INTRODUCTION

In insects, diuresis is the final product of fluid secretion by the malpighian tubules and water reabsorption in the hindgut (1). Both these processes appear to be under hormonal control because putative diuretic and anti-diuretic hormones have been characterized in insects (2-9). Two groups of putative diuretic hormones appear generally present in insects. These are the insect diuretic hormones related to corticotropin-releasing factor (CRF)¹ and the leucokinins. Two CRF-like diuretic hormones have been isolated from the tobacco horn worm moth Manduca sexta (3, 4). Related peptides have been isolated from crickets, locusts, cockroaches, flies, and a beetle (5-9) and are probably generally present in insects. The neuroendocrine cells producing diuretic hormone have been identified in a moth (10, 11) and a locust (12). In these species, diuretic hormone is produced in both median neurosecretory cells in the brain and lateral neurosecretory cells in the abdominal ganglia (10-12).

The leucokinins were initially isolated from the cockroach Leucophaea maderae (13-16), and related peptides have been identified from a cricket (17), a locust (18), mosquitoes (19, 20), and a moth (21). Although the leucokinins were first identified by their ability to stimulate hindgut contractions in L. maderae, these peptides also stimulate fluid secretion by the malpighian tubules in mosquitoes, crickets, locusts, and moths (21-24). Immunohistological studies have found leucokinin-immunoreactive neuroendocrine cells in the abdominal ganglia (11, 24-28). In at least two species, these are the same cells that produce the CRF-like diuretic hormones (11, 24). In some insects a group of median neurosecretory cells in the brain also contains leucokinin-like immunoreactivity (26, 28). So far only a single cDNA sequence encoding one of the CRF-like diuretic homones from M. sexta has been described (29), but no cDNA or genomic sequences are known for any of the leucokinins. Consequently, it is currently unknown whether the leucokinins are produced from different precursors, e.g. as the insect adipokinetic hormones (30), or from a single precursor like the Phe-Met-Arg-Phe-amides (31).

Diuresis in mosquitoes has been well studied, particularly in Aedes aegypti. This species is known to show an extensive diuresis after eclosion (32, 33), as well as after a blood meal (34). Three diuretic factors from the head have been characterized, two of which are able to stimulate fluid secretion in isolated malpighian tubules (35). Three leucokinins were also recently isolated from this species (19). Here, we report the structure of a cDNA encoding the Aedes leucokinins and show that in the mosquito the leucokinins are active on both the malpighian tubules and the hindgut of this species.

EXPERIMENTAL PROCEDURES

A. aegypti were reared as described (36).

Abdominal ganglia were dissected from 200 adult females, 0-4 days after emergence. Dissected ganglia were stored frozen at 70 °C until mRNA was extracted using a Micro-FastTrack^TM kit (Invitrogen, San Diego, CA). The mRNA was reverse transcribed using Superscript and cloned into -ZipLox (Life Technologies, Inc.). The phage was packaged using Gold packaging extract (Stratagene, La Jolla, CA) and amplified on Escherichia coli Y1090(ZL) (Life Technologies, Inc.).

The following oligonucleotides were synthetized by the Division of Biotechnology at The University of Arizona: 7575, 3-AAYAAYCCIAAYGTITTYTAYCCITGGGG-5, in which I stands for 2-deoxyinosine; 8242, 3-CCCCACGCATGGAATGGGTTCCG-5; 8365, 3-GCCTGGAATGTGTTCTTGG-5; 8430, 3-GCAACTCCAAGTACGTCTCCAAGC-5; 8432, 3-GTGTGTGCCGTGCATGAATGG-5; M13, 3-TGTAAAACGACGGCCAGT-3; T7, 5-TAATACGACTCACTATAGGG3; and M13 reverse, 5-AGGAAACAGCTATGACCATG-3.

