Olfactory receptor neurons in Atlantic salmon (Salmo salar) appear to use a phosphoinositide-directed phospholipase C (PLC) in odorant signal transduction. The consequences of odor-activated PLC depend on its product, inositol 1,4,5-trisphosphate (IP3). Therefore, a plasma membrane rich (PMR) fraction, previously characterized from salmon olfactory rosettes, was used to study binding sites for IP3 and its phosphorylation product, inositol 1,3,4,5-tetrakisphosphate (IP4). Binding sites for IP3 were present at the lower limit for detection in the PMR fraction but were abundant in a microsomal fraction. Binding sites for IP4 were abundant in the PMR fraction and thus colocalized in the same subcellular fraction with odorant receptors for amino acids and bile acids. Binding of IP4 was saturable and high affinity (  = 83 nM). The rank order for potency of inhibition of IP4 by other inositol polyphosphates (InsPx) followed the phosphorylation number with InsP6 > InsP5 > other InsP4 isomers > InsP3 isomers > InsP2 isomers, with the latter showing no activity. The consequences of PLC activity in this system may be dictated in part by a putative receptor for IP4.

1. Introduction

Adenylyl cyclase and cAMP appear to dominate odor signal transduction in mammals (for reviews, see [13]). Phosphoinositides may play a divergent role in olfaction, mediating inhibitory signaling through phosphoinositide-3-kinase [4] or excitatory signaling through phospholipase C [1, 5]. For fish, components of a phospholipase C-based olfactory signal transduction system have been characterized in catfish [613] and are seen in carp [14, 15], zebrafish [16], and Atlantic salmon [17, 18].

As potent olfactory stimuli for Atlantic salmon, amino acids and bile acids interact with distinct subclasses of olfactory receptors to begin the process of olfactory reception [18, 19]. The amino acid and bile acid receptors appear to be coupled through G proteins to the activation of phospholipase C (PLC) and the breakdown of phosphatidylinositol 4,5-bisphosphate (PIP2) to generate diacylglycerol (DAG) and inositol 1,4,5-trisphosphate (IP3) [17, 18]. Early biochemical data characterizing these as G protein-coupled receptors is now supported by molecular studies characterizing olfactory receptor gene sequences from Atlantic salmon [2023]. Underscoring the importance of these receptors in salmon physiology, odorant receptor expression has been shown to change during the parr-smolt transformation, a period characterized by increased olfactory sensitivity and olfactory-based learning [24].

The significance of olfactory PLC activity resides in part with the location and characteristics of receptors for IP3. In most cells, IP3 receptors mediate the release of Ca2+ from internal stores in the endoplasmic reticulum (for review, see [25]). However, in association with PLC-based olfactory signal transduction, IP3 receptors have been found in olfactory cilia of catfish [6], carp [14], and lobster [26, 27]. From this position, IP3 may gate Ca2+ influx through the plasma membrane rather than the release from intracellular stores. Another important part of IP3 signaling in other systems has been its metabolism, including phosphorylation by a 3-kinase to generate the biologically active inositol 1,3,4,5-tetrakisphosphate (IP4) [2830]. While IP4 continues to be studied in mammalian systems for roles as diverse as regulating nuclear calcium signaling [31], tyrosine kinase [32], and mitochondrial permeability and apoptosis [33, 34], Fadool and Ache [26] showed that olfactory receptor neurons of lobster express an IP4 receptor acting as a functional channel in the plasma membrane. In lobster, plasma membrane IP3 and IP4 receptors may interact reciprocally to regulate Ca2+ entry in olfactory neurons.

The goal of the present study was to characterize further the PLC-based olfactory signal transduction system of Atlantic salmon, beginning with the hypothesis that IP3 binding sites would colocalize with odor receptor binding sites in a plasma membrane rich fraction (PMR) that we characterized previously [1720, 35]. Finding that binding of IP3 was marginal in this fraction, we proceeded to detect and characterize PMR binding sites for IP4 which may play a critical role in salmon olfactory transduction. Binding sites for IP3 were subsequently detected in the endoplasmic reticulum-rich microsomal fraction.

