type autosomal alleles, suggests that the reduction of IK_DR is due to the rutabaga mutation. F2 flies were selected for forked. rutabaga is nearly 20 map units from the marker locus forked and crossing over could only have occurred in F1 females (there is ∼0.2 probability of crossing over for each F2 fly selected). Fortunately, the perfect agreement of the r35 embryos with the r33 and r34 phenotype is consistent with avoidance of any crossover recombination between rutabaga and forked.

Mechanism of IK_DR Reduction in rutabaga

The amplitude of IK_DR and the slope of the I-V relation (Fig. 2) were both reduced in rutabaga neurons compared with wild-type neurons. This may be due to a smaller number (density) and/or altered properties of K⁺ channels. With respect to kinetic properties of the K⁺ channel, it is plausible that the low level of intracellular cAMP in rutabaga neurons reduces phosphorylation of the K⁺ channel in the long term, altering the channel open-time, single-channel conductance, or resistance to blockage by TEA. The present study was an investigation of the whole-cell K⁺ current; it demonstrated that IK_DR activation time and steady-state inactivation were unchanged in rutabaga neurons, but clearly, single-channel studies are required to determine any change in kinetics.

With respect to the number of K⁺ channels, we must first specify the type of channel. Our results suggest that IK_DR is carried by two different types of channels, one which is TEA-sensitive, whereas the other is TEA-insensitive. Concerning the TEA-sensitive channel, the single-cell studies of blockage of IK_DR with TEA demonstrated that the percentage of IK_DR blockage is greater in wild-type (46 ± 5%, P < 0.001) than in rutabaga (28 ± 3%, P = 0.048) neurons (Fig. 4). Apparently, the TEA-sensitive channel is less expressed in rutabaga neurons than in wild-type neurons. The low intracellular cAMP concentration in rutabaga may cause a specific reduction in the expression of TEA-sensitive channels. It is also plausible that CREB (cAMP-response element binding protein)-dependent gene expression is altered in rutabaga neurons, and that the expression of the TEA-sensitive channel may be CREB dependent.

Concerning the TEA-insensitive channel, the population studies of IK_DR blockage with TEA demonstrated that IK_DR amplitudes become similar in wild-type (356 ± 44 pA) and rutabaga (338 ± 38 pA) neurons after the block (Fig. 3). Furthermore, the single-cell studies of IK_DR blockage with TEA also demonstrated that IK_DR amplitudes become similar in wild-type (310 ± 52 pA) and rutabaga (302 ± 42 pA) neurons after the block. This suggests that the TEA-insensitive channels are essentially the same in wild-type and rutabaga. In summary, the difference in IK_DR between wild-type and rutabaga neurons may be explained by a smaller number of TEA-sensitive channels in rutabaga. Studies using pharmacological and immunohistochemical tools are required to determine the type and density of channels in rutabaga neuronal membrane.

Functional Implications of IK_DR Reduction in rutabaga

The reduced IK_DR in rutabaga neurons contrasts the increased IK_DR in dunce giant neurons (Alshuaib and Mathew 1998). Altered somal IK_DR in dunce and rutabaga can impact presynaptic facilitation that has been proposed to mediate learning (Kandel et al. 1983). Moreover, this altered IK_DR in dunce and rutabaga is in general agreement with the abnormal spontaneous spikes and altered firing patterns in dunce and rutabaga cleavage-arrested giant neurons (Zhao and Wu 1997). Voltage-dependent K⁺ currents have been shown to be crucial in the regulation of neuronal firing patterns in many species (Hille 1992). For example, blockage of IK_DR with TEA broadened the duration of the action potential and reduced the rate of repolarization (Zhao and Wu 1997). In fact, spontaneous spikes, erratic firing, and long-lasting plateau action potentials were more abundant in rutabaga than in dunce giant neurons. Thus, our results of reduced IK_DR in normal rutabaga neurons coupled with the aberrant firing patterns in giant rutabaga neurons (Zhao and Wu 1997) suggest the possibility of defective frequency coding in rutabaga normal culture neurons. Conceivably, such frequency coding might be important for learning, because neuronal activity is often modulated by previous neuronal activity. In other words, modulation of neuronal excitability may be a mechanism underlying learning and memory.

MATERIALS AND METHODS

Genotypes and Construction of Fly Stocks

The Drosophila melanogaster (Oregon-R) strain was used as control wild-type. The homozygous rutabaga¹f (rut¹f) mutant allele in an Oregon-R genetic background was kindly supplied by Yi Zhong (Cold Spring Harbor Laboratory). f (forked) is a morphological marker. rutabaga is an X-linked recessive, single-gene mutation that was isolated originally using ethyl methane sulfonate (EMS) mutagenesis (Livingstone et al. 1984). To control the possible accumulation of unidentified autosomal mutations (modifiers) that might contribute to the phenotypes examined, we reconstructed a homozygous rut¹f/rut¹f stock (stock r35, below). In addition, a heterozygous wild-type/rut¹f stock (stock w25, below) was generated for examination. The characteristics of the different fly stocks are summarized below.

Wild-Type

Stocks w23 and w24 are two copies of the wild-type/wild-type stock. They have been maintained separately for 13 mo.

Stock w25. This stock of wild-type rutabaga phenotype is a heterozygous stock made by crossing the wild-type and rut¹f stocks. In the parental generation, wild-type (w24) males were crossed with rut¹f (r34) females to produce the F1 generation, wild-type/rut¹f. Embryos were harvested for neuron cultures from the F1 generation (the first generation of progeny). Thus, wild-type/rut¹f embryos were used to study wild-type IK_DR phenotype (wild-type rutabaga allele is dominant over mutant rutabaga allele).

rut¹f

Stocks r33 and r34 are two copies of the rut¹f/rut¹f stock. They have been maintained separately for 13 mo.

