Metal Ions Suppress the Abnormal Taste Behavior of the Drosophila Mutant malvolio

The gustatory pathway of Drosophila melanogaster is a useful model for the analysis of gene function in the nervous system. An assay which measures the ability of flies to detect and respond to sugars has been used to isolate mutations in several genes affecting taste perception (Isono and Kikuchi, 1973; Falk and Atidia, 1975; Tompkins et al. 1979; VijayRaghavan et al. 1992; Inamdar et al. 1993). Among them, a mutation in the malvolio (mvl) gene affects taste behavior in Drosophila. The gene is expressed in mature neurons in the central and peripheral nervous system as well as in macrophages (Rodrigues et al. 1995). It was shown that the electrophysiological responses of the peripheral neurons to taste stimuli are normal in these flies. This suggests that the abnormal taste behavior of the mutant resulted from a defect in information processing rather than in the reception of the stimulus (Vidal et al. 1993; Rodrigues et al. 1995). Since the amino acid sequence of the malvolio protein (MVL) is 65% identical to that of the mammalian natural resistance-associated macrophage protein (NRAMP) (Vidal et al. 1993, 1995b; Cellier et al. 1995), it is likely that they both have a similar function. These two proteins show a very limited homology to a 22-amino-acid sequence of the nitrate transporter CRNA from Aspergillus nidulans (Vidal et al. 1993; Rodrigues et al. 1995). Consequently, it was proposed that the NRAMP protein, as well as the MVL protein, may function as nitrite (NO₂⁻) or nitrate (NO₃⁻) transporters. These ions are subsequently converted to nitric oxide (NO) by dismutation (Rodrigues et al. 1995). It was suggested that the NO produced in this way may function in the signal transduction of taste perception in Drosophila. The discovery of a yeast protein (Smf1p) that is 30% identical in its amino acid sequence to NRAMP and MVL, and which was shown to function as a metal-ion transporter, raised the possibility that the mammalian and Drosophila proteins also have a similar function (Supek et al. 1996). According to this hypothesis, metal-ion homeostasis is impaired in the MVL mutant, resulting in a loss of taste perception for sugars. To assess this hypothesis, we attempted to suppress the abnormal taste perception phenotype of the malvolio mutant by adding metal ions to the growth medium.

Unless otherwise specified, flies were reared on molasses–cornmeal–yeast medium at 25°C, under constant illumination. The P-element insertional mutant Mvl^97f was kindly provided by Dr William Chia (Rodrigues et al. 1995). The Canton-S strain (CS) was used as a wild-type control, since the Mvl mutant was isogenized to the CS genetic background.

Flies were used for the experiments 3–5 days after eclosion. Flies were collected upon eclosion and aged on medium containing Mn²⁺. Prior to the test, flies were starved for 20 h by transferring them into bottles containing no food but supplied with distilled water soaked in 3MM paper. The feeding preference test was carried out as described by Tanimura et al. (1982) with some modifications according to Rodrigues et al. (1995). Alternate wells of a 6×10 microtiter plate (Nunc, Denmark) were filled with 10 μl of 1% agar (Difco, Noble) containing 100 mmol l⁻¹ trehalose (Sigma) as a stimulus. The rest of the wells contained 0.2% Acid Red (C₂₇H₂₉O₇N₂Na, Sigma) in 1% agar (Fig. 1A). The control experiment ascertained that Acid Red dye at the concentration used (0.2%) was neither toxic nor metabolized and did not interfere with the test. Approximately 100–150 flies were introduced into each plate and were left to feed for 2 h in the dark at 25°C. Thereafter, flies were immobilized by cooling and were visually scored for the color in their abdomens. All the values were obtained by blind scoring. The acceptance response of the stimulus was calculated as the percentage of flies with uncolored abdomens in the population. Means and standard deviations of each data point were obtained from at least six independent experiments.

