Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
1 Interactions between Paramoeba perurans, the causative agent of amoebic gill disease, 2 and the blue mussel, Mytilus edulis 3 Christine Rolin1,2*, Jennifer Graham2, Una McCarthy2, Samuel A. M. Martin1, Iveta 4 Matejusova2* 5 1 School of Biological Sciences, University of Aberdeen, Tillydrone Avenue, Aberdeen AB24 2TZ, UK 6 2 Marine Scotland Science, Marine Laboratory, 375 Victoria Road, Aberdeen, AB11 9DB, UK 7 8 Abstract: Amoebic gill disease (AGD) is caused by the ectoparasite Paramoeba perurans 9 found free-living in seawater. In recent years outbreaks of AGD have occurred in most 10 salmon farming countries causing significant economic losses. Mussels co-cultured with 11 salmon in integrated multi-trophic aquaculture (IMTA) systems may change pathogen 12 dynamics on sites by acting as reservoirs or biological controls. Through the use of an 18S 13 rRNA gene quantitative real-time PCR we tested the interactions between P. perurans and 14 blue mussels (Mytilus edulis) under experimental conditions by means of water-borne 15 transmission. Quantification of DNA from water samples revealed a rapid decrease in P. 16 perurans DNA over two weeks in the presence of mussels under experimental conditions. P. 17 perurans was detected on swabs from mussel shells up to 48 hours post-exposure. 18 Additionally, no P. perurans were detected in mussels collected from natural mussel beds 19 and fish farms. These results indicate mussels are not a likely reservoir host for P. perurans 20 but may in fact actively remove water-borne P. perurans. 21 22 Keywords: Blue mussel, Paramoeba perurans, amoebic gill disease, biological control, 23 Atlantic salmon, integrated multi-trophic aquaculture. 24 *Corresponding 25 [email protected] Present address: NAFC Marine Centre, Port Arthur, Scalloway, 26 Shetland, ZE1 0UN, UK. Iveta Matejusova: E-mail address: 27 [email protected] . authors: Christine Rolin: Tel.: + 44 (0) 1595 772300; E-mail address: 28 29 1. Introduction 30 Amoebic gill disease (AGD) is an economically significant disease of Atlantic salmon 31 (Salmo salar Linnaeus, 1758) caused by an amoeboid protozoan parasite Paramoeba 32 perurans (synonymous to Neoparamoeba perurans, according to Feehan et al., 2013) (Shinn 33 et al., 2015). In Tasmania, AGD has long been one of the main problem in salmonid 34 aquaculture, with typical losses of 10-20 % of production costs (Munday et al., 2001). In 35 Scotland typical mortalities on Atlantic salmon farms range between 10-20 % but have been 36 reported to be as high as 70 % at some sites (Marine Scotland Science, 2012). In recent years 37 a global emergence of this disease has been documented with frequent outbreaks across the 38 Northern and Southern Hemispheres i.e. Ireland, Scotland, Norway, Spain, France, North 39 America, Japan, Chile and South Africa (Rodger & McArdle, 1996; Palmer et al., 1997; 40 Findlay & Munday, 1998; Steinum et al., 2008; Crosbie et al., 2010; Bustos et al., 2011; 41 Mouton et al., 2014). Paramoeba perurans is a free-living amoeba, however, when it 42 colonises Atlantic salmon gills it can cause white swollen lesions, epithelial hyperplasia, and 43 increased mucus production leading to lethargy, apoxia and eventually death (Munday et al., 44 2001). 45 Very little is known about the origin of P. perurans responsible for disease outbreaks. 46 Fish import and subsequent spread between farms is not generally thought to be a feasible 47 explanation for emergence of this disease in Europe (Steinum et al., 2008). On the other 48 hand, prolonged periods of increased seawater temperature and/or high salinities combined 49 with densities of fish in aquaculture sites most likely provided an advantage to P. perurans 50 naturally present in marine environment (Nowak, 2007; Bridle et al., 2010; Wright et al., 51 2015). In aquaculture, infections with P. perurans were reported from a range of farmed fish 52 in addition to Atlantic salmon (Crosbie et al., 2010; Karlsbakk et al., 2013; Mouton et al., 53 2014) while wild fish collected in the proximity of infected Atlantic salmon farms in 54 Australia were tested negative (Douglas-Helders et al., 2002). 55 Diversification of aquaculture, reduction of potential adverse environmental impacts of 56 Atlantic salmon farm systems and expansion of the shellfish industry are major aspirations 57 intrinsic to the successful expansion of the aquaculture sector in Scotland. Integrated multi- 58 trophic aquaculture (IMTA) systems are being developed in Atlantic salmon farming 59 countries to test the concept of integration of Atlantic salmon, blue mussels and/or seaweed 60 production. The aim is to deliver a diverse aquaculture system with a parallel reduction in 61 organic waste from fish and dissolved inorganic nutrients (Barrington et al., 2009; Chopin et 62 al., 2012; Wang et al., 2012). Co-culturing different species can change disease dynamics on 63 farm sites and create new risks which need to be addressed to ensure good biosecurity in the 64 production systems. Mussels are filter-feeders and have the potential to digest and inactivate 65 some pathogens such as sea lice (Lepeophtheirus salmonis Krøyer, 1837) (Molloy et al., 66 2011; Webb et al., 2013) or infectious salmon anaemia virus (ISAV) (Molloy et al., 2012). 