Effect of dietary oil from Camelina sativa on the growth performance, fillet fatty acid profile and gut microbiome of gilthead Sea bream (Sparus aurata)

Introduction

In the last two decades, research has focused on testing cheaper and sustainable alternatives to fish oil (FO) in aquafeeds that do not compromise fish growth and the omega-3 content in the fillet that benefits human health. Numerous vegetable oils (VOs) have been tested and adopted to reduce FO inclusion in aquafeeds, which can maintain or even enhance growth performance of farmed fish (Turchini, Trushenski & Glencross, 2019; Kwasek, Thorne-Lyman & Phillips, 2020; Cottrell et al., 2020). Among these VOs, rapeseed, soybean, palm and sunflower oils have been frequently used replace FO without negatively affecting fish growth (Turchini, Torstensen & Ng, 2009; Hixson, Parrish & Anderson, 2014b).

However, only a small proportion of FO can be replaced by VOs due to their lack in omega-3 (n-3) long-chain polyunsaturated fatty acids (LC-PUFA), particularly eicosapentaenoic (EPA; 20:5n-3) and docosahexaenoic (DHA; 22:6n-3) acids. On the other hand, VOs are rich in omega-6 (n-6) and omega-9 (n-9) PUFA, mainly linoleic acid (LA; 18:2n-6), and oleic acid (OA; 18:1n-9). Therefore, a consequence of high replacement levels of FO with VOs would be a reduction of EPA and DHA levels in cultured fish tissues that can eventually, compromise the animal’s health (Turchini, Torstensen & Ng, 2009) and the nutritional value of fish as food (Karakatsouli, 2012). Omega-3 LC-PUFA typically enter the human diet through the consumption of oily fish (Ruiz-Lopez et al., 2014b). Therefore, the challenge for the aquaculture industry is not to simply replace FO with alternative lipid sources, such as VOs, but to find sources with a the fatty acid profile that mimics that of pelagic oceanic fish that is typically consumed by farmed fish.

The oilseed plant, Camelina sativa, is a member of the Cruciferae (Brassicaceae) family and has potential to replace a higher proportion of FO in aquafeeds, which gives it a competitive advantage over other commercial oilseed crops, e.g., rapeseed and sunflower oils (Budin, Breene & Putnam, 1995). In particular, C. sativa seeds can produce oil containing up to 40% of the omega-3 precursors α-linolenic acid (ALA; 18:3n-3) and LA (Budin, Breene & Putnam, 1995; Eidhin, Burke & O’Beirne, 2003). The Camelina oil (CO) also contains high levels of PUFA and MUFA (monounsaturated fatty acids) and high amounts of antioxidants, e.g., γ-tocopherol. Lastly, CO represents a sustainable alternative to FO in aquafeeds due to its low environmental footprint (Budin, Breene & Putnam, 1995).

The use of CO from both wild type and transgenic C. sativa plants has been widely investigated in the last decade in diets for different fish species and at different replacement levels of FO (Hixson, Parrish & Anderson, 2013; Hixson, Parrish & Anderson, 2014a; Hixson, Parrish & Anderson, 2014b; Hixson & Parrish, 2014; Haslam et al., 2015; Betancor et al., 2015b; Betancor et al., 2015a; Betancor et al., 2016a; Betancor et al., 2016b; Betancor et al., 2018; Toyes-Vargas et al., 2020). Genetically modified C. sativa is capable of producing EPA or even both EPA and DHA in its seeds due to the insertion of five microalgal genes codifying for fatty acyl desaturase and elongase that are involved in n-3 LC-PUFA biosynthesis (Ruiz-Lopez et al., 2014a).

