Cell Physiology

Regulation of chloride permeability by endogenously produced tyramine in the Drosophila Malpighian tubule

Edward M. Blumenthal


The Malpighian (renal) tubule of Drosophila melanogaster is a useful model for studying epithelial transport. The purpose of this study was to identify factors responsible for modulating transepithelial chloride conductance in isolated tubules. I have found that tyrosine and several of its metabolites cause an increase in chloride conductance. The most potent of these agonists is tyramine, which is active at low nanomolar concentrations; the pharmacology of this response matches that of the previously published cloned insect tyramine receptor. In addition, the tubule appears capable of synthesizing tyramine from applied tyrosine, as shown by direct measurement of tyrosine decarboxylase activity. Immunohistochemical staining of tubules with an antibody against tyramine indicates that the principal cells are the sites of tyramine production, whereas previous characterization of the regulation of chloride conductance suggests that tyramine acts on the stellate cells. This is the first demonstration of a physiological role for an insect tyramine receptor.

  • tyrosine decarboxylase
  • yohimbine
  • tyramine receptor
  • biogenic amines

the malpighian tubulescomprise the initial component of the insect excretory system and have been a useful model system for studying the regulation of epithelial ion transport (4, 17, 37). The fruit flyDrosophila melanogaster possesses two pairs of tubules; these are blind-ended simple tubular epithelia that empty into the digestive tract (17). The main segment of the tubule, in which the secretion of a potassium-rich primary urine occurs, contains two morphologically defined cell types, principal cells and stellate cells (52). Recent work has shown that the principal cells are both genetically and functionally heterogeneous (43,49), although neither the molecular nor the physiological bases for this heterogeneity have been determined.

In Drosophila, as in other insects, cations move into the tubule lumen by active transport through the principal cells (4,16, 25, 32). The electromotive force for this transport is generated by a vacuolar-type proton pump located in the apical membrane. Chloride moves passively into the lumen down its electrochemical gradient through a pathway that lies outside of the principal cells (32, 34). The anatomical site of this chloride shunt conductance, paracellular or transcellular through the stellate cells, has been controversial; however, regulation of chloride conductance in the Drosophila tubule is linked to intracellular calcium levels in the stellate cells (33). The net effect of ion transport in the Drosophila tubule is the establishment of a lumen-positive transepithelial potential (TEP) with an amplitude of ∼50–60 mV.

Ion transport across the Drosophila tubule is a highly regulated process, falling under the control of at least three different second messenger systems. Cation transport through the principal cells can be stimulated by both cGMP and cAMP (32). The former is produced after treatment of tubules with the diuretic hormone cardioacceleratory peptide 2b (CAP2b) (15); production of cAMP is stimulated by a calcitonin-related peptide (14). In many other insect species, a corticotropin-releasing factor (CRF)-like diuretic peptide also acts through cAMP (13). In addition, the diuretic hormone leucokinin stimulates urine secretion by raising calcium levels in the stellate cells and increasing the amplitude of the chloride shunt conductance (33). Recently, a receptor for leucokinin has been cloned and found to be expressed exclusively in stellate cells (41). The biogenic amines serotonin and octopamine have diuretic activity on the tubules of multiple insect species (26, 31, 48); however, no aminergic modulation ofDrosophila renal function has previously been reported.

In the tubules of many insects, including Drosophila, the TEP measured in vitro is not constant but instead undergoes large oscillations in amplitude (15, 30, 39, 53). Recent studies in the mosquito Aedes aegypti and in Drosophilahave shown that these voltage oscillations are caused by fluctuations in transepithelial chloride conductance (5, 7), and inDrosophila they are dependent on increased intracellular calcium levels in the stellate cells (7). No function has yet been determined for these oscillations; neither have they been found to be caused by any particular stimulus. However, the occurrence of such oscillations in the tubules of several different insect species raises the possibility that they may be important for the regulation of normal ion transport. The initial purpose of this study was the identification of the factor responsible for the triggering of these oscillations; the result, as described below, was the identification and characterization of tyramine as a novel insect diuretic factor.



All chemicals were obtained from Sigma (St. Louis, MO).

Drosophila maintenance.

Drosophila melanogaster (Canton-S) were maintained on a 12:12-h light-dark cycle at 23–25°C on cornmeal-molasses-yeast medium.

Tubule isolation.

