Ionic Ca2+ functions as a second messenger to control several intracellular processes. It also influences intercellular communication. The release of Ca2+ from intracellular stores through the inositol 1,4,5-trisphosphate receptor (InsP3R) occurs in both excitable and nonexcitable cells. In Drosophila, InsP3R activity is required in aminergic interneurons during pupal development for normal flight behavior. By altering intracellular Ca2+ and InsP3 levels through genetic means, we now show that signaling through the InsP3R is required at multiple steps for generating the neural circuit required in air puff-stimulated Drosophila flight. Decreased Ca2+ release in aminergic neurons during development of the flight circuit can be compensated by reducing Ca2+ uptake from the cytosol to intracellular stores. However, this mode of increasing intracellular Ca2+ is insufficient for maintenance of flight patterns over time periods necessary for normal flight. Our study suggests that processes such as maintenance of wing posture and formation of the flight circuit require InsP3 receptor function at a slow timescale and can thus be modulated by altering levels of cytosolic Ca2+ and InsP3. In contrast, maintenance of flight patterns probably requires fast modulation of Ca2+ levels, in which the intrinsic properties of the InsP3R play a pivotal role.
Neuronal control of behavior arises from the activity of underlying neural circuits, which in turn are specified by the interaction of various signaling pathways. Over the past few years, evidence has accumulated demonstrating that ionic calcium (Ca2+) plays an important role in the development of activity in neural circuits (Spitzer, 2002; Borodinsky et al., 2004) and hence contributes significantly to the formation of functional neuronal networks. The Ca2+ signal is characterized by a rapid increase in the concentration of free cytosolic calcium ([Ca2+]i) attributable to the opening of calcium channels located in the plasma membrane and in the membranes of intracellular calcium stores. Free Ca2+ is then rapidly sequestered by the action of calcium pumps and exchangers and also with buffering by cytosolic calcium binding proteins. The functional properties and localization of the channels as well as the pumps, exchangers, and buffers are thought to play a crucial role in generating a range of calcium signals that vary from a brief local increase to repetitive calcium spikes and waves spreading over a larger region. The kinetics, amplitude, and duration of these signals in turn determine the cellular specification of neurotransmitters, receptors, and channels that influence synaptic activity and plasticity in cognate neurons (Berridge, 1998; Spitzer et al., 2000).
Our goal is to understand the contribution of the intracellular Ca2+ release channel, the InsP3 receptor (InsP3R), in the development and function of neural circuitry. From studies in Drosophila, we have shown that flight is critically dependent on normal activity of the InsP3R in aminergic interneurons during pupal development suggesting that InsP3-mediated Ca2+ release is required during normal development of the flight circuit (Banerjee et al., 2004). Mutants in the gene encoding the InsP3R (itpr) in Drosophila exhibit a range of defects including altered wing posture, increased spontaneous firing, and loss of rhythmic flight patterns in response to an air puff stimulus. Together, these phenotypes contribute to the loss of flight behavior observed in itpr mutants. An obvious question that arises from these studies is whether the multiple phenotypes arise as a consequence of a single early neuronal dysfunction. Alternately, each phenotype could be attributable to independent events requiring InsP3R activity at individual and distinct steps of flight circuit development. Here, we have addressed these questions by altering intracellular Ca2+ signals in flight-deficient itpr mutants through genetic means. In the first instance, a dominant mutant for the sarco-endoplasmic reticular Ca2+-ATPase (SERCA) pump was introduced into itpr mutant backgrounds. SERCA is responsible for pumping cytosolic Ca2+ into the endoplasmic reticulum (ER) store and thus maintaining the intracellular concentrations of Ca2+ both in the cytosol and in the ER store. The second class of mutants are in genes that encode (1) the α subunit of the heterotrimeric G-protein Gαq (Scott et al., 1995; Scott and Zuker, 1998) and (2) phospholipase Cβ (PLCβ) (Bloomquist et al., 1988; Shortridge et al., 1991). These mutants are expected to reduce InsP3 levels and thus reduce the activity of the InsP3R, in circumstances in which InsP3 is generated by the activation of seven transmembrane receptors. Our results show that neuronal phenotypes of itpr mutants arise from at least two distinct classes of intracellular Ca2+ signals. One class can be modulated by the strength of InsP3 signaling and the rate of Ca2+ uptake into the ER. The second class appears to depend primarily on the Ca2+ release properties of the InsP3 receptor.
Materials and Methods
Different mutant alleles for the itpr gene were tested for flight. The viable heteroallelic combinations used in this study were itprwc703/wc361, itprka1091/ug3. These are single point mutants in the itpr gene that were generated in an EMS (ethyl methanesulfonate) screen. Detailed molecular information on these alleles has been published recently (Joshi et al., 2004; Srikanth et al., 2004b). For rescue experiments, embryonic wild-type itpr cDNA (UASitpr+) (Venkatesh et al., 2001) and α3 splice variant of the dgq gene (UASdgqα3+) (Ratnaparkhi et al., 2002) were used. Ca-P60AKum170ts (referred as Kum170 throughout the text) was obtained from Dr. K. S. Krishnan (Sanyal et al., 2005a), dgq221c was generated by Pinky Kain in collaboration with Veronica Rodrigues, whereas dgq18745 was procured from the Bloomington Stock Center (Bloomington, IN). plc21cP319/11 (referred to as plc21cP319 in this study) was obtained from Dr. S. Leevers (Weinkove et al., 1999), and norpAP24 is a null allele of norpA gene (Pearn et al., 1996). DdcGAL4 (Li et al., 2000), and hsp70GAL4 strains [referred to as hsGAL4L in Banerjee et al. (2004)] were obtained from the Bloomington Stock Center. hsGAL4L exhibits basal green fluorescent protein (GFP) expression at 25°C, when crossed to UASGFP. Because the UASGq3 strain described previously (Ratnaparkhi et al., 2002) was lost, a similar strain was remade. In the newly generated transgenic UASGq3 flies (referred to as UASdgqα3+), the insert was mapped to chromosome 2. The following fly strains were generated by standard genetic methods using individual mutant and transgenic fly lines described above: (1) Kum170/CyoG;itpr ug3/TM6Tb, (2) UASitpr+;+/+;itprka1091/TM6Tb, (3) UASitpr+;+/+;itprwc703/TM6Tb, (4) ElavGAL4/CyoG;itprug3/TM6Tb, (5) dgq221c/CyoG; itprwc703/TM6Tb, (6) dgq18745/CyoG;itprwc703/TM6Tb, (7) plc21cP319;itprwc703/TM6Tb, (8) plc21cP319;itprwc361/TM6Tb, (9) Kum170-plc21cP319; itprwc361/TM6Tb, (10) hsGAL4L-dgq221c/CyoG; itprwc361/TM6Tb, (11) DdcGAL4- dgq221c/CyoG; itprwc361/TM6Tb, (12) UASdgqα3+/CyoG; itprwc703/TM6Tb, and (13) UASdgq1f1; itprwc703/TM6Tb.
