Crassostrea gigas oysters were screened from three different shellfish farms located along the Irish coast (Figure 1). Oysters were locally produced and grown at Site A or imported from an external hatchery and settled at Site B and Site C. Three consecutive bags per site, containing juveniles (12 months old, 6.9 ± 1.5 cm in length) were sampled every fortnight. Percentage of mortality was estimated in 100 oysters per bag by counting coupled empty shells. After removal of all dead animals, 20 oysters per bag were collected, transferred to the laboratory facilities in refrigerated boxes and processed within 24 h for tissue collection and histology.

An initial sample of 30 spat of 4-months old (1.02 ± 0.35 cm in length) produced in Ireland (Site A) was tested by standard PCR to confirm the absence of the virus before being transferred to two infected sites (Site B and C). Primer pairs OHVA and OHVB (Lynch et al., 2013) gave no amplification in any sample. 500 spat were settled by triplicates in bags in the same lines where juveniles were simultaneously sampled. Collection of oysters and the estimation of mortality were carried out as described above.

Temperature loggers (DST CTD, Star Oddi), placed in the same area where the oysters were sampled, recorded temperature data every hour. Maximum temperature reached per day was also included in the analysis.

Shell length and wet weight were measured before dissection. A piece of gill tissue in juveniles was immediately collected after dissection and kept frozen (−20°C) until use. A cross section including the gill, mantle and gonad was fixed in Davidson's solution for 48 h and preserved in Ethanol (70%) before being processed for histology. Depending on the size of the animal and in order to obtain enough DNA material, whole organism excluding the digestive gland or a mix of gill and mantle were collected in spat oysters.

Genomic DNA was extracted from gills (juveniles) or mix of tissues (spat) using the DNeasy Blood and Tissue kit (Qiagen) following the manufacturer's instructions. DNA concentration and quality was assessed by spectrophotometer (NanoDrop, Thermo-Scientific). The number of samples to be analyzed per sampling point was decided attending to the percentage of mortality, 15 animals were processed when normal mortality was observed and 30 in the case of abnormal mortality. A total of 375 juveniles were processed for DNA extraction in five sampling times and 176 spat oysters collected in three sampling times.

According to the method of Lynch et al. (2013) standard PCR using OHVA and OHVB primers (Table 1) was assayed in undiluted DNA samples to determine infection. A total of 105, 120, and 150 juveniles and 45, 90, and 41 spats were screened from Site A, B, and C, respectively. Prevalence of infection was estimated as the mean percentage of positive samples per oyster bag (n = 3). A selection of positive samples (40) was subsequently sequenced to identify the OsHV-1 microvariant. The amplicon obtained with C2-C6 primers (Renault and Arzul, 2001) (Table 1) was diluted in distilled water (1:10) and used as a template in a second round of PCR. Two combinations of primers were used to identify microvariants: (1) OHVC and OHVD primers (Lynch et al., 2013) giving a final fragment of 296 bp to amplify the microsatellite area and (2) C2-C2rev (Lynch et al., 2013) giving a fragment of 400 bp to identify the characteristic polymorphism of Irish OsHV-1 μvar variant. These sequences overlapped and were assembled to obtain a longer fragment for homology searches. PCR products were cleaned up using QIAquick gel extraction kit (Qiagen) and sequenced by the Sanger method using the corresponding primers pairs used for amplification, (SourceBioScience). Raw chromatograms were analyzed using Chromas 231 software (Technelysium). ExPaSy tools (http://us.expasy.org/genomics) and CAP3 (http://doua.prabi.fr/software/cap3) were used for sequence assembly. Multiple sequence alignment were executed by Muscle (Edgar, 2004) and visualized and edited using BioEdit v.7.2.5 (http://www.mbio.ncsu.edu/bioedit/page2.html). Searches of homology were performed on GenBank database using the Blast algorithm (http://ncbi.nlm.nih.gov/blast/). Sequences were deposited on GenBank with the accession number: KX147758.

The quantification of OsHV-1 was carried out by real-time PCR following the Standard Operating Procedure “OsHV-1 detection and quantification by Real Time Polymerase Chain Reaction using OsHV-1 DNA polymerase sequence” (http://www.eurl-mollusc.eu/SOPs) using the HVDP-F and HVDP-R primers (Webb et al., 2007) Each sample was analyzed in triplicate with 5 μl of genomic DNA (5 ng/μl) as template. The reaction contained 12.5 μl of Brilliant SYBR Green Mater Mix reagent (Agilent), 2.5 μl of HVDP-F (5 μM), 2.5 μl of HVDP-R (5 μM) (Table 1) and 2.5 μl of distilled water. Amplification was carried out in 384-microwell plates in an ABI 7900HT Real Time PCR system (Applied Biosystems). The standard cycling conditions consisted of an incubation at 95°C for 10 min allowing enzyme activation followed by 40 cycles of product melt at 95°C for 30 s, primer annealing at 60°C for 1 min and polymerase extension at 72°C for 45 s. A final melting curve analysis was developed following the instrument specifications. Negative controls containing distilled water were included on the plate and showed no amplification. The quantification of virus copies was extrapolated from a standard calibration curve prepared with serial dilutions of a known suspension of OsHV-1 genomic DNA (5 × 10⁵ copies/μL).

