Dynamics of the Pacific oyster pathobiota during mortality episodes in Europe assessed by 16S rRNA gene profiling and a new target enrichment nextgeneration sequencing strategy

Dynamics of the Pacific oyster pathobiota during mortality episodes in Europe assessed by 16S rRNA gene profiling and a new target enrichment nextgeneration sequencing strategy Aide Lasa , 1,2 Andrea di Cesare, 3 Giovanni Tassistro, 1 Alessio Borello, 1 Stefano Gualdi, 4 Dolors Furones, 5 Noelia Carrasco, 5 Deborah Cheslett, 6 Amanda Brechon, 6 Christine Paillard, 7 Adeline Bidault, 7 Fabrice Pernet , 8 Laura Canesi, 1 Paolo Edomi, 9 Alberto Pallavicini, 9 Carla Pruzzo 1 and Luigi Vezzulli 1 * 1 Department of Earth, Environmental and Life Sciences (DISTAV), University of Genoa, Genoa, Italy. 2 Department of Microbiology and Parasitology, Universidade de Santiago de Compostela, Santiago de Compostela, Spain. 3 National Research Council-Water Research Institute (CNR-IRSA), Largo Tonolli 50, 28822 Verbania, Italy. 4 Department of Plant and Microbial Biology, University of Zürich, Zürich, Switzerland. 5 IRTA, Sant Carles e la Ràpita, Tarragona, Spain. 6 Fish Health Unit, The Marine Institute, Rinville Oranmore, Galway, Ireland. 7 Laboratoire des sciences de l’Environnement Marin, Institut Universitaire Européen de la M, Université de Bretagne Occidentale –UMR6539 CNRS/UBO/IRD/Ifremer, Plouzané, France. 8 Ifremer, Physiologie Fonctionnelle des Organismes Marins, UMR 6539 LEMAR (CNRS/Ifremer/IRD/UBO) Technopole Iroise, CS 10070, 29280, Plouzane, France. 9 Department of Life Sciences, University of Trieste, Trieste, Italy. Summary Infectious agents such as the bacteria Vibrio aestuarianus or Ostreid herpesvirus 1 have been repeatedly associated with dramatic disease outbreaks of Crassostrea gigas beds in Europe. Beside roles played by these pathogens, microbial infections in C.gigas may derive from the contribution of a larger number of microorganisms than previously thought, according to an emerging view supporting the polymicrobial nature of bivalve diseases. In this study, the microbial communities associated with a large number of C.gigas samples collected during recurrent mortality episodes at different European sites were investigated by real-time PCR and 16SrRNA genebased microbial profiling. A new target enrichment nextgeneration sequencing protocol for selective capturing of 884 phylogenetic and virulence markers of the potential microbial pathogenic community in oyster tissue was developed allowing high taxonomic resolution analysis of the bivalve pathobiota. Comparative analysis of contrasting C.gigas samples conducted using these methods revealed that oyster experiencing mortality outbreaks displayed signs of microbiota disruption associated with the presence of previously undetected potential pathogenic microbial species mostly belonging to genus Vibrio and Arcobacter. The role of these species and their consortia should be targeted by future studies aiming to shed light on mechanisms underlying polymicrobial infections in C.gigas. Introduction Recent advances in DNA sequencing technology is enabling new quantitative insights into the microbial community diversity associated with human and animal tissues (Yatsunenko et al., 2012; Yarza et al., 2014). It is now recognized that hostassociated microbial communities (also named the microbiota) (Lederberg and McCray, 2001) are playing an important role in animal health by providing prominent services ranging from nutrient processing to protection from diseases. Marine bivalves host high microbial abundance and diversity and alteration of the microbiota due to stressful conditions and/or environmental changes was previously linked with a condition of a compromised health status and susceptibility to diseases (Lokmer and Wegner, 2015). Diseases affecting the Pacificoyster(Crassostrea