Received 26 December 2007; accepted after revision 14 April 2008. Correspondence: Patricia V. Pietrantonio, Associate Professor, Department of Entomology, Texas A&M University, 2475 TAMU, College Station, TX 77843-2475 USA. Tel.: +1 979 845 9728; fax: +1 979 845 6305; e-mail: firstname.lastname@example.org
In the mosquito Aedes aegypti (L.), the molecular endocrine mechanisms underlying rapid water elimination upon eclosion and blood feeding are not fully understood. The genome contains a single predicted diuretic hormone 44 (DH44) gene, but two DH44 receptor genes. The identity of the DH44 receptor(s) in the Malpighian tubule is unknown in any mosquito species. We show that VectorBase gene ID AAEL008292 encodes the DH44 receptor (GPRDIH1) most highly expressed in Malpighian tubules. Sequence analysis and transcript localization indicate that AaegGPRDIH1 is the co-orthologue of the Drosophila melanogaster DH44 receptor (CG12370-PA). Time-course quantitative PCR analysis of Malpighian tubule cDNA revealed AaegGPRDIH1 expression changes paralleling periods of excretion. This suggests that target tissue receptor biology is linked to the known periods of release of diuretic hormones from the nervous system pointing to a common up-stream regulatory mechanism.
The female of the mosquito Aedes aegypti (L.) is the vector of the arboviruses causative of yellow fever and dengue fever. The transmission of these human pathogens directly relates to the obligatory haematophagia of the female for reproductive success. Although similarities in M tubule signalling are evident between Drosophilidae (fruit flies) and Culicidae (mosquitoes), the aquatic habitat of mosquito larvae and pupae (Donini et al., 2006), and the haematophagia of the adult female impose unique physiological constraints for rapid elimination of water. Fluid excretion is elevated immediately upon adult emergence. The excretion rate then rapidly drops within minutes, followed by a gradual increase peaking approximately 11–17 h later (post-eclosion diuresis) before decreasing steadily to a basal level on the second day post-emergence (Gillett, 1983). The rate of fluid excretion also spikes immediately after consuming a sodium-rich blood meal (postprandial diuresis) (Stobbart, 1977; Williams et al., 1983), the volume of which may exceed the female's haemolymph volume by 10 times (Briegel, 1990; Beyenbach, 2003). These two periods of increased excretory water loss show physiological differences, both in duration and urine composition, indicating different mechanisms of hormonal control (Williams et al., 1983; Coast et al., 2005).
These hormones (and factors) activate intracellular signalling cascades by interacting with unique ‘G protein’-coupled receptors (GPCRs) in the basolateral membrane. For both mosquitoes and fruit flies, the M tubule kinin receptor has been characterized (Radford et al., 2002, 2004; Pietrantonio et al., 2005). The known insect diuretic hormone receptors, first cloned and described from Manduca sexta (Lepidoptera; Reagan, 1994) and Acheta domestica (Orthoptera; Reagan, 1996), now also include those from the silk moth Bombyx mori (Lepidoptera; Ha et al., 2000) and the planthopper Nilaparvata lugens (Hemiptera; Price et al., 2004). Genomic sequences with high identity to diuretic hormone receptors are predicted from the jewel wasp, Nasonia vitripennis (Hymenoptera), the honey bee, Apis mellifera (Hymenoptera; Weinstock et al., 2006) and the red flour beetle, Tribolium castaneum (Coleoptera; Hauser et al., 2008). Prediction of the genomic repertoire of GPCRs from the fruit fly, Drosophila melanogaster (Diptera; Brody & Cravchik, 2000; Hewes & Taghert, 2001; Hauser et al., 2006) and the malaria mosquito, Anopheles gambiae (Diptera; Hill et al., 2002), identified putative DH44 and DH31 receptor genes. Drosophila receptors for DH44 (Johnson et al., 2004) and DH31 (Johnson et al., 2005) have been cloned. In the yellow fever mosquito Ae. aegypti, neither DH44 nor DH31 receptor cDNAs have been cloned; however, in the first released annotation of the Ae. aegypti genome (http://www.vectorbase.org), incomplete sequences (AAEL008292, AAEL005894, AAEL008287) with high similarity to D. melanogaster DH44 receptor transcripts (CG8422 and CG12370) are present (Nene et al., 2007).
Confirming activity of predicted GPCRs potentially involved in diuresis and characterizing their temporal and spatial expression in target tissues is essential for a complete understanding of the regulatory mechanisms of water balance, especially for insects such as the female mosquito in which blood-feeding behaviour is fundamental to reproductive success.
Using molecular techniques previously described (Holmes et al., 2000; Pietrantonio et al., 2001, 2005), we cloned a M tubule cDNA (EU273351) encoding a putative DH44 receptor (AaegGPRDIH1) from Ae. aegypti. Additionally, a partial cDNA clone (EU273352) of a paralogue receptor (AaegGPRDIH2) was isolated from female head cDNA. Methods in silico were used to predict the sequences for the corresponding co-orthologue receptors from the non-annotated genome of the Southern house mosquito, Culex (pipiens) quinquefasciatus Say. Our results indicate that the AaegGPRDIH1 transcript is the most abundant DH44 receptor transcript in Ae. aegypti M tubules and quantitative PCR analysis (QPCR) revealed that the transcript's abundance fluctuates during periods that correlate with the physiological demand for urination, mirroring urine excretion rate patterns discussed above. Ours is the first study to show temporal variation of diuretic hormone receptor transcript abundance in a target tissue from an arthropod, specifically a blood-feeding insect.
