Functional Properties of the Drosophila melanogaster Inositol 1,4,5-Trisphosphate Receptor Mutants

Abstract

The inositol (1,4,5)-trisphosphate receptor (InsP₃R) is an intracellular calcium (Ca²⁺) release channel that plays a crucial role in cell signaling. In Drosophila melanogaster a single InsP₃R gene (itpr) encodes a protein (DmInsP₃R) that is ∼60% conserved with mammalian InsP₃Rs. A number of itpr mutant alleles have been identified in genetic screens and studied for their effect on development and physiology. However, the functional properties of wild-type or mutant DmInsP₃Rs have never been described. Here we use the planar lipid bilayer reconstitution technique to describe single-channel properties of embryonic and adult head DmInsP₃R splice variants. The three mutants chosen in this study reside in each of the three structural domains of the DmInsP₃R—the amino-terminal ligand binding domain (ug3), the middle-coupling domain (wc703), and the channel-forming region (ka901). We discovered that 1), the major functional properties of DmInsP₃R (conductance, gating, and sensitivity to InsP₃ and Ca²⁺) are remarkably conserved with the mammalian InsP₃R1; 2), single-channel conductance of the adult head DmInsP₃R isoform is 89 pS and the embryonic DmInsP₃R isoform is 70 pS; 3), ug3 mutation affects sensitivity of the DmInsP₃Rs to activation by InsP₃, but not their InsP₃-binding properties; 4), wc703 channels have increased sensitivity to modulation by Ca²⁺; and 5), homomeric ka901 channels are not functional. We correlated the results obtained in planar lipid bilayer experiments with measurements of InsP₃-induced Ca²⁺ fluxes in microsomes isolated from wild-type and heterozygous itpr mutants. Our study validates the use of D. melanogaster as an appropriate model for InsP₃R structure-function studies and provides novel insights into the fundamental mechanisms of the InsP₃R function.

INTRODUCTION

The inositol (1,4,5)-trisphosphate receptor (InsP₃R) is an intracellular calcium (Ca²⁺) release channel that plays a critical role in Ca²⁺ signaling (Berridge, 1993). Three mammalian isoforms of the InsP₃R—type I (InsP₃R1), type II (InsP₃R2), and type III (InsP₃R3)—have been identified (Furuichi et al., 1994). All three mammalian isoforms share a common domain structure and 60–70% sequence identity (Furuichi et al., 1994). The differences in functional properties between mammalian InsP₃R isoforms are beginning to be elucidated (Miyakawa et al., 1999; Thrower et al., 2001). InsP₃Rs are subjected to multiple levels of regulation (Berridge, 1993; Bezprozvanny and Ehrlich, 1995; Taylor, 1998). Structural determinants responsible for InsP₃R1 conductance and gating properties (Boehning et al., 2001; Ramos-Franco et al., 1999) and modulation by Ca²⁺ (Miyakawa et al., 2001; Nosyreva et al., 2002; Tu et al., 2003), ATP (Tu et al., 2002), and phosphorylation (Tang et al., 2003b; Wagner et al., 2003) have been uncovered in recent functional experiments with InsP₃R1 wild-type and mutant channels.

A single copy of the InsP₃R gene (itpr) is present in the genome of Drosophila (Hasan and Rosbash, 1992; Yoshikawa et al., 1992), a model organism, that has been used widely for genetic screens (Myers et al., 2000). Sequence analysis reveals that the Drosophila InsP₃R protein (DmInsP₃R) shares the same domain structure and ∼60% sequence identity with mammalian InsP₃R isoforms, with the highest level of conservation in the amino-terminal ligand binding and the carboxy-terminal channel regions. Thus, it is likely that analysis of the DmInsP₃R mutant isoforms, identified in genetic screens (Deshpande et al., 2000; Joshi et al., 2004; Venkatesh and Hasan, 1997), will provide important insights into fundamental mechanisms of the InsP₃R function. However, in the absence of information regarding the functional properties of the wild-type DmInsP₃R isoforms, the interpretation of results obtained from DmInsP₃R mutants are limited.

Here we describe the functional properties of two wild-type DmInsP₃R splice variants (Sinha and Hasan, 1999) and three DmInsP₃R mutants (ug3, wc703, and ka901) (Deshpande et al., 2000; Joshi et al., 2004). The mutant isoforms affect residues conserved among all known InsP₃Rs (Joshi et al., 2004). For functional analysis of DmInsP₃R we expressed DmInsP₃R wild-type and mutants in sf9 cells by baculoviral infection and analyzed the properties of resulting channels in planar lipid bilayers. This is the same approach that we used previously for structure-function analysis of the mammalian InsP₃R1 (Nosyreva et al., 2002; Tang et al., 2003a, b; Tu et al., 2002, 2003). Our results indicate that the main functional properties of InsP₃Rs are conserved in evolution from flies to mammals. Moreover, changes in the DmInsP₃R properties induced by mutations provide novel insights into fundamental mechanisms of InsP₃R function. We also demonstrate that functional properties observed in vitro correlate well with in vivo function of DmInsP₃R mutants by measurement of Ca²⁺ fluxes from wild-type and heterozygous mutant flies.

MATERIALS AND METHODS

Ca²⁺ release assay

A modified protocol from Bramley et al. (1990) was used. 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. The Canton-S strain (CS) was used for studying properties of the wild-type InsP₃R and all point mutants were generated in the background of this wild-type strain. The tissues were homogenized in TSHMo buffer (10 mM Tris pH 7.4, 0.25 M sucrose, 12 mM monothioglycerol, 10 mM sodium molybdate, 200 μM CaCl₂) and a mixture of protease inhibitors (1 mM PMSF, 10 μg/ml leupeptin, and 10 μM E64) by 10–12 strokes of a 15-ml glass Teflon homogenizer. The homogenate was centrifuged immediately at 20,000 g for 20 min at 4°C. The supernatant was recentrifuged at 109,000 g on a tabletop ultracentrifuge (Rotor TLA 45 of Beckman centrifuges, Beckman Coulter, Fullerton, CA) for 1 h. The pellet was resuspended in fresh Ca²⁺-free TSHMo buffer and centrifuged again at 109,000 g for 20 min. The pellet was resuspended in 150 μl of Ca²⁺-free TSHMo buffer and used for Ca²⁺ release assays and for protein estimation by the Bradford's method. 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 (see below for details of Ca²⁺-EGTA standard curve). The endogenous leak rate of each microsomal vesicle's preparation was checked by a standard fluorescence assay (Mathew et al., 1982, data not shown) before using them in the Ca²⁺-release assay. Only vesicles with minimal Ca²⁺ leak, measured over a period of 10 min, were used for the Ca²⁺ release assay. No obvious difference in the leak rates between wild-type and mutant vesicles was observed.