gigas) have been rising over the past decades, representing a Received 29 May, 2019; revised 12 July, 2019; accepted 14 July, 2019. *For correspondence. E-mail [email protected]; Tel. +39-01035338018. © 2019 The Authors. Environmental Microbiology published by Society for Applied Microbiology and John Wiley & Sons Ltd. This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited. Environmental Microbiology (2019) 21(12), 4548–4562 doi:10.1111/1462-2920.14750 significant threat for commercial exploitation of both farmed and natural stocks (Alfaro et al., 2018). In Europe, mass mortality episodes of C.gigas in farming areas are attributed to complex interactions among oysters, microbial pathogens and environmental variables (Pernet et al., 2012). In particular, stressful environmental conditions, such as warmer seawater temperatures, were observed to favour a shift in C.gigas bacterial communities toward pathogendominated communities also promoting colonization by secondary opportunistic pathogens and non-resident microbial species (Lemire et al., 2015; Lokmer and Wegner, 2015). Mortality of C.gigas spat and juvenile is mainly associated with infection by the herpes-like virus Oyster Herpesvirus type 1 (OsHV-1) usually when water temperature reaches 16C (Segarra et al., 2010). Infected oysters show a reduction in feeding and swimming activities and mortality can reach 100% in a few days. In contrast, mortality of adult oysters in Europe has been mainly reported in association with the detection of the bacterium Vibrio aestuarianus (Travers et al., 2015). In this case, infection seems to last for a long period, reaching a cumulative mortality rate up to ~30% at the end of the farming cycle. Although these pathogens have been identified to play a role in oyster diseases, there is an emerging view that microbial infections, in marine bivalves, may derive from the contribution of different microbial species/strains that act as a ‘community of pathogens’(hereafter referred to as the ‘pathobiota’) rather than a single species/strain as the only etiological agent (Lemire et al., 2015; de Lorgeril et al., 2018). Under this perspective evidence has been provided supporting the view that oyster infections might be seen as infectious disorders caused by the contribution of a larger number of pathogens (e.g. populations or consortia) than previously thought (Lemire et al., 2015). The question now is no longer whether microorganisms are involved in the pathogenesis of such diseases, but which specific microbial species or strains are involved. Although addressing this issue will be of great value in improving the general understanding of bivalve diseases and to drive future studies to unravel their mechanisms, answering the above question is not straightforward. In fact, accurate phylogenetic analysis of the bivalve microbiota, including the potential pathogenic community, has been historically hampered by methodological constrains, such as the use of traditional culture-dependent protocols, as a large fraction of the bivalve-associated microbial community (e.g. unculturable bacteria) may go undetected using these methods (Garnier et al., 2007). More recently, 16S rRNA gene-based analysis of the microbial diversity, based on next-generation sequencing protocols, was employed for taxonomic identification of bivalve-associated bacteria (Lokmer and Wegner, 2015; Lokmer et al., 2016). However, although this approach allows us to overcome the cultivation bias, 16S rRNA profiling could not be successfully applied for high taxonomic resolution analysis of the microbial community (e.g. taxonomic resolution to the level of species or strains) due to the lack of phylogenetic value of the 16S rRNA gene for many bacterial groups (Rajendhran and Gunasekaran, 2011). This is particularly true for bacteria belonging to the Vibrio genus