The goals of our study were to determine which DH44 receptor participates in diuresis in female Ae. aegypti M tubules and to estimate relative transcript abundance during distinct physiological stages. M tubule cDNA was used as template for DNA amplification using degenerate and specific primers designed based on insect DH44 receptor sequences (Table 1). The 2715-bp cDNA sequence (EU273351) (Fig. 1) contained an open reading frame (ORF) encoding a 443 amino acid residue protein, hereafter referred to as AaegGPRDIH1, with a predicted molecular mass of 49.98 kDa. This sequence was confirmed using Ae. aegypti genomic sequences (AAGE02014128, AAGE02014129 and AAGE02014130). Domain searching (protein family database) identified the protein as a secretin-type (Family B) GPCR with seven transmembrane (TM) regions as predicted by transmembrane hidden Markov model (TMHMM) (Fig. 1). The predicted receptor has residues in the first extracellular domain consistent with Family B GPCRs, including six conserved cysteine residues (C44, C57, C66, C80, C99, C114), two tryptophan residues (W67, W105) and residues likely to be responsible for the formation of a salt bridge (D62, R97; Hoare, 2005). The receptor's N-terminus is not predicted to contain a signal peptide (data not shown). The first extracellular domain contains four predicted Asn-glycosylation sites (N-X-S/T) at residues N2, N37, N48 and N94. The ‘NetPhosK’ tool (http://www.cbs.dtu.dk/services/) predicted potential protein-kinase A phosphorylation sites at threonine residues T393, T419, T429 and T430, and possible protein-kinase C (PKC) phosphorylation sites at threonine and serine residues T393, S413, S425, T431, T432, T436 and T437. Specifically, residue S425 is absolutely conserved among insect DH44 receptor sequences identified to date (Figs 1, 2 and other insect receptors not shown) and may be sites for PKC phosphorylation and subsequent receptor desensitization (Hauger et al., 2003). The unique presence of 10 consecutive threonine residues (T429–T438) in the C-terminus was confirmed by analysis in silico of genomic DNA. Sixteen of the 31 terminal amino acids are either serine or threonine, providing ample potential locations for GPCR kinase phosphorylation and β-arrestin recruitment (Teli et al., 2005; Reiter & Lefkowitz, 2006; Oakley et al., 2007). The receptor possesses residues that within Family B GPCRs are responsible for G-protein interaction, both with the Gs- (residues T317, K318 and E389) and Gq- (residues V314, L315 and K318) mediated signalling pathways (Fig. 1) (Huang et al., 1996; Couvineau et al., 2003). Blast analysis comparing dipteran receptor sequences is summarized in Table 2.
Table 1. DNA primers used for AaegGPRDIH1 and -GPRDIH2 cloning and transcriptional analysis
GPRDIH1 and GPRDIH2 cloning
Tissue expression analysis of AaegGPRDIH1 and -GPRDIH2 transcripts
Amplicon design for quantitative PCR
Table 2. Blast search results for AaegGPRDIH1 (EU273351). Only dipteran database hits are shown
The AaegGPRDIH1 sequence used as a Blast query against Ae. aegypti whole genome sequences (WGS database, NCBI) identified several fragments of a potential paralogue gene. Specific primers were designed to clone a 2791-bp partial cDNA (EU273352) from female head cDNA encoding a DH44 receptor hereafter referred to as AaegGPRDIH2 (Fig. 2A). Blast analysis is summarized in Table 3. AaegGPRDIH1 and -GPRDIH2 sequences were utilized for prediction in silico of the respective co-orthologue-encoding sequences for CquiGPRDIH1 (BK006347) and CquiGPRDIH2 (BK006348) from the genome of the Southern house mosquito, Culex (pipiens) quinquefasciatus (Fig. 2A). However, the exon encoding approximately 26 amino acid residues corresponding to portions of TM region IV and extracellular loop III for CquiGPRDIH2 could not be predicted, probably because of current incomplete identification and assembly of the C. quinquefasciatus genomic sequences (see CquiGPRDIH2 sequence in Fig. 2A, the missing region is indicated by ‘X’ residues at positions 237 and 238). Coincidentally, the annotated An. gambiae GPRDIH2 (AGAP005465-PA) sequence is also incorrectly predicted in the same region. However, when we used the AaegGPRDIH2 sequence as a tBlastx query of the An. gambiae genome, a sequence was identified (LIFVGAWAIAKPFFGSVSNLEHPSKV) that has high similarity. We propose that this sequence complements the incomplete AgamGPRDIH2 annotation from residues 322 to 348 in Fig. 2A.
