This study was approved by the ethics review board of the Institutional Animal Care and Use Committee at Guangdong Ocean University in Zhanjiang, China. Hybrid groupers (♀ Epinephelus fuscoguttatus × ♂ E. lanceolatu) were purchased from a local farm at Donghai Island (Zhanjiang, China) and were temporarily maintained in an outdoor cement pond (5 m × 4 m × 1.8 m) to test specifications. Water temperature was maintained at 30 ± 1°C, salinity was 28, pH was 7.9 ± 0.1, and the dissolved oxygen concentration was kept at above 5.0 mg/L. The concentrations of total ammonia and nitrite remained at <0.05 mg/L.

A pretest was done before the formal test to calculate the semi-lethal concentration (LC) of ammonia nitrogen for 24 h (LC_{50,24 h}), 48 h (LC_{50, 48 h}), 72 h (LC_{50, 72 h}), and 96 h (LC_{50, 96 h}). A mother liquor of 10 g/L was prepared using analytically pure NH₄Cl as an ammonia nitrogen source. The mother liquor was diluted to 7.8, 12.8, 15.8, 19.2, 24.8, 26.4, 28, 29.6, 35.2, 41, and 44.8 mg/L measured values in accordance with the experimental requirements. A total of 330 healthy, active, and uniform groupers were randomly selected and assigned to 33,100 L white plastic buckets (n = 10). No feeding was done during the experiment, and the ammonia nitrogen concentration in each barrel was measured every 6 h using a PTF-001B multiparameter water quality detector (WBD Biotechnology Co., Ltd., China) and 10 g/L NH₄Cl mother liquor to adjust concentrations. The toxicity symptoms of the fish were observed, and the number of deaths was recorded at 24, 48, 72, and 96 h. By fitting the broken-line model or the quadratic curve model, the LC_{50, 24 h}; LC_{50, 48 h}; LC_{50, 72 h}; and LC_{50, 96 h}, respectively, were obtained.

After the pretest, 300 healthy and uniformly sized (51.4 g ± 2.57 g) of grouper juveniles were randomly and evenly distributed into 15 100 L plastic buckets. The concentrations of total ammonia nitrogen in each treatment group were LC_{50, 24 h} (39.5 mg/L), LC_{50, 48 h} (27.3 mg/L), LC_{50, 72 h} (26.5 mg/L) and LC_{50, 96 h} (25 mg/L), and the control group was treated without stress. The test was done in triplicate over 96 h, during which no feed was given. There was natural illumination, the temperature of the water was 30°C ± 1°C, salinity was 28, pH was 7.9 ± 0.1, and dissolved oxygen was >5 mg/L. The safe concentration (SC) was calculated as (Sprague, 1971):

Each treatment was sampled at 0, 24, 48, 72, and 96 h after ammonia stress. At the beginning of the stress, samples of the control group were taken. The blood, liver, and intestines were taken for enzyme activity analysis, mRNA expression determination, and 16S omics determination. Blood was placed in 1.5 mL microcentrifuge tubes and stored at 4°C for 12 h. The blood was later centrifuged (4,000 × g for 15 min at 4°C) and the serum collected and stored at −20°C for biochemical index and immunoenzyme activity analysis. After then, liver and intestine were immediately separated and loaded in 2 mL enzyme-free centrifuge tubes for enzyme activity analysis and 16S omics determination, and part of the liver was placed in an enzyme-free centrifuge tube containing RNA later for mRNA tissue expression determination. All tissue samples were stored at –80°C. Samples of fish surviving at every half-lethal time (24, 48, 72, and 96 h) were collected every 24 h until the end of the test. The above steps were repeated for the sampling method at each time interval.

