A unique Malpighian tubule architecture in Tribolium castaneum informs the evolutionary origins of systemic osmoregulation in beetles

Significance Beetles are the most diverse animal group on the planet. Their evolutionary success suggests unique physiological adaptations in overcoming water stress, yet the mechanisms underlying this ability are unknown. Here we use molecular genetic, electrophysiology, and behavioral studies to show that a group of brain neurons responds to osmotic disturbances by releasing diuretic hormones that regulate salt and water balance. These hormones bind to their receptor exclusively localized to a unique secondary cell in the Malpighian tubules to modulate fluid secretion and organismal water loss. This tubule architecture, common to all higher beetle families, is novel within the insects, and provides an important clue to the evolutionary success of the beetles in colonizing an astounding range of habitats on Earth.

HF® (New England BioLab, MA, USA) linearized pIRESZ2_ZsGreen1 vector using In-Fusion® HD cloning kit (TaKaRa Bio Inc, Kusatsu, JP). After sequence validation (Eurofin, Luxemburg, LU), these plasmids were transfected into competent CHO/G16 cells to develop separate stable clone lines, which were subsequently used in a bioluminescence assay as described in (5).
Tissue-specific cAMP detection. MT cAMP production following ligand stimulation was measured using the time-resolved fluorescence energy transfer (TR-FRET) based LANCE ULTRA cAMP Kit (PerkinElmer, MA, USA) in combination with an EnSight Multimode Plate Reader (Perkin Elmer, MA., USA). In brief, whole MTs were acutely dissected from adult T. castaneum as described above. Then, exactly 10 full-length MTs in stimulation buffer (control), or stimulation buffer supplemented with either DH37 or DH47 at a concentration ranging from 10 -13 M to 10 -6 M, were transferred to individual wells on an OptiPlate-384 (Perkin Elmer, MA., USA). Each sample concentration was setup in 3-6 biological replicates. The loaded plate was then left to incubate at room temperature for 30 min, before adding 5μL 4X EU-cAMP tracer and 5μL of 4X Ulight-anti-cAMP working solutions. The plate was then left to incubate with TopSeal-A sealing film for 1 h, before being measured on the EnSight Multimode Plate Reader, using the TR-FRET program. To calculate absolute changes in cAMP production, we additionally constructed a standard curve, which allowed us to plot a dose-response curve in nM cAMP/tubule. Peptide Synthesis. Synthetic analogues of all peptides used were synthesized by Cambridge Peptides (Birmingham, UK) at a purity of >90%. For T. castaneum DH37 and DH47 ligands, versions with an N-terminal cysteine were additionally made in order to conjugate a TMR-C5maleimide Bodipy dye (BioRad, CA, USA), to make fluorescent TMR-C5-maleimide-SPTISITAPIDVLRKTWAKENMRKQMQINREYLKNLQamide (DH37-F) and TMR-C5-maleimide-AGALGESGASLSIVNSLDVLRNRLLLEIARKKAKEGANRNRQILLSLamide (DH47-F). All peptide concentrations were corrected according to peptide purity.

Generation of Antibodies and visualization of Ligands and Receptor distribution.
