A de novo transcriptome of the Malpighian tubules in non-blood-fed and blood-fed Asian tiger mosquitoes Aedes albopictus: insights into diuresis, detoxification, and blood meal processing

Blood feeding and isolation of Malpighian tubules for cDNA library preparation

For a given experimental trial, 90 adult female mosquitoes were transferred from the main colony to two small containers (45 females per container) and starved for 24 h by removing their sucrose solution. After 24 h, one container of mosquitoes was provided with access to 10% sucrose for 30 min (non-blood-fed treatment, NBF) and the other was provided with access to heparinized rabbit blood (Hemostat Laboratories, Dixon, CA, USA) through a membrane feeder (Hemotek, Blackburn, UK) for 30 min. After 3 h, 12 h, or 24 h, mosquitoes were immobilized on ice for ∼15 min and the Malpighian tubules were isolated with fine forceps under Ringer solution (150 mM NaCl, 3.4 mM KCl, 1.7 mM CaCl₂, 1.8 mM NaHCO₃, 1.0 mM MgCl₂, 5 mM glucose, and 25 mM HEPES; pH 7.1). The tubules were transferred immediately to ice-cold TRIzol Reagent (Life Technologies, Carlsbad, CA, USA) after isolation. Each replicate consisted of ∼200 Malpighian tubules isolated from 40 adult females. Twelve replicates were collected for NBF mosquitoes and 3 replicates each were collected for BF mosquitoes at 3 h, 12 h, and 24 h post-blood meal.

RNA isolation and cDNA synthesis

Immediately after a replicate of tubules was isolated, total RNA was extracted, cleaned, and purified as previously described (Esquivel, Cassone & Piermarini, 2014). The quality and concentration of the RNA were determined with an Experion Automated Electrophoresis System (Bio-Rad, Hercules, CA, USA) and a Qubit 2.0 Fluorometer (Life Technologies), respectively. The TruSeq DNA Sample Prep Kit V2, Set A and B (Illumina, San Diego, CA) was used for the synthesis of double-stranded cDNA libraries as previously described (Esquivel, Cassone & Piermarini, 2014). These libraries were non-strand specific, barcoded with unique indexed adapters, and consisted of ∼270 bp fragments. The quality and concentration of the double-stranded cDNA was determined with an Agilent 2100 Bioanalyzer High Sensitivity DNA Chip (Agilent Technologies, Santa Clara, CA, USA) and Qubit 2.0 Fluorometer (Life Technologies), respectively.

Sequencing of cDNA libraries

The 21 cDNA libraries were pooled to form a multiplexed library (18 nM each). This library was diluted to 36 fM and sequenced (paired-end reads of 100 bp in length) on a single lane of a flow cell using the Illumina HiSeq 2000 platform at The Ohio State University Comprehensive Cancer Center (Columbus, OH, USA). CASAVA (version 1.8.2) was used to demultiplex the sequencing data, and ‘basecall’ files were used to generate FASTQ files. The paired-end reads can be retrieved at the SRA (Sequence Read Archive) of the NCBI (National Center for Biotechnology Information) under the BioProject accession number PRJNA279095(experiment SRX964103).

The reads were trimmed, filtered, assembled, and annotated as described in the Supplemental Information 1, resulting in a total of 8,232 non-redundant contigs or transcripts (hereafter referred to as ‘Ae. albopictus transcripts’), which comprise the adult female Ae. albopictus Malpighian tubule transcriptome. This de novo transcriptome was deposited into the NCBI Transcriptome Shotgun Assembly (TSA) database under BioSample number SRS883933and accession number GCZT00000000.

Differential expression analysis

The transcripts from each of the 21 paired-end cDNA libraries were aligned onto the annotated Ae. albopictus Malpighian tubule transcriptome using the ‘Burrow-Wheeler Aligner’ version 1.2.3 (Li & Durbin, 2009). The SAM files generated were submitted to ‘SAM/BAM To Counts’ version 1.0.0 in the MCIC-Galaxy pipeline, for counting the reads aligned onto the Ae. albopictus transcripts. The ‘SAM/BAM To Counts’ output was submitted to the data package ‘DESeq2’ version 1.4.5 (Love, Huber & Anders, 2014), imported from Bioconductor (Gentleman et al., 2004) and customized onto the R programming language. The number of reads from the NBF libraries were pooled together and used as a universal control for comparison with each of the BF libraries. The relative expression changes of transcripts between the NBF control and a BF library were considered statistically significant if the FDR-adjusted P-value was less than 0.05.

