Genetic dissection of attractant-induced conductances in Paramecium

Paramecium tetraurelia employ multiple chemosensory pathways in order to move to areas populated by their bacterial prey(Van Houten, 1994). A number of bacterial metabolites are attractant stimuli, including glutamate, folate,biotin, cAMP, acetate and ammonium (Bell et al., 1998; Yang et al.,1997; Van Houten,1994) and, except for ammonium, these stimuli bind to cell surface receptors to initiate the responses. In contrast, the attractant NH₄Cl diffuses as ammonia across the cell membrane, which results in alkalinization, hyperpolarization and altered swimming behavior(Davis et al., 1998). The response of the cells to the application of any of the other stimuli is an immediate decrease in frequency of turning and an increase in swimming speed. The removal of the stimuli results in an increase in turning and decreased speed. The abrupt turns in the Paramecium swimming path are attributed to calcium action potentials from activation of ciliary voltage-gated calcium channels, and consequent change in power stroke of the cilia due to the transiently increased ciliary calcium(Eckert, 1972).

The changes in swimming behavior associated with the application of attractant (on-responses) and removal of attractant (off-responses) are characteristic of relative hyperpolarization and depolarization of the cells(Bonini et al., 1986; Eckert, 1972; Machemer, 1989). Electrophysiological recording from cells confirms that the cells hyperpolarize upon application of attractant stimuli, using concentrations and conditions that would not cause artifacts of junction potentials(Van Houten, 1979; Preston and Van Houten, 1987). The most likely mechanism of on-response hyperpolarization is activation of a K⁺ conductance, and, indeed, a transient K⁺ conductance was identified upon stimulation with glutamate(Preston and Usherwood,1988).

The on-response hyperpolarization in attractants is sustained as long as the stimulus is present, and we attribute this to a calcium plasma membrane pump (PMCA) conductance. We have several indirect lines of evidence to support this attribution (reviewed in Van Houten,1998). Together, the evidence points to a PMCA electrogenic conductance during the on-response to attractants.

The removal of attractant elicits an immediate turn of the cell and slower swimming, both consistent with depolarization in the off-response. A turn is caused by the transient increase in ciliary calcium from voltage-gated calcium channels, and a depolarization brings the membrane potential closer to threshold for action potential, increasing the frequency of turns(Nakaoka and Machemer, 1990). Depolarization is also associated with a decrease in ciliary beat frequency,and slower swimming (Nakaoka and Machemer,1990).

We use the T-maze for an assay of chemoresponse behavior(Van Houten et al., 1982). This simple apparatus allows cells to swim between two arms containing relative attractant and repellent solutions. Cells repeatedly leave and re-enter the attractant, spending the majority of time in the attractant solution. Using computer simulations of the T-maze, we found that the immediate responses of the cells upon crossing the step gradient boundary between attractant and control were important parameters for attraction. As simulated cells enter and distribute between the two arms of the T-maze,decreased turning and smooth fast swimming as the cells leave the control solution and move into attractant, and an immediate turn response as the cells leave attractant and move into the control solution are crucial to successful attraction.

While we know a great deal from others' work on the biophysics of the ciliary beat, we do not know the identities of the ion channels that participate in the chemoresponse to stimuli like acetate and biotin, with the exception of the voltage-gated calcium channels of the cilia. Using Pawn mutants that have no functional voltage-gated calcium channels, we showed that without the ability to generate an abrupt turn the mutants are not normally attracted or repelled (Van Houten,1978). Here we present observations of a group of mutants in P. tetraurelia that have defects in specific conductances in order to dissect the conductances that are responsible for on- and off-responses to two stimuli, acetate and biotin.

Paramecium tetraurelia (Wenrich and Sonneborn) strains 51S (sensitive to killer), Cam 1-1, Cam 11(Kink et al., 1990) Cam 4 and eccentric (XntA)(Ling et al., 1994) (R.R.P.,unpublished data) were grown at 28°C in a wheat-grass medium inoculated with Klebsiella pneumoniae(Sasner and Van Houten,1989).

Solutions used in behavioral assays contained 1 mmol l^-1Ca(OH)₂, 1 mmol l^-1 citric acid and approximately 1.3 mmol l^-1 Tris base in addition to the attractant and control compounds noted. Experiments were conducted at either pH 6.7 or 7.0 with the appropriate cation (K⁺, Na⁺, Ca²⁺ or Mg²⁺) for that particular experiment matched with either acetate,biotin or Cl^- as a control (all chemicals from Sigma Chemical, St Louis, MO, USA, unless otherwise noted). For example, a Na-biotin solution would be tested against a NaCl solution in a behavioral assay.