Membrane-impermeant Ca²⁺-sensitive ratiometric fluorescent dye, Fura-2 (Molecular Probes, Eugene, OR) was prepared in calcium-free water. Steady-state fluorescence measurement was 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 or ∼50 μg of larval microsomal vesicles 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. To take into account the buffering capacity of the microsomes, the standard curve was obtained for each experiment in the presence of wild-type or mutant microsomes, made in parallel in Ca²⁺-free buffer. For heparin-mediated inhibition studies, a nonratiometric dye, Fluo-4, was used at 5 μM concentration. To examine the amount of Ca²⁺ trapped in microsomal vesicles of all genotypes fluorescence measurements were done before and after the addition of 0.1% Triton X-100, which was used for disrupting the microsomal vesicles, in the presence of 2 μM Fura-2.

Expression of the DmInsP₃R in sf9 cells

Full-length embryonic DmInsP₃R coding sequence (Sinha and Hasan, 1999) was cloned into pFastBac1 expression vector (Invitrogen, Carlsbad, CA). The 5′ UTR of the DmInsP₃R clone was replaced with the Kozak sequence by polymerase chain reaction to generate the DmInsP₃R-pFastBac1 construct. The Nhe1(1415)/Nhe1(3210) fragment of the DmInsP₃R-pFastBac1 clone was replaced by the corresponding fragment from an adult head (AH) AH-DmInsP₃R clone (Sinha and Hasan, 1999) to yield the AH-DmInsP₃R-pFastBac1 construct. Recently described ka901 (G2630S), wc703 (G2117E), and ug3 (S224F) mutations (Joshi et al., 2004) were generated in the DmInsP₃R-pFastBac1 construct by megaprimer polymerase chain reaction or by using the QuikChange Site-Directed Mutagenesis kit (Stratagene, La Jolla, CA), and verified by sequencing. DmInsP₃R, AH-DmInsP₃R, ka901, wc703, and ug3 baculoviruses were generated and amplified using Bac-to-Bac system (Invitrogen) as described previously (Nosyreva et al., 2002; Tu et al., 2002). The DmInsP₃R wild-type and mutants were expressed in Spodoptera frugiperda (sf9) cells as previously described (Nosyreva et al., 2002; Tu et al., 2002). Expression of the DmInsP₃R in wild-type and mutant flies and sf9 cells was confirmed by Western blotting with the affinity purified anti-DmInsP₃R rabbit polyclonal antibody (IB-9075) raised against KLH-conjugated peptide CEQRKQKQRLGLLNTTANSLLPFQ derived from the DmInsP₃R sequence. Lysates from D. melanogaster larvae and adult heads were used as controls for the specificity of IB-9075 antibody (see Results).

Single-channel recordings and analysis of the DmInsP₃R activity

Recombinant DmInsP₃R wild-type and mutants expressed in sf9 cells were incorporated into the bilayer by microsomal vesicle fusion as previously described for InsP₃RI (Nosyreva et al., 2002; Tu et al., 2002). The cis chamber contained 250 mM HEPES-Tris, pH 7.35 and the trans chamber, which was held at virtual ground, contained 55 mM Ba(OH)₂, dissolved in 250 mM HEPES, pH 7.35. Single-channel conductance values of the DmInsP₃R were determined from the slope of a linear fit to unitary current amplitude versus transmembrane voltage data in the range between +10 mV and −30 mV. For calcium-dependence experiments, free Ca²⁺ concentration in the cis chamber was controlled in the range of 10 nM (pCa 8) to 10 mM (pCa 2) by mixture of 1 mM EGTA, 1 mM HEDTA, and variable concentrations of CaCl₂ or Ca-HEPES₂. To construct InsP₃- and Ca²⁺-dependence curves for the DmInsP₃R, the determined values of P_o were averaged across several independent experiments at each InsP₃ or Ca²⁺ concentration. For InsP₃-dependence experiments the averaged values of P_o are presented as mean ± SE (n = number of independent experiments) and fit by equation modified from Lupu et al. (1998), where P_max is a maximal P_o value, n is a Hill coefficient, and k_InsP3 is the apparent affinity of the DmInsP₃R for InsP₃. For Ca²⁺-dependence experiments the averaged values of P_o are presented as mean ± SE (n = number of independent experiments) and fit by the bell-shaped equation , modified from Bezprozvanny et al. (1991), where P_m is a parameter proportional to the maximal P_o value, n is a Hill coefficient, k is the apparent affinity of Ca²⁺ activating site, and K is the apparent affinity of Ca²⁺ inhibitory site (Tu et al., 2003). The parameters of the optimal fits to InsP₃- and Ca²⁺-dependence are in the text.

³H-InsP₃ binding assay

Scatchard analysis of ³H-InsP₃ binding to the DmInsP₃R and ug3 microsomes was performed as previously described for InsP₃RI (Kaznacheyeva et al., 1998). Briefly, microsomes (20 μg protein) were incubated on ice with 10 nM ³H-InsP₃ and variable amounts of nonlabeled InsP₃ (from 10 nM to 8 μM) in the binding buffer (50 mM Tris-HCl, pH 9.0, 1 mM EDTA, 1 mM DTT, 100 mM NaCl) and precipitated with 12.5% PEG and 1.2 mg/ml γ-globulin at 14,000 g. Precipitates were quickly washed with the binding buffer, dissolved in Soluene (PerkinElmer, Boston, MA) and their [³H] content was determined by liquid scintillation counting. Nonspecific counts, determined in the presence of 25 μM unlabeled InsP₃, were subtracted from the total to yield specific binding. The binding affinity and the density of specific ³H-InsP₃ binding sites was determined from the linear fit to the Scatchard plot. The specific activity of ³H-InsP₃ used in this assay (Amersham Biosciences, Piscataway, NJ) was 45 Ci/mmol.