67 On the other hand Vibrio anguillarum (Pietrak et al., 2012), Loma salmonae (Putz, Hoffman 68 & Dunbar, 1965) Morrison & Sprague, 1981 (McConnachie et al., 2013) or infectious 69 pancreatic necrosis virus (IPNV) (Molloy et al., 2013) may persist and accumulate in 70 shellfish posing a threat of disease transmission to farmed fish. 71 The aim of the study was to assess mussels as a potential wild reservoir of P. 72 perurans and investigate interactions between Mytilus species (Linnaeus, 1758) and P. 73 perurans in terms of pathogen accumulation and transmission. Previously published 74 molecular tests (real-time PCR) (Fringuelli et al., 2012) were applied to (1) examine the 75 persistence of P. perurans in water, on shell surfaces and in mussel tissues after experimental 76 exposure and (2) test for the presence of P. perurans in mussel tissues from wild mussel 77 individuals and from bio-fouling mussels collected on fish farms in Scotland. 78 79 2. Materials and methods 80 2.1. P. perurans culture maintenance and dilution curve for quantification 81 Paramoeba perurans, maintained at the Marine Scotland Science (Aberdeen, UK), 82 originated from gills of infected farmed Atlantic salmon (west coast of Scotland) collected in 83 2012. A culture was established following the protocol of Crosbie et al., (2012). The P. 84 perurans culture was maintained routinely at 15 – 18 oC on malt yeast agar (MYA) (0.1 % 85 malt, 0.1 % yeast, 2 % Bacteriological agar, 35 ‰ 0.22 μm filtered seawater) overlaid with 7 86 ml 0.22 μl filtered seawater (SSW, 35 ‰) collected offshore of Stonehaven (Aberdeenshire, 87 UK). The culture was subcultured every two weeks by transferring the seawater overlay 88 containing a free-floating population of amoebae to a new MYA plate. The old MYA plate 89 containing a population of attached amoebae received a new seawater overlay. The culture of 90 P. perurans was first isolated in November 2012 and used in experimental challenges 91 described in the present study in May and June 2014. The pathogenicity of this culture has 92 been confirmed under experimental conditions by challenging naïve Atlantic salmon at 7, 12 93 and 24 months post-isolation (Marine Scotland Science, unpublished data). 94 P. perurans were harvested from culture plates by centrifugation of culture 95 supernatants at 6000 x g for 5 min. Pellets were pooled and resuspended in SSW. Before 96 harvesting, quadruplicate log10 dilutions of were made in 96-well plates for enumeration of 97 the total number of amoebae harvested under light microscope. After centrifugation, log10 98 dilutions were made in duplicate from a starting cell density of P. perurans of 7 x 103 cells. 99 Then, homogenised mantle-gill tissue was spiked with the prepared dilutions and total DNA 100 was extracted using two different DNA extraction protocols: automated extraction using the 101 QIAsymphony DNA DSP mini kit (Qiagen) or manual extraction using the MasterPure DNA 102 and RNA Purification kit (Epicentre). Negative controls representing mussel tissue 103 homogenate without P. perurans and blank controls were set up. P. perurans DNA was 104 quantified using the quantitative real-time PCR (qPCR) (see section 2.4.) and Ct values 105 recorded to create a standard dilution curve in R’ statistical package (version 3.0.2, R Core 106 Team 2013). 107 108 2.2. Experimental exposure of mussels to P. perurans culture 109 2.2.1. Mussel maintenance 110 Depurated market-size mussels were bought from a commercial supplier or vendor 111 and kept in outdoor covered seawater tanks at 9 oC. The seawater was from Nigg Bay 112 (Aberdeen, Scotland) was filtered (through a 1 mm wedge wire screen and 15 μm particle 113 filter) before pumping to storage tanks. This water was also used in experimental beakers and 114 tanks. Mussels were naturally feeding on nanophytoplankton present in the sea water, 115 considered as one of significant dietary components for bivalves (Shumway et al., 1985) and 116 no additional food was added to the holding tanks prior to the experiments. 117 118 2.2.2. Experiment 1: Pilot experiment to establish the time frame for P. perurans ingestion by 119 mussels and distribution in mussel tissues 120 Forty eight beakers containing 500 ml 35 ‰ seawater (10 oC) with a small air stone 121 were placed in a temperature-controlled room (18 oC). By the end of the experiment (144 122 hours after exposure) the water temperature had risen to 14 oC. Five mussels were placed in 123 each duplicate treatment beaker and allowed to acclimatise before adding P. perurans. 124 Mussels were exposed to P. perurans (104 cells L-1 in suspension) in 500 ml and sampled at 125 1, 2, 3, 4, 6, 24, 48 and 144 hours post-exposure (hpe). The mussel filtering activity was 126 monitored throughout the experiment. 127 At each time point, four beakers were sampled without replacement: two treatment 128 beakers (containing P. perurans and mussels), one positive control beaker (P. perurans only) 129 and one negative control beaker (mussels only). At each time point, all mussels were 130 removed and seawater from the entire beaker was filtered through 1.0 µm cellulose nitrate 131 membrane filters (25 mm diameter, GE Healthcare Life Sciences) which were stored dry in 2 132 ml Safelock tubes (Eppendorf). The shell of each individual mussel was swabbed with a 133 sterile cotton-tip swab on one side using four strokes lengthwise. Additionally, the bottom 134 inside edge of each beaker was swabbed using two circular strokes to test adhesion and 135 replication of P. perurans on plastic surfaces. All swabs were stored in RNAlater (VWR). 136 The whole digestive gland and a section of mantle and gill tissue were sampled from each 137 mussel and placed in separate tubes containing 100 % ethanol (Sigma). All samples were 138 frozen at -20 oC immediately after sampling. 139 140 2.2.3. Experiment. 