Table 3. Blast search results for AaegGPRDIH2 (EU273352). Only dipteran database hits are shown
Multiple ORF alignment of cloned and predicted DH44 receptor sequences from Diptera indicated structural conservation throughout the receptor sequence (Fig. 2A). Additionally, visual inspection of aligned sequences allowed the detection of conserved C-terminal motifs (Fig. 2B) differentiating the DIH1 and DIH2 receptors. While all the dipteran receptors share the motif identified as region I in Fig. 2B, which also contains the already alluded to conserved serine residue (S425 in the Ae. aegypti sequence), only the mosquito DIH2 receptors and Drosophila CG8422-PA share two other well-conserved motifs (regions II and III; Fig. 2B), probably relating to their as yet undiscovered intracellular protein−protein interactions and their potential to form similar signalling complexes or ‘receptosomes’ (Bockaert et al., 2004; Appert-Collin et al., 2006).
Bootstrap phylogenetic analysis (Fig. 3) of DH44 receptor sequences from Ae. aegypti, An. gambiae, C. quinquefasciatus and D. melanogaster indicated that the mosquito DIH1 receptors form a well supported cluster most similar to the Drosophila CG12370-PA, known to be enriched in M tubules (Wang et al., 2004; Chintapalli et al., 2007). The mosquito DIH2 receptors group with Drosophila CG8422-PA, which is apparently exclusively expressed in the nervous system (Johnson et al., 2004, 2005; Chintapalli et al., 2007) (Fig. 3).
Because of the differential tissue expression profile observed for CG12370 and CG8422 in Drosophila, we asked if AaegGPRDIH2 was expressed in the Ae. aegypti M tubule. Specific cDNA amplification reactions by RT-PCR were assembled for AaegGPRDIH1 and -GPRDIH2 transcripts using female head and M tubule cDNAs as template. Both transcripts are present in head tissues, but only the AaegGPRDIH1 is present in sufficient abundance to be amplified from M tubule cDNA and detected by gel electrophoresis analysis (Fig. 4A). Supporting QPCR analysis showed that AaegGPRDIH1 transcript is significantly more abundant than AaegGPRDIH2 transcript in M tubules, while the latter transcript is barely detected (Fig. 4B). Based on sequence analysis, tissue localization and expression abundance, we contend that AaegGPRDIH1 and the Drosophila CG12370 receptor are co-orthologues, as are AaegGPRDIH2 and the Drosophila CG8422 receptor.
QPCR analysis of relative AaegGPRDIH1 transcript levels measured from the time of adult eclosion throughout the previtellogenic (before blood feeding) and vitellogenic (after blood feeding) periods of Ae. aegypti females showed that transcript abundance does fluctuate in M tubules depending upon age and physiological condition (Fig. 5). Twenty-four hour post-eclosion transcript levels were at least double those observed immediately before blood feeding (48–72 h post-eclosion) and these differences were statistically significant. Blood feeding also significantly induced transcription of AaegGPRDIH1 back to post-eclosion levels or higher, but abundance significantly declined 6 h post-blood feeding. Unexpectedly, we observed the most abundant transcript levels 24 h after blood feeding, when relative abundance was about five times the level in 48–72 h previtellogenic females (Fig. 5).
Water excretion is tightly controlled to maintain homeostasis, both of water and ions (Pannabecker, 1995; O'Donnell & Spring, 2000; Beyenbach, 2003; Coast, 2007). In addition to the excess water evacuated immediately upon eclosion and within the first day of adult life (Gillett, 1982, 1983), female Ae. aegypti must consume a water- and sodium-rich bloodmeal to obtain the protein necessary for the production of eggs (Colless & Chellapah, 1960; Telang et al., 2006). After taking a bloodmeal, the female's body weight may increase between two and four times (Stobbart, 1977), although about 40% of the water ingested with the meal is eliminated in the first hour after feeding (Williams et al., 1983). This is a critical time for mosquito predation and evolution has favoured a fast elimination of excess water (and weight) for flight (Jackson et al., 2005; Nelson & Jackson, 2006). The high rates of water excretion coincidental with the high rates of primary urine formation almost immediately upon adult eclosion and post-blood feeding imply a highly coordinated and precise hormonal control of the signalling process and downstream effector proteins for ion and water transport. Despite these obvious and extreme physiological events, there has been no global analysis of gene transcripts levels in the M tubules of Ae. aegypti across life stages or physiological conditions, as has been carried out with other tissues (Sanders et al., 2003; Feitosa et al., 2006) or M tubules in nonblood-feeding dipterans (Wang et al., 2004). Molecular characterization of insect M tubules, especially those from small insects such as mosquitoes, is also hindered by the low amount of tissue recoverable from each individual. In Ae. aegypti the short length, 3 mm (Yu & Beyenbach, 2004), of each of the five functionally homogeneous M tubules (Beyenbach et al., 1993), and low cell count per tubule, 54 ± 2 total cells (Cabrero et al., 2004), require the dissection of many individuals for the detection of low abundance transcripts, such as GPCRs (Fredriksson & Schiöth, 2005).
In order to better understand the mechanisms by which mosquitoes regulate M tubule water transport, we have continued the molecular characterization of the receptors involved. We cloned a M tubule cDNA for AaegGPRDIH1 (EU273351) (Fig. 1), a predicted diuretic hormone 44 (DH44) receptor. A query of the genomic data available prior to the publication of the Ae. aegypti genome (Nene et al., 2007) allowed the identification of ORF fragments encoding a second DH44 receptor, AaegGPRDIH2, which was subsequently cloned from Ae. aegypti heads (this paper) (EU273352). The An. gambiae genome also contains two genes predicted to encode DH44 receptors (AgamGPRDIH1: AGAP005464-RA; AgamGPRDIH2: AGAP005465-RA) (Hill et al., 2002; Holt et al., 2002); however, neither has been cloned to confirm the sequence predictions.