RESULTS

Ca²⁺ release properties of the DmInsP₃R wild-type and mutants

The domain structure of the InsP₃R (Mignery and Sudhof, 1990) is conserved between the DmInsP₃R and mammalian InsP₃R isoforms. Similar to mammalian InsP₃Rs, the DmInsP₃R can be divided into an amino-terminal ligand binding (DmInsP₃R-N, M1-H650) domain, a carboxy-terminal channel-forming (DmInsP₃R-C, S2360-Q2829) domain, and a middle-coupling (DmInsP₃R-M, N651-W2359) domain (Fig. 1). Two alternate splice events (at positions 980 and 1390) were originally described for the embryonic and adult head (AH) forms of the DmInsP₃R (Sinha and Hasan, 1999). From subsequent sequencing of genomic DNA of the itpr gene it is now clear that alternate splicing occurs only at residue 980 to give the two forms shown in Fig. 1, which differ by the nine-amino-acid insertion in the AH isoform. There is no intron near residue 1390, and consequently both forms are identical in this region and contain the sequence GVGHSV. A number of lethal DmInsP₃R mutants were recently obtained by chemical mutagenesis using ethyl methane sulfonate (Deshpande et al., 2000). Among 14 lethal mutant alleles of the DmInsP₃R obtained in the ethyl methane sulfonate screen, genomic DNA in the region of the open reading frame has recently been sequenced from seven alleles (Joshi et al., 2004). In this article we focus on three DmInsP₃R point mutants (ug3, wc703, and ka901; see Fig. 1) in which the altered residues are conserved between all known InsP₃Rs (Joshi et al., 2004).

FIGURE 1.

Schematic of the DmInsP₃R showing point mutants used in this study. The amino-terminal ligand-binding domain (DmInsP₃R-N, M1-H650) consists of an inhibitory domain (ID, M1–E231) and the core-binding domain (CD, H232–H650). The middle-coupling domain (DmInsP₃R-M, N651–W2359) contains a site with a nine-amino-acid insertion in the adult head (AH) isoform in position E980 (Sinha and Hasan, 1999) and a putative Ca²⁺-sensor region (Cas, G1986–S2354). The carboxy-terminal channel-forming domain (DmInsP₃R-C, S2360–Q2829) consists of M1–M6 transmembrane domains and a pore-forming region (P). Also shown are positions of three DmInsP₃R point mutants identified in an ethyl methane sulfonate screen (Joshi et al., 2004), i.e., ug3 (S224F), wc703 (G2117E), and ka901 (G2630S).

To determine consequences of identified mutations for the DmInsP₃R function, a series of InsP₃-induced Ca²⁺ release experiments were performed with microsomes prepared from adult head and larval homogenates of wild-type and mutant flies. For generation of Ca²⁺-loaded microsomal vesicles the buffer contained known amounts of free Ca²⁺ (Materials and Methods). Mutant vesicles were obtained from heterozygous flies, due to the fact that all homozygous mutant combinations are lethal. Since the functional InsP₃ receptor is a tetramer, in which interactions between monomers are known to affect Ca²⁺ release, it was reasoned that heterozygotes might exhibit mutant properties. In these experiments DmInsP₃R function was quantified by measuring the InsP₃ concentration required to achieve half-maximal Ca²⁺-release (K_m, nM InsP₃) and maximal amount of released Ca²⁺ (R_max, nmol Ca²⁺/mg microsomal protein) values. As seen in Fig. 2 A, increasing amounts of InsP₃ showed an increase in fluorescence which was converted to the amount of Ca²⁺ released based on a calibration curve generated using a standard Ca-EGTA buffering system. The amount of Ca²⁺ released with increasing amounts of InsP₃ was seen to follow typical Michaelis-Menten kinetics (Fig. 2 B). We found that in wild-type organisms, the K_m values are similar in microsomes from larvae (405 ± 8 nM InsP₃) and adult heads (398 ± 14 nM InsP₃) (Fig. 2 B). In contrast, the R_max value is much higher in microsomes from adult heads (635 ± 42 nmol Ca²⁺/mg protein) than from larvae (319 ± 22 nmol Ca²⁺/mg protein; see Fig. 2 B). The K_m and R_max values were calculated as the average from at least three independent experiments. The graphs on Fig. 2 B show a single fit to the averaged data points.

FIGURE 2.

Ca²⁺ fluxes supported by the DmInsP₃R mutants. (A) Representative traces of Ca²⁺ release assay in microsomes isolated from the wild-type adult heads in the absence of InsP₃ (trace a) and in the presence of 20 nM (trace b), 200 nM (trace c), and 2 μM (trace d) InsP₃. The fluorescence values at 340 and 380 nm were utilized to calculate the amount of Ca²⁺ released (see Materials and Methods). Further, the free Ca²⁺ present without addition of InsP₃ (trace a) was subtracted from those with InsP₃ additions. (B) InsP₃-dependent Ca²⁺ fluxes in microsomal vesicles isolated from whole larvae (○) and adult heads (•) of the wild-type Canton-S (CS) strain. (C) Inhibition of InsP₃-mediated Ca²⁺ − release by heparin. Arrows indicate the time of addition of various reagents to microsomal vesicles in a buffer containing the fluorescent Ca²⁺ indicator dye Fluo-4. The two traces shown in black (− heparin) and gray (+ heparin) were obtained with the same batch of microsomes made from heads of the wild-type CS strain. The blips observed close to each arrow occur due to interference of the light beam by the syringe used to add reagents to the cuvette. Similar results have been observed independently with two other batches of microsomal vesicles. (D–F) InsP₃-dependent Ca²⁺ fluxes in microsomes isolated from adult heads (•) of ug3/+ (D), wc703/+ (E), and ka901/+ (F) flies. Wild-type adult head data from B are shown for comparison by the thin line in each graph. The difference in K_m between the wild-type and ug3/+ heterozygotes in D is statistically significant (P < 0.001). The concentrations of InsP₃ required for half-maximal Ca²⁺ release (K_m) are shown by arrowheads for wild-type and lines for mutants in B and D–F. All data in B and D–F are shown as the mean ± SD from at least three independent batches of microsomes. The data were fit to an exponential rise using SigmaPlot5 (SPSS, Chicago, IL).