2: Tank experiment to quantify removal rates of P. perurans from the 141 water column 142 Mussels were exposed to P. perurans (3 x 104 cells L-1 in suspension) in tanks 143 containing 20 L of 35 ‰ seawater. These tanks were placed within a larger tank connected to 144 a flow-through seawater system to maintain the experimental tanks at 12 oC throughout the 145 experiment. Thirty mussels were placed in four tanks: three treatment tanks (containing P. 146 perurans and mussels), one positive control tank (P. perurans only) and one negative control 147 tank (mussels only). The mussels were fed with algae (Shellfish diet 1800 from Reed 148 Mariculture Inc.) according to the manufacturer’s instructions, when a reduced filtering 149 capacity was observed (day 7 post-exposure). 150 Water samples and shell swabs were taken at 1, 4, 7, 24, 48, 168 and 384 hpe. A 500 151 ml water sample was taken from each tank and filtered using the method described above. 152 The entire shell of three mussels in each tank was swabbed thoroughly; sampled mussels 153 were placed in a net (3 mm diameter nylon mesh), within the tank to avoid repeated swabbing 154 at different time points. The swabs were placed in a dry microcentrifuge tube (opposed to 155 RNAlater as in experiment 1), immediately stored at -20 °C and processed within 48 hours. 156 Finally, a section of net (1 x 1 cm) (retaining sampled mussels) from each tank was placed in 157 a dry tube to test P. perurans adhesion. Mussels were checked daily and filtering activity and 158 mortalities were recorded. 159 160 2.3. Sampling of wild and biofouling mussels 161 To assess presence of P. perurans in wild mussels fifty individuals were collected 162 from each of three sampling points on the west coast of Scotland (Lunderston Bay, Loch 163 Long (Ardgarten) and Rascarrel Bay) in 2013. To assess presence of P. perurans in mussels 164 in aquaculture farm environment, a sample of 150 mussels, settled directly on the surface of 165 Atlantic salmon fish cages, were collected in 2012 from a farm on the west coast of Scotland. 166 This particular farm had an ongoing P. perurans infection in the time of sample collection. A 167 pooled sample of gill and mantle was collected from each individual and stored in 100 % 168 ethanol (Sigma) until further processing. All field samples were homogenised (see section 169 2.4.) and stored at -80 oC for up to two years before processing in the current study. 170 171 2.4. DNA extraction and qPCR 172 2.4.1. Water filters 173 DNA captured on water filters was extracted manually using the MasterPure 174 Complete DNA and RNA Purification kit (Epicentre) following the protocol for DNA 175 purification of tissue samples. Each filter was processed by adding 125 mg of glass beads 176 (710 – 1180 μm, Sigma), 1 μl proteinase K (600 mAU/ml solution) and 300 μl of Tissue and 177 Cell Lysis solution, then placed in the TissueLyser (Qiagen Ltd.) and homogenised at 25 Hz 178 for 5 min. A negative control contained 1 μl proteinase K and 300 μl Tissue and Cell Lysis 179 solution. Due to a large quantity of debris, protein precipitation was repeated once. As pellets 180 were difficult to resuspend, samples were placed in the fridge for 30 min and then 181 resuspended by pipetting. Genomic DNA was stored at -20 oC prior to further analysis. 182 183 2.4.2. Shell swabs 184 Swabs fixed in RNAlater or dry tubes were transferred to an empty 2 ml Safelock 185 tube containing 20 μl Proteinase K and 180 μl ATL buffer (QIAsymphony DNA DSP mini 186 kit, Qiagen), vortexed briefly and lysed by incubating at 56 oC for a minimum of 3 hours. A 187 negative extraction control of 20 μl proteinase K and 180 μl ATL buffer was included. The 188 tubes were centrifuged for 2 min at 13,500 rpm and supernatant was transferred into a clean 189 1.5 ml tube. Total DNA was extracted using a QIAsymphony DNA DSP mini kit and the 190 QIAsymphony SP robot (Qiagen) according to manufacturer’s instructions and resuspended 191 in 100 μl of distilled water. Samples of nets at 384 hpe in experiment 2 were processed 192 according to the same methods. 193 194 2.4.3. Mussel tissues 195 Mussel tissue (from experimental and field studies) was homogenised in 1 ml of ATL 196 buffer (Qiagen) using a 7 mm metal bead at 25 Hz for 5 min, in a Tissuelyser homogeniser 197 (Qiagen). DNA was extracted from a volume of tissue equivalent to 5 mg of tissue as 198 described in the section above. All samples were stored at -20 oC until further analysis. 199 200 2.4.4. Quantitative Real-Time PCR (qPCR) 201 All DNA samples were tested for the presence of P. perurans using previously 202 published qPCR assays targeting a 139-bp fragment specific to the P. perurans 18S rRNA 203 gene (Fringuelli et al., 2012). As an endogenous control to confirm the quality of extracted 204 DNA, ME15/16 Mytilus spp. qPCR assays were run in multiplex for simultaneous detection 205 of all three Mytilus species (M. edulis, M. galloprovincialis, M. trossulus) according to Dias 206 et al., (2008). 207 The final qPCR reaction mixture contained 2 x Quanta Custom Toughmix (VWR), 208 900 nM of each primer and 250 nM of each TaqMan® probe, 1 μl genomic DNA at a 209 concentration of 50 or 100 ng ul-1 in final volume of 20 μl. The real-time PCR was performed 210 on the LC480 Thermal Cycler (Roche). The reproducible limit of detection (RLOD) for the 211 P. perurans assay was 36.22 Ct (Fringuelli et al., 2012) and this value was used to distinguish 212 between P. perurans negative and positive samples. An increase of 3 Ct was considered a 10- 213 fold decrease. 214 215 216 217 2.5. Data analyses All data analyses were performed in R’ statistical package (version 3.0.2, R Core Team, 2013) and plotted with ggplot2 package (Wickham, 2009). 