Effect of two insect diets on the intestinal commensal microbiome of healthy Salmo trutta vr. trutta


 Background The balance of the intestinal commensal microbiome of fish and other animals plays an important role in the physiological processes of healthy animals, contributes to the defense against pathogens, stimulates the immune system and facilitates nutrient metabolism. In the last decade, the use of insects in fish nutrition has been increasing exponentially, although little is known regarding the effects of insect meals on the fish gastrointestinal tract. The aim of this study was to evaluate the effect of two insect diets containing mealworm (MW) and superworm (SW) on the microbiome of the digesta of sea trout fingerlings and the relative abundances of different taxa among communities under controlled conditions.Results The insect meals produced a similar weight gain and survival rate to sea trout fed fishmeal. The most abundant bacterial phylum in all the treatment groups was Firmicutes followed by Proteobacteria and Actinobacteria, and significant differences in the amount of Cyanobacteria were observed in the SW group. Conclusions The insect meals did not produce differences in the three most abundant phyla in the sea trout digesta. However, the effect of each type of meal on the lower taxonomic levels was evident, particularly in the case of the superworm meal. These microbiome differences indicated that mealworm meal was more related to fishmeal than superworm meal. Our results highlight the potential effects of insect meals, such as mealworm and superworm meals, on the microbiota of sea trout.


Background
Fish, as well as other animals, must maintain the microbiome (bacteria, archaea, fungi, and viruses) in their intestinal tract in a balanced state, preserving the mutualistic relationship along their life cycles. The microbiome contributes to the defense against pathogens, stimulates the immune system and assists with nutrient metabolism [1]. Recently, several studies have related the importance of maintaining a balanced gut microbiota to the maintenance of a healthy status, and related a higher alpha diversity with healthy sh, which leads to improving the studies of the gut microbiota to develop methods to improve sh health [1], [2], [3], [4].
In the last decade, the use of insects in sh nutrition has been increasing exponentially, although little is known about the effects of insect meals on the sh gastrointestinal tract. Moreover, the use of different species of insects to feed a sh species, as well as different methods of insect meal preparation, are in most cases also unknown. Several studies have analyzed the sh microbiota using the next generation sequencing (NGS) technique. The microbial communities are affected by certain factors such as the species, the stage of development, the type of food consumed, and the intestinal morphology [5], [6]; environmental and physiological factors also modify the gut microbiota of sh [7]. The type of microbiome will also be conditioned by the feeding habit of the species; in salmonids such as rainbow trout, the predominant phyla are Proteobacteria, Firmicutes, Bacteriodetes, Fusobacteria and Actinobacteria [8], [9]. However, Rimoldi et al. (2018) found that rainbow trout fed higher levels of plant meals and rendered animal meals, and lower levels of shmeal exhibited more Fusobacteria and Bacteroidetes, and this difference was related to the lower growth performance [10]. Furthermore, Huyben et al. (2019) observed a variation in the abundance of a group of bacteria present in the gut of the rainbow trout according to the stage of development of the insect meal (larvae, prepupae, and pupae) [11]. The Proteobacteria and Firmicutes have been detected as dominated phyla in all gut regions of brown trout (Salmo trutta L.) [12]. Moreover, Michl et al. (2019) observed signi cantly increased abundances of Proteobacteria and Fusobacteria following the consumption of shmeal, whereas plantderived proteins increased the abundance of Firmicutes and Bacteriodetes [8].
In general, plant-based protein meals markedly modify the microbiome, as Kononova et al. (2019) showed the effects of soybean protein and carbohydrates, which are associated with some antinutritional factors, on the autochthonous microbiota, provoking in ammatory processes in the intestine of salmonids [13]. Regarding insect meals, Antonopoulou et al. (2019) found that rainbow trout fed 0 and 60% mealworm meal did not differ in the bacterial species or their amounts [14]. A possible explanation for this nding is that insects are part of the natural diet of this species.
Because insects are also part of the natural diet of sea trout, at least in the rst stages of development, the aim of this study was to evaluate the effects of two insect meal diets on the microbiome of the digesta of sea trout ngerlings and the abundances of different taxa among communities.

