Immune-Suppression by Oshv-1 Viral Infection Causes Fatal Bacteraemia in Paci Fic Oysters

ARTICLE
DOI: 10.1038/s41467-018-06659-3 OPEN
Immune-suppression by OsHV-1 viral infection

causes fatal bacteraemia in Pacific oysters
Julien de Lorgeril1, Aude Lucasson1, Bruno Petton2, Eve Toulza1, Caroline Montagnani1, Camille Clerissi1,
Jeremie Vidal-Dupiol1, Cristian Chaparro1, Richard Galinier1, Jean-Michel Escoubas1, Philippe Haffner1,
Lionel Dégremont3, Guillaume M. Charrière1, Maxime Lafont1, Abigaïl Delort1, Agnes̀ Vergnes1,
Marlène Chiarello 4, Nicole Faury3, Tristan Rubio1, Marc A. Leroy1, Adeline Pérignon5, Denis Régler5,
Benjamin Morga3, Marianne Alunno-Bruscia2, Pierre Boudry6, Frédérique Le Roux7,
1234567890():,;
Delphine Destoumieux-Garzόn1, Yannick Gueguen 1 & Guillaume Mitta1
Infectious diseases are mostly explored using reductionist approaches despite repeated
evidence showing them to be strongly influenced by numerous interacting host and envir-
onmental factors. Many diseases with a complex aetiology therefore remain misunderstood.
By developing a holistic approach to tackle the complexity of interactions, we decipher the
complex intra-host interactions underlying Pacific oyster mortality syndrome affecting
juveniles of Crassostrea gigas, the main oyster species exploited worldwide. Using experi-
mental infections reproducing the natural route of infection and combining thorough mole-
cular analyses of oyster families with contrasted susceptibilities, we demonstrate that the
disease is caused by multiple infection with an initial and necessary step of infection of
oyster haemocytes by the Ostreid herpesvirus OsHV-1 µVar. Viral replication leads to the
host entering an immune-compromised state, evolving towards subsequent bacteraemia by
opportunistic bacteria. We propose the application of our integrative approach to decipher
other multifactorial diseases that affect non-model species worldwide.
1 IHPE, Université de Montpellier, CNRS, Ifremer, Université de Perpignan Via Domitia, Place E. Bataillon, 34095 Montpellier, France. 2 LEMAR UMR 6539,
UBO/CNRS/IRD/Ifremer, 11 presqu’île du vivier, 29840 Argenton-en-Landunvez, France. 3 Laboratoire de Génétique et Pathologie des Mollusques Marins,
Ifremer, Avenue du Mus de Loup, 17930 La Tremblade, France. 4 Marine Biodiversity, Exploitation and Conservation (MARBEC), Université de Montpellier,
CNRS, IRD, Ifremer, Place E. Bataillon, 34095 Montpellier, France. 5 CRCM, Comité de la Conchyliculture de Méditerranée, Quai Baptiste Guitard, 34140
Mèze, France. 6 LEMAR UMR6539, CNRS/UBO/IRD/Ifremer, ZI pointe du diable, CS 10070, F-29280 Plouzané, France. 7 Sorbonne Universités, UPMC Paris
06, CNRS, UMR 8227, LBI2M, Ifremer, Station Biologique de Roscoff, CS 90074, F-29680 Roscoff, France. These authors contributed equally: Julien de
Lorgeril and Aude Lucasson. Correspondence and requests for materials should be addressed to Y.G. (email: ygueguen@ifremer.fr)
or to G.M. (email: mitta@univ-perp.fr)
NATURE COMMUNICATIONS | (2018)9:4215 | DOI: 10.1038/s41467-018-06659-3 | www.nature.com/naturecommunications 1

ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/s41467-018-06659-3
F
or decades, methodological limitations have restricted the Results
study of infectious diseases to simplified experimental Production of oyster families with contrasted phenotypes. To
pathosystems in which the influences of host and pathogen characterize the complex dynamics determining the outcome of
diversity and biotic and abiotic environments have been mini- Pacific oyster mortality syndrome, we produced 15 biparental
mized. Such reductionist approaches have made diseases with oyster families with highly contrasted resistance phenotypes with
complex aetiologies difficult to characterize. Thus, there is an regard to the disease. These families were produced from genitors
incomplete understanding of diseases in which a conserved that had experienced different selective filters. The genitors hav-
consortium of micro-organisms co-operates to induce patho- ing experienced a high level of natural selection by the disease
genesis, diseases involving pathogens that cause immune defi- were either wild oysters collected from farming areas or
ciency followed by secondary infections, and diseases that are oysters issued from mass selection programmes28. The genitors
influenced by a series of host and environmental factors. that had experienced low-level natural selection were wild oysters
There is a lack of understanding of some diseases triggering recruited outside farming areas (Supplementary Table 1). The
recurrent mass mortalities in non-model species of ecological juvenile offspring of the 15 biparental families produced were
and/or economic interest, such as pollinators, corals and marine subjected to an ‘natural’ experimental infection mimicking dis-
molluscs1–3. These dramatic epizooties remain incompletely ease transmission in the wild12,13,29 (Supplementary Fig. 1). High
characterized because epidemiological descriptions require hol- variability in the dynamics of mortalities and percentages
istic approaches to decipher the whole pathosystem. of survival (ranging from 0 to 97.4% at 330 h post-infection) were
The objective of the present work was to examine one disease observed among families (Fig. 1). Two families showing highly
of complex aetiology affecting one of the most utilized inverte- contrasted phenotypes, Susceptible Family 11 (SF11) and Resistant
brate species in the world, the Pacific oyster Crassostrea gigas, Family 21 (RF21) (Mantel–Cox log-rank test, p < 0.0001), were
applying a holistic approach to decipher the phenomenon. This selected for the first set of molecular and histological analysis.
species has been introduced from Asia to numerous countries During the experimental infection, SF11 oysters mortality began at
and is now the main farmed oyster species worldwide4. Intro- 66 h and increased dramatically. At the end of the experiment
duced to France in the 1970s, C. gigas suffers mass mortalities (330 h), only 0.7% of SF11 oysters had survived. In contrast,
associated with complex interactions between the host, the almost all oysters of the RF21 family (97.4%) survived following
environment and pathogens5. The severity of these mortality exposure to the same infectious environment. The survival rates
outbreaks has dramatically increased since 2008. They mainly of SF11 and RF21 oysters were also measured concomitantly in a
affect juvenile stages, decimating up to 100% of young oysters in batch of oysters left on oyster farms; they showed identical
French farms6. In recent years, this mortality syndrome, called phenotypes, with 2 and 98.1% survival after 384 h of exposure to
Pacific oyster mortality syndrome (POMS)7, has become pan- the infectious environment for SF11 and RF21, respectively. These
zootic, being observed in all coastal regions of France and results confirm that the disease that developed in our experi-
numerous other countries worldwide6,8. mental set-up resembles the disease contracted in the natural
Research efforts have revealed a series of factors contributing to environment with the same outcomes. Thereafter, the dynamics
the disease, including infectious agents interacting with seawater of the disease in SF11 and RF21 oysters were investigated by
temperature and oyster genetics6,9–14. The dramatic increase in thorough molecular analyses.
mortality since 2008 coincided with the recurrent detection of
Ostreid herpesvirus (OsHV-1) variants in moribund oysters both
in France14–16 and worldwide6,17–20. This increase has naturally OsHV-1 µVar infection occurs early in disease development.
driven research efforts to focus on the viral aetiology of the dis- OsHV-1 load and transcriptional activity were monitored in
ease. However, the involvement of other aetiological agents is SF11 and RF21 oysters throughout the experiment using qPCR
suspected. In particular, strains assigned to the genus Vibrio have and RNA-seq, respectively (Fig. 2a). Whereas OsHV-1 DNA and
been shown to be associated with the disease21. Among these, a RNA were detected in both families as early as 12 h after the
Vibrio crassostreae population carrying a plasmid required for beginning of experimental infection, very intense replication
virulence has been repeatedly identified in diseased oysters22. occurred in SF11 oysters, with DNA and RNA levels 3-log higher
Recent studies have also proposed that the stability of the natural
bacterial microbiota, which is abundantly present in healthy
oysters, influences their resistance to stress, or invasion by 100 RF21
pathogens23,24. However, the roles of host genetics, pathogens, RF23
RF48
and opportunistic and commensal microbes have been studied 80
Percent survival
F37
with a focus on a restricted number of factors tested in F28
isolation10,21–27. Thus, the dynamics, relative weight and inter- 60 F42
F44
actions of these different parameters in the disease remain to be F33
established. 40 F32
F9
In the present study, we concomitantly characterize the F2
transcriptional responses of oysters and the dynamics of their 20 F1
SF14
associated micro-organisms after exposure to an infectious SF15
environment using an experimental infection system that repro- 0 SF11
duced the natural route of infection in biparental families of 0 100 200 300
Oyster Resistance
oysters selected to display contrasted phenotypes (susceptible vs. Time (hours) families
resistant). This holistic approach allows us to establish that

0 6 12 24 48 60 72 Hours
OsHV-1 replication in haemocytes is the initial step of the
infectious process leading to an immune-compromised state of Fig. 1 Production of oyster biparental families with contrasted resistance
the host, which evolves towards subsequent bacteraemia by phenotypes. Kaplan–Meier survival curves of the 15 families of recipient
opportunistic bacteria. By identifying critical intra-host interac- oysters (n = 1000 for each family) during the ‘natural’ experimental
tions between micro-organisms and host immunity, this study infection. At each time indicated on the arrow, 3 triplicates of 10 oysters
cracks the code of POMS. were sampled from each tank for further molecular analysis
2 NATURE COMMUNICATIONS | (2018)9:4215 | DOI: 10.1038/s41467-018-06659-3 | www.nature.com/naturecommunications

NATURE COMMUNICATIONS | DOI: 10.1038/s41467-018-06659-3 ARTICLE
a b 100
108 OsHV-1 SF11 106 Total bacteria
Bacterial 16S relative