that are thought to play a primary role in oyster infections and mortality outbreak (Vezzulli et al., 2015). The same holds true for shotgun metagenomic techniques due to difficulties in separation of host DNA from microbial DNA resulting in non-optimal detection and taxonomic resolution of microbial species. Thus, investigating polymicrobial nature of OsHV-1 disease and the unknown nature of V. aestuarianus infection affecting C.gigas deserve new tools. The aim of this study was to overcome these constrains and be able to provide new insights into the high-level (species) taxonomic composition of the oyster pathobiota in C.gigas samples collected during mortality outbreaks in Europe. To this end, a new target enrichment next-generation sequencing approach was developed and applied to target 884 phylogenetic and virulence markers of the potential pathogen community associated to bivalve tissues. 16S rRNA gene profiling of the microbial community was also conducted on a large number of contrasting (e.g. infected vs not infected) C.gigas samples to support the ‘pathobiota’analysis and to extensively investigate the patterns and dynamics of oyster microbiota during mortality events. Results OsHV-1 and V.aestuarianus detection in C.gigas samples A total of 525 C.gigas samples collected from 2016 to 2017 in three European sites, i.e., Ebro delta, Dungarvan Bay and Bay of Brest (Supplementary Material and Methods, Fig. S1, Table S1) were screened for the presence of OsHV-1 and V.aestuarianus using quantitative real-Time PCR (Table 1). In general, samples collected during oyster mortality episodes tested positive for at least one of the two pathogens suggesting a strong link between presence of these microorganisms in the oyster tissues and development of disease (Labreuche et al., 2010; Segarra et al., 2010). In particular, V.aestuarianus was found associated to adult C.gigas mortality observed in the Ebro delta on 13th April 2016 and 31st May 2017 and in Dungarvan Bay on 3rd October 2016 (Table 1). OsHV-1 DNA was detected in spat C.gigas sampled during 12 mortality episodes in the Ebro delta, Dungarvan Bay and Bay of Brest (Table 1). © 2019 The Authors. Environmental Microbiology published by Society for Applied Microbiology and John Wiley & Sons Ltd., Environmental Microbiology,21, 4548–4562 Pacific oyster pathobiota 4549 Table 1. Crassostrea gigas samples collected at different European sites and analysed in this study (n+ number of samples/total number of analysed samples scoring positive to species-specific qPCR targeting the pathogen V.aestuarianus or OshV-1 in oyster tissues). Sampling date Code Age Origin Ploidy Mortality (%) OshV-1 (Real-time PCR) V. aestuarianus (Real-time PCR) Ebro Delta (Spain) 13_04_2016 EbDva13apr16/EbH13apr16 Adult Nature Diploid 23% (Up to 50%) 0+/30 22+/30 26_04_2016 EbDoshV26apr16(c3)/EbH26apr16(c4) Spat Nature Diploid 76% (Up to 90%) 17+/30 0+/30 26_04_2016 EbDoshV26apr16(c5)/EbDva26apr16(c6)/EbH26apr16(c6) Spat Nature Diploid 46.6% (Up to 80%) 21+/30 4+/30 19_07_2016 EbDoshV19jul16/EbH19jul16 Spat Hatchery Diploid 3% (light mortality) 1+/30 0+/30 19_07_2016 EbH19jul16(c2) Spat Nature Diploid No mortality 0+/30 0+/30 24_11_2016 EbH24nov16/EbDoshV24nov16/EbDva24nov16 Spat Nature Diploid 30% recent mortality 7+/30 1+/30 25_01_2017 EbH25jan17 Spat Hatchery Diploid No mortality 0+/30 0+/30 26_01_2017 EbH26jan17 Spat Hatchery Diploid No mortality 0+/30 0+/30 05_05_2017 EbDoshV5may17/EbH5may17 Spat Hatchery Triploid 87.83% mortality 10+/18 0+/18 05_05_2017 EbH5may17(c1) Spat Hatchery Triploid No mortality 0+/30 0+/30 31_05_2017 EbDva31may17/EbH31may17 Spat Hatchery Triploid 85% mortality 0+/30 9+/30 Dungarvan Bay (Ireland) 05_07_2016 DrDva7jul16/DrH7jul16 Adult Hatchery Triploid 20% 0+/30 2+/30 05_07_2016 DrDoshV7jul16SPAT/DrDva7jul16SPAT/DrH7jul16SPAT Spat Hatchery Triploid 70%–100% 