Growth performance
At the end of the experimental period, no signi cant differences in body weight gain and survival rates were observed among groups fed the different experimental diets, as shown in Fig. 1.

Microbiota diversity
In the sea trout gut microbiota, signi cant differences were observed in the number of bacteria and archaea among treatment groups, as sh fed the SW diet presented the lowest amount of bacteria and the highest amount of archaea (Fig. 2). In general, 99.95% of the microbiota was constituted by bacteria, 0.03% by archaea and 0.01% by viruses and 0.02% of the microorganisms were not identi ed.
Considering bacteria, 24 phyla were identi ed, and 20 were represented with an abundance less than 1%, whereas the most predominant bacteria among the remaining phyla were Firmicutes (Fig. 3). In general, signi cant differences in the abundances of Actinobacteria, Proteobacteria, and Firmicutes were not observed, although the abundance of Cyanobacteria exhibited signi cant differences, as the SW group presented the lowest content of this phylum as well as the combined data for the 20 phyla (p ≤ 0.05). After observing the class distribution in sea trout digesta, Bacilli and Clostridia were the most predominant classes in the groups treated with the three experimental diets, although the SW group exhibited the highest percentage of Bacilli, but the lowest percentages of Clostridia, Nostocophycideae and other classes compared to groups fed the CON and MW diets (Table 1). A similar trend was observed for the class distribution, where Bacilli and Clostridia classes were the most predominant and the SW group exhibited the highest percentage of Bacilli and the lowest percentage of Clostridia among all three treatment groups. Superworm meal also reduced the amount of Nostocophycideae and the grouped orders (Table 1). In terms of the family distribution, sh fed the SW diet exhibited lower percentages of Bacillaceae, Clostridiaceae, Lachnospiraceae, Rickettsiaceae, and Rivulariaceae than sh fed the CON and MW diets. In contrast, higher abundances of Corynebacteriaceae, Lactobacillaceae, and Leuconostocaceae were observed in the SW group than in sh fed the CON and MW diets. However, the percentage of Enterococcaceae was signi cantly lower in sh fed the CON diet than in sh fed the SW and MW diets (p < 0.05). Five hundred forty-one genera were identi ed, which represented the 85.10 ±3.51% of the total samples. After comparing the most predominant genera present in sh digesta (Fig. 3), sea trout fed the CON diet presented higher amounts of Clostridium and Lactobacillus, and the highest content was observed for Enterococcus followed by Clostridium in the MW group, but the most representative genera in the SW group were Pediococcus and Enterococcus.
The total amount of bacterial species identi ed varied among treatment groups. In the CON diet group, only 67.75 ± 4.63% of the bacteria were identi ed, whereas the MW diet group exhibited the highest percentage of identi cation of species at 73.25 ± 3.36%, followed by the SW diet group at 70.07 ± 7.30%. When observing the most predominant species in each treatment group (Table 2), the sh fed the CON diet presented the highest percentage of Clostridium cadaveris and Calothrix parietina. In the case of the MW group, the most abundant species were also C. cadaveris and C. parietina along with Enterococcus silesiacus. Moreover, the highest abundances were observed for Pediococcus pentosaceus, C. cadaveris, and Enterococcus durans. In addition, signi cant differences were detected in the SW group, with the lowest levels observed for Alkaliphilus crotonatoxidans, C. parietina, C. cadaveris, Lactobacillus antri, L. delbrueckii, and Streptococcus gallinaceus, and the highest values observed for E. durans, E. gallinarum, E. gilvus, Lactococcus garvieae, P. pentosaceus, P. acidilactici, P. stilesii and Weisella cibaria (p ≤ 0.05).
Regarding the species relations ( Fig. 4), 40.26% of the species were shared among treatments and 11.96 to 13.17% of the species are unique to each treatment; a lower number of shared species were observed between two treatments. Additionally, the amount of lactic acid bacteria (LAB) increased signi cantly in the sh fed the insect meals, namely, 36.16% and 47.98% in the MW and SW groups, respectively, compared to sh fed the control diet at 27.29% (p ≤ 0.05). Regarding species richness, 1328 species were identi ed, and the species richness was 905 species for the CON group, 879 species for the MW group, and 905 species for the SW group. At the same time, the Margalef index (D Mg ) showed no signi cant differences among treatments. The alpha diversity indexes, such as Simpson Diversity Index (D), Menhinick index, and Shannon index (H), revealed that insect meals did not affect the species richness. Meanwhile, the Evenness index (e^H/S), and equitability Brillouin index also showed no signi cant differences among treatments. The dominance index (Berger-Parker) and Dominance D displayed similar values among meal-fed groups, and the dominance was low among groups. Regarding the abundance estimator, the Chao1 calculation showed no signi cant differences among treatments as well ( Table 3).
The Bray-Curtis analysis of beta diversity is presented in Table 4. Additionally, the nonmetric multidimensional scaling (NMDS) analysis (Fig. 5) and the clusters showed that CON and MW groups were much more related, with 76.6% similarity, than the SW group (61.2%) (Fig. 6).