OsHV-1 RNA reads
107 105
OsHV-1 DNA load

SF11
quantification
106 104
105 103 10
104 102
103 101 RF21
102 RF21 100
101 10–1 1
0 20 40 60 80 0 20 40 60 80
Time (hours) Time (hours)
c 106
d 102
Total vibrios V. crassostreae SF11
relative quantification
SF11
Vibrio crassostreae
Total vibrio 16S
101
quantification
105
100
104 RF21
10–1
RF21
103 10–2
0 20 40 60 80 0 20 40 60 80
Fig. 2 OsHV-1 and bacterial colonization in the susceptible SF11 oysters during the ‘natural’ experimental infection. a The OsHV-1 load was quantified
by qPCR and expressed as viral genomic units per ng of oyster DNA (plain lines, filled symbols); viral replication was estimated by the total number
of RNA-seq reads mapped on the OsHV-1 genome (dotted lines, open symbols). Relative quantification of total bacteria (b), total Vibrio load (c) and
Vibrio crassostreae (d) abundance were measured by qPCR. Dots represent distinct pools of 10 oysters. The mean values (plain and dotted lines) are
displayed
than in RF21 oysters at 24 h after exposure initiation (Fig. 2a). The oysters (Supplementary Fig. 5). A principal coordinate analysis
time-course of OsHV-1 ORF expression is shown in Supple- (PCoA) of the Bray–Curtis dissimilarity matrix (beta diversity)
mentary Fig. 2 for both oyster families. In SF11 oysters, the consistently revealed a higher microbiota dispersion in the SF11
maximum viral load and transcriptional activity were reached at family (multivariate homogeneity of group dispersions, d.f. = 1; p
48 h and remained stable until the first deaths (at 66 h). The viral = 0.016) than in the RF21 family (Fig. 3a). This disruption of the
load remained low in RF21 throughout the experiment without bacterial community structure occurred in SF11 oysters between
any remarkable mortality until the end of the experiment (330 h). 24 h and 48 h concomitantly with the active replication of OsHV-
Alignment of the Illumina reads to the OsHV-1 genome available 1 µVar. Similarly, the Chao1 and Shannon’s H indexes of alpha
in the NCBI database30 revealed that the virus used in our diversity increased significantly in the SF11 family during the
experiments corresponded to an OsHV-1 µVar variant, as indi- infectious process (Chao1: analysis of variance (ANOVA), d.f. =
cated by (i) the deletion of 3 ORFs (ORF36, ORF37 and part of 6; p = 1.39e-07 and Shannon’s H index: ANOVA, d.f. = 6; p =
ORF38), (ii) the deletion of an adenosine upstream of ORF43 and 3.54e-05), whereas they remained stable in the RF21 family
(iii) a 12-nt deletion in the microsatellite locus H10, which are (Supplementary Fig. 6). All the bacterial genera that changed
shared characteristics of µVar variants14,20,31. These variants have significantly over the ‘natural’ experimental infection in SF11 and
been associated with mass mortalities of juvenile oysters since RF21 oysters are reported in Supplementary Data 2. Among these
20086,14,20,32. Taken together, these results show that OsHV-1 modified genera, those representing more than 4% of the bacteria
µVar infection occurs in both families, but only RF21 oysters in at least one sample of SF11 oysters are shown in Fig. 3b. The
successfully control viral replication. corresponding OTUs, which represent only 2.07% of the total
bacteria at the beginning of the experiment, represent 59.07% of
the SF11 bacterial community at 72 h, when the first mortalities
Dysbiosis and bacteraemia occur in susceptible oysters. To occurred. Some of these OTUs were assigned to the genera Vibrio
investigate the dynamics of oyster microbiota in the two families and Arcobacter, which have been previously associated with
showing contrasting resistance to the disease, we analysed oyster mortalities22,24. In contrast, in RF21 oysters, the same taxa
total bacterial communities using 16S rDNA metabarcoding over did not vary significantly over time (Supplementary Fig. 7 and
the first three days of the experiment. Overall, 6,061,881 clusters Supplementary Data 2).
were obtained from 42 libraries (2 families, 7 sampling times, in In addition, a significant increase in total bacterial abundance
triplicate). After cleaning steps and singleton filtering, was observed in SF11 oysters only, which started at 48 h
4,238,989 sequences affiliated with 10,080 OTUs were kept for and continued to rise until the end of the experiment (ANOVA,
further analyses (Supplementary Data 1). A sufficient sequencing d.f. = 6; p < 0.0001; Fig. 2b). Compared with T0, the total bacteria
depth was confirmed by species richness rarefaction curves were 13- and 17-fold higher at 60 h and 72 h, respectively,
(Supplementary Fig. 3). Changes in microbiota composition suggesting a massive bacterial proliferation in SF11 oysters.
(Supplementary Fig. 4 and Supplementary Fig. 5) were much Simultaneously, we observed a high increase in total Vibrio load
higher in SF11 oysters than in RF21 oysters throughout the and abundance of Vibrio crassostreae, a previously characterized
experiment. Indeed, 105 OTUs significantly differed in terms of bacterial pathogen of oysters associated with the disease21,22
relative proportions in SF11 oysters, as opposed to 0 in RF21 (Fig. 2c, d). Accordingly, histological analyses revealed bacterial

a b
60h RF21
Arcobacter
72h
0.2 SF11 Salinirepens
0h 6h 48h Psychrobium
PC2 (11.8%)
Marinobacterium
48h
24h Psychrilyobacter
0 Marinomonas
12h
72h Vibrio
0h 24h 60h Fusibacter
–0.2 12h Cryomorphaceae
Psychromonas
6h 0 6 12 24 48 60 72
–0.2 0 0.2 0.4 0.6 Time (hours)
0 0.05 0.15
PC1 (37.3%) Frequency
Fig. 3 Microbiota changes in the susceptible SF11 and the resistant RF21 oysters during the ‘natural’ experimental infection. a Principal coordinate analysis
(PCoA) plot of the Bray–Curtis dissimilarity matrix of the microbiota. Each point of the triangles corresponds to one of the 3 replicates at one kinetic time.
Dots represent distinct pools of 10 oysters. b Heatmap of the bacterial taxa that were significantly modified in susceptible oysters during the course
of infection. Analyses were performed at the genus level. Only genera with a relative proportion superior to 4% in at least one sample are shown. The
intensity level of the blue represents the relative abundance of bacterial taxa. At each time, the analysis was performed on 3 distinct pools of 10 oysters
a b
c d
Fig. 4 Accumulation of bacteria in tissues of the susceptible SF11 oysters. Giemsa staining was performed on paraffin wax-embedded tissue histological
sections of SF11 oysters sampled at different time points to visualize tissue colonization by bacteria. Oyster tissues and cells were coloured in shades of pink
to purple, and most bacteria appeared in deep blue (scale bars corresponds to 20 µM). a At 54 h after the beginning of the ‘natural’ experimental infection,
bacteria accumulated in the gills of the SF11 oysters (filled arrowheads). Rounded cells reminiscent of haemocytes were found in both gill tissues and
outside tissues in the vicinity of the bacteria (open arrowheads). At 78 h, gill tissues appeared massively degraded, and bacteria were found in most tissues
of the SF11 oysters, e.g., in b gills, c adductor muscle and d interstitial tissues near the digestive tract. No bacteria or tissue damage were observed on
sections of SF11 or RF21 either at the beginning of the experimental infection or at any time points for the RF21 oyster sections (Supplementary Fig. 8)
colonization and tissue damage in SF11 oyster histological sections times, in triplicate) yielded between 20.4 and 32.3 million
(Fig. 4). At 54 h, bacterial accumulation accompanied by Illumina paired reads per sample; 69.7–75.3% of them mapped
infiltrating haemocytes was visible in different areas both inside to the C. gigas V9 reference genome. RNA-seq results were
and outside the gill tissues (Fig. 4a). At 78 h, the gill tissue validated by RT-qPCR on 30 genes at all sampling times in both
structure was completely disrupted and bacteria had spread oyster families (r² = 0.936) (Supplementary Fig. 9). The tran-
throughout the body (Fig. 4b, c, d and Supplementary Fig. 8). In scriptomic responses of SF11 and RF21 oysters can be divided
contrast, in RF21 oysters, total bacteria and vibrios remained low into two phases: an early response (before 24 h) and a later
and stable (Fig. 2b–d), and no tissue damage or tissue response (24 h to 72 h) (Fig. 5a). In the early response, the total
colonization by bacteria were observed at any sampled time number of differentially expressed genes was higher in RF21 than
point (Supplementary Fig. 8). in SF11 oysters (Fig. 5a). To gain access to the functions enriched
during this early phase, we used gene ontology (GO) enrichment
Resistant oysters display an early antiviral response. To identify analyses (Fig. 5b and Supplementary Fig. 10). At these time
key host factors controlling the infection, we compared the points, transcriptomes of RF21 oysters were more enriched in
transcriptomic responses of SF11 and RF21 oysters by RNA-seq functional categories related to immunity than SF11 tran-
over the time-course of the ‘natural’ experimental infection. In scriptomes (Fig. 5b and Supplementary Fig. 10). Among those
total, the sequencing of 42 samples (2 families, 7 sampling genes, 40.9% were involved in antiviral defence (Supplementary

a c SF11 RF21
10,000 Early response Late response
9426 Time (h) 6 12 24 48 60 72 6 12 24 48 60 72
Number of differentially expressed genes
1
SF11
7925
8000 RF21 7382
2
6000 3
4
5
6
4000 7
8
9
2218
1821 1725 10
2000 1614
1267 11
866
498
173 311
0 12
6 12 24 48 60 72
Time (hours)
13
14
b 6h 12h
SF11 RF21 SF11 RF21
Innate immune response
Immune response
Organophosphate catabolic process
Purine-containing compound catabolic process
Regulation of type I interferon production
Deoxyribonucleotide metabolic process
Biological regulation
Positive regulation of biological process
Response to external stimulus
Response to virus
Regulation of multicellular organismal process
Response to stimulus 15
Response to stress
Regulation of defence response
Regulation of response to stimulus
Cellular response to stimulus
Immune system process
Defense response to other organism
Intracelullar signal transduction
Immune effector process
Protein modification by small protein conjuction removal
Deoxyribonucleotide catabolic process
Regulation of cell death
Response to wounding
Positive regulation of cell death
Cellular response to oxydative stress 16
Cellular response to oxygen containing compound
Opsonization 17
Response to inorganic substance 18
19
–0.3 0 0.3 –0.8 0.0 0.8
Fig. 5 The resistant RF21 oysters display an early antiviral response that is delayed in the SF11 susceptible oysters. a The number of genes that were
differentially expressed during the ‘natural’ experimental infection was higher in RF21 (green) than in SF11 (blue) oysters before 24 h (early response). This
trend was reversed after 24 h and until the end of the infection (late response). Upregulated genes are represented as filled coloured bars and
downregulated genes are represented as hatched bars. b Heatmap focusing on the two clusters containing 29 immune-related GO categories that were
significantly enriched at 6 and 12 h (FDR < 1%; clusters A and B, Supplementary Fig. 10). c Heatmap of the 220 significantly enriched GO categories (FDR <
1%, biological processes root) clustered according to the Pearson correlation (numbered filled bar). The enrichment intensity was expressed as the ratio
between the number of genes that were significantly up- (yellow heat) or downregulated (blue heat) in the category compared with the total number of
genes in the category. If the intensity was equal to zero (black heat), then the enrichment was not significant for the corresponding condition. Detailed
results (cluster number, GO terms and enrichment values) are presented in Supplementary Data 4
Data 3), encoding elements of the RLR/STING (e.g., RIG-1, IRF The antiviral response of susceptible oysters is inefficient.
and cGAS) and JAK/STAT (e.g., STAT and SOCS2) pathways, During the second phase of the response (after 24 h), SF11 oysters
antiviral effectors (e.g., IFI44, Viperin and SAMHD-1) and pro- displayed a major reprogramming of their transcriptome, with
teins involved in the apoptosis (e.g., TNF and caspase-3) and significant changes in the expression of 9426 genes (33.6% of the
maintenance of cellular homoeostasis under viral infection (e.g., transcriptome, Fig. 5a) and a large number of enriched functional
poly(ADP-ribose) polymerase). Taken together, these data show categories as determined by GO enrichment analyses (Fig. 5c and
that the antiviral response is more intense in RF21 oysters than in Supplementary Data 4). This major transcriptomic reprogram-
SF11 oysters during the first 12 h following exposure to the ming was not observed in RF21 oysters (Fig. 5a, c). During this
examined infectious environment. second phase, immunity and antiviral defence genes (cluster 15