2. Materials and Methods

2.1. Isolation of the Plasma Membrane Rich (PMR) and Microsomal Fractions

Atlantic salmon (Salmo salar) were raised under conditions of simulated natural photoperiod and temperature in the aquaculture facility of University of Rhode Island. Using a modification of a method devised originally for rainbow trout by Cagan and Zeiger [36], a plasma membrane rich (PMR) fraction was obtained from the olfactory rosettes as described previously [19]. Rosettes were pooled from ten salmon for each analysis. The microsomal fraction was isolated from the olfactory rosettes using the method of Kalinoski et al. [6]. For comparative purposes, PMR fractions and microsomal fractions were also prepared from salmon brain and rat brain. Concentrations of proteins were determined by the method of Bradford (Bio-Rad Laboratories, Hercules, CA) with bovine serum albumin as a standard.

2.2. IP3 Binding

Binding of [3H]IP3 ([inositol-1-3H]; 21.0 Ci/mmol; New England Nuclear, Boston, MA) was measured using conditions described by Kalinoski et al. [6] except that microsomal fractions (100 μg protein per assay) or PMR fractions (100–300 μg protein per assay) were from salmon olfactory rosettes or from salmon or rat brain. Digitonin (50 μg/mL) was added to permeabilize any membrane vesicles and insure that all binding sites are accessible [6]. The incubation buffer consisted of 110 mM KCl, 1 mM EGTA/0.2 mM CaCl2 (free Ca2+ concentration = 20 nM), and 10 mM HEPES, pH 7.4. Incubations were carried out for 30 min at 4°C. Separation of bound and free [3H]IP3 was achieved by rapidly filtering through Whatman GF/C filters and washing 3 times with assay buffer. Filters were extracted in scintillation cocktail for 4 hr, and the amount of associated radioactivity was determined by scintillation spectrometry. The amount of binding was determined in the absence (total binding) and presence (nonspecific binding) of excess (120 μM) unlabeled InsP3. Two concentrations of [3H]IP3 (7 and 14 nM) were tested. The calculated difference between total and nonspecific binding was operationally defined as a specific binding.

2.3. IP4 Binding

The binding assay for [3H]IP4 ([Inositol-1-3H]; 21.0 Ci/mmol; New England Nuclear, Boston, MA) was performed under conditions identical to those described by Challiss et al. [37]. The assay buffer consisted of 25 mM CH3COONa, 25 mM KH2PO4, 5 mM NaHCO3, 1 mM EDTA, pH 5.0, and the indicated concentrations of [3H]IP4. Nonspecific binding was defined by the inclusion of 120 μM unlabeled IP4. To characterize the binding specificity, competition assays were conducted with a minimum of three concentrations of other inositol polyphosphates (InsPx): InsP6, Ins(1,3,4,5,6)P5, Ins(3,4,5,6)P4, Ins(1,4,5,6)P4, Ins(1,3,4)P3, Ins(1,4,5)P3, Ins(1,4)P2, and Ins(4,5)P2 (all generously provided by Dr. Ching-Shih Chen, School of Pharmacy, University of Rhode Island). Reactions were initiated by the addition of PMR fraction (100 μg protein), and samples were maintained at 4°C for 30 min with gentle rocking. Separation of bound and free [3H]IP4 was achieved by rapidly filtering through Whatman GF/C filters and washing 3 times with assay buffer. Filters were extracted in scintillation cocktail for 4 hr, and radioactivities were determined.

Binding assays for both [3H]IP3 and [3H]IP4 were based on conditions optimized by others ([6, 36], resp.). To rule out any effect of the different incubation conditions (most notably pH) on conclusions regarding binding of [3H]IP3 or [3H]IP4, each was tested at the conditions that had been optimized for the other. As expected, binding was negligible when measured at nonoptimal conditions.