In the work described in this paper, we attempted to use the feeding preference assay in order to try to recover the phenotypic expression of the taste mutant malvolio. In this assay, flies are presented with an option to choose between food containing an attractant, usually a sugar, and medium containing only a food dye (Acid Red) which is neither attractive nor repellent to the flies. Wild-type flies will almost always prefer food containing the sugar over food containing the Acid Red dye. Fig. 1B depicts two flies, one with an uncolored abdomen, since it was able to taste the sugar and preferred it, and the second with a red abdomen, resulting from consumption of the red dye. The ability of flies to discriminate between food containing sugar and food lacking sugar is manifested in the percentage of flies with an uncolored abdomen in the population tested.

Adult wild-type (Canton-S) and mutant (homozygous Mvl^97f) Drosophila were transferred to bottles containing a series of media with increasing concentrations of MnCl₂, allowed to lay eggs for 4 days and then removed. The eggs developed into adult flies on this medium. The resulting adult flies were then tested for their behavioral response in a feeding preference assay of sugar versus dye solution. When the flies developed in the absence of Mn²⁺, the control wild-type flies had an acceptance rate of more than 40%, whereas the Mvl mutants chose randomly (acceptance rate of approximately 2% only). When wild-type and Mvl mutants were reared throughout development on standard food supplemented with different concentrations of Mn²⁺ and then tested for trehalose acceptance, Mn²⁺ was able to reverse the mutant phenotype in a dose-dependent manner (see Fig. 2). When the same experiment was repeated, but this time by adding Ca²⁺, Mg²⁺ or Zn²⁺ as a replacement for Mn²⁺ in the medium, there was no reversion of the mutant phenotype.

It has previously been shown that, in S. cerevisiae, Zn²⁺ inhibits Mn²⁺ transport by Smf1p, which is a yeast homolog of malvolio (Supek et al. 1996). In flies too, when Zn²⁺ is present in the growth medium together with Mn²⁺ throughout development the ameliorating effect of Mn²⁺ on the mutant flies was suppressed, and these flies again lacked taste preference. At the same Zn²⁺ concentrations, the taste behavior of wild-type flies was normal (Fig. 3). Recently, we observed that Smf2p and Smf3p, which are two yeast homologs of Smf1p, may function in the transport of additional metal ions other than Mn²⁺ and Zn²⁺ (A. Kahan, A. Cohen and N. Nelson, unpublished observations). We therefore tested whether Fe²⁺ affects the taste behavior of the Mvl mutant. As shown in Fig. 3, Fe²⁺ had a similar effect to Mn²⁺ in suppressing the mutant taste behavior. Moreover, when both Fe²⁺ and Mn²⁺ were present in the medium throughout development, an additive effect on the taste preference of the mutant flies was observed.

We next set out to determine whether normal taste perception requires the presence of metal ions throughout the development of the nervous system in the fly. Alternatively, it may be sufficient for these metal ions to be supplied to the adult flies, in which the nervous system has already developed. If so, these metal ions may be implicated in signal transduction of taste perception in the adult. To test this, mutant flies that had developed on standard medium (no elevated concentration of any metal ions) were transferred, upon eclosion, to vials containing food supplemented with Mn²⁺, Fe²⁺ or a combination of Mn²⁺ and Zn²⁺ for 3 days prior to testing. As shown in Fig. 4, these flies had the same pattern of behavior as flies that had been reared throughout development in the presence of the same metal ions. These findings indicate that the metal ions are not required for correct development of the nervous system components involved in taste perception. To corroborate these results, mutant flies that had developed on standard medium were tested on plates in which all the wells, those containing sugar and those containing dye, were supplemented with Mn²⁺, Fe²⁺ or a combination of Mn²⁺ and Zn²⁺. The suppression of the abnormal taste perception of the Mvl mutant by Mn²⁺ and Fe²⁺ was even stronger in this case than when the mutant flies had developed in the presence of the ions or had been fed these metal ions for 3 days prior to testing. Under these conditions, the inhibitory effect of Zn²⁺ was almost undetectable (Fig. 5). These results implicate Mn²⁺ and Fe²⁺ in a signal transduction pathway underlying taste perception. The lack of effect of Zn²⁺ may be explained by the slow accumulation of this ion and the consequent failure to reach the required critical amounts.