In the above description of genetic strains and in the rest of the text, the symbol “/” separates two homologous chromosomes, “;” separates two different chromosome, and “-” designates alleles of two genes or transgenes recombined onto one chromosome.
Larval staging and lethality measurements.
To obtain molting profiles of heteroallelic mutant larvae, staging experiments were performed as described previously (Joshi et al., 2004). Heteroallelic and heterozygous larvae were identified based on the Tubby phenotype, which is visible in larvae from 60 h after egg laying (AEL). Timed and synchronized egg collections were done for a period of 6 h at 25°C. The cultures were then allowed to grow at 17.5 or 25°C depending on the experiment. The temperature of 17.5°C was chosen because larval development takes precisely double the time at this temperature compared with development at 25°C. Heteroallelic mutant larvae were selected at 56–64 h AEL when the experiments were done at 25°C and at an interval of 116–124 h AEL in cases in which the experiments were done at 17.5°C. Larvae were selected based on the dominant markers Tubby (on TM6Tb) and CyoGFP and transferred into vials of cornmeal medium lacking agar. These larvae were grown at the desired temperatures of 17.5 or 25°C and screened at appropriate time points, for number of survivors and their stage of development. Larval stages were determined by the morphology of the anterior spiracles (Ashburner, 1989). For each time interval, a minimum of 150 larvae were screened in batches of 50 larvae each.
Flight tests were performed as described by Banerjee et al. (2004) following minor modifications of the “cylinder drop assay” described previously (Benzer, 1973). A batch of 20 flies was tested each time. A minimum of five batches were tested for each genotype. Flies that dropped directly down the cylinder were collected in a vial kept on ice underneath the lower opening of the cylinder. These were counted as flight defective. Flies that were able to hold onto the walls of the cylinder were considered as fliers. The percentage of flight-defective organisms was determined as F/T × 100, where F is the number of flies that dropped into the vial at the bottom of the cylinder and T is the total number of flies tested. Computation of means, SDs, and t tests was performed using Origin software (Origin Lab, Northampton, MA).
Electrophysiological preparation and recordings.
Physiological recordings were performed on the dorsal longitudinal muscles (DLMs) of the giant fiber pathway (Tanouye and Wyman, 1980). Recording methods for the muscle have been described previously (Engel and Wu, 1992; Lee and Wu, 2002). For measurement of responses to an air puff stimulus, flies were anesthetized briefly with diethyl ether and glued to a thin metal wire between the neck and the thorax with nail polish. Flies were allowed to recover from anesthesia for ∼4 h. After recovery, an uninsulated tungsten electrode (0.5 μm), which had been sharpened by electrolysis, was inserted carefully into the DLM (fiber a) just beneath the cuticle. A similar tungsten electrode was inserted in the abdomen as a reference. Flies were rested for at least 10 min after insertion of electrodes before beginning the recording. For measurement of spontaneous activity, flies were kept undisturbed and recordings were performed for 2 min. To record air puff responses, a gentle mouth-blown air puff stimulus was delivered to the fly kept in a tethered condition. Responses were measured from DLM “a” with a tungsten electrode (specification mentioned above), using an ISODAM8A (World Precision Instruments, Sarasota, FL) amplifier with filter set up 30 Hz (low pass) to >10 kHz (high pass). Gap free mode of pClamp8 (Molecular Devices, Union City, CA) was used to digitize the data (10 kHz) on a Pentium 5 computer equipped with Digidata 1322A (Molecular Devices). Data were analyzed using Clampfit (Molecular Devices) and plotted using Origin (MicroCal, Northampton, MA).
Microsomal vesicle Ca2+ release assay.
A modified protocol from Bramley et al. (1990) was used, which has been described recently (Srikanth et al., 2004a). In brief, microsomes were prepared in the presence of 200 μm free Ca2+, from either 75 wandering third-instar larvae or ∼200 adult heads, obtained from the appropriate genotypes. Microsomes were made in parallel, in Ca2+-free buffer and otherwise identical conditions, and added to the buffers used for determining the standard curve for Ca2+ for each individual experiment. Membrane-impermeant Ca2+-sensitive ratiometric fluorescent dye, Fura-2 (Invitrogen, Eugene, OR) was prepared in calcium-free water. Steady-state fluorescence measurements were performed in a SPEX Fluorolog-2 spectrofluorometer (SPEX Industries, Edison, NJ) at 20°C so as to minimize nonspecific calcium leak. For each run, ∼15 μg of adult head microsomes were added to 2 ml of assay buffer (20 mm Tris, pH 7.4, and 80 mm NaCl in calcium-free water) containing 5 μm Fura-2. Steady-state kinetics of Ca2+ release were measured at various concentrations of InsP3 and quantified by plotting a standard curve with known amounts of free Ca2+ using the standard Ca2+-EGTA buffering system.
Neuron culture and calcium imaging.
Primary cultures of Drosophila larval neurons were plated in 200 μl of Schneider’s medium supplemented with 10% fetal bovine serum (Invitrogen), 50 U/ml penicillin, 50 μg/ml streptomycin, and 10 μg/ml Amphotericin B as described previously (Wu et al., 1983). Briefly, brain and the ventral ganglion complex were dissected from surface-sterilized Drosophila third-instar larvae of the appropriate genotypes. The brain tissue was mechanically dissociated using syringe needles in Schneider’s medium containing collagenase (0.75 μg/μl) and dispase (0.4 μg/μl) and incubated in the proteolytic medium for 1 h to allow complete dissociation of the tissue. The lysate containing essentially single cells was then spun down, resuspended in Schneider’s medium (200 μl of the medium was added for every four brains dissected), and plated onto 35 mm culture dishes with a poly-l-lysine-coated coverslip for the bottom. The cells were incubated at 22°C for 14–16 h before imaging.
Calcium imaging in larval neurons.
