Midgut of the non-hematophagous mosquito Toxorhynchites theobaldi (Diptera, Culicidae)

In most mosquito species, the females require a blood-feeding for complete egg development. However, in Toxorhynchites mosquitoes, the eggs develop without blood-feeding, and both females and males exclusively feed on sugary diets. The midgut is a well-understood organ in blood-feeding mosquitoes, but little is known about it in non-blood-feeding ones. In the present study, the detailed morphology of the midgut of Toxorhynchites theobaldi were investigated using histochemical and ultrastructural methods. The midgut of female and male T. theobaldi adults consists of a long, slender anterior midgut (AMG), and a short, dilated posterior midgut (PMG). The AMG is subdivided into AMG1 (short, with folds) and AMG2 (long, without folds). Nerve branches and enteroendocrine cells are present in AMG and PMG, respectively. Compared with the PMG of blood-feeding female mosquitoes, the PMG of T. theobaldi is smaller; however, in both mosquitoes, PMG seems be the main region of food digestion and absorption, and protein secretion. The epithelial folds present in the AMG of T. theobaldi have not been reported in other mosquitoes; however, the midgut muscle organization and endocrine control of the digestion process are conserved in both T. theobaldi and blood-feeding mosquitoes.

Scientific RepoRts | 5:15836 | DOi: 10.1038/srep15836 The midgut in blood-feeding female mosquitoes is the site of blood digestion and the gateway for establishment of various human pathogen, including viruses, protozoa, and nematodes [17][18][19] . This explains why the midgut is one of the most understood organs in mosquitoes. However, there has been little research on the midgut of non-hematophagous mosquitoes, such as Toxorhynchites. Therefore, in the present study, the midgut morphological and functional characteristics of female and male Toxorhynchites theobaldi were investigated, and the differences between this species and blood-feeding mosquito species were discussed. Additionally, this study will also help in understanding the overall morphophysiology of the Culicidae midgut.

Results
General morphology and histology. The midguts of both female and male T. theobaldi consist of a long, slender AMG, and a smaller, dilated PMG. In both females and males, the AMG is divided into two distinct parts: AMG1, with folds on the surface and located in the thorax; and AMG2, without folds and located in abdomen ( Fig. 1a and Sup. Fig. a). The total length of the midgut was 6.1 mm in females and 4.5 mm in males, however, length and width of the different regions of the midgut were proportional between females and males. The length of the AMG corresponded to ~84% of the total midgut length. The length of AMG1 corresponded to a quarter of the total length of the AMG. The width of PMG was higher than AMG1 or AMG2 (Fig. 1b).
In the three regions of the T. theobaldi midgut (AMG1, AMG2, and PMG) there was a single cell layer epithelium with cells displaying brush borders (Figs 2a,g and 3a). The AMG1 epithelium was continuous  with the cardia epithelium (proventriculus or the transition between the foregut and midgut) and had many wrinkles or folds (Fig. 2a,d). In AMG2 and PMG, no folds were seen, but undulations occurred in the basal region of the epithelium, where the circular muscles are inserted (Figs 2g and 3a). In AMG1, digestive cells were approximately of the same height (Fig. 2a), unlike AMG2, where cells exhibited different heights, forming a narrow lumen with an "X" shape when cross-sectioned (Fig. 2f). In the PMG, digestive cells were aligned with the nucleus positioned at the same height (Fig. 3a). The height of the epithelium and the thickness of brush border in each of the three-midgut regions were similar between females and males (p > 0.05). Digestive cells in PMG were higher (43.34 μ m) and had thicker brush borders (11.31 m) than AMG1 and AMG2, which had similar measurements (22.78 m epithelial height and 3.59 m brush border) (Fig. 1c).