Posterior Malpighian tubules were dissected under dissecting saline from adults females 6–8 days posteclosion and placed in a tissue culture dish in which a 100-μl drop of 0.125 mg/ml poly-l-lysine had been dried to promote adhesion of the tubule to the dish, and the solution was replaced with recording medium (32). The dissecting/recording saline contained (in mM) 85 NaCl, 20 KCl, 3 CaCl2, 12 MgSO4, 7.5 NaHCO3, 4 NaH2PO4, 15 glucose, and 10 HEPES, pH 6.75 [osmolality = 255–270 mosmol/kgH2O as measured with a vapor pressure osmometer (Wescor, Logan, UT)]. Standard bathing medium (SBM) consisted of a 1:1 mixture of Schneider's Drosophila medium (Invitrogen, Carlsbad, CA) and a “diluting saline” containing 36 NaCl, 21 KCl, 15 MgCl2, 5 CaCl2, 4.8 NaHCO3, 2 NaH2PO4, 11.1 glucose, and 15 HEPES, pH 6.75. The composition of SBM, based on the compositions of the diluting saline and Schneider's medium (Invitrogen) is (in mM) 36 NaCl, 21 KCl, 5.2 CaCl2, 7.5 MgSO4, 7.5 MgCl2, 4.8 NaHCO3, 1.3 KH2PO4, 3.4 sodium phosphate, 7.5 HEPES, 0.7 α-ketoglutaric acid, 11.1d-glucose, 0.43 fumaric acid, 0.38 malic acid, 0.42 succinic acid, 2.9 trehalose, 2.8 β-alanine, 1.15l-arginine, 1.5 l-aspartic acid, 0.25l-cysteine, 0.21 l-cystine, 2.7l-glutamic acid, 6.15 l-glutamine, 1.67 glycine, 1.29 l-histidine, 0.58 l-isoleucine, 0.58 l-leucine, 4.5 l-lysine HCl, 2.7l-methionine, 0.45 l-phenylalanine, 7.4l-proline, 1.19 l-serine, 1.47l-threonine, 0.25 l-tryptophan, 1.38l-tyrosine, 1.32 l-valine, and 1,000 mg/l yeastolate. The osmolality of SBM was 255–270 mosmol/kgH2O. The osmolality of these solutions is equal to that reported for adult Drosophila hemolymph (47) but is much lower than that used in other studies ofDrosophila tubule physiology. This difference in osmolality explains the higher urine secretion rates and more prominent TEP oscillations than those reported by others (8). For the chloride replacement experiment, the low chloride media consisted of 10% normal dissecting/recording saline and 90% dissecting/recording saline in which NaCl, KCl, and CaCl2 had been replaced with 85 mM sodium isethionate, 10 mM K2SO4, and 3 mM CaSO4. Experiments were conducted within 2 h of tubule dissection.


The tubule lumen and principal cells were impaled with a sharp electrode (R > 25 MOhms) pulled from theta-glass (Sutter Instruments, Novato, CA) and filled with 3 M KCl. Potentials were amplified (Axopatch 200B; Axon Instruments, Foster City, CA), digitized at 100 Hz, and stored online. Recording and analysis were conducted using pCLAMP software (Axon Instruments). The peritubular bath was continuously perfused during recording. Drugs were applied to and removed from the bath during recording by switching perfusion lines.

Quantitation of electrophysiological responses.

To compare the responses of tubules to various agonists, a response index was calculated as follows. A straight line was fitted to the TEP record for the 30 s preceding the drug application and then extrapolated through the time of the application. The area under the TEP trace and under the extrapolated line was measured for a period beginning 15 s after the beginning of the drug application and ending 15 s after the end of the drug application. The baseline for the integral was −10 mV, based on the observation that during maximal responses, the TEP drops to approximately this value. The response index was calculated as 1 − (area under the TEP trace/area under the extrapolated line). In occasional traces, small instabilities in the TEP led to extrapolated lines that did not accurately reflect the behavior of the TEP; in those cases, the 30-s fitting interval was adjusted to give a more accurate line.


RT-PCR from whole fly and Malpighian tubule cDNA was performed as previously described (7). Primer sequences for the tyramine receptor (accession no. AB-073914) were ACAAGGACTCAGCGGGAGAATG and CGATGATGGTCAGCACGATAATG.

Urine secretion assay.

The rate of urine secretion from isolated tubules was measured as described (16). Posterior tubules were dissected from adult females under SBM or saline (depending on the experiment) and then moved into a 15 μl of droplet of SBM or saline under mineral oil. One tubule was pulled out of the droplet and wrapped around a dissecting pin such that the cut end of the ureter and the lower section of the other tubule branch were pulled into the oil. Periodically, the secreted urine droplet was removed from the ureter with a fine glass rod, and its diameter was measured with an ocular micrometer. The volume of the droplet was calculated assuming spherical geometry. Collection intervals ranged from 12–20 min. Secretion rate was calculated as urine volume/collection time. For drug application, 12 μl of the bathing droplet was replaced twice with 12 μl of bathing solution + drug. The bathing droplet of control tubules was replaced twice with bathing solution alone.

