bs_bs_banner
Environmental Microbiology (2015) 17(10), 3914–3924
doi:10.1111/1462-2920.12888
Selection in the host structures the microbiota associated with developing cod larvae (Gadus morhua)
Ingrid Bakke,1,2* Eivind Coward,3 Tom Andersen4 and Olav Vadstein1,2 1 Department of Biotechnology, Faculty of Natural Sciences and Technology, NTNU Norwegian University of Science and Technology, Sem Saelands v. 6/8, N-7491 Trondheim, Norway. 2 NTNU Centre of Fisheries and Aquaculture, NTNU Norwegian University of Science and Technology, Trondheim, Norway. 3 Department of Cancer Research and Molecular Medicine, Faculty of Medicine, NTNU Norwegian University of Science and Technology, Trondheim, Norway. 4 Department of Bioscience, University of Oslo, Box 1066 Blindern, Oslo 0316, Norway. Summary Marine fish larvae are immature upon hatching, and share their environment with high numbers of bacteria. The microbial communities associated with developing fish larvae might be structured by other factors than those important in developing terrestrial animals. Here, we analysed the beta (β)-diversity of the microbiota associated with developing cod larvae and compared it with the bacterial communities in water and live feed by applying pyrosequencing of bar coded v4 16S rDNA amplicons. A total of 15 phyla were observed in the cod larval microbiota. Proteobacteria was the most abundant, followed by Firmicutes, Bacteroidetes and Actinobacteria. The composition and diversity of the cod larval microbiota changed considerably with age. The temporal and spatial patterns of β-diversity could not be explained by stochastic processes, and did not coincide with changes in the rearing conditions. Furthermore, the larval microbiota was highly distinct from the water and the live feed microbiota, particularly at early developmental stages. However, the similarity between larval and water microbiota increased with age. This study suggests that strong selection in the host structures the cod larval Received 4 November, 2014; revised 22 April, 2015; accepted 22 April, 2015. *For correspondence. E-mail ingrid.bakke@ biotech.ntnu.no; Tel. (+47) 73 59 7859/(+47) 932 47872.
© 2015 Society for Applied Microbiology and John Wiley & Sons Ltd
microbiota. The changes in community structure observed with increasing age can be explained by altered selection pressure due to development of the intestinal system. Introduction Vertebrate embryos develop in bacteria-free environments. After birth or hatching, the gastrointestinal (GI) tract is rapidly colonized by bacteria. This colonization is essential for proper development of the immune system and gut function (Fraune and Bosch, 2010). The GI microbiota contributes to efficient nutrient digestion, and it serves as a barrier against pathogenic bacteria (Cheesman and Guillemin, 2007; Fraune and Bosch, 2010; Kanther and Rawls, 2010). Studies of gnotobiotic zebrafish and mice have revealed that many host responses to the GI microbiota are conserved between fishes and mammals (Rawls et al., 2004; Kanther and Rawls, 2010). The principles governing the development of a functional GI microbial community in juvenile host individuals are not well understood. Ecological processes like selection inside the host, dispersal limitation and stochastic processes probably influence the community structure (Dethlefsen et al., 2006; Robinson et al., 2010; De Schryver and Vadstein, 2014). Selection inside the host is multi-faceted and is affected by e.g. diet (resources), competition between taxa in the community, host genetics and developmental processes. In human and mouse, the initial GI microbiota is transferred from the mother during birth, and temporal shifts in community structure associated with dietary changes are observed before the microbiota stabilizes and achieves the characteristics typical for adult host individuals (Palmer et al., 2007; Koenig et al., 2011; Pantoja-Feliciano et al., 2013). In a meta-study including diverse fish species, Sullam and colleagues (2012) concluded that salinity, trophic level and possibly host phylogeny were the most important determinants for the composition of fish gut microbiota. That study was based on data obtained from adult fishes. For developing fish larvae, it is not known which factors are most important for structuring the microbiota. For most fish species, the maternal influence on the offspring’s microbiota is likely of limited
Ontogeny of cod larval microbiota significance (Sullam et al., 2012). As soon as a newly hatched larva opens its mouth, the GI system is exposed to the bacteria in the environment and it becomes colonized (Hansen and Olafsen, 1999; Bates et al., 2006). The water with its high densities of bacteria is probably the source of bacteria for the first colonization of the fish larvae. Moreover, marine fish larvae actively take up bacteria from the water at rates 100 times faster than by the passive uptake that would be expected from their drinking rates (Reitan et al., 1998). How this intimate relationship between the fish larvae and the environmental bacteria influences the community assembly in the developing fish larvae is not known. The structure of the GI microbiota is probably further affected by the changing selective conditions inside the gut of the growing larvae. This may be of particular importance for marine fish larvae, as they are immature at hatching and undergo large developmental changes. Although the importance of physico-chemical conditions in the gut has been recognized, the GI microbiota of the fish larvae in early developmental stages has been suggested to be a reflection of the environmental microbial communities, particularly that of the live feed (Austin, 2006; Korsnes et al., 2006; Bjornsdottir et al., 2009; Nayak, 2010b; Llewellyn et al., 2014). Some recent studies however indicate that diet is a less important determinant to the fish larval microbiota than previously assumed (Yan et al., 2012; Bakke et al., 2013; Sun et al., 2013). Stochastic processes may also influence the community structure, such as random sampling from the microbial communities associated with water and diet (De Schryver and Vadstein, 2014). The relative importance of stochastic and deterministic processes in community assembly in fish larvae has been little studied, but findings by Yan and colleagues (2012) suggested that deterministic processes structured the GI microbiota of zebrafish at early developmental stages, followed by increased significance of stochastic processes at later developmental stages. Detrimental fish microbe interactions have been shown to limit the production of high quality juveniles for many species in aquaculture (Vadstein et al., 1993). Recently, increased attention has been given to the interactions between the host-associated microbiota and the environmental bacterial communities and the potential consequences for a healthy development of the fish larvae (Nayak, 2010a; Vadstein et al., 2013; De Schryver and Vadstein, 2014). In this perspective, understanding the processes determining the community structure of developing fish larvae is of essential interest from both a scientific and commercial point of view. In this study, we characterized the microbiota associated with larval stages of cod by 16S rDNA 454 amplicon pyrosequencing. In addition to describing the ontogeny of the cod larval microbiota, we examined microbial commu-
3915
nity assembly mechanisms by analysing patterns of β-diversity of the larval microbiota and comparing it to bacterial communities in water and live feed. In these analyses, we focused on the significance of stochastic versus deterministic factors, such as environmental influences and selection in the host.