Larval neuron cultures were washed twice after growth for 14–16 h with Drosophila M1 medium (20 mm HEPES, 150 mm NaCl, 5 mm KCl, 1 mm MgCl2, 1 mm CaCl2, and sucrose, pH 6.9). They were loaded in the dark with 2.5 μm Fluo-3AM (see Fig. 2Aa,Ab) or Fluo-4AM (Ac) in M1 medium containing 0.002% Pluronic F-127 for 30 min at room temperature. The fluorescent dyes were obtained from Invitrogen. After washing three times with M1, the cells were finally covered with 100 μl of M1 and imaged within 40 min of loading. For quantitative analysis, a field with several cells was selected and imaged using the epifluorescence optics of a Nikon TE2000 inverted wide-field microscope with an oil objective (60× and 1.4 numerical aperture) lens. Excitation of Fluo-3/4 was performed using 488 nm wavelength illumination from a mercury arc lamp. Emitted light was detected through a 505 nm bandpass filter (FITC filter set, 41001-exciter HQ480/40, dichroic Q505LP, emitter HQ535/50; Chroma, Brattleboro, VT). Image acquisition was performed using the Evolution QEi CCD camera and QED imaging software (Media Cybernetics, Silver Spring, MD). The time lapse acquisition mode of the software was used to follow fluorescence changes in the cells every 5 s for a period of 1 min. Each exposure was for 200 ms. The cells were depolarized with 70 mm KCl within the first 5 s after the start of data acquisition. As a control, a series of images were acquired under the same imaging protocol without addition of KCl. A total of ∼70 cells was analyzed from 10–15 dishes imaged for each genotype. Finally, ∼45 cells for each genotype were selected for the graph in Figure 2Ac. The cells were selected based on the observation that the Ca2+ signal decayed following the expected first-order exponential decay kinetics.
For measuring increase in fluorescence with time, images were processed using the NIH ImageJ software, version 1.33. Fluorescence intensity before (Fbasal) and after depolarization (Fmax) with KCl, was determined for all of the time points shown in Figure 2A. Background fluorescence (an area without any cells) was subtracted from the fluorescence values for each cell. The data were plotted using Origin 6.0 software as follows: ΔF/F = Fmax − Fbasal/Fbasal for each cell. The mean and SE for ΔF/F values of the cells were calculated for every time point and compared between two appropriate genotypes using the Student’s t test. The mean values for each time point were normalized to the mean peak value reached for each individual genotype taken as 100. No significant difference was observed between the mean peak values of the six genotypes tested.
Protein extracts from ∼100 first-instar larvae (aged 28–36 h after egg laying) of indicated genotype were made by homogenization in 100 μl of homogenizing buffer containing 50 mm Tris, pH 8.0, 150 mm NaCl, 1 mm EGTA, 1.5 mm MgCl2, Triton X-100 (1%), 6 m urea. Third-instar larvae (10–15) were used for making lysates from dgq221c/+, dgq18745/+, ArmGAL4/UASdgq1f1, and wild-type animals. The samples were run on a 12% SDS-polyacrylamide gel and transferred to nitrocellulose membranes by standard Western blotting protocols. Anti-Gq antiserum from Santa Cruz Biotechnology (SC-3912; Santa Cruz, CA) was used at a dilution of 1:500. Secondary anti-rabbit (donkey) antibody was used at 1:10,000 dilution. To estimate amount of protein loaded, β-tubulin levels were detected using anti-β-tubulin (E7; Developmental Studies Hybridoma Bank, Iowa City, IA) at a dilution of 1:330 and an anti-mouse secondary (catalog #7076; Cell Signaling Technology, Beverly, MA) was used at a dilution of 1:1000. Secondary antibodies were conjugated to horseradish peroxidase, and the detection of protein in the blot was done by addition of a chemiluminescence substrate from Pierce (catalog #34075; Rockford, IL). Typically, 1–2 s exposures were given to develop the blot.
PCR mapping of dgq coding region deletion in dgq221c.
Genomic DNA was isolated from first-instar larvae of the appropriate genotypes mentioned in Figure 3B according to standard protocol. Amplification of three fragments (see Fig. 3B, denoted as F1, F2, F3) was by following standard PCR protocols. Primers used for amplification of the F1 fragment are P11 (5′-AACCATCACTCTCAGC-3′) and P12 (5′-GCACTCCATGCTAACTA-3′), for F2 are P1 (5′-AGCGGTTACTCGGACGAGGACA-3′) and P3 (5′-CTCAAGAATGCCAGTTGTCGGCAC-3′), and for F3 are P17 (5′-GTGGTCAGCGATCCGAG-3′) and P8 (5′-GCACACGTGAAATGAGAATAGA-3′).
Generation of a dsRNA construct for the dgq gene.
The dgqα3 cDNA isolated from an appendage cDNA library (Ratnaparkhi et al., 2002) was used as a template to generate an inverted repeat construct coding for a dsRNA, which would encompass a region common to all dgq splice variants (approximate position of the primers used is shown in Fig. 3A). Briefly, two PCR products were created of ∼770 from dgqα3 cDNA. The first product was created using a 5′ primer containing KpnI site (5′-GGTACC(538)TCACGATACTAGCAGCATCCC-3′) and a 3′ primer containing a BamH1 site (5′-GGATCC(1314)CGGTGTAAGCGAGCGAAG-3′) at the ends of the respective primers. This product was then cloned into the KpnI and BamH1 of pUAST (Brand et al., 1994). A linker region of 12 nucleotides was introduced during the PCR [6 bp overhang from the cDNA (1308–1314) and 6 bp from the BamH1 site]. A second PCR product was created using a reverse 5′ primer containing an XbaI site (5′-TCTAGA(539)GTGCTATGATCGTCGTAGGGA-3′) and a reverse 3′ primer containing a BamH1 site (5′-GGATCC(1308)TTCGCTCGCTTCTCAATTCT-3′) at the 5′ end. This was cloned in the pUAST KpnI-BamH1 constituent (containing the previous insert) using BamH1 and XbaI sites, thus creating an inverted repeat. Germline transformants with this construct were obtained following the standard protocol of Drosophila embryonic microinjection.