Serum and liver biochemical indices were determined by a commercial kit (Nanjing Jiancheng Biotechnology Co., Ltd., Nanjing, China). The activities of superoxide dismutase (SOD), catalase (CAT), glutathione peroxidase (GSH-PX), lysozyme (LYZ), malondialdehyde (MDA), immunoglobulin M (IgM), complement C₃ (C₃), complement C₄ (C₄), lactic acid (LD), cortisol (COR) and intestinal lipase (LPS), amylase (AMS), and trypsin (TRYP) were analyzed using commercial ELISA kits (Shanghai Enzyme-linked Biotechnology Co., Ltd., Shanghai, China), and the corresponding operations were done according to their instructions. Each treatment with triplicate, and each sample was tested twice.

Transzol UP (TransGen Biotech, Beijing, China) of 1 mL was added to the liver samples, and the total RNA was extracted according to the manufacturer’s protocol. The quantity and quality of isolated RNA were detected at 260 nm and 280 nm using a NanoDrop 2000 spectrophotometer (Gene Company Limited, Guangzhou, China) and by electrophoresis in 1% agarose gel, respectively. The first-strand cDNA was synthesized using PrimeScript RTreagent Kits with cDNA Eraser (Takara, Japan) according to the manufacturer’s instructions. The cDNA was stored at −20°C for real-time quantitative polymerase chain reaction (RT-qPCR). Real-time quantitative polymerase chain reaction (RT-qPCR) was performed in a 384-well plate with a 10 μL reaction volume containing 5 μL of SYBR Green Real-time PCR Master Mix, 0.8 μL of each primer, 1 μL of cDNA sample, and 3.2 μL of RNse Free dH₂O. The PCR conditions were set using a thermal programmer at 95°C for 30 s, 40 cycles of 95°C for 5 s, and 60°C for 34 s. Each sample was tested in triplicate. Primers of the reference gene (β-actin) and target gene were designed according to published sequences of groupers (Table 1). Threshold cycle (Ct) values were collected from each sample after finishing the process. The relative expression levels were calculated using the 2^–ΔΔC_t method (Livak and Schmittgen, 2001).

Total genomic microbial DNA was extracted by an MN NucleoSpin 96 SOI kit, then the V3/V4 region of the 16S rDNA gene was amplified by polymerase chain reaction (PCR) at 95°C for 5 min, and then 25 cycles were done at 95°C for 30 s, 50°C for 30 s, and 72°C for 40 s. Finally, primers 338F: ACTCCTACGGGAGGCAGCA and 806R: GGACTACHVGGGTWTCTAAT were used to extend at 72°C for 7 min. PCR products were mixed in accordance with the results of electrophoretic quantification (ImageJ software), and then purified by OMEGA DNA purification columns. After 120 V 40 min electrophoresis, 1.8% agarose gel, the fragments were cut and recycled. The original data were spliced (FLASH, version 1.2.11), and the mosaic (UCHIME algorithm, version 8.1) was removed to obtain a high-quality tag sequence. Intestinal microbe data analysis was done using BMKCloud¹.

All data were analyzed using SPSS version 20.0 (SPSS Inc., United States). First, a homogeneity test of variance was done, and then the significance analysis was done (one-way analysis of variance followed by Tukey tests). The results are shown as means ± standard error of the means.

At the beginning, the reaction of the groupers exposed to the ammonia solution was more intense, and the frequency of the operculum opening and closing was from rapid to slow. The sensitivity of the fishes’ bodies to sound, which showed violent swings, was stronger than that of the control group. Approximately 30 min after the experiment, the body color of the groupers began to lighten. At the same time, their bodies secreted much mucus, which floated in the water in strips. In the LC_{50, 24 h} group, the symptoms of poisoning were obvious after 6 h of stress: the fish did not try to escape when frightened, and they floated in the water, sluggish and spastic. As time passed, some fish lost their balance and eventually died. The specific manifestations were as follows: the dorsal fin was upright, the pectoral fin was stretched forward, the mouth was exaggerated, the gills were bright red or dark red, and the body was rigid and arched.