3. Results

3.1. IP3 Binding

At a radioligand concentration of 7 nM, no specific binding of IP3 was detectable with the olfactory PMR fraction. At 14 nM radioligand, IP3 binding to the olfactory PMR fraction was at the lower limit of detection in the assay (see data labeled IP3-PMR in Figure 1). Nonspecific binding accounted for almost 90% of the small amount of total binding of [3H]IP3 to the PMR fraction. Similar results were obtained with a salmon brain PMR fraction, analyzed as a negative control. The specific binding of [3H]IP3 corresponded to a maximum of 16 fmol bound per mg olfactory PMR protein and 10 fmol per mg salmon brain fraction.

In contrast, specific binding sites for [3H]IP3 were readily detected in a microsomal (MS) preparation from salmon olfactory rosettes (see data labeled IP3-MS in Figure 1). In this preparation, specific binding accounted for at least 75% of the total binding of [3H]IP3 and corresponded to 1.2 pmol IP3 bound per mg MS protein, a level nearly 100 times higher than the PMR fraction. This compares favorably to the level of IP3 binding measured in a rat brain microsomal fraction that was analyzed as a positive control.

3.2. IP4 Binding

While IP3 binding to the salmon olfactory PMR fraction was at the lower limit for detection in our assay, binding sites for IP4 were readily detected and were present at high density (see data labeled IP4-PMR in Figure 1). At comparable ligand concentration (14 nM), the olfactory PMR fraction supported binding of 364 fmol IP4 per mg protein (contrasted with 16 fmol IP3 per mg protein). Nonspecific binding represented less than 20% of total binding. In a single trial with the microsomal preparation from salmon olfactory rosettes, specific binding of [3H]IP4 was at the lower limit of detection (not shown). Thus, IP4 sites were readily detected in the PMR but not the microsomal fraction, a result opposite of that for IP3 binding.

Experiments performed with increasing concentrations of [3H]IP4 demonstrated that specific binding was saturable (Figure 2). Scatchard analysis of the binding data (Figure 2, inset) yielded 83 nM for the and 3811 fmol/mg protein for the for IP4 binding to the olfactory PMR fraction.

To further characterize the specificity of IP4 binding to the olfactory PMR fraction, competition experiments were performed using 14 nM [3H]IP4 and various other inositol polyphosphates (InsPx) differing in degree and position of phosphorylation (Figure 3). If an analog competes with IP4, then binding will decrease as the concentration of the analog increased (Figure 3).

InsP5 and InsP6 showed reasonably potent inhibition of [3H]IP4 binding. Other IP4 analogs (Ins(3,4,5,6)P4 and Ins(1,4,5,6)P4) were intermediate in potency as inhibitors, while the IP3 analogs (Ins(1,3,4)P3 and Ins(1,4,5)P3) showed little or no activity. Similarly, Ins(1,4)P2 and Ins(4,5)P2, the dephosphorylation products formed from the inactivation of Ins(1,4,5)P3, had no inhibitory effect on [3H]IP4 binding when incubated at 10 μM (data not included in Figure 3). From these competition assays, the effective concentration of analog giving 50% inhibition of [3H]IP4 binding was estimated (Table 1). From this analysis, the rank order for potency of inhibition to [3H]IP4 binding was InsP6 > Ins(1,3,4,5,6)P5 > Ins(3,4,5,6)P4 > Ins(1,4,5,6)P4 > Ins(1,3,4)P3 = Ins(1,4,5)P3.

4. Discussion

Previous characterization of the PMR fraction showed high levels of the plasma membrane marker Na, K-ATPase and binding sites for amino acid [19, 20] and bile acid [18] odors. This fraction had minimal contamination with endoplasmic reticulum as suggested by the absence of thapsigargin-sensitive Ca2+-ATPase [35]. The low level of observed IP3 binding can also be considered as evidence of the lack of ER especially when compared to the microsomal fraction which is traditionally used as a source of endoplasmic reticulum and IP3 receptors [25]. A comparison of IP3 binding to the two subcellular fractions is consistent with the presence of IP3 receptors in endoplasmic reticulum rather than plasma membranes. This does not rule out the possibility that IP3 receptors would be detected at a higher level in isolated cilia [6] rather than the PMR fraction, but the low level of IP3 binding to the olfactory PMR fraction contrasts sharply with the high density of binding sites corresponding to odorant amino acid receptors [19, 20]. Clearly, IP3 receptors do not colocalize with odorant receptors in this fraction. Thus, our initial hypothesis that IP3 binding sites would be abundant in the PMR fraction from the olfactory rosettes of Atlantic salmon was not supported by this study.