Differential suppression of phenotypes arising from mutant InsP3Rs by a dominant mutant in dSERCA or CaP60A
Among the existing set of itpr mutant alleles, are a series of heteroallelic itpr mutant animals, which when grown at 25°C exhibit altered wing posture and severe flight defects. Interestingly, these heteroallelic itpr mutants are lethal when grown at 17.5°C (Joshi et al., 2004). These mutant combinations thus serve as a sensitized genetic background that can be used for assessing the role of modifiers of InsP3R phenotypes. The identification of such modifiers is likely to lead to a better understanding of how InsP3-mediated Ca2+ release modulates neuronal function. Among these heteroallelic mutant combinations, one of the best studied is itprka1091/ug3 (Banerjee et al., 2004; Joshi et al., 2004). A change in the Ca2+ release properties of itprka1091/ug3 channels was first identified by measuring the kinetics of Ca2+ release in response to InsP3 in microsomal vesicles obtained from itprka1091/ug3 organisms and compared with microsomal vesicles from wild-type heads (Fig. 1A). Decreased levels of Ca2+ release was observed after InsP3 stimulation of microsomes prepared from either adult heads or larvae of itprka1091/ug3 when compared with microsomal vesicles from wild-type animals of a comparable developmental stage. These in vitro data suggest that it might be possible to suppress phenotypes associated with itprka1091/ug3 in vivo by elevating [Ca2+]i through genetic means. One way to achieve this would be to reduce the uptake of cytosolic Ca2+ into intracellular stores. Toward this end, a dominant mutant allele (Kum170) for the gene (Ca-P60A) encoding the SERCA, was introduced in itprka1091/ug3 organisms.
Although Kum170 was originally isolated as a temperature-sensitive dominant paralytic whose focus lay in muscle tissue (Sanyal et al., 2005a), recent work has established it as a dominant mutant that affects SERCA function at permissive temperatures also (Sanyal et al., 2005b). To examine the effect of Kum170 on itpr mutant phenotypes, animals of the genotype Kum170/+; itprka1091/ug3 were grown at either 25 or 17.5°C. Interestingly, the presence of a single copy of Kum170 could suppress the altered wing posture seen in itprka1091/ug3 organisms at 25°C (Fig. 1B). The suppression was not evident in Kum170/+; itprka1091/ug3 flies grown at 17.5°C. However, Kum170 could effectively suppress larval lethality seen in itprka1091/ug3 animals at 17.5°C (Fig. 1C), indicating that the Kum170 allele exerts its mutant effect at temperatures as low as 17.5°C. The cellular focus of itpr mutant lethality at 17.5°C lies in neurons, because expression of an itpr+ transgene under the control of a pan-neuronal promoter rescues it completely (Fig. 1C). We therefore tested whether there was an effect of the Kum170 allele on bringing cytosolic Ca2+ ([Ca2+]i) back to basal levels in larval neurons after a depolarization stimulus. As shown in Figure 2A, larval neurons cultured from Kum170/+ animals return to basal levels of [Ca2+]i at a rate slower than what is observed for neurons from wild-type larvae, after depolarization by the addition of 70 mm KCl. Thus, in agreement with previous data from cardiac and muscle cells, in neurons too the Kum170 mutation in SERCA slows down the rate of Ca2+ entry into intracellular stores. We have not confirmed whether the change in this rate depends on temperature, but based on the differential suppression of neuronal phenotypes at 17.5 and 25°C described above, this seems likely.
We then tested the effect of Kum170 on calcium sequestration in larval neurons from itpr mutants after KCl depolarization. The rate of return to basal [Ca2+]i is similar to wild type in both itpr mutants tested (itprka1091/ug3 and itprwc361/wc703), whereas it is slower (and comparable with Kum170/+) in Kum170/+; itprka1091/ug3 and Kum170/+; itprwc361/wc703 neurons (Fig. 2Ac). The InsP3R is known to alter [Ca2+]i only in response to the generation of intracellular InsP3. Any change in cytosolic sequestration in InsP3R mutants on depolarization with KCl would very likely reflect a change in the basal concentration of store calcium. Our data do not support this idea.
Suppression of larval lethality (Fig. 1C) and wing posture by Kum170 led us to examine the status of other neuronal phenotypes associated with itpr mutants (Banerjee et al., 2004; Joshi et al., 2004). Initially, the flight behavior of Kum170/+; itprka1091/ug3 animals was compared with that of itprka1091/ug3. The percentage of organisms with flight defects was no different between the two strains, when measured in a “cylinder drop test” assay (Fig. 2B). Most simplistically, these data indicate that itpr mutant phenotypes of larval lethality and wing posture on one hand and flight behavior on the other hand, arise from differential intracellular calcium requirements. This idea is strongly supported by results of the next set of experiments, which looked at the status of spontaneous firing and air puff-induced flight patterns from the indirect flight muscles in Kum170/+; itprka1091/ug3 organisms.
In wild-type flies, recording of spontaneous activity from the dorsal longitudinal flight muscles reveals occasional bursts of action potentials. itpr mutants display a considerably higher frequency of such spontaneous action potentials (Banerjee et al., 2004). In Kum170/+; itprka1091/ug3 organisms (grown at 25°C), spontaneous firing from the DLMs was significantly reduced compared with itprka1091/ug3 animals (Fig. 2C,D). No obvious effect on spontaneous firing is observed in Kum170/+ animals at 25°C (Fig. 2C,D). Thus, the Kum170/+ mutant allele can suppress this aspect of flight physiology despite its inability to suppress flight behavior in the cylinder drop test. Next, we tested the flight response to a gentle air puff in these flies. Unlike itprka1091/ug3 flies, which are unable to initiate any rhythmic flight patterns, the Kum170/+; itprka1091/ug3 animals exhibit a few cycles of wing beating when stimulated with an air puff. Simultaneous recordings made from the DLMs, showed the rhythmic generation of action potentials immediately after the air puff delivery (Fig. 2E; supplemental movie 1, available at www.jneurosci.org as supplemental material). However, unlike wild-type flight patterns, the air puff response in Kum170/+; itprka1091/ug3 terminated in ∼5 s. The Kum170 mutant allele of SERCA can thus partially suppress the inability of itpr mutants to respond to an air puff.
The differential suppression of these neuronal phenotypes by Kum170 could arise from a difference in the level of [Ca2+]i required for suppression of each of these phenotypes. We therefore tested suppression of flight behavior in another itpr heteroallelic combination, itprwc703/wc361, in which flight defects are limited to ∼30% as measured by the cylinder drop test assay (Banerjee et al., 2004) and in which release of intracellular Ca2+ on InsP3 stimulation is not observably reduced as judged by the microsomal vesicle assay (Fig. 1A). Here, too, the introduction of a single copy of Kum170 did not suppress defects in flight behavior (Fig. 2B). The nonfliers from among itprwc703/wc361 animals exhibit brief rhythmic flight patterns that are usually not sustained beyond 5 s. In organisms of the genotype Kum170/+; itprwc361/wc703, flight physiology was no different from itprwc361/wc703 flies (data not shown). Thus, whereas Kum170 can suppress InsP3R mutant phenotypes of larval viability, wing posture, spontaneous firing, and initiation of the air puff response, it is unable to suppress maintenance of flight patterns for periods beyond 5 s. The latter phenotype remains unchanged in itpr mutants with subnormal (itprka1091/ug3) or near-normal (itprwc361/wc703) levels of Ca2+ release, suggesting that it is the nature of the Ca2+ signal rather than levels of intracellular Ca2+ that dictates the maintenance of flight patterns.