Spat survivors from the natural outbreak were transferred to the laboratory facilities and let acclimatize before carrying out an experimental trial with a purified suspension of OsHV-1 μVar. 48 animals from each site (Site A and Site B) were settled in 4 tanks, 2 for virus injected animals and 2 for controls. Spat were injected in the adductor muscle with 50 μl of OsHV-1 μVar suspension (10⁴ DNA viral copies/μl) after being anesthetized using Magnesium Chloride (Suquet et al., 2009) and set up at constant temperature (22°C) under photoperiod regimen. Although, the number of infective particles was unknown, the suspension was previously tested to check infectivity. Control spat were injected with ultraviolet treated and 0.22 μm filtered sea water and kept under the same conditions. Mortality was recorded twice daily and dead animals were removed and screened for OsHV-1μvar detection and estimation of prevalence as described above.

Fixed body sections of infected oysters were processed for histology and in situ hybridization (ISH). Tissues were dehydrated and embedded in paraffin wax. Blocks were cut in 7 μm sections. Standard histology procedures were carried out before staining with haematoxylin and eosin (Sigma) and covered using DPX mounting medium (Sigma). A selection of juveniles confirmed positive by PCR were processed for ISH to detect OsHV-1 in tissues. DNA digoxigenin-labeled probes were synthesized by PCR with OHVC- OHVD primers (Lynch et al., 2013) using the DIG Nucleid Acid Detection Kit (Roche). Solutions and buffers were freshly prepared following Sambrook and Russell (Sambrook and Russell, 2001). Procedure for hybridization, detection of DIG labeled DNA and counterstaining was made following Lynch et al. (2010) with minor modifications for OsHV-1 detection. Briefly, incubation of slides in moist chamber at 37°C for tissue preparation was reduced to 7 min, a higher volume of probe was used to detect OsHV-1 (10 μl) and incubation with nitroblue tetrazolium and 5-bromo-4-chloro-3-indolyl phosphate (NBT/BCIP) was extended to 2 h in the dark. Slides were observed and photographed under light microscopy at 40 × and 100 × magnification (Nikon Eclipse 80i).

Data were analyzed following t-Students Test. Chi square test was used to compare percentage of mortality, prevalence of infection and sea water temperature. Significant differences were considered at p ≤ 0.05. Data are presented as mean ± standard error. Statistical analyses on sea water temperature were carried out on BoxPlotR.

The highest temperature per day varied between 13.7°C and 23.6°C at Site B with a mean value over the survey period of 18.3 ± 0.25°C. Mean temperature was significantly higher at Site C (20.4 ± 0.31°C), varying between 15.8°C and 28.1°C (Figure 2). Temperature varied by 3°C between 1st and 3rd quartiles for both sites. 75% of values recorded were below 19.6°C at Site B and below 21.9°C at Site C. Analysis of temperature per hour showed that oysters from Site C were exposed to 16°C or higher temperatures for longer than oysters from Site B.

OsHV-1 was detected by PCR in juveniles from Site B (6.1 ± 0.8 cm) and Site C (5.8 ± 0.9 cm). OsHV-1 was not detected in naïve oysters from Site A (8.5 ± 0.9 cm) during the survey. OsHV-1 was detected in both infected areas from sampling 2 to sampling 5 (Figures 3A,B). Overall results showed significantly higher prevalence in juveniles from Site B (25%) compared to Site C (10%). Maximum prevalence of infection was 80% at Site B and 30% at Site C. No significant differences were observed on prevalence of infection per sampling point.

Maximum values of mortality reached 34.6% at Site B and 14.2% at Site C. Although, overall mortality was not significantly different between the two locations, comparison per sampling point showed significantly higher mortality at Site B (28.6%) compared to Site C (6.5%) at sampling point 3. No abnormal mortality was observed at Site A.

Episodes of high temperatures, considered in this study as a period of 14 consecutive days with highest maxima over 16°C, were not associated with mortality in juveniles in any of the infected sites. Prevalence of infection over 20% was significantly associated with abnormal mortalities at Site B (p = 0.03). However, a higher significance (p = 0.006) was obtained when temperature was included in the analysis. No significant correlation between mortality, temperature and prevalence was found at Site C.

The overall size of spat transferred to Site B (1.19 ± 0.2 cm) was significantly lower than spat from Site C (1.4 ± 0.3 cm) and spat that remained at Site A (1.5 ± 0.5 cm). Prevalence of infection reached 70% at Site B (Figure 4A) and 100% at Site C (Figure 4B), whereas OsHV-1 was not detected in spat from Site A. No abnormal mortality was observed at Site A.

