Calcium (Ca2+) signaling is known to regulate the development, maintenance and modulation of activity in neuronal circuits that underlie organismal behavior. In Drosophila, intracellular Ca2+ signaling by the inositol 1,4,5-trisphosphate receptor and the store-operated channel (dOrai) regulates the formation and function of neuronal circuits that control flight. Here, we show that restoring InsP3R activity in insulin-producing neurons of flightless InsP3R mutants (itpr) during pupal development can rescue systemic flight ability. Expression of the store operated Ca2+ entry (SOCE) regulator dSTIM in insulin-producing neurons also suppresses compromised flight ability of InsP3R mutants suggesting that SOCE can compensate for impaired InsP3R function. Despite restricted expression of wild-type InsP3R and dSTIM in insulin-producing neurons, a global restoration of SOCE and store Ca2+ is observed in primary neuronal cultures from the itpr mutant. These results suggest that restoring InsP3R-mediated Ca2+ release and SOCE in a limited subset of neuromodulatory cells can influence systemic behaviors such as flight by regulating intracellular Ca2+ homeostasis in a large population of neurons through a non-cell-autonomous mechanism.
Cellular signals of ionic calcium (Ca2+) are generated using an extensive ‘Ca2+ toolkit' that consists of calcium channels and pumps on the plasma membrane and the membrane of intracellular stores that help in assembling signaling systems with very different temporal and spatial dynamics (Berridge et al., 2003). Cell surface receptor stimulation activates two closely coupled components of this toolkit; the inositol 1,4,5-trisphosphate receptors (InsP3R) followed by store operated Ca2+ (SOC) channels. InsP3Rs rapidly release Ca2+ from intracellular stores such as the endoplasmic reticulum (ER) while SOC channels function to replenish ER stores from the extracellular milieu and contribute to a longer-term Ca2+ signal. The plasma membrane four-transmembrane-spanning Orai protein was recently identified as the long-elusive SOC channel (Feske et al., 2006; Vig et al., 2006; Zhang et al., 2006). The exact mechanism of SOC entry (SOCE) is still under investigation but several reports have established that Ca2+ depletion of ER stores is sensed by STIM (stromal interaction molecule), an ER membrane protein, which upon store Ca2+ depletion oligomerizes in the ER, translocates close to the plasma membrane and organizes the Orai channel into clusters to bring about SOCE (Liou et al., 2005; Zhang et al., 2005; Cahalan, 2009). Recent findings strongly suggest that Orai and STIM together are necessary and sufficient for SOCE (Hewavitharana et al., 2007 and references therein).
Neuronal Ca2+ signals play an important role in the development, maintenance and modulation of activity in neural circuits (Berridge, 1998; Spitzer, 2002; Borodinsky et al., 2004). The importance of neuronal InsP3R-mediated Ca2+ signaling for neural circuit function is well recognized (Banerjee and Hasan, 2005; Mikoshiba, 2006). We have previously shown that flight and associated defects observed in itpr gene (which encodes the only Drosophila InsP3R) mutants can be rescued by restoring InsP3R function in aminergic neurons (Banerjee et al., 2004) and that InsP3R-mediated Ca2+ signaling is required at multiple steps for generating the neural circuit responsible for air-puff stimulated Drosophila flight (Banerjee et al., 2006). We also recently identified a requirement for SOCE in Drosophila neurons for regulating flight behavior (Venkiteswaran and Hasan, 2009).
Here, the contribution of intracellular Ca2+ signaling in function and development of the neural circuitry underlying Drosophila flight behavior, has been further investigated. We find that restoring InsP3R signaling in the insulin-producing neurons (referred to as DILP2; Drosophila insulin-like peptide 2 neurons) during pupal development can rescue flight and flight physiology defects in itpr mutants. The DILP2 neurons do not overlap with aminergic neurons in pupal and adult brains, though they are closely apposed. Interestingly, expressing wild-type dSTIM in either DILP2 or aminergic neurons can also rescue itpr mutant flight phenotypes suggesting that SOCE can compensate for compromised InsP3 signaling. Moreover, this rescue of systemic itpr mutant phenotypes by expression of wild-type InsP3R or dSTIM in DILP2 neurons results in a global restoration of neuronal calcium homeostasis. These results suggest that modulation of calcium signaling in a small neuronal subset can influence both cellular and systemic phenotypes.
Materials and Methods
Drosophila melanogaster strains.
itprka1091/ug3 and itprwc703/ug3 are heteroallelic combinations of 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 (Joshi et al., 2004). The UASdSTIM+ transgenic strain on chromosome 2 was generated by injecting embryos using a standard Drosophila embryo microinjection protocol with a pUASTdSTIM construct. The construct was generated from a full-length dSTIM cDNA clone (BDGP LD45776). The other strains used are as follows: Dilp2GAL4, provided by Dr. E. Rulifson (Rulifson et al., 2002), embryonic wild-type itpr cDNA (UASitpr+) (Venkatesh et al., 2001), GAL80ts with two inserts on the second chromosome (generated by Albert Chiang, NCBS, Bangalore, India), UASdOrai+ (Venkiteswaran and Hasan, 2009), DdcGAL4 referred in the text as aminergic (Li et al., 2000), a recombinant of Dilp2GAL4 and DdcGAL4 referred to as doubleGAL4 and UASmAChR (Venkiteswaran and Hasan, 2009). ElavC155GAL4 (referred in the text as pan-neuronal), UASdicer(III) and UASmCD8GFP(II) were obtained from the Bloomington Stock Centre, Bloomington, IN. UASRNAi strain for dOrai (12221) and dSTIM (47073) (referred in the text as dsdOrai and dsdSTIM, respectively) were obtained from the Vienna Drosophila RNAi Centre, Vienna, Austria (Dietzl et al., 2007), while that for itpr (1063-R2) referred in the text as dsitpr was obtained from the National Institute of Genetics Fly Stocks Centre. The other fly strains used were generated by standard genetic methods using individual mutant and transgenic fly lines described above.
