Nanocomposite of Half-Fin Anchovy Hydrolysates/Zinc Oxide Nanoparticles Exhibits Actual Non-Toxicity and Regulates Intestinal Microbiota, Short-Chain Fatty Acids Production and Oxidative Status in Mice

The nanocomposite of half-fin anchovy hydrolysates (HAHp) and zinc oxide nanoparticles (ZnO NPs) (named as HAHp(3.0)/ZnO NPs) demonstrated increased antibacterial activity compared to either HAHp(3.0) or ZnO NPs as per our previous studies. Also, reactive oxygen species (ROS) formation was detected in Escherichia coli cells after treatment with HAHp(3.0)/ZnO NPs. The aim of the present study was to evaluate the acute toxicity of this nanocomposite and to investigate its effect on intestinal microbiota composition, short-chain fatty acids (SCFAs) production, and oxidative status in healthy mice. The limit test studies show that this nanoparticle is non-toxic at the doses tested. The administration of HAHp(3.0)/ZnO NPs, daily dose of 1.0 g/kg body weight for 14 days, increased the number of goblet cells in jejunum. High-throughput 16S ribosomal RNA gene sequencing of fecal samples revealed that HAHp(3.0)/ZnO NPs increased Firmicutes and reduced Bacteriodetes abundances in female mice. Furthermore, the microbiota for probiotic-type bacteria, including Lactobacillus and Bifidobacterium, and SCFAs-producing bacteria in the Clostridia class, e.g., Lachnospiraceae_unclassified and Lachnospiraceae_UCG-001, were enriched in the feces of female mice. Increases of SCFAs, especially statistically increased propionic and butyric acids, indicated the up-regulated anti-inflammatory activity of HAHp(3.0)/ZnO NPs. Additionally, some positive responses in liver, like markedly increased glutathione and decreased malonaldehyde contents, indicated the improved oxidative status. Therefore, our results suggest that HAHp(3.0)/ZnO NPs could have potential applications as a safe regulator of intestinal microbiota or also can be used as an antioxidant used in food products.


Introduction
ZnO, generally regarded as a safe (GRAS) material by the United States Food and Drug Administration (USFDA, 21CFR182.8991), is the most widely used nanoparticle [1]. The antimicrobial property of ZnO nanoparticles (NPs) makes them a viable approach to prevent infectious diseases [2]. To date, ZnO NPs have been developed as functional nanofillers that are incorporated into packaging and food contact materials to inhibit target bacteria [3,4]. However, it is also known that NPs might pose deleterious effects to living beings. In a study by Wang et al. (2006), the gastrointestinal (GI) CK, normal control. Each group contained 10 female (F) and 10 male (M) mice. Different superscript letters in the same column indicate significant differences between two samples (p < 0.05).

Histopathology
After continuous oral administration of HAHp(3.0)/ZnO NPs for 14 days, the morphological changes of the jejunum were observed with hematoxylin-eosin (HE) stain ( Figure 1). In the female normal control CK(F) and male normal control CK(M) groups, typical microvilli lined by enterocytes bearing brush border membranes and intestinal crypt regions having different types of cells can be seen in Figure 1a,b. The gavage administration of HAHp(3.0)/ZnO NPs in mice for 14 days at a dose of 1.0 g/kg BW did not lead to marked degeneration and desquamation of the intestinal villi and crypt regions of the jejunum. However, a profoundly increased number of goblet cells were observed in microvilli lines after HAHp(3.0)/ZnO NPs administration (Figure 1c,d).

Species Abundance and Diversity
A total of 649,051 valid sequences were determined for the 20 specimens. The average number of sequences of each specimen was more than 30,000 sequences. Of the sequence detected, 100% were >400 bp in length. The plateau phase of rarefaction curve suggests the adequate sequencing reads in the experiment [35]. Obviously, most specimens in the rarefaction curves reached the plateau phases (Figure 2a and 14th days, the observed OTUs in all of the Z specimens were higher than the same sex specimens in the CK groups. After continuous oral administration of HAHp(3.0)/ZnO NPs for 14 days, the morphological changes of the jejunum were observed with hematoxylin-eosin (HE) stain ( Figure 1). In the female normal control CK(F) and male normal control CK(M) groups, typical microvilli lined by enterocytes bearing brush border membranes and intestinal crypt regions having different types of cells can be seen in Figure 1a,b. The gavage administration of HAHp(3.0)/ZnO NPs in mice for 14 days at a dose of 1.0 g/kg BW did not lead to marked degeneration and desquamation of the intestinal villi and crypt regions of the jejunum. However, a profoundly increased number of goblet cells were observed in microvilli lines after HAHp(3.0)/ZnO NPs administration (Figure 1c,d).