Our experiments suggest that the yeast Smf1p and the Drosophila MLV protein are metal-ion transporters. On the basis of the sequence homology between the MVL protein and the mammalian NRAMP protein, it is tempting to suggest that all the other members of this family, including the mammalian ones, have a similar function. In mice, natural resistance to infection with unrelated intracellular parasites such as Mycobacteria, Salmonella and Leishmania is controlled by a single gene on chromosome 1, designated Bcg, Ity or Lsh. The gene product of Bcg, designated natural resistance-associated macrophage protein (NRAMP), has been shown to be altered in susceptible animals (Vidal et al. 1993, 1995a,b; Cellier et al. 1995; Govoni et al. 1995). The discovery of a yeast homolog of Nramp and malvolio that functions as a metal-ion transporter raised the possibility that the mammalian protein has a similar function (Supek et al. 1996, 1997). Following phagocytosis of a parasite into the phagosome, the macrophage produces reactive oxygen and/or nitrogen intermediates that are toxic for the internalized bacteria. Survival of the pathogen during the burst of macrophage respiratory activity is thought to be partly mediated by microbial metalloproteins such as superoxide dismutase (SOD), which contain Mn²⁺, Cu²⁺–Zn²⁺ or Fe²⁺ in their active centers (Chan et al. 1992; Fridovich, 1995; Cooper et al. 1995; Lah et al. 1995). We propose that NRAMP may transport Mn²⁺ and/or other metal ions from the extracellular milieu into the cytoplasm of the macrophage and, following the generation of the phagosome, remove the metal ions from the organelle (Supeck et al. 1997). Thus, metal-ion depletion of the microenvironment in the phagosome by NRAMP may limit the rate of production of metalloenzymes by the engulfed bacteria. This limitation will restrict the pathogen’s ability to produce metalloenzymes such as SOD and prevent the propagation of the ingested microorganisms. Conversely, an increased concentration of metal ions in the phagosome, caused by a defective NRAMP transporter (Bcg^s), may promote the growth of the parasites and render the organism sensitive to the pathogen.

Several lines of evidence point to a direct involvement of metal ions in neurotransmission. Zn²⁺ has been implicated in several processes in the nervous system. For example, Zn²⁺ can interact strongly with a variety of ligands including sulfur in cysteine, nitrogen in histidine and oxygen in acidic amino acids (Berg and Shi, 1996). In mammalian brain cells, Zn²⁺ is accumulated in presynaptic vesicles of excitatory neurons and is released during synaptic activity (Assaf and Chung, 1984; Howell et al. 1984; Palmiter et al. 1996). Zn²⁺ interacts with some ionotropic receptors in the brain, such as the ionotropic ATP receptor (P2x3), which is potentiated by Zn²⁺ (Seguela et al. 1996). Zn²⁺ blocks currents mediated by N-methyl-D-aspartate (NMDA) or γ-aminobutyric acid (GABA) as well as voltage-gated Ca²⁺ channels (Westbrook and Mayer, 1987; Peters et al. 1987). It also interacts with neurotransmitter uptake systems, such as the dopamine transporter, and inhibits dopamine uptake and cocaine binding (Richfield, 1993). Finally, it has been demonstrated that Zn²⁺ may play a role in neuronal death after transient cerebral ischemia (Koh et al. 1996).