Given the scarce desaturation and elongation capacity of most marine fish species to produce EPA and DHA starting from ALA, efforts have focused on enhancing the composition of CO to include these important fatty acids. In comparison to marine fish species, salmonids usually show greater endogenous capacity to synthesize EPA and DHA from ALA (Hixson, Parrish & Anderson, 2014b). Historically, marine fish have inhabited environments and consumed pelagic fish rich in n-3 LC-PUFA, therefore, they have had no evolutionary pressure to retain the ability to endogenously produce EPA and DHA. In contrast, phytoplankton in freshwater are characterized by having high levels of LA and ALA, moderate levels of EPA and low levels of DHA, which has contributed to maintaining high selective pressures on freshwater fish to produce DHA (Tocher, 2010). High levels of n-3 LC-PUFA in CO are vital for effective replacement of FO in diets for marine species, such as sea bream, that cannot endogenously produce DHA and EPA and rely on their inclusion in the diet.

Alternative ingredients must also be tested for the modulation of fish intestinal microbiota and ramifications on fish health. Gut microbiota plays an important role in key functions in fish, including the supply of nutrients, intestinal integrity and interactions with the immune system (Llewellyn et al., 2014; Ringøet al., 2016). As suggested by Castro et al. (2019), the different fatty acid composition of VO, with respect to FO, could modify the lipid composition, function and fluidity of intestinal cell membranes. Consequently, this could affect bacteria adherence to enterocytes, alter metabolic processes and modify the intestinal microbiota profile.

Due to advances from DNA sequencing technologies in the last two decades, researchers can utilize next-generation sequencing (NGS) to investigate the microbes in and on the outside of fish for both basic and applied applications. Recently, several studies focused on gut microbiome have been conducted in a wide range of fish species (Rimoldi et al., 2019; Rimoldi et al., 2020b; Terova et al., 2019; Legrand et al., 2020; Blaufuss et al., 2020; Pérez-Pascual et al., 2020; Zhao et al., 2020; Panteli et al., 2020). However, despite the large number of studies, the effects of replacement of FO with novel lipid sources on fish gut microbiota structure have not been sufficiently examined (Piazzon et al., 2017; Torrecillas et al., 2017; Castro et al., 2019; Magalhães et al., 2020). Furthermore, few studies have used NGS to characterize the gut microbiota of gilthead sea bream (Sparus aurata) in response to diet (Estruch et al., 2015; Parma et al., 2016; Parma et al., 2020; Piazzon et al., 2017; Rimoldi et al., 2018; Rimoldi et al., 2020a; Piazzon et al., 2020), and even less have described the functional profile of gut microbiome in this species (Piazzon et al., 2019; Piazzon et al., 2020). Indeed, dietary changes can affect the structure of fish intestine’s associated microbial community (what’s there?) but, to understand the intrinsic processes that lead to similar functionality, it is necessary to search the connections between individual microbiota (what are they doing?) and the corresponding metabolic phenotype.

The aim of the present study was to determine the effects of replacing FO with CO on the growth performance, fillet fatty acid composition and gut microbiota of gilthead sea bream. We used advanced nutritional techniques to characterize fatty acid profiles, the NGS Illumina platform to characterize microbial populations in the gut and the meta-genomic tool PICRUSt to determine the functional potential of the gut microbiome.

Materials & Methods

Results

Alpha-diversity of bacteria in the gut and diet

After filtering sequences, the Illumina MiSeq generated 2.7 million sequences for a library of 82 samples (32,721 reads per sample) and was further subsampled to 10,171 sequences per sample. Alpha-diversity of bacteria in the faeces (F0, F20, F40, F60) was not significantly different between diets in terms of number of OTUs (p = 0.940), Shannon (p = 0.926) and Chao1 (p = 0.780) indices (Figs. 1A–1C). For mucosa (M0, M20, M40, M60), there were no differences between diets for the above indices (p = 0.535, 0.521 and 0.930) and the same was found for diet (D0, D20, D40, D60) (p = 0.760, 0.234 and 0.791). Between each sample type, there were significant differences in terms of number of OTUs (p = 0.030) (Fig. 1A), Shannon (p = 0.003) (Fig. 1B) and Choa1 (p < 0.001) (Fig. 1C). All three types of sample were different for number of OTUs, Shannon and Chao1 (p < 0.05), except Shannon diversity was similar between mucosa and diet (p = 0.874).