The serum biochemical indices of groupers under acute ammonia stress are shown in Figure 1A. After ammonia stress, the serum biochemical indices of the experimental groups increased relative to that of the control group. Aspartate transaminase (AST) of the 48 and 96 h groups and Alanine transaminase (ALT) of the 48, 72, and 96 h groups were higher than those in the control group (P < 0.05). The glucose (GLU) contents of the 24, 72, and 96 h groups were also higher than those in the control group (P < 0.05), but the GLU content of the 48 h group was not markedly different from that of the control group (P > 0.05). Meanwhile, the contents of serum COR and LD first increased and then decreased during the stress period and reached their maximum values at 72 h, which were higher than those of the control group (P < 0.05).

The liver biochemical indices of the groupers under acute ammonia stress are shown in Figure 1B. The activities of AST and ALT in the liver increased in the 24 h group (P < 0.05). However, there was no significant difference in liver ALT activity between the other groups and the control group (P > 0.05).

As shown in Figure 2A, the alkaline phosphatase (AKP) activity at each point of time was higher than that of the control group (P < 0.05). Meanwhile, the acid phosphatase (ACP) activity of the 24, 72, and 96 h groups was also higher than that of the control group (P < 0.05), but the ACP activity of the 48 h group was not significantly different from that of the control group (P > 0.05). Relative to the control group, CAT activity increased in the 24 h group and decreased in the 72 and 96 h groups (P < 0.05). After the ammonia stress, SOD and GSH-PX activity decreased first and then increased, reaching the minimum value in the 48 and 72 h groups (P < 0.05). Meanwhile, LYZ activity also decreased after the stress compared with that of the control group (P < 0.05), but the stress duration did not affect LYZ activity (P > 0.05). However, the MDA content increased first and then decreased with stress time, which in the 24, 48, and 72 h groups was higher than that in the control group and in the 96 h group was lower than in the control group (P < 0.05).

Liver immune indices are shown in Figure 2B. In the 24 h group, AKP activity was higher than that in the control group, while that of the other groups was lower than that of the control group (P < 0.05). Ammonia stress increased ACP activity in the liver, and the highest value was found in the 96 h group (P < 0.05). Liver CAT, GSH-PX, SOD, and LYZ activities in the 24 h group were significantly higher than other groups (P < 0.05); meanwhile, CAT, SOD and GSH-PX in the 96 h group were significantly higher than those in the 48 and 72 h groups, but CAT and GSH-PX had no significant difference between the 96 h group and the control group (P > 0.05). However, liver SOD activity from ammonia stress in the 96 h group was higher than that in the other groups (except the 24 h group) (P < 0.05). After ammonia stress, the liver MDA content increased relative to that in the control group (P < 0.05) and reached the highest level in the 48 h group, which was higher than that in the other groups (P < 0.05). In the 24 h group under ammonia stress, the content of IgM in the liver had the maximum value, which was markedly higher than in the other groups (P < 0.05). However, the content of IgM decreased after more than 48 h of stress (P < 0.05). At the same time, the contents of C₃ in the 96 h group and C₄ in the 72 h and 96 h groups were lower than those in the control group (P < 0.05).

After ammonia stress, the activity of the intestinal digestive enzymes decreased (P < 0.05). There was no significant difference in LPS between the control group and the stressed 96 h group (P > 0.05), but the rest had lower levels than those in the control group (P < 0.05). Meanwhile, the minimum values of LPS, AMS, and TRYP appeared in the 48 and 72 h stress groups (Figure 3).

As shown in Figure 4A, ammonia stress markedly up-regulated the mRNA expression of tumor necrosis factor α (TNF-α) and interleukin 6 (IL6) (P < 0.05), and the highest expression was found in the 72 and 24 h groups. The expressions of interleukin 1β (IL-1β) and interleukin 8 (IL-8) mRNA were higher in the 24 h group than in the other experimental groups under ammonia stress (P < 0.05). Meanwhile, the expression of IL-8 mRNA in the 72 h group was also higher than in the other three experimental groups (P < 0.05). However, the expression of IL-1β mRNA in the 48 and 96 h groups was lower than that in the other experimental groups (P < 0.05). The expression of IL-8 mRNA in the 24 h group was higher than that of the control group (P < 0.05), but there was no significant difference in the other groups (P > 0.05).