The role of Mn²⁺ and Fe²⁺ in neurotransmission and taste perception is not clear. Their possible involvement in neural development has been ruled out since Mn²⁺ and Fe²⁺ suppressed the mutant phenotype even when supplied to flies only at the adult stage, when the nervous system had already been developed. Furthermore, the short period that was sufficient for the mutant flies to recover their taste (see Fig. 5) prompts us to suggest that Mn²⁺ and/or Fe²⁺ participate in a novel signal transduction pathway involved in taste perception. However, it is possible that an appropriate metal-ion concentration in specific nerve cells is required for the optimal performance of these cells in one of the known signal transduction pathways. In Drosophila, the inhibition by Zn²⁺ of the restoration of taste behavior by metal ions can be explained by at least two different mechanisms. Zn²⁺ may compete with Mn²⁺ for uptake by metal-ion transporters other than MVL and, by so doing, prevent the suppression of the malvolio mutant phenotype by the added high concentrations of Mn²⁺. Alternatively, Zn²⁺ may interact directly with the site in the signal transduction pathway that requires Mn²⁺ for its activity. The steps in the signal transduction sequence that may involve metal ions include the modulation of receptor activity in the nervous system, the interaction of receptors with second messengers and the expression of signals in the modulation of Ca²⁺ concentration. Recent studies on the effect of metal ions on neuronal receptors (Shuto et al. 1997) makes it conceivable that Mn²⁺ and Fe²⁺ may modulate a subset of excitatory receptors involved in taste perception.

Recently, a novel mechanism for regulating ion concentrations in yeast cells has been discovered (Liu et al. 1997). It was demonstrated that in Saccharomyces cerevisiae a mutation in the BSD2 gene suppresses oxidative damage in cells lacking superoxide dismutase (Liu and Culotta, 1994). The mechanism for this effect was explained when it was subsequently found that BDS2 prevents metal hyperaccumulation by exerting negative control over the SMF1 and SMF2 metal transport systems (Liu et al. 1997). The gene product of BSD2 (Bds2p) is situated in the endoplasmic reticulum, while Smf1p and presumably Smf2p function in the plasma membrane (Supek et al. 1996, 1997; Liu et al. 1997). The most plausible explanation for the function of Bds2p is that it acts as an endoplasmic reticulum receptor that binds several plasma membrane transporters and, by doing so, prevents them from over-accumulating on the plasma membrane. Null mutations in the BSD2 gene cause an increase in the number of metal-ion transporters in the plasma membrane, resulting in over-accumulation of metal ions, presumably by chemically catalyzing the superoxide dismutation reaction. Bsd2p must also sense, directly or indirectly, the metal-ion concentration inside the cells in order to be able to distribute the specific transporters appropriately between the endoplasmic reticulum and the plasma membrane. When the metal-ion concentration is elevated above a specific threshold, the transporters may be tightly bound and remain in the endoplasmic reticulum. When the metal-ion concentration decreases, the transporters are released and distributed to the plasma membrane. This kind of regulation may be particularly useful for the nervous system, the blood barriers and the digestive tract. We also anticipate the involvement of such transporter receptors in mutants of taste perception in which the MVL protein is intact. Very recently, NRAMP2 (DCT1) was shown to transport Fe²⁺ and other metal ions (Gunshin et al. 1997), confirming our findings that the members of this family of membrane proteins are metal-ion transporters (Supek et al. 1996, 1997). A genetic disorder that causes hemochromatosis due to elevated rates of uptake of iron from the intestine was localized to a gene that may produce a protein analogous to Bsd2p (Feder et al. 1996). We suggest that this mechanism of holding excess and ready-to-use transporters in the endoplasmic reticulum and/or the Golgi body is widespread.

The ramification of this work lies far beyond the fact that the addition of metal ions corrects a defect in Drosophila taste behavior. The data not only suggest that Drosophila MVL, like yeast Smf1p, is a metal-ion transporter but that all the other family members are also metal-ion transporters. Furthermore, it suggests that these family members function in key metabolic processes that are involved in systems such as signal transduction, neuronal activity and resistance to bacterial infection. We anticipate that the other members of this family will play key roles in metal-ion homeostasis in a variety of processes that are yet to be discovered.

We wish to thank Dr William Chia for providing us with the Drosophila strains. This work was supported by a grant from the Israel Science Foundation.