Likewise, two putative DH44 receptor-encoding genes have been cloned from D. melanogaster: (1) the CG12370 transcript is present and enriched in M tubules of D. melanogaster (Wang et al., 2004; Chintapalli et al., 2007) but has not been functionally characterized; and (2) the CG8422 transcript is not enriched in the fruit-fly M tubule transcriptome relative to either other tissues or fly carcasses (Wang et al., 2004; Chintapalli et al., 2007), nor is the receptor protein detected by immunohistochemistry of M tubules (Johnson et al., 2005). CG8422 was cloned from head cDNA and functionally characterized by expression in mammalian cells (Johnson et al., 2004). In adult flies it is expressed primarily in nervous tissue (Chintapalli et al., 2007).
Insect and vertebrate Family B GPCRs share a complex regulation, both with respect to the G-proteins to which they couple, and their phosphorylation sites for internalization and degradation. Similarly, the receptors AaegGPRDIH1 and AaegGPRDIH2 contain conserved residues in their third intracellular loops and C-terminal ends which in the human CRF receptor interact with G-protein subunits controlling both the cAMP (Gs) and inositol phosphate/Ca++ (Gq) signalling pathways (Huang et al., 1996; Couvineau et al., 2003) (Fig. 1). Whilst studying the effect of CRF-like peptide from another mosquito (Culex salinarius) on M tubules of Ae. aegypti, Clark & Bradley (1998) theorized that the activation of different intracellular signalling pathways was responsible for the differential effects observed when using low (10−9 mol/l) and high (10−7 mol/l) peptide concentrations. They found that a low concentration reduced transepithelial voltage without increasing the short-circuit current resulting in a mild diuresis, whereas a high concentration significantly stimulated fluid and ion secretion and increased the short-circuit (intracellular) current (Clark & Bradley, 1998). Additionally, activation of D. melanogaster CG8422 expressed in human embryonic kidney (HEK)-293 cells elicited a cAMP response at nanomolar concentrations of DH44 (EC50 = 1.5 nM) and a calcium response at higher concentrations (EC50 = 300 nM) (Johnson et al., 2004), but receptor activation by DH44 in vivo on D. melanogaster M tubules showed no effect on intracellular calcium levels (Cabrero et al., 2002). Overexpression of the CG8422 receptor in HEK cells originates a skewed stoichiometry of the available G-proteins that could explain the observation of dual signalling (Maudsley et al., 2005). As the population of preferred Gs subunits available for coupling to the overexpressed DH44 receptor decreased, the likelihood of receptor coupling to the Gq subunits, of lower affinity, would increase. Based on our analysis of Ae. aegypti receptor sequence motifs (Fig. 1), the ability of the DH44 receptors to interact with either the Gs- or Gq-mediated signalling pathways appears to be ancestral, because sequence residues responsible for the interaction are also conserved in vertebrate CRF-receptors (Wietfeld et al., 2004) and within the broader Family B GPCR receptor family (Huang et al., 1996).
AaegGPRDIH1 also contains multiple serine and threonine residues on the third intracellular loop and C-terminus, providing a multiplicity of potential sites for GPCR kinase phosphorylation and β-arrestin recruitment (Teli et al., 2005; Reiter & Lefkowitz, 2006; Oakley et al., 2007), and a sequence motif [-S425-I426-R427-] implicated in PKC phosphorylation and desensitization (Hauger et al., 2003). These diverse potential phosphorylation sites appear partially conserved among dipteran DH44 receptors (Figs 1, 2) as well as among the greater CRF-receptor family (Teli et al., 2005; Oakley et al., 2007), and reflects the potential complexity in post-translational regulation in these insect receptors.
Unlike vertebrate CRF receptors (Hofmann et al., 2001; Perrin et al., 2001; Alken et al., 2005), there is no experimental evidence for the presence of N-terminal signal sequences for any insect DH44 receptor identified to date. Reagan claimed that the Ac. domestica DH44 receptor possessed a signal sequence (Reagan, 1996); however, no description of his prediction method was given. The ‘SignalP 3.0’ prediction tool (http://www.cbs.dtu.dk/services/) does not indicate either a signal sequence for the Ac. domestica receptor or for any insect DH44 receptor (data not shown); therefore, Reagan's claims are currently unverified. The lack of an N-terminal signal sequence indicates intracellular trafficking differences may exist between insects and vertebrates for the localization of these receptors to the cell membrane or, alternatively, that the current algorithms for prediction are biased for vertebrate signal peptide identification.