Functional clustering analysis of transcripts

Transcripts from the NBF libraries with reads per kilobase per million mapped reads (RPKM) values >0 were subjected to a Database for Annotation, Visualization and Integrated Discovery (DAVID, version 6.7) functional clustering analysis (Huang da, Sherman & Lempicki, 2009) as described in the Supplemental Information 1. Transcripts that were differentially expressed between the BF and NBF libraries were also subjected to a DAVID analysis. Only the functional clusters with enrichment values >1.3 (equivalent to a P-value < 0.05) were retained (Esquivel, Cassone & Piermarini, 2014).

In vivo diuresis assay

The diuretic capacities of NBF and BF adult female Ae. albopictus mosquitoes were determined using a previously described assay (Raphemot et al., 2013). In brief, NBF and BF mosquitoes were prepared as described in “Blood feeding and isolation of Malpighian tubules for cDNA library preparation.” At 3 h, 12 h, and 24 h after feeding, five mosquitoes were immobilized on ice and their hemolymph was injected with 900 nl (100 nl/sec) of a phosphate-buffered saline (PBS) using a Nanoject II nanoliter injector (Drummond Scientific, Broomall, PA, USA). The PBS consisted of 137 mM NaCl, 2.7 mM KCl, 10 mM Na₂HPO₄, and 2 mM KH₂PO₄ (pH 7.5) (Fisher Scientific, Hampton, NH, USA). Immediately after injection, the 5 mosquitoes were transferred to a graduated packed-cell volume (PCV) tube (MidSci, St. Louis, MO, USA) and placed into an incubator at 28 °C. After 30 min, the mosquitoes were removed and the PCV tube was centrifuged to collect the urine in the graduated column at the bottom of the tube. The amount of urine excreted was measured visually from the gradations. At least 9 replicates of 5 mosquitoes each were performed for each treatment. The baseline excretion volumes of uninjected NBF and BF mosquitoes for each time point were also determined and subtracted from those of the corresponding injected mosquitoes.

Glutathione S-transferase (GST) activity

GST activity in Malpighian tubules was measured by adopting elements of standard biochemical assays (Brogdon & Barber, 1990; Habig, Pabst & Jakoby, 1974; Hemingway, 1998). This assay is based on the GST-mediated conjugation of reduced glutathione (GSH) to 1-chloro-2,4-dinitrobenzene (CDNB) to generate 3-(2-chloro-4-nitrophenil)-glutathione; CDNB absorbs light at 340 nm. The amount of CDNB (nmol) conjugated is calculated based on (1) the rate of the absorbance increase at 340 nm, and (2) the extinction coefficient of CDNB (0.00503 µM⁻¹) (Habig, Pabst & Jakoby, 1974).

In the present study, NBF and BF mosquitoes were prepared and ∼100 Malpighian tubules were isolated from 20 females, as described in “Blood feeding and isolation of Malpighian tubules for cDNA library preparation.” The Malpighian tubules were stored at −80 °C until the day of the assay. At least nine biological replicates were collected for each treatment/time point. Immediately before the assay, the Malpighian tubules were thawed on ice and homogenized with a plastic pestle in 100 µl of a potassium phosphate buffer, which consisted of 100 mM potassium phosphate (pH 7.0), and 0.1% Triton X-100 (Fisher Scientific, Hampton, NH, USA). The crude tubule lysates were centrifuged at 10,000 g for 15 min at 4 °C. The supernatant was transferred to a clean 1.5 ml microcentrifuge tube on ice.

For the analysis of a single sample, the following were combined in the well of a UV transparent 96-well microtiter plate (Thermo Fisher Scientific, Waltham, MA, USA): 4 µl of the tubule supernatant, 186 µl of 200 mM GSH (Sigma-Aldrich, St. Louis, MO, USA), and 10 µl of 20 mM CDNB (Acros Organics, Geel, Belgium). GSH and CDNB were diluted in 17 MΩ-water and 100% ethanol, respectively. The GSH was dissolved in PBS. On a given microplate, each sample was represented in triplicate. Immediately after adding the CDNB, the microtiter plate was placed into a Multiskan GO Microplate Spectrophotometer (Thermo Scientific, Waltham, MA, USA) and submitted to a ‘slow’ mixing protocol for 10 s at 25 °C. Thereafter, the absorbance at 340 nm was measured every minute for six minutes. The calculated GST activity (nmol CDNB/min) of each sample was normalized to its total protein content, which was measured with a Bradford protein assay (Bio-Rad, Hercules, CA, USA).