Discussion
In the last several decades, the importance of the gut microbiota has been documented in numerous studies showing that growth performance and sh health are closely related to the microbiota. As Butt and Volkoff (2019) commented, feeding habits in uence the structure and composition of the gut microbiota [15]. Additionally, plant-based proteins change the content and structure of the autochthonous microbiota of carnivorous species (Kononova et al. 2019) such as sea trout; in contrast, the use of a natural source of protein such as insect meals may play a role in maintaining the amount of these types of microorganisms that are part of the gut environment of the sh and enhance sh health [13]. Furthermore, insect meals would be able to modulate the microbiota of these animals due to the chitin and antimicrobial peptide contents [16], [17].
In this trial, more than 40% of FM ( sh meal) was replaced with insect meals in the two diets, although the insect meals produced similar growth and survival rates when observing the weight gain. When analyzing the bacteria present in the digesta, the dominant phylum in all the treatment groups was Firmicutes, in contrast to brown trout fed a commercial diet, in which the dominant phylum was Proteobacteria, ranging from 88.4 to 92.6% [12].  [13]. The results from the present trial showed that the abundance of Firmicutes would be conditioned by the amount of plant meal in the three diets, which was approximately 47% of the total, but not the inclusion of insect meals.
After performing a detailed analysis of the classes present in the digesta, Bacilli was the most abundant in all treatment groups, followed by Clostridia, both of which belong to the Firmicutes phylum, but the sum of Alphaproteobacteria and Gammaproteobacteria, which belong to the Proteobacteria phylum, presented similar amounts in all treatment groups, indicating that MW and SW meals exerted the same effect as FM on the digesta of sea trouts. In addition, that the phylum Firmicutes and class Clostridia have been repeatedly identi ed in the digestive tracts of herbivorous sh, and as described above, the higher abundance of this class would be related to the higher amount of plant meal present in the diets [6].
The order and family distribution followed a similar trend as the class distribution. Although the bacterial genera exhibited changes based on the type of protein meal source, the most representative genera in the CON group were Clostridium and Lactobacillus and those in the MW group were Enterococcus and Clostridium, but the most representative genera in the SW group were Pediococcus and Enterococcus.
With the exception of Pediococcus, the other genera are used as probiotics in aquaculture, increasing bacterial diversity [15], which probably occurred in sh fed these insect meals. The type of meals exerted a direct effect on the abundance of different genera in the intestinal digesta.
An analysis of the species abundance showed that sh fed the diet with SM exhibited a decrease in the relative abundance of C. cadaveris compared to the CON and MW groups; this species is known as a component of the normal fecal ora of humans and animals, which affects people with a poor overall condition of immunosuppression [19]. On the other hand, C. cadaveris is one of the most prominent bacterium present during the decay of dead bodies [19]. Moreover, C. cadaveris might trigger bacteremia that is related to a high mortality rate in humans. In the present study, the relative abundance of the commensal species C. cadaveris was decreased in the SW group, but signi cant differences were not observed between the CON and MW groups. Therefore, the reduction in the relative abundance C. cadaveris in the gastrointestinal tract of sh induced by the diet containing SW should may considered as a positive effect on public health.
C. parietina belongs to phylum Cyanobacterium and was previously detected in alkaline and oxygenated freshwaters [20]. The growth of Cyanobacteria is stimulated by the hypoxia of water reservoirs. Moreover, the contamination of dry food and feed with Cyanobacterium is considered a risk of toxin prevalence. Moreover, C. parietina is the bacteria with a higher potential for endotoxin production. The diet containing SW caused a decrease in the abundance of the bacterial species C. parietina in the sh GIT, which may reduce possible cyanobacterial toxin reservoirs in the sh GIT.
The diet containing SW improved the commensal probiotic microbiome in intestinal digesta of Salmo trutta vr. trutta. The SW diet increased the abundance of some bacterial genera, such as Pediococcus that is considered a positive sh GIT bacteria. Pediococcus is a genus of gram-positive lactic acid-producing bacteria belonging to the family Lactobacillaceae. In the SM group, an increase in the abundance of pediococci was observed, with the most abundant species identi ed as P. pentosaceus in the SW group. The bacterial species P. pentosaceus exerts bacteriocynogenic effects on Staphylococcus aureus and Escherichia coli [21]. P. pentosaceus is mostly associated with food fermentation; it produces pendocins that are safe for food preservation and is used as a starter culture in the fermentation of meat products.