SF11 RF21
Time
(hours) 6 12 24 48 60 72 6 12 24 48 60 72 0 10
ORF42
ORF87 Viral IAPs
ORF99
ORF106
CGI_10026772
CGI_10007759
CGI_10005393
CGI_10005392
CGI_10024671
CGI_10000338
CGI_10019869
CGI_10026316 Oyster IAPs
CGI_10019361
CGI_10026857
CGI_10026860
CGI_10007421
CGI_10026855
CGI_10021439
CGI_10026771
CGI_10016355
CGI_10016364
CGI_10018117
CGI_10019017
CGI_10020879
CGI_10003770
CGI_10027244
CGI_10013106
CGI_10000852 Oyster
CGI_10019549
CGI_10006488 baculoviral
CGI_10026858 IAPs
CGI_10019372
CGI_10018493
CGI_10018493
CGI_10011513
CGI_10017592
CGI_10009762
Fig. 6 Viral and oyster endogenous inhibitors of apoptosis (IAPs) are strongly induced in the susceptible SF11 oysters (but not in the resistant RF21 oysters)
during the ‘natural’ experimental infection. The fold changes in virus IAPs, oyster IAPs, and oyster and baculoviral IAPs were calculated between each time
point of the kinetics and the T0. Analyses were conducted with the RNA-seq data through mapping against the C. gigas genome for oyster IAPs and against
the OsHV-1 genome for viral IAPs30. The intensity of the colour indicates the magnitude of the differential expression (log2 fold change). The heatmap was
constructed with Multiple Array Viewer software
Fig. 5c) were strongly enriched in SF11 oysters. However, this has not been previously reported, was particularly strong after
intense antiviral response was inefficient, as the replication of the 24 h, which corresponds to the period of microbiota destabiliza-
virus in SF11 oysters was very high at these time points (Fig. 2a). tion preceding bacterial proliferation in SF11 oysters (Figs. 2, 3
Among the functional categories that were highly induced in SF11 and 4). Interestingly, both Cg-BigDefs and Cg-PRPs are produced
oysters only, we found the negative regulation of cell death only by haemocytes34. We also showed that the transcript
category to be of particular interest (Supplementary Data 4). We abundance of Cg-EcSOD, a specific marker of haemocytes35,
performed a detailed analysis of the expression of the corre- decreased over time in SF11 oysters only (Fig. 8). Taken together,
sponding genes encoding endogenous inhibitors of apoptosis our data strongly suggest that by invading haemocytes of sus-
proteins (IAPs and Baculoviral IAPs, Fig. 6). These genes were ceptible oysters, OsHV-1 infection alters haemocyte functions
found to be highly induced in SF11 oysters, but only from 24 h to and thereby disrupts an important component of the oyster
the end of the experimental infection. In addition, viral tran- antibacterial shield.
scriptome analysis revealed an over-representation of viral tran-
scripts encoding putative IAPs (ORFs 42, 87, 99 and 106) at the
same time points in SF11 oysters (Fig. 6). These data indicate that Similar events lead to similar phenotypes across families. To
SF11 oysters mount a delayed inefficient antiviral response and determine whether the main molecular events observed in SF11
have an impacted apoptosis regulation during the late phase and RF21 oysters were conserved in other oyster families sharing
response. the same phenotypes but with distinct genetic backgrounds,
we monitored viral replication, bacterial load and gene expression
in two additional susceptible families, SF14 and SF15, which died
OsHV-1 infection alters the antibacterial defence of oysters. at more than 96% (Fig. 1), and in two other resistant families
Histological sections of SF11 oysters showed accumulation of (RF23 and RF48), which died at less than 18% (Fig. 1). Consistent
OsHV-1 in the haemocytes (Fig. 7), which are oyster immune with our previous observations for RF11 and SF21, viral infection
cells that play a major role in controlling bacterial infections33. occurred early (12 h) after the beginning of the experimental
Interestingly, histological sections of RF21 oysters were devoid of infection in all oyster families, but OsHV-1 replicated intensely in
OsHV-1 positive cells (Supplementary Fig. 11). As a proxy to susceptible families (SF14 and SF15) only (Fig. 9a). Subsequently,
monitor the antibacterial defence of oysters, we analysed the these two families were heavily colonized by bacteria (pairwise
expression of genes encoding antimicrobial peptides (AMPs) in t-test at T72h; d.f. = 10; p = 0.0027), including vibrios and
both oyster families. AMPs showed highly contrasted expression V. crassostreae (Fig. 9b-d). Like SF11, the susceptible families SF14
patterns in RF21 and SF11 oysters during the experiment (Fig. 8). and SF15 exhibited a high induction of both oyster and viral
In resistant (RF21) oysters, expression of Cg-BigDef1 and 2 was IAP (pairwise t-test at T72h; d.f. = 10; p = 0.0025 and p < 0.0001,
induced. Conversely, expression of Cg-BigDef1, Cg-BigDef3 and respectively), together with lower expression or repression of
Cg-PRP transcripts decreased significantly and regularly over haemocyte genes encoding AMPs and Cg-EcSOD (Fig. 9e, f). In
time in SF11 oysters (Kruskal–Wallis, p = 0.009, 0.009, and 0.005, contrast, oysters from resistant families RF23 and RF48 displayed
respectively). This overall repression of AMP expression, which neither such a massive induction of IAP nor a repression of

OsHV-1 in situ hybridization Anti-SOD immune staining Cg-EcSOD or AMP expression, which were instead induced
(Fig. 9e, f), confirming earlier results for the RF21 family (Figs. 6
a b and 8). Finally, we showed that early induction of antiviral genes
(Viperin, cGAS, IRF, TNF and SOCS2) at 6 h (pairwise t-test; d.f.
= 10; p = 0.1841, p = 0.0063, p = 0.0317, p = 0.0024, p = 0.0409,
respectively) and 12 h (pairwise t-test; d.f. = 10; p = 0.0012,
p = 0.0011, p = 0.0143, p = 0.0009, p < 0.0001, respectively) was
the hallmark of resistant families (Fig. 9g), as also evidenced
by the RNA-seq data from RF21 oysters (Supplementary Data 3).
Viral infection and bacteraemia are needed to kill oysters. To

c d disentangle the respective roles of OsHV-1 and bacteria in
pathogenesis, we designed a series of experimental infections
using genetically diversified juveniles (Fig. 10). In a first set of
experiments, we mimicked the simultaneous transmission of
OsHV-1 and vibrios from infected to healthy oysters. Recipient
(healthy) oysters were simultaneously exposed to two types of
donors: the first ones were injected with V. crassostreae (patho-
genic strain J2-921), and the second ones were injected with
OsHV-1 (Fig. 10a–j). The mortality of the donors is presented in
Supplementary Fig. 12. Recipient oysters were either injected with
e f poly(I:C), which restricts OsHV-1 replication36, with sterile sea-
water (SW) as a control (Fig. 10a–e), or treated with chlor-
amphenicol, a broad-spectrum bacteriostatic antibiotic that is
used to limit oyster colonization by bacteria, particularly vibrios12
(Fig. 10f–j). The mortality and pathogen load of recipient oysters
were monitored throughout disease development. Injection of
poly(I:C), as opposed to SW, was sufficient to completely block
OsHV-1 replication (Fig. 10b), bacterial colonization (Fig. 10c, d
and e) and the death of recipient oysters (Fig. 10a). Moreover,
antibiotic treatment significantly reduced the load of vibrios
Fig. 7 OsHV-1 invades haemocytes of the susceptible SF11 oysters. OsHV-1 (pairwise t-test at T72h; d.f. = 2; p = 0.0009, Fig. 10i) including V.
was detected in SF11 by in situ hybridization (a, c, e). Paraffin wax- crassostreae (pairwise t-test at T72h; d.f. = 2; p = 0.028, Fig. 10j)
embedded sections of oysters that were fixed 54 h after the and oyster mortality (Mantel–Cox log-rank test, p = 0.0075,
beginning of the ‘natural’ experimental infection were hybridized with a Fig. 10f) without affecting OsHV-1 replication (Fig. 10g) and the
specific probe labelled with digoxigenin and revealed using alkaline total bacterial load associated with oysters (Fig. 10h). Importantly,
phosphatase-conjugated antibodies and NBT/BCIP (dark blue when we mimicked the transmission of only one pathogen
precipitate). Haemocytes were localized on consecutive tissue (Fig. 10k–o), the mortality of recipient oysters was observed only
sections by immunostaining with an antibody specific to the SOD when they were exposed to virus-injected donors (Fig. 10k).
haemocytic protein (b, d, f). Immunostaining was revealed using In this condition, viral replication (Fig. 10l) was accompanied
alkaline phosphatase-conjugated secondary antibodies. Labelling of virus- by increases in the total bacterial (Fig. 10m) and total vibrio
infected cells was observed in areas enriched in haemocytes in different loads (Fig. 10n). Nevertheless, V. crassostreae, which was not
tissues (indicated by arrowheads), such as the connective tissue
included in this last experimental infection set-up, was not
surrounding the digestive gland (a, b, scale bars 50 µM) or in the gills detected in oyster flesh (Fig. 10o). Two additional rationalized
(c–f, scale bars 20 µM)
experimental infections were performed using the same design
with susceptible juveniles from a biparental family (family H1210)
(Supplementary Fig. 13). Both experiments confirmed the
SF11 RF21
Time (hours) 6 12 24 48 60 72 6 12 24 48 60 72
Cg-BigDef2
Cg-BigDef1
Cg-BigDef3
Cg-DefM
Cg-Defh
Cg-PRP
Cg-EcSOD
–2 0 2
Fig. 8 Antimicrobial peptides and Cg-EcSOD expression is repressed in the susceptible SF11 oysters. Time-course of antimicrobial peptide (AMP) and Cg-
EcSOD expression in SF11 and RF21 oysters during the ‘natural’ experimental infection. The relative expression of AMPs was analysed by comparing the
number of reads (calculated by alignment using DIAMOND 0.7.9) between each time point and time zero. Analyses were performed using the RNA-seq
data by mapping against the C. gigas genome for Cg-EcSOD. The intensity of the colour from blue to yellow indicates the magnitude of the differential
expression (log2 fold change). The heatmap was constructed with Multiple Array Viewer software

