Introduction
Perkinsus marinus is a protistan parasite that infects the eastern oyster (Crassostrea virginica) and is notorious for causing dermo disease (Mackin, Reference Mackin1951; Ray, Reference Ray1954). It was first identified in the 1940s after causing severe mass mortalities of wild oyster populations in the Gulf of Mexico (Ray, Reference Ray1954). Its current geographic range spans the east coast of the Americas, from northern South America (da Silva et al., Reference da Silva, Vianna, Guertler, Ferreira, Santana, Fernandez-Boo, Ramilo, Cao and Villalba2013), to the Gulf of Maine (Cook et al., Reference Cook, Folli, Klinck, Ford and Miller1998), including Central America (Pagenkopp Lohan et al., Reference Pagenkopp Lohan, Hill-Spanik, Torchin, Aquirre-Macedo, Fleischer and Ruiz2016), and the Gulf of Mexico (Ray, Reference Ray1954). In the wild, infective P. marinus cells are taken into the oyster through filter feeding, with many infections occurring through the mantle from infected pseudofeces (particles that accumulate near the mantle after being expelled by the oyster without digestion) (Allam et al., Reference Allam, Carden, Ward, Ralph, Winnicki and Pales Espinosa2013; Ben-Horin et al., Reference Ben-Horin, Bidegain, Huey, Narvaez and Bushek2015). While P. marinus epizootics have historically resulted in the collapse of entire oyster reefs (Hewatt and Andrews, Reference Hewatt and Andrews1954), enzootic populations of the parasite are currently common (Calvo et al., Reference Calvo, Dungan, Roberson and Burreson2003). Low-level infections can be sustained for years with only slight effects on growth and reproduction (Ray, Reference Ray1954; Sunila and LaBanca, Reference Sunila and LaBanca2003), with parasites building up over time in older oysters (Andrews and Hewatt, Reference Andrews and Hewatt1957).
On oyster reefs, P. marinus prevalence, intensity and resulting oyster mortality increase with temperature and salinity (Ray, Reference Ray1954; Andrews, Reference Andrews1988). Epizootics are therefore seasonally cyclical, with the highest prevalence and intensity in summer and fall months (Andrews and Hewatt, Reference Andrews and Hewatt1957) and lowest in winter months (Andrews, Reference Andrews1965). The severity of outbreaks also varies across years, with particularly intense outbreaks after La Niña events, when both temperature and salinity increase (Soniat et al., Reference Soniat, Klinck, Powell and Hofmann2005). Geographic variation in temperature and salinity also drives P. marinus distribution, with warming sea temperatures likely contributing to the northern range expansion of P. marinus (Ford, Reference Ford1996; Cook et al., Reference Cook, Folli, Klinck, Ford and Miller1998; Ford and Chintala, Reference Ford and Chintala2006). In the laboratory, Chu and coworkers found that both higher temperature (Chu and La Peyre, Reference Chu and La Peyre1993) and salinity (Chu et al., Reference Chu, La Peyre and Burreson1993) increased infection prevalence and intensity in oysters challenged with P. marinus. In the same experiments, temperature and salinity also altered the presumed cellular defences of the hosts, such as the abundance of haemocytes, suggesting a potential impact on host immune processes. In culture, high temperature and salinity influenced two infectious stages of P. marinus: increasing development of P. marinus hypnospores (Chu and Greene, Reference Chu and Greene1989b) and metabolic activity and proliferation of trophozoites (La Peyre et al., Reference La Peyre, Casas, Villalba and La Peyre2008), with salinities below 7 ppt causing severe reductions in viability (La Peyre et al., Reference La Peyre, Casas and La Peyre2006).