25+/30 3+/30 03_10_2016 DrDva3oct16/DrH3oct16 Adult Hatchery Triploid End of mortality 0+/30 11+/30 03_10_2016 DrDoshV3oct16SPAT/DrDvaV3oct16SPAT/DrH3oct16SPAT Spat Hatchery Triploid End of mortality 12+/30 1+/30 10_01_2017 DrH10jan17 Adult Hatchery Triploid No mortality 0+/30 0+/30 10_01_2017 DrH10jan17SPAT Spat Hatchery Triploid No mortality 0+/30 0+/30 Bay of Brest (France) 10_07_2017 BbDoshV10jul17/BbDva10jul17/BbH10jul17 ad/sp/jv Hatchery Diploid Mortality 2+/12 2+/12 05_10_2017 BbDoshV5oct17/ BbDva5oct17/BbH5oct17 ad/sp/jv Hatchery Diploid No mortality 3+/15 1+/15 Code legend = [Eb, Ebro delta; Dr, Dungarvan bay; Bb, Bay of Brest]/[D, Infected; H, Not Infected]/[va, Vibrio aestuarianus; oshV, Ostreid herpesvirus 1]/[collection date]. © 2019 The Authors. Environmental Microbiology published by Society for Applied Microbiology and John Wiley & Sons Ltd., Environmental Microbiology,21, 4548–4562 4550 A. Lasa et al. Microbiota analysis One hundred one contrasting C.gigas samples were selected in the Ebro Delta (n= 50), Dungarvan bay (n=40) and the Bay of Brest (n= 11) for microbiota analysis based on results from PCR screening analysis. In the experimental design, most interesting sampling periods were chosen (e.g. periods with high mortality or absence of mortality), and for each period, 10 contrasting individual oyster samples (five samples infected by OsHV-1 or V.aestuarianus and five noninfected control samples) were selected. Sequencing of the samples by Ion Torrent technology produced more than 14.000.000 amplicon sequence reads spanning the V4 hypervariable region of the bacterial 16S rRNA gene. Barcode and adapters sequences were removed and raw sequences were trimmed to minimize bias associated with PCR amplification of target genes. In particular, reads that contained one or more ambiguous bases, had errors in the barcode or primer sequence, were atypically short (100 bp), and had an average quality score <0.05 were removed from the data set. This process produced 13.143.153 trimmed reads corresponding on average to 146.000 sequence reads per analysed sample. In general, the composition of the bacterial community associated to C.gigas was dominated by the classes of Gamma and Alphaproteobacteria accounting on average for 28% and 15% of the total bacterial diversity followed by Epsilonproteobacteria (11%), Mollicutes (10%) and Flavobacteria (9%) (Fig. 1). Rarefaction curves computed for total operational taxonomical units (OTUs) abundance almost reached a plateau indicating that sequencing effort was good enough to describe the majority of phylotypes in most of the samples (Fig. S2). Generally, alpha diversity was lower in adult oysters infected by V.aestuarianus whilst it was higher in oyster spat infected by OsHV-1 (Fig. S2). Different physiological or environmental conditions including animal health status and age, geographic location and season significantly influenced the composition of the oyster microbiota as showed by beta diversity analysis and PERMANOVA testing (p< 0.01) (Fig. 2). Analysis of contrasting bivalve samples based on the presence/absence of microbial pathogens showed that a significant shift in the microbiota community was observed in adult oyster infected by V.aestuarianus (Fig. 2). In particular, an increase in abundance of bacteria belonging to the genus Vibrio and Arcobacter was observed in Fig. 1. Taxonomical composition at the class (A) and genus (B) level of 101 C.gigas microbiota grouped by oyster health status (H, not infected oysters; Va, Vibrio aestuarianus-infected oysters; OshV1, Ostreid herpesvirus 1-infected oysters) and geographic locations (Eb, Ebro delta; Dr, Dungarvan bay; Br, Bay of Brest), as revealed by sequencing of the V4 hypervariable region of the 16S rRNA gene. [Color figure can be viewed at wileyonlinelibrary.com] © 2019 The Authors. Environmental