Membrane currents were measured under two-electrode voltage clamp as described (Preston et al.,1990). Glass electrodes were filled with either 0.5-1.0 mol l^-1 KCl or 0.5-1.0 mol l^-1 CsCl, depending on the current being measured. The cell membrane was voltage clamped near resting potential (-40 mV) and currents were recorded as the contents of the experimental chamber were changed with a perfusion system at a rate of 1 ml s^-1 in a 1 ml bath. Recordings from among 20 cells that are shown in Fig. 1 were selected to illustrate the inward and outward conductances that suggested to us models that we then explored with studies of mutant behaviors.

T-maze assays were conducted as previously described(Van Houten et al., 1982). Paramecia in control solution were placed in a stopcock and allowed to swim freely between the two glass arms of the T-maze apparatus that contained the test and control solutions. After 30 min the stopcock was closed, isolating the test and control arms, and aliquots from each were counted. The number of cells in the test arm was divided by the total number of animals in both arms to yield an index of chemokinesis (I_che). Values greater than 0.5 indicate attraction.

Videotapes of cells entering control or test solution were analyzed with Expert Vision software (version 3.14, Motion Analysis Corp., Santa Rosa, CA,USA) and a modified user program provided by K. Clark and D. L. Nelson. Analysis of swimming speed and turning frequency (% directional change) were conducted as described by Clark and Nelson(Clark and Nelson, 1991).

Cells were incubated in the appropriate buffer for at least 30 min before transfer to the buffer in which they were analyzed. Paramecia were pipetted in 2-3 μl into a 50 μl volume of test solution on a glass slide placed on a StereoZoom 7 dissecting microscope (Bausch and Lomb, Rochester, NY, USA). Cells were just off screen when put on the slide and videotaped using a video camera (Cohu 6410, San Diego, CA, USA) attached to a Sony SLV-R5UC video cassette recorder as they swam into view (∼1 s)(Davis et al., 1998). Cell images were processed using a VP 110 video processor (Motion Analysis Corp.).

We measured conductances under two general voltage clamp conditions: in the absence of K⁺ and Na⁺ and with CsCl electrodes to block K⁺ conductances, and in the presence of K⁺ with KCl electrodes. Cells were superfused with the stimulus acetate or biotin in buffer and both the on- and off-responses were recorded.

Cells show a very small and transient inward conductance of about 1-5 pA upon both addition and removal of acetate (see arrows in representative recordings in Fig. 1A). If the small inward conductances in acetate were carried by Ca²⁺, and if they were sufficiently large, they should activate the Na⁺ or Mg²⁺ currents, hence magnifying any inward conductance elicited by acetate because, in P. tetraurelia, inward Mg²⁺ and Na⁺ conductances are activated by a rise in intracellular Ca²⁺ levels (Preston,1990). However, in the presence of Mg²⁺, the small inward conductances initiated by acetate are not changed. When extracellular K⁺ is present, upon addition of acetate and at times upon its removal, there are small outward conductances of 10-30 pA, which are likely to be K⁺ conductances activated by intracellular Ca²⁺ from I_Ca.

Cells in biotin show a very different pattern of conductances(Fig. 1B). There is a sustained outward conductance upon addition of biotin (about 5-10 pA) and a larger,transient inward conductance of approximately 10 pA upon removal of biotin. This off-response conductance can be enhanced to about 40 pA with the addition of Mg²⁺ to the bath. In Mg²⁺ solutions, even 100 μmol l^-1 biotin, a tenfold lower concentration of biotin than we generally use for behavior tests, can elicit a measurable off-response conductance. K-biotin elicits a transient outward conductance followed by a sustained outward conductance during the on-response. The replacement of K-biotin with KCl (stimulus off) elicits an inward conductance, usually followed by an outward conductance. For K-biotin, two different pH conditions were used because of our observation that K-biotin is a better attractant at low pH (6.7) and, unless otherwise stated, the recording and behavior are measured using biotin solutions of pH 6.7 (see below). Consistent with this observation about behavior and pH, there is little conductance change at relatively high pH (7.3). Since the calcium-activated Na channel carries H⁺ as well as Na⁺, we attribute the inward conductances at low pH in K⁺ to Ca²⁺ activation of the Na⁺channel.