Enterococcus is a key component of the intestinal ora of humans and is widespread in the intestines of most animals, including sh. Some species belonging to the Enterococcus genus, such as Enterococcus faecalis from sh intestine, are use as aquatic probiotics [22]. The SM diet increased the abundance of some Enterococcus species in the fecal digesta, among which E. durans may be considered a possible probiotic, because it potentially produces bacteriocins, namely, durancins [23]. Another species with probiotic potential that have been isolated from sh is Enterococcus gallinarum that regulates the innate immune response [24]. An increase in the abundance of Enterococcus gilvus was also observed in the fecal digesta of the analyzed SW group. The analysis of gene expression in Enterococcus gilvus has identi ed novel carotenoid biosynthesis genes that improve the multistress tolerance of Lactococcus lactis and promotes their activity toward methicillin-resistant S. aureus (MRSA) and vancomycin-resistant enterococci (VRE) [25]. Additionally, W. cibaria, which was more abundant in the SW group, has shown to be an effective probiotic in hybrid surubim [26]. Although signi cant differences among certain groups of bacteria were observed, the composition of the most representative species shows that they are part of the digestive tract ora, the environment, or part of the protein sources with probiotic properties that help the sh to thrive and achieve target growth and survival rates. In addition, Gajardo et al. (2016) commented that LAB are more abundant in salmon fed a plant-based diet than in sh fed a shmealbased diet [27], although, Ringø and Gatesoupe (1998) commented that LAB, such as the Lactobacillus, Carnobacterium, and Streptococcus genera, are also commonly detected in healthy sh microbiota of different sh species, including salmonids [28]. However, insect meals also increase the amount of LAB, as observed in the present study.
Our research diet containing SW decreased the relative abundance of Streptococcus gallinaceus in the intestinal digesta. S. gallinaceus was rst described in 2002 and was isolated from clinical samples of chickens. In 2003, S. gallinaceous was isolated from an outbreak of septicemia associated with a high prevalence of endocarditis in a ock of broiler parents. Chad eld et al. (2005) reported an association of this species with septicemia and endocarditis in chickens [29]. The decrease in the relative abundance of S. gallinaceus detected in the intestinal digesta might be considered a positive dietary effect of SW.
Furthermore, when comparing alpha diversity parameters, the inclusion of insect meal in the diet did not modify the different parameters measured, such as richness, evenness, and dominance. The Shannon H values were similar to those obtained in rainbow trout fed only FM and greater than 60% of MW meal [14]. Additionally, the Chao1 values obtained in the present study were similar to those observed in the digesta of the proximal intestine of salmon fed 45% FM and 38% plant meals [27]. These authors obtained a higher Shannon H index than observed in our results. Moreover, in brown trout fries fed three experimental diets, 100% FM, 50% and 90% plant-based diets followed by a crossover feeding design, plant-based diets produced higher Chao1 and Shannon indexes than the FM diet, although the Chao1 values were lower than the values reported for sea trout in this experiment [8]. In general, the diversity among treatments was similar. Additionally, the NMDS analysis and the similarities of the clusters showed that the microbiome of the MW group is more similar to the FM group than the SW group, which would be more useful for salmonid nutrition, as described by Antonopoulou et al. (2019) [14].
As mentioned above, the two insect meals exerted a similar effect to FM on maintaining the alpha diversity, and the values of dominance, equitability, and evenness were similar between all treatment groups, showing a balanced microbiome population that varied in abundance among bacterial classes, orders, genus and species as a natural consequence of the type of protein sources used. Nevertheless, the different meals that the sh consumed exerted positive effects on the microbiome, growth and survival performance of the sea trout, although the predominance of phylum Firmicutes in all treatment groups would be a consequence of the amount of plant meals, which were higher (47.17%) than animal meals (33.5%) in the diets, particularly for soybean meal, as highlighted by Kononova [8], [13]. Nevertheless, we cannot forget that plant meals are part of all commercial diets because of their availability and lower prices than shmeal, and they are used to study the effects of alternative meals, such as insect meals.