Microbiology published by Society for Applied Microbiology and John Wiley & Sons Ltd., Environmental Microbiology,21, 4548–4562 Pacific oyster pathobiota 4551 V.aestuarianus-infected oysters compared to either V. aestuarianus non-infected or OsHV-1 infected oysters. In oyster spats infected by OsHV-1 an increase in the Vibrio fraction was also observed compared with noninfected oysters. The analysis of the core microbiota showed that the microbial community of oyster samples is dominated by two main bacterial groups, namely Vibrio andanuncultured bacterial group (uncultured-053) (Supplementary Material and Methods, Table S2). When analysing adult C.gigas samples infected with V.aestuarianus, core taxa were assigned to Arcobacter and Vibrio genera. In contrast, core taxa of V.aestuarianus non-infected samples were assigned to the uncultured-053 group. When OsHV-1 DNA was detected in spat oysters, the core microbiota included Vibrio but also the uncultured-053 group. Interestingly, the Fig. 2. Comparative analysis of C.gigas microbiota (Beta diversity analysis) under different conditions: health status (A), geographic location (B), season (C) and animal age (D) (Va, Vibrio aestuarianus-infected oysters; OshV1, Ostreid herpesvirus 1-infected oysters). [Color figure can be viewed at wileyonlinelibrary.com] © 2019 The Authors. Environmental Microbiology published by Society for Applied Microbiology and John Wiley & Sons Ltd., Environmental Microbiology,21, 4548–4562 4552 A. Lasa et al. core microbiota of oyster spats and adult was clearly different. While Vibrio and Arcobacter were members of the core microbiota in adult oysters, in oyster spats the core group was dominated by Vibrio, uncultured-053, Haliea and Marinicella. Differences in the core microbiota were also evident across the different geographic locations. In oysters from Dungarvan bay the core groups were Sulfitobacter, uncultured-053, Arcobacter,Pseudoalteromonas, Marinicella,Vibrio and Borrelia, while in Ebro Delta the core microbiota was formed only by Vibrio and uncultured-053. On the other hand, the core microbiota of the samples from the Bay of Brest was more diverse than that of oysters from other sites, with more than 20 different bacterial groups (Arcobacter,Vibrio,Mycoplasma, Winogradskyella,Haliea,Sulfitobacter,Polaribacter, Marinicella,Lutimonas,Aquibacter,Roseovarius, Table 2. Phylogenetic and virulence markers targeted by the target enrichment NGS protocol to study the potential microbial pathogenic community (pathobiota) in C.gigas tissues (Rosenberg et al., 2007; Travers et al., 2015). Target Taxonomy Main host Marker gene No. of allelic variants Length (nt) Vibrio spp. Vibrio spp. Human, Animal gyrB 243 400 Vibrio spp. Human, Animal recA 204 400 Vibrio spp. Human, Animal atpA 133 400 Vibrio spp. Human, Animal dnaJ 56 400 Vibrio spp. Human, Animal pyrH 113 400 V.tasmaniensis Crassostrea gigas LGP32 probes 10 400 V.cholerae O1 El Tor Human ctxA 1 777 V.cholerae O1 El Tor Human ctxB 1 375 V.cholerae O139 Human ctxA-B 1 938 V.cholerae O1 el Tor Human tcpA 1 675 V.cholerae O1 classical Human tcpA 1 675 V.cholerae O1 el Tor Human rstR 1 339 V.cholerae O1 classical Human rstR 1 336 V.cholerae O139 Human wbf 1 449 V.cholerae O1 el Tor Human gbpA 1 400 V.parahaemolyticus Human toxR 1 552 V.parahaemolyticus Human tdh,trh 3 570 V.vulnificus Human vvhA 1 1416 V.vulnificus Human rtxA1 1 400 V.tasmaniensis Crassostrea gigas vsm (LGP32 strain) 1 1824 V.tasmaniensis Crassostrea gigas ompU (LGP32 strain) 1 400 V.aestuarianus Crassostrea gigas vam 1 1836 V.tapetis Ruditapes philippinarum djlA 1 1826 V.coralliilyticus Paramuricea clavata vcpA 15 1824 V.harveyi Stony corals vhhA 1 1260 V.crassostreae Crassostrea gigas R-5.71 2397 V.tubiashii Crassostrea gigas Metalloprotease 1 1821 Vibrio spp. Human, Animal MSHA 12 400 Arcobacter spp. Arcobacter spp. Human