We tested chemoresponse behavior using T-maze assays of mutants that have specific defects. Each mutant has a single site mutation that leads to characterized conductance defects (Table 1). Mutant Cam 11 has a mutation in the gene for calmodulin, and consequently has lost the calcium activated Na⁺conductance (I_Na(Ca))(Saimi, 1986; Kink et al., 1990). Mutant Cam 4 likewise has a mutation in the gene for calmodulin, and has lost a calcium activated K⁺ conductance I_K(Ca,d). The mutation in Cam 1-1's calmodulin gene produces defects in both of the major Ca-dependent K⁺conductances (I_K(Ca,d) and I_K(Ca,h))(Preston et al., 1990). Mutant XntA has lost the calcium activated Mg²⁺ conductance(I_Mg(Ca))and also has a reduced calcium activated Na⁺ current (I_Na(Ca))(Preston and Kung, 1994). In all cases below, the Mann-Whitney U-test was used to determine statistically significant differences from wild type at the P<0.05 level (Tables 1 and 2).

Wild-type cells (51-S) are about equally well attracted to K⁺ and Na⁺ salts of acetate and to K-acetate with Mg²⁺ in the bath (Table 1). Mutant Cam 4, likewise is attracted to K⁺and Na⁺ salts of acetate, indicating that I_K(Ca,d) plays little or no role in the acetate on- or off-responses.

Cam 1-1, which shares with Cam 4 the loss of I_K(Ca,d) but has also lost I_K(Ca,h),shows significantly reduced chemoresonse to K- or Na-acetate(Table 1). Considering that Cam 4 is not defective in its chemoresponse to these stimuli, we attribute the decrease in chemoresponse of Cam 1-1 cells to the loss of I_K(Ca,h). I_K(Ca,h) may be the conductance implicated in the initial hyperpolarization of the on-response.

Cam 11, with its loss of I_Na(Ca), shows chemoresponse to K-acetate within the normal range, but has a complete loss of chemoresponse to Na-acetate (Table 1). Despite the loss of attraction in Na-acetate, we believe that I_Na(Ca) plays no significant role in chemoresponse to acetate because the loss of chemoresponse in K⁺-free (Na) acetate solutions is a secondary effect of the mutation. Cam 11 in K⁺-free media has an extremely low resting membrane potential [near the equilibrium potential for K⁺ (E_K)(Satow and Kung, 1976)], which prevents the cell from responding with an adequate hyperpolarization for the on-response or depolarization of the off-response.

To test further the potential role of I_Na(Ca) in acetate chemoresponse, we tested mutant XntA, which has lost I_Mg(Ca) and has a greatly reduced I_Na(Ca), and found that XntA is normally attracted in K- and Na-acetate solutions(Table 1).

Cam 4 mutants that have lost the I_K(Ca,d)conductance show significantly stronger attraction than wild type in Na-biotin(Tables 2 and 3). Cam 1-1, which has lost both the depolarization- and hyperpolarization-activated I_K(Ca), shows significantly stronger attraction in K-biotin (pH 6.7) and Na-biotin pH 7 than wild-type cells (Tables 2 and 3). Thus, when the I_K(Ca,d) is lost in Cam 4 or when both I_K(Ca,d) and I_K(Ca,h) are lost in Cam 1-1, there is an enhancement of chemoresponse.

To test the hypothesis that the depolarizing off-response is initiated by I_Ca and, therefore, can potentially be enhanced with activation of I_Na(Ca) or I_Mg(Ca) in Na, low pH or Mg solutions, we tested wild-type cells in Na-biotin, Mg-biotin and K-biotin at pH 7 and 6.7 (Table 2). Wild-type cells show significantly stronger attraction to biotin at pH 6.7 than at pH 7. Wild-type cells are also significantly better attracted in Mg-biotin pH 7 than in K-biotin at pH 7, or even Na-biotin at pH 6.7 or 7.

Cam 11, which has lost the I_Na(Ca), and XntA, which has defects in I_Na(Ca) in addition to its loss of I_Mg(Ca), both show in K-biotin at pH 6.7 slightly depressed chemoresponses that nonetheless are within the wild-type range (Table 3). Cam 11 shows significant loss of chemoresponse in Na-biotin at pH 6.7, but this result cannot be used to implicate its lost I_Na(Ca)in an important function like the off-response depolarization because, as described above, Cam 11 in K⁺-free Na⁺solutions has a severely reduced resting potential. However, resting membrane potential of Cam 11 appears to be normal in Mg²⁺solutions, judging from the swimming speed and behavior in K⁺,Na⁺ and Mg²⁺ solutions (W.E.B. and J.L.V.H., unpublished observation) and direct measurements (R.R.P., unpublished observations). Cam 11 shows strong attraction to Mg-biotin.