In the three regions of the midgut, digestive cells had elongated or rounded nuclei, and their supranuclear portion was predominantly acidic (Figs 2c and 3a). The subapical cell region, immediately below the brush border, was negative for the PAS reaction, but rich in proteins. This region was thicker or more evident in PMG (Figs 2d and 3d-3e). The brush border on all regions of the midgut was positive for the PAS reaction (Figs 2a,f and 3d). Besides digestive cells, other cells were seen in the basal region of the midgut epithelium (Figs 2c and 3a). These basal cells could correspond to regenerative or enteroendocrine cells. In whole mounts, they had small nuclei with no defined shape, unlike the digestive cells with large, elongated, and regular nuclei ( Fig. 3f-g). Externally, the midgut had elongated and large nuclei of muscle cells (Fig. 3h).
The lumen of AMG1 and AMG2 was narrow, sometimes with opposite brush borders very close to each other (Fig. 2a,f), while the PMG lumen was large (Fig. 3c). In AMG1 and PMG, structures protruding from the apical side of the digestive cells towards midgut lumen were visualized, characterizing the process of aprocrine secretion. These structures were positive for proteins (Figs 2d and 3b), and negative for PAS reaction ( Supplementary Fig. d,e). In AMG1, these structures were basic, while in PMG, they were acidic (Figs 2c and 3a). This type of cell process was not seen in AMG2.
In midguts where the food bolus is being transferred to the hindgut, digestive cells of AMG and PMG exhibited intense staining for the PAS reaction. This staining was observed in the apical and basal regions, or only in the basal region of cells (Figs 2e and 3d).
A PAS-positive material, corresponding to the PM, was also seen in the lumen of midgut, including inside epithelial folds of AMG1 (Figs 2b,g and 3d). This PM was thin and loosely organized (Figs 2b and 3d). The presence of PM in the lumen of all midgut regions of T. theobaldi were confirmed by WGA-FITC staining ( Fig. 4a-f). In the AMG1 (Fig. 4a) and AMG2 (Fig. 4b), the labelling was seen just above the brush border. Differently, the labelling in PMG were more diffused or located in the center of the lumen (Fig. 4c).
By phalloidin-FITC labeling (actin marker), muscle bundles were seen in the midgut of T. theobaldi, forming a network covering the outer wall of the organ. This labeling revealed the arrangement of circular and longitudinal bundles in the three regions of the midgut (Fig. 5a-c). The longitudinal bundles had similar width in the AMG and were apparently thicker in the PMG, whereas the circular bundles were thicker in AMG1, becoming narrower in the passage between AMG1 and AMG2 ( Fig. 5d-f).
The circular bundles were organized orthogonally to the longitudinal ones. Each circular bundle was interconnected to neighbor circular bundles. The interconnection of the circular muscle bundles always occurred at the same position along the midgut length (Fig. 5d,e,g).
The longitudinal bundles were parallel and with few branches, seen only in some muscle bundles of AMG1 and PMG (Fig. 5d,g). The longitudinal bundles were not all continuous from the beginning to the end of the midgut, with some of them terminating early in the PMG, while others originated from the rear end of the PMG, extending the transition of AMG2 to PMG (Fig. 5f).
Transmission electron microscopy. AMG1. The AMG1 digestive cells had densely-packed microvilli in the apical region and invaginations in the basal region, forming an extensive and sparse labyrinth, which occupied nearly half the cell (Fig. 6a,f). The microvilli were thin, tall, numerous, and contained extracellular material with granular aspect on its ends, corresponding to the PM (Fig. 6b).
The digestive cells of AMG1 had many mitochondria and lamellar bodies in the apex (Fig. 6c). Golgi apparatus, small autophagic vacuoles, lamellar, and multilamellar bodies were also seen (Fig. 6d,e,g). The basal lamina was compact and continuous, and had undulations and depressions (Fig. 6f,h). Muscle cells were located adjacent to the basal lamina (Fig. 6a,g).