Beta-diversity of bacteria in the gut and diet

ANOSIM analysis revealed that the composition of bacteria was significantly different between sample types (faeces vs mucosa, p < 0.001), but it was not modulated by diet neither in faeces nor in mucosa samples (p > 0.05). The non-metric multidimensional scaling (NMDS) confirmed the statistical analysis as shown in Fig. 2. At the phyla level, Firmicutes and Proteobacteria were the most dominant with a relative abundance up to 94% in faeces, 92% in mucosa and 95% in diet (Fig. 3). The ratio of Firmicutes:Proteobacteria was highest in the diets and lowest in the mucosa. Actinobacteria, Bacteroidetes, Spirochaetes and other bacteria made up the small remainder. At the genus level, Lactobacillus was the most dominant with up to 60% in faeces, 47% in mucosa and 77% in diets (Fig. 4). Photobacterium, Vibrio, Enterovibrio, Corynebacterium, Brevinema and other bacteria made up the remainder.

In regards to specific indicator OTUs, there were differences between dietary treatments based on LEfSe analysis. In faeces, genera of Peptostreptococcus, Vagococcus and Massilia increased in fish fed diets with 60% replacement with CO (Table 5). Lactobacillus was most associated with the 0% CO diet. In mucosa, there were more OTUs associated with diet type. Increased abundance of Corynebacterium, Rubrobacter, Ruegeria, Acetobacter and Enhydrobacter was found in fish fed diets with 60% replacement with CO (Table 6). Several genera in the phylum of Proteobacteria, such as Psychrobacter, were associated with the mucosa of fish fed the control diet (0% CO).

Table 5:

Indicator OTUs identified by LEfSe.

OTUs found in the faeces of S. aurata (n = 9) fed diets that replaced FO with 0% CO (Ctrl) and 60% CO.

Phyla	Order	OTU	%CO Diet	LDA	P-value
Firmicutes	Bacillales	Cerasibacillus	0	3.27	0.003
	Clostridiales	Peptostreptococcus	60	3.40	<0.001
	Lactobacillales	Lactobacillus	0	4.87	0.010
		Pediococcus	0	3.78	0.007
		Vagococcus	60	3.16	0.001
Proteobacteria	Burkholderiales	Massilia	60	3.43	<0.001

DOI: 10.7717/peerj.10430/table-5

Table 6:

Indicator OTUs identified by LEfSe.

OTUS found in the intestinal mucosa of gilthead sea bream (n = 9) fed diets that replaced fish oil with camelina oil (CO).

Phyla	Order	OTU	%CO Diet	LDA	P-value
Actinobacteria	Corynebacteriales	Corynebacterium	60	3.82	<0.001
	Rubrobacterales	Rubrobacter	60	3.19	0.027
Firmicutes	Bacillales	Exiguobacterium	0	3.32	0.011
Proteobacteria	Alteromonadales	Idiomarina	0	3.84	0.001
	Cellvibrionales	Haliea	0	3.05	0.005
	Oceanospirillales	Halomonas	0	3.98	<0.001
		Kangiella	0	3.04	0.021
	Pseudomonadales	Psychrobacter	0	4.26	<0.001
	Rhodobacterales	Loktanella	0	3.34	0.029
		Ruegeria	60	3.52	0.008
	Rhodospirillales	Acetobacter	60	3.83	<0.001
		Enhydrobacter	60	3.92	0.009
	Salinisphaerales	Salinisphaera	0	3.08	0.004