The precise change in the nature of the Ca2+ signal in itprwc703/wc361 organisms remains to be understood. However, based on our previous findings with WC703 single channels, it is expected that the Ca2+ dependence of InsP3-mediated Ca2+ release from WC703/WC361 channels will be altered in vivo. Whereas wild-type Drosophila InsP3 receptors are activated and inhibited over a wide range of [Ca2+], WC703 channels exhibit this activation and inhibition over a very narrow range of calcium concentrations (Srikanth et al., 2004a). These data support the idea that Ca2+ release through the InsP3R results in at least two kinds of intracellular Ca2+ signals. This idea was tested further by analyzing the effect of mutants predicted to reduce InsP3 signaling.
Drosophila Gαq and PLCβ21C mutants enhance InsP3R phenotypes in larvae
The differential suppression of itpr phenotypes by Kum170 led us to first investigate the mode of InsP3 receptor activation in Drosophila neurons in each phenotypic context. InsP3 signaling in vertebrates can be initiated either by the activation of seven transmembrane domain receptors or receptor tyrosine kinases (Berridge, 1993). The former class of receptors act via heterotrimeric G-proteins, consisting of an α-subunit of the Gq class. In Drosophila, Gαq is encoded by the dgq gene (Lee et al., 1990, 1994; Scott et al., 1995).
To test whether activation of the InsP3R occurs through Gαq, we obtained mutant dgq alleles predicted to affect all known dgq splice forms (Fig. 3). An existing mutant of dgq (dgq393) specifically affects splicing of the eye-specific splice variant (dgqα1) and consequently the process of visual transduction (Scott et al., 1995). However, larval and adult brains express the dgqα3 splice variant, which is also present in other tissues (Ratnaparkhi et al., 2002). Based on its expression pattern, we predicted that mutant alleles affecting dgqα3 would be lethal. An excision allele of dgq (221c) was generated by hopping out a P element (BL-14073) from the 5′ untranslated region (UTR) of dgq. The insertion site of P14073 lies ∼3.5 kb upstream of the translation start site of all Dgq isoforms (Fig. 3A) [details regarding the generation of dgq221c will be published independently (P. Kain, G. Susinder, T. Senthil, V. Rodrigues, and G. Hasan, unpublished results)]. PCR analysis performed on genomic DNA obtained from dgq221c homozygous larvae indicated the absence of a 359 bp fragment in the region of the translation start site (Fig. 3A,B). This suggested that dgq221c homozygotes would be devoid of zygotic Gqα protein. To verify this, the level of Dgqα3 was determined in larval lysates by carrying out Western blots in which Dgqα3 levels were detected with an antibody specific for this isoform (Ratnaparkhi et al., 2002). As expected, the intensity of the band corresponding to Dgqα3 was considerably reduced in dgq221c homozygotes on comparison with the level of another protein (β-tubulin), and with the levels observed in dgq221c/+ heterozygotes (Fig. 3C). Dgqα3 levels are also reduced to a similar extent in a newly designated dgq allele (dgq18745), obtained from the public stock center. The P-insert in this allele is in the first intron of dgq as determined from the sequence of the neighboring genomic region, obtained from the BDGP (Fig. 3). We attribute the low level of Dgqα3 seen in the mutant homozygotes to a maternal contribution. However, this has not been tested rigorously.
Animals homozygous for dgq221c and dgq18745 survive for 10–12 h posthatching, indicating an essential requirement of this gene in larvae (data not shown). A role for the itpr locus in larval neurons, in the context of viability, is known (Joshi et al., 2004). Consequently, we tested the effect of a single copy of dgq221c on the viability of itprwc361/wc703 organisms. The viability profile of animals of the genotype dgq221c/+; itprwc703/wc361 appeared similar to that of itprwc703/wc361 (Table 1). To further reduce signaling through Gqα and PLCβ, a single copy of an existing hypomorphic allele of plc21C [P319; (Weinkove et al., 1999)] was introduced to generate a strain of the genotype dgq221C/plc21CP319; itprwc703/wc361. Viability in this triple mutant strain was indeed reduced at all developmental stages tested, and only 40% of the animals survived until adulthood (Table 1). Thus, in the relatively broad context of larval viability, signaling through the InsP3R is compromised by reducing levels of Gαq and PLCβ. The rationale of this experiment was based on biochemical and pharmacological studies from vertebrates in which it is known that activation of Gqα leads to stimulation of the enzyme PLCβ, which cleaves phosphatidylinositol 1,4-bisphosphate (PIP2) to generate InsP3. The Drosophila genome has two genes coding for PLCβ, referred to as plc21C and norpA. Among these, plc21C is expressed ubiquitously (Shortridge et al., 1991), whereas norpA is expressed strongly in the eyes and is essential for visual transduction (Bloomquist et al., 1988). The effect of norpA mutant alleles on the viability of itpr mutants is currently under investigation.
In vertebrates, cleavage of PIP2 by PLCγ has been shown to occur after activation of receptor tyrosine kinases (Rebecchi and Pentyala, 2000). In Drosophila, a single gene with homology to PLCγ has been described (Thackeray et al., 1998). This form of PLCγ shares an equal degree of similarity with two forms of PLCγ found in mammals (Manning et al., 2003). We tested whether mutant alleles in the PLCγ gene (small wing—sl1; sl2) enhanced larval lethality in itpr mutants. Both sl mutant alleles have been designated as null (Thackeray et al., 1998). We monitored the survival of sl1; itprka1091/ug3 or sl2; itprka1091/ug3 males at different stages of development. The survival profile of sl; itpr double mutant animals is similar to animals carrying mutations in the itpr locus alone (Table 1). Moreover, adult phenotypes described for the sl mutants were not enhanced (data not shown). Thus, Drosophila PLCγ appears to play no discernable role in the activation of the InsP3R. This finding is consistent with previously described phenotypes of sl mutants, none of which overlaps with phenotypes described for adult viable itpr mutants (Thackeray et al., 1998; Manning et al., 2003; Joshi et al., 2004).