a 109 b 106
SF15
Total vibrio 16S quantification

8
10
SF14
OsHV-1 DNA load

SF15
107
105
RF23 SF14
106
10 5 RF48 RF23
4 104
10 RF48
3
10
102 103
0 20 40 60 80 0 20 40 60 80
c 100 d 102
SF15 SF15
Bacterial 16S relative
relative quantification
Vibrio crassostreae
101
SF14
quantification
10 SF14
100
RF23
RF48 RF48
1 10–1
RF23
10–2
0 20 40 60 80 0 20 40 60 80
e 6h 12 h 24 h 48 h 60 h 72 h f 6h 12 h 24 h 48 h 60 h 72 h
SF14 SF14
SF15 SF15
Cg-PRP Cg-IAP
RF23 RF23
RF48 RF48
SF14 SF14
SF15 SF15
Cg-EcSOD OsHV-1 IAP
RF23 RF23
RF48 RF48
SF14
SF15 –6 0 6
Cg-BigDef2
RF23
RF48
–3 0 3
g
–3 0 3
SF14 SF15 RF23 RF48 SF14 SF15 RF23 RF48
Viperin
cGAS
IRF
TNF
SOCS2
6h 12 h
Fig. 9 Conservation of molecular events during the ‘natural’ experimental infection in the susceptible SF14 and SF15, and in the resistant RF23 and RF48
oysters. a The OsHV-1 load was quantified by qPCR. Quantification of total Vibrio (b), total bacteria (c) and Vibrio crassostreae (d) were performed by qPCR.
For a to d, dots represent distinct pools of 10 oysters. The mean values (plain lines) are displayed. Heatmap of antimicrobial peptide (Cg-Bigdef2, Cg-PRP),
Cg-EcSOD (e) and IAPs (f) expression measured by RT-qPCR in each oyster family. g Heatmap of antiviral gene (Viperin, cGAS, IRF, TNF and SOCS2)
expression measured by RT-qPCR in each oyster family at early time points (6 and 12 h) of the experimental infection. The intensity of the colour from blue
to yellow indicates the magnitude of the differential expression (log2 fold change). The heatmap was constructed with Multiple Array Viewer software
essential role of OsHV-1 replication in the bacterial colonization microbiota changes in recipient oysters using 16S rDNA
and death of recipient oysters (mortality data for donors are metabarcoding (Supplementary Data 5). All bacterial genera that
provided in Supplementary Fig. 14). changed significantly during these experimental infections are
To identify the bacterial species involved in the secondary reported in Supplementary Data 6. Among them, the most
bacterial infection in our rationalized experiments, we studied abundant in oysters (more than 4% of the total bacterial

Donor Recipient Antibiotic Donor Recipient Antibiotic Donor Recipient Antibiotic

Os+Vc SW Cm– Os+Vc Ø Cm+ Os Ø Cm–
Os+Vc PIC Cm– Os+Vc Ø Cm– Vc Ø Cm–
100 100 100

a f k
Percent survival
80 80 80
60 60 60
10 10 10
0 0 0
0 40 80 120 160 0 40 80 120 160 0 40 80 120 160
8.0 × 106
b 8.0 × 106 g 8.0 × 106
l
OsHV-1 DNA load
6.0 × 106 6.0 × 106 6.0 × 106
4.0 × 106 4.0 × 106 4.0 × 106
2.0 × 106 2.0 × 106 2.0 × 106
0 0 0
0 72 0 72 0 72
8 8 8
c h m
Total bacteria
6 6 6
4 4 4
2 2 2
0 0 0
0 72 0 72 0 72
60,000 d 60,000 i 60,000 n
Total vibrio
40,000 40,000 40,000
20,000 20,000 20,000
0 0 0
0 72 0 72 0 72
120
e 120 j 120
o
100 100 100
V. crassostreae
80 80 80
60 60 60
40 40 40
20 20 20
0 0 0
0 72 0 72 0 72
Time (hours) Time (hours) Time (hours)
Fig. 10 Rationalized experimental infections demonstrate that OsHV-1 replication is required for bacterial colonization and oyster death. Experimental
infection by OsHV-1 and/or V. crassostreae was performed as follows: oyster donors were injected with either 3.88 × 108 genomic units of OsHV-1 (Os) or
5 × 107 cfu V. crassostreae (Vc). a–e Recipient oysters were injected with poly(I:C) (PIC) or sterile seawater (SW) before exposition to both Vc and Os
donors. f–j Recipients were exposed to both Os and Vc donors in the presence (Cm + ) or absence (Cm−) of chloramphenicol in the tanks. k–o Recipients
were exposed to Os or Vc donors. The OsHV-1 DNA load (viral genomic units per ng of total oyster DNA), relative quantification of total bacteria, and total
Vibrio and Vibrio crassostreae abundance were measured by qPCR. No mortality was observed, and no OsHV-1 DNA was detected in recipient oysters when
untreated donors were used as a control. Each dot represents a measure done on a distinct pool of 10 oysters
communities in at least one sample) (Supplementary Fig. 15) these results indicate that viral infection is necessary to initiate
included OTUs from different genera, which increased signifi- the infectious process and that opportunistic bacteria are
cantly in the four treatments in which mortalities occurred responsible for a secondary infection necessary to complete
(Supplementary Fig. 15b, c, d and e). Two of them (Vibrio and disease development.
Arcobacter) were common to three out of four treatments
(Supplementary Fig. 15b, c and e). These two genera were also
found to colonize the recipient oysters in our initial ‘natural’ Discussion
experimental infection (Fig. 3b). Finally, although chloramphe- In this study, we deciphered the complex intra-host interactions
nicol treatment reduced and delayed mortality (Fig. 10b), underlying the mortality syndrome affecting juveniles of
secondary bacterial infection could still occur involving bacteria Crassostrea gigas, providing a comprehensive view of the patho-
from other genera (Supplementary Fig. 15d). Taken together, genic processes underpinning a disease that has remained

incompletely understood until now. The entire sequence of events support this theoretical prediction and identified the early
leading to oyster death was traced and identified. induction of AMP expression as a key determinant of an efficient
We showed that infection by the OsHV-1 µVar is the first response against infection43.
event that occurs during the infectious process and that the In conclusion, the present work enabled us to decipher the
intense replication of this virus is a prerequisite for development mechanisms underlying a complex pathosystem affecting juvenile
of the disease. We showed that the immune cells of oysters, the oysters. We found that pathogenesis was caused by multiple
haemocytes, are one of the cell types targeted by the virus. This infections involving a virus and opportunistic bacteria. Indirect
localization of OsHV-1 in haemocytes during the infectious intra-host interactions were shown to occur during pathogenesis,
process has also been recently reported37. Notably, infection of enabling bacterial colonization of oysters that were already
haemocytes by OsHV-1 impacted haemocyte physiology and immune-compromised by the virus. As in a suprainfection pat-
particularly impaired the expression of AMPs either through tern44, we found that the bacteria could not infect oysters in the
transcriptional regulation or indirectly through the induction of absence of the virus, and we never observed oysters infected by
cell death or lysis processes, as previously reported37. Following only one of the pathogens. Future studies are needed to validate
the repression of antibacterial defences, profound changes in this suprainfection model and explore the genetic and physiolo-
oyster-associated microbiota were observed, followed by bacter- gical attributes underlying initial and subsequent colonization
aemia and mortalities. These results clearly identify OsHV-1 as a waves in oysters. Characterization of this multifactorial disease
necessary actor triggering the disease as a whole. represents a breakthrough that was made possible by holistic
Bacterial colonization was shown to be the second event approaches developed by combining ‘natural’ experimental pro-
necessary to complete the infectious process leading to death in cedures, using oyster biparental families with contrasted sus-
oysters. This finding supports previous studies identifying bac- ceptibilities to the disease and developing thorough molecular
teria as important aetiological agents of the disease12. From our analyses of host responses, the microbiota structure and pathogen
histological analysis, gills could be the initial route of bacterial monitoring. This holistic view of diseases as a system provides
infection in OsHV-1 immune-compromised oysters before bac- an exceptionally well-adapted framework for studying the
terial dissemination to the rest of the tissues. Such a scenario is factors governing the progression of infections. We believe that
reminiscent of immune suppression by viruses such as HIV and such an integrative and holistic approach could now be applied
secondary colonization by opportunists38. Bacteria of the Vibrio to a series of multifactorial diseases that affect non-model
genus were associated with the disease, supporting previous invertebrate species worldwide.
studies that identified V. crassostreae as an important pathogenic
population in the field21,22. However, (i) V. crassostreae was Methods
not required to complete the infectious process leading to Production of biparental oyster families. In 2015, fifty different biparental
oyster death, and (ii) vibrios were not the only bacteria system- Crassostrea gigas oyster families were produced from wild seed broodstocks
sampled in farming and non-farming areas in two geographic regions (French
atically associated with dying oysters. Finally, although chlor- Mediterranean and Atlantic coasts, Supplementary Table 1). In the Atlantic area,
amphenicol treatments controlling vibrios reduced and delayed 73 oysters were collected at Logonna Daoulas (farming area) and 70 oysters at
mortality, secondary bacterial infection could still occur involving Dellec (non-farming area). In the Mediterranean area, 125 oysters were collected in
bacteria from other genera. These results clearly showed that the Thau lagoon (farming area) and 65 at the Vidourle river mouth (non-farming
the secondary bacterial infection of oysters already immune- area). In addition, 84 oysters issued from a mass selection programme to enhance
their resistance to mortality syndrome were used28. All the collected oysters were
compromised by viral infection could be engaged by a series of transferred to the Ifremer facility at Argenton (Brittany, France) between 6 and
opportunistic bacteria present in the environment. 8 January 2015 and treated for 6 days with chloramphenicol (8 mg/l).
The inability of susceptible oysters to control viral replication For gametogenesis induction, animals were held for 8 weeks in 500 l flow-
and further bacterial colonization was associated with a strong, through tanks with seawater enriched with a phytoplankton mixture at a constant
temperature of 17 °C13,22. Seawater was UV-treated and filtered through 10-μm
but late, antiviral response. Importantly, the molecular function mesh. The daily mixed diet consisted of Tisochrysis lutea (CCAP 927/14; 40 μm3,
negative regulation of cell death was highly mobilized con- 12 pg cell−1) and Chaetoceros muelleri (CCAP 1010/3; 80 μm3, 25 pg cell−1). Once
comitantly with the intense replication of the virus in susceptible the oysters were reproductively mature, gametes from 91 individuals (46 males,
oysters. The majority of the genes belonging to this function 45 females) were obtained by stripping. Gametes from one male and one female
from the same origin were mixed in a 5-l cylinder at a ratio of 50 spermatozoids per
encode endogenous IAPs that were strongly induced in suscep- oocyte (day 0). The fertilized oocytes completed their embryonic development in
tible oysters. Remarkably, intense OsHV-1 replication was also 5-l tubes filled with 1-μm-filtered, UV-treated seawater at 21 °C for 48 h. The
associated with high expression of IAPs of viral origin. Such viral D-larvae (day 2) were then collected and reared in flow-through rearing systems
proteins of the BIR family are known to have anti-apoptotic at 25 °C45. At the end of the pelagic phase (day 15), all the larvae were collected
on a 100-μm sieve and allowed to settle on cultch. Post-larvae were maintained
activities favouring viral replication39. These results indicate that in downwelling systems, where they were continuously supplied with enriched
both endogenous and exogenous anti-apoptotic processes, which seawater until the experiments began. In the larval and post-larval stages, the
are strongly activated in susceptible oysters, may play a key role oysters were fed the same diet as the broodstock at a concentration between
in the success of OsHV-1 infection26,40. 1500–2000 μm3 μl−145.
Oyster resistance was associated with an early limitation of Of the 50 families of oyster seed produced, 3 families from each location were
kept, along with 3 other families from the mass selected broodstock, for ‘natural’
viral replication during pathogenesis. Although the genomic experimental infections. These 15 oyster families were maintained under highly
determinants of the antiviral response of resistant oysters remain controlled biosecured conditions to be sure that no oyster pathogens would
to be identified and validated functionally, antiviral pathways that interfere with further experiments. The ‘pathogen-free’ status of the animals was
are known to be necessary for resistance to herpesviruses in confirmed by (i) the absence of OsHV-1 DNA detection by qPCR and (ii) a low
Vibrio presence (~10 cfu−1 tissue) determined by isolation on selective culture
vertebrates41 were shown to be highly induced at early times in medium (thiosulfate-citrate-bile salts-sucrose agar, TCBS)12. Oysters were observed
resistant oysters only. Genes involved in these pathways could be to remain free of any abnormal mortality throughout the larvae until the beginning
valuable candidates for future selective breeding. Overall, our data of the ‘natural’ experimental infections.
indicate that the time required for an oyster to establish effective
immunological control is a key indicator of disease outcome. This ‘Natural’ experimental infections. Our experimental infection protocol consists
finding is in agreement with theoretical predictions indicating of a cohabitation between C. gigas oysters (‘donors’) carrying the disease and’pa-
thogen-free’ C. gigas oysters (‘recipients’)13,22. ‘Pathogen-free’ oysters used as
that variations in parameters such as pathogen expansion or host donors (a mixture of 116-day-old oysters from 15 families, 17,700 g with a mean
response dynamics can affect the outcome of the infection42. individual weight of 1.1 g corresponding to a weight of flesh without shell of ~0.2 g)
Recent studies in an insect model of bacterial infection also were first deployed in a farming area (Logonna Daoulas, (lat 48.335263—long