218 Experiment 1 was designed as a pilot experiment to test the sampling methods and 219 detection of P. perurans in water, swabs and tissues, as well as establish the rate at which P. 220 perurans might be extracted from water by filtering mussels. Therefore, no models were 221 applied and the results were analysed by calculating means and standard deviations of P. 222 perurans Ct values for each treatment group. 223 In experiment 2 all tanks were independent from each other and were sampled 224 repeatedly over the duration of the experiment. Only the data from water extraction were 225 modelled, to determine any change in P. perurans over time and between treatments (P. 226 perurans with and without mussels). Models were tested for effects and interactions between 227 tanks before pooling data in the respective treatment groups. Data transformations were 228 performed and the best model was chosen from linear models, including interactions and 229 fixed factors, and from mixed effect models with tank as a random factor. The best fit model 230 was chosen based on an Akaike Information Criterion (AIC) comparison, summary plots and 231 residual plots. If the difference in AIC was ≤ 2 between two models, the model with less 232 variables was considered the better fit. Analysis of Variance (ANOVA) was used to test any 233 significant differences over time and between treatments. 234 235 236 The shell swabs were analysed as binomial data based on proportion of swabs positive and tested with a Chi test to detect significant changes over time. 237 3. Results 238 3.1. Detection of P. perurans in spiked mussel tissue 239 The dilution curves generated from P. perurans-spiked mussel tissue showed that the 240 MasterPure extraction protocol has a greater extraction efficiency than the DNA DSP 241 protocol and confirm that P. perurans can be detected down to 70 cells for the DNA DSP and 242 7 cells for the MasterPure kit (Table 1). A linear relationship between the concentration of P. 243 perurans-spiked mussel tissue and the Ct values was observed for both the DNA DSP mini 244 kit (y = -4.28x + 43.5, R2 = 0.99) and MasterPure kit (y = -3.38x + 37.91, R2 = 0.99) (Fig. 1). 245 However, DNA yield from the DNA DSP kit might be underestimated (by up to 1 Ct) as only 246 200 μl of lysed solution was extracted using the QIAsymphony compared to 300 μl used for 247 manual MasterPure kit extraction. 248 249 3.2. Experimental exposure of mussels to P. perurans culture 250 Observed mortality in experiment 1 and 2 reached 6 % and 22 %, respectively, by the 251 end of the experiment. In experiment 2, all mortalities occurred in one out of three individual 252 treatment tanks, 3 mussels died before day 7 and 15 mussels by the end of the experiment. 253 Causes of mussel mortality were not investigated. 254 255 3.2.1. Detection of P. perurans in water samples 256 In experiment 1, a clear reduction in quantity of P. perurans (103-fold) was detected 257 in water after 1 hour and it remained low until day 16 (P. perurans Ct = 33.4 ± 2.0 0 with 258 mussels and Ct = 25.7 ± 0.9 without mussels) (Fig. 2). 259 In experiment 2, only data from 1 to 168 hpe (i.e. 7 days) were modelled as P. 260 perurans Ct values were below RLOD (i.e. positive) in only one of three with mussels on 384 261 hpe. Due to a non-linear relationship between Ct values and time, the values for hpe were 262 log10 transformed. Despite a significant variation in the rate of the reduction of P. perurans 263 between tanks (F = 3.69, df = 2, p < 0.05) data from all tanks were pooled in the final model 264 as the model excluding the tanks provided the same or better fit based on summary plots, 265 residual plots and the AIC values (< 2 AIC difference). The final model included an 266 interaction between treatment tank and hpe as explanatory variables. 267 A rapid and significant decrease in detectable P. perurans was recorded in the 268 presence of mussels, after 1 hour the difference between the tanks with and without mussels 269 was > 10-fold (Table 2). P. perurans were reduced a further 102-fold between 1 and 24 hours 270 and a 104-fold between 1 hour and 7 days (F = 86.7, df = 1, p < 0.001; y = 3.73x + 25.29, R2 271 = 0.83) (Fig. 3). Some reduction in P. perurans with time was also observed in the positive 272 control tank (in the absence of mussels) with a 10-fold change (range 21.6 – 23.7 Ct) across 273 the 164 hpe (F = 15.25, df = 1, p < 0.05; y = 0.96x + 21.47, R2 = 0.74) (Fig. 3). The rate of 274 reduction in P. perurans was significantly greater in the mussel treatment group compared to 275 the positive control group (F = 169.79, df = 1, p < 0.001). No P. perurans DNA was detected 276 in water samples from negative control beakers or tanks. Endogenous Mytilus spp. multiplex 277 assays indicated that Mytilus DNA levels in the water of all tanks were stable throughout the 278 experiment (range: 26.5-32.5 Ct). 279 280 3.2.2. Shell swabs 281 Mussel shells were swabbed to determine if P. perurans attaches to shells and uses 282 them as a surface for attachment, survival or replication. The proportion of shells testing 283 positive in both experiments are given in Table 3. All shell swabs from negative control 284 beakers and tanks tested negative for P. perurans DNA. Additionally, all swabs from the 285 positive control beakers tested positive. 286 In experiment 1, 1/10 shell swabs tested positive between 1 and 4 hpe with most Ct 287 values below the RLOD (mean: 36.01 + 1.46, range: 33.70-38.41 Ct). In experiment 2, where 288 a modified sampling protocol was used 9 out of 10 swabs tested positive between 1 and 4 hpe 289 and lower Ct values were obtained for P. perurans than in experiment 1 (mean: 33.98 + 2.14 290 Ct, range: 28.94 – 38.71 Ct). 