Wing posture and flight behavior defects in itpr mutants are enhanced by a single copy of dgq mutant alleles
The newly characterized dgq mutant alleles were tested for their effect on flight phenotypes associated with itpr mutants. For this purpose, the itpr mutant combination of itprwc703/wc361 was chosen. On introduction of a single copy of dgq221c in itprwc703/wc361 organisms, a slight change in wing posture was evident. This was further enhanced in flies of the genotype dgq221c/plc21cP319; itprwc703/wc361 (Fig. 4). Similar results were obtained with dgq18745 (data not shown). Quantification of the wing posture seen in these genotypes was not possible primarily because of the fact that wing posture undergoes alterations during anesthetization in wild-type and mutant flies.
To study the effect of dgq mutant alleles on flight behavior, flies of different genotypes were tested in the cylinder drop test. The 30% flight defects seen in itprwc703/wc361 animals were enhanced to 80% by introducing a single copy of either dgq221c or dgq18745 (Fig. 4B). Next, we assessed the effect of mutant alleles of plc21c and norpA on flight behavior of dgq and itpr mutant combinations. As expected, from the altered wing posture of the dgq221c/plc21cP319; itprwc703/wc361 flies (Fig. 4A), these organisms exhibit poor flight (∼95% are flight defective) (Fig. 4B). In comparison, flight defects in males of the genotype norpAP24; dgq221c/+; itprwc703/wc361 were no different from dgq221c/+; itprwc703/wc361 (76 ± 8 and 80 ± 5%, respectively). None of the PLCβ mutant strains has flight defects by itself (Fig. 4B) (data not shown). Furthermore, the two mutant alleles tested, plcβ21CP319 and norpAP24, did not enhance flight defects in itprwc703/wc361 organisms (data not shown). Because plcβ21CP319 is a viable, weak hypomorph, its effect is presumably evident only in a sensitized genetic background that includes dgq and itpr mutant alleles. In contrast, norpAP24 is a null allele (Pearn et al., 1996). The absence of a genetic interaction between this allele and either itprwc703/wc361 or dgq221c/+; itprwc703/wc361 suggests that there is a nonredundant role for the plc21c gene product in adult wing posture and flight behavior.
Effects of dgq and plc21c mutants on flight physiology
The enhancement of flight behavior defects in dgq221c/plc21cP319; itprwc703/wc361 animals could arise from either the enhanced wing posture defect or a change in the neural circuits underlying flight behavior or both. To distinguish between these possibilities, the patterns of spontaneous firing and air puff-induced flight were recorded from the DLMs of flies of the genotype itprwc703/wc361 and dgq221c/+; itprwc703/wc361. As shown in Figure 5, A and B, the spontaneous firing rate recorded from the DLMs of itprwc703/wc361 flies is no different from that of wild-type flies. In contrast, the DLMs of dgq221c/+; itprwc703/wc361 animals exhibit mild endogenous hyperactivity (Fig. 5A,B). Interestingly, the rate of spontaneous firing was altered only minimally in dgq18745/+;itprwc703/wc361 animals (Fig. 5), whereas animals of the genotype dgq221c/plc21cP319; itprwc703/wc361 showed no additional increase in their spontaneous firing rate compared with dgq221c/+;itprwc703/wc361 animals (data not shown). These observations are consistent with previous results in which altered wing posture and increased spontaneous firing could be rescued independently (Banerjee et al., 2004). Next, all organisms were tested in the cylinder drop test and separated into two classes referred to as “fliers” and “nonfliers,” based on their performance in the test. The “flier” population of itprwc703/wc361 flies responded to an air puff with a rhythmic train of action potentials similar to that observed in wild-type flies. This rhythmic response lasts for 30 s or more. Nonfliers of the same genotype exhibit an initial rhythmic response after the air puff, but fail to maintain the response after 5 s or less (Fig. 5C). Nearly identical patterns of response were obtained from nonfliers of the genotypes dgq221c/+; itprwc703/wc361 and dgq18745/+; itprwc703/wc361. Moreover, the response from DLMs of triple mutant animals (dgq221c/plc21cP319; itprwc703/wc361) was also no different from itprwc703/wc361 (Fig. 5C). From these data, it can be inferred that the enhanced flight defects in dgq, itpr double mutants and dgq, plc21C, itpr triple mutants arise from a combination of the altered wing posture and changes in the neural circuit underlying flight.
Kum170 can suppress phenotypes enhanced by dgq and plc21C in itpr mutants
The results obtained so far show that itpr mutant phenotypes of altered wing posture and increased spontaneous firing from the DLMs can be suppressed by reducing SERCA activity and enhanced by reducing levels of Dgqα and PLC21C. These observations are consistent with the idea that the observed enhancement and suppression arise from opposing effects on the cellular output of InsP3 signaling viz. Ca2+ release from ER stores. To test this idea more rigorously, we asked whether the Kum170 mutation could suppress increased spontaneous firing from DLMs, wing posture, and flight defects in dgq221c/+;itprwc703/wc361 and dgq221c/plc21cP319;itprwc703/wc361 animals. As shown in Figure 4, the Kum170 mutation in SERCA could suppress wing posture and flight defects in dgq; itpr double mutants (dgq221c/Kum170; itprwc703/wc361) and in dgq/plc21c; itpr triple mutants (dgq221c/plc21cP319-Kum170; itprwc703/wc361). Furthermore, the enhanced spontaneous firing recorded from the DLMs of dgq221c/+; itprwc703/wc361 flies was also suppressed in dgq221c/Kum170; itprwc703/wc361 and dgq221c/plc21cP319-Kum170; itprwc703/wc361 flies (Fig. 5A,B). In all genotypes tested, flight patterns in response to a gentle air puff seen in nonflier animals were comparable with that of flight-defective itprwc703/wc361 animals (Fig. 5C).