Uric acid assay

The Eton^® Uric Assay Kit (Eton Bioscience, Inc., San Diego, CA, USA) was used to measure the concentration of uric acid in isolated Malpighian tubules of NBF and BF mosquitoes. Malpighian tubules were isolated from NBF and BF mosquitoes as described above. Each replicate consisted of ∼100 isolated Malpighian tubules from 20 females. In brief, isolated tubules were homogenized in 300 µl of dH₂O and 50 µl of the homogenate was analyzed for uric acid in UV transparent 96-well plates (Thermo Fisher Scientific) following the manufacturer’s protocol. On a given microplate, each sample was represented in duplicate. For each sample, the uric acid content (mmol uric acid) was normalized to its total protein content, which was measured with a Bradford protein assay (Bio-Rad).

Statistical analyses for functional assays

For the diuresis, GST, and uric acid assays, the data from the NBF and BF treatments were analyzed via a one-way ANOVA with a Newman-Keuls posttest using GraphPad Prism 6.0 (San Diego, CA).

Na⁺∕Ca²⁺ exchangers

To our surprise, the most abundant transcript encoding an ion transporter in the Malpighian tubules of NBF Ae. albopictus was that for a putative K-dependent Na⁺∕Ca²⁺ exchanger (NCKX) of the SLC24 superfamily (Table 4). Notably, NCKX1 was the fifth highest-expressed transcript in the tubules of NBF mosquitoes and over 3 times more abundant than the most-highly expressed subunit of the V-type H⁺-ATPase (i.e., ‘Locus135v1rpkm1149.21,’ subunit G, in Table S8). In addition, the transcripts for another NCKX (NCKX2) and two putative Na⁺∕Ca²⁺ exchangers (NCXs) of the SLC8 superfamily were expressed in tubules of NBF Ae. albopictus, but were of much lower abundance compared to NCKX1 (Table 4). In the Malpighian tubules of adult female An. gambiae, the NCKX1 ortholog (AGAP010975) is highly expressed and enriched (http://Moztubules.org; http://mozatlas.gen.cam.ac.uk).

To date, a NCKX or NCX is not included in the current model of transepithelial fluid secretion in Malpighian tubules of mosquitoes (Fig. S5). In vertebrates, NCKXs and NCXs play critical roles in regulating concentrations of intracellular Ca²⁺ by exporting Ca²⁺ from the cytosol (Khananshvili, 2013; Schnetkamp, 2013). In mosquito Malpighian tubules, kinin diuretic peptides utilize intracellular Ca²⁺ as a second messenger (Pietrantonio et al., 2005; Yu & Beyenbach, 2002). Thus NCKXs and NCXs could potentially play an important role in (1) maintaining low intracellular Ca²⁺ prior to kinin activation and/or (2) returning intracellular Ca²⁺ to normal levels after kinin stimulation. It is also possible that these transporters contribute to the electrogenic entry of Na⁺ into the epithelium, assuming that they localize to the basolateral membrane. Elucidating the localization and functional role(s) of NCKXs and NCXs in mosquito Malpighian tubules should be a priority of future studies.

GPCRs

At least three neuroendocrine peptides regulate the rates of transepithelial fluid secretion in mosquito Malpighian tubules: diuretic hormone 31 (DH₃₁), DH₄₄, nin (Beyenbach & Piermarini, 2009; Coast, 2007; Coast, 2009). DH₃₁ is also known as the mosquito natriuretic factor/peptide, which was first identified and characterized in the Malpighian tubules of Ae. aegypti by the Beyenbach laboratory (Beyenbach & Petzel, 1987; Petzel, Hagedorn & Beyenbach, 1985; Petzel, Hagedorn & Beyenbach, 1986; Sawyer & Beyenbach, 1985).