Sequence and phylogenetic analysis (Figs 2 and 3) of dipteran DH44 receptor transcripts clearly identify the mosquito DIH2 receptors as the co-orthologues of the Drosophila CG8422 receptor, consistent with the cloning from head cDNA of both AaegGPRDIH2 (this report) and D. melanogaster CG8422 (Johnson et al., 2005). In addition, the CG8422 transcript is apparently lacking in Drosophila M tubules (Wang et al., 2004; Chintapalli et al., 2007), while AaegGPRDIH2 transcript is minimally expressed in M tubules as detected by QPCR (Fig. 4B). We speculate that low level GPRDIH2 expression in Ae. aegypti M tubules may be localized in the tip cell, which is of neuronal origin in dipterans (Skaer, 1989) or perhaps the tracheolar cells (Pietrantonio et al., 2000), which as part of the respiratory system, interface with the M tubules. These analyses also identify the mosquito M tubule receptor GPRDIH1 as the co-orthologue of the Drosophila M tubule receptor encoded by CG12370. Both the AaegGPRDIH1 transcript (Fig. 4) and CG12370 (Chintapalli et al., 2007) are the most abundant DH44 receptor-encoding transcripts present in M tubules from their respective species, and ORF alignments reveal similarities in C-terminal protein motifs indicating conservation of intracellular interaction (Fig. 2B).
We hypothesized that the relative abundance of the AaegGPRDIH1 transcript in Ae. aegypti M tubules varied according to the age and physiological state of the female given that: (1) the AaegGPRDIH1 is indeed the most abundant receptor transcript in the Ae. aegypti M tubule; (2) the DH44 receptor contributes to the production of primary urine; and (3) the diuretic need of female mosquitoes is not constant (Cabrero et al., 2002; Beyenbach, 2003; Coast et al., 2005). Therefore, using QPCR we verified that the relative abundance of the AaegGPRDIH1 transcript in M tubules significantly increased paralleling known periods of high diuresis in the previtellogenic and vitellogenic stages (Fig. 5).
In newly emerged female Ae. aegypti, Gillett observed two periods of water excretion: immediately upon eclosion, and again 11–17 h after eclosion (Gillett, 1982, 1983). During this first day it is necessary for the female to eliminate excess water and/or nitrogenous wastes remaining from the larval stages (von Dungern & Briegel, 2001b), but because blood feeding has yet to occur, there is no physiological demand for differential cation loss to maintain homeostasis. This corresponds to the high AaegGPRDIH1 transcript abundance observed during the first day of adult life (Fig. 5).
Upon blood feeding, three different phases of urination are observed in female Ae. aegypti: peak phase (within 10 min of feeding), post-peak phase (10 min to 50 min) and late phase (50 to 120 min), with the most significant water loss during the immediate peak phase (Williams et al., 1983). Ion composition of the urine varies during these three postprandial phases with greatest natriuresis (Na+ elimination) during the peak phase, with potassium excretion delayed until the post-peak and late-peak phases. The pattern of water and ion loss indicates that multiple control mechanisms act simultaneously during the postprandial period (Williams et al., 1983). Using blood-fed females, Coast (2001b) continuously measured water loss from individual mosquitoes and observed a pattern of water excretion similar to Williams et al. (1983). From isolated M tubules, Coast et al. (2005) also observed that the calcitonin-like peptide (DH31) stimulated sodium transport preferentially over potassium ions in both An. gambiae and Ae. aegypti, whereas the CRF-like peptide (DH44) nonspecifically signalled transport of both cations. While haemolymph concentrations of these peptides have not been determined, the data from Williams et al. and Coast (2005) suggest that after a blood meal, DH31 is released immediately into circulation, stimulating the rapid loss of sodium ions and excess water. Somewhat delayed DH44 release after blood feeding maintains the elevated, but decreasing, water loss whilst conserving sodium ions. We observed that AaegGPRDIH1 transcript expression doubled after blood feeding compared to expression levels at 48–72 h in nonfed animals (Fig. 5). Although it remains to be determined if the highest DH44 receptor transcript pools measured are for receptor protein synthesis or alternatively reflect new transcript synthesis for replacement of the depleted transcript pool, we favour the first possibility because the high levels of receptor transcript after 1 h of blood feeding are consistent with a delayed release of DH44, further suggesting that the DH44 receptor is involved in the post-peak and late phase of diuresis. Interestingly, the highest transcript abundance was observed at 24 h post-feeding, a point when there is no observed increase in water loss from the mosquito. This perhaps reflects the need to transport large quantities of nitrogenous waste from the digestion of the bloodmeal into the M tubules (Briegel & Lea, 1975; Cole & Gillett, 1979; Briegel, 1980; Van Handel & Klowden, 1996; von Dungern & Briegel, 2001a) with the hindgut then acting to reabsorb and recycle the water. Female mosquitoes also begin eliminating the indigestible haematin portion of the blood meal 24 h post-feeding (Cole & Gillett, 1979) without additional water loss. The abundance of AaegGPRDIH1 transcript in M tubules 24 h after taking a bloodmeal suggests an important yet-uninvestigated role for DH44 signalling in the elimination of nitrogenous waste.
The release of diuretic hormones has been demonstrated immediately after eclosion in locusts (Audsley et al., 1997b) and during and after feeding, such as serotonin release in the bug Rhodnius (Lange et al., 1989), CRF-like diuretic hormone in the locust (Audsley et al., 1997a) and diuretic hormones in the mosquito Anopheles freeborni (Nijhout & Carrow, 1978). Ours is the first study to show temporal variation of a diuretic hormone receptor transcript in a target tissue from an arthropod, specifically a blood-feeding insect, providing evidence that receptor biology (transcriptional expression) in the M tubule may be linked to diuretic hormone levels during peaks of diuresis after eclosion and feeding.