Results Richness and diversity of the microbiota associated with cod larvae, rearing water and live feed cultures The microbiota associated with individual cod larvae reared in two rearing tanks (T1 and T2), representing different feeding regimes, was characterized using pyrosequencing of bar coded v4 16S rDNA amplicons. For both tanks, a pooled amplicon sample was generated, consisting of tagged amplicons representing five individual larvae and rearing water sampled four times [8 days post-hatching (dph), 17 dph, 32 dph and 61 dph], and the relevant live feed cultures (copepod + Artemia culture for T1, rotifer + Artemia cultures for T2). After quality trimming and chimera removal, 169 771 and 181 860 sequence reads were obtained for the T1 and the T2 samples respectively. The highest numbers of reads per amplified sample were obtained for water and feed samples (average > 8000; Table 1). The average number of reads across all larval samples was 6248 ± 1196. Comparison of estimated richness (Chao1) and the observed number of operational taxonomic units (OTUs; at 97% sequence similarity level) demonstrated that the sequencing effort across samples on average covered close to 70% of the estimated richness (Fig. 1). Both richness and diversity appeared to be highest in the bacterial communities of rearing water and live feed cultures, whereas the larval microbiota exhibited the lowest diversity at 17 dph and 32 dph, and the highest at 61 dph, where the diversity was similar to that of the rearing water and live feed communities (Fig. 1).
Table 1. Average number of reads (±standard deviation) per sample category after quality trimming and removal of chimeric sequences. Average number of reads (±STD) Sample category
Tank 1
Tank 2
Larvae D8 (n = 5) Larvae D17 (n = 5) Larvae D32 (n = 5) Larvae D61 (n = 5) Water D8, 17, 32, 61 Live feed (Cop/Rot + Art)*
5671 ± 979 5743 ± 917 5287 ± 1515 6841 ± 665 8535 ± 698 8313 ± 305
6264 ± 804 6746 ± 741 5602 ± 1377 7572 ± 1412 8312 ± 310 8847 ± 940
* Including amplicons representing copepod and Artemia cultures for T1, and rotifer and Artemia cultures for T2.
© 2015 Society for Applied Microbiology and John Wiley & Sons Ltd, Environmental Microbiology, 17, 3914–3924
3916 I. Bakke, E. Coward, T. Andersen and O. Vadstein
Fig. 1. Average richness and diversity indices for microbial communities associated with individual cod larvae, water and live feed samples in the two tanks T1 and T2. A. Chao1 richness and observed number of OTUs. B. Shannon’s diversity index.
Beta-diversity of cod larval microbiota: temporal trends Fifteen phyla were observed in the cod larval microbiota. Actinobacteria, Bacteroidetes, Firmicutes and Proteobacteria were observed in all individuals, and also in all water and live feed samples (Fig. 2). Proteobacteria was the most abundant phylum in all larvae, and particularly Gammaproteobacteria dominated larval and water microbiota. Representatives of the Acidobacteria, Chlamydiae, Cyanobacteria, Fusobacteria, Gemmatimonadetes, Nitrospira, Planctomycetes, Verrucomicrobia and the candidate phyla TM7, OD1 and SR1 were sporadically present in the cod larval microbiota. The composition of the larval microbiota clearly changed with age, both at the phylum level and at higher taxonomic levels (Fig. 2). However, the microbiota of larvae at 17 dph and 32 dph appeared to be strikingly similar. The temporal trends were similar for larvae reared with different diets. Characteristic for the 8 dph larval microbiota was a high abundance of Pseudomonas (mainly represented by one OTU), the presence of diverse Betaproteobacteria (representing at least 13 different genera), and a relatively high abundance of Bacilli. At 17 dph, the composition of the
larval microbiota was remarkably changed. The diversity of the larval microbiota was lowest at 17 and 32 dph (Figs 1B and 2), and Arcobacter (Epsilonproteobacteria) represented by one OTU together with Gammaproteobacteria (Vibrionaceae, Pseudomonadaceae and Oceanospirillaceae were the most abundant families) typically constituted more than 90% (up to 98%) of the reads. At 61 dph, the larval microbiota was considerably more diverse, and the Gammaproteobacteria was represented by a variety of genera dominated by Colwellia, Photobacterium, Leucothrix, Vibrio and Pseudomonas. Another characteristic of the 61 dph larval microbiota was a relatively high abundance of diverse Rhodobacteraceae (Alphaproteobacteria), represented by more than 80 different OTUs. A principal coordinate analysis (PCoA) plot based on Bray–Curtis dissimilarities corroborates the patterns described above: The larval community profiles cluster according to larval ages, and the 17 dph and 32 dph microbiota overlaps (Fig. 3). The average Bray–Curtis dissimilarity between 17 dph and 32 dph larval microbiotas was only around 0.4, compared with approximately 0.9 for comparisons between other larval age groups. A one-way permutational multivariate analysis of variance (PERMANOVA) test confirmed that there were significant differences in the larval mirobiota between all