The T-maze results for the conductance mutants lead us to ask about the swimming and turning behavior of individual cells, which we measure as percent directional change (turning) and speed. In these experiments that examine the on-response, control cells are videotaped as they swim across a boundary from control solution to more control solution and we refer to test cells as those that swim from control to attractant solution. For the off-response, control cells swim from attractant into more attractant while test cells swim out of attractant into the control solution. We are not attempting to use the same solutions for wild type and all the mutants even though strength of attraction can vary with counter-ion. Here we are using a selection of strains and conditions to correlate turning or speed with strength of attraction. Also, Cam 11 must be tested in K⁺ to avoid artifacts of abnormal resting potential of this mutant in the absence of K⁺. Therefore,the statistical analysis that we used in this part of the study was regression analysis.

We assayed behavior of wild type, Cam 11, XntA and Cam 1-1 in acetate solutions. The data in Fig. 2 are arranged by descending order of T-maze attraction to acetate: wild type in Na-acetate 0.84±0.03; Cam 11 in K-acetate 0.70±0.02; XntAin Na-acetate 0.70±0.03; Cam 1-1 in K-acetate 0.62±0.03). Fig. 2Ashows the average numbers of turns (expressed as % directional change or PDC)of cells entering acetate compared to control cells. Fig. 2B shows the turns upon leaving acetate, and Fig. 2Cshows the speed of control cells and those entering acetate. The wild type, Cam 11 and XntA all turn less frequently and speed up as they enter acetate. Cam 1-1 shows no significant change in turning or speed upon entering acetate. Upon leaving acetate, wild-type cells turn more,as do Cam 11 cells. However, XntA and Cam1-1 cells do the opposite and depress the frequency of turning compared with controls.

When the data are treated not as absolute frequencies of turning, but rather as net changes in turning upon entering acetate(Fig. 3A), the differences among the strains of cells become more apparent. Wild type shows larger changes in turning and speed than Cam 11. XntA shows a very large decrease in turning upon entering acetate, but an unexpected decrease in turning upon leaving acetate as well. Cam 1-1 shows small changes in turning and speed, which are opposite to those shown by wild-type cells. A comparison of Fig. 3B,C shows that there is a significant correlation of the speed change upon entering acetate with magnitude of T-maze responses (R²=0.945,Pearce r=0.972), but other parameters do not show statistically significant correlations.

For turning and swimming speed, we observed wild type moving into and out of solutions in Na-, Mg- and K-biotin solutions; Cam 1-1 and XntA in Na-biotin; and Cam 11 in K-biotin. Results for turning are shown in (Fig. 4)and the strains are arranged by descending T-maze values (wild type in Mg-biotin 0.89±0.01; Cam 1-1 in K-biotin 0.85±0.02;wild type in Na-biotin 0.79±0.02; wild type in K-biotin 0.72±0.01; Cam 11 in K-biotin 0.66±0.03; XntAin Na-biotin 0.63±0.04; and XntA in Mg-biotin 0.50±0.02). Comparison of Fig. 4A and B shows that the absolute values for turning frequency moving into biotin correlate significantly with the magnitude of T-maze results(Fig. 4A, R²=0.844, Pearce r=0.7126). Turning frequency in the off-response (Fig. 4C) and speed (not shown) in the on-response do not correlate significantly with the T-maze data.

The transformation of the data into net differences from control in turning frequency of the on- and off-responses highlight the differences among the strains (Fig. 5). Cam 11 and XntA that have both lost I_Na(Ca) show very small changes in both the on- and off-responses in K- and Na-biotin,respectively. XntA that is not attracted at all to Mg-biotin shows statistically significantly different and opposite responses to biotin than wild type in Mg-biotin.

The limited number of conductances that we measured in acetate and biotin did not constitute an exhaustive study, but we present them here because they suggested to us two different models for chemoresponses in these stimuli.