−4.317922) during the infectious period until the first mortalities occurred (0.01% obtained by alignment against a protein database using DIAMOND 0.7.955. The
on 17 July 2015). Then, donor oysters were transferred back to the laboratory and AMP database was prepared by retrieving C. gigas sequences from GenBank that
placed in contact with ‘pathogen-free’ recipient oysters in a controlled environ- were manually inspected to discard irrelevant or incomplete sequences. The reads
ment11–13,22 (Supplementary Fig. 1). The experiment was conducted by placing the for each sample/replicate were compared with the database using DIAMOND
same biomass (1120 g) of donors in cohabitation in 15 independent tanks (500 l), blastx. Alignments were filtered for the best hit and e‐value < 1e‐6. Read counts for
with each containing one of the 15 families (recipient oysters with a mean indi- each AMP were normalized against both the transcript size and the total sequence
vidual weight of 1.1 g corresponding to a weight of flesh without shell of ~0.2 g) number for each sample/replicate and used for the differential gene expression
acclimatized in these structures for 2 weeks. In parallel, a control cohabitation analysis with DESeq2.
experiment was performed under identical conditions but using donors that had
not spent time in the farming areas. The ‘natural’ experimental infection began on Gene ontology annotation and enrichment analysis. To work with current
17 July 2015 and ended on 31 July 2015. Mortality was monitored in laboratory functional annotations of the C. gigas gene set, we performed a de novo functional
tanks. When recipients were exposed to the donors (17 July 2015), 2 replicates of annotation. Blastx comparison against the NR database was performed for the
100 ‘pathogen-free’ oysters of each family were placed in the farming area, and 28,027 genes annotated in the genome, with a maximum number of target hits of
mortality was monitored daily. 20 and a minimum e-value of 0.001. XML blast result files were loaded onto
During the experimental infection, 10 oysters in triplicate were randomly Blast2GO56 for GO mapping and annotation with the b2g_sep13 version of the
sampled without blinding protocols from each tank and at each time (0, 6 h, 12 h, B2G database. These results were used as inputs for GO enrichment analysis, which
24 h, 48 h, 60 h and 72 h) of the kinetics. The shell was removed, and pools of 10 was performed using adaptive clustering and a rank-based statistical test
oysters were flash frozen in liquid nitrogen. Oyster pools (10 individuals per pool) (Mann–Whitney U-test combined with adaptive clustering). The R and Perl scripts
were ground in liquid nitrogen in 50-ml stainless steel bowls with 20-mm-diameter used57 can be downloaded [https://github.com/z0on/GO_MWU]. The following
grinding balls (Retsch MM400 mill). The obtained powders (stored at −80 °C) parameters were used for adaptive clustering: largest = 0.5; smallest = 10;
were then used for extraction of RNA and DNA. In addition, 5 oysters were clusterCutHeight = 0.25. For the continuous value characterization of each gene in
sampled at 54 h and 78 h and fixed in Davidson fixative46 for histological analyses. the data set, we used a strategy aiming to take into account both the level of
expression and the significance of the differential expression. To combine these
Rationalized experimental infections. The experiments were performed using two factors, the log2 fold change was attributed to genes that were significantly
genetically diversified C. gigas oysters (4 months old with a mean individual whole differentially expressed (adjusted p < 0.05), while a zero was attributed to the
weight of 1 g corresponding to a weight of flesh without shell of ~0.2 g) or others. A category was considered enriched with a FDR < 1%. To represent the
susceptible biparental family H1210 C. gigas oysters (4 months old with a mean results synthetically, the intensity of the enrichment was calculated with the
individual whole weight 1.5 g corresponding to a weight of flesh without shell of following ratios: i) for the upregulated enriched categories, ‘number of genes
~0.3 g). Oysters from these genetic backgrounds were ‘pathogen-free’ and were significantly upregulated in the category/total number of genes in the category’;
produced as previously described13. For the experiments, the oysters were (ii) for the downregulated enriched categories, ‘−1 × (number of genes significantly
anesthetized in hexahydrate MgCl2 (ACROS, 50 g l−1, 100 oysters l−1) for 2 h47. downregulated in the category /total number of genes in the category)’. These ratios
Then, they were injected using a 26-gauge needle attached to a multi-dispensing were then displayed on a heatmap using the Multiple Experiment Viewer and
hand pipette in the adductor muscle to allow spreading into the circulatory system clustered according to the Pearson correlation58.
with either 20 µl viral or 40 µl bacterial inoculum. The OsHV-1 inoculum (3.88 ×
108 OsHV-1 genomic units μl−1) was prepared according to previously described Virus transcriptome analyses. RNA-seq reads that did not align with the C. gigas
protocols48, and injections were performed 24 h before the start of the experiment. genome were collected using SAMtools59. These reads were aligned with the
Vibrio crassostreae J2-9 was grown under agitation at 20 °C in Luria-Bertani (LB) viral genome sequence of OsHV-1 (Refseq NC_005881.230) using bowtie2 with
+ NaCl 0.5 M for 18 h21. The culture was centrifuged (1000× g, 10 min, 20 °C), single-end global alignment and default parameters60. The abundance of viral
suspended in culture medium to an optical density (OD600) of 1 and injected transcripts was calculated using HTSeq-count53 and the viral genome GFF3 file.
(5 × 107 CFU) immediately before the start of the experiment. The rationalized The counts for each ORF were normalized by dividing by the size of the ORF and
experimental infections were performed by exposing the recipients (n = 100) to multiplying by a library normalization factor (calculated as the average library size
injected donors (n = 100) in 50-l tanks at 21 °C. A control experiment was carried for all times
and controls divided by the library size for the specific time point).
out by placing the non-injected donors in contact with recipient oysters. During
C ¼ csorf ´ LL , where corf represents the raw counts for the ORF, sorf represents
each experimental treatment, oysters were sampled at 0 and 72 h, and cumulative orf i
the size of the ORF, L represents the average reads for all the time points and Li
mortality was monitored up to 72 h and 157 h for both donors and recipients, corresponds to the number of total reads for the specific time point.
respectively. Recipient oyster pools were ground in liquid nitrogen in 50-ml
stainless steel bowls with 20-mm-diameter grinding balls (Retsch MM400 mill).
The powders obtained (stored at −80 °C) were then used for DNA extraction. In Quantification of bacteria and viruses. Quantification of OsHV-1, total 16S
one of these rationalized experimental infections, chloramphenicol (8 mg l−1) was rDNA Vibrio and Vibrio crassostreae was performed using quantitative PCR
added to the tank every 2 days to restrict bacterial proliferation. In another (qPCR). All amplification reactions were analysed using a Roche LightCycler 480
experiment, the recipients were treated with poly(I:C), which has been shown to Real-Time thermocycler (qPHD-Montpellier GenomiX platform, Montpellier
block OsHV-1 replication36. In this case, 20 µl poly(I:C) (HMW, InvivoGen, cat. University, France). The total qPCR reaction volume was 1.5 μl and consisted of
code: tlrl-pic - 1 mg ml−1) was injected into the adductor muscle of recipient 0.5 μl DNA (40 ng µl−1) and 1 μl LightCycler 480 SYBR Green I Master mix
oysters 24 h before the start of the cohabitation experiment. The experiment with (Roche) containing 0.5 μM PCR primer (Eurogenetec SA). Virus-specific primer
genetically diversified oysters was performed one time. The experiments using pairs targeted a region of the OsHV-1 genome predicted to encode a DNA poly-
oysters from the H12 family were replicated two times. Three pools of 10 indivi- merase catalytic subunit (ORF100, AY509253): Fw-ATTGATGATGTGGATAAT
duals for each experiment were randomly sampled without blinding protocols, and CTGTG and Rev-GGTAAATACCATTGGTCTTGTTCC30. Total bacteria specific
submitted for further molecular analysis. primer pairs were the 341F-CCTACGGGNGGCWGCAG and 805R-GACTACHV
GGGTATCTAATCC primers targeting the variable V3V4 loops for bacterial
communities61. Total Vibrio specific primer pairs were Fw-GGCGTAAAGCGCA
Oyster transcriptome analyses. RNA was extracted from the powdered oysters TGCAGGT and Rev-GAAATTCTACCCCCCTCTACAG62, and Vibrio crassos-
using the Direct-Zol RNA Miniprep kit (Proteigene) according to the manu- treae-specific primer pairs were Fw-ATGACCATCCAACAACCCG and Rev-AGC
facturer’s protocol. RNA concentration and purity were checked using a Nanodrop CGTAATTGATACGCACG. A Labcyte Acoustic Automated Liquid Handling
ND-1000 spectrometer (Thermo Scientific), and their integrity was analysed by Platform (ECHO) was used for pipetting into the 384-well plate (Roche). A
capillary electrophoresis on a BioAnalyzer 2100 (Agilent). RNA-seq library con- LightCycler® 480 Instrument (Roche) was used for qPCR with the following
struction and sequencing were performed by the Fasteris Company (Switzerland). program: enzyme activation at 95 °C for 10 min, followed by 40 cycles of dena-
Directional cDNA libraries were constructed using a TruSeq mRNA Stranded kit turation (95 °C, 10 s), hybridization (60 °C, 20 s) and elongation (72 °C, 25 s). A
(Illumina) and sequenced on a Hiseq in paired-end reads of 2 × 75 bp. All data subsequent melting temperature curve of the amplicon was performed to verify the
treatments were carried out under a local galaxy instance49. Phred scores were specificity of the amplification. Absolute quantification of viral and bacterial DNA
checked using the Fastq-X toolkit50 and were higher than 26 over 90% of the read copies were estimated by comparing the observed Cq values to a standard curve of
length for all the sequences. All the reads were thus kept for subsequent analyses. the DP amplification product cloned into the pCR4-TOPO vector for OsHV-1 and
Mapping to the C. gigas reference genome (assembly version V951) was performed from total DNA extraction of V. crassostreae J2-9 for total vibrio 16S rDNA. For
using RNAstar (Galaxy Version 2.4.0d-252). The HTSeq-count was used to count total bacteria and V. crassostreae 16S rDNA, we used the relative quantification
the number of reads overlapping annotated genes (mode Union) (Galaxy Version calculated by the 2−ΔΔCq method63 with the mean of the measured threshold
v0.6.1)53. Finally, the differential gene expression levels were analysed with the cycle values of two reference genes (Cg-BPI, GenBank: AY165040 and Cg-actin,
DESeq2 R package54. Fold changes between each time point of the kinetics and the GenBank: AF026063).
T0 control condition were considered significant when the adjusted p-value (Padj)
for multiple testing with the Benjamini–Hochberg procedure, which controls the
false discovery rate (FDR), was < 0.05. Giemsa staining and in situ hybridization. Crassostrea gigas tissues were fixed for
Because not all known C. gigas AMPs were present in the C. gigas reference 24 h in Davidson fixative46. Tissues were embedded in paraffin wax, serially sec-
genome (assembly version V951), read counts for all of them were specifically tioned to a thickness of 5 μm and collected on polylysine-coated slides (performed