age groups for both the T1 and the T2 dataset (P < 0.01), except between the 17 dph and the 32 dph larvae (P = 0.14 for the T1 and P = 0.45 for the T2 dataset respectively). To investigate the changes in larval microbiota with age at a more detailed level, we determined the total number of OTUs observed for the first time at each age and the number of OTUs that persisted from the younger larval stages (data not shown). Even though the 17 dph microbiota was very dissimilar from the 8 dph microbiota, 60 (T2) to 70% (T1) of the OTUs present at 17 dph persisted from the 8 dph microbiota. At 61 dph, 26% (T2) to 33% (T1) of the OTUs observed for the first time at 8 dph were still present in the larval microbiota. Beta-diversity of cod larval microbiota: inter-individual variation Despite the fact that larvae of a given age exhibit a characteristic microbiota, and the potential dispersal of bacteria between host individuals through the water, the individual community profiles revealed variation among individuals (Fig. 2). Generally, the dominating taxa were the same for individuals at identical age, but their relative abundance varied among individuals. For example, the Arcobacter OTU that was abundant in 17 dph and 32 dph larval microbiota constituted from 17% to 77% of the OTUs for individual larvae. Whereas the
© 2015 Society for Applied Microbiology and John Wiley & Sons Ltd, Environmental Microbiology, 17, 3914–3924
Ontogeny of cod larval microbiota # 0 # # 0 # # # # # # # # # # # # # # # 0 # # # # # 0 # # 0 # # 0 # # # 0 # # # # 0 # # #
# 0 # # 0 # # # # # # # # # # # # # # # 0 # # # # # 0 # # 0 # # 0 # # # 0 # # # # 0 # # #
Larvae 8 dph # 0 # # 0 # # # # # # # # # # # # # # # 0 # # # # # 0 # # 0 # # 0 # # # 0 # # # # 0 # # #
# 0 # # 0 # # # # # # # # # # # # # # # 0 # # # # # 0 # # 0 # # 0 # # # 0 # # # # 0 # # #
# 0 # # 0 # # # # # # # # # # # # # # # 0 # # # # # 0 # # 0 # # 0 # # # 0 # # # # 0 # # #
0 0 0 0 0 0 0 0 0 0 0 # 0 0 # 0 0 0 0 # 0 0 0 0 # 0 0 0 0 0 0 0 0 # # 0 0 # 0 0 0 0 0 # 0
0 0 0 0 0 0 0 0 0 0 0 # 0 0 # 0 0 0 0 # 0 0 0 0 # 0 0 0 0 0 0 0 0 # # 0 0 # 0 0 0 0 0 # 0
0 0 0 0 0 0 0 0 0 0 0 # 0 0 # 0 0 0 0 # 0 0 0 0 # 0 0 0 0 0 0 0 0 # # 0 0 # 0 0 0 0 0 # 0
0 0 0 0 0 0 0 0 0 0 0 # 0 0 # 0 0 0 0 # 0 0 0 0 # 0 0 0 0 0 0 0 0 # # 0 0 # 0 0 0 0 0 # 0
0 0 0 0 0 0 0 0 0 0 0 # 0 0 # 0 0 0 0 # 0 0 0 0 # 0 0 0 0 0 0 0 0 # # 0 0 # 0 0 0 0 0 # 0
# 0 # # 0 # # # # # # # # # # # # # # # 0 # # # # # 0 # # 0 # # 0 # # # 0 # # # # 0 # # #
# 0 # # 0 # # # # # # # # # # # # # # # 0 # # # # # 0 # # 0 # # 0 # # # 0 # # # # 0 # # #
Larvae 17 dph # 0 # # 0 # # # # # # # # # # # # # # # 0 # # # # # 0 # # 0 # # 0 # # # 0 # # # # 0 # # #
# 0 # # 0 # # # # # # # # # # # # # # # 0 # # # # # 0 # # 0 # # 0 # # # 0 # # # # 0 # # #
# 0 # # 0 # # # # # # # # # # # # # # # 0 # # # # # 0 # # 0 # # 0 # # # 0 # # # # 0 # # #
0 0 0 0 0 0 0 0 0 0 0 # 0 0 # 0 0 0 1 # 0 0 0 0 # 0 0 0 0 0 0 0 0 # # 0 0 # 0 0 0 0 0 # 0
0 0 0 0 0 0 0 0 0 0 0 # 0 0 # 0 0 0 0 # 0 0 0 0 # 0 0 0 0 0 0 0 0 # # 0 0 # 0 0 0 0 0 # 0
0 0 0 0 0 0 0 0 0 0 0 # 0 0 # 0 0 0 1 # 0 0 0 0 # 0 0 0 0 0 0 0 0 # # 0 0 # 0 0 0 0 0 # 0
0 0 0 0 0 0 0 0 0 0 0 # 0 0 # 0 0 0 0 # 0 0 0 0 # 0 0 0 0 0 0 0 0 # # 0 0 # 0 0 0 0 0 # 0
0 0 0 0 0 0 0 0 0 0 0 # 0 0 # 0 0 1 0 # 0 0 0 0 # 0 0 0 0 0 0 0 0 # # 0 0 # 0 0 0 0 0 # 0
# 0 # # 0 # # # # # # # # # # # # # # # 0 # # # # # 0 # # 0 # # 0 # # # 0 # # # # 0 # # #
# 0 # # 0 # # # # # # # # # # # # # # # 0 # # # # # 0 # # 0 # # 0 # # # 0 # # # # 0 # # #
Larvae 32 dph # 0 # # 0 # # # # # # # # # # # # # # # 0 # # # # # 0 # # 0 # # 0 # # # 0 # # # # 0 # # #
# 0 # # 0 # # # # # # # # # # # # # # # 0 # # # # # 0 # # 0 # # 0 # # # 0 # # # # 0 # # #
# 0 # # 0 # # # # # # # # # # # # # # # 0 # # # # # 0 # # 0 # # 0 # # # 0 # # # # 0 # # #
0 0 0 0 0 0 0 0 0 0 0 # 0 0 # 0 0 0 0 # 0 0 0 0 # 0 0 0 0 0 0 0 0 # # 0 0 # 0 0 0 0 0 # 0
0 0 0 0 0 0 0 0 0 0 0 # 0 0 # 0 0 0 1 # 0 0 0 0 # 0 0 0 0 0 0 0 0 # # 0 0 # 0 0 0 0 0 # 0
0 0 0 0 0 0 0 0 0 0 0 # 0 0 # 0 0 0 0 # 0 0 0 0 # 0 0 0 0 0 0 0 0 # # 0 0 # 0 0 0 0 0 # 0
0 0 0 0 0 0 0 0 0 0 0 # 0 0 # 0 0 0 0 # 0 0 0 0 # 0 0 0 0 0 0 0 0 # # 0 0 # 0 0 0 0 0 # 0
0 0 0 0 0 0 0 0 0 0 0 # 0 0 # 0 0 0 0 # 0 0 0 0 # 0 0 0 0 0 0 0 0 # # 0 0 # 0 0 0 0 0 # 0
# 0 # # 0 # # # # # # # # # # # # # # # 0 # # # # # 0 # # 0 # # 0 # # # 0 # # # # 0 # # #
# 0 # # 0 # # # # # # # # # # # # # # # 0 # # # # # 0 # # 0 # # 0 # # # 0 # # # # 0 # # #
Larvae 61 dph # 0 # # 0 # # # # # # # # # # # # # # # 0 # # # # # 0 # # 0 # # 0 # # # 0 # # # # 0 # # #
# 0 # # 0 # # # # # # # # # # # # # # # 0 # # # # # 0 # # 0 # # 0 # # # 0 # # # # 0 # # #
# 0 # # 0 # # # # # # # # # # # # # # # 0 # # # # # 0 # # 0 # # 0 # # # 0 # # # # 0 # # #
0 0 0 0 0 0 0 0 0 0 0 # 0 0 # 0 0 0 0 # 0 0 0 0 # 0 0 0 0 0 0 0 0 # # 0 0 # 0 0 0 0 0 # 0
0 0 0 0 0 0 0 0 0 0 0 # 0 0 # 0 0 0 0 # 0 0 0 0 # 0 0 0 0 0 0 0 0 # # 0 0 # 0 0 0 0 0 # 0
0 0 0 0 0 0 0 0 0 0 0 # 0 0 # 0 0 0 0 # 0 0 0 0 # 0 0 0 0 0 0 0 0 # # 0 0 # 0 0 0 0 0 # 0
0 0 0 0 0 0 0 0 0 0 0 # 0 0 # 0 0 0 0 # 0 0 0 0 # 0 0 0 0 0 0 0 0 # # 0 0 # 0 0 0 0 0 # 0
0 0 0 0 0 0 0 0 0 0 0 # 0 0 # 0 0 0 0 # 0 0 0 0 # 0 0 0 0 0 0 0 0 # # 0 0 # 0 0 0 0 0 # 0