291 In both experiments P. perurans was detected on swabs only during the first 48 hpe. 292 In experiment 2, all shells tested positive 1 hpe but only 50 % tested positive after 48 hpe. As 293 there was no significant tank effect on the proportion of P. perurans shell swabs testing 294 positive, data were pooled. The reduction in the proportion of positive shell swabs was 295 significant over time (Chi test: p < 0.001). Additionally, swabs taken from the bottom of the 296 beakers in experiment 1 all tested negative when mussels were present (range ≥ 36.58 Ct), 297 when mussels were absent 3 out of 8 beakers tested positive (range ≥ 32.24 Ct). Samples of 298 nets taken at 385 hpe in experiment 2 tested negative when mussels were present and positive 299 when mussels were absent (range: 26.0 – 27.27 Ct). Nets were also positive for Mytilus spp. 300 with lower Ct values (range: 34.73 – 36.83 Ct) than for P. perurans. 301 302 3.2.3. Detection of P. perurans in mussel tissues 303 To determine a potential of P. perurans accumulation in mussel tissues and to 304 establish a time frame for accumulation and/or release of P. perurans, the gill, mantle and 305 digestive gland were tested for presence of the parasite by qPCR. No P. perurans DNA was 306 detected in any tissues in either experiment. All mussels from the negative control group 307 were negative for P. perurans DNA. Endogenous Mytilus spp. multiplex assay indicated that 308 all DNA extracted from mussels and tissues was of good quality (range: 20.6 – 28.0 Ct). No 309 P. perurans or Mytilus DNA was detected in their respective negative controls. 310 311 3.3. Detection of P. perurans in naturally settled and biofouling mussels 312 The combined mantle-gill tissue samples collected from natural mussel beds (Lunderston, 313 Long, Rascarrel Bay) and fish farm cage (west coast of Scotland) tested negative for P. 314 perurans. Endogenous Mytilus spp. multiplex assay indicated that all DNA extracted from 315 mussels tissues was of good quality (all values < 30 Ct). No P. perurans or Mytilus DNA was 316 detected in their respective negative controls. 317 318 4. Discussion 319 Paramoeba spp. have previously been reported in shellfish such as lobsters (Mullen et 320 al., 2005), sea urchins (Feehan et al., 2013) and oysters (Sühnel et al., 2014) but to date P. 321 perurans has not been detected in any shellfish and the only known environmental reservoir 322 is free-living in sea water (Bridle et al., 2010, Wright et al., 2015). Similarly, the present 323 study did not detect any P. perurans in mussels collected from natural beds nor selected fish 324 farms on the west coast of Scotland. 325 Mussel shell surfaces have not previously been considered as an environmental 326 reservoir for fish diseases. To assess the mussel shell’s potential as a substrate for P. 327 perurans attachment and replication, the shells from mussels experimentally exposed to high 328 numbers of P. perurans were swabbed according to the protocol developed for fish gill swabs 329 (Bridle et al., 2010). The method yielded Ct values below RLOD in experiment 1. A change 330 in the protocol (increased sample area and storage in dry tubes) was more effective, with 331 more positive detections in experiment 2. The improvement in detection of P. perurans may 332 be attributed to increasing the quantity of DNA on each swab or storage in dry tubes may 333 have prevented loss of DNA into the fixative solution. 334 The present study suggested that mussel shells only provide a suitable surface for P. 335 perurans up to 48 hours post exposure when the parasite is present in high numbers. Shell 336 swabs had Ct values for P. perurans comparable to those from gill swabs from AGD-infected 337 fish with gill scores of 1 or 2 (Bridle et al., 2010), which is considered a light infection 338 (Powell et al., 2001). Using the calibration curve from the present study, the Ct values 339 detected on the surface of mussel shells at 24 hpe correspond to approximately 100 amoebae, 340 and 10 amoebae after 48 days and none at the end of the experiment (144 hpe). Therefore, it 341 can be concluded that mussel shells most likely do not act as a long-term environmental 342 reservoir for P. perurans. In comparison, swabs of the inside of the beaker surface 343 (experiment 1) and net samples taken at 384 hpe (experiment 2) tested positive for P. 344 perurans when mussels were absent, but not when mussels were present, indicating P. 345 perurans do attach to surfaces in the absence of mussels. When the ambient water had a 346 concentration of approximately 10,000 amoebae L-1, the Ct values of the net were equivalent 347 to as much as 1,000 amoebae which is more than expected considering the small size of the 348 net sample (1 x 1 cm). Materials such as plastic and net surfaces are abundant on salmon 349 farms and should be considered as a potential substrate for P. perurans attachment and/or 350 replication. 