by Histalim Company, France). To visualize bacteria that had infiltrated the the maximum number of differences between two OTUs68. Chimeras were
tissues, tissue sections were stained using Giemsa, which coloured the different removed using VSEARCH69. We filtered out the data set for singletons and
tissues and cells in shades of pink to purple and most bacteria in deep blue performed an affiliation using Blast + against the Silva 16S rDNA database
(performed by Histalim Company, France). The presence of OsHV-1 in tissue (release 128, Sept 2016) to produce an OTU and affiliation table in the standard
sections was detected by in situ hybridization following a previously published BIOM format. Rarefaction curves of species richness were produced using the
protocol64. The slides were hybridized with 5 ng µl−1 of OsHV-1-specific R package and the rarefy_even_depth and ggrare functions70. We used phyloseq to
digoxigenin-labelled (DIG) antisense probes designed based on the C2-C6 frag- obtain relative abundances at different taxonomic ranks (from genus to phylum)
ment of the ORF4 of the OsHV-1 reference genome (GenBank: NC_005881.2), (tax_glom function). In our analyses, we only kept taxa that had a true annotation
which is also present in the OsHV-1 µVar sequence. As a negative control, for the for each corresponding taxonomic rank (from genus to phylum).
specificity assessment of the OsHV-1 probe, the samples were also hybridized with
a GFP probe with no homology to the OsHV-1 or C. gigas genomes. The primers
Statistical analyses. Statistical analyses were performed using R v3.3.1 (R: a
used for the probe synthesis were as follows: OsHV-1 probe, C2- language and environment for statistical computing, 2008; R Development Core
CTCTTTACCATGAAGATACCCACC and C6-GTGCACGGCTTACCATTTTT;
Team, R Foundation for Statistical Computing, Vienna, Austria [http://www.R-
GFP probe, GFP-Fw-ACGTAAACGGCCACAAGTTC and GFP-Rev-AAGTCGT
project.org]). Survival curves were used to analyse mortality kinetics, and the non-
GCTGCTTCATGTG. After hybridization, the tissues were counter-stained with
parametric Kaplan–Meier test was used to estimate log-rank values for comparing
a solution of Bismark Brown yellow. Haemocyte localization in tissue sections
conditions. Principal coordinate analyses (PCoA, {phyloseq}) were computed to
was performed by immunohistology. After dewaxing the tissue sections in xylene
represent dissimilarities between samples using the Bray–Curtis distance matrix
followed by rehydration in an ethanol series and distilled water, a heat-induced
(ordinate, {phyloseq}). Multivariate homogeneity of group dispersions was tested
antigen retrieval procedure was conducted by incubating the sections for 12 min in
between bacterial assemblages of SF11 and RF21 using 999 permutations (permutest,
sodium citrate solution (10 mM, pH 6) in a microwave (800 W). The sections were
betadisper, {vegan}). To identify candidate taxa with changes in abundances
then incubated at room temperature for 1 h in a 5% non-fat dry milk (NFDM) and
between the initial and the final time points of the experiment, we used DESeq2
0.1% Triton X100 solution as a blocking agent and permeabilization solution.
(DESeq54) from the OTUs to the higher taxonomic ranks. Heatmaps of significant
Immunodetection of the haemocyte-specific protein Cg-EcSOD was performed genera were computed using relative abundances and the heatmap.2 function
using a primary antibody produced in-house from mouse ascites35. Sections were
({gplots}). We performed one-way ANOVA or non-parametric Kruskal–Wallis
incubated overnight in a humidified chamber at 4 °C in a 1/500 dilution (in PBS-
tests when the normality of residuals was rejected (Shapiro test) to compare alpha
5% NFDM) of the primary antibody and rinsed 3 × 5 min in TBS (pH 7.4). A 1/
diversity metrics for bacterial microbiota, along with OsHV-1 (RNA and DNA
5000 dilution of the secondary antibody (goat anti-mouse polyvalent immu-
level) and bacterial absolute abundances over the experiment after logarithmic
noglobulin alkaline phosphatase-conjugated, SIGMA A0162) was then applied for
transformations. When the ANOVA or Kruskal–Wallis tests were significant, we
2 h at room temperature in a humidified chamber, and the sections were rinsed
then computed pairwise comparisons between group levels (post-hoc analyses)
2×5 min in TBS. Alkaline phosphatase enzymatic activity was revealed by incu-
with Bonferroni corrections for multiple testing using the pairwise t-test and the
bating slides in NBT/BCIP solution for 20 min in the dark, and reactions were Dunn test, respectively. For all analyses, the threshold significance level was set
stopped by thoroughly rinsing the slides in distilled water. As a negative control for
at 0.05.
specificity assessment of the anti-Cg-EcSOD signal, the sections were incubated
with the secondary antibody only. The slides were finally mounted in Dako
mounting medium (DAKO S3023). Images were acquired using a Zeiss microscope Data availability
equipped with a Zeiss colour camera and managed with ZEN software (Montpellier RNA-seq data and amplicon sequences for microbiota analysis have been made available
RIO imaging platform). For in situ hybridization, labelling was observed only through the SRA database (BioProject accession number PRJNA423079 with SRA
with the virus-specific probe, and no labelling was detected in the sections treated accession SRP130264). RNA-seq data are available under SRA accessions SRR6679052-
with the GFP probe (Supplementary Fig. 16). For Cg-EcSOD immunostaining, SRR6679093. Amplicon sequences for microbiota analysis during ‘natural’ experimental
labelling was observed in sections treated with anti-Cg-EcSOD antibody, and infections are available under SRA accessions SRR7786101–SRR7786142. Amplicon
no labelling was observed for slides treated with the secondary antibody only sequences for microbiota analysis during rationalized experimental infections are
(Supplementary Fig. 17). available under SRA accessions SRR6675783–SRR6675803. Other data analyzed during
this study are included in this published article and its supplementary information files.
RNA extraction and RT-qPCR analysis. RNA extraction was performed Complementary information is available from the corresponding authors on reasonable
using Direct-zoltm RNA MiniPrep according to manufacturer’s instructions request.
(Zymo Research). Frozen oyster powder (20 mg) was homogenized in 1 ml
TRIzol by vortexing 1 h at 4 °C. Prior to extraction, insoluble materials were
removed by centrifugation at 12,000× g for 10 min at 4 °C. The quantification Received: 12 February 2018 Accepted: 18 September 2018
and integrity of the total RNA were checked using a NanoDrop spectrophotometer
(Thermo Scientific) and 1.5% agarose gel electrophoresis, respectively. Total
RNA (300 ng) was reverse-transcribed in 20 μl using Moloney Murine Leukaemia
Virus Reverse Transcriptase (MMLV-RT) according to the manufacturer’s
instructions (Invitrogen). Amplification and pipetting were performed with a
Roche LightCycler 480 and a Labcyte Acoustic Automated Liquid Handling References
Platform (ECHO), as previously described (see above, Quantification of 1. Goulson, D., Nicholls, E., Botias, C. & Rotheray, E. L. Bee declines driven by
bacteria and viruses). The total RT-qPCR reaction volume was 1.5 μl and consisted combined stress from parasites, pesticides, and lack of flowers. Science 347,
of 0.5 μl of cDNA (dilution 1/6) and 1 μl of LightCycler 480 SYBR Green I Master
1255957 (2015).
mix (Roche) containing 0.5 μM of PCR primer (Eurogenetec SA). The amplifica-
2. Bourne, D. G. et al. Microbial disease and the coral holobiont. Trends
tion efficiency of each primer pair (Supplementary Tables 2 and 3) was validated by
Microbiol. 17, 554–562 (2009).
serial dilution of a pool of all cDNAs. Related expression was calculated as the
3. Guo, X. & Ford, S. E. Infectious diseases of marine molluscs and host
threshold cycle (Cq) values of selected genes minus the mean of the measured
threshold cycle (Cq) values of three constitutively expressed genes (Cg-EF1, responses as revealed by genomic tools. Philos Trans R Soc Lond B Biol Sci
Cg-RPL40 and Cg-RPS6). 371, 20150206 (2016).
4. FAO. Crassostrea gigas (Thunberg, 1793). Species Fact Sheets-Food and
Agriculture Organization of the United Nations, (Food and Agriculture
Bacterial microbiota analysis. DNA from powdered oyster tissues was extracted
Organization of the United Nations, Italy, 2014).
with DNA from the tissue Macherey-Nagel kit (reference 740952.250) according to
the manufacturer’s protocol. Prior to 90 min of enzymatic lysis in the presence of 5. Samain, J.-F. & Mc Combie, H. Summer mortality of Pacific oyster Crassostrea
proteinase K, an additional 12-min mechanical lysis (Retsch MM400 mill) was gigas. The MOREST project. 400 p. (QUAE, France, 2008).
performed with zirconia/silica beads (BioSpec). DNA concentration and purity 6. EFSA, Po. A. H. W. Oyster mortality. FASEB J. 13, 4u22–4n/au22 (2015).
were checked with a Nanodrop ND-1000 spectrometer (Thermo Scientific). For 7. Paul-Pont, I., Dhand, N. K. & Whittington, R. J. Influence of husbandry
each sample, 16S rDNA amplicon libraries were generated using the 341F- practices on OsHV-1 associated mortality of Pacific oysters Crassostrea gigas.
CCTACGGGNGGCWGCAG and 805R-GACTACHVGGGTATCTAATCC pri- Aquaculture 412, 202–214 (2013).
mers targeting the variable V3V4 loops for bacterial communities61. Paired-end 8. Barbosa Solomieu, V., Renault, T. & Travers, M. A. Mass mortality in bivalves
sequencing with a 250-bp read length was performed at McGill University (Gén- and the intricate case of the Pacific oyster, Crassostrea gigas. J. Invertebr.
ome Québec Innovation Centre, Montréal, Canada) on a MiSeq system (Illumina) Pathol. 131, 2–10 (2015).
using v2 chemistry according to the manufacturer’s protocol. The FROGS pipeline 9. Pernet, F. et al. Mass mortalities of Pacific oysters Crassostrea gigas reflect
(Find Rapidly OTU with Galaxy Solution) implemented on a galaxy instance infectious diseases and vary with farming practices in the Mediterranean
[http://sigenae-workbench.toulouse.inra.fr/galaxy/] was used for data processing65. Thau lagoon, France. Aquacult Env Inter. 2, 215–237 (2012).
In brief, paired reads were merged using FLASH66. After denoising and primer/ 10. Azema, P. et al. Genetic parameters of resistance to Vibrio aestuarianus,
adapter removal with cutadapt67, clustering was performed using SWARM, which and OsHV-1 infections in the Pacific oyster, Crassostrea gigas, at three
uses a novel clustering algorithm with a threshold (distance = 3) corresponding to different life stages. Genet. Sel. Evol. 49, 23 (2017).