Acute ammonia stress up-regulated the expression of antioxidant gene mRNA in the liver (P < 0.05) (Figure 4B), but the expressions of heat shock protein 90 (Hsp90), heat shock cognate protein 70 (Hsc70), peroxiredoxin 6 (Prx6), and catalase (CAT) in the 48 h group and peroxiredoxin 4 (Prx4), Prx6, ferritin (FI), and copper/zinc superoxide dismutase (SOD) in the 72 h group were not significantly different from those in the control group (P > 0.05).

The relative expression of apoptosis-related gene mRNA in the liver is shown in Figure 4C. Acute ammonia stress up-regulated the expression of apoptosis-related genes (P < 0.05). However, 48 h of ammonia stress did not significantly affect the expressions of tumor suppressor protein p53 (p53), caspase 1 (casp1), caspase 3 (casp3), or caspase 9 (casp9) mRNA (P > 0.05). Except for the highest expression of p53 in the 72 h group, the highest expression of the other gene mRNA was found in the 24 h group (P < 0.05).

16S sequencing was done on the intestines of the groupers before and 24 h after stress; a total of 419,490 pairs of reads were obtained from 6 samples. After splicing and filtering, 403,798 clean tags were generated. Each sample produced at least 47,227 clean tags with an average of 67,300 clean tags. The two groups of operational taxonomic units (OTUs) were analyzed by a Venn diagram. The results showed that there were 320 repeated OTUs in the two groups, and the abundance of OTUs increased after 24 h of stress (Figure 5A). At the phylum level, the dominant intestinal microbes of a hybrid grouper are Firmicutes (35.2%), Proteobacteria (30.4%), and Bacteroidetes (9.94%). The secondary dominant microbes are Actinobacteria (6.51%), Cyanobacteria (5.66%), Acidobacteria (4.21%), and Chloroflexi (2.70%). In our experiment, ammonia stress increased the abundance of Proteobacteria and Bacteroidetes but decreased the abundance of Firmicutes (Figure 5B). At the family level, the dominant intestinal microbes of the hybrid grouper are Bacillaceae (10.4%), Enterobacteriaceae (7.33%), Camellia_oleifera (4.82%), Rhizobiaceae (4.15%), Ruminococcaceae (4.02%), and the secondary dominant microbes are Weeksellaceae (2.91%), Moraxellaceae (2.61%), Lachnospiraceae (2.41%), and Burkholderiaceae (2.10%). Ammonia stress increased the abundance of Enterobacteriaceae, Weeksellaceae, Lachnospiraceae, and other non-dominant bacteria, but decreased the abundance of Bacillaceae, Ruminococcaceae, Camellia_oleifera, and Rhizobiaceae (Figure 5C). Also, the relative abundances of Fusobacteria and Gemmatimonadetes were different before and after ammonia stress at the phylum level (P < 0.05) (Figure 5D).