Rules and conventions for genetic features, gene products and the role of species designations in gene names were as proposed in VectorBase for Family Culicidae (aaegypti.vectorbase.org/Docs; under Naming Conventions). Similar nomenclature is accepted for An. gambiae GPCRs (Hill et al., 2002) and used for Ae. aegypti genome supplementary information (Nene et al., 2007).
Mosquito rearing, dissection and cDNA synthesis
Aedes aegypti L. (Diptera: Culicidae) (Rockefeller strain) larvae were reared at 27 °C (Shapiro & Hagedorn, 1982). Larvae were fed high protein Purina ONE® Natural Blends dog food (Nestle Purina PetCare, St. Louis, MO, USA); density was kept low to maximize adult size. Adults were provided 10% sucrose solution for feeding ad libitum. M tubules from 50 previtellogenic 3–5-day-old females were dissected under phosphate-buffered saline (PBS) and placed into RNAlater® (Ambion, Austin, TX, USA) RNA stabilization solution. mRNA was isolated using the DynaBead® mRNA Direct kit (Invitrogen, Carlsbad, CA, USA) as per kit specifications. The final elution volume was 20 µl; other details as in Pietrantonio et al., 2005. cDNA concentration was considered as relative ‘tissue equivalents’ per unit volume for all experiments.
To ensure reverse transcription of full-length cDNA, RNA ligase-mediated (RLM) cDNA synthesis was with the GeneRacer™ Kit (Invitrogen) as per the manufacturer's protocol. cDNA was similarly synthesized using mRNA extracted (DynaBead® kit) from heads of 50 3–5-day-old, previtellogenic females, and M tubules from 24 h vitellogenic females. Females were blood-fed on an anesthetized mouse following NIH guidelines (Approved Animal Use Protocol TAMU 2007-141).
A pair of nested, degenerate sense primers (P133 and P134; Table 1) were designed for 3′RACE which corresponded to the TM region II (residues 169–179 in the final AaegGPRDIH1 sequence) and TM region V (residues 308–317), respectively (Fig. 1), of insect DH44 receptors (D. melanogaster: CG8422, CG12370; An. gambiae: AGAP005464-RA; AGAP005465-RA; Manduca sexta: U03489; Ac. domestica: U15959).
All rapid amplification of cDNA ends (RACE) reactions contained 1 µl Advantage® 2 Polymerase Mix (Clontech, Mountain View, CA, USA), enzyme buffer to 1× and 0.2 µM of each dNTPs. For 3′RACE, a 50 µl primary reaction was assembled as follows: to 2.5 µl (1 M tubule equivalent) of diluted GeneRacer™ M tubule cDNA was added 0.6 µM P133 and 0.3 µM GeneRacer™ 3′ adaptor primer. After initial heating at 94 °C for 3 min, amplification was for 45 cycles (94 °C 30 s, 65 °C 60 s, 72 °C 90 s) with final incubation at 72 °C for 5 min. A 50 µl secondary reaction was assembled by adding 1 µl primary reaction to 0.4 µM P134 and 0.2 µM GeneRacer™ 3′ nested adaptor primer. Amplification was as above except that the time for the extension step was 120 s. Products were visualized by agarose electrophoresis with GelStar™ dye (Lonza Group Ltd, Basel, Switzerland). A 1900-bp band was purified using the QIAquick® Gel Extraction Kit (Qiagen, Valencia, CA, USA), and cloned using the TOPO TA Cloning® Kit (Invitrogen). Plasmid DNA was prepared using the QIAprep® Spin Miniprep Kit (Qiagen). Plasmid inserts were confirmed by EcoRI restriction analysis, and sequenced as detailed below. Sequence similarity to known DH44 receptors was confirmed by Blast analyses. Specific antisense primers, P144 and P143, were designed based on this 3′RACE product sequence (Table 1). For 5′RACE, a primary asymmetric reaction was assembled using 2.5 µl of the same M tubule cDNA to which was added 0.4 µM Primer P144. Final volume was 49 µl. After an initial incubation at 94 °C for 3 min, asymmetric amplification proceeded for 10 cycles (94 °C 30 s, 64 °C 60 s, 72 °C 120 s) before the addition of 0.2 µM GeneRacer™ 5′ adaptor primer for 45 additional cycles, with final incubation at 72 °C for 5 min. A secondary nested reaction combined 1 µl primary reaction with 0.4 µM Primer P143 and 0.2 µM GeneRacer™ 5′ nested adaptor primer in 50 µl volume. Amplification was for 45 cycles (94 °C 30 s, 64 °C 60 s, 72 °C 120 s), prior to a final 72 °C extension step for 5 min. A 1900-bp product was purified and cloned as above. Sequencing analyses identified a large sequence overlap between the 3′RACE and 5′RACE clones, confirming that they were obtained from the same cDNA. The PCR amplification product corresponding to the full length ORF is shown in Fig. 4A.