11. Le Roux, F., Wegner, K. M. & Polz, M. F. Oysters and vibrios as a model for 39. Miller, L. K. An exegesis of IAPs: salvation and surprises from BIR motifs.
disease dynamics in wild animals. Trends Microbiol. 24, 568–580 (2016). Trends Cell Biol. 9, 323–328 (1999).
12. Petton, B. et al. Crassostrea gigas mortality in France: the usual suspect, a 40. Green, T. J., Rolland, J. L., Vergnes, A., Raftos, D. & Montagnani, C. OsHV-1
herpes virus, may not be the killer in this polymicrobial opportunistic disease. countermeasures to the Pacific oyster’s anti-viral response. Fish. Shellfish.
Front. Microbiol. 6, 686 (2015). Immunol. 47, 435–443 (2015).
13. Petton, B., Pernet, F., Robert, R. & Boudry, P. Temperature influence on 41. Gao, F. X. et al. Distinct herpesvirus resistances and immune responses of
pathogen transmission and subsequent mortalities in juvenile Pacific oysters three gynogenetic clones of gibel carp revealed by comprehensive
Crassostrea gigas. Aquacult Env Inter. 3, 257–273 (2013). transcriptomes. BMC Genom. 18, 561 (2017).
14. Segarra, A. et al. Detection and description of a particular Ostreid herpesvirus 42. Alizon, S., Luciani, F. & Regoes, R. R. Epidemiological and clinical
1 genotype associated with massive mortality outbreaks of Pacific oysters, consequences of within-host evolution. Trends Microbiol. 19, 24–32 (2011).
Crassostrea gigas, in France in 2008. Virus Res. 153, 92–99 (2010). 43. Duneau, D. et al. Stochastic variation in the initial phase of bacterial infection
15. Martenot, C., Oden, E., Travaille, E., Malas, J. P. & Houssin, M. Detection of predicts the probability of survival in D. melanogaster. Elife 6, e28298 (2017).
different variants of Ostreid Herpesvirus 1 in the Pacific oyster, Crassostrea 44. Sofonea, M. T., Alizon, S. & Michalakis, Y. Exposing the diversity of multiple
gigas between 2008 and 2010. Virus Res. 160, 25–31 (2011). infection patterns. J. Theor. Biol. 419, 278–289 (2017).
16. Renault, T. et al. Analysis of clinical ostreid herpesvirus 1 45. Rico-Villa, B., Pouvreau, S. & Robert, R. Influence of food density and
(Malacoherpesviridae) specimens by sequencing amplified fragments from temperature on ingestion, growth and settlement of Pacific oyster larvae,
three virus genome areas. J. Virol. 86, 5942–5947 (2012). Crassostrea gigas. Aquaculture 287, 395–401 (2009).
17. Peeler, E. J. et al. Investigation of mortality in Pacific oysters associated with 46. Gonzalez, M. et al. Evidence of a bactericidal permeability increasing protein
Ostreid herpesvirus-1 muVar in the Republic of Ireland in 2009. Prev. Vet. in an invertebrate, the Crassostrea gigas Cg-BPI. Proc. Natl Acad. Sci. USA
Med. 105, 136–143 (2012). 104, 17759–17764 (2007).
18. Lynch, S. A., Carlsson, J., Reilly, A. O., Cotter, E. & Culloty, S. C. A previously 47. Suquet, M. et al. Anesthesia in Pacific oyster, Crassostrea gigas. Aquat. Living
undescribed ostreid herpes virus 1 (OsHV-1) genotype detected in the pacific Resour. 22, 29–34 (2009).
oyster, Crassostrea gigas, in Ireland. Parasitology 139, 1526–1532 (2012). 48. Schikorski, D. et al. Experimental infection of Pacific oyster Crassostrea gigas
19. Abbadi, M. et al. Identification of a newly described OsHV-1 microvar from spat by ostreid herpesvirus 1: demonstration of oyster spat susceptibility.
the North Adriatic Sea (Italy). J. Gen. Virol. 99, 693–703 (2018). Vet. Res. 42, 27 (2011).
20. Burioli, E. A. V., Prearo, M. & Houssin, M. Complete genome sequence of 49. Goecks, J., Nekrutenko, A., Taylor, J. & Galaxy, T. Galaxy: a comprehensive
Ostreid herpesvirus type 1 microVar isolated during mortality events in the approach for supporting accessible, reproducible, and transparent
Pacific oyster Crassostrea gigas in France and Ireland. Virology 509, 239–251 computational research in the life sciences. Genome Biol. 11, R86 (2010).
(2017). 50. Blankenberg, D. et al. Manipulation of FASTQ data with Galaxy.
21. Lemire, A. et al. Populations, not clones, are the unit of vibrio pathogenesis in Bioinformatics 26, 1783–1785 (2010).
naturally infected oysters. ISME J. 9, 1523–1531 (2015). 51. Zhang, G. et al. The oyster genome reveals stress adaptation and complexity of
22. Bruto, M. et al. Vibrio crassostreae, a benign oyster colonizer turned into a shell formation. Nature 490, 49–54 (2012).
pathogen after plasmid acquisition. ISME J. 11, 1043–1052 (2017). 52. Dobin, A. et al. STAR: ultrafast universal RNA-seq aligner. Bioinformatics 29,
23. Lokmer, A., Kuenzel, S., Baines, J. & Wegner, K. M. The role of tissue-specific 15–21 (2013).
microbiota in initial establishment success of Pacific oysters. Environ 53. Anders, S., Pyl, P. T. & Huber, W. HTSeq—a Python framework to work
Microbiol 18, 970–987 (2015). with high-throughput sequencing data. Bioinformatics 31, 166–169 (2015).
24. Lokmer, A. & Mathias Wegner, K. Hemolymph microbiome of Pacific oysters 54. Love, M. I., Huber, W. & Anders, S. Moderated estimation of fold change and
in response to temperature, temperature stress and infection. ISME J. 9, dispersion for RNA-seq data with DESeq2. Genome Biol. 15, 550 (2014).
670–682 (2015). 55. Buchfink, B., Xie, C. & Huson, D. H. Fast and sensitive protein alignment
25. Wendling, C. C. & Wegner, K. M. Adaptation to enemy shifts: rapid resistance using DIAMOND. Nat. Methods 12, 59–60 (2015).
evolution to local Vibrio spp. in invasive Pacific oysters. Proc. Biol. Sci. 282, 56. Conesa, A. et al. Blast2GO: a universal tool for annotation, visualization and
20142244 (2015). analysis in functional genomics research. Bioinformatics 21, 3674–3676
26. Segarra, A. et al. Ostreid herpesvirus type 1 replication and host response in (2005).
adult Pacific oysters, Crassostrea gigas. Vet. Res. 45, 103 (2014). 57. Wright, R. M., Aglyamova, G. V., Meyer, E. & Matz, M. V. Gene expression
27. He, Y. et al. Transcriptome analysis reveals strong and complex antiviral associated with white syndromes in a reef building coral, Acropora
response in a mollusc. Fish. Shellfish. Immunol. 46, 131–144 (2015). hyacinthus. BMC Genom. 16, 371 (2015).
28. Degremont, L., Nourry, M. & Maurouard, E. Mass selection for survival and 58. Saeed, A. I. et al. TM4: a free, open-source system for microarray data
resistance to OsHV-1 infection in Crassostrea gigas spat in field conditions: management and analysis. Biotechniques 34, 374–378 (2003).
response to selection after four generations. Aquaculture 446, 111–121 (2015). 59. Li, H. et al. The Sequence alignment/map format and SAMtools.
29. Le Roux, F. Diagnosis of vibriosis in the era of genomics: lessons from Bioinformatics 25, 2078–2079 (2009).
invertebrates. Rev. Sci. Tech. 35, 259–269 (2016). 60. Langmead, B. & Salzberg, S. L. Fast gapped-read alignment with Bowtie 2.
30. Davison, A. J. et al. A novel class of herpesvirus with bivalve hosts. J. Gen. Nat. Methods 9, 357–359 (2012).
Virol. 86, 41–53 (2005). 61. Klindworth, A. et al. Evaluation of general 16S ribosomal RNA gene PCR
31. Renault, T. et al. Genotyping of a microsatellite locus to differentiate clinical primers for classical and next-generation sequencing-based diversity studies.
Ostreid herpesvirus 1 specimens. Vet. Res. 45, 3 (2014). Nucleic Acids Res. 41, e1 (2013).
32. Renault, T., Bouquet, A. L., Maurice, J. T., Lupo, C. & Blachier, P. Ostreid 62. Mansergh, S. & Zehr, J. P. Vibrio diversity and dynamics in the Monterey Bay
herpesvirus 1 infection among Pacific oysters, Crassostrea giga, spat: virus upwelling region. Front. Microbiol. 5, 48 (2014).
replication and circulation related to water temperature prior the onset of 63. Pfaffl, M. W. A new mathematical model for relative quantification in real-
mortality. Appl Environ Microbiol 80, 5419–5426 (2014). time RT-PCR. Nucleic Acids Res. 29, e45 (2001).
33. Bachere, E. et al. The new insights into the oyster antimicrobial defense: 64. Bueno, R., Perrott, M., Dunowska, M., Brosnahan, C. & Johnston, C. In situ
cellular, molecular and genetic view. Fish. Shellfish. Immunol. 46, 50–64 hybridization and histopathological observations during ostreid herpesvirus-
(2015). 1-associated mortalities in Pacific oysters Crassostrea gigas. Dis. Aquat. Organ.
34. Schmitt, P. et al. The antimicrobial defense of the Pacific Oyster, Crassostrea 122, 43–55 (2016).
gigas. How diversity may compensate for scarcity in the regulation of resident/ 65. Escudie, F. et al. FROGS: find, rapidly, OTUs with galaxy solution.
pathogenic microflora. Front. Microbiol. 3, 160 (2012). Bioinformatics 34, 1287–1294 (2018).
35. Gonzalez, M. et al. Evidence in oyster of a plasma extracellular superoxide 66. Magoc, T. & Salzberg, S. L. FLASH: fast length adjustment of short reads to
dismutase which binds LPS. Biochem. Biophys. Res. Commun. 338, 1089–1097 improve genome assemblies. Bioinformatics 27, 2957–2963 (2011).
(2005). 67. Martin, M. Cutadapt removes adapter sequences from high-throughput
36. Lafont, M. et al. Long-lasting antiviral innate immune priming in the sequencing reads. EMBnet j. 17, 10–12 (2011).
Lophotrochozoan Pacific oyster, Crassostrea gigas. Sci. Rep. 7, 13143 (2017). 68. Mahe, F., Rognes, T., Quince, C., de Vargas, C. & Dunthorn, M. Swarm:
37. Martenot, C., Gervais, O., Chollet, B., Houssin, M. & Renault, T. Haemocytes robust and fast clustering method for amplicon-based studies. PeerJ 2, e593
collected from experimentally infected Pacific oysters, Crassostrea gigas: (2014).
Detection of ostreid herpesvirus 1 DNA, RNA, and proteins in relation with 69. Rognes, T., Flouri, T., Nichols, B., Quince, C. & Mahe, F. VSEARCH: a
inhibition of apoptosis. PLoS ONE 12, e0177448 (2017). versatile open source tool for metagenomics. PeerJ 4, e2584 (2016).
38. Yang, L. et al. HIV-induced immunosuppression is associated with 70. McMurdie, P. J. & Holmes, S. phyloseq: an R package for reproducible
colonization of the proximal gut by environmental bacteria. AIDS 30, 19–29 interactive analysis and graphics of microbiome census data. PLoS ONE 8,
(2016). e61217 (2013).