Determination of chloride concentration in urine.

The urine secretion assay was performed as described above. After the diameter of each urine droplet was measured, the droplet was taken up into a 250 nl of capillary tube (Drummond Scientific, Broomall, PA). The volume of the urine was again determined by measuring the length of the aqueous phase within the capillary. Any samples for which the two volume measurements disagreed by >10% were discarded from further analysis. The urine droplet was expelled into a tube containing 20 μl of H2O. Capillary electrophoresis was performed on the samples as described using a Waters Quanta 4000 (23). The elution time and area of the ultraviolet absorbance peaks were compared with standards of known composition and concentration, and the ion concentrations in the original urine droplet were then calculated.

Tyramine immunohistochemistry.

Tubules were dissected from 7- to 9-day-old females under PBS and then incubated for 20 min in PBS containing 1 mM l-tyrosine before fixation. Immunostaining was performed as previously described (11), with the following modifications: the concentration of NaBH4 was 1%, and 0.1% BSA was added to the PMT solution. The primary antibody was a rabbit anti-tyramine antibody (Chemicon International, Temecula, CA), diluted 1:2,500. Immunostaining was detected with a biotinylated anti-rabbit secondary antibody, diluted 1:250 (Jackson ImmunoResearch, West Grove, PA), followed by the Vectastain peroxidase ABC kit (Vector Laboratories, Burlingame, CA) and incubation with DAB/peroxide/nickel.

Tyrosine decarboxylase enzyme assays.

Assays for tyrosine decarboxylase (TDC) activity were performed on tubule extracts as described (28). Tubules from 18 adult females (4–7 days posteclosion) were homogenized on ice in 90 μl of buffer (50 mM Tris, pH 7.5, and 1 mM phenylthiourea). Protein content was measured in an aliquot of homogenate using the Bio-Rad protein assay reagent (Bio-Rad, Hercules, CA). Tubule homogenate (12 μl) was mixed with 48 μl of assay mix [100 mM sodium phosphate, pH 6.8, 0.1 mM pyridoxal phosphate, 1 mM β-mercaptoethanol, 0.1 mM EDTA, and 20 μCi/ml l-[3,5-3H]tyrosine (Amersham, Piscataway, NJ)] and either incubated for 60 min at 30°C or placed on ice and processed immediately. Reactions were stopped by the addition of 300 μl of sodium phosphate, pH 8.0, and extracted with 100 μl of chloroform containing 100 mM diethylhexylphosphoric acid. The organic phase was reextracted with an additional 300 μl of sodium phosphate, and then 80 μl of the organic phase was placed into a scintillation vial and allowed to dry before the addition of scintillant and counting. Each experiment consisted of two zero-timepoint reactions (background) and three 60-min reactions; TDC activity was calculated by subtracting the averaged background counts from the averaged 60-min counts.


Isolated Drosophila Malpighian tubules exhibit a lumen-positive TEP that undergoes large oscillations in amplitude, but the cause of these oscillations has not yet been determined (7,15). During the characterization of these oscillations, it became clear that although they were always present in tubules bathed in SBM, which contains tissue culture medium, oscillations were rarely seen when tubules were bathed in a simple saline. This effect of SBM is demonstrated most clearly by switching the bathing medium from saline to SBM; a typical response is shown in Fig.1 A. As is clear from this trace, SBM has two effects on the TEP. The first is a rapid and reversible depolarization and induction of oscillations. The appearance of oscillations is quantified by an increase in the coefficient of variation of the TEP (Fig. 1 B, left). The second effect of SBM is an increase in the amplitude of the TEP that persists after the SBM is washed out (Fig. 1 B, right). This study focuses on the depolarizing activity of SBM (however, see Fig. 7).

Fig. 1.

A: electrophysiological response of a tubule to standard bathing medium (SBM). The bathing solution was switched from saline to SBM and back to saline, as indicated by the bar.B: quantitation of the effects of SBM on transepithelial potential (TEP) amplitude and coefficient of variation. * Significantly different from saline value; P < 0.05; paired t-test. **P < 0.0001; n.s., not significantly different from saline value; n = 6 tubules. Error bars in this and subsequent figures indicate SE.