351 Development of IMTA has a clear economic and environmental benefit as it widens 352 the aquaculture product portfolio and removes organic waste associated with unused fish feed 353 (Stirling & Okumus, 1995; Whitmarsh et al., 2006; Macdonald et al., 2011; Handå et al., 354 2012). While species interactions in polyculture may be exploited for bio-mitigation of fish 355 farm waste or as biological controls of pathogens, they may also pose new potential pathways 356 for disease transmission. Uptake and release of variety of a finfish pathogens by mussels and 357 scallops have been tested under experimental conditions. Some pathogens appear to be taken 358 up and inactivated by mussels e.g. sea lice copepodids (Molloy et al., 2011; Bartsch et al., 359 2013; Webb et al., 2013) and ISAV (Skår & Mortensen 2007; Molloy et al., 2012) while 360 others such as Vibrio anguillarum (Pietrak et al., 2012), IPNV (Mortensen et al., 1992; 361 Molloy et al., 2013) and L. salmonae (McConnachie et al., 2013) can be accumulated, remain 362 viable in tissues and can be shed in faeces long after exposure. For example, up to 70 % 363 mortality were observed in Atlantic cod (Gadus morhua) exposed to V. anguillarum-infected 364 mussel faeces under experimental conditions (Pietrak et al., 2012). On the other hand, no 365 mortality of Atlantic salmon cohabited with IPNV-positive mussels was observed, although 366 the rate of transmission (determined by qRT-PCR and virus culture) between mussels and 367 Atlantic salmon, and between fish infected with IPNV by intraperitoneal injection and naïve 368 fish was comparable, suggesting a risk of disease transmission at potential IMTA sites 369 (Molloy et al., 2013). However, caution must be used when interpreting and utilising results 370 of experimental challenges in estimating risk of disease transmission under the dynamic 371 conditions of open sea cage farming. For example, in a direct immersion experiment mussels 372 were able to take up ISAV but no mussels were found positive when cohabited with 373 artificially infected Atlantic salmon exhibiting mortality and clinical signs of the ISA disease 374 or during ISA outbreaks at farm sites (Skår & Mortensen, 2007). 375 In the present study, no evidence was found to suggest that P. perurans is capable of 376 persisting or accumulating in mussels under experimental conditions. Despite exposing 377 mussels to high numbers of viable P. perurans (experiments 1 and 2: 2 x 104 amoebae L-1 and 378 3 x 104 amoebae L-1 respectively) no values above the RLOD were detected in gill, mantle or 379 digestive gland using a qPCR assay (Fringuelli et al., 2012). Sensitivity of the qPCR assay 380 was tested in the present study and it was confirmed that P. perurans can be detected down to 381 70 amoeba in spiked mussel tissue spiked material. Similar studies considered an adverse 382 effect of PCR inhibitors on the ability of qPCR assays to detect accumulated pathogens in 383 mussels, particularly in the digestive gland (Schrader et al., 2012; Molloy et al., 2013). 384 However, the endogenous Mytilus spp. assay used in the current study did not indicate that 385 inhibition was occuring, with the range of Ct values in the digestive gland (22.6 ± 1.3 Ct) 386 comparable to other tissues such as mantel and gill (23.3 ± 1.6 Ct). 387 Preservation and storage of samples can affect the quality of DNA and the final result 388 of the analysis (Ørpetveit et al., 2010). Gill swabs are commonly stored in ethanol before 389 AGD diagnosis (Young et al., 2008; Crosbie et al., 2010; Karlsbakk et al., 2013). The storage 390 of tissue samples in ethanol with the addition of a freezing step may have led to lysis of P. 391 perurans cells and loss of material into the fixative solution. Water filters and shell swabs 392 were stored in dry tubes instead of ethanol and consistently tested positive in the experiments. 393 Other studies have successfully detected sea lice, ISAV and IPNV from digestive gland 394 stored directly in lysis buffer after sampling (Molloy et al., 2011, 2012, 2013) and P. 395 perurans on water filters stored at room temperature (Bridle et al., 2010; Wright et al., 2015). 396 Despite the absence of P. perurans DNA in mussel tissues, in both experiments a 397 significant decrease of P. perurans DNA signal over time was observed in water from the 398 beakers/tanks where mussels were kept compared to the positive control tank. It should be 399 emphasised that the P. perurans Ct values in the positive control beakers/tanks did not rapidly 400 change, and remained stable increasing by approximately 10-fold across 16 days (Table 2) 401 and therefore reduction of P. perurans DNA signal in treatment tank cannot be explained by 402 DNA degradation over time. When compared to the positive control, after only 1 hour P. 403 perurans decreased by 103-fold in experiment 1 (0.1 L of water per mussel) and 10-fold (2 404 litres of water per mussel) in experiment 2 (overall 104-fold across 7 days), indicating a very 405 rapid removal that increases with mussel density. Furthermore, to support the hypothesis that 406 higher density increases removal rate it can be mentioned anecdotally that the tank with the 407 highest mussel mortalities (i.e. lowest filtering capability) experienced a lower removal rate 408 than the other two tanks. A rapid uptake and re-release of P. perurans could explain the lack 409 of accumulation/detection of P. perurans in any analysed mussel tissues, however, as the 410 water samples would contain faeces and pseudofaeces, it is unlikely any accumulation in the 411 waste products occurred as there was no secondary spike in detection of P. perurans target 412 DNA. Changes to mussel feeding behaviour observed towards the end of tank experiment 413 could have affected results of the study as filtering activity was reduced at this point. 414 However, no P. perurans was detected in the mussel tissues even in the very early time points 415 where active feeding was observed, in both beaker and tank experiments. 