To generate antibodies specific against proteins of interest, we analyzed the amino acid (aa) sequence of the proteins to identify the most optimal immunizing peptide region according to a previously described method (6). For Urn8R, this analysis resulted in the selection of a peptide corresponding to aa 3-17 (WSEPLPQEPEPVDAD) in the N-terminal region of the full-length parent protein, which was then submitted for a custom immunization protocol carried out by Genosphere Biotechnologies (Paris, France). Additionally, aa 9-23 (PIDVLRKTWAKENMRK) and aa 27-41 (IARKKAKEGANRNRQILLSL) of the mature DH37 and DH47 peptides, respectively, were also selected for preparation of polyclona antisera. Epitope specificity of the different antisera was established by comparing wild type and RNAi animals by immunostaining as well as by co-application of the pre-immune serum. Additionally, the specificity of the anti-Urn8R antiserum was tested by western blotting. Dissected MTs were lysed in 50ul of RIPA buffer (25mM Tris, 150mM NaCl, 0.5% sodium deoxycholate, 1% Triton X-100) + Halt protease and phosphatase inhibitor cocktail 100X (100:1; ThermoFisher, MO, USA), and homogenized using a beadmill. Next, the sample was centrifuged at 14,000 x g at 4°C for 15 min to pellet debris, before adding x2 Laemmli buffer (Bio-Rad, CA, USA) in a ratio of 1:1 and heat-treating at 95°C for 5 min. The sample was then electrophoresed through a 4-20% precast polyacrylamide gradient gel (Bio-Rad, CA, USA). Proteins were transferred to polyvinylidene difluoride membrane (Millipore, MA, USA), and the membrane was blocked with Odyssey Blocking Buffer (LI-COR, NE, USA). Next, the membrane was incubated with rabbit anti-Urn8R (1:1000) and mouse anti-tubulin (1:2500; Sigma-Alrich, MO, USA) in blocking buffer supplemented with 0.2% Tween 20 (w/v). Primary antisera were detected with goat secondary antibodies -IRDye 680RD anti-mouse and IRDye 800CW anti-rabbit diluted (1:10000)and bands visualized using an Odyssey Fc imaging system. Immunocytochemistry (ICC) was performed as in (7). In brief, appropriate tissues were dissected and fixed in 4% paraformaldehyde in PBS for 20 min. Tissues were then washed foursix times in PBST (PBS + 0.1% Triton X-100), blocked with PBST containing 3 % normal goat serum (blockPBST; Sigma-Aldrich, MO, USA) for 1 h, and incubated in primary antibodies. Primary antibodies used were polyclonal rabbit anti-Urn8R (1:200), polyclonal rat α-DH37 (1:500) and polyclonal guinea pig α-DH47 (1:500). The subcellular localization of the endogenous proteins were visualized by applying Alexa Fluor 488/555/647anti-rabbit, anti-rat or anti-guinea pig secondary antibodies (1:500; Sigma Aldrich, MO, USA) in combination with DAPI (1:1000) and Rhodamine-conjugated Phalloidin (1:500; Sigma Aldrich, MO, USA) in blockPBST overnight at 4°C. Following several washes, first in PBST and then in PBS, the different tissues were mounted on poly-L-lysine coated dishes 35mm glass bottom dishes (MatTek Corporation, MA, USA) in Vectashield (Vector Laboratories Inc., CA, USA) and imaged on an inverted Zeiss LSM800 confocal microscope equipped with airy scan technology (Zeiss, Oberkochen, DE). Where necessary, immunofluorescence levels were quantified using the FIJI software package from images acquired using identical microscope settings as described in (8) Environmental stress exposure. In fed (control) conditions, animals were housed individually in 96-well plate with standard Tribolium medium as described above. For drinking only (water) treatments, animals were kept in 96-well plates with a small block of 1% agar with 0.05% bromophenol blue (BPB); drinking was verified by the presence of blue deposits. For desiccation treatments, animals were kept in 96-well plates with a piece of filter paper without any nutritional and water sources with individual plates being kept at 5, 50 or 90% RH, respectively, at 30ºC.
Hemolymph collection and quantification. Hemolymph was collected according to a modified protocol (9) from animals exposed to the different environmental stress exposures as described above. In brief, animals were washed and subsequently dried on tissue paper for 2 hours to remove moisture. Then, beetles were anesthetized by CO2 and their cuticle pierced between the pronotum and elytron before being transferred to an ice-cold 0.5-ml tube with a small hole in the bottom in groups of 10. This tube was then placed in a larger 1.5-ml collecting tube, which was centrifuged at 12,000 x g for 15 min at 4°C. Hemolymph from three separate tubes were combined into each collecting tubes (containing 500 µl paraffin oil to prevent oxygen-induced melanization) from each environmental condition. Following sample collection, each sample was diluted to a final volume of 50 µl with ddH2O and the osmotic pressure of each sample was measured in triplicates on VAPRO Vapor Pressure Osmometer Model 5600 (Wescor Inc., UT, USA) with each measurement corrected according to the dilution factor of the sample.