DOI: 10.7717/peerj.10430/table-6

Discussion

Reduced final weight of gilthead sea bream fed 60% CO (Table 2) demonstrated that this level of dietary inclusion is too high, although diets that replace 40% of FO with CO are feasible without negative effects on growth. This result contradicts a previous study on the same species where full replacement of FO by CO did not elicit differences in growth (Betancor et al., 2016a). In this previous study, the experimental diets were rich in fishmeal (FM; i.e., 49%), which may have compensated for the low levels of LC-PUFA in the diet. On the contrary, full replacement of FO with CO did not negatively affect growth of rainbow trout (Oncorhynchus mykiss) after 12-weeks (Hixson, Parrish & Anderson, 2014a). However, in another study by the same authors (Hixson, Parrish & Anderson, 2014b), the growth performance of Atlantic cod (family Gadidae), but not that of Atlantic salmon (Salmo salar) and rainbow trout (family Salmonidae), was negatively affected when fed diets that replaced 100% of FO with CO. As attested by Turchini, Torstensen & Ng (2009), a successful replacement of FO with terrestrial plant oil may be more readily achieved for fatty fish, such as salmonids, that have a higher and larger variety of fatty acids in their tissues than in lean fish, such as Atlantic cod (Gadus morhua), European sea bass (Dicentrarchus labrax), and gilthead sea bream.

In regards to transgenic CO, a study by Betancor et al. (2015b) fed post-smolt Atlantic salmon for 7-weeks with one of three experimental diets containing either FO, wild-type CO or transgenic CO as sole lipid source and showed contrasting results compared to our study. Atlantic salmon fed both wild and transgenic CO diets had similar final weight compared to salmon fed the FO diet. In a lean species such as European sea bass, the use of a high EPA+DHA genetically modified-Camelina sativa oil has instead been shown to promote adequate growth even at moderate dietary inclusion of FM (i.e., 25%) (Betancor et al., 2021). In the present study, FM was included at 19.4%, with no differences observed in growth with a 40% substitution of FO by CO. Other studies in the same species, did not find an effect on growth or feed utilization when FO was replaced up to 75% with VO at FM levels between 15–20% (Sánchez-Moya et al., 2020). The poorest growth in fish fed with 60% CO could be due to the presence of glucosinolates, which are responsible for the bitter or sharp taste of many cruciferous vegetables (Clarke, 2010) and are known to be prevalent in CO (Berhow et al., 2013). Indeed, poor palatability has been observed when using 100% CO substitution in European sea bass during the first two months of the experimental trial (Betancor et al., 2021).

We hypothesized that dietary CO alters the intestinal microbiota and its metabolic pathways. The NGS of gut microbiota in gilthead sea bream has been performed previously in a few studies and in general, the results agree with our findings. For instance, the replacement of marine with vegetable ingredients, i.e., vegetal meal (VM) and/or VO, in the gilthead sea bream diet did not influence the overall alpha-diversity of gut bacteria in previous studies by Estruch et al. (2015), Parma et al. (2016) and Castro et al. (2019). However, unlike our results, almost 2-fold higher alpha-diversity values (e.g., No. OTUs, Shannon and Choa1) were found in the hindgut of gilthead sea bream in these previous studies (Estruch et al., 2015; Parma et al., 2016). This could be explained by the different age of fish used in the present study since there is evidence that alpha-diversity decreases with age in fish (Llewellyn et al., 2016; Heys et al., 2020). Rimoldi et al. (2018) found no differences in the alpha-diversity of gut bacteria in gilthead sea bream of similar size compared to the current study. No significant differences in richness and diversity of gut bacteria was found in the study of Piazzon et al. (2020) that fed a control or a plant-based diet to three groups of genetically selected gilthead sea bream families. Equally, dietary levels of FM, FO or their combination did not affect faeces and mucosa species richness or Shannon-Weaver diversity index in European sea bass (Torrecillas et al., 2017). Conversely, a recent analysis of gut microbial community in gilthead sea bream fed a diet with high VM and VO level has highlighted a significant increase of microbiota biodiversity compared to control fish fed a FM/FO based-diet (Parma et al., 2020). In the same fish species, Piazzon and colleagues (2017) reported instead that 58% and 84% of substitution of dietary FO by VO lowered autochthonous gut microbiota diversity leading to a dominance of the Photobacterium genus.