Stock r35. This stock of mutant rutabaga phenotype was outcrossed and reconstructed. In the parental generation, wild-type (w23) males were crossed with rut¹f (r33) females to produce the F1 generation. The product of this breed (wild-type/rut¹f ) were crossed with each other to produce the F2 generation. The F2 homozygous flies (rut¹f/rut¹f) were crossed with each other (F2 flies were selected for forked), and embryos were harvested for neuron cultures from the F3 generation.

Preparation of Cultures

Eggs were collected over a 1.5 hr period from Drosophila flies maintained in pint milk bottles at 26°C. Each culture was prepared from the cells of 1–3 gastrulating embryos in a modified Schneider's Drosophila medium (DM) (Salvaterra et al. 1987). Five hours after the beginning of egg collection, the embryos were placed in a 50% ethanol/50% Clorox solution for 2 min to sterilize and dechorionate them. The embryos were then repeatedly washed with DM. Two or three embryos were transferred to a drop of DM on a 35-mm tissue culture dish (Falcon 3001). Each embryo was impaled by a hand-held micropipette (tip diameter, ∼100 μm), the cells were collected by suction, and blown onto the surface of the dish. The cells were further dispersed by repeated passage through the tip of a smaller pipette (tip diameter, 50 μm). The cells adhered to the surface of the dish within minutes of dispersal. The culture dish containing the embryonic cells (in a single drop of DM) was kept in a humid container at room temperature (23°C). All cell cultures were studied electrophysiologically 2 d (43–49 h) later at room temperature. The culture dish was used as the recording chamber with a Sylgard form insert in the dish to confine the extracellular solution to a small volume (0.3 mL). Cells were viewed using Carl Zeiss bright-field optics.

Patch-Clamp Techniques

The conventional whole-cell patch-clamp technique (Hamill et al. 1981) was used to study the membrane currents of neurons. A patch-clamp amplifier measures the membrane current while keeping the membrane potential at a specific level. Electrodes were pulled from 100-μL micropipettes (VWR), coated with Sylgard resin near the tip, and polished to a bubble number (Corey and Stevens 1983) of 3.0–4.0. When filled with potassium aspartate solution, these electrodes had resistances of 6–12 MΩ. The application of patch-clamp to cultured embryonic Drosophila neurons has been described in detail previously (Alshuaib and Byerly 1996). Typically, pipette potential was nulled, gigaohm seal was formed using gentle suction, pipette capacitance was compensated, and the whole-cell configuration was obtained with the application of further suction.

Experiments were performed with an Axopatch 200 A-patch-clamp amplifier (Axon Instruments). Data acquisition and analysis were performed using Digidata 1200 (Axon Instruments) and pCLAMP software (version 5.5, Axon Instruments) on a 486 HP personal computer. Current recordings were filtered (four-pole Bessel) at 5 kHz (capacitative currents) or 1 kHz (ionic currents), and digitized at 20 or 200 μs intervals, respectively. Passive (leakage) currents, determined from negative pulses of one-quarter the amplitude of the test pulse (−P/4), were subtracted from all of the ionic currents. None of the voltages given in the Results have been corrected for the liquid junction potential that existed between the KAsp-filled electrode tip and the Drosophila saline bath solution. The nominal zero potential level is taken to be the level that gives zero current at the beginning of the experiment. By use of a 3 mole/l KCl reference electrode, the junction potential between the KAsp internal solution and the Drosophila saline external solution was −10.5 mV (Byerly and Leung 1988). Thus, the true somal membrane potential was actually 10.5 mV more negative than that given. In addition, there are errors due to uncompensated series resistance of ∼5 mV per 1 nA.

Solutions

K⁺ currents were measured in an external 6K/0Ca Tris Drosophila saline, which contained (in mmole/L) 6 KCl, 10 MgCl₂, 140 TrisHCl, 10 HEPES, and 10 glucose. The pH was adjusted to 7.4 with TrisOH. To test the effect of Ca²⁺ on the K⁺ current, a Ca²⁺-containing external solution was used in other experiments, it contained (in mmole/L) 125 NaCl, 6 KCl, 5 CaCl₂, 5 MgCl₂, 10 HEPES, and 10 glucose. The pH was adjusted to 7.4 with NaOH. For these two external solutions, the osmolarity was ∼296 mOsm /L (measured using a 3 MO micro-osmometer, Advanced Instruments). The external solution was changed during experiments by pipetting 3 mL of the new solution into the 0.3-mL bath; excess solution was removed by a continuous, vacuum-powered exhaust. The pipette internal solution was potassium aspartate, it contained (in mmole/L) 3 KCl, 139 L-aspartic acid, 1 MgCl₂, 10 HEPES, 0.1 CaCl₂, and l EGTA (final Ca²⁺ concentration is ∼10 mmole/L). The internal solution was adjusted to pH 7.3 with KOH (final K⁺ concentration is ∼156 mmole/L). The osmolarity of the internal solution was ∼10% lower than that of the external solution to improve seal formation (Hamill et al. 1981).

Statistical Analysis and Data Presentation

It is possible to detect the effect of a mutation only if it causes a change in membrane current that is large compared with the inherent variability of cultured neurons (Alshuaib and Byerly 1996). Throughout the Results, population data are presented as the mean ± SEM. The means of the two populations were compared by use of a two-tailed Student's t-test for independent samples. A difference was considered statistically significant if the probability that both samples came from the same distribution was at least <5% (P < 0.05). Graphics were generated with Excel (Microsoft) and SigmaPlot (Jandel Scientific) software packages.