SBM contains 50% Schneider's Drosophila medium, which is a mixture of yeast extract, salts, sugars, metabolic intermediates, and 20 amino acids (see methods for precise composition). To determine which component of the medium is responsible for the induction of oscillations, tubules were exposed to solutions containing different constituents of the complete medium. TEP oscillations were observed in the presence of either yeast extract or a mixture of all of the amino acids (data not shown). Individual testing of 10 of these amino acids revealed that tubules respond only tol-tyrosine. As shown in Fig.2 A, application ofl-tyrosine causes the TEP to depolarize and oscillate in a manner indistinguishable to that induced by complete SBM.

Fig. 2.

Responses of tubules challenged with different concentrations of l-tyrosine (A) and tyramine (B) in saline. C: dose-response profiles of tyramine, l-tyrosine, dl-octopamine, and dopamine. n = 5–7 tubules/point. Individual tubules were challenged with up to 3 different concentrations of a single agonist; applications lasted 45 s and were separated by 3 min of saline. See methods for quantitation of the responses.

The amino acid tyrosine serves as a precursor for the synthesis of several biogenic amines (29, 45); tyrosine metabolism in insects is diagrammed in Fig. 3. Unlike in mammals, where a single aromatic amino acid decarboxylase acts upon both tyrosine and DOPA, insects possess separable TDC and DOPA decarboxylase (DDC) enzymes (54). Because it seemed possible that one of these metabolites could be mediating the effect of tyrosine, I tested three biogenic amines–dopamine, octopamine, and tyramine–on tubules. All three compounds elicited qualitatively similar electrical responses that also resembled those caused by tyrosine (Fig. 2 B). At low doses, these amines caused the TEP to oscillate. Intermediate doses elicited a transient depolarization followed by oscillations, and high doses resulted in a sustained depolarization. This pattern of responses is quite similar to that observed in Aedes tubules to increasing doses of leucokinin (51). Strikingly, tyramine was several orders of magnitude more potent than any other compound tested, giving a measurable response at a concentration of 1 nM and a maximal response at ∼100 nM (Fig. 2 C). In contrast, octopamine caused measurable responses only at concentrations above 1 μM. The preferential response of tubules to tyramine over octopamine is particularly interesting, given that for many years, tyramine was thought simply to be a metabolic intermediate on the octopamine synthesis pathway. Only relatively recently have several lines of evidence suggested that tyramine can function as an independent signaling molecule; chief among these has been the cloning fromDrosophila and other insect species of a receptor that is preferentially activated by tyramine over octopamine (2, 6, 40,42, 46, 50).

Fig. 3.

Pathways for metabolism of tyrosine in insects. TDC, tyrosine decarboxylase; TβH, tyramine β-hydroxylase; TH, tyrosine hydroxylase; DDC, DOPA decarboxylase.

To determine whether the antagonist pharmacology of the tubule response also matched that of the cloned tyramine receptor, I assayed the ability of several tyramine/octopamine receptor antagonists to block the depolarization elicited by an application of 50 nM tyramine. A full dose-response profile was generated for yohimbine, which is the most potent inhibitor of tyramine binding in the locust brain and to the cloned receptor (20, 46, 50); yohimbine blocked the response of tubules to 50 nM tyramine with a half-maximal concentration of ∼300 nM. The rank-order potency of the three antagonists tested was yohimbine ≥ phentolamine > metoclopramide (Fig.4). This profile is very similar to that of an insect tyramine receptor stably expressed in aDrosophila cell line (50) and of the tyramine binding site in locust brain (20) but does not resemble that of any of the octopamine receptor subtypes (21). Consistent with the pharmacology, RT-PCR analysis shows that the gene encoding the tyramine receptor is expressed in the Malpighian tubules (Fig. 5).

Fig. 4.

Antagonist pharmacology of the tyramine response. Tubules were treated with two 45-s applications of 50 nM tyramine (TA) separated by 3 min of saline. Yohimbine [1 μM (A) or 10 μM (B)] was added to the saline during the interval after the first TA application and during the second TA application.C: dose-response curve for yohimbine and single concentrations of phentolamine and metaclopramide, illustrating the relative potencies of the 3 antagonists. The ratio of the second TA response/first TA response is given; application profiles were identical to those shown in the first 2 panels. n = 4–6 tubules/point.

Fig. 5.

RT-PCR of tyramine receptor transcript from tubules.Lane 1: whole fly cDNA. Lane 2: tubule cDNA.Lane 3: tubule negative control (no reverse transcriptase).Lane 4: 100-bp ladder (Invitrogen). Expected products are 141 bp from mRNA and 700 bp from genomic DNA.