Species Abundance and Diversity
A total of 649,051 valid sequences were determined for the 20 specimens. The average number of sequences of each specimen was more than 30,000 sequences. Of the sequence detected, 100% were >400 bp in length. The plateau phase of rarefaction curve suggests the adequate sequencing reads in the experiment [35]. Obviously, most specimens in the rarefaction curves reached the plateau phases (Figure 2a), confirming the adequate sequencing depth of the experiment. The observed OUTs numbers between the CK and the HAHp(3.0)/ZnO NPs treated (Z) specimens were further compared, shown in Figure 2b. Significant differences were observed in CK(F) and CK(M) at 0 day. At the seventh and 14th days, the observed OTUs in all of the Z specimens were higher than the same sex specimens in the CK groups. Similar to the results shown in Figure 2b, the CK groups showed significant differences in the Chao1 indices between the female and male specimens at 0 day (p < 0.05), and consistent tendencies with the results shown in Figure 2b were observed at 7 day and 14 day ( Figure 2c). A greater Shannon index indicates a greater diversity of intestinal microbiota [36]. The Shannon indices of all of the specimens changed with the feed time. The Shannon index of Z(M) at 7 day was 4.17 ± 0.14 and it increased to 4.51 ± 0.05 at 14 day. In comparison, the Shannon index of Z(F) was dramatically increased from 3.18 ± 0.05 at 7 day to 4.58 ± 0.02 at 14 day (p < 0.05) (Figure 2d). Similar to the results shown in Figure 2b, the CK groups showed significant differences in the Chao1 indices between the female and male specimens at 0 day (p < 0.05), and consistent tendencies with the results shown in Figure 2b  index indicates a greater diversity of intestinal microbiota [36]. The Shannon indices of all of the specimens changed with the feed time. The Shannon index of Z(M) at 7 day was 4.17 ± 0.14 and it increased to 4.51 ± 0.05 at 14 day. In comparison, the Shannon index of Z(F) was dramatically increased from 3.18 ± 0.05 at 7 day to 4.58 ± 0.02 at 14 day (p < 0.05) (Figure 2d).

Principal Coordinates Analysis (PCoA)
As shown in Figure 3, PC1, PC2, and PC3 accounted for 34.3%, 19.33%, and 14.77% of the variation, respectively. In the scatter of PC1 and PC2 (Figure 3a), the CK specimens at 0 day were distinguished from the CK specimens at 14 day. This indicated that the distributions of genera in the CK were changed with the feed time. Furthermore, the female specimens of HAHp(3.0)/ZnO NPs at 14 day (14d_Z_1 and 14d_Z_2) were separated from the female ones of CK (14d_CK_1 and 14d_CK_2) in the scatters of PC1 and PC2 (Figure 3a), PC1 and PC3 (Figure 3b), and PC2 and PC3 (Figure 3c). These results suggested that the administration of HAHp(3.0)/ZnO NPs could result in changes of intestinal microbiota at the genus level for female mice. By comparison, the overlap of male specimens at 14 day between CK (14d_CK_3 and 14d_CK_4) and HAHp(3.0)/ZnO NPs (14d_Z_3 and 14d_Z_4) indicated their similar intestinal microbiota composition. As shown in Figure 3, PC1, PC2, and PC3 accounted for 34.3%, 19.33%, and 14.77% of the variation, respectively. In the scatter of PC1 and PC2 (Figure 3a), the CK specimens at 0 day were distinguished from the CK specimens at 14 day. This indicated that the distributions of genera in the CK were changed with the feed time. Furthermore, the female specimens of HAHp(3.0)/ZnO NPs at 14 day (14d_Z_1 and 14d_Z_2) were separated from the female ones of CK (14d_CK_1 and 14d_CK_2) in the scatters of PC1 and PC2 (Figure 3a), PC1 and PC3 (Figure 3b), and PC2 and PC3 (Figure 3c). These results suggested that the administration of HAHp(3.0)/ZnO NPs could result in changes of intestinal microbiota at the genus level for female mice. By comparison, the overlap of male specimens at 14 day between CK (14d_CK_3 and 14d_CK_4) and HAHp(3.0)/ZnO NPs (14d_Z_3 and 14d_Z_4) indicated their similar intestinal microbiota composition. ). However, the relative abundance of phylum Firmicutes in 14 day of Z(F) (14 d Z(F)) was 68.32 ± 5.98%, dramatically higher than other specimens (p < 0.05). In contrast, a markedly decreased relative abundance of phylum Bacteroidetes (16.98 ± 5.63%) was found in 14 d Z(F) (p < 0.05).
The intestinal microbiota composition of 14 d CK(F) (14d_CK_1 and 14d_CK_2) and 14 d Z(F) (14d_Z_1 and 14d_Z_2) specimens was further evaluated at the genus level. As shown in Figure 4c, more than 38 bacteria at the genus level were identified in these specimens, of which Bacteroidales_S24−7_group_norank and Lactobacillus were the two most abundant bacteria communities in each specimen. The relative abundance shifts of 37 dominant bacteria between 14 d CK(F) and 14 d Z(F) at genus level plus Bifidobacteria, one of most important probiotic species of the intestinal tract that is associated with a healthy status in animals or humans [37], are listed in Table 2.