In the wild, P. marinus is believed to be primarily transmitted between oysters through the water column (Ray, Reference Ray1954). The parasite is able to survive for weeks outside the host due to its ability to synthesize fatty acids, facilitating both short and long distance (i.e. reef-to-reef) transmission (Chu et al., Reference Chu, Lund, Soudant and Harvey2002). Transmission can also occur through ectoparasitic snails feeding on infected oysters (White et al., Reference White, Powell, Ray and Wilson1987). Live P. marinus cells have been found in the stomachs of multiple species of snails and fish, and on the setae of mud crabs (Andrews et al., Reference Andrews, Hoese and Wood1962), although the relative frequency of transmission by animal vector vs through the water column is unknown. Even with continuous exposure, detectable infections develop slowly, likely because a high infective dose is required (Bidegain et al., Reference Bidegain, Powell, Klinck, Ben-Horin and Hofmann2016). While timing can vary, naive oysters placed in water where the parasite is endemic or adjacent to heavily infected oysters will generally develop visible infections over the course of weeks (Ray, Reference Ray1954; Andrews, Reference Andrews1965). For this reason, understanding the factors that contribute to the abundance of infective particles in the water column is crucial to understanding disease dynamics, including rates of infection.
The majority of infective P. marinus in the water column is historically believed to come primarily from the death of heavily infected oysters, as the abundance of free-living parasites is highly correlated with oyster mortality (Calvo et al., Reference Calvo, Dungan, Roberson and Burreson2003; Audemard et al., Reference Audemard, Ragone Calvo, Paynter, Reece and Burreson2006). Large numbers of cells are released from dead and dying oysters, especially when tissues are shredded by scavengers, providing a logical mechanism for these observed patterns (Hoese, Reference Hoese1962). However, death of heavily infected oysters is not the only avenue for the reintroduction of P. marinus cells to the water column, and live cells can be released from infected oysters in pseudofeces (Bushek et al., Reference Bushek, Ford and Chintala2002). In the wild, parasite abundance is not only correlated with mortality but also with weighted prevalence in live oysters, suggesting that heavily infected live oysters may also contribute significantly to P. marinus abundance in the water (Audemard et al., Reference Audemard, Ragone Calvo, Paynter, Reece and Burreson2006). In addition, naïve oysters can pick up infections from the water even when oyster mortality is low (Calvo et al., Reference Calvo, Dungan, Roberson and Burreson2003). Because infection and development of disease signs happen over the course of weeks and months, studies of P. marinus abundance in the water column have focused on these time scales (Calvo et al., Reference Calvo, Dungan, Roberson and Burreson2003; Audemard et al., Reference Audemard, Ragone Calvo, Paynter, Reece and Burreson2006). However, on shallow tidal oyster reefs, environmental parameters such as temperature and salinity often fluctuate over much shorter timescales. Little is known about how short-term changes in these parameters affect the release of P. marinus cells from live oysters. Here, we conducted two experiments to examine the effects of time of day, salinity and temperature on the release of P. marinus from infected oysters. We found that both time of day and temperature strongly influenced the number of cells released. Further, we found that while generally more cells were released from dead or dying oysters than from live oysters, similar numbers of cells can be released from live oysters at high water temperatures. We surmise that live oysters contribute significantly to the population of P. marinus in the water column and should be considered in transmission dynamics.
Materials and methods
Experimental oysters
In order to measure the release of P. marinus by infected oysters, we obtained 200 wild oysters (C. virginica) from the Nanticoke River, Maryland in July 2019. Mean fall salinity for this location is 9.98 ppt (s.d. 3.47) and temperature is 23.5°C (s.d. 1.75). We measure the intensity of Perkinsus infections using the Mackin scale (Ray, Reference Ray1954), which rates infection from 1 (light infection) to 5 (very heavy infection). This location was chosen due to previous findings of high prevalence, oysters from this river had a 73% prevalence in 2018, with one of the highest recorded mean intensities in the Maryland portion of the Chesapeake Bay (2.3 on the Mackin scale) (Tarnowski, Reference Tarnowski2018). Oysters had a mean two-dimensional area of 76.46 cm (s.d. 20.82) and mean wet weight (including shell) of 237.21 g (s.d. 92.46). Prior to the start of the experiments, we sacrificed 55 oysters and measured P. marinus infection using Ray's Fluid Thioglycollate Method (RFTM) as described below. We found a prevalence of 43% and a mean intensity of 0.75 (s.d. 1.16). The remaining oysters were randomly divided into three groups kept in Rhode River water at different salinities: low (5–10 ppt), medium (10–15 ppt) and high (20–25 ppt) manipulated by adding freshwater or aquarium salt (Instant Ocean, Blacksburg, VA, USA) as needed (Fig. S1A). These levels were chosen to mimic the extremes of the salinity range found in the Maryland portion of the Chesapeake Bay. All oysters were fed Frozen Shellfish Diet for Broodstock (Reed Mariculture, Campbell, CA, USA) until the start of Experiment #1. The authors assert that all procedures contributing to this work comply with the ethical standards of the relevant national and institutional guides on the care and use of laboratory animals.