The analysis of DmInsP₃R mutants was carried out using AH microsomes, which release more Ca²⁺ (twofold higher R_max) than the larval microsomes (Fig. 2 B). Qualitatively similar results were obtained in experiments with microsomes from mutant larvae (data not shown). The specificity of the InsP₃R-mediated release in this assay was verified by using heparin, a competitive inhibitor of the InsP₃R. For this purpose, a nonratiometric dye, Fluo-4, was used. As seen in Fig. 2 C, InsP₃-mediated Ca²⁺ release is completely inhibited in the presence of 50 μg/ml of heparin. Microsomes isolated from the ug3/+ adult heads show a dominant phenotype of lowered K_m for InsP₃ (K_m = 274 ± 34 nM InsP₃) without affecting the R_max value (Fig. 2 D). Microsomes from adult heads of wc703/+ heterozygotes show an increased R_max (923 ± 29 nmol Ca²⁺/mg protein; Fig. 2 E) but an unchanged K_m compared to wild-type (Fig. 2 E). Microsomes obtained from ka901/+ adult heads released over twice as much Ca²⁺ (R_max = 1678 ± 124 nmol Ca²⁺/mg protein), with a significantly lower K_m (K_m = 308 ± 63 nM), compared to wild-type microsomes (Fig. 2 F).

From these experiments we concluded that each tested mutant allele has distinct Ca²⁺-release properties in the Ca²⁺ flux assay. The observed changes in R_max could be due to different levels of the DmInsP₃R protein among mutants or different amounts of Ca²⁺ trapped in vesicles from different genotypes. By Western blotting the levels of DmInsP₃R protein were found to be equivalent in the wild-type and all mutant microsomes (Fig. 3 A). The amount of Ca²⁺ trapped in vesicles of all the genotypes was also found to be equivalent as determined by measuring Ca²⁺ release on addition of 0.1% Triton X-100 (Fig. 3 B). However, as expected from Fig. 2, E and F, the amount of InsP₃-releasable Ca²⁺ was found to be higher in vesicles from itp^wc703 and itpr^ka901 heterozygotes (Fig. 3 B). Therefore, the observed changes in R_max and K_m values most likely reflect changes in DmInsP₃R functional properties induced by each mutation. To test this idea conclusively an alternate approach was adopted.

FIGURE 3.

Levels of the DmInsP₃R protein and Ca²⁺ in wild-type and mutant vesicles. (A) Western blot of wild-type and mutant adult head microsomes with anti-DmInsP₃R polyclonal antibody. Lane 1, CS adult head lysate; lanes 2–5, microsomal vesicles from adult heads of the genotypes as indicated. Five micrograms of total protein was loaded in each lane. The blot was stripped and reprobed with antiserum to α-spectrin (278 kDa) as a loading control. When levels of the DmInsP₃R were compared with the levels of the loading control by densitometry analysis, no significant changes were observed between the various genotypes. The representative blot of three independent experiments is shown. (B) Fraction of maximal Ca²⁺ released from microsomes in response to 2 μM InsP₃ (black bars) as a fraction of total vesicular Ca²⁺ (estimated by lysing the microsomes with 0.1% Triton X-100, open bars). The fraction of InsP₃-releasable Ca²⁺ was significantly higher in itprwc703/+ and itprka901/+ (P < 0.001 for both) microsomes than in wild-type (CS).

Single-channel properties of the DmInsP₃R embryonic and adult head isoforms

Single-channel recordings offer by far the most detailed level of characterization of functional properties of channels. To obtain single-channel recordings of the DmInsP₃R, we adopted a previously described approach developed for the structure-functional analysis of the mammalian InsP₃R1 (Nosyreva et al., 2002; Tang et al., 2003a, b; Tu et al., 2002, 2003). The DmInsP₃R (embryonic form) and AH-DmInsP₃R (adult head form) were expressed in sf9 cells by baculovirus infection and the expression was confirmed by Western blotting (Fig. 4) using affinity-purified anti-DmInsP₃R rabbit polyclonal antibodies. The apparent molecular weight of DmInsP₃R and AH-DmInsP₃R expressed in sf9 cells (280 kDa) was identical to endogenous DmInsP₃R detected by the same antibodies in Drosophila larval and adult head lysates (Fig. 4). No signal was observed in samples prepared from noninfected sf9 cells (Fig. 4).

FIGURE 4.

Expression of the DmInsP₃R in sf9 cells. Western blot of Drosophila larval (LL) and adult head (AHL) lysates and microsomes isolated from noninfected sf9 cells (sf9) and sf9 cells infected with DmInsP₃R, AH-DmInsP₃R, ug3, wc703, and ka901 baculoviruses as indicated. The DmInsP₃R was detected using an affinity-purified anti-DmInsP₃R polyclonal antibody. For the larval lysate 30 μg of total protein was loaded whereas 20 μg was loaded in the lane containing the adult head lysate. For each microsomal preparation, 100 ng of total microsomal protein was loaded on the gel except for ug3, where 300 ng total microsomal protein was loaded. The representative blot of three independent experiments is shown.