# 0 # # 0 # # # # # # # # # # # # # # # 0 # # # # # 0 # # 0 # # 0 # # # 0 # # # # 0 # # #
# 0 # # 0 # # # # # # # # # # # # # # # 0 # # # # # 0 # # 0 # # 0 # # # 0 # # # # 0 # # #
Water # 0 # # 0 # # # # # # # # # # # # # # # 0 # # # # # 0 # # 0 # # 0 # # # 0 # # # # 0 # # #
# 0 # # 0 # # # # # # # # # # # # # # # 0 # # # # # 0 # # 0 # # 0 # # # 0 # # # # 0 # # #
0 0 0 0 0 0 0 0 0 0 0 # 0 0 # 0 0 0 0 # 0 0 0 0 # 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 # 0 0 # 0 0 0 0 # 0 0 0 0 # 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 # 0 0 # 0 0 0 0 # 0 0 0 0 # 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 # 0 0 # 0 0 0 0 # 0 0 0 0 # 0 0 0 0 0 0 0 0
# 0 0 # 0 0 0 0 0 # 0
# 0 0 # 0 0 0 0 0 # 0
# 0 0 # 0 0 0 0 0 # 0
# 0 0 # 0 0 0 0 0 # 0
3917
Live feed # 0 # # 0 # # # # # # # # # # # # # # # 0 # # # # # 0 # # 0 # # 0 # # # 0 # # # # 0 # # #
0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 # 0
# 0 # # 0 # # # # # # # # # # # # # # # 0 # # # # # 0 # # 0 # # 0 # # # 0 # # # # 0 # # #
0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 # 0
Bacteria, other Proteobacteria, other Gammaproteobacteria, other Colwelliaceae, Alteromonadales Idiomarinaceae, Alteromonadales Moritellaceae, Alteromonadales Alteromonadales, other Pseudoalteromonadaceae, Alteromonadales Psychromonadaceae, Alteromonadales Enterobacteriaceae, Enterobacteriales Oceanospirillaceae, Oceanospirillales Oceanospirillales, other Moraxellaceae, Pseudomonadales Pseudomonadaceae, Pseudomonadales Thiotrichales, other Piscirickettsiaceae, Thiotrichales Thiotrichaceae, Thiotrichales Vibrionaceae, Vibrionales Campylobacteraceae, Campylobacterales Burkholderiales, other Comamonadaceae, Burkholderiales Oxalobacteraceae, Burkholderiales Methylophilaceae, Methylophilales Betaproteobacteria, other Alphaproteobacteria, other Bradyrhizobiaceae, Rhizobiales Rhizobiales, other Rhodobacteraceae, Rhodobacterales Sphingomonadaceae, Sphingomonadales Lachnospiraceae, Clostridiales Clostridiaceae, Clostridiales Clostridiales, other Carnobacteriaceae, Lactobacillales Lactobacillales, other Paenibacillaceae, Bacillales Bacillales, other Planococcaceae, Bacillales Flexibacteraceae, Sphingobacteriales Sphingobacteriaceae, Sphingobacteriales Cryomorphaceae, Flavobacteriales Flavobacteriaceae, Flavobacteriales Flavobacteriales, other Bacteroidetes, other Opitutaceae, Verrucomicrobiales Actinobacteridae, Actinomycetales
Environmental Microbiology (2015) 17(10), 3914–3924
doi:10.1111/1462-2920.12888
Selection in the host structures the microbiota associated with developing cod larvae (Gadus morhua)
Ingrid Bakke,1,2* Eivind Coward,3 Tom Andersen4 and Olav Vadstein1,2 1 Department of Biotechnology, Faculty of Natural Sciences and Technology, NTNU Norwegian University of Science and Technology, Sem Saelands v. 6/8, N-7491 Trondheim, Norway. 2 NTNU Centre of Fisheries and Aquaculture, NTNU Norwegian University of Science and Technology, Trondheim, Norway. 3 Department of Cancer Research and Molecular Medicine, Faculty of Medicine, NTNU Norwegian University of Science and Technology, Trondheim, Norway. 4 Department of Bioscience, University of Oslo, Box 1066 Blindern, Oslo 0316, Norway. Summary Marine fish larvae are immature upon hatching, and share their environment with high numbers of bacteria. The microbial communities associated with developing fish larvae might be structured by other factors than those important in developing terrestrial animals. Here, we analysed the beta (β)-diversity of the microbiota associated with developing cod larvae and compared it with the bacterial communities in water and live feed by applying pyrosequencing of bar coded v4 16S rDNA amplicons. A total of 15 phyla were observed in the cod larval microbiota. Proteobacteria was the most abundant, followed by Firmicutes, Bacteroidetes and Actinobacteria. The composition and diversity of the cod larval microbiota changed considerably with age. The temporal and spatial patterns of β-diversity could not be explained by stochastic processes, and did not coincide with changes in the rearing conditions. Furthermore, the larval microbiota was highly distinct from the water and the live feed microbiota, particularly at early developmental stages. However, the similarity between larval and water microbiota increased with age. This study suggests that strong selection in the host structures the cod larval Received 4 November, 2014; revised 22 April, 2015; accepted 22 April, 2015. *For correspondence. E-mail ingrid.bakke@ biotech.ntnu.no; Tel. (+47) 73 59 7859/(+47) 932 47872.
© 2015 Society for Applied Microbiology and John Wiley & Sons Ltd