I next sought to characterize the ionic basis of the tyramine-induced depolarization. Previous work has shown that the TEP oscillations seen in the presence of SBM are due to fluctuations in the transepithelial chloride conductance (7). Because the Nernst potential for chloride between the peritubular bath and the lumen is near 0 mV, an increase in chloride conductance will cause the TEP to depolarize. It was therefore of interest to determine whether the tyramine-induced depolarizations were also associated with an increase in chloride conductance. Such was likely to be the case, both because high concentrations of tyramine cause the TEP to depolarize to near 0 mV and also because the electrical response of the tubule to tyramine resembles the response to leucokinin, which is known to increase chloride conductance (32). The chloride substitution experiment shown in Fig. 6 shows that tyramine does indeed cause a large increase in transepithelial chloride conductance, as indicated by the increased chloride diffusion potential in the presence of tyramine.

Fig. 6.

Tyramine increases transepithelial chloride conductance.A: trace recorded in saline. At bars marked “low Cl,” saline was switched to a saline in which 90% of the chloride was replaced with isethionate (see methods). At bars marked “TA,” saline was switched to a saline containing 10 μM tyramine, first with normal chloride, then with 10% chloride during low Cl/TA bar. A final application of low chloride saline was given after the electrode was withdrawn from the tubule to measure the diffusion potential at the open electrode. B: average responses to low chloride saline in the absence and presence of 10 μM TA. Responses were corrected for the electrode diffusion potential. n = 7 tubules; P < 5 × 10−7; paired t-test.

It is evident that the treatment of tubules with tyramine receptor agonists causes an electrical response indistinguishable from that induced by SBM. To determine whether the components of SBM act solely through the tyramine receptor, I measured the ability of yohimbine to block the response of the tubule to SBM. Figure7 A shows the TEP coefficient of variation measurements from tubules recorded in SBM in the presence of yohimbine; oscillations are blocked with a dose dependence very similar to that shown in Fig. 4 for the antagonism of tyramine responses. At the highest concentrations of yohimbine, TEP oscillations are completely absent, although the coefficient of variation is nonzero due to small fluctuations and drift in the TEP. Importantly, long-term (1–2 h) exposure to yohimbine did not appear to be toxic to the tubules, because there was no reduction in TEP amplitude with increasing concentrations of yohimbine. These results demonstrate that all of the oscillation-inducing activity of SBM is due to the activation of yohimbine-sensitive tyramine receptors.

Fig. 7.

A: yohimbine eliminates SBM-induced oscillations. The graph shows the TEP coefficient of variation for tubules incubated in SBM with various concentrations of yohimbine for 1–1.8 h and then recorded in the same medium. n = 4–8 tubules/point. B and C: same experiment as in Fig. 1, but all recording solutions contain 100 μM yohimbine.n = 7 tubules. There is no significant effect of SBM + yohimbine on the TEP coefficient of variation; however, the increase in TEP amplitude is significant (P < 0.05; paired t-test). The starting TEP amplitudes of these tubules were not significantly different from those in Fig. 1, nor was the relative increase in amplitude after SBM treatment different (unpairedt-test).

Interestingly, the ability of SBM to increase the amplitude of the TEP, shown in Fig. 1, is not blocked by yohimbine. Figure 7, Band C, shows a similar experiment to that in Fig. 1, but in the presence of 100 μM yohimbine. Clearly, the blockade of the tyramine receptor prevents the depolarization and oscillations seen previously; however, the sustained increase in TEP amplitude is identical to that seen in Fig. 1. The time course of this increase is much more apparent in the absence of the depolarization and oscillations. This augmentation of the TEP could be due to a specific component of the SBM acting on a yohimbine-insensitive receptor; alternatively, exposure of the tubule to a rich medium such as SBM may result in a more general stimulation of epithelial transport. Also seen in the trace in Fig. 7 B are a rapid hyperpolarization and depolarization when SBM is added and removed. These rapid effects are most likely due to differences in the ionic composition of saline and SBM (see methods). Most notably, SBM contains significantly less sodium than the saline, and removal of sodium from the peritubular bath leads to a hyperpolarization of the TEP (unpublished results).

Because yohimbine appears to be a selective blocker of tyraminergic signaling in the tubule, it is possible to use this antagonist to examine the role of tyramine in the function of the intact tubule. Urine secretion rates were measured in isolated tubules bathed in SBM; the addition of 100 μM yohimbine caused a 60% reduction in the rate of urine production (Fig. 8 A). Measurements of the chloride concentration in the secreted urine droplets showed no effect of yohimbine (Fig. 8 B). Thus the 60% reduction in secretion rate indicates a reduction in transepithelial chloride flux of the same magnitude. In isolated tubules bathed in SBM, therefore, approximately half of the transepithelial chloride transport occurs through a yohimbine-sensitive pathway, whereas the other half of the chloride transport persists in the absence of tyramine signaling.

Fig. 8.