A partial cDNA was isolated by using the polymerase chain reaction (PCR) with Taq polymerase (Boehringer Mannheim) and two primers, 7575 (based on the sequence of Aedes leucokinin 3) and M13 (based on the cloning vector). As a substrate for the PCR phenol choroform extracted DNA from the amplified library was used. Thirty cycles were programmed consisting of 2 min of denaturation at 94 °C, primer annealing at 44 °C for 2 min, slow rise to 72 °C over 2 min, and extension at 72 °C for 3 min. Full-length clones were isolated by plating the A. aegypti abdominal ganglia cDNA library on LB plates and lifting the plaques on nitrocellulose filters (Biotrace NT from Gelman Sciences, Ann Arbor, MI). The probe was a PCR product labeled with [-³²P]dATP using Klenow polymerase and primers 7575 and 8242. Positive clones were plaque purified and in vivo excised by infecting E. coli DH10B(ZIP) from Life Technologies, Inc. Sequencing reactions were performed with Taq polymerase and fluorescent dideoxynucleotides (Applied Biosystems), and the reaction products were electrophoresed and analyzed on an automated DNA sequencer (Applied Biosystems model 373) by the Division of Biotechnology of the University of Arizona. Sequence was obtained in both directions using primers 8242, 8365, 8430, 8432, T7, and M13.

Total RNA was isolated from whole adult mosquitoes that had been killed by freezing them rapidly at 70 °C. They were pulverized in liquid nitrogen using a mortar and pestle and subsequently rapidly extracted using the guanidinium thiocyanate method (37). After centrifugation through a CsCl₂ cushion, poly(A⁺) RNA was isolated from total RNA using an oligo(dT) spin column from New England Biolabs (Beverly, MA). RNA was electrophoresed in a 1.1% agarose gel containing formaldehyde and blotted onto Nytran membrane (Schleicher & Schuell). Hybridization was performed under high stringency conditions using the cDNA, cut from the plasmid with restriction enzyme Mlu1, and labeled with [-³²P]dATP using a random primer kit (Life Technologies, Inc.). RNA molecular size markers were from Life Technologies, Inc.

Synthesis, purification, and quantification of Aedes leucokinin 1 (Asn-Ser-Lys-Tyr-Val-Ser-Lys-Gln-Lys-Phe-Tyr-Ser-Trp-Gly-amide), Aedes leucokinin 2 (Asn-Pro-Phe-His-Ala-Trp-Gly-amide), and Aedes leucokinin 3 (Asn-Asn-Pro-Asn-Val-Phe-Tyr-Pro-Trp-Gly-amide) have been described (18).

The transepithelial voltage in the malpighian tubules was measured using the method of Burg et al. (38). Voltage was recorded in the lumen with respect to ground in the bath using Ag-AgCl electrodes. The recording electrode was placed in the inner perfusion pipette. No distal holding pipette was used, but because the electrical length constant of these tubules (about 300 µm) is considerably less than the length of the tubules used (more than 1.8 mm), this does not significantly affect voltage measurements. Peptides were applied by adding fixed concentrations of the peptides to the superfusing bath saline.

Female 3-7-day-old mosquitoes were cold-anesthetized and decapitated, and the digestive tract and adhering malpighian tubules dissected. The proximal ends of the malpighian tubules were severed from the pylorus, and the digestive tract was cut at the pyloric valve. The hindgut with the adhering malpighian tubules was transferred to a 20-µl drop of saline covered with water-saturated light white paraffin oil (Fisher). The proximal ends of the malpighian tubules were pulled into the paraffin oil, and the volume of the fluid secreted into the oil was estimated by measuring the long and short axes of the droplet with an ocular micrometer and by approximating the volume of the droplet as a prolate spheroid. Fluid secretion was measured during a 30-min control period, after which known concentrations of the Aedes leucokinins were added to the saline, and fluid secretion was measured over another 30 min. Results are expressed as an increase in fluid secretion. Each tubule was used for only one concentration.