Ex-vivo organ culture. For organ culture experiments, brains were dissected from 2-week old mature adults in cold Tribolium saline. Brains were then divided into groups of 8-10 brains, and each group incubated in 500 µL Tribolium saline of different osmotic strengths (-200, -100, 0 or +200mOsm) containing 5% feta bovine serum (Sigma-Aldrich, MO, USA) prepared according to a previously described protocol (10). The samples were incubated for different durations (0 h, 1 h or 4 h) in humidity chambers at room temperature. After the respective incubation periods, the brains were removed for ICC, and the DH37 and DH47 retention levels were measured as described above.
ELISA detection of circulating DH37 levels. To detect circulating DH37 peptide levels, hemolymph was collected from adult T. molitor exposed to either high (RH 90%) or low (RH 5%) humidity conditions for a period of 7 days. In brief, animals were anesthetized by CO2 and then pierced on the dorsal side as described above, before being gently squeezed to collect the clear hemolymph using a p20 micropipette. The collected hemolymph was then immediately transferred to a pre-chilled collection tube containing a small amount of N-Phenylthiourea (to prevent hemolymph melanization) and placed on dry ice, with hemolymph pooled from 5-10 animals to achieve approximately 50 µl per sample; a minimum of 5 samples were collected in total for each condition. Next, the samples were heat in-activated at 60°C for 5 min and subsequently centrifuged at 2,000×g for 2 min at 4 °C. The supernatant was transferred to a new tube and stored at -80°C until further processing. Following sample collection, the DH37 levels in hemolymph was quantified using a modified ELISA protocol (8). Briefly, wells of a 96-well plate was coated using a polyclonal rat α-DH37 (1:20) diluted in coating buffer (2.12 g of Na2CO3 and 6.72 g of NaHCO3 in 500 ml of ddH2O) overnight at 4°C. Next, the wells were rinsed 5 times in PBT and blocked for 1 hour in blocking buffer (PBT + 4% milk powder). Hemolymph samples were then diluted (3:2) in blocking buffer and 100 µl of sample (N=4) and standards (duplicates of 10 -7 M to 10 -9 M of DH37 peptide) were pipetted into appropriate wells and incubated overnight at 4°C on gentle shaking. Following several washes in PBT, a 100 µl of detection solution containing rabbit α-DH37 (11) in blocking buffer (1:1000: generous gift from Dr. Liliane Schoofs, KU Leuven, Belgium) was added to each well and incubated for 2 hours at room temperature. Each well was then washed several times in PBT and then incubated in HRP-conjugated anti-rabbit IgG (1:20000; SigmaAldrich) in blocking buffer on gentle shaking. The solution was then aspirated and the wells washed a final time. Finally, each well was added a 100 µl of TMB substrate solution (SigmaAldrich) and incubated for 30 min, before being added 100 µl stop solution (2M sulfuric acid). The absorbance of each well was measured at 450 nm on an EnSight Multimode Plate Reader Ensight, and the amount of DH37 in each sample was then calculated from the standard curve.
Ramsay Fluid Secretion Assay. Fluid secretions were measured according to a modified version of the method described in (1). In brief, intact MTs were carefully dissected from whole animals and set-up as in vitro preparations by isolating them in drops of Schneider's medium and Tribolium saline (1:1, v/v) under water-saturated liquid paraffin oil, with both ends wrapped around two oppositely placed minuten pins and the middle region bathed in the saline. Next, a small hole was introduced mid-way between the saline drop and the pin, thereby allowing the secreted fluid to accumulate as a discrete droplet. The volume of the secreted fluid was then collected at distinct time intervals and the volume calculated according to (4/3)πr 3 by measuring the diameter of the droplet using an ocular micrometer. An increase in fluid secretion rate following DH37 or DH47 application compared to unstimulated basal conditions was taken as an indication of a diuretic effect. For each species, the above-mentioned protocol was modified to accommodate the vast difference in size and function of the tubules.