Flight tests were performed as described previously (Banerjee et al., 2004), following minor modifications of the “cylinder drop assay″ (Benzer, 1973). Flies were tested in batches of 20 by dropping them into a 1-m-long glass cylinder. Flies that fell through directly into a chilled conical flask kept below were scored as nonfliers (NF), and those that flew and sat on the walls of the cylinder were scored as fliers (Fl). The percentage of fliers was then determined. Computation of mean and SEM was performed on results obtained from at least 100 flies, using Origin 7.5 software (MicroCal). Statistically significant differences between 2 groups were determined by Student's t tests for independent populations. Significant differences were taken at p < 0.05.
Electrophysiological and video recordings.
For electrophysiological and video recordings of responses to an air puff stimulus, flies were anesthetized briefly with diethyl ether or on ice and glued to a thin metal wire between the neck and the thorax with nail polish. To record air puff responses, a gentle mouth-blown air puff stimulus was delivered to the fly kept in a tethered condition. Physiological recordings were performed on the dorsal longitudinal muscles (DLMs) of the giant fiber pathway as described previously (Banerjee et al., 2004). After recovery from anesthesia, 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 5–10 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. Air puff-induced recordings were performed for 30 s. All recordings were made using an ISODAM8A (World Precision Instruments) amplifier with filter set up 30 Hz (low pass) to >10 kHz (high pass). Gap-free mode of pClamp8 (Molecular Devices) 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 7.5 software (MicroCal). Response to air puff and spontaneous activity were recorded from ∼10 or more flies for every genotype.
Immunohistochemistry was performed on Drosophila pupal and adult brains, expressing a membrane bound GFP (UASmCD8GFP) with the Dilp2GAL4, after fixing the dissected tissue in 4% paraformaldehyde. The following primary antibodies were used: rat anti-Ddc (1:400; provided by Dr. J. Hirsh, University of Virginia, VA), rabbit anti-GFP antibody (1:10,000; Invitrogen) and a monoclonal anti-serotonin antibody (NeoMarkers, 1:50; Lab Vision Corp.). Fluorescent secondary antibodies were used at a dilution of 1:400 as follows: anti-rabbit Alexa Fluor 488 and anti-rat Alexa Fluor 633 (Invitrogen), anti-mouse Rhodamine Red (Jackson ImmunoResearch Laboratories Inc.). Confocal analysis was performed on a Zeiss LSM 510 Meta microscope (Carl Zeiss) or an Olympus Confocal FV1000 microscope using 20× 0.9 numerical aperture (NA) or 63× 1.4 NA objectives. Confocal data were acquired as image stacks of separate channels and combined and visualized as three-dimensional projections using the LSM5 version 3.2/SP2 software (Carl Zeiss) or FV10-ASW 1.3 viewer (Olympus Corporation).
Primary neuronal cultures from Drosophila larvae.
Procedures for culturing Drosophila larval neurons have been described (Venkiteswaran and Hasan, 2009). Primary cultures of dissociated Drosophila larval neurons were plated in 200 μl of Drosophila M1 medium (30 mm HEPES/150 mm NaCl/5 mm KCl/1 mm MgCl2/1 mm CaCl2/35 mm sucrose, pH 7.2) supplemented with 10% Fetal Bovine Serum (Invitrogen), 50 U/ml penicillin (Invitrogen), 50 μg/ml streptomycin (Invitrogen), and 10 μg/ml amphotericin B (Invitrogen) as described previously (Wu et al., 1983). Briefly, brain and the ventral ganglion complex were dissected from Drosophila third instar larvae of the appropriate genotypes. The brain tissue was mechanically dissociated using syringe needles in Schneider's medium (Invitrogen) containing 0.75 μg/μl collagenase (Invitrogen) and 0.4 μg/μl dispase (Roche) and incubated in the proteolytic medium for 20 min to allow complete dissociation of the tissue. The lysate containing essentially single cells was spun down, resuspended in M1 medium (200 μl of M1 was used for lysates of four brains), and plated onto polylysine-coated coverslips which formed the base of 35 mm culture dishes. The cells were incubated at 22°C for 14–16 h before imaging. All chemicals for cell culture were obtained from Sigma-Aldrich unless mentioned otherwise.
Calcium imaging in larval neurons.
Calcium imaging in larval neuron cultures was performed after growth for 14–16 h. The cultures were washed with fresh Drosophila M1 medium twice and loaded in the dark with 2.5 μm Fluo-4 acetoxymethyl ester (AM; Invitrogen) in M1 medium containing 0.002% Pluronic F-127 (Sigma-Aldrich), for 30 min at room temperature. After washing 3 times with M1, the cells were finally covered with 100 μl of Ca2+-free M1 (CaCl2 was replaced with equal concentration of MgCl2) containing 0.5 mm EGTA 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. Fluo-4 was excited for 20 ms using 488 nm wavelength illuminations 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 Technology Corp.). Fluorescent images were acquired using the Evolution QExi CCD camera and In Vivo imaging software (Media Cybernetics). The time lapse acquisition mode of the software was used to follow fluorescence changes in the cells every 10 or 15 s for 15 frames. Different concentrations of carbachol (Sigma-Aldrich), 10 μm thapsigargin (Invitrogen), or 10 μm ionomycin (Calbiochem, Merck) were added manually ≈15 s after the start of data acquisition. For measurement of store-operated Ca2+ entry, 1 mm CaCl2 was added to the cells 225 s after thapsigargin addition. Images were acquired every 15 s. As controls, a series of images were acquired with the same imaging protocol without any additions. A total of 150–200 cells were imaged for each genotype for each experiment. All chemicals for calcium imaging were from Sigma-Aldrich unless mentioned otherwise.