Tissue specificity of Gαq–InsP3 receptor signaling
Although the data are consistent with a model in which all mutants act within the same cell (or cells) to eventually result in the observed phenotypes, given the pleiotropic nature of calcium signaling, the observed genetic interactions could arise from differential effects in multiple cell types. In a previous report, we showed that defective flight behavior and related phenotypes in itpr mutants can be rescued by expression of a wild-type itpr+ cDNA in aminergic interneurons (Banerjee et al., 2004). To test the cellular of focus of dgq enhancement of itpr mutant flight defects, a UASdsRNA construct was generated against the dgq gene (UASdgq1f1) (for details of the construct, see Materials and Methods). Ubiquitous expression of this construct with ArmGAL4 results in marginally lower levels of Dgqα3 in larvae (Fig. 3D). The UASdgq1f1 transgene was turned on specifically in aminergic neurons of itprwc703/wc361 organisms with the Dopa-decarboxylase GAL4 strain (DdcGAL4), which drives expression in all aminergic cells and neurons in Drosophila (Li et al., 2000). A significantly greater percentage of these flies are flightless (60%), compared with itprwc703/wc361 (30%) (Fig. 6B). However, flies expressing UASdgq1f1 in aminergic neurons do not show any change in wing posture and the rate of spontaneous firing (data not shown). Moreover, the introduction of UASdgq1f1 does not alter the air puff response of itprwc703/wc361 (compare Figs. 6D, 5C). These data suggest that the enhanced flight defects in dgq;itpr double mutants are primarily a consequence of reducing Dgqα3 levels in the DdcGAL4 domain, but the reduction in Dgqα3 is probably less than that obtained with a single copy of dgq mutant alleles. The absence of flight phenotypes in DdcGAL4/UASdgq1f1 organisms (Fig. 6B) is consistent with the observation that this transgene lowers Dgqα3 levels to a modest extent only (Fig. 3D). It also agrees with the phenotypes of dgq221c/+ and dgq18745/+ organisms, which exhibit normal flight behavior yet are able to enhance itpr mutant phenotypes in single copy (Fig. 4) (data not shown).
In an alternate approach, cell and tissue specificity of Gqα and InsP3 signaling was addressed by expression of an UASdgqα3+ transgene in the DdcGAL4 domain of dgq221c/+; itprwc703/wc361 animals (Fig. 6). In DdcGAL4-dgq221c/UASdgqα3+; itprwc703/wc361 animals, wing posture was rescued partially (Fig. 6A). The wing-rescued flies were selected visually and tested for flight in the flight column, in which ∼45% passed the flight test compared with 15% of dgq221c/+:itprwc703/wc361 animals (Fig. 6B). The rescued animals also showed a reduction in spontaneous firing (Fig. 6C). Flight patterns, in response to an air puff, were restored in the majority of fliers of the genotype DdcGAL4-dgq221c/UASdgqα3+; itprwc703/wc361 (Fig. 6D), similar to what is observed in fliers from itprwc703/wc361 (Fig. 5B). As controls, wing posture, flight, and related responses were measured in a number of related genotypes (Fig. 6). Interestingly, the expression of UASdgqα3+ in the DdcGAL4 domain resulted in a slightly defective wing posture (data not shown) suggesting that overexpression of Dgqα3 in DdcGAL4 cells can have deleterious effects. This observation also provides a possible explanation for why UASdgqα3+ rescue of flight in dgq221c/+; itprwc361/wc703 organisms is poor compared with rescue by the UASitpr+ transgene. Ubiquitous expression to a low level by the hsGAL4L driver at 25°C results in similar levels of rescue of flight by both transgenes (Fig. 6).
Intracellular Ca2+ release through the InsP3R is a feature of all multicellular organisms in which it is thought to shape the temporal and spatial aspects of calcium signaling in both excitable and nonexcitable cells (Berridge et al., 2003). The results presented here demonstrate that calcium signals generated through the InsP3R can have distinct attributes that lead to different phenotypes at the systemic level. First, the InsP3R elevates [Ca2+]i concentration in a manner that can be compensated in InsP3R mutants by decreasing the rate of calcium reentry back into the ER store. Second, it generates a class of Ca2+ signals that appear to depend on the intrinsic properties of the InsP3R. These are neither suppressed by slower entry of Ca2+ into the ER nor enhanced by reducing the activity of components that lie upstream of the InsP3R (supplemental Fig. 1, available at www.jneurosci.org as supplemental material). The precise nature of these two classes of calcium signals needs to be elucidated further. Intuitively, the two might activate separate sets of signaling molecules in the cytosol. For example, the information in a given Ca2+ signal derives both from the amplitude of Ca2+ elevation and from the frequency of Ca2+ oscillations (Berridge, 1993; Thorn et al., 1993a,b) that ensue. Although amplitude and frequency based signals can both be decoded by transcription factors (Dolmetsch et al., 1997, 1998) with differential sensitivity to Ca2+, oscillatory Ca2+ signals are decoded by cytosolic effectors like CaMKII (Ca2+/calmodulin-dependent protein kinase II) (De Koninck and Schulman, 1998). At a cellular level, oscillatory signals are known to control secretion (Hille et al., 1994) and neuronal differentiation (Spitzer et al., 2000). Our study shows that different Ca2+ signals generated through the InsP3R act at multiple levels to finally control systemic phenotypes such as flight behavior.
Interaction between the InsP3R and SERCA
An interaction between the InsP3R and SERCA is not unexpected but has not been demonstrated to our knowledge in vivo. Pharmacological blockage of the SERCA pump is known to cause store depletion and a gradual elevation of cytosolic Ca2+ because of an unopposed Ca2+ leak from intracellular Ca2+ stores (Thastrup et al., 1990; Mason et al., 1991). Although complete depletion of store Ca2+ would abolish intracellular Ca2+ release through the InsP3R on agonist stimulation, a partial blockage of SERCA activity may result in slower uptake of released Ca2+, leading to modest but prolonged elevation of [Ca2+]i. The fact that Kum170 suppresses rather than enhances InsP3R phenotypes suggests that this mutant allele (at the temperatures tested) slows down calcium uptake into ER stores after agonist stimulation. Slower return to basal [Ca2+]i after KCl depolarization supports this idea. It is possible that Kum170 also lowers luminal Ca2+ content in the ER. We have not directly measured the luminal Ca2+ concentration in Kum170/+ animals. However, because Kum170 did not enhance any of the described itpr phenotypes, it seems unlikely that luminal Ca2+ concentrations in this strain have fallen to a level that negatively impacts InsP3R function.