Composition of the Intestinal Microbiota at the Phylum and Genus Levels
A total of 15 intestinal microflora were identified at the phylum level in the specimens (Figure 4a), of which the phyla Firmicutes, Bacteroidetes, Proteobacteria, Saccharibacteria, and Actinobacteria were most abundant, accounting for >97% of the total sequences. The other phyla had comparatively small numbers. The relative abundance of the five dominant phyla at 14 day was compared between the CK and Z groups ( Figure 4b). Obviously, Firmicutes and Bacteroidetes were the two most dominant phyla in the specimens. At 14 day, the relative abundance of phylum Firmicutes were 42.89 ± 5.85% for 14 day of CK(F)(14 d CK(F)), 48.40 ± 1.83% for 14 day of CK(M) (14 d CK(M)), and 45.35 ± 5.89% for 14 day of Z(M) (14 d Z(M)). However, the relative abundance of phylum Firmicutes in 14 day of Z(F) (14 d Z(F)) was 68.32 ± 5.98%, dramatically higher than other specimens (p < 0.05). In contrast, a markedly decreased relative abundance of phylum Bacteroidetes (16.98 ± 5.63%) was found in 14 d Z(F) (p < 0.05).
The intestinal microbiota composition of 14 d CK(F) (14d_CK_1 and 14d_CK_2) and 14 d Z(F) (14d_Z_1 and 14d_Z_2) specimens was further evaluated at the genus level. As shown in Figure 4c, more than 38 bacteria at the genus level were identified in these specimens, of which Bacteroidales_S24−7_group_norank and Lactobacillus were the two most abundant bacteria communities in each specimen. The relative abundance shifts of 37 dominant bacteria between 14 d CK(F) and 14 d Z(F) at genus level plus Bifidobacteria, one of most important probiotic species of the intestinal tract that is associated with a healthy status in animals or humans [37], are listed in Table 2.     The microbes of groups 14 d CK(F) and 14 d Z(F) were derived from six different phyla, 10 different classes, 12 different orders, and 18 different families. Although an increase in the phylum Firmicutes was noted in 14 d Z(F) (as shown in Figure 4b), increase and decrease of particular taxa were also observed in the Bacilli, Clostridia, and Erysipolotrichi classes ( Table 2).
In the Bacilli class, an increase in the relative abundance of genus Lactobacillu was observed in 14 d Z(F) ( Table 2). Specific members including Lachnospiraceae_unclassified, Lachnospiraceae _UCG-001, Ruminococcaceae_UCG-014, and Ruminococcus_1, were markedly increased in the Clostridia class. Alternatively, some members of this class such as Acetitomaculum and [Eubacterium]_coprostanoligenes _group were reduced by the administration of HAHp(3.0)/ZnO NPs. In the Erysipelotrichia class, the relative abundance of the genus Faecalibaculum was as small as 0.13% in 14 d CK(F) and increased to 5.95% in 14 d Z(F).
Within the phylum Bacteroidetes, the relative abundance of all of the members of four different families, including Bacteroidaceae, Bacteroidales_S24-7_group, Porphyromonadaceae, and Rikenellaceae were declined, and significant changes were found for the genera Bacteroides, Bacteroidales_S24-7_group_norank, Alistipes, Rikenella, and Rikenellaceae_RC9_gut_group. In phylum Actinobacteria, the relative abundance of genera Enterorhabdus and Bifidobacterium were dramatically increased. In addition to significant shifts of some genera of the phyla Firmicutes, Bacteroidetes, and Actinobacteria, a markedly increased relative abundance of the genus Vibrio (phylum Proteobacteria) was observed in 14 d Z(F).

Function Prediction, SCFAs and Amino Acid Concentrations in Feces
Since different species of intestinal microbiota might perform same function, the function differences among individual feces can be discovered through metagenomic study. As shown in Figure 5a, the relative abundances of five categories, namely, function D (Cell cycle control, cell division, chromosome partitioning), function G (Carbohydrate transport and metabolism), function K (Transcription), function R (General function prediction only) and function V (Defense mechanisms) were markedly increased in 14 d Z(F) (p < 0.05). In addition to significantly increased functions, the functions of I (Lipid transport and metabolism), M (Cell wall/membrane/envelope biogenesis), P (Inorganic ion transport and metabolism) and U (Intracellular trafficking, secretion, and vesicular transport) were remarkably decreased in 14 d Z(F) (p < 0.05) (Figure 5a). An evident increase of SCFAs concentrations was observed in the 14 d Z(F) group, and significant increases were found in propionic and butyric acids (p < 0.05) (

Changes of Liver Oxidative Status
No significant differences were found in the total superoxide dismutase (SOD) activity (p > 0.05) (Figure 6a). Similar results were noted for glutathione peroxidase (GPx) and catalase (CAT) activities (Figure 6b,c), though decreased activities were observed in 14 d Z(F). However, the content of glutathione (GSH) was increased to 0.58 μmol/gprot in 14 d Z(F), markedly higher than that in the 14 d CK(F) (0.39 μmol/gprot) (p < 0.05) (Figure 6d). In contrast, the malonaldehyde (MDA) content of 14 d Z(F) was dramatically reduced to 3.63 nmol/mgprot, significantly lower than that in the 14 d CK(F) (6.15 nmol/mgprot) (p < 0.05) (Figure 6e). The total sulfhydryl (SH) content, related to GSH level and/or used as substrates for GSH biosynthesis in vivo, were further determined in HAHp(3.0)/ZnO NPs, as well as the concentrations of protein and peptide were measured (Figure 6f). Interestingly, HAHp(3.0)/ZnO NPs had significantly increased total SH content compared to HAHp(3.0) (p < 0.05). Meanwhile, decreased protein and increased peptide concentrations were detected in HAHp(3.0)/ZnO NPs.

Changes of Liver Oxidative Status
No significant differences were found in the total superoxide dismutase (SOD) activity (p > 0.05) (Figure 6a). Similar results were noted for glutathione peroxidase (GPx) and catalase (CAT) activities (Figure 6b,c), though decreased activities were observed in 14 d Z(F). However, the content of glutathione (GSH) was increased to 0.58 µmol/gprot in 14 d Z(F), markedly higher than that in the 14 d CK(F) (0.39 µmol/gprot) (p < 0.05) (Figure 6d). In contrast, the malonaldehyde (MDA) content of 14 d Z(F) was dramatically reduced to 3.63 nmol/mgprot, significantly lower than that in the 14 d CK(F) (6.15 nmol/mgprot) (p < 0.05) (Figure 6e). The total sulfhydryl (SH) content, related to GSH level and/or used as substrates for GSH biosynthesis in vivo, were further determined in HAHp

Discussion
In the present study, the acute oral toxicity of HAHp(3.0)/ZnO NPs with zinc content of 912.74 mg/kg BW was consistent with a recent report by Jayarambabu, Rao and Rajendar (2018); they described that the ZnO NPs, which were synthesized by a green method using Lawsonia inermis plant extract, showed no mortality at 2000 mg/kg BW in an acute oral toxicity study [38]. At a dose of 1.0 g/kg BW for 14 days, a profoundly increased number of goblet cells were observed in microvilli lines. It is well known that goblet cells secrete mucins and play an important role in the innate immunological defense [39]. The positive alteration of intestinal microbiota could interfere with the production of mucus, and finally contribute to the exertion of protective effects on the integrity of the mucosal barrier function [40,41]. The increased number of goblet cells (as shown in Figure 1c,d) should be associated with the changes of intestinal microbiota composition induced by HAHp(3.0)/ZnO NPs administration. Furthermore, the results of Chao1 index and Shannon index also verified that species abundance and diversity were changed with feed time. In the PCoA analysis, the more similar two specimens are in composition in genus level, the closer they are on the PCoA diagram [36]. The result of Figure 3 revealed that the administration of HAHp(3.0)/ZnO NPs resulted in intestinal microbiota changes in mice, especially in female mice.
Dramatically increased Firmicutes and decreased Bacteroidetes were found in 14 d Z(F) compared to 14 d CK(F) or 14 d Z(M). According to some documents, in obese hosts, the relative abundance of phylum Firmicutes tends to increase, while the relative abundance of Bacteroidetes tends to decrease [42]. However, in this study, increased Firmicutes and decreased Bacteroidetes in 14 d Z(F) did not result in overweight mice (as seen in Table 1). Similar to our results, an increase in Firmicutes and a reduction in Bacteroides were observed in the immunological improvement of human fecal samples in response to the addition of whole grains to the diet [43]. In the latest study by Byerley et al. (2017), they also reported walnut consumption increased the abundance of Firmicutes and reduced the abundance of Bacteriodetes in rats, and they speculated that the reshaping of the gut microbial community may play a physiological role in promoting walnut's health benefits [36]. The ratio of Firmicutes to Bacteroidetes (F/B) is commonly used to indicate the composition of the gut microbiota [35]. In the present study, the average F/B ratio for 14 d Z(F) was