Experiment #1: effect of salinity and time of day on P. marinus release over 24 h
In order to test the effect of salinity on the release of P. marinus cells over a short period of time, we took oysters (N = 36) from the medium salinity holding treatment and placed them in Rhode River water adjusted to three salinity levels (n = 12), low (5 ppt), medium (15 ppt) and high (25 ppt) as above (Fig. S1). Water was filtered through a 1 μm mesh filter to ensure that no P. marinus cells were introduced via the water. Oysters were acclimated to their experimental salinity levels over the course of 2 days. At the start of the experiment (24:00 on 14 August 2019), oysters were placed in individual 1 L containers with water and an air stone. To ensure continuous filtration by the oysters, the pre-filtered water was supplemented with a small amount of dilute shellfish diet (Ehrich and Harris, Reference Ehrich and Harris2015). The mesocosms were placed outside of the wetlab at the Smithsonian Environmental Research Center (SERC) in the shade in order to mimic the change in water temperature over a 24 h period. To monitor the temperature throughout the experiment, HOBO sensors were randomly placed in three containers and began temperature measurements at 0:00. The water was sampled as described below and replaced with new water every 6 h, resulting in four sampling points (Fig. S1A). All oysters were sacrificed for RFTM at the end of the experiment.
Experiment #2: effect of temperature on P. marinus release
In order to test the long-term effects of salinity on P. marinus release, the surviving oysters in the high (n = 35) and low (n = 36) salinity holding treatments were kept for 2 months before the start of Experiment #2 in 1135 L tanks where water was changed and salinity was adjusted every 3 days. Oysters were labelled with bee tags (Better Bee Greenwich, NY, USA) in order to identify the individuals. To measure short-term temperature impacts on Perkinsus release, each oyster was placed in an individual 1 L container with filtered water for a 6 h period on four separate days (hereafter referred to as trials). Trials were conducted in water at either high or low salinities consistent with the long-term salinity treatments. Individual containers were placed in a communal water bath to ensure consistent temperatures across replicates. Oysters were kept in communal holding tanks inside at appropriate salinity levels and ambient temperature between trials with timed artificial lights. All trials ran from 9:00 to 15:00 inside the wetlab at SERC under artificial light (Fig. S1B). At the conclusion of each 6 h trial, the oyster was removed and the entire contents of the container were sampled, including water and any particles, feces and pseudofeces. All oysters experienced a total of four trials: two trials where the water temperature was held at ambient temperatures (23–25°C) followed by two trials where the water bath was heated (30–32°C) (Fig. S1B). We chose to measure the same oysters multiple times in order to minimize the effects of individual variability in release. We conducted ambient trials prior to heated trials in order to minimize any carry-over effects of heating. While oysters are constantly releasing P. marinus cells, the parasite is also reproducing inside the oyster, and so we would not expect to see any reduction in release over time. Due to space constraints, oysters were divided into two groups and trials were run on separate days. For heated treatments, temperature was slowly ramped up and down over the course of 24 h to not shock the oysters. For a given group, trials were always separated by a minimum of 3 days.