gyrB 27 400 Nocardia crassostrea Nocardia crassostrea Crassostrea gigas,Ostrea edulis rpoB,hsp65,gyrB 3 400 Marteilia refringens Marteilia refringens Ostrea edulis,Mytilus edulis,M. galloprovincialis ITS1O, ITS1M, probe 3 400 Bonamia ostreae Bonamia ostreae Ostrea edulis 5.8S-ITS rDNA, hsp90,act1 3 400 OsHV-1 Ostreid herpesvirus 1 Crassostrea gigas C2/C6 (2), IA1-IA2, orf4, Hyp. Protein, RING finger protein gene 7 400 ORF100 1 198 C9-C10 1 197 B3-B4 1 207 Enterococcus spp. E.faecalis,E.faecium,E. avium,E.gallinarum,E. casseliflavus,E.durans, E.raffinosus,E.mundtii Human atpA 8 400 Roseovarius crassostrea Roseovarius crassostrea Crassostrea virginica dnaJ,pyrH 6 400 Escherichia coli Escherichia coli Human dnaJ,pyrH,atpA,gyrB 4 400 Aspergillus sydowii Aspergillus sydowii Gorgonia ventalina, Human TUB2, trpC, ITS, calmodulin gene 4 400 Aurantimonas coralicida Aurantimonas coralicida Corals atpD,gyrB,recA,rpoB 4 400 Serratia marcescens Serratia marcescens Acropora palmata gyrB,recA,dnaJ 3 400 Pseudoalteromonas sp. Pseudoalteromonas sp. Rhopaloeides odorabile gyrB 1 400 Total 884 29,292 © 2019 The Authors. Environmental Microbiology published by Society for Applied Microbiology and John Wiley & Sons Ltd., Environmental Microbiology,21, 4548–4562 Pacific oyster pathobiota 4553 uncultured-104, uncultured-079, uncultured-053, uncultured- 015, OM60(NOR5) clade, uncultured bacterium-088, uncultured-072, uncultured-050 and uncultured bacterium- 248) probably reflecting the smaller number of samples analysed in this site. Seasonality also affected the composition of the core microbiota, and a transition of specific genera was identified (e.g. Sulfitobacter and Marinicella from cold seasons to warmer seasons) (Supplementary Material and Methods, Table S2). Pathobiota analysis A target enrichment next-generation sequencing protocol was for the first time applied for high taxonomic resolution analysis of the bivalve pathobiota on a total of 12 selected contrasting C.gigas samples collected during mortality episodes (e.g. C.gigas samples infected and noninfected by V.aestuarianus or OsHV-1) and in the absence of mortality. A mock community sample (positive control) and a nuclease-free water sample (negative control) were also included in the analysis (see methods section for details). To run the protocol, a ‘pathobiota’ sequence database containing 884 phylogenetic and virulence markers of the bivalve microbial pathogen community was built and used to produce a total of 12.114 biotinylated RNA baits for selective capturing of target DNA via hybridization as described in the methods section (Table 2). Sequencing of enriched DNA libraries Fig. 3. Results from target enrichment NGS analysis investigating C.gigas pathobiota in contrasting samples infected or not infected (control) by the bacteria V.aestuarianus. Relative abundance is calculated from the number of reads specifically mapping on target sequences and expressed as percentage (see main text for details on mapping parameters settings). Presence of virulence genes is also shown. [Color figure can be viewed at wileyonlinelibrary.com] © 2019 The Authors. Environmental Microbiology published by Society for Applied Microbiology and John Wiley & Sons Ltd., Environmental Microbiology,21, 4548–4562 4554 A. Lasa et al. produced a total of 67.614.544 sequence reads corresponding to about 5.634.545 reads for each sample. On average, less than 5% of the reads mapped against reference sequences from the pathobiota database and were used to produce consensus sequences for pathobiota taxonomic classification with the NCBI BLAST function. Results of the target enrichment protocol allowed the detection and relative quantification of members of the bivalve pathogen community (pathobiota) associated to oyster tissues with a taxonomic resolution up to the species level. Generally, although