Discussion
In the present study, the acute oral toxicity of HAHp(3.0)/ZnO NPs with zinc content of 912.74 mg/kg BW was consistent with a recent report by Jayarambabu, Rao and Rajendar (2018); they described that the ZnO NPs, which were synthesized by a green method using Lawsonia inermis plant extract, showed no mortality at 2000 mg/kg BW in an acute oral toxicity study [38]. At a dose of 1.0 g/kg BW for 14 days, a profoundly increased number of goblet cells were observed in microvilli lines. It is well known that goblet cells secrete mucins and play an important role in the innate immunological defense [39]. The positive alteration of intestinal microbiota could interfere with the production of mucus, and finally contribute to the exertion of protective effects on the integrity of the mucosal barrier function [40,41]. The increased number of goblet cells (as shown in Figure 1c,d) should be associated with the changes of intestinal microbiota composition induced by HAHp(3.0)/ZnO NPs administration. Furthermore, the results of Chao1 index and Shannon index also verified that species abundance and diversity were changed with feed time. In the PCoA analysis, the more similar two specimens are in composition in genus level, the closer they are on the PCoA diagram [36]. The result of Figure 3 revealed that the administration of HAHp(3.0)/ZnO NPs resulted in intestinal microbiota changes in mice, especially in female mice.
Dramatically increased Firmicutes and decreased Bacteroidetes were found in 14 d Z(F) compared to 14 d CK(F) or 14 d Z(M). According to some documents, in obese hosts, the relative abundance of phylum Firmicutes tends to increase, while the relative abundance of Bacteroidetes tends to decrease [42]. However, in this study, increased Firmicutes and decreased Bacteroidetes in 14 d Z(F) did not result in overweight mice (as seen in Table 1). Similar to our results, an increase in Firmicutes and a reduction in Bacteroides were observed in the immunological improvement of human fecal samples in response to the addition of whole grains to the diet [43]. In the latest study by Byerley et al. (2017), they also reported walnut consumption increased the abundance of Firmicutes and reduced the abundance of Bacteriodetes in rats, and they speculated that the reshaping of the gut microbial community may play a physiological role in promoting walnut's health benefits [36]. The ratio of Firmicutes to Bacteroidetes (F/B) is commonly used to indicate the composition of the gut microbiota [35]. In the present study, the average F/B ratio for 14 d Z(F) was 4.32, significantly greater (>4.2-fold) than that of the corresponding 14 d CK(F) (F/B = 1.02) group. We assumed the changes of intestinal microbiota composition, especially for the phyla Firmicutes and Bacteroidetes, could be responsible for the distinctive PCoA plot profiles between 14 d CK(F) and 14 d Z(F) (as shown in Figure 3).
Although an increase in the phylum Firmicutes was noted in 14 d Z(F) (as shown in Figure 4b), particular taxa increase and decrease were also observed within the phylum Firmicutes. Significant changes were observed in the Bacilli, Clostridia, and Erysipolotrichi classes. In the Bacilli class, the Lactobacillaceae family is known to participate in the breakdown of carbohydrates into SCFAs [29]. The positive functions of various strains of lactic acid bacteria (LAB) derived from the Lactobacillaceae family have attracted great attention in the last few years.  reported that two species of Lactobacillus, Lactobacillus G15 and Q14, which are used in the preparation of Chinese traditional fermented dairy foods, alleviated type 2 diabetes in a gut microbiota-dependent way via down regulating G − bacteria-related lipopolysaccharide secretion and promoted the enrichment of SCFA-producing bacteria [44]. Yu and Li (2016) gathered the studies of lactic acid production of LAB that could contribute to the prevention of colon cancer and modulate host immunity, anti-inflammatory, and antipathogenic activity [45]. Similar to our results, Yu, Amorim, Marques, Calhau and Pintado (2016) found the relative abundance of Lactobacillus spp. was significantly increased by whey peptide extract in rats fed with a standard diet [46]. In the present study, a significant increase in the relative abundance of genus Lactobacillus in 14 d Z(F) (as seen in Table 2) suggested potential benefits could be expected by consuming enriched HAHp(3.0)/ZnO NPs.
The proportions of bacteria in the Clostridia class were 16.48% for 14 d CK(F) and 32.19% for 14 d Z(F) based on their relative abundances in Table 2. The Clostridia class is known for its production of butyrate [36], which has multiple functions as a cellular mediator, including gut cell regulation, gut tissue development, cell differentiation, immune modulation, oxidative stress reduction, inactivation of carcinogens, as well as acting as a primary nutrient to provide energy for colonocytes [47]. The results shown in Table 2 revealed that the administration of HAHp(3.0)/ZnO NPs contributed to a higher amount of bacterial counts in the Clostridia class, which could result in an increased production of butyrate and promote positive effects on mice. Similarly, in the Erysipelotrichia class, the relative abundance of the genus Faecalibaculum was 0.13% in 14 d CK(F) in contrast with an increase to 5.95% in 14 d Z(F). Faecalibacterium was characterized by the anti-inflammatory activity within the gut [28]. Therefore, an increase in this genus indicated an elevated anti-inflammatory activity induced by the administration of HAHp(3.0)/ZnO NPs.
Although a reduction in Bacteroides abundance is assumed to be associated with some metabolic diseases, such as obesity and diabetes [30,31], the enriched microbiota for probiotic-type bacteria, including Lactobacillus, Ruminococcaceae, and Roseburia, in rat feces was induced by intake of walnuts, while it was significantly reduced in Bacteroides and Anaerotruncus [36]. Similarly, in the present study, the increased relative abundances of Lactobacillus, Ruminococcaceae, and Roseburia were found in 14 d Z(F), as well as a reduced relative abundance of Bacteroides (Table 2). Furthermore, a decrease in Alistipes is considered to reduce the frequency of abdominal pain in patients with irritable bowel syndrome, and thus decrease gut inflammation [48]. The decrease in the relative abundance of Alistipes in 14 d Z(F) also suggested its contribution to the anti-inflammatory function.
In phylum Actinobacteria, the relative abundance of genera Enterorhabdus and Bifidobacterium were dramatically increased. In particular, the genus Bifidobacterium had as high as 53-fold relative abundance in 14 d Z(F) after administration of HAHp(3.0)/ZnO NPs. The genus Bifidobacterium can secrete lacto-N-biosidase to digest oligosaccharides, such as galactooligosaccharide, to produce SCFAs [49,50]. A number of studies have proven the beneficial effects of the Bifidobacterium species on the immune system and allergic diseases later in life [51,52]. However, the number of Bifidobacterium dominant in an infant's gut microbiota, decreased with solid food consumption and old age in animals and humans [53]. The enriched genus Bifidobacterium in 14 d Z(F) gut microbiota should be attributed to the prebiotic effect induced by HAHp(3.0)/ZnO NPs administration.