It is interesting to note the lower alpha-diversity found in the mucosa in comparison to the faeces samples. This agrees with previous studies in salmonid fish that have reported a higher biodiversity and richness of bacterial species hosted by intestinal lumen in comparison to the gut mucosa (Gajardo et al., 2016; Huyben et al., 2018; Rimoldi et al., 2019; Terova et al., 2019). In European sea bass, autochthonous microbiota was instead characterized by higher species richness than allochthonous (transient) microbiota (Rimoldi et al., 2020b).

Previous NGS studies on gilthead sea bream have found similar gut microbial profile dominated by Firmicutes and to a lesser extent by Proteobacteria, Actinobacteria and Bacteroidetes phyla, and by Lactobacillus, Photobacterium, Corynebacterium and Propionibacterium genera (Estruch et al., 2015; Parma et al., 2016; Parma et al., 2020). In the present study, Proteobacteria and Firmicutes were the dominant phyla in both, faeces and mucosa samples regardless of the diet, as previously reported in gilthead sea bream and other fish species (Estruch et al., 2015; Parma et al., 2016; Parma et al., 2020; Piazzon et al., 2017; Rimoldi et al., 2018; Rimoldi et al., 2019; Rimoldi et al., 2020a; Rimoldi et al., 2020b; Terova et al., 2019; Magalhães et al., 2020). Beta diversity analysis showed that neither resident nor transient intestinal communities were affected by dietary treatment. In contrast, a clear distinguished microbial profile was found in the faeces of gilthead sea bream fed a diet with a total replacement of FO with VO (Castro et al., 2019).

At the genus/family taxa level, LEfSe analysis revealed that fish fed 60% CO diet showed a higher abundance of Corynebacterium in the gut mucosa, indicating a possible association of this taxon with the 60% CO diet. The same bacterial genus increased in gut mucosa of slow-growth gilthead sea bream fed with a plant-based diet (Piazzon et al., 2020). This genus contributed to higher predictive pathways of fatty acid synthesis, since it, along with other genera belonging to Actinobacteria and Firmicutes phyla (e.g., Lactobacillus), is capable of producing short-chain fatty acids, such as acetate and butyrate, which act as an energy source and stimulate enterocytes proliferation in fish (Kihara & Sakata, 1997; Asaduzzaman et al., 2018). In European sea bass the genera Clostridium, Corynebacterium and Staphylococcus were only identified in fish fed with low FM/FO levels (Torrecillas et al., 2017). In the gut of Arctic charr, dietary marine-derived oil reduced viable counts of Corynebacterium in comparison to fish fed soybean and linseed oils (Ringo et al., 2002). In contrast, a study by Estruch et al. (2015) showed that Corynebacterium in the hindgut of gilthead sea bream remained the same when FM was replaced with VM. Therefore, either these results indicate that high inclusion of CO may change the mucosal environment that benefits the colonisation of Corynebacterium, or fish may select Corynebacterium to produce fatty acids to compensate for fish oil replacement in the diet. A study in Atlantic salmon, showed that most bacteria in the gut are transient with no evidence of adaptation to host environment (Heys et al., 2020), thus suggesting that Corynebacterium may have been selected by the host based on the demand for fatty acids.

Reduced weight gain in fish fed the 60% CO diet may be associated with a reduced abundance of Lactobacillus in the gut, which were instead enriched in fish fed control diet 0% CO. Lactobacillus is a lactic acid bacterium (LAB; Lactobacillales order) often used as a probiotic in aquafeeds and therefore, considered favourable due to its properties to stimulate and enhance gut development, digestive function, mucosal tolerance, immune response and disease resistance in fish (Merrifield et al., 2010; Ringøet al., 2018). Indeed, a higher Firmicutes:Proteobacteria ratio is generally regarded as beneficial because it is positively correlated to the abundance of LAB. The high abundance of Firmicutes (32–68%) found in the gut agrees with most of the previous studies on gilthead sea bream (Estruch et al., 2015; Parma et al., 2016; Rimoldi et al., 2018). Only Piazzon et al. (2017) found a low percentage of Firmicutes (1–28%) in the gut of gilthead sea bream and these authors suggested Firmicutes may be more transient than the other phyla.