The reduced rate of Ca2+ uptake into the ER could compensate for reduced InsP3R function by different mechanisms. The simplest explanation, which we favor, is that slower Ca2+ uptake from the cytosol attributable to Kum170 compensates directly for attenuated Ca2+ release in the neurons of itprka1091/ug3 animals. This would be similar to the priming of oxytocin release on application of thapsigargin in the dendritic terminals of oxytocin neurons (Ludwig et al., 2002). However, at this stage, our results do not show unambiguously that the suppressor effect of Kum170 in itpr mutants is within the same cell or cells. A transgene encoding mutant Kum170 is available and experiments with this are in progress. Clearly, although reduced, SERCA activity in Kum170/+ animals can suppress several defects in itpr mutants including larval viability and certain aspects of development of the air puff-induced flight circuit responsible for wing posture, neuronal hyperactivity, and flight initiation. Our data suggest that, for the cellular processes underlying these phenotypes, maintaining a certain cytosolic level of [Ca2+]i on agonist stimulation is sufficient. Precisely which cellular process(es) is affected in itpr mutants remains unclear at this stage. It could include dendritic remodeling during pupal development (Duch and Levine, 2002), targeting of growth cones as observed in vitro in Xenopus (for review, see Gomez and Zheng, 2006) and the specification of neurotransmitters and their receptors during development of Xenopus (Borodinsky et al., 2004).
The failure of Kum170 to sustain a rhythmic train of action potentials (necessary for a normal flight response) suggests that this phenotype does not depend solely on an increase in cytosolic [Ca2+]i on InsP3-mediated Ca2+ release. Possibly sustained air puff-induced flight patterns are dependent on the intrinsic properties of the InsP3 receptor (as discussed further below).
Interestingly, the interplay of SERCA and InsP3R goes beyond the phenotypes described here. In vertebrate models of two neurodegenerative conditions, spinocerebellar ataxia (SPA) and Niemann–Pick A disease, a downregulation of InsP3R and SERCA is observed well before the manifestation of disease symptoms (Lin et al., 2000; Ginzburg and Futerman, 2005). Both diseases are characterized by the selective loss of Purkinje neurons in the cerebellum known to express high levels of InsP3R and SERCA. Several other proteins implicated in Ca2+ signaling remain unaffected in SPA (Lin et al., 2000). These data suggest the existence of a widespread mechanism for Ca2+ signaling in neurons, which requires both InsP3-mediated Ca2+ release and SERCA activity. Altering this might lead to neurodegeneration. The Drosophila model of altered flight behavior thus opens up a new avenue of investigating how these two proteins together control cellular Ca2+ levels and downstream Ca2+-dependent processes. The effect of itpr mutants on Ca-P60A levels and vice versa is currently under investigation.
The Gq-PLCβ pathway regulates InsP3 receptor function
The idea that the InsP3R can generate different classes of calcium signals, which result in specific cellular and systemic phenotypes, is not new. However, it has not been tested before in a systemic context. One possibility that we considered for the differential suppression of itpr phenotypes by Kum170 was that InsP3 is generated in the two contexts by the activation of either PLCγ or PLCβ. Both these isoforms of PLC are known to generate InsP3, leading to activation of the InsP3R in vertebrates and invertebrates (Rebecchi and Pentyala, 2000). Interestingly, our results in Drosophila do not support this idea, because none of the itpr mutant phenotypes was enhanced by reducing PLCγ activity (Table 1). Rather, they suggest that activation of the InsP3R is primarily through Gqα and PLCβ. All itpr mutant phenotypes, except for the maintenance of flight patterns, are enhanced by the dgq alleles described here. Because an extended flight pattern is also the one phenotype that cannot be suppressed by Kum170, we attribute this phenotype to an intrinsic property of the InsP3R.
From suppression of the dgq- and plc21C-mediated enhancement of itpr phenotypes by Kum170, we can be reasonably certain that changes in cytosolic Ca2+ levels are the basis of all the phenotypes. Our data thus support the idea that InsP3-mediated Ca2+ release is further compromised in itpr mutants by the introduction of dgq and plc21c alleles. However, we cannot rule out the possibility that dgq and plc21c mutants act through a parallel pathway such as the generation of DAG and reduced Ca2+ entry through a DAG-dependent plasma membrane Ca2+ channel (Hardie, 2003). Additional interaction studies with mutants in this arm of the pathway are required to test this possibility rigorously.
InsP3-mediated Ca2+ release and development of the flight circuit
Our analysis of flight in itpr mutants has demonstrated that InsP3 signaling in the DdcGAL4 domain, which consists primarily of aminergic interneurons, is necessary for development of the flight circuit (Banerjee et al., 2004). From loss of the rhythmic response to an air puff in itpr mutants, it seems likely that neurons sensitive to InsP3 signaling form part of the central pattern generator (CPG) for flight in Drosophila. Based on the genetic and phenotypic analysis presented here, we now propose that there are several steps in the formation of the flight CPG, which are differentially sensitive to InsP3 signaling. For example, the suppression of increased spontaneous firing, recorded from the DLMs of itpr mutants by the introduction of Kum170, is one such step. The importance of rhythmic spontaneous electrical activity in the development of activity in neural circuits is well known (Zhang and Poo, 2001). Another step is the connection(s) required for initiating flight patterns. Finally, to obtain normal air puff-induced flight, appropriate connectivity is necessary to maintain flight patterns for longer periods of 30 s or more. From the genetic paradigm used here, it is clear that the nature of the InsP3 signal at this step is distinct from the others. One possibility is that it requires a specific frequency and/or amplitude of calcium waves, which are dependent on the intrinsic calcium release properties of the InsP3R (supplemental Fig. 1, available at www.jneurosci.org as supplemental material). Elucidating what each step and phenotype denotes in terms of CPG connectivity and how this is altered in itpr mutants, is a major challenge and is likely to require several different approaches in the future.
This work was supported by core grants from the National Center for Biological Sciences and a grant from the Department of Biotechnology to G.H. We thank Aditya Venugopal for PCR mapping of the deletion in dgq221c as part of an undergraduate training program at the National Centre for Biological Sciences. We are grateful to Prof. V. Rodrigues and K. S. Krishnan for sharing reagents before publication and to Prof. M. K. Mathew for his help with data analysis. We are also grateful to Prof. C.-F. Wu, Dr. J. Lee, Prof. U. Bhalla, and Dr. R. Rajan for their help with the electrophysiology rig and data analysis.
R. Joshi’s present address: Department of Biochemistry and Molecular Biophysics, Columbia University Medical Center, HHSC1104, 701 West 168th Street, New York, NY 10032.
S. Srikanth’s present address: The Center for Blood Research Institute for Biomedical Research and Department of Pathology, Harvard Medical School, Boston, MA 02115.
- Correspondence should be addressed to Gaiti Hasan, National Center for Biological Sciences, Tata Institute of Fundamental Research, Bellary Road, Bangalore 560 065, India. Email: