Ammonia-N exposure alters neurohormone levels in the hemolymph and mRNA abundance of neurohormone receptors and associated downstream factors in the gills of Litopenaeus vannamei

Ammonia-N (NH₃/NH₄⁺) is the most common contaminant in aquaculture systems, mainly enriched by the decomposition of residual baits, the excretion of cultured aquatic animals and the remains of aquatic animals. Some benthic crustaceans are more susceptible to exposure to high ambient ammonia levels, owing to the decomposition of nitrogenous organic compounds in the bottom water and relatively low water exchange rates. Ammonia-N is highly toxic to crustaceans and can result in oxidative stress (Liang et al., 2016), immune responses (Zhang et al., 2018) and disruptions to ion regulation (Romano and Zeng, 2007). Indeed, high ammonia levels could heavily damage the hepatopancreas, induce apoptosis and even result in death of individual Litopenaeus vannamei (Liang et al., 2016). Importantly, there are several ammonia detoxification strategies in Portunus trituberculatus, such as detoxification of ammonia to glutamine, urea and uric acid (Pan et al., 2018); however, as major ammonotelic animals, ammonia accounts for 95% of the total excreted nitrogen in Palaemonetes varians and Crangon crangon (Snow and Williams, 1971; Regnault, 1983). Numerous experiments have revealed that ammonia is excreted across the gills of crabs by several transporters, including Rhesus (Rh) protein, Na⁺/K⁺-ATPase (NKA), K⁺ channels and Na⁺/K⁺/2Cl⁻ cotransporter (NKCC) (Ren et al., 2015; Weihrauch et al., 1998; Weihrauch and Donnell, 2015). Furthermore, a microtubule-dependent ammonia excretion mechanism with V-type H⁺-ATPase (V-ATPase) and vesicle associated membrane protein (VAMP) was proposed in Carcinus maenas (Weihrauch et al., 2002), and it was also verified in P. trituberculatus (Ren et al., 2015). As alluded to earlier, a complex ammonia detoxification network exists in crustaceans. However, how ammonia exactly might be excreted in L. vannamei and what commands the gill cells of crustaceans to perform active ammonia transport is still unknown.

Ammonia-N is the most common stressor in water, and is primarily perceived by sensors of the nervous system, which involves a signaling cascade triggered by the activation of neural circuits in the central nervous system (CNS) (Barton, 2002). The hypothalamic–pituitary–interrenal (HPI) axis plays a major role in fish, and directly exerts its effects through neural hormones such as corticotrophin-releasing hormone (CRH), adrenocorticotropic hormone (ACTH) and cortisol. There is evidence that the plasma cortisol level is positively correlated with the ammonia level in rainbow trout exposed to ammonia-N (Ortega et al., 2005). Moreover, CRH and ACTH were proven to be present in L. vannamei, and regulate the release of biogenic amines, such as dopamine (DA), noradrenaline (NA) and 5-hydroxytryptamine (5-HT) (Ottaviani and Franceschi, 1996; Zhao et al., 2016). In crustaceans, some studies have been conducted on biogenic amines (Briffa and Elwood, 2007; Christie, 2011). It has been reported that ammonia-N or salinity stress could result in an increase in the concentrations of DA, NA and 5-HT in crustacean hemolymph (Zhang et al., 2018; Péqueux et al., 2002; Zhao et al., 2016). In arthropods, there are some studies on biogenic amine receptors, which consist of G protein-coupled receptors (GPCRs) that can activate signaling pathways such as cyclic adenosine monophosphate (cAMP), cyclic guanosine monophosphate (cGMP), protein kinase A (PKA) and protein kinase C (PKC), and can produce unique cellular signaling effects in crustaceans (Blenau and Baumann, 2001; Buckley et al., 2016). Particularly, FXYD2 (the gamma subunit of the NKA) and 14-3-3 protein, phosphorylated by PKA, can participate in downstream signal transduction in crabs and regulate enzyme activity (Silva et al., 2012; Jayasundara et al., 2007). In vertebrates, the cyclic AMP response element-binding protein (CREB) acts as a key transcription factor and is mainly phosphorylated by PKA, thereby facilitating its binding to various promoters of target genes. Transcription of CREB has been shown to be activated in L. vannamei in response to ammonia-N (Zhen et al., 2016; Zhang et al., 2018). Furthermore, the Wnt/β-catenin signaling pathway can be activated by CRH and ACTH (Khattak et al., 2010; Kuulasmaa et al., 2008), which may stimulate the secretion of cortisol in mammals (Schinner et al., 2007). Of note, recent work has highlighted the fact that the Wnt/β-catenin pathway is closely associated with ammonia metabolism and transport in vertebrate livers (Merhi et al., 2015). In crustaceans, there is a scarcity of data to support the hypothesis of the existence of an HPI axis and Wnt/β-catenin signaling pathway. It remains unclear whether cortisol, biogenic amines and Wnt/β-catenin, which are regulated by CRH and ACTH, can play a regulatory role in crustaceans under ammonia-N stress.

In crustaceans, an important neuroendocrine organ resembles the hypothalamus–pituitary system of vertebrates, the X-organ/sinus gland (XO/SG) complex, which secretes crustacean hyperglycemic hormone (CHH). This seems to depend on the stimulus produced by some neurotransmitter-like biogenic amines (Camacho-Jiménez et al., 2017). CHH has been extensively studied and appears to be an important multifunctional hormone for L. vannamei; primarily involved in metabolism, the neuropeptide also controls the immune response (Wanlem et al., 2011) and osmoregulation (Camacho-Jiménez et al., 2017). Because CHH is essentially an adaptive hormone, we proposed that any perceived stressful change will immediately trigger the release of CHH, to meet increased energy demands. It has been shown that CHH stimulates the activity of membrane-bound guanylyl cyclase (GC) (but not cytosolic) and results in a significant increase in cGMP levels, thereby modulating protein kinase G (PKG) in the target tissue (Lee et al., 2014). Furthermore, ion concentrations (essentially Na⁺ and Cl^–) in the hemolymph and NKA mRNA abundance in the gills of L. vannamei can be regulated by CHH (Camacho-Jiménez et al., 2018). Collectively, these studies revealed the involvement of CHH in the control of ion regulation. However, when crustaceans are exposed to ammonia-N, there is less information about the regulatory role of CHH in ammonia detoxification and ion transport.

The white shrimp L. vannamei, which is widely distributed on the west coast of the Pacific Ocean, is of great economic value and has excellent properties in breeding. It shows a high tolerance to changes in environmental factors, such as temperature and salinity (González et al., 2010; Chong-Robles et al., 2014). In the present study, we investigated the regulatory effects of neuroendocrine factors on ammonia transport in L. vannamei exposed to ammonia-N, and evaluated the histological alterations in the hepatopancreas. The study aimed to gain insight into the neuroendocrine–ammonia transport network in L. vannamei under ammonia-N stress. This will provide not only a theoretical basis for understanding the neuroendocrine regulation mechanism when exposed to ammonia-N, but also a technical reference for assessing the tolerance of ammonia toxicity in shrimp.

Adult shrimp Litopenaeus vannamei (Boone 1931) (body length 7.5±0.5 cm, mass 6.5±1.0 g) were obtained from Shazikou farm (Qingdao, China). The shrimps were acclimated in tanks (40×50×60 cm) with 90 l of sand-filtered seawater (salinity 31‰, pH 8.2) at 25±0.5°C and aerated continuously using air-stones for 7 days prior to experiments. During the acclimation period, half of the tank water was renewed twice daily and the shrimp were fed with a commercial diet daily (Haiyue Company, Qingdao). Apparently healthy shrimp (physical integrity without injury, normal color and good viability) at the intermolt stage were chosen for the following experiment. The molt stage was discerned by observing partial retraction of the epidermis in uropoda. All shrimp were fasted for 2 days before the experiment to prevent any change in ammonia level owing to metabolic ammonia production following feeding. All procedures were in accordance with the guidelines of the respective Animal Research and Ethics Committees of Ocean University of China and did not involve endangered or protected species.

According to the ammonia-N concentration of bottom water in aquaculture tanks measured by Chen et al. (1988) (0–46 mg l⁻¹) and Zhong et al. (1997) (0–29.3 mg l⁻¹), as well as the current status of the shrimp’s aquaculture environment, four different ammonia-N concentrations of 0.05 (seawater as a control), 2, 10 and 20 mg l⁻¹ were chosen for the exposure experiment. Seawater used for the exposure experiment was the same as that used during the acclimation period. Ammonia concentrations of three experimental groups were prepared by adding 10 g l⁻¹ NH₄Cl stock solution to the seawater. Shrimp were randomly allocated to the experimental groups and the control group with three replicates per sampling point. Each replicate contained five shrimp, initially collecting 360 shrimp, and there was no death during the exposure experiment. One half of the water was renewed with seawater with the same ammonia-N concentration every 12 h, and the water conditions were kept the same as those for the acclimation. Addition of 2, 10 and 20 mg l⁻¹ NH₄Cl caused a slight decrease in pH of the seawater, which was considered negligible, and no measures were taken for correction during the exposure experiment. During the experimental period, ammonia-N concentrations were measured every 12 h and varied by 0.05±0.03, 2.16±0.12, 10.32±0.20 and 20.41±0.28 mg l⁻¹, respectively, for the four experimental concentrations, according to hypobromite oxidation (GB17378.4; State Oceanic Administration, 2007). Five shrimp were sampled randomly from each replicate group at 0, 3, 6, 12, 24 and 48 h ammonia exposure.

Gills were collected after the shrimp were placed on ice for approximately 30 min. Samples used for RNA extraction were dissected using RNase-treated scissors and forceps, and then milled in liquid nitrogen. A total of 80–100 mg of tissue powder was placed into 1.5 ml RNase-free tubes and lysed with 1 ml RNAiso Plus reagent (TaKaRa, Dalian, China). After full mixing, the lysed samples were centrifuged at 12,000 g for 15 min at 4°C and supernatant fluid was stored at −80°C. Hemolymph sample was obtained from the first abdominal segment of each shrimp using a sterilized syringe with an equal volume of the anti-coagulant (450 mmol l⁻¹ NaCl, 10 mmol l⁻¹ KCl, 10 mmol l⁻¹ EDTA-Na₂ and 10 mmol l⁻¹ Hepes, pH 7.45, osmolarity of 780 mOsm kg⁻¹) modified from the anti-coagulant devised by Söderhäll and Smith (1983). After collection, the sample was immediately centrifuged in a refrigerated centrifuge (4°C) for 10 min at 700 g and the supernatant was collected as the plasma sample and frozen at −80°C until analysis.

CHH concentration in plasma was determined by an enzyme-linked immunosorbent assay (ELISA). Polyclonal antibodies (Pabs) to purified recombinant CHH protein of L. vannamei were developed in two male New Zealand White rabbits, 6 months old and 2–3 kg in mass. The experimental protocol was approved by the respective Animal Research and Ethics Committees of Ocean University of China. Briefly, during the first week, the rabbits were injected subcutaneously at six locations with 0.5 ml of disrupted antigen (the protein concentration of the antigen, 100 μg ml⁻¹) intensively mixed with complete Freund's adjuvant in a 1:1 ratio. In the third week after the initial injection, the same immunization was carried out but with incomplete Freund's adjuvant. Subsequently, the same operation was adopted at two weekly intervals. One week after the fourth injection, to determine anti-rLvCHH antibody titers, approximately 0.1−0.5 ml serum was collected from the auricular vein of the rabbits and assayed by ELISA. Antibody dilutions were used instead of rabbit sera as a blank control. Serum samples collected from an unimmunized rabbit were used as negative controls. The negative control revealed the non-specific background noise of the system, which was subtracted from all values of the plate. One of the two rabbits with higher antibody titer and sensitivity was killed and antiserum was harvested from the auricular artery of the rabbit with medical hemostix and stored at −20°C until use.

The ELISA was used to plot the rLvCHH standard curve, and was carried out as previously described by Zou et al. (2003). The purified rLvCHH suspension was serially diluted at 10⁻⁴, 5×10⁻⁴, 10⁻³, 5×10⁻³, 10⁻², 5×10⁻², 10⁻¹ μg ml⁻¹ in phosphate-buffered saline (PBS), and 100 μl of the solution was pipetted into the wells of an ELISA plate. ELISA plate wells were coated with 100 μl of solution and incubated at 4°C overnight, then washed three times with 300 μl of PBS containing 0.05% (v/v) Tween 20 (PBST). Subsequently, the ELISA plate wells were blocked with 200 μl of PBS containing 3% BSA, incubated at 37°C for 1 h, then washed with PBST as described above. The test was performed with rabbit sera diluted in PBST containing 0.5% (w/v) Bio-Rad Blocker (PBST-B) at 1:400; HRP-labeled goat polyclonal anti-rabbit IgG antibody (Beijing Haplen and Protein Biomedical Institute, China) was considered as a second antibody. The wells were incubated with diluted sera at 37°C for 1 h, washed and then incubated with 1:1000 diluted goat polyclonal anti-rabbit IgG antibody in PBST at 37°C for 1 h. After washing, the color development was achieved by substrate solution tetramethylbenzidine (TMB). The substrate solution was incubated in the dark at 37°C for 20 min and color development was stopped by adding 50 μl of 2 mol l⁻¹ H₂SO₄, the absorbance was measured at 450 nm in a microplate reader (SpectraMax 190, Molecular Devices), and the results were evaluated as an index value calculated as P/N, where P/N=(A_450sample−A_{450blank control})/(A_{450negative control}−A_{450blank control}). As P/N≥2.1, the sample was positive for antigen.

CRH, ACTH and cortisol concentrations in plasma were measured using a shrimp corticotropin-releasing hormone ELISA kit (BP-E94005), a shrimp adreno-cortico-tropic-hormone ELISA kit (BP-E94055) and a shrimp cortisol ELISA kit (BP-E94332; all Shanghai Lengton Bioscience, China), respectively. All steps were carried out according to the manufacturer's instructions.

For the biogenic amine concentration assay, 1 ml of plasma sample, 0.5 ml of Tris buffer and 10 mg of acidic alumina were mixed in a 2 ml plastic tube, then centrifuged and the supernatant was discarded. Then, 1 ml of ultrapure water was added and centrifuged to retain precipitation. Finally, 0.5 ml of 0.1 mol l⁻¹ perchloric acid was added to the precipitate and centrifuged at 5000 g for 15 min at 4°C, and the supernatant was collected.

Standard samples of DA and NA (Sigma) were diluted to 1 mg ml⁻¹ and 5-HT was diluted to 2 mg ml⁻¹ with 0.01 mol l⁻¹ NaOH before use. Standard curves were established with 0.1 mol l⁻¹ perchloric acid at six concentrations ranging from 4 to 400 ng ml⁻¹ for each biogenic amine. The DA, NA and 5-HT concentrations of the supernatant were assayed (n=6) by high performance liquid chromatographic-electrochemical detection (HPLC-ECD) with a 20 μl injection valve. The stationary phase consisted of an Agilent column (150×3.2 mm) containing 2 μm octadecylsilane particles. Temperature of the column was kept constant at 28°C. Pumping rate was 0.2 ml min⁻¹. The potential for electrochemical detection was adjusted to 700 mV with a sensitivity of 2 nA. The mobile phase (pH 4.3, adjusted by phosphoric acid) was prepared with 200 mmol l⁻¹ monobasic sodium phosphate, 0.75 mmol l⁻¹ octansulphonic acid, 15 μmol l⁻¹ EDTA, 7% methanol, 4% acetonitrile and 0.25% tetrahydrofurane, diluted with HPLC-grade water. The mobile phase was degasified with a BAS degasifier. The concentration of hemolymph biogenic amines was determined by comparing the chromatograms with those generated by the biogenic amine standard.

Total RNA was extracted from 80–100 mg of gills using RNAiso Plus reagent (Takara, Dalian, China). RNA quantity, purity and integrity were examined by both native RNA electrophoresis on 1.0% agarose gel and the UV absorbance ratio at 260 and 280 nm (Multiskan Go 1510, Thermo Scientific, Finland). After detection, cDNA was synthesized based on 1 μg of total RNA using the PrimeScript RT reagent kit with gDNA Eraser (Takara, Dalian, China). Oligonucleotide primers (Table S1) were designed by Primer Premier 5.0 and synthesized by Sangon (Shanghai, China). Stability of β-actin and ribosomal 18S genes (candidate housekeeping genes) was evaluated using the BestKeeper method (Pfaffl et al., 2004). β-actin was found to have the lower variation. Therefore, it was selected as the housekeeping gene for the rest of the analysis. PCR amplification was carried out according to the method of Zhao et al. (2016). The relative mRNA abundance ratio (R) was calculated using the equation: R=(E_target)^△CP_target^{(control–sample)}/(E_ref)^△CP_ref^{(control–sample)} (Pfaffl, 2001), where CP is defined as the point at which the fluorescence rises appreciably above the background fluorescence. PCR efficiency (E) was determined by running standard curves for 10-fold serial dilutions of cDNA templates, and calculated according to E=10^–1/slope−1 (Rasmussen, 2001). For all standard curves, the primer amplification efficiencies of genes were 91–110% and 0.986<R²<0.997.

Ammonia concentration in the hemolymph was determined directly using a hemolymph ammonia assay kit (Nanjing Jiancheng Bioengineering Institute, A086, China) following the manufacturer's instructions. Hemocyanin was precipitated and the enzyme activity was disrupted with a protein precipitant to prevent the production of free ammonia, while most of the color-interfering substances were removed. Ammonia was prepared in protein-free filtrate by the Berthelot reaction, and then the hemolymph ammonia concentration was calculated by comparison with the standard solution.

After exposure to ammonia-N for 48 h, four groups (control, 2, 10 and 20 mg l⁻¹ ammonia-N treatments) of L. vannamei hepatopancreas were sampled for histological evaluation. Five replicates were set for each group with hepatopancreas tissues from five shrimp. The small pieces of tissues were fixed in Bouin's fixative for 24 h. Following dehydration through an ascending ethanol series, the tissues were embedded in paraffin, sectioned at 6 µm, stained with Ehrlich's haematoxylin and eosin (H&E), and documented with a Nikon microscope.

All raw data were first tested for normality of distribution (Kolmogorov−Smirnov or Shapiro−Wilks tests) and homogeneity of variance (Levene’s test). A one-way ANOVA was utilized to examine the differences between control and treated groups at the same sampling time with SPSS software (version 17.0). All data are presented as means±s.e.m. Significant differences were considered at P<0.05. Tukey's test was used to identify the differences when significant differences were found.

The CHH concentrations of the 2, 10 and 20 mg l⁻¹ ammonia treatment groups were significantly increased, and reached the highest levels at 12 h, which were 1.91-, 2.61- and 3.97-fold higher than the control group, respectively. The 2 mg l⁻¹ ammonia group recovered to the control level at 24 h, but the 10 and 20 mg l⁻¹ ammonia groups were still higher than the control level at 48 h (Fig. 1A). Both CRH and ACTH concentrations increased significantly and reached the maximum at 6 h, then decreased rapidly to the control level after 24 h exposure (Fig. 1B,C). The cortisol concentration in the 2, 10 and 20 mg l⁻¹ ammonia groups increased dramatically at 3 h, which was the most variable at this time point compared with other hormones, with a 2.50-, 2.68- and 2.87-fold increase, respectively. It reached a peak at 6 h, then declined gradually and returned to the control level at 48 h (Fig. 1D). It also showed a dose-dependent effect between the concentration of hemolymph hormone and the concentration of ammonia-N.

The DA (Fig. 2A), NA (Fig. 2B) and 5-HT (Fig. 2C) concentrations in hemolymph increased significantly within 12 h when exposed to 2, 10 and 20 mg l⁻¹ ammonia-N. DA and 5-HT concentrations increased significantly in all treatment groups at 3 h, while the NA concentration in the 2 and 10 mg l⁻¹ treatments increased but was not significant. Their peak values all occurred at 6 h and then recovered to control levels at 24 h. It was also obvious that the concentration of biogenic amines showed a dose-dependent effect with ambient ammonia concentration.

Compared with the control group, GC mRNA abundance in the 2, 10 and 20 mg l⁻¹ ammonia groups was upregulated significantly during the first 12 h. The peak values occurred at 12 h (1.55-, 2.28- and 2.51-fold) and then decreased, but the 20 mg l⁻¹ ammonia group was still higher than the control at 48 h (Fig. 3A). The mRNA abundance of the dopamine receptor D₄ in the ammonia treatment groups decreased significantly, reached the lowest level at 6 h, and then recovered to the control level at 48 h (Fig. 3B). The mRNA abundance of the a₂ adrenergic receptor varied similarly (Fig. 3C). The mRNA abundance of the 5-HT₇ receptor in the 2, 10 and 20 mg l⁻¹ ammonia groups was upregulated significantly, and a peak was noted at 6 h, which was 1.64-, 2.28- and 2.36-fold higher than the control, respectively. Afterwards, levels recovered to the control level at 24 h (Fig. 3D).

Ammonia-N exposure caused a significant change in the mRNA abundance of protein kinases (Fig. 4). The PKA mRNA abundance in ammonia treatment groups was upregulated significantly and reached a maximum at 6 h, up to a 2.92-fold increase in the 20 mg l⁻¹ group, and then recovered to the control level and tended to be stable after 24 h (Fig. 4A). The PKC mRNA abundance in the 2, 10 and 20 mg l⁻¹ ammonia groups decreased significantly, reached the lowest level at 6 h, and then increased slowly to the control level at 48 h (Fig. 4B). The PKG mRNA abundance in the 2, 10 and 20 mg l⁻¹ ammonia groups reached a peak value at 12 h (1.70-, 2.37- and 2.68-fold increase, respectively), then decreased sharply and returned to the control level (except for the 20 mg l⁻¹ group) at 48 h (Fig. 4C).

The mRNA abundance of the signaling protein β-catenin associated with the Wnt signaling pathway was significantly upregulated under ammonia-N stress. In the 2, 10 and 20 mg l⁻¹ ammonia groups, it reached a peak value at 12 h (2.29-, 3.25- and 4.00-fold increase, respectively). However, only the 2 mg l⁻¹ group returned to the control level at 48 h (Fig. 4D). The downstream effector proteins in signal transduction that regulate the activity of ion transporters were also investigated, including 14-3-3 and FXYD2 (Fig. 4E,F). The mRNA abundance of 14-3-3 in the 2, 10 and 20 mg l⁻¹ ammonia groups was significantly upregulated and reached the highest level at 6 h (1.81-, 2.39- and 2.71-fold increase, respectively). Subsequently, the mRNA abundance of all three ammonia groups declined rapidly but was higher than the control group at 48 h (Fig. 4E). The mRNA abundance of FXYD2 also showed a peak change and reached a maximum at 6 h (2.82-, 4.24- and 5.25-fold increase, respectively) after exposure to 2, 10 and 20 mg l⁻¹ ammonia-N, and then recovered to the control level at 48 h (Fig. 4F).

The mRNA abundance of nuclear transcription factors CREB and TCF in the gills of L. vannamei significantly increased from 3 h after exposure to 2, 10 and 20 mg l⁻¹ ammonia-N (Fig. 4G,H). In the 2, 10 and 20 mg l⁻¹ ammonia-N groups, the peak value of CREB occurred at 6 h (2.07-, 3.23- and 4.02-fold increase, respectively), whereas the peak value of TCF occurred at 12 h (2.25-, 2.96- and 3.97-fold increase, respectively). CREB mRNA levels returned to control levels at 48 h, whereas the TCF mRNA level of the 20 mg l⁻¹ ammonia treatment group was still significantly higher than that of the control group.

A significant increase of NKA (Fig. 5A) and K⁺ channel (Fig. 5B) mRNA abundance was found in treatment groups after 3 h. The peak values occurred at 6 h (2.61- and 3.46-fold in the 20 mg l⁻¹ group, respectively). NKA mRNA abundance then recovered to control levels at 48 h, but K⁺ channel mRNA abundance in the 20 mg l⁻¹ group was still significantly higher than the control level after 48 h exposure. As shown in Fig. 5C, ammonia exposure strongly suppressed the mRNA abundance of NKCC during 3–24 h, especially in the 10 and 20 mg l⁻¹ treatments, and the minimum value appeared at 6 h, but this decrease returned to the control level at 48 h. Relative mRNA abundance of Rh in the 2, 10 and 20 mg l⁻¹ ammonia-N groups was significantly upregulated from 3 h, and the highest mRNA level occurred at 12 h (1.96-, 2.30-, 3.12-fold increase, respectively), then at 48 h, levels in the 2 and 10 mg l⁻¹ ammonia-N groups were restored to the control level (Fig. 5D).

To demonstrate the existence of an ammonia excretion mechanism centered on vesicle transport and exocytosis in L. vannamei, the relative mRNA abundance of V-ATPase and VAMP was investigated. As shown in Fig. 5E, a significant increase in V-ATPase mRNA abundance appeared during 2, 10 and 20 mg l⁻¹ ammonia-N exposure and peaked at 6 h (1.73-, 2.34- and 2.66-fold increase, respectively). The mRNA abundance of VAMP, which shared a similar trend with V-ATPase, was significantly upregulated in ammonia treatment groups from 3 h, and reached a maximum at 6 h (1.75-, 2.31- and 3.16-fold increase; Fig. 5F). Afterwards, the 2 and 10 mg l⁻¹ ammonia groups recovered gradually to the control level at 48 h.

The hemolymph ammonia concentration in the 2 mg l⁻¹ ammonia-N group was significantly increased from 24 to 48 h, while the hemolymph ammonia concentration in the 10 and 20 mg l⁻¹ ammonia-N groups was significantly higher than the control group throughout the experimental period. The hemolymph ammonia concentration in the three ammonia treatment groups reached a maximum at 24 h and then decreased slightly (Fig. 6).

The hepatopancreas of control shrimp exhibited a well-organized glandular tubular structure. The tubule lumen was found to have an asterisk-like appearance, and a single layer of epithelial cells was observed to line the tubules (Fig. 7A). After exposure to ammonia (2 mg l⁻¹) for 48 h, the structure of the hepatopancreas duct was not significantly disturbed, but the stellate tubule lumen was dilated (Fig. 7B). In the 10 mg l⁻¹ ammonia-N group, full-scale vacuoles were generated and the borderlines between cells were obscure in the hepatopancreas (Fig. 7C). The hepatopancreas cells of L. vannamei exposed to 20 mg l⁻¹ ammonia-N were severely dissolved, and connective tissue congestion and haemolymph cell infiltration were observed (Fig. 7D).

Teleost fish are amongst the evolutionarily oldest vertebrates and have a neuroendocrine system similar to that of mammals. The best-studied example of neuroendocrine regulation is the interaction between the components of the hypothalamus–pituitary–adrenal (HPA) axis in mammals and the HPI axis in fish, which plays a central role in acclimation to stress. In fish, modulation is based on the established hierarchical control of the axis. When a stress signal is perceived, the hypothalamic region of the nucleus pre-opticus of the brain responds by releasing CRH into the pituitary. The binding of CRH to its receptor stimulates the release of ACTH into the circulation. Subsequently, ACTH stimulates the production and release of cortisol by interrenal cells of the head kidney (Nardocci et al., 2014). For invertebrates, a notable increase of CRH and ACTH concentrations were observed in L. vannamei under salinity stress (Zhao et al., 2016). Similar results were observed in Scophthalmus maximus: exposure to high concentrations of ammonia-N enhanced the levels of CRH, ACTH and cortisol (Jia et al., 2017). Consistent with previous studies, our results showed that the CRH, ACTH and cortisol concentrations of L. vannamei under ammonia-N exposure increased significantly. This indicates that a mechanism similar to the HPI axis might be present in L. vannamei. Furthermore, the concentration of cortisol returned to control levels later than the concentrations of CRH and ACTH. As the final effector of the HPI axis, cortisol exerts a negative feedback on the CRH and ACTH secretory, and participates in the control of systemic homeostasis and the response to stress in organisms. The presence of cortisol has also been reported in invertebrates; in this case, relatively little data are available, i.e. that regarding the presence of cortisol-like molecules in the lobster Homarus americanus and Panulirus homarus (Dixon and Atwood, 1983; Kirubagaran et al., 2002). In the present study, the level of cortisol was evaluated using ELISA, which revealed cortisol-like molecules of similar structure and molecular size. Indeed, studies on cortisol in invertebrates have been neglected but deserve more attention.

It was of interest to consider the release of biogenic amines regulated by CRH and ACTH in crustaceans (Zhao et al., 2016). In the present study, the changes in DA, NA and 5-HT concentration were consistent with the patterns of CRH and ACTH. CRH and ACTH can be secreted into the circulatory system and may subsequently induce the release of biogenic amines from pericardial organs as a response to the ammonia-N stress (Morris, 2001). The results of Zhang et al. (2018) are similar, indicating that DA, NA and 5-HT play regulatory roles in physiological adaption. DA, NA and 5-HT modulate physiological effects through the specific receptors – G protein-coupled receptors (GPCRs). The dopamine receptor D₄ and α₂ adrenergic receptor are GPCRs associated with the G_i protein, which inhibits the activity of adenylyl cyclase (AC), thereby reducing the intracellular concentration of the second messenger cAMP. Finally, a decrease in the concentration of cAMP inhibits PKA. The 5-HT₇ receptor is a member of the GPCRs but coupled to G_s protein, which stimulates the production of cAMP by inducing the intracellular concentration of AC and then activate PKA (Neve et al., 2004; Ruuskanen et al., 2004; Vanhoenacker et al., 2000). In the present study, the mRNA abundance of the D₄ dopamine receptor and the α₂ adrenergic receptor in ammonia-N groups decreased significantly, whereas the 5-HT₇ receptor mRNA abundance increased significantly, suggesting that the elevation of biogenic amines inhibited the suppression effect of PKA and, ultimately, activated the PKA pathway. A significant increase in PKA mRNA abundance in ammonia-N groups detected in this study also supports this hypothesis. However, the PKC mRNA abundance in treatments was significantly downregulated in this study. Additionally, the DA receptor has been identified in the rat brain, which activates the G_q protein and phospholipase C (PLC) (Ng et al., 2010). Kim et al. (1989) also showed that stimulation of AC could inhibit the activity of PLC. Furthermore, 5-HT could decrease PKC activity by raising the concentration of Ca²⁺ (Zhang et al., 2018). Combining these results with our findings, the downregulated DA receptor and upregulated 5-HT receptor may play a role in inhibiting PKC. The present study supports these analyses by demonstrating that: (1) the increasing CRH and ACTH levels may stimulate the release of DA, NA and 5-HT; (2) the increasing DA, NA and 5-HT levels transduce signals through their receptors on the cell membrane of gills; (3) the downregulated dopamine receptor D₄ binding to the G_i protein may increase the mRNA level of PKA but decrease that of PKC; (4) the downregulated α₂ adrenergic receptor binding to the G_i protein depresses the inhibition of PKA; and (5) the upregulated 5-HT₇ receptor activates the PKA pathway, but may eventually reduce the mRNA abundance of PKC.

In addition, there are some downstream effector proteins in signal transduction involved in regulating the activity of ion transporters, such as the 14-3-3 protein and FXYD2. Kagan et al. (2002) reported that the 14-3-3 family of proteins were highly conserved and associated with ion channels in a PKA-dependent phosphorylation manner. The 14-3-3 protein has been detected in L. vannamei (Wanna et al., 2012), Procambarus clarkii (Zhu et al., 2014) and Scylla paramamosain (Shu et al., 2015). Furthermore, Silva et al. (2012) identified the FXYD2 protein in the posterior gills of Callinectes danae, and showed it was phosphorylated by PKA but not by PKC. At a deeper level, CREB can be considered as a transcription factor that regulates the mRNA abundance of ion transporters. It is directly mediated by PKA, whereas it does not appear to be directly phosphorylated by PKC (Johannessen and Moens, 2007). In the present study, changes in the mRNA abundance of 14-3-3, FXYD2 and CREB showed a tendency similar to that of the concentrations of biogenic amines and the mRNA abundance of PKA, and the 14-3-3 protein has a longer duration of effectiveness. Therefore, we assumed that increased biogenic amines can influence the PKA pathway in the gills, and then activate the effector proteins 14-3-3 and FXYD2, as well as the transcription factor CREB to regulate physiological responses under ammonia-N stress.

The Wnt pathway is a highly conserved signaling pathway that is found in all phyla of the animal kingdom. There is evidence that in mammals, the Wnt/β-catenin pathway may be modulated by CRH, and Wnt mRNA abundance is clearly upregulated by ACTH (Khattak et al., 2010; Kuulasmaa et al., 2008). The Wnts are a family of secreted glycoproteins that include identified 19 Wnt ligands in mammals and 12 Wnt ligands in L. vannamei (Logan and Nusse, 2004; Du et al. 2018). The canonical Wnt/β-catenin pathway is centered on regulating the levels of its major effector, β-catenin. The β-catenin accumulated in the cytoplasm translocates to the nucleus, where it forms a complex with the T-cell factor/lymphoid enhancer factor (TCF/LEF) transcription factors and activates target gene expression (MacDonald et al., 2009). To date, little research has been reported on the Wnt/β-catenin pathway in aquatic animals under environmental stress. Transcriptomic analysis of the gills in Oreochromis mossambicus revealed that the Wnt signaling pathway was enriched in freshwater relative to those stressed by salinity (Lam et al., 2014). Proteomics indicated that the Wnt signaling pathway was involved in the process of salinity adaption in L. vannamei (Xu et al., 2017). In the present study, similar changes were detected in the mRNA levels of β-catenin and TCF, suggesting that the Wnt/β-catenin signaling pathway can be activated by ammonia-N stress, which may play a role in ammonia detoxification.

The crustacean-specific hormone CHH functions like the HPI axis of fish. Previous studies have reported that a membrane-bound GC acts as the CHH receptor, and CHH can activate cellular signaling by directly activating GC without binding to G proteins. Then cGMP plays the role of a second messenger, which can regulate the activity of PKG (Chung and Webster, 2006). As was shown in the present study, the relative mRNA abundance of GC and PKG was consistent with the trend of CHH concentration, and they were significantly upregulated in a dose-dependent manner. Similarly, an increase in hemolymph CHH was observed in H. americanus under conditions of acute hypoxia, elevated temperature and altered salinity (Chang, 2005). When C. maenas was exposed to dilute seawater, cGMP and glucose levels in gills and circulating CHH levels increased significantly (Chung and Webster, 2006). In the present study, we speculated that the upregulated hemolymph CHH can activate GC on the membrane, resulting in an increase in PKG by activating cGMP activity. Initially, the CHH pathway appeared to be separated from other signaling pathways, but later study suggested that biogenic amines may stimulate the release of CHH (Camacho-Jiménez et al., 2017). Moreover, the CHH pathway remained unrecovered at the end, and it peaked after biogenic amines and cortisol, indicating that it may be associated with long-term regulation.

Shrimp in aquaculture are not only challenged by ammonia influx, but are also faced with accumulating levels of endogenous ammonia over time. Nevertheless, in our study, the shrimp were able to maintain hemolymph ammonia concentrations well below ambient ammonia levels, suggesting that there were many strategies to cope with elevated ammonia. Currently, there have been many reports that several ammonia transporters can transport ammonia through gills, such as Rh protein, K⁺ channel, NKA and NKCC (Si et al., 2018; Weihrauch et al., 1998; Weihrauch and O'Donnell, 2015). In the present study, the increase in mRNA abundance of Rh protein suggested that Rh protein could facilitate ammonia excretion. A similar report exists for P. trituberculatus, where exposure to high external ammonia resulted in a significant increase in the mRNA level of Rh protein (Ren et al., 2015). However, Martin et al. (2011) detected a downregulation of Rh protein in Metacarcinus magister after long-term high environmental ammonia exposure, which may also serve as a protective measure to minimize passive NH₃ influx. Theoretical considerations argued that NH₄⁺ in the hemolymph may substitute for K⁺ via the K⁺ channel and NKA, and transport NH₄⁺ across the basolateral membrane into the epithelial cells (Choe et al., 2000; Wood et al., 2013). The K⁺ channel we investigated was classified as a potassium voltage-gated channel (K_V) or a delayed rectifier potassium channel, which was a Ba²⁺-sensitive K⁺ channel. A significant increase in K⁺ channel mRNA abundance was detected in all ammonia-N treatment groups, suggesting that the K⁺ channel may also be involved in ammonia excretion, which is consistent with a previous report in C. maenas (Weihrauch et al., 1998). NKA mRNA abundance was markedly upregulated after ammonia-N challenge, suggesting that NKA may be participating in mediating the transport of NH₄⁺ from hemolymph into gill epithelial cells. Wilkie (1997) demonstrated that substitution of NH₄⁺ for K⁺ on basolateral NKCC could contribute to branchial ammonia excretion by marine fish, while Riestenpatt et al. (1996) suggested that an apical NKCC was present in the gills of C. maenas. Our results showed that NKCC was significantly downregulated in the gills of L. vannamei under ammonia-N stress, suggesting that the shrimp may reduce ammonia influx via downregulation of the relative mRNA abundance of apical NKCC. Interestingly, V-ATPase was thought to acidify cytoplasmic NH₃ as NH₄⁺ and trap NH₄⁺ in the vesicles for exocytotic release from the gill (Weihrauch et al., 2002). The increase in V-ATPase mRNA abundance provided evidence for the presence of a vesicular ammonia-trapping mechanism. Together with V-ATPase, VAMP is required to actively excrete ammonia by vesicles for membrane fusion (Trimble et al., 1988). A significant upregulation of VAMP was observed in the present study, which was in agreement with our hypothesis that an exocytotic ammonia excretion mechanism may be enhanced by ammonia-N stress. The possibility of an exocytotic ammonia excretion mechanism should be considered in L. vannamei. In this situation, toxic ammonia was trapped in vesicles rather than diffusing through the entire cytoplasm, which could reduce the damage.

At a deeper level, the mechanism of ammonia transport can be controlled by the neuroendocrine system. Some ion channels, such as NKA, K⁺ channel and NKCC, were regulated by PKA and PKC. Also, both FXYD2 and 14-3-3 protein, which are phosphorylated by PKA, can activate NKA (Zhang et al., 2008; Yuan et al., 2002; Yang et al., 2001; Silva et al., 2012; Jayasundara et al., 2007). It has also been demonstrated that CHH exerts ionoregulatory actions on the gill of L. vannamei by regulating the hemolymph Na⁺ and Cl⁻ and the mRNA abundance of NKA (Camacho-Jiménez et al., 2018). Kiilerich et al. (2007) demonstrated the important role of cortisol in upregulating the transcript levels of NKA and NKCC in Atlantic salmon gills. Merhi et al. (2015) also proposed the human ammonia permease gene RHBG as a direct target of Wnt/β-catenin signalling, and showed that siRNA-mediated β-catenin knockdown resulted in significant reduction of RHBG mRNA. Likewise, cortisol stimulated mRNA levels of Rh protein and V-ATPase, as well as ammonia permeability in cultured trout gill epithelium (Tsui et al., 2009). In addition, exocytosis can be stimulated by direct activation of PKA and PKC as well as by 14-3-3 protein (Jung et al., 2004; Burgoyne et al., 1993). Taking these results into account along with those from our study, we suggest that NKA mRNA abundance was mainly regulated by PKA and PKG, and signal proteins 14-3-3 and FXYD2 as well as cortisol also play multiple regulatory roles. PKA may have upregulated the mRNA abundance of the K⁺ channel, and its abundance was not restored to the control level, probably regulated by 14-3-3 protein (Kagan et al., 2002). Inhibition of NKCC mRNA abundance may be primarily regulated by PKC or other pathways. Importantly, Rh protein may be mediated by cortisol and the Wnt/β-catenin signaling pathway. The relative mRNA abundance of PKC was suppressed, whereas V-ATPase and VAMP were upregulated, indicating that exocytosis might be regulated by multiple effects of PKA, 14-3-3 protein and cortisol under ammonia-N stress. Based on the results of this study, we propose a hypothetical model of the signal transduction pathways under the ammonia-N stress in the gills of L. vannamei (Fig. 8).

The present study showed that the hemolymph ammonia concentration in ammonia-N treatments increased significantly throughout the experiment, and similar changes were found in P. trituberculatus (Ren et al., 2015) and M. magister (Martin et al., 2011). The gills of crustaceans are the main organ for ammonia excretion, whereas the hepatopancreas is an important detoxification and metabolic organ that is very sensitive to water-borne pollutants. The hepatopancreas is one of the indicators in shrimp that can be used to identify the condition of the animal the (Manan et al., 2015). Although there are many detoxification strategies in crustaceans, we observed severe damage of hepatopancreas in the high ammonia-N concentration group. Therefore, we hypothesized that the activated ammonia excretion pathways are still insufficient to completely eliminate the toxicity of ammonia in vivo, which is consistent with the case of hemolymph ammonia.

In general, these results demonstrated that L. vannamei possessed an efficient mechanism to activate ammonia transport to deal with elevated hemolymph ammonia levels. After perceiving the ammonia-N stress, L. vannamei can secrete CRH and ACTH into the hemolymph and stimulate the release of cortisol and biogenic amines. Simultaneously, CHH also participates in physiological responses and potentially plays a role in adaptation to long-term stress. Biogenic amine receptors and GC could transduce the signals from biogenic amines and CHH into gill cells and regulate the intracellular PKA, PKC and PKG levels. 14-3-3 protein, FXYD2 and CREB can control target genes mainly through PKA. Downstream effectors of the cortisol-regulated signaling pathway are unclear, including receptors messengers, transcription factors etc, but Wnt regulates mRNA abundance of target genes mainly through β-catenin binding to TCF. Finally, the ammonia transporters (NKA, K⁺ channel and Rh protein) and exocytosis were activated, except for NKCC, which was was inhibited. However, high concentrations of ammonia-N eventually damage the hepatopancreas, although ammonia excretion is conducted as much as possible.

We thank the staff at the Laboratory of Environmental Physiology of Aquatic Animals for their continuous technical advice and helpful discussions.