In the Malpighian tubules of NBF Ae. albopictus, we detected the expression of transcripts encoding putative DH₃₁ and kinin GPCRs, but not one encoding a putative DH₄₄ receptor (Table 5). In Malpighian tubules of Ae. aegypti and An. gambiae, transcripts for DH₄₄ receptors are expressed (Jagge & Pietrantonio, 2008; Overend et al., 2015), which suggests that the endocrine regulation of fluid secretion in Ae. albopictus may differ from other mosquitoes. The most abundant GPCR in the tubules of NBF Ae. albopictus was a putative tachykinin receptor (Table 5). To our knowledge, the effects of tachykinin on Malpighian tubule function in mosquitoes are unknown, but in Malpighian tubules of locusts (Locusta migratoria, Schistocerca gregaria) and a lepidopteran (Manduca sexta), tachykinin promotes diuresis (Johard et al., 2003; Skaer et al., 2002).

We detected the expression of at least 16 other transcripts encoding putative GPCRs in the Malpighian tubules of NBF Ae. albopictus (Table 5). Some encode receptors for unknown ligands (i.e., orphaned receptors) or ligands not previously studied in mosquito Malpighian tubules, such as allatostatin. Notably, five transcripts encode putative Methusaleh GPCRs (Table 5); these receptors are linked to reproduction, aging, and stress resistance in D. melanogaster (Araújo et al., 2013; Caers et al., 2012). Moreover, two transcripts encode putative frizzled receptors (Table 5), which play key roles in Wnt signaling and tissue development/morphogenesis in D. melanogaster (Hanlon & Andrew, 2015). Thus, GPCRs are likely involved in regulating a wide range of physiological processes in mosquito Malpighian tubules, beyond their conventional roles in salt and water balance.

Septate junctional proteins

Septate junctions are hypothesized to play roles in the paracellular transport of Cl⁻ and H₂O in mosquito Malpighian tubules (Fig. S5) and likely play important epithelial barrier functions. In insects, there are two types of septate junctions. Pleated septate junctions (pSJs) are typically found in tissues of ectodermal origin (e.g., epidermis, hindgut), whereas smooth septate junctions (sSJs) are typically found in tissues of endodermal origin (e.g., midgut). Malpighian tubules are exceptional in that they are derived from ectoderm, but contain sSJs (Izumi & Furuse, 2014).

In D. melanogaster, several transmembrane proteins have been shown to play key roles in the formation of pSJs, including claudin-like proteins (e.g., megatrachea, sinuous, and kune-kune), fasciclins (i.e., fas2 and fas3), gliotactin, lachesin, macroglobulin complement-related protein, melanotransferrin, neurexin IV, neuroglian, and subunits of the α and β subunits of the Na,K-ATPase (Izumi & Furuse, 2014). One fasciclin (fas3) is also a component of sSJs in the midgut of D. melanogaster (Izumi, Yanagihashi & Furuse, 2012). The only transmembrane proteins known to be specifically-associated with sSJs in D. melanogaster are snakeskin and mesh, which are expressed in the midgut and Malpighian tubules (Izumi & Furuse, 2014; Yanagihashi et al., 2012).

As shown in Table 6, snakeskin and mesh were among the most abundantly-expressed transcripts encoding putative transmembrane sSJ proteins in Malpighian tubules of NBF Ae. albopictus. Moreover, several transcripts encoding fasciclins were abundantly expressed, suggesting a potential role of these proteins in the formation of sSJs in mosquito Malpighian tubules. Transcripts encoding claudin-like proteins and other key elements of pSJs (e.g., gliotactin, neurexin IV) were of lower abundance. Although transcripts encoding α and β subunits of the Na,K-ATPase were also abundant (Table S8), evidence for a structural role of these subunits in septate junctions of Malpighian tubules is lacking. The α subunit localizes to the basolateral membrane of stellate cells and ouabain inhibits transepithelial fluid secretion and ATPase activity in isolated Malpighian tubules (Hine et al., 2014; Patrick et al., 2006; Tiburcy, Beyenbach & Wieczorek, 2013), which suggests that the Na,K-ATPase primarily participates in membrane transport.