For measuring fluorescence changes over time, images were processed using the ImagePro plus software, V1.33 (Media Cybernetics). Fluorescence intensity before (F′basal) and at various time points after addition of carbachol, thapsigargin, or CaCl2 (F′t) were determined. Background fluorescence (an area without any cells) was subtracted from the values of F′t and F′basal for each cell to obtain Ft and Fbasal. The data were plotted using Origin 6.0 software as follows: ΔF/F = (Ft − Fbasal)/Fbasal for every time point. The maximum value of ΔF/F was obtained for every cell and a box chart representing the data spread was plotted. The rectangular boxes represent the spread of data points between 25% and 75% of cells, the horizontal line is the median, and the small square within represents the mean. Significant differences between multiple groups of data were analyzed by one-way ANOVA test of significance as indicated in the figure legends. Significant differences were taken at PANOVA < 0.05.
Restoring itpr+ function in DILP2 neurons rescues flight defects of itpr mutants
Expression of the wild-type Drosophila InsP3R cDNA (UASitpr+) in the aminergic domain can rescue both larval and adult phenotypes of itpr mutants (Banerjee et al., 2004; Joshi et al., 2004). Larval itpr mutant phenotypes of growth, feeding and viability can also be rescued to a significant extent by itpr+ expression in the DILP2 neurons (Agrawal et al., 2009). To investigate the effect of restoring itpr function in the DILP2 neurons on adult itpr mutant phenotypes, UASitpr+ was expressed with Dilp2GAL4 (Rulifson et al., 2002) in adult viable itpr mutant heteroallelic combinations, itprka1091/ug3 and itprwc703/ug3. itprka1091 and itprwc703 alleles each have single point mutations in the modulatory domain of the InsP3R, while itprug3 has a single point mutation in the N-terminal ligand binding domain of the InsP3R (Joshi et al., 2004). Surprisingly, complete rescue of the altered wing posture in itprka1091/ug3 and itprwc703/ug3 was seen in UASitpr+/+;Dilp2GAL4/+;itprka1091/ug3 and UASitpr+/+;Dilp2GAL4/+;itprwc703/ug3 animals (Fig. 1A). Moreover, flight ability was rescued significantly as measured by the “cylinder drop test” assay (Fig. 1C). Flight defects are not significantly different between males and females of either itpr mutants or of itpr mutants expressing UASitpr+ with the Dilp2GAL4 (data not shown). The physiological correlate of loss of flight in itprka1091/ug3 and itprwc703/ug3 flies is an inability to initiate a rhythmic train of action potentials in response to an air-puff stimulus and a considerably higher frequency of spontaneous action potentials as measured by electrophysiological recordings from the dorsal longitudinal set of indirect flight muscles (DLMs) (Banerjee et al., 2004). The delivery of an air puff stimulus to itprka1091/ug3 flies expressing UASitpr+ with the Dilp2GAL4 led to a rhythmic generation of action potentials similar to what is observed in wild-type flies (Fig. 1B). Though, a majority of the flies tested had a sustained air puff response (10/16), in some flies (6/16) the response was more erratic and could not be maintained for more than five seconds. itprwc703/ug3 flies expressing UASitpr+ with the Dilp2GAL4 also led to rhythmic generation of action potentials immediately after the delivery of an air puff stimulus, but in most flies (14/16) the air puff response lasted for a shorter period of time compared with the control flies (Fig. 1B,D; supplemental Video 1, available at www.jneurosci.org as supplemental material). Increased spontaneous firing from the DLMs, which is characteristic of itprka1091/ug3 and itprwc703/ug3, was also significantly reduced in the Dilp2GAL4 rescued condition (Fig. 1E,F). Flight and associated physiology was not rescued in itpr mutants by expression of UASitpr+ in neuropeptidergic domains other than Dilp2GAL4. The neuropeptidergic domains tested were P386GAL4 (Taghert et al., 2001), c929GAL4 (Hewes et al., 2003), AKHGAL4 (Kim and Rulifson, 2004) and CCAPGAL4 (Park et al., 2003; data not shown).