Alpha diversity indices as shown in Table 2: the Chao1, ACE, Shannon, Simpson, and Coverage index of each sample at 97% similarity level were calculated. The results of paired sample t-tests showed that there was a difference between OTU and Shannon indices before and after ammonia stress (P < 0.05). However, there were no significant differences in ACE, Chao1, Simpson, and Coverage indices (P > 0.05). Based on the binary Jaccard PCoA analysis shown in Figure 5E, the ammonia stress of the 24 h group reduced the intestinal microbial diversity of the hybrid grouper. In accordance with the different distance matrix information, the samples were classified for the unweighted pair group method with arithmetic mean (UPGMA) clustering analysis; the more similar the samples were, the shorter the common branches were. The UPGMA clustering result of a weighted UniFrac is shown in Figure 5F, in which the samples before and after ammonia stress were clustered in the same branch. Both UPGMA and PCoA showed that the samples were strongly aggregated under environmental influence, indicating that ammonia stress markedly affected the composition of intestinal microbes in groupers. The results of RDA–CCA analysis showed that there was a negative correlation between Cloacibacterium, Escherichia_Shigella, ANcylostoma, Bacteroides, and ammonia. Bacillus, Enterococcus, Clostridium_sensu_stricto_1, Ochrobactrum, and Camellia_oleifera were positively correlated with ammonia (Figure 5G). Meanwhile, Figure 5H shows that ammonia in water affected the intestinal microbes of the tested groupers (P < 0.05).

Toxic reaction is the most obvious characteristic of ammonia stress in grouper. Ammonia refers to nitrogen existing in the form of free ammonia (NH₃) and the ammonium ion (NH₄⁺) in water, among which free ammonia (non-ionic ammonium) is the main hazard to fishes. Most bony fishes are sensitive to ammonia (Cheng et al., 2015; Ren et al., 2016; Xing et al., 2016), and there are many reports on how the toxic concentration and mechanism of free ammonia affects fishes (Zhang et al., 2018; Kr et al., 2019). Exposure of fishes to high ammonia can lead to ammonia poisoning; typical symptoms include decreased growth performance (Elshafai et al., 2004), liver tissue lesions, immunosuppression, hyperactivity, coma, and even death (Benli et al., 2008). In this study, we found that under acute ammonia stress, hybrid groupers appeared to have local hyperemia, and some stress behaviors appeared in the early stage of poisoning, including manic restlessness, frequent collision or rollover, shortness of breath, and increased frequency of gill flap agitation. As the ammonia stress continued and the ammonia concentration rose, the activity of the fish body decreased, respiration was blocked, activity was slow, the secretion of body surface mucus increased, and spasms occurred at the verge of death. This was consistent with ammonia poisoning in other fishes (Rodrigues et al., 2007). At the early stage of ammonia stress, the groupers’ rapid respiration and increased exercise were due to the rapid increase in blood ammonia concentration caused by the high concentration of environmental ammonia (Zhang and Wood, 2009), which affected the blood–oxygen transport capacity of the groupers, so it is necessary to improve the air exchange rate of the body to a certain extent (Cheng and Chen, 1998). The results of this study showed that the acute toxicity of ammonia to the groupers was linear with ammonia concentration and stress duration. With the increase in ammonia concentration, mortality increased sharply and the time to death shortened. Similarly, when the stress time was prolonged, the toxic effects of ammonia increased and the LC₅₀ decreased. This suggests that a high concentration of ammonia stress has a significant effect on the immunity of aquatic organisms; in particular, long-term high-concentration stress will lead to a decline in immunity.

Acute ammonia stress can not only cause ammonia poisoning in fish, but also change biochemical indexes of fish. AST and ALT are important transaminases widely found in animals; they play an important role in amino acid synthesis and catabolism. The liver is the main gland of nutrient digestion in animal bodies and also the main site of urea synthesis, with a detoxification function. Under normal conditions, there is only a small amount of AST and ALT in serum. When the membranes of hepatocytes and mitochondria are damaged, the membrane permeability increases, then AST and ALT are released into the blood (Feng et al., 2016). In this study, the activity of serum AST and ALT increased after ammonia stress, and the activity of liver AST and ALT in the 24 h stress group was markedly higher than those in the other experimental groups, which suggested that the two enzymes were induced by stress, and their activities increased, which enhanced the excretion of ammonia. When the experimental groupers were ammonia-stressed for 48–96 h, AST and ALT in their serum were markedly higher than those in the control group, but ALT in the liver was not markedly different from that in the control group. However, in the 48 and 72 h groups, ALT was markedly lower than that in the control group. The reason may be that with the prolongation of ammonia stress duration, the hepatocytes were damaged.

COR is a kind of hormone produced by the adrenal gland and was produced in a large amount during the stress reaction. Generally, COR enters the blood directly and is not stored; therefore, the content of COR in the blood can be used as an index to measure the stress intensity of fish (Hsieh et al., 2003). The results showed that the content of COR in the serum of each experimental group was markedly higher than that before the stress, which may indicate that the free ammonia in the water entered the organism during the stress process of ammonia, making the fish produce a stress response.

GLU is an important energy substance for many kinds of tissue metabolism, and stable blood GLU concentration plays an important role in maintaining the normal life activities of fishes. The secretion of COR can release GLU from the liver, which is transported to the blood for energy use in a stress response. Studies have shown that stress can lead to a marked increase in blood GLU in fishes (Yang et al., 2019); this is consistent with the findings on the 24, 72, and 96 h groups.

LD is a kind of fatigue substance produced by glycolysis of muscles when there is an insufficient oxygen supply. With the extension of stress time, the LD content increased gradually, similar to that of Atlantic salmon (Salmo salar L.) (Iversen et al., 2005) and chinook salmon (Oncorhynchus tshawytscha) (Schreck et al., 1995) after environmental stress. Sulikowski et al. (2006) showed that the presence of GLU, LD, and COR in blood can indicate the degree of stress, thus it can be concluded that the stress on the groupers was the strongest after 72 h.

Meanwhile, acute ammonia stress can stimulate non-specific immunity of fish. The immune system is divided into the specific immune system and the non-specific immune system. The non-specific immune system of fishes plays a leading role in coping with adverse environmental stress; in particular, ACP, AKP, SOD, CAT, GSH-PX, LYZ, and other enzymes play an important role in immune regulation (Ming et al., 2015). AKP is an important index of fish disease diagnosis and physiological activities (Giannini et al., 2005). ACP is an important part of lysosome, which can ensure that the fish is not invaded by pathogens (Yan et al., 2019). SOD and CAT are the enzymes of oxidative metabolism in the body; they work together to make the production and elimination of free radicals reach a dynamic balance and reduce oxidative damage (Li et al., 2007). They are important protective enzyme systems in the body (Martinezalvarez et al., 2005). GSH-PX is a ubiquitous enzyme that catalyzes the decomposition of H₂O₂, thus protecting the integrity of cell membrane structures and functions. LYZ is involved in a variety of metabolic processes and is an indicator of the health and immune status of aquatic animals (Liu et al., 2011). MDA is the end product of lipid oxidation; the MDA content in the body shows the degree of oxidative damage of lipid by reactive oxygen species and free radicals (Peng et al., 2009). The results of this study show that AKP in all stress groups and ACP in the 24, 72, and 96 h groups were higher than that in the control group. This may be related to the marked increase in MDA content after 24–72 h of stress. Meanwhile, serum SOD and LYZ markedly decreased after stress, and GSH-PX markedly decreased under high concentrations of acute stress. This shows that the activity of antioxidant enzymes was inhibited under the stress of a high concentration of ammonia or a low concentration of ammonia for a long time, so that the antioxidant capacity of the groupers’ bodies was reduced, and the bodies could not effectively remove the oxygen free radicals produced over time. This led to oxygen free radicals accelerating lipid peroxidation and increasing MDA. As time passed, the ammonia accumulation in the groupers became more toxic, the defense system of free radicals was destroyed, and the concentration of hydrogen peroxide and singlet oxygen increased, resulting in a rapid decline in enzyme activity and the imbalance of the immune system. Therefore, the activity of SOD and CAT was markedly lower than that of the 24 h stress group. Meanwhile, the levels of liver CAT, SOD, and GSH-PX were markedly higher in the 24 h group than those in other groups; this was consistent with the results of other stress studies (Burmester et al., 2012). Those levels were markedly lower at 48 and 72 h of ammonia stress than those in other groups. The reason may be that under this condition, the level of active oxygen free radicals and the content of MDA increased, and the physiological regulation of external stress to maintain the balance of the antioxidant level of the body tissue led to a compensatory increase in the activity of antioxidant enzymes (Liu et al., 2018a). In an LC, the extension of stress duration may have caused some damage to the antioxidant system, which showed that relative to the control group, MDA content increased and antioxidant enzyme activity decreased in the 48 and 72 h groups. When fishes are exposed to high concentrations of environmental ammonia, many immunosuppressive factors are produced in the lymphoid tissues (Barton and Iwama, 1991). These factors usually lead to a reduction about immune response of lymphocytes and phagocytes (Mock and Peters, 1990). This is consistent with the increase in IgM content in this study after 24 h of stress. However, ammonia stress for a long time may have an irreversible toxic effect on a grouper, damaging its normal physiological function. Its body’s ability to regulate ammonia stress becomes weaker and weaker. The accumulated toxicity of ammonia in the body increases and the defense system of free radicals is damaged, resulting in an immune system imbalance (Liu et al., 2018a). These groupers also show a decrease in IgM and C₄ in the liver. This is consistent with the immune suppression of rainbow trout (Oncorhynchus mykiss) (Vedel et al., 1998), Atlantic halibut (Hippoglossus hippoglossus) (Paust et al., 2011), and darkbarbel catfish (Pelteobagrus fulvidraco) (Li et al., 2014) after acute or periodic ammonia stimulation.