A partial ORF for AaegGPRDIH2 was predicted based on the sequence present in the Ae. aegypti genome. Based on this, a specific sense primer, P197F (encompassing amino acid residues 69–82 in the final GPRDIH2 sequence), was designed (Table 1). All RACE reactions contained 1 µl Advantage® 2 Polymerase Mix (Clontech), enzyme buffer to 1× and 0.2 µM of each dNTPs. 3′RACE reactions were in 50 µl containing 1 µl (0.4 head equivalent) diluted GeneRacer™ head cDNA, 0.4 µM P197F and 0.4 µM GeneRacer® 3′ adaptor primer. Amplification proceeded for 40 cycles using ‘Touch Down’: 94 °C for 3 min initial denaturation, followed by 20 cycles of 94 °C for 30 s, 72 °C (decreasing 0.5 °C/cycle) for 60 s and 72 °C for 120 s; then following 20 cycles of 94 °C for 30 s, 65 °C for 60 s and 72 °C for 120 s. The reaction was then incubated at 72 °C for 5 min. Products were visualized by gel electrophoresis, cut from the gel, cleaned and cloned as above. Plasmid DNA was prepared, and inserts were confirmed by restriction analysis as above. The 2130-bp DNA 3′RACE product was sequenced from both directions as detailed below, with identity confirmed by comparison with the available genomic prediction.
For 5′RACE, specific antisense primers P197R (complementary to P197F) and P206 (encompassing the stop codon and a portion of the 3′ UTR) (Table 1) were designed and RACE carried out in a 50 µl amplification reaction containing 1 µl (0.4 head equivalent) diluted GeneRacer™ head cDNA to which was added 0.4 µM GeneRacer® 5′ adaptor primer and 0.4 µM P206 primer. After initial denaturation and enzyme activation at 94 °C for 3 min, amplification was for 40 cycles (94 °C 30 s, 62 °C 60 s, 72 °C 120 s) with a final extension step at 72 °C for 5 min. Gel electrophoresis analysis revealed no products of the expected size, so a secondary nested reaction was assembled using 1 µl of a 1 : 100 dilution of the previous reaction as template to which was added 0.4 µM GeneRacer® 5′ nested adaptor primer and 0.4 µM P197R primer. The amplification procedure was as for the primary reaction except that the annealing temperature was raised from 62 °C to 66 °C. Products were visualized, cut, gel extracted and cloned as above. Plasmid DNA was prepared and inserts were confirmed as above. The 5′RACE 909-bp DNA insert was sequenced as below and a region of identical DNA sequence overlap with the 3′-RACE clone confirmed this sequence as part of the same cDNA.
All sequencing reactions used ~400 ng DNA template and proceeded for 50 cycles using ABI Big Dye® (Applied Biosystems, Foster City, CA, USA) with other details as described (Pietrantonio et al., 2005). Sequence data were analysed using the DNASTAR software suite (DNASTAR Inc., Madison, WI, USA). The Blast search algorithms at NCBI (http://www.ncbi.nlm.nih.gov/) and VectorBase (http://www.vectorbase.com) were used to identify DH44 receptor-like sequences from other organisms. Protein domain searching was with PFAM (http://www.sanger.ac.uk/cgi-bin/Pfam) (Finn et al., 2006). Transmembrane region predictions were made using TMHMM (http://www.cbs.dtu.dk/services/). Potential glycosylation sites and kinase-specific phosphorylation sites located within the AaegGPRDIH1 sequence were predicted with ‘NetNGlyc’ and ‘NetPhosK’ (http://www.cbs.dtu.dk/services/), respectively. Potential N-terminal signal peptides were investigated using ‘SignalP 3.0’ (http://www.cbs.dtu.dk/services/) (Bendtsen et al., 2004). Sequences were aligned using Megalign (DNASTAR) and ClustalW (align.genome.jp/). Bootstrap analysis was performed using paup version 4.0b10 (Swofford, 2002) with 10 000 bootstrap replications and 50% majority-rule consensus. Alignment gaps were treated as missing data. Trees were rooted using the Caenorhabditis elegans C18B12.2 secretin-type receptor (Harmar, 2001). Sequence predictions for C. quinquefasciatus GPRDIH1 and GPRDIH2 were obtained from the WGS database (NCBI) and the CpipJ1.0_5 gene assembly (http://www.vectorbase.org) by using AaegGPRDIH1 and AaegGPRDIH2 as query sequences with the tBlastn algorithm to identify genomic DNA fragments with the highest similarities. Genomic DNA fragments encoding potential ORFs were imported into Mapdraw (DNASTAR) and translated in all reading frames. tBlastn results were manually mapped to the Mapdraw output, and potential exon−intron boundaries were confirmed using the splice site prediction tool of the Berkeley Drosophila Genome Project (http://www.fruitfly.org/seq_tools/splice.html). Identified ORF-containing exons were manually arranged using Editseq (DNASTAR) and translated in silico.