The predictive function of the gut microbiome of fish has only been investigated in a small handful of studies (Lyons et al., 2017; Wei et al., 2018; Yang et al., 2019; Piazzon et al., 2019; Qiao et al., 2020; Piazzon et al., 2020) and to our knowledge this is one of the first investigations of gilthead sea bream. The absence of significant effects of diet or gut section indicate that replacing FO with CO did not drastically alter the composition and function of microbes in the gut. It also reveals that the function of allochthonous (transient) and autochthonous (resident) bacteria are similar and analysing each gut section separately may not be necessary.

In regards to metabolic pathways, the predictive functions of gut microbes in the present studies agree with previous findings that they are mainly associated with the metabolism of carbohydrates, amino acids, energy and to a lesser extent lipids (Liu et al., 2016; Dehler, Secombes & Martin, 2017; Lyons et al., 2017; Wei et al., 2018; Yildirimer & Brown, 2018). Actinobacteria, such as Corynebacterium and Gordonia (Fig. 7), dominated the pathway of fatty acid synthase, which have low abundances in the gut but may be key contributors to lipid metabolism. In Arctic charr, feeding VOs increased viable counts of Corynebacterium in comparison to fish fed marine-derived oils (Ringo et al., 2002). A similar result was found in the present study especially with regard to the mucosa of fish fed with 60% CO diet. High VO diets may increase the abundance of fatty acid producing bacteria to compensate for the lack of fatty acids in the diet. Piazzon et al. (2020) found striking changes in the metabolic capacity of intestinal bacteria of gilthead sea bream in response to plant-based diets. In particular, 59, 84 and 15 pathways were found to be changed in fast growth, intermediate growth, and slow growth sea bream families, respectively. Among them, pathways related to infection, inflammation or immune activation, were the most influenced by diet and genetic selection. However, it should be noted that these results are only functional predictions and the deep shotgun sequencing of all bacterial genomes is preferred (Langille et al., 2013). Indeed, recently Sun, Jones & Fodor (2020) reported that the utility of PICRUSt is likely limited outside of human samples and that the development of tools for gene prediction specific to different non-human samples is needed. However, it is also true that the use of this meta-genomic tool represents the only available strategy to determine metabolic capacities of non-human microbiomes. For this reason, we only focused on the metabolic pathways instead of the disease related pathways. Functional metagenomics data should be validated or complemented with empirical approaches, such as sequencing the bacterial transcriptome, to give a more detailed description of metabolic pathways that are actually active in the gut microbial community.

Conclusions

This study demonstrated that diets can replace up to 40% of fish oil with CO without negative effects on growth performance, fillet fatty acid profile and gut microbiome of gilthead sea bream. Final weight and fillet EPA were significantly reduced in fish fed the 60% CO diet. Alpha and beta-diversities of gut bacteria both in the faeces and mucosa were not affected by any dietary treatment, although a few indicator bacteria, such as Corynebacterium and Rhodospirillales, were associated with the 60% CO diet. This study highlighted the difference in alpha-diversity between sample type, which was highest in the feeds and lowest in the gut mucosa. Reduced growth of fish fed 60% CO diet may be due to the lack of EPA and DHA in the CO and a reduction in lactic acid bacteria, specifically Lactobacillus, in the feaces and gut mucosa. Feed conversion was not affected by CO inclusion, thus reduced growth was not due to reduced digestibility but rather driven by nutrient demand, e.g., n-3 PUFA. Metagenomic analysis revealed similar predictive function of the gut microbes, although a higher abundance of Corynebacterium in the mucosa of fish fed 60% CO diet increased the KEGG pathway of fatty acid synthesis and may act to compensate for the lack of fatty acids in the diet. These results indicate that 40% of FO can be replaced with CO in diets for gilthead sea bream and demonstrates that CO can be an alternative oil ingredient in aquafeeds.

Supplemental Information