The application of acetate activates a Ca²⁺-dependent K⁺ conductance (I_K(Ca)) that initiates the hyperpolarization in the stimulus. There is a small, sustained outward conductance, which we attribute to the plasma membrane calcium pump (PMCA). Upon removal of acetate, there is an influx of Ca²⁺(I_Ca), large enough at times to activate a small outward K⁺ conductance, but not large enough to activate an inward Na⁺ or Mg²⁺ conductance when these ions are in the bath. Alternatively, the Ca²⁺ influx does not occur in close enough proximity to the Ca²⁺-activated Na⁺ and Mg²⁺channels to activate them.

In the case of biotin, it is difficult to observe an initial K⁺conductance that probably initiates the hyperpolarization, but there is a sustained outward conductance, which we attribute to the PMCAs. There appears to be an inward Ca²⁺ (I_Ca) conductance upon removal of biotin, and this off-response inward conductance is amplified in low pH, Na⁺ (not shown), or Mg²⁺. This is consistent with the activation of I_Na(Ca,d), which also carries H⁺, or activating I_Mg(Ca) when there is Na⁺ or Mg²⁺ in the bath or the bath is of acidic pH(R.P.P., unpublished observations).

Our studies of mutants in T-maze assays implicate I_K(Ca,h) and rule out a role for I_K(Ca,d) in attraction to acetate because mutant Cam 4 that has lost I_K(Ca,d) shows normal chemoresponse while Cam 1-1, which has lost both I_K(Ca,d) and I_K(Ca,h), shows greatly reduced chemoresponse. Since the Cam 1-1 mutant shows reduced attraction but not neutral responses to acetate, the I_K(Ca,h) must be necessary but not sufficient for optimal chemoresponse behavior. Other mutants with altered I_Na(Ca) or I_Mg(Ca) show no significant differences from wild type in their attraction to acetate.

Mutant strains as well as wild type in various salts of biotin demonstrate a role for I_K(Ca,h) and/or I_K(Ca,d) in chemoresponse to biotin, but a very different role from that in acetate. When I_K(Ca,h) is eliminated by mutation, the attraction response is not reduced as in acetate, but rather is enhanced in biotin. Mutant Cam 1-1 shows significantly higher T-maze results than wild type in K- and Na-biotin, and Cam 4 shows significantly higher attraction in Na-biotin (Table 2). Therefore, a K⁺ conductance appears to be activated in biotin chemoresponse, perhaps normally short circuiting and reducing the effectiveness of the depolarizing off-response. We discuss below that when the K⁺ conductance is eliminated by mutation, the off-response is increased (compare Cam 1-1 in Na-biotin to wild type in Na-biotin off-response in Fig. 4C).

Other conductances that can contribute to but are not necessary for chemoresponse in biotin are I_Na(Ca) and I_Mg(Ca). Wild-type cells can be attracted in K-biotin at pH 7 (Table 2), that is, under conditions when I_Na(Ca) and I_Mg(Ca)are not activated. However, cells are more effectively attracted to K-biotin at pH 6.7 or to Na-biotin at pH 6.7 or 7, and show the strongest attraction to Mg-biotin even at pH 7. When I_Na(Ca) is eliminated as in Cam 11, the attraction is greatly diminished for K-biotin at pH 6.7,presumably because I_Na(Ca) is not functional to carry protons or Na⁺. However, when Cam 11 is in Mg-biotin attraction is strong and normal, presumably because Cam 11 has a functional Mg²⁺ channel allowing for a strong off-response. When I_Mg(Ca) is eliminated as in XntA, there is little attraction to Na-biotin and no attraction to Mg-biotin at all, which emphasizes the importance of the depolarizing off-response conductances.

We interpret these results to support the model that in Na-biotin or biotin solutions of pH 6.7, the off-response depolarization can be augmented with the I_Ca activation of I_Na(Ca), which conducts protons as well as Na⁺. When Mg²⁺ is available,the off-response depolarizing I_Ca activates I_Mg(Ca), which even more effectively depolarizes the cell and causes the largest increase in turning frequency as cells leave biotin.

In previous computer simulation studies(Van Houten and Van Houten,1982), we found that some aspects of swimming behavior are critical to accumulation or repulsion in the T-mazes. Most relevant to these studies is the observation that boundary responses are necessary but not sufficient for accumulation. That is, an immediate response when crossing a step gradient is essential as the cells move into attractant (hyperpolarize and swim smoothly and straight) and out of attractant (depolarize and immediately turn). It appears that the behaviors of wild type and mutants in acetate solutions (Figs 2 and 3) implicate the immediate on-response, particularly the speed increase, in successful attraction to acetate.