These data suggest that DILP2 neurons can participate in the development and/or function of the air-puff stimulated flight circuit. Wing posture, flight and flight physiology can be rescued by ubiquitous expression of the wild-type itpr transgene during pupal development (Banerjee et al., 2004). To address the question of itpr gene requirement in the DILP2 neurons during pupal development the TARGET (temporal and regional gene expression targeting) system (McGuire et al., 2003) was used for regulating the temporal expression of UASitpr+ in the Dilp2GAL4 domain of itprka1091/ug3 organisms. The TARGET system includes a temperature-sensitive GAL80 element (GAL80ts) that regulates GAL4 in a temperature-dependent manner, with optimal repression observed at 19°C and derepression at 30°C (McGuire et al., 2004). Wing posture and air puff-induced flight (assessed in single fly tethered assays) were not rescued when UASitpr+/+; Dilp2GAL4/GAL80ts; itprka1091/ug3 animals were grown at 19°C during pupal and adult stages (Fig. 2A), but were completely rescued when these animals were grown at 30°C (Fig. 2B; supplemental Video 2, available at www.jneurosci.org as supplemental material). UASitpr+/+;Dilp2GAL4/GAL80ts;itprka1091/ug3 organisms, grown at 19°C through pupation and shifted to 30°C as adults, also had defective wing posture and flight ability (Fig. 2C; supplemental Video 3, available at www.jneurosci.org as supplemental material). However, UASitpr+/+;Dilp2GAL4/GAL80ts; itprka1091/ug3 organisms, grown at 30°C during pupal stages but shifted to 19°C as adults, had normal wing posture and air puff-induced flight (Fig. 2D; supplemental Video 4, available at www.jneurosci.org as supplemental material) indicating a requirement for itpr expression in the DILP2 neurons during pupal development, when the flight circuit is formed. An additional requirement during acute adult flight remains possible due to perdurance of the InsP3R in DILP2 neurons from pupae to adults.
Overexpression of the ER store calcium sensor dSTIM+ in DILP2 or aminergic neurons suppresses flight defects of itpr mutants
Certain adult flight phenotypes in itprka1091/ug3, such as wing posture, spontaneous firing frequency and flight initiation can be partially suppressed by expression of the wild-type cDNA for the SOC dOrai+ in DILP2 and aminergic neurons (Venkiteswaran and Hasan, 2009). Together with the rescue by itpr+ expression in these domains and its requirement during pupal development, the data suggest that Ca2+ release through the InsP3R and store-operated Ca2+ entry (SOCE) in DILP2 and aminergic neurons influence Drosophila flight circuit development to a significant extent.
The partial suppression of flight phenotypes by dOrai+ may be due to insufficient levels of the ER Ca2+ sensor STIM which works together with Orai to bring about SOCE (Cahalan, 2009). Coexpression of Orai and STIM results in SOCE which is significantly greater than that observed by the expression of either protein alone (Soboloff et al., 2006). This idea was tested further by coexpressing dSTIM+ and dOrai+ in the DILP2 and aminergic neurons of itprka1091/ug3 organisms. For this purpose transgenic strains were generated with dSTIM+. Coexpression of dSTIM+ and dOrai+ in DILP2 neurons completely suppressed the wing posture, flight and flight physiology defects of itprka1091/ug3 (Fig. 3A–E; supplemental Video 5, available at www.jneurosci.org as supplemental material). Simultaneous expression of dSTIM+ and dOrai+ in aminergic neurons resulted in pupal lethality (data not shown). Interestingly, expression of dSTIM+ alone in DILP2 or aminergic neurons also suppressed the altered wing posture of itprka1091/ug3 flies (Fig. 3A) as well as spontaneous hyperactivity of the DLMs (Fig. 3D,E). Importantly, expression of dSTIM+ in either DILP2 or aminergic neurons of itprka1091/ug3 flies restored flight ability and sustained air-puff-induced rhythmic flight patterns (Fig. 3B,C). This rescue of itpr mutant phenotypes was better than that observed on expression of UASdOrai+ in these domains, where a partial rescue of wing posture and a brief initiation of air-puff-induced flight was obtained (Venkiteswaran and Hasan, 2009). These results suggest that dSTIM+ expression on its own can efficiently restore intracellular calcium homeostasis in DILP2 neurons, and this restoration can bring about systemic suppression of itpr mutant phenotypes.
Pan-neuronal expression of dOrai+ and dSTIM+ in itprka1091/ug3 organisms suppresses wing-posture to varying extents but does not restore flight ability to levels comparable with Dilp2GAL4 and DdcGAL4 driven expression (supplemental Fig. 1, available at www.jneurosci.org as supplemental material). These data are consistent with previous observations that flight in itprka1091/ug3 is not rescued by pan-neuronal expression of itpr+ with the ElavGAL4 (Banerjee et al., 2004). Expression of itpr+ with ElavC155GAL4 (which has a stronger expression in the pupal stages than ElavGAL4; data not shown) can partially rescue wing posture and flight physiology defects but is insufficient for restoring flight ability (supplemental Fig. 1, available at www.jneurosci.org as supplemental material). Differences in temporal and spatial levels of pan-neuronal GAL4 expression coupled with a differential requirement for InsP3R function at multiple steps of flight circuit formation are a likely explanation for these data (see Discussion and Banerjee et al., 2006).
Relation between the Dilp2GAL4 and Ddc domains
The simplest explanation for the rescue of itpr mutant phenotypes by expression of UASitpr+ or UASdSTIM+ in aminergic and DILP2 neurons is an overlap between the two domains. However in larvae, where the two domains have a similar effect upon rescue of growth and viability, DILP2 and aminergic neurons do not overlap (Agrawal et al., 2009). To determine whether an overlap develops at later stages, a membrane bound GFP (UASmCD8GFP) was expressed with Dilp2GAL4 and pupal and adult brains of these animals were stained with an anti-Ddc antibody (Lundell and Hirsh, 1994).