Recombinant DmInsP₃R and AH-DmInsP₃R expressed in sf9 cells were incorporated into the bilayer by microsomal vesicle fusion as described previously for the wild-type and mutant mammalian InsP₃R1 (Nosyreva et al., 2002; Tang et al., 2003a, b; Tu et al., 2002, 2003). InsP₃-gated currents supported by the DmInsP₃R and the AH-DmInsP₃R channels were recorded using 50 mM Ba²⁺ as the charge carrier at 0-mV holding potential (Fig. 5, A and B). InsP₃-gated channels in bilayers were observed in 45 of 50 experiments with DmInsP₃R microsomes and in 40 of 50 experiments with AH-DmInsP₃R microsomes, but were never (n = 10) observed in experiments with microsomes from noninfected sf9 cells. Single-channel analysis revealed that the mean current amplitude at 0-mV holding potential is equal to 1.80 ± 0.04 pA for the DmInsP₃R (Fig. 5 A) and 1.86 ± 0.01 pA for the AH-DmInsP₃R (Fig. 5 B); the mean open dwell time is equal to 4.3 ± 0.3 ms for the DmInsP₃R (Fig. 5 A) and 3.6 ± 0.1 ms for the AH-DmInsP₃R (Fig. 5 B) (Table 1).

FIGURE 5.

Functional properties of the wild-type isoforms of the DmInsP₃R. Representative channel activity of the DmInsP₃R (A) and the AH-DmInsP₃R (B) recorded in the presence of 0.2 μM Ca²⁺, 0.5 mM Na₂ATP, and 2 μM InsP₃ (standard recording conditions). Current records are shown at compressed and expanded timescales as indicated. Unitary current amplitude histograms and open dwell time distributions from the same experiments are shown below the current records. Unitary currents were fitted with a Gaussian function that was centered at 1.83 pA for the DmInsP₃R and at 1.86 pA for the AH-DmInsP₃R. Open time distributions were fit with a single exponential function (curve) that yielded a τ₀ of 4.0 ms for the DmInsP₃R and 3.98 ms for the AH-DmInsP₃R. Similar analysis was performed for at least five independent experiments with the DmInsP₃R and the AH-DmInsP₃R microsomes.

TABLE 1.

Comparison of main functional properties of DmInsP₃R and mammalian InsP₃R1 (Tang et al., 2003b; Tu et al., 2002)

		Conductance properties		Gating properties		InsP3-dependence		Bell-shaped Ca2+ dependence
InsP3R isoform	Splice variant	i, pA	γ, pS	τo, ms	Po max %	kd, nM	nHill	nHill	Kact, μM	Kinh, μM	Peak (pCa)
InsP3R1	SII(+)	1.9 pA	81 pS	4.7 ms	19%	190 nM	1.47	1.3	0.22 μM	0.20 μM	6.6
DmInsP3R	emb	1.8 pA	70 pS	4.3 ms	15%	100 nM	1.13	0.91	0.28 μM	0.53 μM	6.4
DmInsP3R	AH	1.9 pA	89 pS	3.6 ms	17%	ND		Similar to embryonic
wc703	emb	1.75 pA	70 pS	4.5 ms	21%	ND		1.96	0.17 μM	0.18 μM	6.8
ug3	emb	1.72 pA	71 pS	6.4 ms	17%	ND		Similar to embryonic

To generate a current-voltage relationship, the transmembrane potential was varied between +10 mV and −30 mV during current recordings. A linear fit to the obtained results (Fig. 6) yielded single-channel conductances of 70.0 ± 0.1 pS for the DmInsP₃R and 89.0 ± 0.1 pS for the AH-DmInsP₃R. From previous molecular analysis of these isoforms it is known that Drosophila larvae express both splice variants (predominantly the embryonic isoform) whereas only the AH form is present in adult heads. The higher conductance of the AH-DmInsP₃R is thus consistent with the differences in R_max values observed between larval and adult head microsomes in the Ca²⁺-release assay (Fig. 2 B).

FIGURE 6.

Current-voltage relationship of the DmInsP₃R isoforms. Current-voltage relationship of the DmInsP₃R (•) and the AH-DmInsP₃R (○) channels measured at transmembrane voltages between +10 mV and −30 mV. Average data from at least four independent experiments with DmInsP₃R and AH-DmInsP₃R are shown as the mean ± SE at each voltage. The linear fit to the data (lines) yielded single-channel conductance of 70.0 pS for the DmInsP₃R and 89 pS for the AH-DmInsP₃R.

To further characterize functional properties of the DmInsP₃R, its open probability was measured in a range of Ca²⁺ and InsP₃ concentrations. The analysis of the DmInsP₃R InsP₃-sensitivity (Fig. 7) yielded an apparent affinity for InsP₃ equal to 100 nM. We found that similar to mammalian InsP₃R1 (Table 1), DmInsP₃R displays a bell-shaped Ca²⁺ dependence with the peak at pCa 6.4 (Fig. 8). On an average, the single-channel open probability of the DmInsP₃R channels at pCa 6.4 was equal to 0.15 ± 0.015. A fit to the averaged Ca²⁺-dependence data (Fig. 8 B) using a modified bell-shaped equation from Bezprozvanny et al. (1991) yielded the affinities of activating and inhibitory sites of the DmInsP₃R equal to 0.28 μM Ca²⁺ and 0.53 μM Ca²⁺, respectively. Although the affinity of the activating site is similar to mammalian InsP₃R1 (Nosyreva et al., 2002; Tu et al., 2002, 2003), the affinity for the inhibitory site is lower in the DmInsP₃R (Table 1), indicating that higher concentrations of cytoplasmic Ca²⁺ are required for inhibition of the InsP₃R in Drosophila. Similar results were obtained for the AH-DmInsP₃R (data not shown).

FIGURE 7.

InsP₃ dependence of the DmInsP₃R. The DmInsP₃R Po measured in three independent experiments were averaged together at each InsP₃ concentration as described in Materials and Methods and shown as mean ± SE (•). The averaged data were fitted by the equation modified from Lupu et al. (1998) as explained in Materials and Methods. The parameters of optimal fit (smooth curve) yielded k_InsP3 = 100 nM, n_Hill = 1.13, and P_max = 0.078.

FIGURE 8.