CYP450s

At each time point after a blood meal, several CYP450 transcripts from each clade were differentially expressed, with the exception of the CYP2 clade (Fig. 4). A diverse pattern of differential expression occurred in the mitochondrial CYP450 clade. At 3 and 12 h, there were similar numbers of up and downregulated transcripts (3 up and 3 down at 3 h; 2 up and 2 down at 12 h), whereas at 24 h most were upregulated (4 up and 1 down) (Fig. 4, Table S11). Interestingly, all of the differentially-expressed mitochondrial CYP450 transcripts were transiently up or downregulated, and in some cases showed an upregulation followed by a downregulation or vice versa. In the CYP2 clade only one member was differentially-expressed; it was upregulated at 3 h (Fig. 4, Table S11).

The functional implications of such changes in transcript expression are uncertain. However, the general trends we found in the differential expression of CYP302A1, CYP314A1, and CYP15A1 (Table S11) were very similar to those observed in the ovaries of Ae. aegypti where their differential expression correlates with peak periods of ovarian ecdysteroid production (Sieglaff, Duncan & Brown, 2005). Thus, a putative role of the Malpighian tubules in mosquito ecdysteroid synthesis after a blood meal should be further investigated. Consistent with this notion, ecdysteroid-synthesis was detected in the abdomen of adult female Ae. aegypti after the ovaries were removed (Sieglaff, Duncan & Brown, 2005).

The CYP3 clade also showed a dynamic pattern of differential expression after a blood meal (Fig. 4). At 3 h and 24 h, a similar number of CYP3 transcripts were up- and downregulated (13 up and 12 down at 3 h; 11 up and 9 down at 24 h), whereas at 12 h, most CYP3 transcripts were downregulated (5 up and 14 down) (Fig. 4, Table S11). Only 2 CYP3 transcripts were upregulated at all 3 time points, whereas 7 CYP3 transcripts were downregulated at all 3 time points; moreover, several CYP3 transcripts were transiently up- or downregulated (Fig. 4, Table S11). Thus, the response of CYP3 transcripts to blood-feeding is complex and offers no universal trend. In contrast, transcripts in the CYP4 clade showed a concerted pattern of upregulation. A total of 6, 4, and 5 transcripts were upregulated at 3 h, 12 h, and 24 h, while none were downregulated; two of the CYP4 transcripts were upregulated at all 3 time points (Fig. 4, Table S11). These data suggest that members of the CYP4 clade may play important roles in the detoxification of blood meal metabolites.

Interestingly, several of the CYP3 and CYP4 transcripts that exhibited an upregulation at one or more time points after a blood meal in the present study have been implicated in metabolic resistance to insecticides in Ae. aegypti (Bariami et al., 2012; David et al., 2014; Dusfour et al., 2015; Reid et al., 2014; Vontas et al., 2012); i.e., CYP9J26, CYP9J28, CYP6BB2, CYP6M9, CYP6N9, CYP6N12, CYP6Z8, CYP6Z9, CYP9J9 v1, CYP9J10 v2, CYP9J19 v2, CYP4D24, CYP4H29, CYP4H30, CYP4H33 (Table S11). Also, CYP9M9, which is constitutively upregulated in Malpighian tubules after blood feeding (Table S11), has been shown to be upregulated in larvae of Ae. aegypti and Ae. albopictus reared in water containing toxic leaf litter (Kim & Muturi, 2012). Thus, the transcriptional regulation of the mosquito CYP3 and CYP4 members appears to respond to a broad variety of toxic insults from artificial or natural xenobiotics, including vertebrate blood.

GSTs

The cytosolic and microsomal GSTs showed a concerted upregulation after blood feeding (Table S11). At least 18 GST transcripts were upregulated at each time point, while only 2–3 transcripts were downregulated (Table S11). Moreover, 14 transcripts were upregulated at all 3 time points, in contrast to only one (i.e., GSTz1) that was constitutively downregulated (Table S11). Thus, the Malpighian tubules appear to increase their molecular capacity for GST-mediated detoxification after a blood meal, which is consistent with results of our earlier study (Esquivel, Cassone & Piermarini, 2014). Moreover, several GST transcripts that exhibited an upregulation at one or more time points after a blood meal have been implicated in metabolic resistance to insecticides in Ae. aegypti (Bariami et al., 2012; Dusfour et al., 2015; Vontas et al., 2012); i.e., GSTe2, GSTe4, GSTe7, GSTi1, GSTo1, and GSTx2 (Table S11). Thus, the transcriptional regulation of GSTs also appears to respond to a wide variety of xenobiotic stresses.