Effect of tyraminergic signaling on urine secretion and composition. A: inhibition of secretion rate by 100 μM yohimbine. n = 3 control and 7 yohimbine-treated tubules. Tubules were dissected and bathed in SBM. Yohimbine was added to the bathing droplet at the end of the second collection interval. For each tubule, each individual secretion rate measurement was normalized to the average of the first 2 measurements for that tubule. These average initial rates did not differ between the control and yohimbine-treated groups (3.2 ± 0.4 nl/min and 2.7 ± 0.3 nl/min, respectively). B: yohimbine treatment (100 μM) did not alter the concentration of chloride in the urine. n= 3 tubules; P = 0.5; paired t-test.C: stimulation of secretion by 1 μM tyramine.n = 5 control and 6 tyramine-treated tubules. Tubules were dissected and bathed in saline, and tyramine was added at the end of the second collection interval. Secretion rates were normalized as in A. Average initial rates did not differ between the control and tyramine-treated groups (1.5 ± 0.1 nl/min and 1.2 ± 0.2 nl/min, respectively).

If tyramine acts to stimulate transepithelial chloride flux, it should trigger diuresis when added to isolated tubules. Figure 8 Cshows that this is indeed the case. When added to tubules that were bathed in saline, 1 μM tyramine caused a 45% increase in urine secretion (P < 0.005, paired t-test).

The data shown in Fig. 2 demonstrate that tubules respond to tyrosine as well as tyramine, albeit at far greater concentrations. Tyrosine could be acting in one of two ways to elicit this response; it could be a weak agonist of the tyramine receptor, or it could be converted into tyramine by a TDC activity endogenous to the tubule. Two lines of evidence suggest that the latter is true. First, tubules contain a significant level of TDC activity. Assays of tubule extracts show a TDC activity of 18.8 ± 2.6 fmol tyramine · min−1 · mg−1protein (n = 2 experiments). The only basis for comparison of this value with another Drosophila tissue is a report of TDC activity in extracts of adult brain of ∼4 fmol · 10 min−1 · brain−1(28). Converting the tubule activity yields a value for the activity from one fly's set of tubules of 0.21 ± 0.05 fmol/10 min; thus the tubules in a fly contain ∼5% of the TDC activity of the brain. Given that the tubules consist of just over 300 cells (49), the amount of activity per cell is far higher in the tubules than in the brain.

The second line of evidence for endogenous conversion of tyrosine to tyramine is the observation that d-tyrosine, the inactive enantiomer of l-tyrosine, inhibits the response of wild-type tubules to l-tyrosine but not to tyramine. Figure9 shows traces from tubules challenged with two applications of either l-tyrosine or tyramine; between the two applications, the tubule is treated with eitherd-tyrosine or saline. As quantified in Fig. 9 C, the second response to l-tyrosine is nearly eliminated by the d-tyrosine application, whereas the tyramine response is unaffected. This selective antagonism would be virtually impossible to explain if l-tyrosine were a direct agonist of the tyramine receptor. If, however, the tubule is converting tyrosine to tyramine, there are many steps that could potentially be inhibited byd-tyrosine (see discussion). Taken together therefore, the d-tyrosine and TDC assay results provide strong evidence for endogenous production of tyramine by theDrosophila tubule.

Fig. 9.

Effect of d-tyrosine on thel-tyrosine response. A: responses of tubules to two 45-s applications of 500 μM l-tyrosine separated by a 3-min application of either 500 μM d-tyrosine (left trace) or saline (right trace).B: same as in A, except that the agonist is 50 nM tyramine. C: average ratios of the second response/first response, showing the effect of d-tyrosine on the secondl-tyrosine response (P < 0.001 compared with saline; unpaired t-test) and the second tyramine response (not significantly different from saline). n = 5–6 tubules/condition. There were no significant differences in the amplitudes of the first responses from tubules treated withl-tyrosine/d-tyrosine vs.l-tyrosine/saline, tyramine/d-tyrosine vs. tyramine/saline, or l-tyrosine/d-tyrosine vs. tyramine/d-tyrosine.

The tubule is a heterogeneous tissue, and TEP measurements do not give any information as to cellular localization of function; therefore, it is important to determine which cell type is responsible for the production of tyramine. Because TDC is thought to be an intracellular enzyme (19), I reasoned that the tyramine-producing cells should contain relatively high levels of tyramine. Figure10 shows the pattern of immunostaining seen with an antibody against tyramine. The principal cells are labeled, whereas stellate cells are stained either lightly or not at all. Thus the principal cells appear to be the major site of tyramine production.

Fig. 10.