Calcium dependence of the DmInsP₃R. (A) Current recordings of the DmInsP₃R from the same experiment at different Ca²⁺ concentration as indicated. (B) Bell-shaped Ca²⁺ dependence of DmInsP₃R. The P_o measured in five independent experiments with DmInsP₃R were averaged together at each Ca²⁺ concentration as described in Materials and Methods and shown as mean ± SE (•). The averaged data were fitted by the bell-shaped equation modified from Bezprozvanny et al. (1991) as explained in Materials and Methods. The optimal fit (smooth curve) was obtained with n_Hill = 0.91, P_m =0.13, k_act = 0.28 μM Ca²⁺, and K_inh = 0.53 μM Ca²⁺.

Functional properties of the ug3 mutant

To characterize single-channel properties of the ug3 mutant (embryonic form) it was expressed in sf9 cells by baculovirus infection. The expression levels of ug3 mutant channels in sf9 cells were lower than expression levels of the wild-type channels (in Fig. 4, a three-times excess of ug3 microsomes was loaded). When microsomes isolated from ug3-infected sf9 cells were fused to planar lipid bilayers, InsP₃-gated channels were observed less frequently than in experiments with wild-type microsomes (in 20 of 34 experiments; Fig. 9). The mean current amplitude of ug3 channels at 0-mV holding potential is 1.7 ± 0.32 pA (Fig. 9), with a mean open dwell time of 6.4 ± 0.8 ms (Fig. 9), and single-channel conductance of 71.0 ± 0.2 pS (n = 3, data not shown). Thus, ug3 channels display conductance properties similar to the wild-type DmInsP₃R (embryonic form), but have 47% longer mean open time.

FIGURE 9.

Functional properties of ug3 channels. Representative channel activity of ug3 channels recorded in the standard recording conditions. Current records are shown at compressed and expanded timescales as indicated. Unitary current amplitude histograms and open dwell time distributions from the same experiments are shown below the current records. Unitary current distribution was fit with a Gaussian function that was centered at 1.71 pA. Open time distribution was fit with a single exponential function (curve) that yielded a τ₀ of 5.6 ms. Similar analysis was performed for at least three independent experiments with ug3 microsomes.

The amino-terminal region of the InsP₃R forms a ligand-binding domain (Fig. 1). Biochemical and structural studies of mammalian InsP₃R1 indicate that the InsP₃R ligand-binding domain can be subdivided into a core InsP₃ binding domain and an inhibitory domain (Bosanac et al., 2002; Yoshikawa et al., 1999, 1996; see also Fig. 1, this article). Sequences corresponding to the core and inhibitory domains are highly conserved between DmInsP₃R and InsP₃R1, and the ug3 mutation (S224F) is interestingly located at the interface between the predicted inhibitory and core domains of the DmInsP₃R (Fig. 1). From the position of the ug3 mutation in the ligand-binding domain it was expected that ug3 may have an effect on the binding affinity of the DmInsP₃R for InsP₃. Binding of ³H-InsP₃ to microsomes isolated from the DmInsP₃R- and ug3-infected sf9 cells yielded affinity values of 152 nM InsP₃ for the wild-type DmInsP₃R (embryonic form) and 214 nM InsP₃ for the ug3 mutant (Fig. 10). The observed difference in affinity was found to be insignificant upon repetition of the experiment (data not shown). Notably, the affinity for InsP₃ is either the same or slightly lower in ug3 when compared with wild-type DmInsP₃R in ³H-InsP₃ binding experiments (Fig. 10). This conclusion contrasts with higher sensitivity of ug3/+ microsomes to InsP₃ in the Ca²⁺ flux assay as compared to the wild-type microsomes (Fig. 2 D).

FIGURE 10.

³H-InsP₃ binding studies with the DmInsP₃R and ug3 microsomes. Scatchard analysis of ³H-InsP₃ binding to the DmInsP₃R (•) and ug3 (○) microsomes. Linear fit to the data (lines) yielded k_d = 152 nM InsP₃ and B_max = 180 pmol/mg for the DmInsP₃R and k_d = 214 nM InsP₃ and B_max = 124 pmol/mg for the ug3 mutant. The experiment was repeated in duplicate with similar results.

To understand this discrepancy, the sensitivity of ug3 channels to activation by low InsP₃ concentrations was investigated in bilayer experiments. We discovered that ug3 channels displayed maximum open probability (P_o) at InsP₃ concentrations as low as 50 nM (Fig. 11, A and B). Much higher concentrations of InsP₃ are required to fully activate the wild-type DmInsP₃R (Fig. 11, C and D, and Fig. 7). The ug3 channels are expressed in sf9 cells at a much lower level than DmInsP₃R. This was evident from the B_max values (180 pmol/mg for DmInsP₃R versus 124 pmol/mg for ug3 mutant) and the significantly lower abundance of the InsP₃R positive band in ug3 microsomes, as seen in a quantitative Western blot (Fig. 4, data not shown). Because of reduced expression levels of ug3 mutant and resulting difficulties with channel recordings, we have not been able to obtain a complete InsP₃-dependence curve for this mutant in planar lipid bilayers. However, it is clear from Fig. 11 that, unlike the DmInsP₃R, the open probability for ug3 is similar at 50 nM and 2 μM InsP₃, indicating a shift in apparent sensitivity to activation by InsP₃.

FIGURE 11.

InsP₃ sensitivity of ug3 channels. (A) Activity of ug3 channels in bilayers in the presence of 50 nM InsP₃ and 2 μM InsP₃ is shown. Each current trace corresponds to 10 s of current recording from the same experiment. (B) The average open probability (P_o) of ug3 channels is calculated for a 5-s window of time and plotted for the duration of an experiment. The times of 50 nM and 2 μM InsP₃ additions to the bilayer are shown above the P_o plot. Data from the same experiment are shown in A and B. Similar results were obtained in three independent experiments. (C–D) Same data as in A and B for the wild-type DmInsP₃R channels. Similar results were obtained in three independent experiments.