In contrast, IP4 binding sites were abundant in this PMR fraction, which was previously shown to support odor-stimulated PLC activity [17, 18, 20]. Thus, it is an IP4 binding that colocalizes with odor receptors in the PMR fraction from salmon. Although the binding sites for IP4 appear in the PMR fraction with odor binding sites, we cannot confirm from this result alone that they appear together on the same membrane. In the only other olfactory system in which it has been characterized, IP4 gated a calcium channel in the lobster olfactory system [26]. If in salmon, the colocalization of odor and IP4 binding sites in the PMR fraction extends to a common membrane location, then an IP4 receptor could be an important downstream element in salmon olfactory transduction. The pH optimum and high affinity value for IP4 binding are similar to what has been reported in mammalian brain, but the profile for the competition by other InsPx is somewhat different [37]. The for IP4 binding reflects a density of sites comparable to the density of IP3 binding sites in the olfactory plasma membrane of catfish ( = 17.6 pmol/mg protein from Kalinoski et al. [6]). The value for IP4 binding is much lower (i.e., the affinity is much higher) than the for IP3 binding sites in catfish ( μM from Kalinoski et al. [6]), which is consistent with the lower level of IP4 produced relative to IP3 [38].

In essentially all animal cells, IP3 is metabolized in a bifurcate pathway that includes phosphorylation by a 3-kinase to produce IP4 [28, 39]. Higher-order inositol polyphosphates are also produced in cells along with an array of dephosphorylation products. We included many of these inositol polyphosphates in competition analyses to further characterize the olfactory IP4 binding site. Among the inositol polyphosphates tested, InsP5 and InsP6 showed reasonably potent inhibition of [3H]IP4 binding. These are formed by the sequential actions of specific kinases, are inhibitors of IP4 3-phosphatase and IP4 5-phosphatase [40], and are active in other cellular systems [41]. In contrast, Ins(1,3,4)P3, Ins(1,4)P2, and Ins(4,5)P2 showed little or no ability to interact with the IP4 site. This is not surprising because these are regarded as the products of inactivating phosphatases. Marginal inhibition of [3H]IP4 binding by IP3 (Ins(1,4,5)P3) confirmed the independence of the IP4 and IP3 binding sites in this system and supported the conclusions from direct measurements of [3H]IP3 binding at optimal pH that these sites are not present in the PMR fraction.

In summary, we found a unique IP4 binding site that colocalizes with odor receptors in a subcellular fraction derived from the olfactory system of Atlantic salmon. This is the first biochemical evidence of a putative membrane-bound IP4 receptor in a fish olfactory system. The exact plasma membrane location and the colocalization of odor receptors and putative IP4 receptors in the same plasma membrane remain to be shown. In the only other olfactory system in which it has been studied, electrophysiological studies have demonstrated that IP4 gates a calcium channel and helps regulate Ca2+ entry into lobster olfactory neurons [26], a similar role to that ascribed to IP3 in lobster [27], catfish [6], and carp [14]. This provides the only context with which to interpret the significance of finding IP4 binding sites in membranes of the salmon olfactory system and to begin to suggest that IP4 rather than (or in addition to) IP3 may be a key downstream element for olfactory signal transduction in Atlantic salmon.


The authors thank Dr. Terrance Bradley for providing the Atlantic salmon used in this study and Dr. Ching-Shih Chen for providing the inositol phosphate analogs. Initial portions of this paper were supported by the University of Rhode Island. The authors acknowledge the advice and assistance of Dr. Jian Wang and Dr. Chun-Shiang Chung with the brain fractions and Lee-Ju Cheng with IP3 binding assays. The authors also thank Dr. Karl Hartman for helpful discussions and suggestions.