Acknowledgements Additional information

We warmly thank the staff of the Ifremer stations of Argenton (LPI, PFOM) and Supplementary Information accompanies this paper at https://doi.org/10.1038/s41467-
Sète (LER), and the Comité Régional de Conchyliculture de Méditerranée (CRCM) 018-06659-3.
for technical support in the collection of the oyster genitors and reproduction.
We also thank Fabrice Pernet, Marie-Agnès Travers, David Mouillot, Marion Richard Competing interests: The authors declare no competing interests.
and Franck Lagarde for fruitful discussions. The authors are grateful to Philippe
Clair from the qPHD platform/Montpellier genomix for useful advice and Yannick Reprints and permission information is available online at http://npg.nature.com/
Labreuche for the Vibrio crassostreae-specific primers. This work, through the use reprintsandpermissions/
of the GENSEQ platform [http://www.labex-cemeb.org/fr/genomique-envir-
onnementale-2] from the labEx CeMEB, benefited from the support of the National Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in
Research Agency under the ‘Investissements d’avenir’ program (reference ANR-10- published maps and institutional affiliations.
LABX-04-01).The authors also thank the Montpellier RIO imaging platform (https://
www.mri.cnrs.fr). The present study was supported by the ANR project DECIPHER
(ANR-14-CE19-0023), by the EU funded project VIVALDI (H2020 program, n°678589)
and by Ifremer, CNRS, Université de Montpellier and Université de Perpignan via Open Access This article is licensed under a Creative Commons
Domitia. Attribution 4.0 International License, which permits use, sharing,
adaptation, distribution and reproduction in any medium or format, as long as you give
appropriate credit to the original author(s) and the source, provide a link to the Creative
Author contributions Commons license, and indicate if changes were made. The images or other third party
J.D.L, A.L., B.P., C.M., J-.M.E, P.H., L.D., M.L., A.V., N.F., T.R., M.A.L., A.P., D.R., material in this article are included in the article’s Creative Commons license, unless
B.M., M.A.B., Y.G. and G.M. performed oyster experiments. J.D.L., A.L., E.T., C.C.L., indicated otherwise in a credit line to the material. If material is not included in the
M.C., Y.G. and G.M. performed microbiota analyses. J.D.L., A.L., E.T., C.M., J.V.D., C.C. article’s Creative Commons license and your intended use is not permitted by statutory
H., R.G., J-.M.E., Y.G. and G.M. performed RNA-seq analyses. J.D.L., A.L., A.D., A.V. regulation or exceeds the permitted use, you will need to obtain permission directly from
and C.M. performed qPCR analyses. C.M., G.M.C. and A.V. performed histology the copyright holder. To view a copy of this license, visit http://creativecommons.org/
analyses. J.D.L., B.P., P.B., D.D-.G. and G.M. designed experiments. J.D.L., A.L., B.P., licenses/by/4.0/.
F.L.R., D.D-.G., Y.G. and G.M. interpreted results. J.D.L., A.L., F.L.R., D.D-.G., Y.G.
and G.M. wrote the paper. All the authors have revised and approved the manuscript
submission. © The Author(s) 2018

Immune-Suppression by Oshv-1 Viral Infection Causes Fatal Bacteraemia in Paci Fic Oysters

Uploaded by

Copyright:

Available Formats

Immune-Suppression by Oshv-1 Viral Infection Causes Fatal Bacteraemia in Paci Fic Oysters

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Immune-Suppression by Oshv-1 Viral Infection Causes Fatal Bacteraemia in Paci Fic Oysters

Uploaded by

Copyright:

Available Formats

ARTICLE

DOI: 10.1038/s41467-018-06659-3 OPEN

Immune-suppression by OsHV-1 viral infection

Delphine Destoumieux-Garzόn1, Yannick Gueguen 1 & Guillaume Mitta1

NATURE COMMUNICATIONS | (2018)9:4215 | DOI: 10.1038/s41467-018-06659-3 | www.nature.com/naturecommunications 1

resistant). This holistic approach allows us to establish that

2 NATURE COMMUNICATIONS | (2018)9:4215 | DOI: 10.1038/s41467-018-06659-3 | www.nature.com/naturecommunications

Bacterial 16S relative

OsHV-1 DNA load

NATURE COMMUNICATIONS | (2018)9:4215 | DOI: 10.1038/s41467-018-06659-3 | www.nature.com/naturecommunications 3

4 NATURE COMMUNICATIONS | (2018)9:4215 | DOI: 10.1038/s41467-018-06659-3 | www.nature.com/naturecommunications

NATURE COMMUNICATIONS | (2018)9:4215 | DOI: 10.1038/s41467-018-06659-3 | www.nature.com/naturecommunications 5

6 NATURE COMMUNICATIONS | (2018)9:4215 | DOI: 10.1038/s41467-018-06659-3 | www.nature.com/naturecommunications

Viral infection and bacteraemia are needed to kill oysters. To

NATURE COMMUNICATIONS | (2018)9:4215 | DOI: 10.1038/s41467-018-06659-3 | www.nature.com/naturecommunications 7

Total vibrio 16S quantification

OsHV-1 DNA load

8 NATURE COMMUNICATIONS | (2018)9:4215 | DOI: 10.1038/s41467-018-06659-3 | www.nature.com/naturecommunications

Donor Recipient Antibiotic Donor Recipient Antibiotic Donor Recipient Antibiotic

100 100 100

6.0 × 106 6.0 × 106 6.0 × 106

4.0 × 106 4.0 × 106 4.0 × 106

2.0 × 106 2.0 × 106 2.0 × 106

40,000 40,000 40,000

20,000 20,000 20,000

NATURE COMMUNICATIONS | (2018)9:4215 | DOI: 10.1038/s41467-018-06659-3 | www.nature.com/naturecommunications 9

10 NATURE COMMUNICATIONS | (2018)9:4215 | DOI: 10.1038/s41467-018-06659-3 | www.nature.com/naturecommunications

NATURE COMMUNICATIONS | (2018)9:4215 | DOI: 10.1038/s41467-018-06659-3 | www.nature.com/naturecommunications 11

12 NATURE COMMUNICATIONS | (2018)9:4215 | DOI: 10.1038/s41467-018-06659-3 | www.nature.com/naturecommunications

NATURE COMMUNICATIONS | (2018)9:4215 | DOI: 10.1038/s41467-018-06659-3 | www.nature.com/naturecommunications 13

Acknowledgements Additional information

14 NATURE COMMUNICATIONS | (2018)9:4215 | DOI: 10.1038/s41467-018-06659-3 | www.nature.com/naturecommunications

You might also like

Pfad - The Proxy pFad of © 2024 Garber Painting. All rights reserved.