In addition to significant shifts of some genera of the phyla Firmicutes, Bacteroidetes, and Actinobacteria, a markedly increased relative abundance of the genus Vibrio (phylum Proteobacteria) was observed in 14 d Z(F). In the genus Vibrio, there are over 100 individual species, and only a few of them have been associated with human illness [54]. In fact, specific bacteria of Vibrio demonstrated positive effects, for instance, Liu, Zhang, Zhang, and Li (2017) recently reported the bacterium Vibrio sp. V33 isolated from healthy sepia showed strong antagonistic activity towards the pathogenic isolate Vibrio splendidus [55]. During the oral administration of HAHp(3.0)/ZnO NPs in mice, we did not observe treatment-related clinical signs of toxicity or mortality in 14 d Z(F). We speculated that the increase in genus Vibrio might play weak roles in the health of the mice.
In the present study, the potential microbial  Figure 1c), which could facilitate the innate immunological defense. Furthermore, the increase of predicted function G (Carbohydrate transport and metabolism) (Figure 5a) and the decrease of detected amino acids in 14 d Z(F) (Figure 5c) suggested more amount of SCFAs could be produced after HAHp(3.0)/ZnO NPs administration, since SCFAs are major derived from carbohydrate fermentation and degradation of proteins by the intestinal microbiota [36]. As we expected, an evident increase in the SCFAs concentrations was observed in the 14 d Z(F) group (Figure 5b). Production of SCFAs not only can be used as energy source, but also can decrease the intestinal pH values. Different scientific publications have demonstrated that increases of SCFAs content are associated with the enhanced anti-inflammatory activity in host [57,58]. Especially, the SCFAs of acetic, propionic and butyric acids have deep impacts on the physiology and nutrition of the gastrointestinal tract [59]. Though no statistical significances were found in formic and acetic concentrations between 14 d CK(F) and 14 d Z(F) groups, significant increases of propionic and butyric acids were detected in 14 d Z(F) group (as shown in Figure 5b). Propionic and butyric acids can act as attracting molecules for neutrophils by means of the interaction with the anti-inflammatory receptor, and subsequently inhibit the inflammatory response caused by diverse agents [57,60]. The increases of propionic and butyric acids in 14 d Z(F) group meant positive effects of HAHp(3.0)/ZnO NPs in regulating intestinal microbiota composition. Also, the result of SCFAs was consistent with the upregulated genera shifts in 14 d Z(F) group (as seen in Table 2).
Among these reduced amino acids (as shown in Figure 5c), the most decreased amino acid was Phe, with concentration ranging from 13.70 nmol/mL to 4.72 nmol/mL (−65.55%). This result was consisted with the fact that phenyl-containing amino acids are a priority to be metabolized by intestinal bacteria [61]. In contrast to those significantly decreased amino acids, the concentration of Cys in 14 d Z(F) remained stable. This meant the alteration of intestinal microbiota composition induced by HAHp(3.0)/ZnO NPs administration had less effect on Cys concentration in feces.
It is well known that reactive oxygen species (ROS) generation is a predominant contributor to the antibacterial activity of ZnO NPs in vitro. In our previous studies, we proved that the incubation of HAHp(3.0)/ZnO NPs with E. coli cells induced ROS formation in intracellular cells [11]. Excess ROS in vivo might cause oxidative stress which is associated with the development of oxidative damage and chronic inflammation [62]. In this study, the increased propionic and butyric acids detected in the fecal samples in HAHp(3.0)/ZnO NPs fed group (as shown in Figure 5b) suggested that HAHp(3.0)/ZnO NPs could be used as a diet supplement to improve immunity and gut motility in mice. Liver is an important organ for detoxification and metabolism in animals and humans, the oxidative stress of which might be influenced by intestinal microbiota composition changes.
In the mechanism of ROS scavenging, of the key enzymes involved, including SOD, GPx, and CAT, play crucial roles in converting ROS into H 2 O and O 2 to avoid ROS accumulation [63,64]. In the present study, the content of GSH, the most abundant low molecular weight thiol groups that protect the cells from oxidative stress damage [65], were significantly increased in 14 d Z(F) samples. Meanwhile, the MDA content, reflecting the peroxidation of polyunsaturated fatty acid in biological membrane, was decreased dramatically. In general, the MDA level is increased with aggravated oxidative stress [63]. The decreased MDA content as shown in Figure 6e indicated an increase of antioxidative status in mice liver after HAHp(3.0)/ZnO NPs administration for 14 days. Similar to our results, Tang et al. (2014) reported the supplementation with zinc-bearing clinoptilolite (ZnCP) in a basal diet for broiler chickens not only modulated microbial populations in fecal content, but also decreased the MDA content and increased the GSH content in the intestine [41]. A balance of intestinal microbiota in animals can be found to serve for various health benefits, such as modulation of redox status in the host, breakdown and synthesis of nutrients, detoxification of toxic products, prevention of infection by pathogenic organisms, and development of the host immune system [66][67][68]. Therefore, the increased populations of beneficial bacteria, such as Lactobacillus, Faecalibaculum, and Bifidobacterium, and the decreased numbers of harmful bacteria including Bacteroides and Alistipes in the treatment group of 14 d Z(F) (seen in Table 2) could play important roles in preventing hepatic GSH depletion, and consequently inhibited lipid peroxidation.
In addition, some specific peptides were able to elevate cellular GSH levels. For example, Shi et al. (2014) reported eggshell membrane peptides, digested with Alcalase and Protease S (AL-PS), and its ultrafiltered fraction AL-PS-I (Molecular weight cut off < 5 kDa) significantly suppressed the formation of H 2 O 2 -induced MDA and protein carbonyl derivatives, and increased GSH synthesis and antioxidant enzyme activity in H 2 O 2 -stimulated Caco-2 cells [69]. In our previous studies, we found the HAHp(3.0)/ZnO NPs displayed different transmission electron microscopy images compared to their initial counterparts of ZnO NPs, and a dynamic layer of "corona" formed on the surface of ZnO NPs [11]. Proteins and/or peptides with specific residues and conformations could be responsible for binding with ZnO NPs [70]. We speculated that specific groups in HAHp(3.0)/ZnO NPs may be another contributor to biosynthesis or accumulation of GSH in liver. Interestingly, according to the result from Figure 6f, HAHp(3.0)/ZnO NPs had significantly increased SH content compared to HAHp(3.0) (p < 0.05). Meanwhile, decreased protein concentrations and increased peptide concentrations were detected in HAHp(3.0)/ZnO NPs. Accordingly, we supposed that the dynamic "corona" absorbed onto the surface of ZnO NPs in HAHp(3.0)/ZnO NPs should contain large amounts of SH residues, such as Cys, GSH and/or GSH alike peptides. Since the GSH levels are highly regulated and limited by availability of the sulfhydryl amino acid L-Cys [71], the present result as seen from Figure 6f also suggested that HAHp(3.0)/ZnO NPs should be a good source of glutathione or its precursor for glutathione biosynthesis. High level of glutathione in diet is well known to act as excellent antioxidants to provide beneficial effects in vivo [72]. The relatively stable concentration of Cys in the fecal sample of 14 d Z(F) (as seen in Figure 5c) might be due to the higher SH concentration of HAHp(3.0)/ZnO NPs fed in mice. However, further studies need to investigate the adsorption kinetics of HAHp(3.0)/ZnO NPs in the intestinal tract in large animal models.