Tissue expression analysis of AaegGPRDIH1 and -GPRDIH2 transcripts
To determine the presence or absence of GPRDIH1 and GPRDIH2 transcripts in Ae. aegypti heads and M tubules, amplification reactions were assembled using tissue-specific cDNA templates and unique pairs of specific primers (Table 1). Reactions contained either 1 µl (0.4 head equivalent) diluted GeneRacer™ head cDNA prepared from 3–5-day-old, previtellogenic females or 5 µl (0.4 M tubule equivalent/µl) diluted GeneRacer™ M tubule cDNA prepared from 24 h vitellogenic females, to which was added 0.2 µM dNTPs (each), 0.6 µM sense primer, 0.6 µM antisense primer, 1× enzyme buffer and 1 µl Advantage® 2 Polymerase Mix (Clontech). Primer pairs corresponded to sequences encompassing the start (sense) or stop (antisense) codons of each transcript (Table 1). The amplification of another GPCR transcript (5HT7-like receptor) known to be expressed in heads and in trachelolar cells attached to M tubules (Pietrantonio et al., 2000; Lee & Pietrantonio, 2003) was used as a positive control. As the AaegGPRDIH2 sequence is incomplete on the 5′ end, P197F was used as the sense primer. After initial heating at 94 °C for 3 min, amplification was for 35 cycles of 94 °C 30 s, 62 °C 60 s, 72 °C 90 s, with a final extension at 72 °C for 5 min. Amplification products were visualized on a 1% agarose gel, 0.5× TBE, stained with GelStar™ dye (Lonza) and digitally documented using the Foto/Analyst® Investigator photographic system with Image J 1.34s software (Fotodyne, Hartland, WI, USA).
Amplicon design for quantitative PCR
Using Primer Express® software (Applied Biosystems), primers P180 and P181 (Table 1) were designed to amplify a 69-bp amplicon corresponding to the sequence between TM regions V and IV of AaegGPRDIH1. Primers P203 and P204 (Table 1) were designed to amplify a 67 bp region of the AaegGPRDIH2 transcript corresponding to the terminal intracellular domain. There was no sequence similarity between the amplicons. Optimal primer concentrations (900 nM each) were determined according to ABI directions (SYBR® Green PCR Master Mix protocol) using as template M tubule (GPRDIH1) or head cDNA (GPRDIH2). β-Actin gene expression was used to normalize all experimental results, with amplicon primers synthesized as reported (Zhu et al., 2003) and used at 900 nM each.
Time course analysis of AaegGPRDIH1 abundance
M tubules from 50 females (~250 M tubules; about 11 µg total RNA) for each of the following nine time points were dissected into RNAlater® (Ambion) stabilization solution and stored at 4 °C. Previtellogenic period: 0–12, 12–24, 24–48 and 48–72 h posteclosion. Vitellogenic period: 1, 3, 6, 12 and 24 h post-blood feeding. Dissections were repeated in triplicate for each time point. In order to avoid any potential variability from circadianicity, the tissues for each time point were dissected at various times of the day and in no particular order. Blood-fed females were allowed to feed for 30 min and to rest for 30 additional minutes prior to the start of dissections. Total tissues dissected were approximately 7000 M tubules from 1350 female mosquitoes.
For each time point, three independent cDNA synthesis reactions were from the three separate tissue isolations. For each synthesis, M tubules were pelleted by centrifugation, and mRNA was extracted (DynaBead® kit); beads were reconditioned for an additional extraction from the same homogenate resulting in a final eluted volume of 20 µl mRNA which was stored at –80 °C. Single-stranded cDNA was prepared for every isolated mRNA sample using the SuperScript™ III First-Strand Synthesis System (Invitrogen) using oligo(dT)20 as per the kit instructions.
Reactions using SYBR® Green PCR Master Mix (Applied Biosystems) were assembled for three replicate cDNAs for every time point as follows: to 60 µl SYBR® Green reagent was added 6 µl (~30 M tubule equivalents) cDNA and 10.8 µl water. This volume was equally divided (38.4 µl each) for estimation of the AaegGPRDIH1 and β-actin transcripts, respectively. To each aliquot was added either 10.8 µl each of AaegGPRDIH1 amplicon primers P180 and P181, or 10.8 µl each of β-actin amplicon primers for a final volume of 60 µl. On a 96-well MicroAmp® plate (Applied Biosystems), three wells (20 µl) were loaded for each template- primer combination. QPCR was performed for 45 cycles (95 °C 15 s, 60 °C 60 s) using the 7300 RT-PCR System (Applied Biosystems). Relative abundance for each transcript was calculated using the comparative CT (2−ΔΔCt) method and Sequence Detection Software v1.3.1 (Relative Quantification Study Application; Applied Biosystems) according to the manufacturer's directions, with statistical analysis using SPSS v12.0.1 for Windows (SPSS Inc, Chicago, IL, USA). The 48–72 h previtellogenic data point was used as the reference ratio (calibrator) for the time course analysis (Fig. 5).
For time course analysis, the null-hypothesis that there were no differences among sample means with respect to transcript relative abundance was rejected and differences among sample means were tested by one-way analysis of variance (anova) with Bonferroni correction (Fig. 5). Significance was established as P < 0.05. For QPCR supporting analysis of expression in specific tissues (Fig. 4B), relative transcript abundance was determined using respective amplicon primers for AaegGPRDIH1 and AaegGPRDIH2 (Table 1). Abundance of the respective transcripts in head cDNA, which was similar for both receptors, was set as the calibrator (not shown). Dissociation curve analysis indicated that the QPCR amplicons were specific for their respective cDNAs (data not shown). Graphs were prepared using GraphPad Prism v4.0 (GraphPad Software Inc., San Diego, CA, USA).
Funding for this research was partially from NIH-NIAID award 5 R01 AI 46447 to PVP and from Texas AgriLife Research (Formerly TAES). We thank Aleta Sosnik, Department of Nutrition and Food Science, TAMU, for the maintenance of the mice according to NIH guidelines.