Functional properties of the wc703 mutant

To characterize single-channel properties of the wc703 mutant, its embryonic form was expressed in sf9 cells by baculovirus infection. The expression levels of the wc703 mutant in sf9 cells were comparable to expression levels of the wild-type channels (Fig. 4). When microsomes isolated from the wc703-infected sf9 cells were fused to planar lipid bilayers, InsP₃-gated channels were observed in 25 of 40 experiments (Fig. 12 A). Thus, similar to the ug3 mutant, homomeric channels formed by wc703 mutant subunits are functional. The mean current amplitude of wc703 channels at 0-mV holding potential is 1.74 ± 0.08 pA (Fig. 12 A), with a mean open dwell time of 4.5 ± 0.4 ms (Fig. 12 A), and single-channel conductance of 69.0 ± 0.1 pS (n = 3, data not shown). Thus, gating and conductance properties of wc703 channels are not significantly different from the wild-type DmInsP₃R (embryonic form). To determine the effect of wc703 mutation on Ca²⁺ sensitivity of the DmInsP₃R, we measured open probability of wc703 channels in bilayers at different Ca²⁺ concentrations. We discovered that, in contrast to the wild-type DmInsP₃R, the maximal open probability of wc703 was achieved at pCa 6.8 (Fig. 12 B). On an average, the single-channel open probability of wc703 channels at pCa 6.8 was equal to 0.21 ± 0.013. Fit to the averaged Ca²⁺-dependence data (Fig. 12 C, n = 3) using a modified bell-shaped equation from Bezprozvanny et al. (1991) yielded the affinities of activating and inhibitory sites of DmInsP₃R equal to 0.17 μM Ca²⁺ and 0.18 μM Ca²⁺, respectively. The observed effects on the DmInsP₃R Ca²⁺ sensitivity are consistent with the position of the wc703 mutation (G2117E) in a Ca²⁺-sensor region (Tu et al., 2003) of the DmInsP₃R (Fig. 1). The results suggest that the increased R_max observed in the Ca²⁺-release assay with microsomes from wc703/+ flies (Fig. 2 E) can be explained by the approximately twofold increase in an apparent affinity of DmInsP₃R Ca²⁺ sensor and 30% increase in the DmInsP₃R maximal single-channel open probability.

FIGURE 12.

Functional properties of wc703 channels. (A) Representative channel activity of wc703 channels recorded in the standard recording conditions. Current records are shown at compressed and expanded timescales as indicated. Unitary current amplitude histograms and open dwell time distributions from the same experiments are shown below the current records. Unitary current distribution was fit with a Gaussian function that was centered at 1.82 pA. Open time distribution was fit with a single exponential function (curve) that yielded a τ₀ of 3.8 ms. Similar analysis was performed for at least four independent experiments with wc703 microsomes. (B) The current recordings of wc703 from the same experiment at different Ca²⁺ concentration as indicated. (C) Bell-shaped Ca²⁺ dependence of wc703. The P_o measured in three independent experiments with wc703 were averaged together at each Ca²⁺ concentration as described in Materials and Methods and shown as mean ± SE (•). The averaged data were fitted by the bell-shaped equation modified from Bezprozvanny et al. (1991) as explained in Materials and Methods. The optimal fit (thick smooth curve) was obtained with n_Hill = 1.9, P_m = 0.2, k_act = 0.17 μM Ca²⁺, and K_inh = 0.18 μM Ca²⁺. The fit to the Ca²⁺ dependence of the wild-type DmInsP₃R (embryonic form) fromFig. 8 is shown for the reference (thin line).

Absence of channel function in the ka901 mutant

The DmInsP₃R ka901 mutant (embryonic form) was expressed in sf9 cells by baculovirus infection. As determined by Western blotting, expression levels of ka901 in sf9 cells were comparable to expression levels of wild-type channels (Fig. 4). However, InsP₃-gated channel activity was not observed in bilayer experiments with microsomes from ka901-infected sf9 cells (n = 15), indicating that ka901 homomeric channels are nonfunctional. This idea is supported by the observation that ka901 mutants are equivalent to null alleles of the InsP₃R in Drosophila (Joshi et al., 2004). Interestingly, however, we did measure increased Ca²⁺ fluxes in microsomes from ka901/+ flies (Fig. 2 F). Taken together, these observations suggest that ka901 may form functional channels when assembled with the wild-type DmInsP₃R subunits, but not in the homomeric state.

DISCUSSION

Major functional properties of InsP₃Rs are conserved in evolution

The major functional properties of embryonic and adult head splice variants of the Drosophila InsP₃R obtained from planar lipid bilayer experiments (Figs. 5–8) have been described here. To the best of our knowledge, this is the first description of single-channel properties of Drosophila melanogaster InsP₃R. On comparison with the properties of rat InsP₃R1, determined in identical experimental conditions in our previous studies (Tang et al., 2003b; Tu et al., 2002), we conclude that the single-channel properties of the DmInsP₃R are remarkably similar to the single-channel properties of mammalian InsP₃R1, apart from a broader Ca²⁺-dependence curve (Table 1). This conclusion agrees with the conservation of InsP₃R domain structure (Fig. 1), and the high degree of sequence identity (57%) between DmInsP₃R and rat InsP₃R1, and with previous analysis of InsP₃-dependent Ca²⁺ fluxes in Drosophila-derived S2 cells (Swatton et al., 2001). Overall, our results validate the use of Drosophila as a genetic model for InsP₃R structure-function analysis, as results obtained with the DmInsP₃R are likely to have an implication for understanding the mechanisms of mammalian InsP₃R function. Interestingly, an SIII splice variant of the human type I InsP₃R, spliced in a region homologous to the Drosophila adult head isoform, is enriched in brain tissues (Nucifora et al., 1995). This SIII splice variant contains nine additional amino acids, which give rise to a putative protein kinase C phosphorylation site. The nine-amino-acid insert in the fly adult head isoform also contains a putative casein kinase II phosphorylation site. It would thus be interesting to compare the single-channel properties of the SIII isoform of human InsP₃RI with the adult head isoform of DmInsP₃R described here.