416 Prevention is the key to controlling outbreaks of diseases that are difficult to treat 417 and/or manage. An understanding of the source and potential spread of pathogen in natural 418 ecosystems as well as systems such as IMTA is therefore crucial for aquaculture expansion 419 and sustainability. The present study indicates that the presence of mussels on Atlantic 420 salmon farms do not pose an increased risk in terms of accumulation of P. perurans and 421 release of viable parasites. This is an agreement with observations made in Shetland where 422 mussels on rafts appeared unaffected during widespread AGD outbreaks (personal 423 communication, Margaret Nugent, SSQC Ltd). However, P. perurans may be able to settle 424 on hard surfaces such as plastic or net structures on fish farm cages, these should therefore be 425 considered as potential environmental reservoirs for P. perurans. Further work is needed to 426 understand the mechanism leading to reduction of P. perurans in the presence of blue 427 mussels in order to determine if mussels can be considered as biological control against 428 AGD. 429 430 Acknowledgements: The authors would like to thank staff at Marine Scotland Science 431 Patricia White, Rebecca McIntosh, Julia Black and Mark Fordyce for their technical 432 assistance and invaluable feedback on the project. Thanks also go to Alex Douglas at the 433 University of Aberdeen for his advice on data analysis and statistics. For feedback on the 434 manuscript thanks to Lesley McEvoy and Rhiannon Inkster at the NAFC Marine Centre. The 435 study was supported by the Marine Collaborations Forum (MarCRF) which aims to develop 436 cross-disciplinary research between the University of Aberdeen and Marine Scotland 437 Science. Finally, thanks are also due to Scottish Fishermen’s Trust for a student support 438 bursary. 439 440 441 442 443 444 445 446 447 448 449 450 451 452 References 453 Barrington, K., Chopin, T., Robinson, S., 2009. Integrated multi-trophic aquaculture (IMTA) 454 in marine temperate waters. In: Soto D (ed) Integrated Mariculture: A Global Review, p 7– 455 46. FAO Fisheries and Aquaculture Technical Paper No. 529 FAO Rome. 456 Bartsch, A., Robinson, S.M.C., Liutkus, M., Ang, K.P., Webb, J., Pearce, C.M., 2013. 457 Filtration of sea louse, Lepeophtheirus salmonis, copepodids by the blue mussel, Mytilus 458 edulis, and the Atlantic sea scallop, Placopecten magellanicus, under different flow, light and 459 copepodid-density regimes. J. Fish Dis. 36, 361–370. 460 Bridle, A.R., Crosbie, P.B.B., Cadoret, K., Nowak, B.F., 2010. Rapid detection and 461 quantification of Neoparamoeba perurans in the marine environment. Aquaculture 309, 56– 462 61. 463 Bustos, P.A., Young, N.D., Rozas, M.A., Bohle, H.M., Ildefonso, R.S., Morrison, R.N., 464 Nowak, B.F., 2011. Amoebic gill disease (AGD) in Atlantic salmon (Salmo salar) farmed in 465 Chile. Aquaculture 310, 281–288. 466 Chopin, T., Cooper, J.A., Reid, G., Cross, S., Moore, C., 2012. Open-water integrated multi- 467 trophic aquaculture: environmental biomitigation and economic diversification of fed 468 aquaculture by extractive aquaculture. Rev. Aquaculture 4, 209–220. 469 Crosbie, P.B.B., Ogawa, K., Nakano, D., Nowak, B.F., 2010. Amoebic gill disease in 470 hatchery-reared ayu, Plecoglossus altivelis (Temminck and Schlegel), in Japan is caused by 471 Neoparamoeba perurans. J. Fish Dis. 33, 455–458. 472 Crosbie, P.B.B., Bridle, A.R., Cadoret, K., Nowak, B.F., 2012. In vitro cultured 473 Neoparamoeba perurans causes amoebic gill disease in Atlantic salmon and fulfils Koch’s 474 postulates. Int. J. Parasitol. 42, 511–515. 475 Dias, P.J., Sollelis, L., Cook, E.J., Piertney, S.B., Davies, I.M., Snow, M., 2008. 476 Development of a real-time PCR assay for detection of Mytilus species specific alleles: 477 Application to a sampling survey in Scotland. J. Exp. Mar. Biol. Ecol. 367, 253–258. 478 Douglas-Helders, G.M., Dawson, D.R., Carson, J., Nowak, B.F., 2002. Wild fish are not a 479 significant reservoir for Neoparamoeba pemaquidensis (Page, 1987). J. Fish Dis. 25, 569– 480 574. 481 Feehan, C.J., Johnson-Mackinnon, J., Scheilbling, R.E., Lauzon-Guay, J.-S., Simpson, 482 A.G.B., 2013. Validating the identity of Paramoeba invadens, the causative agent of 483 recurrent mass mortality of sea urchins in Nova Scotia, Canada. Dis. Aquat. Org. 103, 209– 484 227. 485 Findlay, V.L. & Munday, B.L., 1998. Further studies acquired resistance to amoebic gill 486 disease (AGD) in Atlantic salmon, Salmo salar L. J. Fish Dis. 23, 193–198. 487 Fringuelli, E., Gordon, A.W., Rodger, H., Welsh, M.D., Graham, D.A., 2012. Detection of 488 Neoparamoeba perurans by duplex quantitative Taqman real-time PCR in formalin-fixed, 489 paraffin-embedded Atlantic salmonid gill tissues. J. Fish Dis. 35, 711–724. 490 Handå, A., Min, H., Wang, X., Broch, O.J., Reitan, K.J., Reinertsen, H., Olsen, Y., 2012. 491 Incorporation of fish feed and growth of blue mussels (Mytilus edulis) in close proximity to 492 salmon (Salmo salar) aquaculture: Implications for integrated multi-trophic aquaculture in 493 Norwegian coastal waters. Aquaculture 356–357, 328–341. 494 Karlsbakk, E., Olsen, A.E., Einen, A.-C.B., Mo, T.A., Fiksdal, I.U., Aase, H., Kalgraff, C., 495 Skår, S.-J., Hansen, H., 2013. Amoebic gill disease due to Paramoeba perurans in ballan 496 wrasse (Labrus bergylta). Aquaculture 412–413, 41–44. 497 Marine Scotland Science, 2012. Amoebic gill disease. Topic sheet no. 96 V1. 498 http://www.scotland.gov.uk/Resource/ 0039/00393037.pdf on the 15/7/2015. 499 McConnachie, S.H., Guselle, N.J., Speare, D.J., 2013. Retention of viable microsporidial 500 (Loma salmonae) spores within the blue mussel (Mytilus edulis): Use of an experimental 501 laboratory model probing pathogen transfer within a multi-trophic aquaculture setting. 502 Aquaculture 376–379, 1–5. 503 Molloy, S.D., Pietrak, M.R., Bouchard, D.A., Bricknell, I., 2011. Ingestion of Lepeophtheirus 504 salmonis by the blue mussel Mytilus edulis. Aquaculture 311, 61–64. 