Conclusion
Insects are part of the sea trout diet in nature, at least in the rst stages of development, before these sh feed on more diverse prey, including other sh. To conclude, this nding may explain why the main phyla present in the digesta were similar in all the treatment groups. However, the effect of each type of meal on the lower taxonomic levels was evident, particularly in the case of superworm meal. These differences were highlighted through the NMDS and the clusters, where sh fed mealworm meal were more related to sh fed shmeal than sh fed superworm meal. Nevertheless, further studies are necessary to corroborate the nding that insect meals are one of the best alternatives to replace shmeal in the diets of carnivorous sh.

Fish rearing conditions and experimental diets
Living insects were provided by HiProMine S.A (Robakowo, Poland). The larvae were euthanized by freezing at -20°C for 24 h, after which the insects were oven-dried at 50°C for 24 h and nely ground.
(Metalchem S-60, Gliwice, Poland) at 110°C to obtain pellets with 1.5-mm and 2.5-mm diameters. After extrusion, pellets were dried in an oven for 48 h at 40°C, and then sh oil was added to the mildly heated pellets. The nutritional values are shown in Table 5. Bacterial DNA Extraction and 16Sr RNASequencing The research was conducted in accordance with the methodology of the Authors' previous research [30], [31].

Metagenomic Analysis
The research was conducted in accordance with the methodology of the Authors' previous research [30], [31].

Statistical Analysis
The research was conducted in accordance with the methodology of the Authors' previous research [30], [31].
Bioinformatic analysis ensuring the classi cation of readings by species level was carried out with the free Infostat software was used for the one-way ANOVA, and if signi cant differences were observed among treatment groups, data were further analyzed using Tukey's post hoc test.
The beta diversity measure was calculated based on the Bray-Curtis method [32].
The Kolmogorov-Smirnov test was used to determine the normality of the data distribution and equality of variances. Data are presented as the means ± standard errors of the means (SEM). Statistical signi cance was declared at p ≤ 0.05. carried out in compliance with the ARRIVE guidelines. All methods were carried out in accordance with relevant guidelines and regulations.
According to Polish law and the EU directive (no 2010/63/EU), the experiments conducted within this study did not require the approval of the Local Ethical Committee for Experiments on Animals in Poznan.

Consent for publication
Not applicable.