the oyster pathobiota was composed of different species in all samples a dominance of primary pathogens such as V.aestuarianus and OsHV- 1 was observed in C.gigas samples collected during mortality episodes linked to these pathogens (Figs 3 and 4). In particular, adult C.gigas samples collected during mortality episodes linked to Vibrio aestuarianus outbreaks in Ebro Delta and Dungarvan bay (e.g. samples EbDva13apr16-1, DrDva3oct16-2, DrDva3oct16-5) showed a large proportion of reads specifically mapping on V.aestuarianus phylogenetic (e.g. gyrB,recA,atpA, dnaJ and pyrH) and virulence (e.g. vam) marker sequences accounting on average for more than 40% of total read sequences. Accordingly, consensus sequences obtained from read mapping were unequivocally assigned to the species V.aestuarianus by BLAST sequence analysis (Fig. 3). In addition, a significant fraction of reads common to all V. aestuarianus infected samples were specifically mapping on Arcobacter haliotis marker sequences (gyrB). Albeit at lower relative abundance other species identified in these samples included V.alginolyticus,V.coralliilyticus,V.crassostreae,V.mediterranei,V. toranzoniae,V.splendidus and V.tasmaniensis (Fig. 3). Virulence genes such as metalloproteases (V.splendidus vsm,V.coralliilyticus vcpA)andompU of V.tasmaniensis were also detected (Fig. 3). Fig. 4. Results from target enrichment NGS analysis investigating C.gigas pathobiota in contrasting samples infected or not infected (control) by Ostreid herpesvirus 1 (OshV1). Samples DrH10jan17-3 and DrH10jan17-SPAT2 were collected during period of no mortality. Relative abundance is calculated from the number of reads specifically mapping on target sequences and expressed as percentage (see main text for details on mapping parameters settings). Presence of virulence genes is also shown. [Color figure can be viewed at wileyonlinelibrary.com] © 2019 The Authors. Environmental Microbiology published by Society for Applied Microbiology and John Wiley & Sons Ltd., Environmental Microbiology,21, 4548–4562 Pacific oyster pathobiota 4555 Similarly, C.gigas samples collected during mortality episodes of spat oysters linked to an OsHV-1 outbreak in Dungarvan bay (e.g. samples DrDoshV7jul16SPAT-4, DrDoshV7jul16SPAT-5) showed a large number of reads (on average more than 60% of the total read sequences) specifically mapping on OsHV-1 marker sequences (e.g. C2/C6 (2), IA1-IA2, orf4, Hyp. Protein, RING finger protein gene, ORF100, C9-C10, B3-B4). An additional fraction of reads found in these samples pointed to the presence of bacterial species belonging to the Splendidus clade including V.splendidus,V.tasmaniensis,V.cyclitrophicus and V.lentus. Other bacterial species found in OsHV-1 infected oyster included V.coralliilyticus,V.alginolyitucs,V. crassostreae,V.cortegadensis and V.pomeroyi (Fig. 4). Virulence genes (e.g. V.tasmaniensis vsm and ompU)and the genetic element R5.7 were also found (Fig. 4). Generally, no reads were observed to map on reference sequences above defined thresholds in control samples, suggesting that target bacterial concentrations were either absent or below the limit of detection in these samples (Figs 3 and 4). An exception to this was observed for sample DrH7jul16SPAT-4 collected in Dungarvan Bay during an OsHV-1 outbreak in July 2016 where a diversified pathobiota community dominated by members of the Splendidus clade such as V.splendidus and V.tasmaniensis was found (Fig. 4). Virulence genes (e.g. vsm and ompU) and the genetic element R5.7 were detected in this sample as well as a number of other Vibrio species including V.cyclitrophicus,V.cortegadensis, V.gallaecicus,V.crassostreae,V.alginolyticus,V.atlanticus and V.lentus (Fig. 4). In contrast, no target sequences were detected in samples DrH10jan17SPAT- 2 and DrH10jan17-3 collected in Dungarvan bay in winter