Ligand-Receptor Binding Assay.
The ex vivo receptor-binding assay was performed as described in (1, 12, 13). Tubules were carefully dissected from specimens of each species under Schneider's and Tribolium saline and then mounted on poly-L-lysine-covered 35mm glass bottom dishes. Next, the tissues were set-up in a matched-pair protocol, in which one batch was incubated in the appropriate insect saline added the labelled neuropeptide analogue (10 -6 M) and DAPI (1 µg ml -1 ), while the other was incubated in just DAPI; the latter batch was used to adjust baseline filter and exposure settings to minimize auto-fluorescence during image acquisition. Images were subsequently recorded on a Zeiss LSM 800 confocal microscope using baseline filter and exposure settings. A concentration of 10 -6 M of the peptide analogues was chosen for assay, as this was shown to be the minimal concentration needed to produce a saturated receptor response, thereby optimizing the conditions for optical detection of ligand-receptor complexes. Competitive displacement of the labelled ligand under identical microscope settings, following co-application of the labelled (10 -7 M) and unlabeled (10 -5 M) ligands was taken as an indication of binding specificity.
Electrophysiological assays. Scanning Ion-selective Electrode Technique (SIET) and transepithelial potential (TEP) measurements were performed on free isolated MTs from Tenebrio molitor as described in detail in (14,15). In brief, the ion-selective microelectrode voltage was measured at a position close to the tissue (3-5 μm) and subsequently at a more distant position (app. 50 μm) from the tubule. The mean measured voltage difference for three replicate measurements between the inner and outer limits of excursion was converted into a concentration difference. Ion flux was estimated from the measured concentration difference using Fick's law. For TEP measurements, isolated MTs were transferred immediately after dissection to a poly-L-lysine-coated Petri dish filled with Tribolium saline. TEP was measured by impaling the tubule lumen with a sharp microelectrode pulled from double-barreled theta-glass (World Precision Instruments, Inc. FL, USA) with reference to the basolateral bath. The TEP depolarization before and after neuropeptide treatment (10 -7 M) was recorded with a high impedance dual channel differential electrometer HiZ-223 (Warner Instruments, CT, USA) that was connected to PowerLab data acquisition system running LabChart software (ADI Instruments, Oxford, UK).

Production of dsRNA and RNAi-mediated Knockdown.
To silence target gene expression by RNAi, transcript sequences covering app. 200-500 bp were selected. Total RNA was then extracted from either tubules or heads (showing highest enrichment of Urn8R or Urn8, respectively) and cDNA synthesis were carried out as described above. Using the cDNA as template, fragments were amplified by PCR using gene-specific primers that were tagged with T7 promoter sequences in both 3-prime and 5-prime ends (see Supplemental Table 1). These gene specific fragments were then cloned into the pUC19 vector individually and subsequently verified by sequencing (Eurofins, Luxembourg, L). Using the cloned vector as template, bidirectional in vitro transcription was carried out using the MEGAscript T7 transcription kit (ThermoFishcer, MA, USA), and the quality of the resulting dsRNA was checked by gel electrophoresis and quantified using NanoDrop. The concentration was adjusted to 2 µg/ul using injection buffer (1.4 mM NaCl, 0.07 mM Na2HPO4, 0.03 mM KH2PO4, 4 mM KCl), and a total of 500 nl dsRNA solution was injected into age-matched adults using a Nanoject II injector (Drummond Scientific, PA, USA). Animals were allowed to recover for 3 days post-injection before used for experimentation.

Gene expression analysis.