Regenerative cells were seen in the basal region of the epithelium of AMG1, AMG2, and PMG (Figs 6h and 7d). These small cells had few organelles and extensive lateral expansions, which establish a connection with the neighboring regenerative cells and the basal lamina. Differentiating regenerative cells had emerging microvilli and basal labyrinths (Fig. 6h).
between epithelial folds (F). (e) AMG1 of a male showing intense labeling for carbohydrates (arrows) in the digestive cells. (f) Cross-section of AMG2 of a female showing cells of different sizes and a X-shaped narrow lumen. (g) AMG2 of a male with an undulated basal lamina (bl) where muscle cells (mc) are inserted. The peritrophic matrix (PM) is thin, PAS-positive and is located next to the brush border (bb). Lu: midgut lumen.
AMG2. AMG digestive cells had long and slender microvilli (Fig. 7a, inset and 7b). As well as in AMG1, microvesicle-like structures were seen close to the PM, with single or double layers (Fig. 7c). The digestive cell cytoplasm had autophagic vacuoles (Fig. 7b, inset), electrondense lysosome-like structures, and many mitochondria and lamellar bodies concentrated in the apex (Fig. 7b). The basal labyrinth was extensive, but less developed than in AMG1, and the basal lamina was compact and continuous ( Fig. 7d).
Enteroendocrine cells were seen in AMG2, and in PMG. In both regions, these cells had electron-lucent nuclei, few mitochondria, and many small electrondense granules. These cells established contact with the basal lamina through extensive cytoplasmic processes (Figs 7d and 8h,i).
PMG. The microvilli of digestive cells of PMG were thin, numerous, and higher than that of AMG and were also associated to microvesicle-like structures (Fig. 8a, inset). Autophagic vacuoles of various sizes, multilamellar bodies, and Golgi apparatus were also found here ( Fig. 8b-f). Digestive cells in the PMG were rich in rough endoplasmic reticulum with their concentric lamellae accumulated in the supranuclear region ( Fig. 8f-g). Large vesicles, or inclusion bodies, containing eletrondense structures or a granular material (Fig. 8g, inset) were seen in the apex of the digestive cells. Basal labyrinth in PMG was less expressive and the basal lamina was thick and compact in some intervals in comparison to AMG (Fig. 8h,i).
Scanning electron microscopy. The midgut topography was similar in T. theobaldi females and males. AMG1 was continuous with the cardia, a dilated structure that connected the esophagus to AMG (Fig. 9a,b). As seen in the histological sections, the AMG1 epithelium had folds (Fig. 9a,d) that were not seen in AMG2 (Fig. 9e).
Ganglia were located just above the cardia and nerve fibers extended along AMG. Nerves ramified and connected to the longitudinal muscle bundles (Fig. 9c). Tracheoles were seen on the entire surface of the midgut and were most commonly found in AMG2 and PMG (Fig. 9e,f).
In AMG1 and AMG2, only the longitudinal muscle bundles were seen (Fig. 9c-e) under SEM. In AMG1, longitudinal bundles were more widely spaced, but the circular bundles still could not be seen, as they were hidden in the furrows formed between the epithelial folds (Fig. 9d). In AMG2, there were many tracheoles and the longitudinal bundles were very close to each other, hiding the circular bundles (Fig. 9e). In the PMG, the longitudinal bundles were widely spaced, allowing the visualization of circular muscle bundles (Fig. 9f,g).
Cell proliferation. Cell proliferation was not detected in any of the three regions in the midgut of 5-to 10-day-old adult mosquitoes under experiment conditions. However, in the positive control, corresponding to the midguts of A. aegypti (4 th larval stage), labeled nuclei were present as expected (Sup. Fig. f).

FMRFamide-like positive cells.
The anti-FMRFamide antibodies labeled neurons and endocrine cells in T. theobaldi midgut. The pattern of this labeling was similar in female and male T. theobaldi adults. FMRFamide-like positive ganglions were seen above the cardia and their ramifications were seen overlying more than half of the AMG (Fig. 10a-c).
Enteroendocrine cells (i.e., FMRFamide-like positive cells) were abundant and scattered among the digestive cells of the extreme end of AMG2 (close to PMG) and throughout PMG (Fig. 10d,e). The

Discussion
The general morphology of the midgut in female and male T. theobaldi resembled that of the midgut of male mosquitoes whose females are hematophagous. In this regard, similar to these males, the AMG of T. theobaldi was thin and long, while the PMG was enlarged and reduced in size. Different of this, in blood-feeding female mosquitoes, the AMG is short and the PMG is expanded (Sup. Figs. b and c) 13,[20][21][22][23] .
The AMG of T. theobaldi was subdivided into two morphologically distinct regions: AMG1 and AMG2. In other Culicidae (both females and males) this subdivision is not evident (or absent), and the AMG is slender and without folds, similar to AMG2 of T. theobaldi. By being wider than AMG2 and containing folds, AMG1 seems to function as a first site of food digestion. The presence of folds increase the contact surface between the food and the digestive epithelium, and probably reduce the speed in which nutrients pass through the lumen, facilitating the digestion and absorption 24 .
The epithelium characteristics of the three midgut regions of T. theobaldi were compatible with the secretory, digestive, absorptive, and nutrient transport functions as reported elsewhere 13,25 . Both AMG1 and PMG seem to be more involved in enzyme secretion and nutrient absorption compared with AMG2. These two regions presented apocrine secretion of proteins and intense labeling for carbohydrates, especially in the basal portion of the digestive cells. However, the acidic character of apocrine secretion in AMG1 versus the basic character of this secretion in PMG indicates that the secreted proteins are probably different in the two regions.
The AMG1 and PMG digestive cells of T. theobaldi showed greater carbohydrates accumulation when food was being transferred to the hindgut. The carbohydrates accumulation, such as glycogen, is common in insect digestive cells during absorption activity 10,26 , and in the PMG digestive cells of larval and adult mosquitoes fed with sugar or blood [27][28][29] . These carbohydrate deposits seen in PMG seem to accumulate because of the digestion process, functioning as energy reserves, or facilitating the subsequent absorption of more carbohydrates 30 .
Apocrine secretions are typically released during the digestive process of insects, and it is speculated that this is also related to regions that perform nutrients absorption 31 . Corroborating this, AMG1 and PMG are apparently more involved with the carbohydrate absorption, and are the regions where apocrine secretion occurred.
Another possibly secretory mechanism present in the midgut of T. theobaldi is the microapocrine secretion. The small single and double membrane structures seen across the midgut lumen of T. theobaldi resemble microapocrine secreted vesicles found in the midgut lumen of various insects 31 . Enzymes, such as amylase, and various peritrofins are released into the midgut lumen by this secretory mechanism 32 . The existence of this type of secretory mechanism is something that needed to be clarified in adults of non-hematophagous mosquitoes.
The abundant rough endoplasmic reticulum (RER) lamellae in the PMG are also found in the PMG of hematophagous females when a blood meal is acquired. The marked presence of these organelles occurs in cells that are specialized in protein secretion 9,13,29 . Accordingly, it is possible to infer that the bloodmeal in blood-feeding female mosquitoes, and the sugar meal in T. theobaldi stimulate intense activity of protein secretion in PMG digestive cells. In addition to the PMG, the AMG1 also had many RER lamellae  in the digestive cells, which is probably related to the apocrine secretion of proteins as demonstrated by histochemistry with bromophenol blue.
Autophagic vacuoles were seen in the digestive cells of all regions of T. theobaldi midgut, being larger in size and quantity in PMG. These vacuoles are related to the recycling of membranes that occurs due to endo-and exocytosis during digestion 29 . Large inclusion bodies were also seen in the digestive cells of T. theobaldi, similar to those observed in the PMG of blood-feeding mosquitoes post bloodmeal 16,29 . The function of these inclusion bodies is unknown, but it has been proposed for recycling membranes, along with the autophagic vacuoles 16 . By containing a large amount of autophagic vacuoles and inclusion bodies compared with AMG, PMG digestive cells may be more involved in vesicular transport than AMG digestive cells.
Lamellar bodies were also abundantly found in digestive cells throughout the T. theobaldi midgut. These structures are composed primarily of lipids and proteins, and their biogenesis involves endocytic and/or autophagic pathways 33 . In some vertebrate digestive epithelia, lamellar bodies are secreted to protect cell membranes from digestive enzymes and the abrasion of food flow 34 . However, the function of these organelles in mosquitoes is unknown.
The PAS reaction and WGA staining 35 confirmed the presence of a thin PM throughout the midgut lumen in T. theobaldi. This PM had a granular/loose appearance under TEM. In hematophagous mosquitoes (adults), the PM is of type I 35,36 , however, for the non-hematophagous T. theobaldi, the classification of PM (if it is type I or II) is still unclear. In comparison with the PM of T. theobaldi, the PM of hematophagous mosquitoes is thicker, more compact, and is found only in the PMG 29,[37][38][39] . The presence of PM in T. theobaldi midgut indicates that this structure is not only related to blood digestion in Culicidae, but also plays a role in sugar digestion of non-hematophagous species of this family. In the latter case, PM would be important to protect the midgut cells against microorganisms, which are found abundantly in sugar-based foods such as nectar and honey 40 .
Cell divisions could not be identified in the midgut of 5-to 10-day-old T. theobaldi adults through antibodies against PH3. The absence of cell division was also reported in adults of A. aegypti (with the exception of newly emerged A. aegypti) and Culex quinquefasciatus 41,42 . The dynamics of stem cells division and differentiation in T. theolbadi midgut need more investigations considering different ages and digestive phases.
The circular muscle bundles of T. theobaldi midgut are interconnected with adjacent circular bundles. This characteristic also occurs in adult female A. aegypti and Anopheles gambiae 15 . In these two hematophagous species, as well as in T. theobaldi, not all PMG longitudinal bundles are continuous. Some bundles emanate from AMG, ending close to the AMG/PMG transition. Additionally, other longitudinal bundles extend from the hindgut towards the PMG surface 15 . These structural similarities between the midgut muscles of T. theobaldi and blood-feeding females indicate the existence of a common organization of muscle fibers in mosquitoes, even with markedly different feeding habits.
The ganglia associated with the cardia are part of the stomatogastric nervous system of insects 43,44 . FMRFamide-like immunoreactive neurons in T. theobaldi innervate only the anterior portion of the midgut, as in other mosquitoes previously studied 43 . Neuropeptides, such as FMRFamide-like peptides (FLPs), are secreted by neurons and endocrine cells, and supposedly these peptides act in the control of the digestive process 43,45,46 . The physiological roles of these peptides in mosquito midguts are unknown, but in some insects the FLPs look like to be involved in the control of gut motility and secretion of digestive enzymes [47][48][49][50] .
FMRF-like immunoreactive (enteroendocrine) cells are found in the PMG and in the final portion of AMG2 of T. theobaldi adults, while in A. aegypti adults, these cells are only seen in the PMG 22,42,43,47 . Despite this minor difference in the location of enterodocrine cells, the neuroendocrine control of the digestive process based on FMRFamide-like neuropeptides, in both blood-feeding and in non-blood-feeding mosquitoes, seems to be performed by the ramifications of the stomatogastric nervous system in AMG, and the endocrine cells in PMG. The enteroendocrine cells of T. theobaldi present abundant small, electrondense secretory granules, as well as the enteroendocrine cells described elsewhere 16 . Thus, the presence of these granules is another feature conserved among adult mosquitoes with different diets.