Since the rescue by Dilp2GAL4 has a temporal focus during pupal development, a possible overlap of DILP- and Ddc-positive cells in the CNS of 36 h pupae was investigated (Fig. 4). No overlap between Ddc cells and the DILP2 neurons was observed in either the developing pupal brain (Fig. 4A–C) or the ventral ganglion (Fig. 4D–F). A cluster of Ddc-positive cells (Fig. 4A, red arrowheads) were present in close proximity to the DILP2 neurons and their processes project in close proximity to processes from insulin-producing cells. A cluster of Ddc-positive cells was observed in the subesophageal ganglion (Fig. 4A, red arrows) which are also seen in larval (Agrawal et al., 2009) and adult brains. However, no cells positive for both GFP and Ddc could be seen in the thoracic and abdominal ganglia, though a cluster of Dilp2::GFP cells in the thoracic ganglia (Fig. 4E, green arrowhead) had processes extending toward (Fig. 4E, green arrows) segmentally labeled Ddc cells in the abdominal ganglia (Fig. 4D, red arrowhead). The overlap of DdcGAL4 expression with the Ddc antibody used here was also ascertained. A few cells in the DdcGAL4 strain did not stain with anti-Ddc (supplemental Fig. 2, available at www.jneurosci.org as supplemental material). However, none of these were in the vicinity of the Dilp2GAL4-positive cells.
In the adult brain too, no overlap between Ddc and DILP2 neurons (Fig. 5C, green arrowhead) was observed (Fig. 5D,H, white arrowheads). Similar to observations in the developing pupal brain, Ddc-positive cells on the lateral part of the subesophageal ganglion (Fig. 5F,J, red arrowheads) had processes that extended toward the medial part of the subesophageal region (Fig. 5F,J, red arrows) in close proximity to where processes from DILP2 neurons terminate (Fig. 5C,G,K, green arrows). Though DdcGAL4 expresses in both serotonergic and dopaminergic neurons, itpr mutant phenotypes are not rescued by expression of UASitpr+ in the dopaminergic domain (with the THGAL4; Friggi-Grelin et al., 2003; our unpublished data) suggesting that rescue of itpr mutants is through serotonergic neurons in the context of the phenotypes under study. These lateral Ddc-positive cells also stained with the anti-serotonin antibody (Fig. 5I, blue arrowheads). Serotonin-positive varicosities were visible near the DILP2 neurons and on the processes extending from them (Fig. 5E, blue arrows). The thoracic and the abdominal ganglia did not show any cells labeled with both GFP and Ddc. These experiments indicate the absence of any detectable overlap between Dilp2 and DdcGAL4 domains in the CNS of pupae and adults.
Pan-neuronal but not tissue-specific knockdown of itpr, dSTIM, or dOrai recapitulates adult itpr mutant phenotypes
The rescue of flight and associated defects observed in itpr mutants independently by either Dilp2GAL4 and DdcGAL4 and the absence of any detectable overlap between them suggests the existence of nonoverlapping neuronal domains that might operate in parallel to regulate flight circuit development. To further understand the requirement of Ca2+ homeostasis in these domains for flight ability, specific knockdown of itpr, dSTIM and dOrai was attempted.
Ubiquitous reduction of InsP3R levels by expression of a dsitpr construct with the Actin5cGAL4 line is larval lethal, while itpr knockdown in either ElavC155GAL4, Dilp2GAL4 or DdcGAL4 domains had no observable effect on larval viability (Agrawal et al., 2009). Pan-neuronal knockdown of itpr however results in partial pupal lethality (data not shown). The surviving ElavC155GAL4; dsitpr flies exhibit wing posture defects (Fig. 6A), reminiscent of those observed in adult viable itpr heteroallelic mutant combinations. The spread out wing posture defect was further exacerbated on enhancing RNA interference by expressing a UASdicer transgene (Fig. 6A). Moreover, the wing posture defect was more pronounced in males compared with females, very likely due to differences in dosage compensation of ElavC155GAL4, an insert on the first chromosome (Fig. 6A). ElavC155GAL4;dsitpr flies exhibit significant levels of flight defects (Fig. 6B) and were unable to initiate rhythmic flight patterns in recordings from the DLMs in response to an air-puff stimulus (0/12; Fig. 6C).
However, no significant change in wing posture or flight ability was observed when dsitpr was expressed with either Dilp2GAL4 or DdcGAL4 (Fig. 6A,B) even when dsitpr expression was enhanced with a dicer transgene. A majority of these flies show a sustained air-puff response (Fig. 6C). Nevertheless, increased spontaneous firing from the DLMs was observed on both pan-neuronal or tissue-specific knockdown of itpr, suggesting that neuronal activity in the flight circuit is mildly perturbed due to reduced InsP3R function in either DILP2 or aminergic neurons (Fig. 6D,E). The effect of itpr knockdown simultaneously in both DILP2 and Ddc domains was tested by generating a recombinant strain from the Dilp2GAL4 and DdcGAL4 strains. This strain is referred to as DoubleGAL4 in Figure 6. Adult flies eclosed normally in the genotype with doubleGAL4 driven expression of dsitpr enhanced by UASdicer. The adults had normal wing posture (data not shown) and passed the flight column test (Fig. 6B). Electrophysiological recordings from the DLMs of these individuals exhibit a high rate of spontaneous firing as in the case of the individual GAL4s (Fig. 6D,E). However, unlike the individual GAL4s, 50% of flies in which the dsitpr transgene was driven by the double GAL4 strain, were unable to maintain flight beyond 15–20 s (Fig. 6C). Presumably, this level of flight maintenance is sufficient for passing the flight column test. Thus the combined knockdown of itpr in DILP2 and Ddc domains affects flight circuit function to a greater extent than knock down in the individual domains.