DmInsP₃R ug3—the InsP₃ efficacy mutant

The ug3 mutation (S224F) is located within the ligand-binding domain of DmInsP₃R (Fig. 1). Thus the observation that ug3 affects the InsP₃ sensitivity of DmInsP₃R channel opening (Fig. 11, A–D), but does not increase InsP₃ binding affinity (Fig. 10), is intriguing. This result suggests that the InsP₃R inhibitory domain plays an important role in coupling InsP₃ binding to the core domain with opening of the InsP₃R channel gate, in agreement with recent biochemical (Boehning and Joseph, 2000) and functional (Uchida et al., 2003) data. Increased open dwell time in ug3 mutants (by 47%) is consistent with the conclusion that ug3 is not a binding mutant, but is, in fact, a gating mutant. These data lead us to conclude that the increased apparent affinity to InsP₃ (change in K_m) observed in Ca²⁺ flux assay with ug3/+ microsomes (Fig. 2 D) results not from an increased affinity of the DmInsP₃R for InsP₃ but from an increased efficacy in coupling InsP₃ binding with DmInsP₃R channel opening. Thus, ug3 is a gain-of-function mutant that acts by increasing sensitivity of the DmInsP₃R to activation by InsP₃ and also by increasing dwell open time of the channels. Since the affinity of InsP₃ binding is not increased by the ug3 mutation (Fig. 10), we classified the ug3 mutant as the efficacy mutant, that is, the mutant with increased efficiency of coupling between InsP₃ binding and channel opening. This conclusion is consistent with the biochemical association observed between InsP₃R1 amino-terminal and carboxy-terminal ends (Boehning and Joseph, 2000) and a recent suggestion that the InsP₃R1 inhibitory amino-terminal domain participates in the channel-gating process (Uchida et al., 2003).

The observation that ug3 homomeric channels are functional is consistent with the weak allelic strength of ug3 mutants. Heterozygotes of the genotype ug3/90B0 die as late second- or early third-instar larvae (Deshpande et al., 2000; Joshi et al., 2004) as compared to ka901/90B0 and 90B0/90B0 (itpr null) mutants, both of which die as early second-instar larvae (Venkatesh and Hasan, 1997).

DmInsP₃R wc703—the Ca²⁺ sensor mutant

Microsomes from wc703/+ flies display increased Ca²⁺ release (1.4-fold increase in R_max) in Ca²⁺ flux experiments (Fig. 2 E). Although single channels formed by wc703 display wild-type conductance and gating properties (Fig. 12 A and data not shown), they also exhibit increased sensitivity to Ca²⁺ and a higher maximal open probability (Fig. 12, B and C, and Table 1). The peak of wc703 channel Ca²⁺ dependence is at pCa 6.8 (compare to pCa 6.4 for wild-type channels) and their maximal open probability is 21% (as compared to 15% for wild-type). A fit of the bell-shaped Ca²⁺ dependence of wc703 channels yielded affinities of activating and inhibitory Ca²⁺ binding sites equal to 0.17 μM Ca²⁺ and 0.18 μM Ca²⁺, respectively (as compared to 0.28 μM Ca²⁺ and 0.53 μM Ca²⁺ for the wild-type DmInsP₃R). These results indicate that wc703 is also a gain-of-function mutant that acts by increasing sensitivity of DmInsP₃R to activation by Ca²⁺. These effects are consistent with the position of wc703 mutation (G2117E) within a putative Ca²⁺ sensor region (Tu et al., 2003) of the DmInsP₃R (Fig. 1). The wc703 homomeric channels are functional, consistent with the weak allelic strength of wc703 mutants, which is similar to that of the ug3 allele (Deshpande et al., 2000; Joshi et al., 2004).

DmInsP₃R ka901—the Ca²⁺ channel mutant

The ka901 mutation lies in the pore-forming region of the InsP₃R and RyanR families among which a putative pore-forming motif GXRXGGGI/VGD (the G2630 residue of the DmInsP₃R mutated in ka901 is underlined) is highly conserved. Mutational studies of the homologous G4826 residue in RyanR2 has shown that the G4826C mutant does not form functional homomeric channels in bilayer experiments (Chen et al., 2002), similar to what we observe with ka901 homomeric channels. The absence of functional channels from ka901 homomers in bilayer experiments agrees with the genetic finding that the allelic strength of ka901 is equivalent to that of the null allele 90B0 (Joshi et al., 2004). To understand why microsomal vesicles from ka901/+ heterozygotes release greater amounts of Ca²⁺, co-infection with wild-type DmInsP₃R and ka901 studies can be done similar to what has been done with RyanR2-G4826C to obtain hybrid channels (Chen et al., 2002; Zhao et al., 1999). These experiments are currently in progress.

CONCLUSION

Genetic and biophysical approaches to the InsP₃R structure-function

In conclusion, single-channel properties of embryonic and adult head splice variants of the Drosophila InsP₃R (Sinha and Hasan, 1999) and three previously isolated and sequenced Drosophila InsP₃R point mutants (Deshpande et al., 2000; Joshi et al., 2004) have been characterized in our study. Our results indicate that the main functional properties of InsP₃R are conserved in evolution from DmInsP₃R to InsP₃R1 (Table 1) and furnish further evidence for differences in function between splice variant isoforms. Changes in the DmInsP₃R properties induced by random mutagenesis have helped identify residues that may not necessarily be identified easily by a targeted mutagenesis approach. Although one of these residues (ka901) in homomeric state supports the data obtained earlier from a similar mutation in the RyanR2 (Chen et al., 2002), the others provide fresh insights into the mechanism of Ca²⁺ dependence and the relationship between ligand binding and gating. Furthermore, the single-channel properties of the wild-type and mutant DmInsP₃Rs have been correlated with characteristics of InsP₃-dependent Ca²⁺ fluxes supported by these channels in experiments with microsomes from wild-type and mutant Drosophila. The approaches used in this work have thus allowed for the extension of genetic studies of the InsP₃R in a model organism to biophysical studies on the channel.