505 Molloy, S.D., Pietrak, M.R., Bouchard, D.A., Bricknell, I., 2012. The interaction of infectious 506 salmon anaemia virus (ISAV) with the blue mussel, Mytilus edulis. Aquacult. Res. 1–10. 507 Molloy, S.D., Pietrak, M.R., Bricknell, I., Bouchard, D.A., 2013. Experimental transmission 508 of infectious pancreatic necrosis virus (IPNV) from the blue mussel, Mytilus edulis, to 509 cohabiting Atlantic salmon (Salmo salar) smolts. Appl. Environ. Microbiol. 79, 5882–5890. 510 Mortensen, S.H., Bachere, E., Le Gall, G., Mialhe, E., 1992. Persistence of infectious 511 pancreatic necrosis virus (IPNV) in scallops Pecten maximus. Dis. Aquat. Org. 12, 221–227. 512 Mouton, A., Crosbie, P., Cadoret, K., Nowak, B., 2014. First record of amoebic gill disease 513 caused by Neoparamoeba perurans in South Africa. J. Fish Dis. 37, 407–409. 514 Mullen, T.E., Nevis, K.R., O'Kelly, C., Gast, R.J., Frasca, Jr. S., 2005. Nuclear small-subunit 515 ribosomal RNA gene-based characterization, molecular phylogeny and PCR detection of the 516 Neoparamoeba from Western Long Island Sound lobster. J. Shellfish Res. 24, 719–731. 517 Munday, B.L., Zilberg, D., Findlay, V., 2001. Gill disease of marine fish caused by infection 518 with Neoparamoeba pemaquidensis. J. Fish Dis. 24, 497–507. 519 Nowak, B., 2007. Parasitic diseases in marine cage culture – An example of experimental 520 evolution of parasites? Int. J. Parasitol. 37, 581–588. 521 Ørpetveit, I., Mikalsen, A.B., Sindre, H., Evensen, Ø., Dannevig, B.H., Midtlyng, P.J., 2010. 522 Detection of Infectious Pancreatic Necrosis Virus in subclinically infected Atlantic Salmon 523 by virus isolation in cell culture or Real-Time reverse transcription polymerase chain 524 reaction: Influence of sample preservation and storage. J. Vet. Diagn. Invest. 22, 886–895. 525 Palmer, R., Carson, J., Ruttledge, M., Drinan, E., Wagner, T., 1997. Gill disease associated 526 with Paramoeba, in sea reared Atlantic salmon in Ireland. Bull. Eur. Assn. Fish P. 17, 112– 527 114. 528 Pietrak, M.R., Molloy, S.D., Bouchard, D.A., Singer, J.T., Bricknell, I., 2012. Potential role 529 of Mytilus edulis in modulating the infectious pressure of Vibrio anguillarum 02β on an 530 integrated multi-trophic aquaculture farm. Aquaculture 326-329, 36–39. 531 Powell, M.D., Parsons, H.J., Nowak, B.F., 2001. Physiological effects of freshwater bathing 532 of Atlantic salmon (Salmo salar) as a treatment for amoebic gill disease. Aquaculture 199, 533 259–266. 534 R Core Team, 2013. R: A language and environment for statistical computing. R Foundation 535 for Statistical Computing, Vienna, Austria. ISBN 3-900051-07-0, URL http://www.R- 536 project.org/. 537 Rodger, H.D. & McArdle, J.F., 1996. An outbreak of amoebic gill disease in Ireland. Vet. 538 Rec. 139, 348–349. 539 Schrader, C., Schielke, A., Ellerbroek, L., Johne, R., 2012. PCR inhibitors – occurrence, 540 properties and removal. J. Appl. Microbiol. 113, 1014–1026. 541 Shumway, S.E., Cucci, T.L., Newell, R.C., Yentsch, C.M., 1985. Particle selection, ingestion, 542 and absorption in filter-feeding bivalves. J. Exp. Mar. Biol. Ecol. 91, 77–92. 543 Shinn, A.P., Pratoomyot, J., Bron, J.E., Paladini, G., Brooker, E.E., Brooker, A.J., 2015. 544 Economic costs of protistan and metazoan parasites to global mariculture. Parasitology 142, 545 196–270. 546 Skår, C.K. & Mortensen, S., 2007. Fate of infectious salmon anaemia virus (ISAV) in 547 experimentally challenged blue mussels Mytilus edulis. Dis. Aquat. Org. 74, 1–6. 548 Steinum, T., Kvellestad, A., Rønneberg, L.B., Nilsen, H., Asheim, A., Fjell, K., Nygård, 549 S.M.R., Olsen, A.B., Dale, O.B., 2008. First cases of amoebic gill disease (AGD) in 550 Norwegian seawater farmed Atlantic salmon, Salmo salar L., and phylogeny of the causative 551 amoeba using 18S cDNA sequences. J. Fish Dis. 31, 205–214. 552 Stirling, H.P. & Okumus, I., 1995. Growth and production of mussels (Mytilus edulis L.) 553 suspended at salmon marine cages and mussel farms in two sea lochs on the West Coast of 554 Scotland. Aquaculture 134, 193–210. 555 Sühnel, S., Ivachuk Cda, S., Schaefer, A.L., Pontinha, V.A., Martins, M.L., Figueras, A., 556 Meyer, G.R., Jones, S.R., Stewart, J.C., Gurney-Smith, H.J., Magalhães, A.R., Bower, S.M., 557 2014. Detection of the parasitic amoeba (Order Dactylopodida) in the female gonads of 558 oysters in Brazil. Dis. Aquat. Org. 109, 241–250. 559 Young, N.D., Dyková, I., Nowak, B.F., Morrison, R.N., 2008. Development of a diagnostic 560 PCR to detect Neoparamoeba perurans, agent of amoebic gill disease. J. Fish Dis. 31, 285– 561 295. 562 Wang, X., Olsen, L.M., Reitan, K.I., Olsen, Y., 2012. Discharge of nutrient wastes from 563 salmon farms: Environmental effects, and potential for integrated multi-trophic aquaculture. 564 Aquacult. Environ. Interac. 2, 267–283. 565 Webb, J.L., Vandenbor, J., Pirie, B., Robinson, S.M.C., Cross, S.F., Jones, S.R.M., Pearce, 566 C.M., 2013. Effects of temperature, diet, and bivalve size on the ingestion of sea lice 567 (Lepeophtheirus salmonis) larvae by various filter-feeding shellfish. Aquaculture 406–407, 568 9–17. 569 Whitmarsh, D.J., Cook, E.J., Black, K.D., 2006. Searching for sustainability in aquaculture: 570 An investigation into the economic prospects for an integrated salmon-mussel production 571 system. Mar. Policy 30, 293–298. 572 Wickham, H., 2009. ggplot2: elegant graphics for data analysis. Springer New York. 573 Wright, D.W., Nowak, B., Oppedal, F., Bridle, A., Dempster, T., 2015. Depth distribution of 574 the amoebic gill disease agent, Neoparamoeba perurans, in salmon cages. Aquacult. Environ. 575 Interact. 7, 67–74.