Hindgut contractions were monitored using a modification of the assay system described before (39, 40). Briefly, the hindgut of an adult female mosquito is held between an immobile suction pipette holding the rectum and a mobile suction pipette attached just anteriorly to the pyloric valve. Movements of the gut cause movements of a small flag attached to the mobile pipette, which interrupts the light from a photoemitter before it reaches a photodetector. The signal was recorded digitally by computer. The tissue is kept in a 50-µl perfusion bath that is constantly perfused with saline at a rate of 100 µl/min with a peristaltic pump.

A. aegypti saline contained the following salts: 1.8 mM CaCl₂, 3.4 mM KCl, 150 mM NaCl, 0.6 mM MgCl₂, 1.8 mM NaHCO₃, 25 mM HEPES, and 5 mM glucose. The pH was adjusted to 7.2 with NaOH.

The Mann-Whitney U test was used for statistical analysis of fluid secretion rates.

RESULTS

Oligonucleotide 7575 end-labeled with [-³²P]ATP was initially intended to be used as a probe to screen the abdominal ganglia cDNA library. However, this primer and one based on the vector gave a distinct band in PCR and automated sequence analysis of the purified excised band yielded sequencing signal, which, although very poor, contained in it the sequence for Aedes leucokinin 2. A primer based on the obtained nucleotide sequence for Aedes leucokinin 2 (8242) was next used in PCR together with the M13 reverse primer. This PCR reaction yielded several products, most of which had a size close to 870 base pairs, of which 120 base pairs were vector sequence, whereas two minor products had sizes of approximately 300 and 400 base pairs. Automated sequence analysis of the different products revealed that all were transcribed from the same gene, with the smaller products being incompletely reverse transcribed cDNAs. The library was subsequently screened with a PCR product generated by primers 7575 and 8242 to obtain a full-length clone, which was excised in vivo into plasmid pZL1 and sequenced in both directions. The sequence of the longest cDNA obtained is shown in Fig. 1.

The first ATG is followed immediately by a stop codon, but the second ATG shows the beginning of a classical signal peptide, which is likely to be cleaved between residues 18 and 19 (41). The putative preproleucokinin also reveals the presence of one copy each of the Aedes leucokinins 1, 2, and 3. Classical Lys-Arg proteolytic processing sites delimit the three Aedes leucokinins, and all the Aedes leucokinins have a C-terminal Gly residue that can be processed into the C-terminal amides present in the mature peptides (42). Apart from the Aedes leucokinins, none of the possible products of the preproleucokinin have significant sequence similarity with any described protein or peptide, as sequence comparisons with the sequences available in the data base revealed.

Analysis of mRNA isolated from whole mosquitoes showed a single message of about 1300 base pairs (Fig. 2), thus suggesting that the isolated cDNA may be lacking about 200 bases at its 5-end.

All three Aedes leucokinins depolarize the malpighian tubules when added to the bathing saline. The concentrations of the peptides needed to obtain a depolarization in 50% of the malpighian tubules are 2.5 ± 1.2 × 10¹¹ M, 3.9 ± 1.7 × 10¹⁰ M, and 2.6 ± 1.4 × 10¹⁰ M for the Aedes leucokinins 1, 2, and 3, respectively. Transepithelial membrane voltages in unstimulated malpighian tubules are generally between 40 and 60 mV (lumen positive). Such voltages may be either stable or show spontaneous depolarizations. In the presence of low concentration of leucokinins, these depolarizations increase in frequency, and with increasing leucokinin concentrations the depolarization may become continuous (Fig. 3). Previous work has shown that cyclic AMP hyperpolarizes the malpighian tubules (43), but in the presence of depolarizations such as being induced by the leucokinins these hyperpolarizations become only visible after washout. Occasionally after washout of the leucokinins from the bath, the malpighian tubule would be temporarily hyperpolarized, e.g. as in Fig. 3 after washout of 10¹⁰ M Aedes leucokinin 3. Although small quantitative differences were noted in the potencies of the three Aedes leucokinins (see above), no consistent qualitative differences were found in the depolarizations induced by the Aedes leucokinins.