microbiota. The changes in community structure observed with increasing age can be explained by altered selection pressure due to development of the intestinal system. Introduction Vertebrate embryos develop in bacteria-free environments. After birth or hatching, the gastrointestinal (GI) tract is rapidly colonized by bacteria. This colonization is essential for proper development of the immune system and gut function (Fraune and Bosch, 2010). The GI microbiota contributes to efficient nutrient digestion, and it serves as a barrier against pathogenic bacteria (Cheesman and Guillemin, 2007; Fraune and Bosch, 2010; Kanther and Rawls, 2010). Studies of gnotobiotic zebrafish and mice have revealed that many host responses to the GI microbiota are conserved between fishes and mammals (Rawls et al., 2004; Kanther and Rawls, 2010). The principles governing the development of a functional GI microbial community in juvenile host individuals are not well understood. Ecological processes like selection inside the host, dispersal limitation and stochastic processes probably influence the community structure (Dethlefsen et al., 2006; Robinson et al., 2010; De Schryver and Vadstein, 2014). Selection inside the host is multi-faceted and is affected by e.g. diet (resources), competition between taxa in the community, host genetics and developmental processes. In human and mouse, the initial GI microbiota is transferred from the mother during birth, and temporal shifts in community structure associated with dietary changes are observed before the microbiota stabilizes and achieves the characteristics typical for adult host individuals (Palmer et al., 2007; Koenig et al., 2011; Pantoja-Feliciano et al., 2013). In a meta-study including diverse fish species, Sullam and colleagues (2012) concluded that salinity, trophic level and possibly host phylogeny were the most important determinants for the composition of fish gut microbiota. That study was based on data obtained from adult fishes. For developing fish larvae, it is not known which factors are most important for structuring the microbiota. For most fish species, the maternal influence on the offspring’s microbiota is likely of limited
Ontogeny of cod larval microbiota significance (Sullam et al., 2012). As soon as a newly hatched larva opens its mouth, the GI system is exposed to the bacteria in the environment and it becomes colonized (Hansen and Olafsen, 1999; Bates et al., 2006). The water with its high densities of bacteria is probably the source of bacteria for the first colonization of the fish larvae. Moreover, marine fish larvae actively take up bacteria from the water at rates 100 times faster than by the passive uptake that would be expected from their drinking rates (Reitan et al., 1998). How this intimate relationship between the fish larvae and the environmental bacteria influences the community assembly in the developing fish larvae is not known. The structure of the GI microbiota is probably further affected by the changing selective conditions inside the gut of the growing larvae. This may be of particular importance for marine fish larvae, as they are immature at hatching and undergo large developmental changes. Although the importance of physico-chemical conditions in the gut has been recognized, the GI microbiota of the fish larvae in early developmental stages has been suggested to be a reflection of the environmental microbial communities, particularly that of the live feed (Austin, 2006; Korsnes et al., 2006; Bjornsdottir et al., 2009; Nayak, 2010b; Llewellyn et al., 2014). Some recent studies however indicate that diet is a less important determinant to the fish larval microbiota than previously assumed (Yan et al., 2012; Bakke et al., 2013; Sun et al., 2013). Stochastic processes may also influence the community structure, such as random sampling from the microbial communities associated with water and diet (De Schryver and Vadstein, 2014). The relative importance of stochastic and deterministic processes in community assembly in fish larvae has been little studied, but findings by Yan and colleagues (2012) suggested that deterministic processes structured the GI microbiota of zebrafish at early developmental stages, followed by increased significance of stochastic processes at later developmental stages. Detrimental fish microbe interactions have been shown to limit the production of high quality juveniles for many species in aquaculture (Vadstein et al., 1993). Recently, increased attention has been given to the interactions between the host-associated microbiota and the environmental bacterial communities and the potential consequences for a healthy development of the fish larvae (Nayak, 2010a; Vadstein et al., 2013; De Schryver and Vadstein, 2014). In this perspective, understanding the processes determining the community structure of developing fish larvae is of essential interest from both a scientific and commercial point of view. In this study, we characterized the microbiota associated with larval stages of cod by 16S rDNA 454 amplicon pyrosequencing. In addition to describing the ontogeny of the cod larval microbiota, we examined microbial commu-
3915
nity assembly mechanisms by analysing patterns of β-diversity of the larval microbiota and comparing it to bacterial communities in water and live feed. In these analyses, we focused on the significance of stochastic versus deterministic factors, such as environmental influences and selection in the host.