Materials
Half-fin anchovy (Setipinna taty) were purchased from Fengmao Aquatic Market in Zhoushan City, China. The extraction kits of bacterial genome DNA were bought from Beyotime Institute of Biotechnology (Haimen, China). The total SOD, GPx, CAT, GSH, MDA, and total sulfhydryl assay kits were purchased from Jiancheng Bioengineering Institute (Nanjing, China). All of the other chemicals were obtained from commercial products. Female and male Kunming species mice (clean), approximate weight 18-22 g, were purchased from the Laboratory Animal Center of Zhejiang Province, Hangzhou, China (SCXK (Zhe)2014-0001, certification number: 1612220002).

Preparation of the Nanocomposite of HAHp and ZnO NPs
The nanocomposite of HAHp and ZnO NPs were prepared according to our previous studies [11,32]. In brief, the pH of the HAHp was adjusted to 3.0 and centrifuged at 5000× g for 10 min. The supernatant was collected and designated as HAHp(3.0). The ZnO NPs were added to HAHp(3.0) to reach the final concentration of 19.78 mg/mL. After 6.9 h of incubation at 37 • C in a water bath oscillator (160 rpm), the mixtures were centrifuged at 5000× g for 10 min. The final supernatants, designated as HAHp(3.0)/ZnO NPs, were collected, freeze dried, and stored at 4 • C for future experiments. The peptide concentration of dried powder of HAHp(3.0)/ZnO NPs was 9.23 g/kg, determined by ortho-phthaldialdehyde (OPA) method [73], with GSH as the standard. The total content of zinc in the HAHp(3.0)/ZnO NPs was 9127.4 mg/kg as determined by inductively coupled plasma-optical emission method [11].

Acute Oral Toxicity Study
The acute oral toxicity of HAHp(3.0)/ZnO NPs was evaluated in mice using the limit test method according to the procedures outlined in the Procedure and Methods of Food Safety Toxicological Assessment, GB15193.3-2014 (in Chinese) (National Health and Family Planning Commission of the People's Republic of China, 2015). After acclimatization, the mice were randomly assigned to two groups. Each group contained 20 animals (10 female and 10 male). The BW of each mouse was measured prior to dosing on the first day. HAHp(3.0)/ZnO NPs conjugates were dissolved in saline water to reach the concentration of 1.0 g/mL and then administered by oral gavage at a dose of 10.0 g/kg BW. The CK mice were orally administered sterile saline (0.2 mL/kg BW). The clinical signs were monitored continuously for the first 1 h after administration. The general condition, toxic symptoms, and mortality were monitored daily for 14 days after administration. On the 15th day, all of the mice were weighed and killed under light ether anesthesia. The liver, thymus, and spleen tissues were removed and weighed. The liver coefficient, thymus coefficient, and spleen coefficient of the CK and HAHp(3.0)/ZnO NPs groups were calculated according to the following equations: Liver coefficient/(g/kg) = Liver weight (g)/(Mouse weight (g) × 1000) Thymus coefficient/(mg/10 g) = Thymus weight (mg)/(Mouse weight (g) × 1) Spleen coefficient/(mg/10 g) = Spleen weight (mg)/(Mouse weight (g) × 1)

Continuous Oral Administration of HAHp(3.0)/ZnO NPs in Mice
After getting acclimatized to the animal facility for 4 days on standard mice chow and having access to water under controlled conditions, mice were divided into two groups of 10 animals each (5 female and 5 male). Animals in the CK group were orally administered 0.2 mL of saline daily. The treated group was given a dose of 1.0 g/kg BW HAHp(3.0)/ZnO NPs daily for 14 days. The fecal samples from the first day (0 day) and the seventh day (7 day) were collected, mixed separately based on sex, and stored at −80 • C for further use. After completing the oral administration on the 14th day (14 day), the mice were not permitted to eat chow, but they were allowed to freely drink water. Animals were killed at 4 h after the last oral administration under light ether anesthesia. The small intestines were removed and processed for histopathology. The fecal samples at the time of the killing were collected aseptically from the descending colon directly, mixed together according to sex, and stored at −80 • C until DNA isolation. The liver tissues were collected after absorbing excess blood with blotting paper. Some of liver sections were blended with ice-cold saline at a ratio of 1:9 (w/v) and homogenized in an ice bath to prepare 10% liver homogenates, followed by centrifugation at 2500× g for 10 min at 4 • C. The supernatants were collected and stored at −80 • C for further oxidative status assay within 1 week.

Histopathology
The histopathology of jejunum was assessed via HE stain following a standard procedure. Briefly, paraffin sections (5 µm-thick) were cut, de-paraffinized in xylene, rehydrated in an alcohol gradient, and stained with HE. Stained sections were observed using a light microscope (CX41-12C02, Olympus, Tokyo, Japan) at 200× magnification.