Immunolabeling of tubules with an antibody against tyramine.A: immunoreactivity is apparent throughout the tubule, indicating staining of the principal cells, which make up the large majority of the epithelium. B: absence of staining in a control tubule treated without the primary antibody. C: higher magnification of tubule in A. A stellate cell (arrow) is visible as an unstained “window” in the tubule. The diameters of the tubules are ∼35 μm.


I have shown that tyramine causes an increase in chloride conductance across the Drosophila Malpighian tubule. The pharmacology of this response matches very closely that of the cloned insect tyramine receptor, both in the selectivity of tyramine over other biogenic amines, particularly octopamine, and in the rank-order potency of several tyramine/octopamine antagonists (40, 42, 46,50). This represents the first report of a physiological response with the same pharmacology as the tyramine receptor and the first well-defined role for tyramine in an insect tissue. Three previous studies have implicated tyramine in other aspects of insect physiology. First, tyramine has been reported to be necessary for the sensitization of Drosophila to repeated applications of cocaine (28), although this function has not yet been localized anatomically. In addition, two brief reports showed differential effects of tyramine, compared with octopamine, in stimulating trehalose metabolism in isolated cockroach fat bodies (18) and in inhibiting the contraction of locust visceral muscle (22). No additional pharmacological characterization was performed in either system. Although I have not shown conclusively in the Drosophila tubule that the cloned tyramine receptor is responsible for the increased chloride conductance, the expression of the receptor in the tubule and the close match in the pharmacology of the response with that of the cloned receptor make this highly likely to be the case.

A recent paper has reported the identification of flies carrying a mutation (hono) in the tyramine receptor gene (24). These flies contain a transposon inserted upstream of the first noncoding exon of the receptor gene; this exon is ∼25 kb upstream of the rest of the gene. I have found that, although the tubules of hono flies are hyposensitive to tyramine, tubules from the parental stock, which does not contain the transposon insertion, are similarly hyposensitive (unpublished results). Indeed, tubules from many common laboratory stocks are hyposensitive to tyramine; the genetic basis for this phenotype is currently under investigation. It is possible that the hono insertion does not affect the expression of the tyramine receptor in the tubule, perhaps due to the utilization of an alternative transcriptional start site or other tissue-specific message processing.

A model for the synthesis and action of tyramine in the tubule is presented in Fig. 11. Three features of this model merit discussion. The first is the synthesis and release of tyramine by the tubule. The evidence for this is as follows: first, tyrosine and tyramine both increase chloride conductance, and the actions of both are blocked by yohimbine with a very similar potency (compare Figs. 4 C and 7 A), suggesting that both stimulate the same receptor. Second, the d-tyrosine effect presented in Fig. 9 argues against a direct action of tyrosine at the tyramine receptor. Finally, the tubule contains significant TDC activity and so is capable of converting tyrosine to tyramine. Thus tyrosine must be taken up from the peritubular bath and decarboxylated into tyramine. This tyramine must then be released; release is most likely across the basolateral membrane because the tyramine receptor is accessible to agents applied to the peritubular bath, although I cannot exclude the possibility that tyramine can also act within the tubule lumen. The molecular mechanisms governing the uptake of tyrosine and subsequent release of tyramine are unknown, although the data shown in Fig. 9 suggest that one of these steps is inhibited byd-tyrosine.

Fig. 11.

Model for tyraminergic signaling in the tubule. See text for details. A principal cell is at left, unshaded, and a stellate cell is at right, shaded.

The second feature of the model is that the principal cells are the site of tyramine production and release. This is a likely but not definitive conclusion from the data. Figure 10 clearly shows that principal cells contain tyramine and that stellate cells do not stain strongly with the antibody, but it remains possible that stellate cells are also capable of synthesizing tyramine. Interestingly, incubation of tubules in the absence of tyrosine, which the electrophysiology suggests should deplete the “releasable” pool of tyramine, does not lead to a noticeable reduction in tyramine immunoreactivity (data not shown), demonstrating that most of the immunostaining does not represent the releasable pool of tyramine. Clearly then, the production of tyramine by the tubule is likely to be compartmentalized, and further studies are needed for its conclusive localization.

The third feature of the model is that tyramine binds to receptors on the stellate cells. Previously, I showed that the TEP oscillations seen in SBM were governed by calcium levels in the stellate cells; the oscillations are eliminated by chelation of intracellular calcium, and treatment of tubules with leucokinin, which increases calcium levels specifically in stellate cells (33, 41), causes a long-lasting suppression of the oscillations (7). The current work has demonstrated that the TEP oscillations are the result of the activation of tyramine receptors, thereby linking this signaling pathway, albeit indirectly, with intracellular calcium levels in the stellate cells. Consistent with this hypothesis, activation of the cloned insect tyramine receptor can cause increases in intracellular calcium levels in several heterologous systems (40, 42,44). It now appears likely that the leucokinin-induced suppression of oscillations resulted from a cross-desensitization of the leucokinin and tyramine signaling pathways and that both pathways converge at some point downstream of receptor activation to stimulate chloride transport. The pathway for this chloride transport, paracellular or transcellular, is still unknown, and both possibilities are shown on the model.