The Aedes leucokinins were also tested for their effects on fluid secretion by the malpighian tubules. Small increases in the rate of fluid secretion were observed for the Aedes leucokinins 1 and 3 but not for Aedes leucokinin 2 in concentrations of 10⁸, 10⁷, and 10⁶ M (Fig. 4).

When exposed to concentrations of 10⁹ or 10⁸ M of the three Aedes leucokinins, the hindgut contractions always increased in frequency. This effect was reversible, because on washout of the peptides the frequency of the contractions returned to the rate before addition of the peptides. In one out of four preparations tested, we were also able to see a significant effect with a concentration of 10¹⁰ M of Aedes leucokinin 2 (Fig. 5). The fragility of the mosquito hindgut and the large variability in the rate of contractions under control conditions prevented us from obtaining sufficient data for meaningful dose-response curves. Nevertheless, the data clearly demonstrated that each of the three Aedes leucokinins increases the frequency of contractions in this tissue in concentrations of 10⁸-10⁹ M (Fig. 5).

DISCUSSION

Peptide isolation efforts yielded only three Aedes leucokinins (19), whose structures are the same as those predicted by the cDNA isolated here. This strongly suggests that A. aegypti has only three leucokinins and not eight or five like the cockroach L. maderae (13-16) and the cricket Acheta domesticus (17), respectively. Three leucokinins were also isolated from the moth Helicoverpa zea (21). The leucokinin cDNA did not code for a CRF-like diuretic hormone, and there is not even a small sequence similarity with the M. sexta CRF-like diuretic hormone cDNA, suggesting that these two different insect peptide families are encoded by different genes.

It has been shown in vertebrates that the PC1/PC3 convertase, which normally cleaves the peptide preucursor at Lys-Arg dibasic sites, is also able to function as a mono-arginyl convertase, if the Arg residue is present in a favorable context, i.e. there is a basic amino acid four or six amino acid residues N-terminal from the cleavage site (47). One mono-arginyl processing site must be cleaved in the proleucokinin to obtain Aedes leucokinin 2. The presence of a Lys five amino acid residues N-terminal of this cleavage site can be expected to be only marginally effective (44, 45), but the presence of an Arg three residues more N-terminal is possibly sufficient to induce cleavage (45). An additional cleavage might occur at the Arg⁵¹-Tyr⁵²-Arg⁵³-Lys⁵⁴ sequence of the preproleucokinin (45).

The sequence of the preproleucokinin between amino acids 19 and 164 has no similarity with any known protein. Thus, the Aedes leucokinins may be the only biologically active peptides produced from this precursor. Nevertheless, it is interesting to note that four Cys residues (or five, depending on whether proteolytic cleavage occurs in the region Arg⁵¹-Tyr⁵²-Arg⁵³-Lys⁵⁴) are present in this part of the precursor. Cys residues in regulatory peptides can be histochemically identified by paraldehyde-fuchsin staining methods (46), and the leucokinin-immunoreactive neuroendocrine cells in the abdominal ganglia of hemimetabolous insect species are stained by paraldehyde-fuchsin (28, 47, 48). This suggests that these Cys residues are responsible for the observed paraldehyde fuchsin staining. Because the paraldehyde-fuchsin staining appears to be conserved between different species, these Cys residues and this part of the precursor are likely to be conserved also. It will thus be of interest to determine the structure of other leucokinin precursors.

Several neuropeptides are produced as multiple copies on a single precursor in both vertebrates and invertebrates. Sometimes a single peptide is present in exactly the same sequence in several copies; however in other cases single copies of structurally related peptides are found, as here for the Aedes leucokinins. It is clear from the work on opioid peptides in vertebrates that different peptides from the same precursor interact preferentially with the various receptors, and a wide variety of different opioid receptors has been found in vertebrates (49). Although there appear to be only three Aedes leucokinins, these are structurally sufficiently diverse to raise the question of whether or not they might have different receptors.