Validation of RNAi-mediated gene knockdown and environmentally induced changes in gene expression was assessed by quantitative Real-Time PCR (qPCR). Total RNA extraction was carried out 3 days post dsRNA injection and cDNA synthesis were carried out as described above. Next, qPCR was performed using the QuantiTect SYBR Green PCR Kit (Fisher Scientific, NH, USA) in combination with a Stratagene Mx3005P qPCR system (Agilent Technologies, CA, USA). Expression levels were normalized against the housekeeping gene rp49. All primers used are listed in Supplemental Table 1.
Desiccation tolerance. Animals were kept on Tribolium medium for 3 days after dsRNA injection. Healthy animals were then transferred to a 96-well plate in a container filled with silica gel beads (Sigma-Aldrich, MO, USA) to produce a low humidity environment (app. RH 5%measured by a custom-build hygrometer). The number of dead animals (not responding to tactile stimuli) were then counted every 4-8 h for 7 days. Data were expressed as percent survival over time.
Quantification of water content. To measure changes in total water content, individual beetles were transferred to a small plastic container and then measured on a Sartorius SE2 ultra micro balance (= WT; Sartorius, Göttingen, DE; 0.1 µg readability). The animals were then housed under low humidity conditions as described above, and after 48 h the beetles were reweighted (= W48). To measure the corresponding dry weight of the animals, they were kept at -20°C over night and then placed in a 65°C incubator for at least 2 days before being weighed a final time (=Wdry). The percent water loss of total body water for each animal was calculated as (WT -WT48)/(WT -Wdry) × 100%, with N=29 animals weighed for each experimental group. Defecation Behavior. To assess the effects of manipulating Urn8-signalling on whole-animal excretory behavior in vivo, dsRNA-injected animals were starved for 2 days followed by refeeding a standard Tribolium medium supplemented with 0.05% (w/w) Bromophenol blue (BPB) sodium salt (Sigma-Aldrich, MO, USA) overnight. This special medium was created by mixing the standard Tribolium medium with BPB and a small amount of water hereby creating a uniform paste, which was left to dry at room temperature overnight. The dried BPB-labelled Tribolium medium was then ground to a fine powder creating a consistency identical to that of the standard medium. Beetles were then placed in individual wells of a 96-well plate fitted with a small piece of filter paper and the number of BPB-labelled deposits produced by each animal over a 4 h period was quantified. The same approach was used to test the physiological effects of DH37 or DH47 hormone stimulation on in vivo excretion, by injecting groups of animals with either PBS or PBS containing DH37 or DH47 peptide corresponding to a final peptide concentration of app. 10 -7 M. A minimum of 18-37 animals was used in each experimental group.
Statistics. The statistical analyses were performed using the data analysis software GraphPad Prism 8 (CA, USA). The normal (Gaussian) distribution of data were tested using D-Agostino-Pearsen omnibus normality test. Data are plotted as mean ± SEM, Tukey's box-and-whisker plots or as violin plots as indicated in each figure legend. Statistical differences between one control group and another group (unpaired samples) or between the same groups at different time points (paired samples) were compared using two-tailed Student t-test, whereas differences between one control group and several other groups were pairwise compared by one-way ANOVA followed by Dunnett's multiple comparisons tests taking P=0.05 (two-tailed) as the critical value. P-values are indicated as: * P < 0.05, ** P < 0.01, *** P < 0.001, **** P < 0.0001. Isoform -RA shows stronger activation by DH37 compared to DH47, which only induces a partial receptor response. Isoform -RB shows high activation by DH47, and a pronounced smaller receptor activity following DH37 stimulation. Both labelled and unlabeled peptides induce similar receptor responses with comparable potencies. (I-J) Dose-response curves of cAMP production as measured by the ultra-sensitive FRET-based LANCE ULTRA method in MTs stimulated with either DH37 or DH47 peptide.     Table S1. Primer sequences used for In-Fusion cloning, RT-qPCR and dsRNA synthesis. Sequences marked in red correspond to vector sequence for In-Fusion cloning primers, and to the T7 promoter sequence for dsRNA primers.