The presence of a midgut with a long AMG and reduced PMG in T. theobaldi adults is likely an adaptation to sugar-rich diets. Other adaptations to this diet include modifications of the stimuli perception 51 , as in the morphophysiology of salivary glands 52 . Compared to blood-feeding females, Toxorhynchites amboinensis females lost some types of chemoreceptors in the palps and antennae, such as a putative ionotropic receptor and various odorant receptors. This simplification of the chemoreceptive repertoire probably resulted from the loss of the blood-feeding habit by Toxorhynchites 51 . Additionally, Toxorhynchites splendens salivary glands lack secretory cavities and are morphologically and biochemically similar between male and female, unlike in blood-feeding species, where these glands are sexually dimorphic, and females synthesize a range of proteins related to blood-feeding habits 52 .
Through our observations, we conclude that the midgut features that are similar among T. theobaldi and hematophagous females include: (1) PMG is rich in specialized organelles for protein secretion; (2) the muscular organization in PMG involves sharing of muscle fibers between neighboring muscle bundles; (3) longitudinal muscle bundles are not continuous along the organ; (4) regenerative cell divisions were not detected in aged adults; and (5) FMRF-like immunoreactive cells, including nervous cells and endocrine cells, are located in AMG and in PMG, respectively. However, the differences between the midguts of T. theobaldi and hematophagous females include: (1) AMG in T. theobaldi is subdivided into two anatomically distinct regions, AMG1 and AMG2, while in hematophagous females this subdivision is not evident; (2) AMG is very long and PMG is small in T. theobaldi when compared with hematophagous females; (3) PM is very thin and is observed in the whole midgut of T. theobaldi, but thick, compact, and is synthesized only after a bloodmeal in hematophagous females.
Our results indicate that the morphophysiology of the midgut of the autogenous and sugar-feeding mosquito T. theobaldi is similar in both males and females, unlike blood-feeding mosquito species, where sexual dimorphism is evident. This similarity can be ascribed to both female and male T. theobaldi having the same feeding behavior. Information on protein synthesis by the midgut of these mosquitoes can unravel the differences in morphology and physiology between the midguts of T. theobaldi and hematophagous mosquitoes; however, to date, nothing is known regarding this in the genus Toxorhynchites. Studies focusing on the enzymatic activity and proteomics of the midgut in Toxorhynchites species are the next steps to improve the understanding of midgut physiology in these insects, providing new insights into the evolutionary adaptations of the family Culicidae related to a sugar-based diet.

Material and Methods
Mosquitoes. Immature specimens of T. theobaldi (larvae of different stages and pupae) were collected from Mata do Paraíso (20°45′ 14″ S, 42°52′ 55″ W), Atlantic Forest region of Viçosa, Minas Gerais, Brazil. Larvae and pupae were collected in black plastic buckets (5 L) containing rainwater at ground level, near tree trunks. The specimens were transferred to the insectary of the Departamento de Biologia Geral/UFV, where the larvae were reared individually in transparent glass vials (200 mL) and fed daily with Aedes aegypti larvae at different stages.
Pupae were transferred to plastic pots containing dechlorinated tap water and kept in cages until the emergence of adults. Adults were fed ad libitum with 10% sucrose solution, and dissected 5-10 d after emergence. The insects were kept at a temperature of 26 ± 2 °C and humidity 60-70%.
Whole mounting and histology. Twenty midguts (10 males and 10 females) were dissected in phosphate buffered saline-PBS pH 7.6 (0.1 M NaCl, 20 mM KH2PO4, and 20 mM Na 2 HP 4 ) and fixed in 2.5% glutaraldehyde solution (sucrose/cacodylate buffer 0.1 M pH 7.2) for 24 h. The fixed midguts were cut with microscissors in three distinct regions: AMG1 (initial portion of the AMG), AMG2 (posterior portion of the AMG), and PMG. The samples were washed with distilled water, dehydrated in ascending ethanol series (70, 80, 90, 95 and 100%) and embedded in 2-hydroxyethyl methacrylate historesin (Leica Microsystems Heidelberg Mannhein, Germany). Serial sections of 2-μ m thickness were obtained using an automatic microtome with glass knives.
The sections were subjected to different staining methods. As conventional staining methods, hematoxylin-eosin (HE) or toluidine blue staining protocols were used according to standard routine laboratory procedures. For histochemistry, the periodic acid Schiff (PAS) reaction 53 was employed for neutral glycoproteins, neutral carbohydrates and glycogen detection; and bromophenol blue 54 for total protein detection. After staining and drying, the slides were mounted using Eukitt (Fluka, St. Louis, MO, USA) mounting medium, analyzed, and photographed under an Olympus BX53 microscope, coupled with an Olympus DP 73 digital camera of the Laboratório de Sistemática Molecular, Departamento de Biologia Animal/UFV.
The length and width of AMG1, AMG2, and PMG of 20 insects (10 females and 10 males) were measured. The total organ length was obtained by adding the length measurements of AMG and PMG. The height of the epithelium (i.e., digestive cells), and brush border were measured separately in AMG1, AMG2, and PMG. The measurements were performed only at regular epithelia, with no cytoplasmic protrusions, and with muscular layer and brush border evident. Measurements were performed using the Image Pro-Plus 4.5 software (Media Cybernetics) in light micrographs of six females and six males.
The midgut cell organization was also analyzed in whole mounts. For this, 10 fixed and washed midguts were stained using diamidino-2-phenylindole (DAPI) (Biotium, Inc., Hayward, CA, USA) for 30 min and mounted with Mowiol (Sigma-Aldrich Brasil Ltda, São Paulo, SP, Brazil) solution. The slide preparations were photographed with a epifluorescence microscope (see above).
To examine the midgut muscle organization, 10 midguts (5 males and 5 females) were dissected and fixed for 2 h in Zamboni's solution, rinsed thrice for 30

Immunofluorescence.
To identify dividing cells in the midgut, a primary antibody against the nuclear protein phospho-histone H3 (PH3) (Cell Signaling Technology, Inc., Beverly, MA, USA) was used 42 . Midguts of 10 males and 10 females were fixed in Zamboni's fixative, washed three times for 30 min each in PBS/1% Triton X-100 (PBST) and incubated for 24 h at 4 °C in anti-PH3 primary antibody (Cell Signaling) (1:100) in 1% PBST. The samples were washed three times and incubated with a secondary antibody conjugated with FITC (Sigma) (1:500) in PBS for 24 h at 4 °C, followed by three washes in PBS of 10 min each.
For detection of cells expressing neuropeptides FMRFamide (neurons and endocrine cells), after fixation, 20 midguts (10 females and 10 males) were washed in PBST and incubated for 24 h at 4 °C with anti-FMRFamide primary antibody (Peninsula Laboratories, Inc., San Carlos, CA, USA) (1:400) in 1% PBST. Samples were washed, incubated with the secondary antibody and washed again, as described above. For both in situ identification of proteins (H3 and FMRF), the cell nuclei were stained with TO-PRO-3 Iodide (Life Technologies) for 30 min, washed in PBST, mounted, and analyzed under CLSM (2.2). The total number of FMRFamide-positive cells in the midgut was determined for each insect manually using the z-stack tool of multiple confocal planes.

Controls.
Midguts from six fourth instar larvae (L4) of A. aegypti were used as a positive control for PH3 identification 42 . As negative controls for the two proteins (H3 and FMRF) above, midguts of male and female (n = 3, each) T. theobaldi were treated as described in Immunofluorescence item, but without primary antibodies. As negative controls for WGA-FITC staining, histological sections of midguts of 4 individuals (2 males and 2 females) were mounted with Mowiol solution and observed under fluorescent microscope.

Statistical analysis.
Measurements of midguts and cell counts were subjected to analysis of variance (ANOVA) for variables with normal distribution, and to the Kruskal-Wallis test when non-normal distribution was found. Results were deemed significant when p < 0.05. Standard deviations (SD) were calculated using GraphPad Prism version 4.0 for Windows (GraphPad Software, San Diego, California, USA), and the data were expressed as replicate means.