Sampling and analysis
In order to count the number of pathogens released into the water column in both experiments, the 1 L container was shaken and 500 mL was removed and filtered through a 0.45 μm Whatman filter (Cytiva Life Sciences, Marlborough, MA, USA) using a glass filter apparatus that was rinsed with bleach in between samples. To measure oyster infection at the end of each experiment, oysters were sacrificed and a piece of anal and mantle tissue was placed in RFTM media. We incubated the filters and tissue in Ray's Fluid Thioglycollate Media (RFTM) at 27°C for 5–7 days. RFTM media causes viable Perkinsus sp. cells to enlarge making them visible under a dissecting microscope (Dungan and Bushek, Reference Dungan and Bushek2015). We placed tissue or filters on a microscope slide and stained with 2–5 drops of Lugol's iodine (Sigma Aldrich, St. Louis, MO, USA). Staining with Lugol's solution allows visualization of cells under a dissecting microscope. For tissue samples, we utilized the Mackin Scale, which quantifies infection intensity on a scale from 1 to 5, with 0 indicating no infection (Ray, Reference Ray1954). For filters where total number of cells was <~500, we counted all cells. For the 12 filters with a high abundance of cells (>500), we took eight pictures of random locations on the filter at 4× magnification and counted the number of cells using IMAGEJ (Schindelin et al., Reference Schindelin, Arganda-Carreras, Frise, Kaynig, Longair, Pietzsch, Preibisch, Rueden, Saalfeld and Schmid2012). At 4× magnification, the filter contains 72 frames, so we multiplied these numbers by 9 to estimate the total number of cells across the entire filter.
Statistics
We looked at the effect of salinity in Experiments #1 and #2 on the number of cells released (log transformed) using analyses of covariance (ANCOVAs) with Mackin score of tissue as a covariate. We conducted a logistic regression with Mackin score as an ordinal response using the polr function in the R package MASS (Ripley et al., Reference Ripley, Venables, Bates, Hornik, Gebhardt, Firth and Ripley2013) to measure the effect of salinity on the intensity of oyster infection for Experiment #2. To confirm that this result was unlikely to be due to uneven distribution of infected oysters into the two treatments, we ran 10 000 simulations randomly distributing 71 oysters into two buckets with a disease rate that follows a binomial distribution with P = 0.43. We used separate GLMMs with a Poisson distribution and ‘replicate’ as a random effect in the R package lme4 (Bates et al., Reference Bates, Maechler, Bolker, Walker, Christensen, Singmann, Dai, Scheipl, Grothendieck and Green2018) to look at the effect of time of day, temperature and sunlight on the number of cells released in Experiment #1. To look at the effect of temperature in Experiment #2 for a given oyster, we first averaged the number of cells observed in the two trials (for each oyster, two ambient and two heated). We then subtracted the ambient average from the heated average. In order to transform the data, we added 105 to every value to make all values positive and conducted a Box–Cox transformation in the r package MASS using a λ of 0.033. We then conducted a one-sample t-test with a null hypothesis of no difference. Finally, to compare the number of cells released by survivors (n = 19) and non-survivors (n = 52) in Experiment #2, we first took the maximum of all available values then conducted a glm with a Gaussian distribution, with ‘treatment’ and ‘survival’ as factors. We conducted a Tukey HSD test using the R package multcomp. All graphs were made using ggplot2 (Wickham, Reference Wickham2009).
Results
Salinity
In Experiment #1, where oysters were exposed to three salinity levels for 3 days, salinity had no effect on the number of P. marinus cells released (ANCOVA d.f. = 6,130, P = 0.17). In Experiment #2, where oysters were held at two levels of salinity for 4 months, salinity also had no effect on the number of cells released (ANCOVA d.f. = 2,45, P = 0.08). Salinity did, however, have a significant effect on the final intensity of P. marinus infections within oyster tissues – while all oysters were infected (100% prevalence), oysters in the high salinity treatment (20–25 ppt) had significantly higher intensity infections at the end of this period (Fig. 1, logistic regression P = 7.47 × 10−06). The probability that the starting prevalence would differ between the two treatments by more than 20%, and therefore drive the final results was 1%.
Salinity treatments affected the final intensity of infection of oyster tissues. Histogram showing the number (frequency) of oysters from Experiment #2 with each Mackin score broken up by salinity treatment (across temperature treatments). We combined 4 and 5 due to the few number of 5s observed. Low salinity water was maintained between 5 and 10 ppt and high salinity water was maintained between 20 and 25 ppt.