DNA Isolation and Polymerase Chain Reaction (PCR) Amplification and Sequencing
The DNA of the total bacteria in the fecal samples separated based on sex at 0 day, 7 day, and 14 day were extracted according to the DNA extraction kit instructions. All of the operations strictly followed the product specifications. The total bacteria DNA of the fecal samples was used as the template for PCR amplification of the bacterial 16S rRNA. All extracted DNA were forwarded to Hangzhou Legenomics Bio-Pharm Technology Co., Ltd. (Hangzhou, China) for PCR analysis and high-throughput sequencing. The specific primers 338F (5 -ACTCCTACGGGAGGCAGCA-3 ) and 806R (5 -GGACTACHVGGGTWTCTAAT-3 ) with a barcode were used for amplification. TransStart FastPfu DNA Polymerase (TransGen AP221-02; 20 µL reaction system) was used for the PCR (Applied Biosystems ® GeneAmp ® 9700, Thermo Fisher Scientific Inc., Waltham, MA, USA). The 16S rRNA fragment V3-V4 variable region was the amplification substance of PCR. The amplification program was based on the following optimal annealing and extension conditions: 95 • C for 3 min, 25 cycles (95 • C for 30 s, 55 • C for 30 s, and 72 • C for 45 s), and 72 • C for 5 min, with a hold at 4 • C. This was repeated three times for each specimen. The PCR amplification products from the same specimen were mixed and subjected to 2% agarose gel electrophoresis. The target products were collected with gel extraction using the AxyPrep DNA Gel Extraction Kit (Axygen, Union City, CA, USA), eluted with Tris-HCl, and analyzed with 2% agarose gel electrophoresis. In order to ensure the accuracy and reliability of data analysis in the subsequent assay, PCR amplification needed to meet the following requirements: (1) used a low cycle number in amplification whenever possible; and (2) ensured that the number of cycles amplified per sample was consistent. Moreover, a representative sample was selected at random to perform preliminary experiments, and the majority of the samples could be amplified at the lowest number of cycles to produce a suitable concentration. The quantitative detection of PCR amplification products was further performed using a QuantiFluor™-ST blue fluorescence quantitative system (Promega, Madison, WI, USA). After the construction of a MiSeq library, the MiSeq high-throughput sequencing procedures were performed. The paired-end (PE) reads obtained in MiSeq sequencing were first mosaicked according to the overlap relationships. After removing the low-quality sequences, the relevant raw data were analyzed with procedures that identified and clustered the 16S rRNA sequences. The number of operational taxonomic units (OTUs, with a similarity of 97%) was calculated for each specimen.

Bioinformatics Analysis
Rarefaction curves of the OTUs were drawn based on the sequencing results, using a specific taxon to represent a particular species. Alpha-diversity was calculated on the Chao1 index (a community richness index) and the Shannon index (a community diversity index). Beta diversity (between communities) was applied using PCoA for the statistical analysis of the OTUs data. The species abundance at the phylum and genus levels was determined. Significant differences between the CK and HAHp(3.0)/ZnO NPs administration groups were compared at various taxonomic levels. The 16S rDNA sequencing data of intestinal microbiota was further analyzed by an online tool PICRUSt (http://picrust.github.io/picrust/tutorials/algorithm_description.html). Potential microbial functions were identified using eggnog functional prediction, and the 16S rRNA sequences were compared with Greengenes database to obtain COG orthology and function abundance. The relative abundance of each function was then calculated based on the total abundance.

Determination of SCFAs in Feces
Fecal samples (20 mg of freeze dried powder) were mixed with 1.2 mL of 0.1 mol/L phosphate buffered saline (PBS) (pH 6.5). The mixture was vortexed for 2 min and centrifuged at 5000× g for 5 min. The supernatant (1 mL) was collected, blended with 2.4 mL of absolute ethanol, and vortexed completely. Then, 850 µL of solution was removed to an ampere bottle, followed by addition of 1 mL of n-hexane and 100 µL of concentrated sulfuric acid. After sealed with alcohol burner, the ampere bottle was kept at room temperature for 1 h, and incubated at 60 • C water bath for 1 h. The upper layer of solutions was finally filtered through a 0.22 µm organic filter into HPLC vials.
The analysis of SCFAs in feces was performed according to the method described by Asarat, Vasiljevic, Ravikumar, Apostolopoulos and Donkor (2016) with slight modifications [74]. Briefly, the HPLC system used was an Agilent 1260 Infinity (Waldbronn, Germany) equipped with a quaternary pump, an autosampler, a column oven and a variable wavelength UV-vis detector (PDA). A Zorbax SB-C 18 column (4.6 × 250 mm, 5 µm) was used at 30 • C for HPLC analysis. Sample solution (10 µL) was injected into the HPLC system and eluted with a mobile phase made of a mixture of 50 mM PBS (pH 2.8) and acetonitrile at a ratio of 99:1 (v/v). The flow rate was 1.0 mL/min running in isocratic conditions for 35 min and the wavelength was set at 210 nm. A mixture of four standards of SCFAs including formic, acetic, propionic and butyric acids, with concentrations ranging from 10,000 to 5 mg/L, was performed at the same conditions. The concentrations of formic, acetic, propionic and butyric acids in fecal solution were calculated according to their corresponding standard curves.

Determination of Amino Acid in Feces
The amino acid composition of fecal samples was determined according to our previous method [75]. In this study, the concentration of amino acid was expressed as µmol per L.

Determination of Oxidative Status
After continuous oral administration of HAHp(3.0)/ZnO NPs for 14 days, the activities of some key enzymes in the liver, including SOD, GPx, and CAT, were measured using their corresponding assay kits. The content of antioxidant compounds and the oxidative degree of lipids are also associated with the oxidative stress. In the present study, the concentration of GSH, one of major antioxidants in organisms, as well as MDA, which reflects the finial oxidation degree of lipids, were also determined. The supernatants of 10% liver homogenates were used for GSH and MDA assays. The 1% and 0.25% liver homogenates were prepared by diluting the supernatant of 10% liver homogenates with saline. The 1% liver homogenates were used for the CAT assay, and the 0.25% liver homogenates were used for the total SOD and GPx assays.
Meanwhile, the concentrations of protein, peptide and total SH were compared between HAHp(3.0) and HAHp(3.0)/ZnO NPs solutions. The protein concentration was determined using Bradford's (1976) method with bovine serum albumin as the standard [76]. The peptide concentration was measured using OPA method [73]. Both protein and peptide concentrations were expressed as mg per mL. The total SH concentration was evaluated using the total SH assay kit and expressed as mmol per gprot.

Statistical Analysis
The data were expressed as mean ± standard deviation. Statistical analysis was performed using the SPSS ® software (SPSS Statistical Software 19.0, Inc., Chicago, IL, USA). The significant difference was set as p < 0.05.