Results Richness and diversity of the microbiota associated with cod larvae, rearing water and live feed cultures The microbiota associated with individual cod larvae reared in two rearing tanks (T1 and T2), representing different feeding regimes, was characterized using pyrosequencing of bar coded v4 16S rDNA amplicons. For both tanks, a pooled amplicon sample was generated, consisting of tagged amplicons representing five individual larvae and rearing water sampled four times [8 days post-hatching (dph), 17 dph, 32 dph and 61 dph], and the relevant live feed cultures (copepod + Artemia culture for T1, rotifer + Artemia cultures for T2). After quality trimming and chimera removal, 169 771 and 181 860 sequence reads were obtained for the T1 and the T2 samples respectively. The highest numbers of reads per amplified sample were obtained for water and feed samples (average > 8000; Table 1). The average number of reads across all larval samples was 6248 ± 1196. Comparison of estimated richness (Chao1) and the observed number of operational taxonomic units (OTUs; at 97% sequence similarity level) demonstrated that the sequencing effort across samples on average covered close to 70% of the estimated richness (Fig. 1). Both richness and diversity appeared to be highest in the bacterial communities of rearing water and live feed cultures, whereas the larval microbiota exhibited the lowest diversity at 17 dph and 32 dph, and the highest at 61 dph, where the diversity was similar to that of the rearing water and live feed communities (Fig. 1).
Table 1. Average number of reads (±standard deviation) per sample category after quality trimming and removal of chimeric sequences. Average number of reads (±STD) Sample category
Tank 1
Tank 2
Larvae D8 (n = 5) Larvae D17 (n = 5) Larvae D32 (n = 5) Larvae D61 (n = 5) Water D8, 17, 32, 61 Live feed (Cop/Rot + Art)*
5671 ± 979 5743 ± 917 5287 ± 1515 6841 ± 665 8535 ± 698 8313 ± 305
6264 ± 804 6746 ± 741 5602 ± 1377 7572 ± 1412 8312 ± 310 8847 ± 940
* Including amplicons representing copepod and Artemia cultures for T1, and rotifer and Artemia cultures for T2.
© 2015 Society for Applied Microbiology and John Wiley & Sons Ltd, Environmental Microbiology, 17, 3914–3924
3916 I. Bakke, E. Coward, T. Andersen and O. Vadstein
Fig. 1. Average richness and diversity indices for microbial communities associated with individual cod larvae, water and live feed samples in the two tanks T1 and T2. A. Chao1 richness and observed number of OTUs. B. Shannon’s diversity index.
Beta-diversity of cod larval microbiota: temporal trends Fifteen phyla were observed in the cod larval microbiota. Actinobacteria, Bacteroidetes, Firmicutes and Proteobacteria were observed in all individuals, and also in all water and live feed samples (Fig. 2). Proteobacteria was the most abundant phylum in all larvae, and particularly Gammaproteobacteria dominated larval and water microbiota. Representatives of the Acidobacteria, Chlamydiae, Cyanobacteria, Fusobacteria, Gemmatimonadetes, Nitrospira, Planctomycetes, Verrucomicrobia and the candidate phyla TM7, OD1 and SR1 were sporadically present in the cod larval microbiota. The composition of the larval microbiota clearly changed with age, both at the phylum level and at higher taxonomic levels (Fig. 2). However, the microbiota of larvae at 17 dph and 32 dph appeared to be strikingly similar. The temporal trends were similar for larvae reared with different diets. Characteristic for the 8 dph larval microbiota was a high abundance of Pseudomonas (mainly represented by one OTU), the presence of diverse Betaproteobacteria (representing at least 13 different genera), and a relatively high abundance of Bacilli. At 17 dph, the composition of the