The action of tyramine by the tubule bears a striking resemblance to dopaminergic signaling in the mammalian kidney. In the latter case, dopamine is employed in the proximal tubule as a paracrine or autocrine hormone to regulate sodium transport (reviewed in Ref. 1). The dopamine is produced from circulating DOPA by an aromatic amino acid decarboxylase (AADC). Interestingly, AADC is also able to decarboxylate tyrosine to produce tyramine. Although DOPA is the preferred substrate, it has been calculated that due to the much higher levels of tyrosine than DOPA in the circulation, tyramine and dopamine should be produced in vivo at comparable rates (10). Indeed, the kidney contains high levels of tyramine (36), and a recently cloned mammalian tyramine receptor is expressed in the kidney (9, 12). There are currently no specific pharmacological tools for manipulating tyraminergic signaling in vertebrates–yohimbine is also an α-adrenergic antagonist–but it is possible that tyramine will prove to be important in vertebrate renal function as well.

The current work raises the possibility that chloride transport in the tubule may be more complicated than was previously appreciated. Until now, there has been no evidence for heterogeneity of chloride transport pathways; hence, agents that increase chloride conductance, such as leucokinin, are thought to increase the amplitude of the conductance pathway that is active in the unstimulated tubule (33,34). The data shown in Fig. 8 suggest an alternative possibility; the tubule may possess multiple, pharmacologically separable pathways for chloride transport. At least one pathway is calcium sensitive and can be stimulated by leucokinin and tyramine. When tyraminergic signaling is blocked by yohimbine, however, there is still a significant level of chloride transport and urine secretion. This could represent a basal level of activation of the calcium-sensitive pathway, but it could also represent an additional, calcium-independent pathway. Unfortunately, the lack of specific pharmacological agents for blocking chloride conductance makes it difficult to distinguish between these two possibilities. It is hoped that the techniques available in Drosophila for the manipulation of gene expression may allow for the resolution of this issue.

Bathing solutions containing Schneider's medium have been used as the control medium in many studies involving Drosophila tubule function. It is clear from the current study that tubules bathed in SBM are not unstimulated relative to saline. Rather, SBM influences tubule physiology in at least two different ways; first, by increasing chloride conductance through the activation of tyramine receptors, and second, by increasing the amplitude of the TEP through an as-yet uncharacterized mechanism.

The stimulation of chloride conductance and urine secretion by tyramine is likely to be physiologically important in the intact fly and not simply an artifact of the recording conditions in vitro.Drosophila hemolymph contains a relatively high level of tyrosine (38), similar to that found in SBM, so the tyraminergic signaling pathway is almost certainly active in vivo. Moreover, because tyramine is a product of bacterial metabolism, it is likely to be present in the decomposing fruit on which the fly feeds. Food containing tyramine would therefore be expected to trigger a postfeeding diuretic response. In the locust, postfeeding diuresis is triggered by a CRF-like diuretic peptide hormone (3, 35); the response to tyramine in Drosophila would be different in that it would be mediated by a component of the food itself. Because the high rate of urine production by the tubules is thought to important for clearing toxins from the hemolymph (27, 37), it is tempting to speculate that the response to tyramine could have evolved to protect the fly from toxic substances that might be present in its food. Whatever the purpose, it is clear that tyramine represents an addition to the multiple signals already known to control renal function in insects.


I thank Dr. Gene Block for support, Dr. Jay Hirsh and the members of his lab for helpful discussions, fly lines, equipment, and reagents, Dr. James Burnette for the tyramine receptor primers, Shannon Cole for assistance with the TDC assay, and Drs. Oliver Schneider and Robert Kelly for assistance with the capillary electrophoresis.


  • This work was supported by the University of Virginia and National Institute of Diabetes and Digestive and Kidney Diseases R21-DK-060860 to E. M. Blumenthal.

  • Address for reprint requests and other correspondence: E. M. Blumenthal, Dept. of Biology, Gilmer Hall, Univ. of Virginia, P.O. Box 400328, Charlottesville, VA 22904-4328 (E-mail:eb5f{at}virginia.edu).

  • The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

  • First published November 20, 2002;10.1152/ajpcell.00359.2002


View Abstract