Pan-neuronal knockdown of either dSTIM or dOrai results in defects in wing posture, flight and flight physiology (Venkiteswaran and Hasan, 2009). However, in observations similar to those obtained on expression of dsitpr, no wing posture defects were observed on expression of dsdSTIM or dsdOrai with either the Dilp2GAL4 or DdcGAL4 (supplemental Fig. 3A, available at www.jneurosci.org as supplemental material). Moreover, the flight ability of these flies was not significantly different from controls (supplemental Fig. 3B, available at www.jneurosci.org as supplemental material). These results support the existence of multiple neural domains that can contribute in parallel to flight circuit development. Perturbing Ca2+ homeostasis in one or two neuronal domains may thus be insufficient for inducing strong flight defects.
Next, the nature of alterations in intracellular calcium signaling was investigated among the various genotypes studied. Pan-neuronal downregulation of the InsP3R by dsRNA is expected to reduce InsP3-mediated Ca2+-release as demonstrated for itpr mutant neurons. In addition, certain itpr mutants also show global dysregulation of intracellular Ca2+ homeostasis by altering [Ca2+]ER and SOCE (Venkiteswaran and Hasan, 2009). To assess Ca2+ homeostasis on pan-neuronal downregulation of itpr, [Ca2+]ER and SOCE was measured in primary neuronal cultures obtained from brains of ElavC155GAL4;dsitpr larvae (Fig. 6F–H). SOCE was monitored by Ca2+ imaging in Ca2+ add-back experiments, after depletion of ER stores with thapsigargin in very low external Ca2+. SOCE was reduced in neurons expressing dsRNA for itpr compared with controls (Fig. 6F–H, PANOVA < 0.05) similar to the results obtained with pan-neuronal expression of dsdSTIM or dsdOrai (Venkiteswaran and Hasan, 2009). However, store Ca2+ or [Ca2+]ER was similar between ElavC155GAL4;dsitpr neurons and controls (Fig. 6H). These data suggest that the primary effect of dsitpr is likely to be on the magnitude of InsP3-mediated Ca2+-release, followed by a milder change in SOCE.
Global restoration of calcium homeostasis on tissue-specific expression of itpr+ in itprka1091/ug3
So far, restoration of itpr+ activity in a small neuronal subset in an itpr mutant background is sufficient for systemic rescue of the observed physiological phenotypes. In neurons derived from itprka1091/ug3 multiple parameters of intracellular Ca2+ homeostasis including the magnitude of Ca2+ released through InsP3R, SOCE and store Ca2+ are perturbed (Venkiteswaran and Hasan, 2009). In the rescued organisms, restoration of Ca2+ homeostasis in either the DILP2 or aminergic neurons might be sufficient for the observed systemic rescue. Alternatively, itpr+ expression in these subneuronal domains may restore global Ca2+ homeostasis by non-cell-autonomous mechanisms. To delineate between these two possibilities, store Ca2+ and SOCE were measured in cultured neurons derived from larval brain and ventral ganglion complex in conditions where itpr+ expression was either pan-neuronal or in DILP2 or aminergic neurons.
Pan-neuronal expression of itpr+ in an itprka1091/ug3 background restored both SOCE and store Ca2+ to levels seen in wild-type (Fig. 7C,D). Interestingly, SOCE and store Ca2+ in neurons from itprka1091/ug3 larvae expressing itpr+ in either DILP2 or aminergic neurons was also similar to wild-type (Fig. 7A–D). The proportion of cells exhibiting detectable SOCE was restored upon expression of UASitpr+ under Ddc (68%) or Dilp2GAL4 (71%) compared with itprka1091/ug3 (<5%; Table 1). These results strongly suggest that Ca2+ signaling through the InsP3R in DILP2 and aminergic neurons helps maintain global intracellular Ca2+ homeostasis through non-cell-autonomous mechanisms.
Overexpressing dSTIM+ in itprka1091/ug3 DILP2 neurons restores global intracellular Ca2+ homeostasis
To understand the cellular basis of suppression of itprka1091/ug3 phenotypes by dSTIM+ expression in the DILP2 neurons, SOCE and store Ca2+ were measured in larval neuronal cultures generated from larvae expressing dSTIM+ or dOrai+ either individually or by coexpressing both transgenes in DILP2 neurons. SOCE was detectable in ∼75% of neurons upon overexpression either dOrai+ or dSTIM+ or both using Dilp2GAL4. The extent of SOCE and store Ca2+ in these neuronal populations was significantly higher than in neurons from itprka1091/ug3 (Fig. 8B,C; PANOVA ≪ 0.05). In comparison with wild-type, SOCE remained significantly lower in neurons from itprka1091/ug3 larvae expressing dOrai+ in DILP2 neurons (PANOVA < 0.05). In contrast, the extent of SOCE and store Ca2+ was similar to wild-type (PANOVA > 0.05), in the neuronal population on expression of dSTIM+ in DILP2 neurons. Thus, expression of dSTIM+ in DILP2 neurons can restore both flight and SOCE in itprka1091/ug3 close to wild-type levels. The significance of elevated store Ca2+ in itprka1091/ug3 is not clear though it might potentially affect neuronal function (see also Venkiteswaran and Hasan, 2009). In itprwc703/ug3 store Ca2+ appears normal while loss-of-flight is equivalent to itprka1091/ug3 suggesting that elevated store Ca2+ and flight deficits may not be related directly (data not shown).