in the absence of significant mortality. All species included in the mock community sample were correctly identified by the target enrichment protocol developed in this study (Supplementary Material and Methods, Fig. S3), whilst no sequences were obtained from analysis of the negative control ruling out laboratory contamination. Discussion Microbiota composition of healthy C.gigas oysters in European farming sites The composition of C.gigas microbiota in this study was dominated by the classes of Gamma and Alphaproteobacteria and was highly variable in relation to health status, geographic location, season and oyster age (Fig. 2). Accordingly, many transient and opportunistic microbial taxa appeared to dominate oyster microbiota while core microbial communities were restricted to only a few, most probably resident bacteria, such as Vibrio and the uncultured-053 group. This may be linked to the filter feeding behaviour of oysters, which expose the animals to colonization by complex and highly variable microbial communities found in the seawater environment. Mycoplasmataceae,Arcobacter,Synechococcus and Spirochaetaceae dominated the microbiota in healthy (non-infected) oysters suggesting these taxa might play a beneficial role in oyster fitness and health status (King et al., 2019). Interestingly, Arcobacter and Mycoplasmataceae were found to represent an abundant fraction of the microbiota in the Eastern oyster (Crassostrea virginica) and Chilean oyster (Tiostrea chilensis) respectively, although their role within the host is still largely unknown (Romero et al., 2002; King et al., 2012). Similarly, Synechococcus was observed in the cytosol of the digestive gland, connective tissue, mantle, and gonad of C.gigas and a host-endobiont relationship between C.gigas and Synechococcus cells was recently suggested (Avila-Poveda et al., 2014). In addition, since cyanoprokaryotes are particularly abundant in the digestive gland (Avila-Poveda et al., 2014) and colonize the oyster microbiota especially in the summer period when seawater concentrations of these bacteria are higher (Fig. 1), it can be speculated that their accumulation in oyster tissues may directly derive from the natural environment. Host-associated spirochetes have been found mainly in the digestive tract of eukaryotes including molluscs (Romero and Espejo, 2001; Duperron et al., 2007) but again little is known on their association and role with their C.gigas hosts. Changes in the microbiota community structure associated to C.gigas infections during mass mortality episodes in Europe Comparative analysis of healthy versus infected C.gigas samples clearly showed that infected oysters displayed signs of community structure disruption and were characterized by a low diversity and proliferation of few bacterial taxa. This was particularly evident for V.aestuarianusinfected oysters where dominance by bacteria belonging to the genus Vibrio and Arcobacter resulted in low microbial diversity compared with healthy oysters. In the case of OsHV-1 infected oyster, loss of Alpha diversity was less clear and mostly linked to proliferation of Vibrio and a significant decline of some bacterial taxa such as cyanoprokaryotes (e.g. Synechococcus) (Fig. 1). Loss of microbiota diversity and proliferation of few OTUs (‘dysbiosis’) has previously been linked with impaired health in oysters (Green and Barnes, 2010; King et al., 2019), including C.gigas (Garnier et al., 2007; Lokmer and Wegner, 2015). It is thus apparent that specific microbial taxa and especially members of the Vibrio genus are likely to play a role in affecting oyster health status during disease outbreaks. Nevertheless, whether © 2019 The Authors. Environmental Microbiology published by Society for Applied Microbiology and John Wiley & Sons Ltd., Environmental Microbiology,21, 4548–4562 4556 A. Lasa et al.