Overall mortality in spat was not significantly different between sites B (64.2%) and C (46.6%). Analysis per sampling point showed that mortality was significantly higher at Site B in samples 3 (49.9%) and 4 (62.3%). At sample 5, mortality increased significantly in spat from Site C. Overall prevalence of infection was significantly higher at Site C (83.3%) compared to Site B (33.3%). However, at sampling point 3 the prevalence of infection was higher at Site B and no significant differences were observed at sampling point 4 and 5. Abnormal mortality in spat was significantly related to the presence of OsHV-1 at both sites (p = 0.01, Site B; p = 0.04, Site C). Episodes of high temperatures were also related to mortalities at Site C (p = 0.04) but not at Site B (p = 0.2).

OsHV-1 was quantified by qPCR in a selection of samples that were positive by standard PCR. Overall viral load was significantly higher in spat (4.7 × 10⁶ DNA viral copies) compared to juveniles (1.35 × 10⁵ DNA viral copies) (Figure 5). No significant differences were observed between Sites for any of the oyster stages analyzed.

In juveniles from Site B and Site C and spat from Site C the highest viral load was observed at the beginning of the survey, reaching between 1.5 × 10⁵ and 1.07 × 10⁷ DNA viral copies. However, maximum viral load in spat from Site B was reached later, at sampling point 4. Number of viral copies was significantly higher in spat compared to juveniles in sampling point 3 and 4. The viral load decreased in all samples, showing no significant differences between samples and reaching a mean of 7.7 × 10³ viral copies at the end of the survey.

Positive amplicons obtained by the combination of different primer pairs were purified from the agarose gel using the QIAquick Gel Extraction Kit (Qiagen) and bidirectionally sequenced with specific primers. Chromatograms were carefully analyzed to discard any erroneous reading. All sequences from the same set of positive samples were identical. Multiple alignment grouped sequences in two main contigs obtaining two consensus sequences (Table 2). One contig containing 57 sequences from Site C juveniles and both spat stocks (sequence 1) and the other one containing only the 25 sequences from Site B juveniles (sequence 2). Nucleotide Blast searches revealed that sequence 1 had higher identity to OsHV-1 μVar than sequence 2 and both sequences had the same identity compared to the reference genome OsHV-1 (Table 2). Differences between consensus sequences described herein and previous microvariants sequences were observed in the nucleotide alignment (Figure 6). The polymorphisms on the C region that characterized the μVar variant were conserved in the two consensus sequences including the microsatellite deletion. One nucleotide mismatch was observed between sequence 1 and 2 (Figure 6, boxed region), consisting on a Guanine deletion at position 178430 (Davison et al., 2005). Compared to OsHV-1 μVar, translation of sequence 2 produced a different amino acid sequence at the beginning of ORF5. However, the protein sequence of ORF4 was not modified by this polymorphism (data not shown).

Figure 7 shows the percentage of survival after intramuscular injection of OsHV-1 μVar in spat survivors from the natural outbreak. A decrease on survival was observed at 24 h post-injection in all injected animals. Dead animals from control conditions were not positive for OsHV-1 μVar detection by standard PCR over the 4 days trial. At 48 h post-injection, virus injected spat from Site A experienced a significant decrease in survival compared to controls (p = 0.02). No significant differences were observed between virus and control exposed animals from Site B (p = 0.05). Comparison between the two stocks revealed that survival was significantly higher in virus injected animals from Site B compared to virus injected animals from Site A (p = 0.03). Prevalence on virus injected animals was observed from day 1 in Site A and from day 2 in Site B reaching a peak of 80 and 100%, respectively.

Cellular changes were predominantly observed in gills (Figure 8A). Hypertrophied cells with a low nucleus-cytoplasm ratio (arrows, Figure 8A) and nuclear abnormalities such as chromatin fragmentation and margination, and pycnosis were observed (arrowheads, Figure 8B). Infiltrated cells found in mantle were apparently intact (Figure 8B). The amount of circulating cells infiltrating this tissue was relatively low compared to gills. A number of epithelial cells showed enlarged and damaged nucleus compared to intact cells (arrowheads, Figure 8B).

Detection of virus DNA by ISH was characterized by blue precipitates into host cells. Uninfected oyster showed no specific labeling (Figure 9A). Although, a remaining light blue could be observed in some areas this was due to some residual staining and not to positive OsHV-1 marking. In further confirmation of histology observations, blue precipitates were very abundant in gill epithelial cells of infected animals (Figure 9B). The intensity of the signal was also strong in enlarged cells that showed clear blue precipitates in the cytoplasm (Figures 9B,D). Degradation of tissue probably due to the in situ hybridization procedure was more evident in mantle (Figure 9C). Cell structures in epithelial cells were not easily distinguished and blue staining was more diffuse occupying mainly enlarged nuclei.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.