In addition, the digestive enzyme activity of intestine was also affected by water environment. The activity of digestive enzymes can reflect the basic characteristics of the digestive physiology of aquatic animals (Lundstedt et al., 2004). The digestive systems of fishes are highly sensitive to reactive oxygen species, and an increase in their digestive enzyme activity is related to their synthesis and secretion of enzymes (Gisbert et al., 2004). In the present study, stress in the 24–72 h groups markedly decreased the activity of three digestive enzymes, but after 96 h of stress, LPS and AMS increased markedly compared with that of the other experimental groups, which suggests that a low concentration of ammonia stress can increase the activity of digestive enzymes, and a high concentration of ammonia stress can inhibit that activity (Liu et al., 2018b). It was speculated that this may be because a grouper adapts better to a low concentration of ammonia over a long time than to a high concentration for a short time.

Apoptosis induced by ammonia poisoning in hybrid groupers may be related to oxidative stress and activation of inflammatory factors. Previous studies have found that hyperglycemia caused brain astrocytes and microglia to swell and produce metabolites by activating N-Methyl-D-aspartic acid receptors and Ca²⁺. Depending on the pathway of nitric oxide synthetase, hyperglycemia can produce many free radicals, accelerate protein oxidation and lipid peroxidation, accumulate aldehydes and ketones, and cause oxidative damage to the body (Rao and Norenberg, 2012). It also releases many pro-inflammatory factors into the blood, causing an inflammatory reaction and apoptosis (Mckenna et al., 2006). A study of Litopenaeus vannamei showed that Microcystis aeruginosa and microcystin-LR exposure can induce damage from exposure and increased oxidative stress and inflammatory reactions (Duan et al., 2020). However, there have been few studies on oxidative damage, inflammatory response, and apoptosis caused by ammonia poisoning in fishes.