The Aedes leucokinins have different threshold concentrations when assayed for depolarizing activity on the malpighian tubules; they also have different potencies on fluid secretion. However, whereas Aedes leucokinin 1 is the most potent depolarizer, Aedes leucokinin 3 is the most potent inducer of fluid secretion, and Aedes leucokinin 2 was without significant effects on fluid secretion. As suggested by an anonymous reviewer, this might be due to the rapid inactivation of Aedes leucokinin 2 during the fluid secretion assay. We therefore performed a fluid secretion assay containing 10⁶ M Aedes leucokinin 2 and tested the bathing saline after the assay for effects on the transepithelial voltage. The results showed no measurable decrease in activity in the transepithelial assay, even when exposure to the malpighian tubules was increased to 1 h, and hence the absence of significant effects of Aedes leucokinin 2 on fluid secretion is not due to the rapid inactivation of this peptide during the fluid secretion assay.

It has been shown previously that strong depolarizing activity is not necessarily correlated with strong effects on fluid secretion. Thus diuretic factor 1 causes the malpighian tubule to depolarize but has little effect on fluid secretion, whereas diuretic factor 3 has only limited depolarizing activity but has strong effects on fluid secretion. The absence of strong effects on fluid secretion suggests that one or more of the Aedes leucokinins may represent the diuretic activity previously described as factor 1 (35). It has been shown elsewhere that the leucokinins regulate the chloride conductance of the malpighian tubule (50, 52). We have shown, that although Aedes leucokinin 2 depolarizes the malpighian tubule and affects the chloride conductance² in the same fashion as the other leucokinins, it does not appear to stimulate fluid secretion by the malpighian tubules. This suggests that Aedes leucokinin 2 may be specifically regulating chloride conductance, and it indicates that increasing chloride conductance in the malpighian tubules by itself may be insufficient to lead to an increase in fluid secretion. Furthermore, these results may also suggest the existence of different receptors for the leucokinins in the malpighian tubules of A. aegypti.

Drosophila melanogaster has malpighian tubules that are morphologically similar to those in A. aegypti, and it is likely that the regulatory mechanisms of fluid secretion by the malpighian tubules in these two Dipteran species are also similar (51). It has recently been shown for Drosophila that the stimulation of fluid secretion by the leucokinins can be augmented by cAMP and cGMP but not by thapsigargin, which induces release of intracellular calcium, whereas thapsigargin is able to augment fluid secretion induced by either cAMP or cGMP. This indicates that the stimulation of fluid secretion by the leucokinins is associated with an increase in intracellular calcium (52). Analysis of the secondary messengers induced by the Aedes leucokinins may clarify whether the three Aedes leucokinins activate more than one receptor.

The maximal effects on fluid secretion of the Aedes leucokinins are small compared with the effects of crude mosquito head extracts, which are able to induce much larger increases in fluid secretion (35, 53), whereas significant increases in fluid secretion happen only at relatively high leucokinin concentrations. The Aedes leucokinins are therefore unlikely to be the major diuretic hormones regulating diuresis after a blood meal, when urine flow is as high as 40 nl/min (34), which is equivalent to a fluid secretion rate of 8 nl/min per malpighian tubule. Thus the leucokinins are probably mere modulators of diuresis, as is 5-hydroxytryptamine, which is able to stimulate fluid secretion by the malpighian tubules only at unphysiologically high concentrations (53). It is interesting to note that in the mosquito the leucokinins are able to stimulate both fluid secretion by the malpighian tubules and hindgut contractions, and this appears to be the first instance in which both effects are known for a single species. This is noteworthy, because it has been reported for the locust that locustakinin (the sole identified locust leucokinin homolog) does not stimulate hindgut contractions (18), although it does stimulate fluid secretion by the malpighian tubules (24).