Behavior of wild type and mutants in biotin solutions likewise implicates the on-response depression of turning as critical in successful attraction to biotin, but speed increases do not correlate well with attraction in T-mazes(Figs 4 and 5). The off-response increase in turning and changes in speed, likewise, do not correlate with successful T-maze attraction for either stimulus. However, we note that the wild-type cells increase their speed significantly (0.93±0.03 in control vs 1.14±0.02 mm s^-1) when cells leave biotin, which is unexpected given the large off-response depolarization and inward conductances.

Our data on behavior of wild type and mutant cells in acetate solutions lead us to a model for acetate attraction that begins with a hyperpolarizing K⁺ conductance to initiate the on-response behavior of fast swimming with few turns. We believe that this is the most important aspect of successful attraction to acetate and the I_K(Ca,h) is the conductance involved. The significant correlation of the on-response increase in swimming speed with T-maze values fits with the initiation of the on-response by I_K(Ca,h). The loss of this conductance in mutant Cam 1-1 correlates with diminished attraction and with speed and turning changes that are the opposite of the wild type. The small residual attraction of the mutant Cam 1-1 to acetate can be attributed to the activation of the plasma membrane calcium pump, which we believe normally sustains the hyperpolarization in wild-type cells. Our observations of cells that have 60% reduced calmodulin levels (by transformation with anti-sense calmodulin expression vectors) show that these cells have no sustained hyperpolarizing conductance and also no significant attraction to acetate(Yano et al., 1996) (R.R.P.,W.E.B., J. Yano and J.L.V.H., unpublished observations).

A cell in biotin appears to experience very different chemosensory conductances. The tests of mutants in T-mazes point to I_{K(Ca,d or h)} short-circuiting a conductance that is critical for attraction because elimination of I_{K(Ca,h and d)} in mutant Cam 1-1 leads to increased attraction, just the opposite as for acetate. The conductances that might be short-circuited by I_K(Ca)appear to be I_Na(Ca) and I_Mg(Ca). When I_Na(Ca) or I_Mg(Ca) is lost by gene mutations in Cam 11 and XntA, which we previously thought to be associated with the depolarizing off-response, to our surprise, Cam 11 and XntA attraction to biotin is compromised and on-response speed and turning changes are small or even opposite to the wild type. We had not anticipated a role for conductances lost in these mutants(I_Na(Ca) and I_Mg(Ca)) in the on-response. We associate I_Na(Ca) with the depolarization off-response and turning upon leaving biotin, as supported by the electrophysiological recordings (Fig. 1). However, we find a significant correlation of decreased turning in the biotin on-response with attraction strength, and we do not see a role for I_K(Ca) in determining the on-response behavior(see Fig. 4A wild type vs Cam 1-1). Additionally, the unexpected transient increase in speed as wild-type cells leave biotin could be explained with an activated I_K(Ca). Therefore, we propose that the hyperpolarization on-response in biotin is initiated by a conductance other than I_{K(Ca,h or d)} and that on-response behavior is important to the attraction outcome. The off-response depolarization is initiated by an I_Ca, which can be large enough or close enough to channels to open I_K(Ca,d), I_Na(Ca) and I_Mg(Ca). The Na and Mg conductances can enhance the off-response depolarization, but I_K(Ca,d) dampens it. When I_K(Ca,d) is eliminated by mutation, attraction is stronger. Therefore, we attribute importance to both the on- and off-responses for optimal attraction to biotin. Strength of attraction correlates to decrease in on-response turning and enhancement or loss of the off-response conductances greatly affect attraction outcomes. With this model we can account for the loss of attraction of XntA and Cam 11, but we cannot account for the loss of on-responses of PDC and speed when these mutants are introduced to biotin.

In summary, attraction signal transduction pathways for two stimuli,acetate and biotin, involve conductances for on- and off-responses differently. We presumed that the off-response would be more important to the outcome of attraction to biotin than the on-response, but we did not foresee the role of the I_K(Ca) in wild-type attraction to biotin on- or off-responses. Likewise, without the mutants, we would not have uncovered a dampening role for I_K(Ca) or the potential roles for I_Na(Ca) and I_Mg(Ca) in the on-response. These studies point to the usefulness of the P. tetraurelia behavioral mutants in dissecting complex pathways.

This work was supported by NIH R01 DC 00721. We thank C. Kung for his isolation and sharing of so many behavioral mutants of P. tetraurelia.