Ammonia-poisoned fishes can incur apoptosis in many ways, and p53 is an important regulator of apoptosis (Soengas et al., 1999). In this study, it was found that the mRNA expression of the p53 gene was markedly up-regulated in the 24, 72, and 96 h stress groups. This was consistent with the fact that free radical production induced by ammonia stress can up-regulate the expression of p53 (Zhang et al., 2018), which suggests that p53 could be involved in ammonium-induced apoptosis. Also, p53 can induce apoptosis by up-regulating the transcription of bcl2-associated X (Bax) (Zeng et al., 2014). Based on the results of this study, it was speculated that apoptosis also might be induced by a p53–Bax pathway in groupers after ammonia poisoning (Zhang et al., 2018). Bax is an important cell death promoting gene in Mammalia and fish that can induce the release of cytochrome c, leading to casp activation (Wei et al., 2001), which is an effective indicator of stress-induced apoptosis in aquatic animals (Fulda and Debatin, 2006). In the present study, casp mRNA expression was markedly up-regulated after stress, except in the 48 h stress group, which further suggested that casp dependent apoptosis may be related to ammonia-poisoning-induced apoptosis in grouper liver cells (Chen et al., 2006). It has been reported that free radicals and inflammation may also participate in apoptosis (Chen et al., 2006). Ammonia stress markedly up-regulated the expressions of Hsp70, Hsp90, Hsc70, thioredoxin (Trx), Prx-4, Prx-6, FI, SOD, CAT, and glutathione peroxidase (GPX) mRNA, which occurred to cope with oxidative stress. An increase in oxidative stress is one of the effects of environmental pollution on aquatic animals (Min et al., 2018). Oxidative stress leads to not only apoptosis but also inflammation (Kim et al., 2011). An inflammatory response is the normal defense response of the body to tissue damage and is an indispensable part of tissue repair (Seyan et al., 2010). In this study, it was found that ammonia stress markedly up-regulated the expressions of TNF-α, IL-1β, IL-6, and IL-8 mRNA, which were consistent with studies of rainbow trout (Vedel et al., 1998), Atlantic halibut (Paust et al., 2011), and darkbarbel catfish (Li et al., 2014). The activation of inflammatory cytokines can activate an NF-κB channel and apoptosis (Lam et al., 2011). NF-κB acts as a messenger in enhancing an environmental stress response and activates a series of genes related to inflammation, including IL and TNF (Jia et al., 2014). These results suggested that the apoptosis of groupers under ammonia stress may be related to oxidative stress and an inflammatory response.

Food intake (Jiang et al., 2020) and the environment (Duan et al., 2020) have a great influence on the composition of intestinal microbes. Intestinal microbes are one of the main factors affecting the metabolism and growth performance of a host (Cani and Delzenne, 2009). Studies have shown that oxidative stress and an inflammatory response are often associated with intestinal microbial imbalance (Ganz et al., 2011). In the present study, acute ammonia stress markedly changed the intestinal microbial diversity of the groupers. Intestinal microbes are closely related to host health (Xiong et al., 2015). In the present study, the intestinal microbes of the hybrid groupers were Firmicutes, Proteobacteria, and Bacteroidetes; the relative abundance of Proteobacteria and Bacteroidetes was higher after acute ammonia stress. Proteobacteria participate in the degradation of the carbon complex in an aquatic ecosystem (Klase et al., 2019), which is an indication of intestinal microbial imbalance and inflammation (Mirpuri et al., 2014). Bacteroidetes are considered as polymeric organic matter degraders (Liu et al., 2019). Therefore, groupers may alleviate oxidative stress and inflammatory reactions by increasing the abundance of intestinal dominant bacteria. After ammonia stress, the abundance of Fusobacteria increased markedly. Fusobacteria can activate the host’s inflammatory response, thus protecting the host from pathogens that promote tumor growth (Kelly et al., 2018). This finding is consistent with that in the present study. RDA–CCA analysis of environmental factors and intestinal microbes showed that Cloacibacterium, Escherichia_Shigella, Ancylostoma, and Bacteroides were negatively correlated with ammonia and that Bacillus, Enterococcus, Clostridium sensu stricto 1, Ochrobactrum, and Camellia_oleifera were positively correlated with ammonia, which also suggests that environmental factors played an important role in the stabilization of intestinal microorganisms.