larval microbiota was remarkably changed. The diversity of the larval microbiota was lowest at 17 and 32 dph (Figs 1B and 2), and Arcobacter (Epsilonproteobacteria) represented by one OTU together with Gammaproteobacteria (Vibrionaceae, Pseudomonadaceae and Oceanospirillaceae were the most abundant families) typically constituted more than 90% (up to 98%) of the reads. At 61 dph, the larval microbiota was considerably more diverse, and the Gammaproteobacteria was represented by a variety of genera dominated by Colwellia, Photobacterium, Leucothrix, Vibrio and Pseudomonas. Another characteristic of the 61 dph larval microbiota was a relatively high abundance of diverse Rhodobacteraceae (Alphaproteobacteria), represented by more than 80 different OTUs. A principal coordinate analysis (PCoA) plot based on Bray–Curtis dissimilarities corroborates the patterns described above: The larval community profiles cluster according to larval ages, and the 17 dph and 32 dph microbiota overlaps (Fig. 3). The average Bray–Curtis dissimilarity between 17 dph and 32 dph larval microbiotas was only around 0.4, compared with approximately 0.9 for comparisons between other larval age groups. A one-way permutational multivariate analysis of variance (PERMANOVA) test confirmed that there were significant differences in the larval mirobiota between all age groups for both the T1 and the T2 dataset (P < 0.01), except between the 17 dph and the 32 dph larvae (P = 0.14 for the T1 and P = 0.45 for the T2 dataset respectively). To investigate the changes in larval microbiota with age at a more detailed level, we determined the total number of OTUs observed for the first time at each age and the number of OTUs that persisted from the younger larval stages (data not shown). Even though the 17 dph microbiota was very dissimilar from the 8 dph microbiota, 60 (T2) to 70% (T1) of the OTUs present at 17 dph persisted from the 8 dph microbiota. At 61 dph, 26% (T2) to 33% (T1) of the OTUs observed for the first time at 8 dph were still present in the larval microbiota. Beta-diversity of cod larval microbiota: inter-individual variation Despite the fact that larvae of a given age exhibit a characteristic microbiota, and the potential dispersal of bacteria between host individuals through the water, the individual community profiles revealed variation among individuals (Fig. 2). Generally, the dominating taxa were the same for individuals at identical age, but their relative abundance varied among individuals. For example, the Arcobacter OTU that was abundant in 17 dph and 32 dph larval microbiota constituted from 17% to 77% of the OTUs for individual larvae. Whereas the
© 2015 Society for Applied Microbiology and John Wiley & Sons Ltd, Environmental Microbiology, 17, 3914–3924
Ontogeny of cod larval microbiota # 0 # # 0 # # # # # # # # # # # # # # # 0 # # # # # 0 # # 0 # # 0 # # # 0 # # # # 0 # # #
# 0 # # 0 # # # # # # # # # # # # # # # 0 # # # # # 0 # # 0 # # 0 # # # 0 # # # # 0 # # #
Larvae 8 dph # 0 # # 0 # # # # # # # # # # # # # # # 0 # # # # # 0 # # 0 # # 0 # # # 0 # # # # 0 # # #
# 0 # # 0 # # # # # # # # # # # # # # # 0 # # # # # 0 # # 0 # # 0 # # # 0 # # # # 0 # # #
# 0 # # 0 # # # # # # # # # # # # # # # 0 # # # # # 0 # # 0 # # 0 # # # 0 # # # # 0 # # #
0 0 0 0 0 0 0 0 0 0 0 # 0 0 # 0 0 0 0 # 0 0 0 0 # 0 0 0 0 0 0 0 0 # # 0 0 # 0 0 0 0 0 # 0
0 0 0 0 0 0 0 0 0 0 0 # 0 0 # 0 0 0 0 # 0 0 0 0 # 0 0 0 0 0 0 0 0 # # 0 0 # 0 0 0 0 0 # 0
0 0 0 0 0 0 0 0 0 0 0 # 0 0 # 0 0 0 0 # 0 0 0 0 # 0 0 0 0 0 0 0 0 # # 0 0 # 0 0 0 0 0 # 0
0 0 0 0 0 0 0 0 0 0 0 # 0 0 # 0 0 0 0 # 0 0 0 0 # 0 0 0 0 0 0 0 0 # # 0 0 # 0 0 0 0 0 # 0
0 0 0 0 0 0 0 0 0 0 0 # 0 0 # 0 0 0 0 # 0 0 0 0 # 0 0 0 0 0 0 0 0 # # 0 0 # 0 0 0 0 0 # 0
# 0 # # 0 # # # # # # # # # # # # # # # 0 # # # # # 0 # # 0 # # 0 # # # 0 # # # # 0 # # #
# 0 # # 0 # # # # # # # # # # # # # # # 0 # # # # # 0 # # 0 # # 0 # # # 0 # # # # 0 # # #
Larvae 17 dph # 0 # # 0 # # # # # # # # # # # # # # # 0 # # # # # 0 # # 0 # # 0 # # # 0 # # # # 0 # # #
# 0 # # 0 # # # # # # # # # # # # # # # 0 # # # # # 0 # # 0 # # 0 # # # 0 # # # # 0 # # #
# 0 # # 0 # # # # # # # # # # # # # # # 0 # # # # # 0 # # 0 # # 0 # # # 0 # # # # 0 # # #
0 0 0 0 0 0 0 0 0 0 0 # 0 0 # 0 0 0 1 # 0 0 0 0 # 0 0 0 0 0 0 0 0 # # 0 0 # 0 0 0 0 0 # 0
0 0 0 0 0 0 0 0 0 0 0 # 0 0 # 0 0 0 0 # 0 0 0 0 # 0 0 0 0 0 0 0 0 # # 0 0 # 0 0 0 0 0 # 0
0 0 0 0 0 0 0 0 0 0 0 # 0 0 # 0 0 0 1 # 0 0 0 0 # 0 0 0 0 0 0 0 0 # # 0 0 # 0 0 0 0 0 # 0
0 0 0 0 0 0 0 0 0 0 0 # 0 0 # 0 0 0 0 # 0 0 0 0 # 0 0 0 0 0 0 0 0 # # 0 0 # 0 0 0 0 0 # 0
0 0 0 0 0 0 0 0 0 0 0 # 0 0 # 0 0 1 0 # 0 0 0 0 # 0 0 0 0 0 0 0 0 # # 0 0 # 0 0 0 0 0 # 0