ABC transporters

In general, transcripts encoding ABC transporters were upregulated after blood feeding (Table S11). At 3 h and 24 h, most transcripts were upregulated (15 up, 8 down at 3 h; 11 up, 4 down at 24 h) (Table S11). At 12 h, similar numbers of transcripts were up- and downregulated (6 up, 7 down) (Table S11). Three transcripts were upregulated at all three time points, while two transcripts were downregulated at all three time points (Table S11). Thus, consistent with our previous study, the Malpighian tubules appear to increase their capacity for ABC transporter-mediated detoxification after a blood meal (Esquivel, Cassone & Piermarini, 2014). Only a few of the transcripts that were up-regulated after blood feeding have been associated with insecticide resistance in Ae. aegypti; i.e., AAEL005937 and AAEL008624 (Bariami et al., 2012; Vontas et al., 2012) (Table S11). Thus, the transcriptional response of ABC transporters to a blood meal appears unique to that found in insecticide-resistant mosquitoes, which may suggest that ABC transporters are more specific in the substrates they transport compared to substrates that are enzymatically modified by GSTs and CYP450s.

GST biochemical assay

Given that transcripts encoding GSTs in Malpighian tubules showed an overwhelming upregulation at each time point after blood feeding (Table S11), we sought to determine whether the biochemical activity of GST was affected. As shown in Fig. 5, the biochemical activity of GST significantly increased from 110.3 ± 6.8 nmol CDNB/min/mg protein in the Malpighian tubules of NBF mosquitoes to 172.6 ± 18.8 nmol CDNB/min/mg protein in the tubules of 3 h BF mosquitoes. The GST activity in the Malpighian tubules of 12 h and 24 h BF mosquitoes (12 h BF = 163.5 ± 11.1 nmol CDNB/min/mg protein; 24 h BF = 182.5 ± 23.1 nmol CDNB/min/mg protein) was significantly elevated compared to NBF controls and similar to the tubules of 3 h BF mosquitoes (Fig. 5). Thus, the upregulation of GST transcripts in Malpighian tubules 3–24 h after a blood meal correlates with an increase in biochemical GST activity by 3 h and persists through 24 h. These findings are consistent with results of recent studies that have found whole mosquito GST activity increases in Ae. aegypti, C. pipiens, and several anopheline species after one or more blood meal (Oliver & Brooke, 2014; Tripathy & Kar, 2015).

Taken together, the above transcriptomic and biochemical data suggest that the Malpighian tubules of mosquitoes enhance their capacity for xenobiotic metabolism and excretion within 24 h post-blood meal. These findings may in part explain why previous studies have found that the tolerance of mosquitoes to insecticides increases following a blood meal (Halliday & Feyereisen, 1987; Oliver & Brooke, 2014).

Uric acid assay

We next sought to determine whether the above changes in transcript expression correlated with changes in the uric acid content of Malpighian tubules following a blood meal. As shown in Fig. 7, the uric acid content in the Malpighian tubules of 3 h BF mosquitoes (1.39 ± 0.12 mmol uric acid/mg) and NBF mosquitoes (1.23 ± 0.06 mmol uric acid/mg protein) were similar. However, in the Malpighian tubules of 12 h BF mosquitoes, the uric acid content significantly increased to 2.19 ± 0.13 mmol uric acid/mg protein (Fig. 7). The uric acid content in the Malpighian tubules of 24 h BF mosquitoes (1.785 ± 0.14 mmol uric acid/mg protein) was significantly greater than that in the tubules of NBF and 3 h BF mosquitoes, but significantly lower than that in the tubules of 12 h BF mosquitoes (Fig. 7).

Taken together, the above transcriptomic and biochemical data suggest that the Malpighian tubules of mosquitoes enhance their capacity to synthesize uric acid within 24 h post-blood meal. These findings are consistent with results from our previous study (Esquivel, Cassone & Piermarini, 2014) and correlate well with the peak rate of uric acid excretion found in Ae. aegypti mosquitoes after a blood meal, which occurs between 12 and 24 h after blood ingestion (Von Dungern & Briegel, 2001). Our data also suggest that the Malpighian tubules likely play a key role in the excretion of uric acid by mosquitoes following a blood meal.