In addition to normal SOCE, flight in itprka1091/ug3 depends upon the extent of Ca2+ release through the InsP3R (Venkiteswaran and Hasan, 2009). This parameter of intracellular Ca2+ signaling was measured by pan-neuronal expression of dSTIM+ and dOrai+ in neurons and simultaneous ectopic expression of the Drosophila muscarinic acetylcholine receptor (mAChR). The Drosophila mAChR responds to carbamylcholine (carbachol) with generation of InsP3 and Ca2+-release through the InsP3R (Millar et al., 1995). Using this assay it has been demonstrated that Ca2+-release is significantly attenuated through mutant itprka1091/ug3 channels (Venkiteswaran and Hasan, 2009). Pan-neuronal expression of either dSTIM+ or dOrai+ in the background of itprka1091/ug3 led to higher Ca2+ release on stimulation with two different concentrations of carbachol (Fig. 8D–G; PANOVA < 0.05). Direct measurement of Ca2+-release through the InsP3R in DILP2 and Ddc neurons is technically not feasible at present. However, based on these observations it is likely that expression of dSTIM+ in DILP2 and aminergic neurons of itprka1091/ug3 has a stimulatory effect on InsP3-mediated Ca2+ release. The mechanism of how STIM/Orai stimulates Ca2+-release through mutant InsP3Rs is under investigation.
In this study, we show that defective flight and associated physiology of itpr mutants can be rescued by expression of either itpr+ or dSTIM+ in DILP2 neurons as well as in nonoverlapping aminergic neurons. This rescue of systemic phenotypes is accompanied by pan-neuronal restoration of intracellular calcium homeostasis, specifically SOCE, to wild-type levels. Moreover, while pan-neuronal knock-down of the InsP3R mirrors itpr mutant phenotypes, knock-down of itpr+, dSTIM+ or dOrai+ by tissue-specific dsRNA expression has a minimal effect on flight ability. These results support an important role for InsP3R activity and SOCE in the function of DILP2 neurons and suggest that DILP2 and aminergic neurons can modulate neuronal function at systemic and cellular levels by non-cell-autonomous mechanisms.
Intracellular calcium homeostasis in DILP2 neurons and the control of flight circuit development
The rescue of flight in itpr mutants by itpr+ expression in DILP2 neurons during pupal development suggests a role for Drosophila DILPs in development of the flight circuit. Several studies in both invertebrates and vertebrates have demonstrated that insulin participates in the neuronal remodeling of developing circuits. Motoneurons that innervate the indirect flight muscles in Drosophila undergo dendritic and axonal remodeling during pupation with extensive dendritic outgrowth after 20 h APF till adult eclosion (Consoulas et al., 2002). DILP release from the DILP2 neurons may thus be modulated by intracellular Ca2+ signaling and function in the process of dendritic outgrowth during the pupal stages. In Drosophila, insulin signaling contributes to guidance of photoreceptor cell axons from the retina to the brain during development of the visual system (Song et al., 2003). Insulin transport into the CNS appears to increase in the neonatal period (Banks, 2004) in mammalian systems, suggesting that peripheral insulin levels influence the formation of circuits in the developing brain (Needleman and McAllister, 2008).
Suppression of flight defects by expression of the SOCE regulator dSTIM+ in DILP2 and aminergic neurons suggests that InsP3R-mediated Ca2+-release and SOCE cooperate to maintain intracellular Ca2+ homeostasis in these neurons, thus enabling dSTIM+ to compensate for impaired InsP3R function. The partial compensation by dOrai+ may be a consequence of reduced expression in the transgenic strain used. Alternately, STIM can also independently function in the regulation of basal and ER Ca2+ levels (Brandman et al., 2007) and this additional role may be relevant in improved maintenance of calcium homeostasis.
Despite rescue of adult itpr mutant phenotypes by aminergic and DILP2 neurons to similar extents, the two cellular domains exhibit no obvious overlap in either pupal or adult brains, indicating that the rescues are not mediated by a shared neuronal subset and suggests the existence of parallel modes of rescue. This is also supported by pan-neuronal knockdown of itpr, dSTIM or dOrai which results in wing posture and flight defects, but a similar effect is not seen on specific knockdown of these molecules in DILP2 or aminergic neurons. Thus perturbing intracellular calcium homeostasis in one neuronal domain is insufficient for inducing flight defects possibly due to compensation by other domains. The mildly enhanced phenotypes observed in the doubleGAL4 itpr knock-down support an interaction between the two expression domains. However, some of these results could also be a consequence of multiple modes of intracellular calcium regulation by alternate pathways in the DILP2 and aminergic neurons. Our studies suggest that both aminergic and DILP2 neurons participate in development of the air-puff stimulated flight network in Drosophila. The existence of connections between serotonin and insulin signaling in Drosophila has been demonstrated as Drosophila Nucleostemin NS3 was found to act specifically within serotonergic neurons to regulate insulin signaling and exert global control over cell size and number (Kaplan et al., 2008). A similar mechanism might exist in the development and/or function of the flight circuit. However, a complete understanding of the individual neuronal components that constitute this flight network is currently lacking. Therefore, at this stage it is difficult to establish whether these domains physically constitute the flight circuit or indirectly influence its formation or function through secreted neuromodulators.
Similar systemic phenotypes are seen by knock-down of itpr, dOrai or dSTIM in the pan-neuronal domain (Venkiteswaran and Hasan, 2009) (Fig. 6). Pan-neuronal expression of wild-type itpr, dOrai and dSTIM however do not rescue itpr mutants to the same extent as rescue from the subneuronal domains, suggesting a developmental hierarchy for itpr mutant phenotypes beginning with wing posture. The rescue of individual phenotypes depends upon the level of pan-neuronal GAL4 expression at different stages of pupal development; normal wing posture is necessary but insufficient for restoration of flight physiology (Banerjee et al., 2004, 2006). Thus, unlike complete rescue, which requires GAL4 driven expression throughout flight circuit development, pan-neuronal knock-down at an early stage of pupal development produces wing posture defects, which abrogate normal flight. Data from the rescue of other adult neuronal phenotypes support the idea of a reduction in ElavC155GAL4 expression through pupal development (Zars et al., 2000) and a decrease in pan-neuronal GAL4 expression from larvae to adults is consistent with our observations of itpr rescue and RNAi knockdown.