# 0 # # 0 # # # # # # # # # # # # # # # 0 # # # # # 0 # # 0 # # 0 # # # 0 # # # # 0 # # #
# 0 # # 0 # # # # # # # # # # # # # # # 0 # # # # # 0 # # 0 # # 0 # # # 0 # # # # 0 # # #
Larvae 32 dph # 0 # # 0 # # # # # # # # # # # # # # # 0 # # # # # 0 # # 0 # # 0 # # # 0 # # # # 0 # # #
# 0 # # 0 # # # # # # # # # # # # # # # 0 # # # # # 0 # # 0 # # 0 # # # 0 # # # # 0 # # #
# 0 # # 0 # # # # # # # # # # # # # # # 0 # # # # # 0 # # 0 # # 0 # # # 0 # # # # 0 # # #
0 0 0 0 0 0 0 0 0 0 0 # 0 0 # 0 0 0 0 # 0 0 0 0 # 0 0 0 0 0 0 0 0 # # 0 0 # 0 0 0 0 0 # 0
0 0 0 0 0 0 0 0 0 0 0 # 0 0 # 0 0 0 1 # 0 0 0 0 # 0 0 0 0 0 0 0 0 # # 0 0 # 0 0 0 0 0 # 0
0 0 0 0 0 0 0 0 0 0 0 # 0 0 # 0 0 0 0 # 0 0 0 0 # 0 0 0 0 0 0 0 0 # # 0 0 # 0 0 0 0 0 # 0
0 0 0 0 0 0 0 0 0 0 0 # 0 0 # 0 0 0 0 # 0 0 0 0 # 0 0 0 0 0 0 0 0 # # 0 0 # 0 0 0 0 0 # 0
0 0 0 0 0 0 0 0 0 0 0 # 0 0 # 0 0 0 0 # 0 0 0 0 # 0 0 0 0 0 0 0 0 # # 0 0 # 0 0 0 0 0 # 0
# 0 # # 0 # # # # # # # # # # # # # # # 0 # # # # # 0 # # 0 # # 0 # # # 0 # # # # 0 # # #
# 0 # # 0 # # # # # # # # # # # # # # # 0 # # # # # 0 # # 0 # # 0 # # # 0 # # # # 0 # # #
Larvae 61 dph # 0 # # 0 # # # # # # # # # # # # # # # 0 # # # # # 0 # # 0 # # 0 # # # 0 # # # # 0 # # #
# 0 # # 0 # # # # # # # # # # # # # # # 0 # # # # # 0 # # 0 # # 0 # # # 0 # # # # 0 # # #
# 0 # # 0 # # # # # # # # # # # # # # # 0 # # # # # 0 # # 0 # # 0 # # # 0 # # # # 0 # # #
0 0 0 0 0 0 0 0 0 0 0 # 0 0 # 0 0 0 0 # 0 0 0 0 # 0 0 0 0 0 0 0 0 # # 0 0 # 0 0 0 0 0 # 0
0 0 0 0 0 0 0 0 0 0 0 # 0 0 # 0 0 0 0 # 0 0 0 0 # 0 0 0 0 0 0 0 0 # # 0 0 # 0 0 0 0 0 # 0
0 0 0 0 0 0 0 0 0 0 0 # 0 0 # 0 0 0 0 # 0 0 0 0 # 0 0 0 0 0 0 0 0 # # 0 0 # 0 0 0 0 0 # 0
0 0 0 0 0 0 0 0 0 0 0 # 0 0 # 0 0 0 0 # 0 0 0 0 # 0 0 0 0 0 0 0 0 # # 0 0 # 0 0 0 0 0 # 0
0 0 0 0 0 0 0 0 0 0 0 # 0 0 # 0 0 0 0 # 0 0 0 0 # 0 0 0 0 0 0 0 0 # # 0 0 # 0 0 0 0 0 # 0
# 0 # # 0 # # # # # # # # # # # # # # # 0 # # # # # 0 # # 0 # # 0 # # # 0 # # # # 0 # # #
# 0 # # 0 # # # # # # # # # # # # # # # 0 # # # # # 0 # # 0 # # 0 # # # 0 # # # # 0 # # #
Water # 0 # # 0 # # # # # # # # # # # # # # # 0 # # # # # 0 # # 0 # # 0 # # # 0 # # # # 0 # # #
# 0 # # 0 # # # # # # # # # # # # # # # 0 # # # # # 0 # # 0 # # 0 # # # 0 # # # # 0 # # #
0 0 0 0 0 0 0 0 0 0 0 # 0 0 # 0 0 0 0 # 0 0 0 0 # 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 # 0 0 # 0 0 0 0 # 0 0 0 0 # 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 # 0 0 # 0 0 0 0 # 0 0 0 0 # 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 # 0 0 # 0 0 0 0 # 0 0 0 0 # 0 0 0 0 0 0 0 0
# 0 0 # 0 0 0 0 0 # 0
# 0 0 # 0 0 0 0 0 # 0
# 0 0 # 0 0 0 0 0 # 0
# 0 0 # 0 0 0 0 0 # 0
3917
Live feed # 0 # # 0 # # # # # # # # # # # # # # # 0 # # # # # 0 # # 0 # # 0 # # # 0 # # # # 0 # # #
0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 # 0
# 0 # # 0 # # # # # # # # # # # # # # # 0 # # # # # 0 # # 0 # # 0 # # # 0 # # # # 0 # # #
0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 # 0
Bacteria, other Proteobacteria, other Gammaproteobacteria, other Colwelliaceae, Alteromonadales Idiomarinaceae, Alteromonadales Moritellaceae, Alteromonadales Alteromonadales, other Pseudoalteromonadaceae, Alteromonadales Psychromonadaceae, Alteromonadales Enterobacteriaceae, Enterobacteriales Oceanospirillaceae, Oceanospirillales Oceanospirillales, other Moraxellaceae, Pseudomonadales Pseudomonadaceae, Pseudomonadales Thiotrichales, other Piscirickettsiaceae, Thiotrichales Thiotrichaceae, Thiotrichales Vibrionaceae, Vibrionales Campylobacteraceae, Campylobacterales Burkholderiales, other Comamonadaceae, Burkholderiales Oxalobacteraceae, Burkholderiales Methylophilaceae, Methylophilales Betaproteobacteria, other Alphaproteobacteria, other Bradyrhizobiaceae, Rhizobiales Rhizobiales, other Rhodobacteraceae, Rhodobacterales Sphingomonadaceae, Sphingomonadales Lachnospiraceae, Clostridiales Clostridiaceae, Clostridiales Clostridiales, other Carnobacteriaceae, Lactobacillales Lactobacillales, other Paenibacillaceae, Bacillales Bacillales, other Planococcaceae, Bacillales Flexibacteraceae, Sphingobacteriales Sphingobacteriaceae, Sphingobacteriales Cryomorphaceae, Flavobacteriales Flavobacteriaceae, Flavobacteriales Flavobacteriales, other Bacteroidetes, other Opitutaceae, Verrucomicrobiales Actinobacteridae, Actinomycetales