Correlating intracellular calcium homeostasis with flight behavior
We show that SOCE and store Ca2+are restored to wild-type levels in cultured larval neurons from itprka1091/ug3 in conditions where flight and flight physiology are also rescued. Thus, expression of dSTIM+ alone or dSTIM+ and dOrai+ simultaneously with Dilp2GAL4 suppresses flight and associated physiological defects while expression of dOrai+ alone in the DILP neurons partially rescues wing posture and flight physiology but does not rescue flight ability and the maintenance of air-puff stimulated flight patterns (Venkiteswaran and Hasan, 2009). Interestingly, in neurons derived from Dilp2GAL4/UASdOrai+; itprka1091/ug3, SOCE remains significantly different from wild-type. Though correlative, these observations support the idea that wing posture defects, increased frequency of spontaneous firing and the inability to initiate flight can be suppressed by partial restoration of intracellular Ca2+ homeostasis in neurons. However, the maintenance of flight patterns and flight ability require normal SOCE in addition to the above parameters. The results obtained in this study parallel our previous observations, where a comprehensive rescue of flight and flight physiology is observed when InsP3-stimulated Ca2+-release and SOCE is brought back to wild-type levels in itprka1091/ug3 (Venkiteswaran and Hasan, 2009).
A direct link between InsP3-mediated Ca2+ release and SOCE has been contested in vertebrates based primarily on data from DT40 cells in which all three InsP3Rs have been deleted (Broad et al., 2001; Ma et al., 2002). Interestingly, in Drosophila macrophages, SOCE through dOrai is coupled to intracellular Ca2+ release through the ryanodine receptor (RyR; Cuttell et al., 2008). Thus in Drosophila there appears to exist cell-type-specific coupling of either InsP3R- or RyR-mediated Ca2+-release to SOCE. A similar dichotomy may exist among the kind of intracellular Ca2+-release channel used to trigger SOCE in different cell types in vertebrates. The extent and mechanism of activation of SOCE, by InsP3-mediated Ca2+ release, in Drosophila neurons needs further elucidation. These experiments will also help in understanding how dSTIM overexpression modulates mutant InsP3R function in cell-autonomous and non-cell-autonomous modes. At present these mechanisms can only be conjectured, since there does not appear to be a direct effect of itpr mutants and the rescued genotypes on dOrai and dSTIM levels (supplemental Fig. 4, available at www.jneurosci.org as supplemental material). Cell-autonomous rescue of mutant itpr function by dSTIM overexpression may be through the observed changes in luminal and/or cytosolic [Ca2+]. Non-cell-autonomous mechanisms are discussed below.
Non-cell-autonomous role of DILP2 neurons and aminergic neurons in regulating flight and calcium homeostasis
Ca2+ measurements have been performed from a non-selective population of larval neurons in this study. The numbers of DILP or aminergic neurons in such conditions are a very small percentage of the total number of neurons analyzed (data not shown). Therefore, these results demonstrate that expression of itpr+ or dSTIM+ in a limited set of DILP2- or Ddc-expressing neurons restores calcium homeostasis in a wider number of unrelated neurons. As both these domains are predominantly neurosecretory and synthesize and release either insulin-like peptides or monoamines, it is likely that they can influence intracellular calcium homeostasis by secreted neuromodulators.
High levels of Drosophila insulin receptor (dIR) mRNA are present in the larval and adult nervous system (Garofalo and Rosen, 1988) and dIR protein has been localized in the larval brain (Gorczyca et al., 1993). Drosophila serotonin receptors 5-HT1BDro (d5-HT1B) and 5-HT2Dro have been observed in both larval and adult brains (Yuan et al., 2005; Nichols, 2007). Therefore, a number of neurons could be targeted by secreted DILP or monoamines which might indirectly regulate the development and/or function of other neurons in the flight circuit.
Similar circuit outputs can be generated by multiple mechanisms (Marder and Goaillard, 2006). In the DdcGAL4 and Dilp2GAL4 rescue conditions the system may not be restored to a wild-type state at a cellular level but instead achieve new stable states in which a wild-type output is preserved (Greenspan, 2001). Our results demonstrate that for Drosophila flight, cellular calcium homeostasis in individual cells needs to be restored to a wild-type state and this then ensures the preservation of a normal physiological output. These observations are significant in the context of diseases where intracellular calcium signaling is deranged such as certain neurodegenerative and metabolic disorders (Verkhratsky, 2005; van de Leemput et al., 2007; Bezprozvanny and Mattson, 2008; Chan et al., 2009).
This work was supported by core funding from the National Centre for Biological Sciences and a grant from the Department of Biotechnology to G. H. We thank Swarna Mathre for help with pupal dissections, Shalini V. for making the UASdSTIM construct, and Dr. H. Krishnamurthy and the National Centre for Biological Sciences Central Image-Flow Facility for help with imaging.
- Correspondence should be addressed to: Gaiti Hasan, National Centre for Biological Sciences, Tata Institute of Fundamental Research, Bangalore 560065, India.