Diurnal cycle
Oysters used in Experiment #1, where water was sampled four times over a 24 h period, had a prevalence of 39% (14/36) and a mean intensity of 3.18. Perkinsus marinus cells were released from nine oysters during at least one of the four sampled time points, with six oysters releasing cells at two timepoints. Cells were only released from infected oysters (i.e. those with a Mackin score >0). We did not find a significant relationship between oyster size and total number of cells released (glm, P = 0.498). Because there was no effect of salinity on the number of cells released, salinity was excluded from further analyses. Across all individuals, we observed a diurnal cycle of P. marinus release, with the greatest number of cells released between 12:00 and 18:00. For seven out of nine oysters, the majority of parasite cells were released during this time period (Fig. 2). The observed effect of time of day on release was correlated to both measured temperature (glmm, P < 2 × 10−16) and inferred sunlight based on the angle of the sun (glmm, P < 2 × 10−16), with both variables peaking at 18:00 (Fig. 2, Fig. S2).
Time of day drove P. marinus release. Nine of the oysters used in Experiment #1 released cells. Number of cells found during a timepoint is coloured by oyster replicate and points are connected by lines. The time when the sample was taken is shown, meaning that cells were released in the 6 h period beforehand. Temperature (°C) as measured by a hobo datalogger is displayed on the right y-axis and denoted by triangles.
Temperature
In Experiment #2, we compared the number of parasite cells released when oysters were in ambient (23–25°C) vs heated (30–32°C) water temperatures. For 76% (36 out of 47) of the oysters that survived to the end of the experiment, more P. marinus cells were released in heated water than in ambient water (Fig. 3, one-sample t-test, d.f. = 46, P = 0.036). Neither salinity nor Mackin score of the oyster influenced this change in release [ANOVA, d.f. = 40, P = 0.418 (Mackin score), P = 0.99 (salinity)].
Most oysters released more cells in the heated than ambient treatments. Difference between mean heat and mean ambient cells released for each individual oyster in Experiment #2 shown by bar. Difference was calculated as mean heated cells released − mean ambient cells released for each oyster. Plotted values are the log10 of the absolute value of that difference then adjusted by sign depending on non-adjusted direction of change (i.e. negative differences remain negative). Oysters with no difference are displayed as 0.
Survival
There were no mortalities in Experiment #1. In Experiment #2, 27% (19 out of 71) of oysters died before the conclusion of the experiment and were excluded from the analyses presented above. Salinity did not impact mortality (11 non-survivors were in high salinity and eight were in low salinity). Fourteen oysters died in the first 20 days of the experiment before being exposed to elevated temperatures and two oysters died after being exposed to one heated trial. We were only able to collect tissues from five of the oysters that died, four of which had a Mackin score ⩾4. In order to compare cells released from survivors and non-survivors, we calculated the maximum number of cells for each oyster and compared across treatments in order to eliminate issues with the number of datapoints (e.g. oysters that died after one trial compared to oysters that survived all four). For 76% (39 out of 51) of survivors, the maximum number of cells were released during a heated trial. The highest maximum observed was by a non-survivor in a heated trial (70 173 cells, Fig. 4). Overall, non-survivors in heated treatments had significantly higher maxima than all survivors as well as non-survivors in ambient treatments (Fig. 4, glm, interaction P < 0.001, Tukey HSD). However, the highest maximum for survivors (14 247 cells released in a heated treatment) was higher than the maxima of 14 non-survivors.
Heated non-survivors released the highest maximum number of cells. Boxplots display mean maxima for a given treatment with outliers indicated by dots. Letters (A and B) are results of Tukey HSD post-hoc test.
Discussion
Release of P. marinus cells from infected oysters determines the abundance of infectious cells in the water column and the likelihood of new infections. Variation in the rate of release can therefore drive overall disease dynamics, particularly within reefs. We build upon previous work showing that viable infective P. marinus cells can be released from live infected oysters in feces and pseudofeces (Bushek et al., Reference Bushek, Ford and Chintala2002). By sampling all water surrounding the oyster including (but not limited to) any feces and pseudofeces released during the experiment, we were able to more reliably estimate the number of cells released by a given oyster over time. We examined the influence of environmental factors on the rate of release and directly compared the number of P. marinus cells released from live to dead and dying oysters. We find that time of day and temperature, but not salinity, have strong effects on the number of cells released in a 6 h period (Figs 2 and 3). Because oysters live in estuaries, where temperature and salinity change on daily to yearly cycles, it is important to understand how these factors influence the parasite abundance in the water column on different time scales and the consequences for transmission and disease dynamics in host populations. Our results suggest that temperature may be an important driver of parasite abundance in the water column at short time scales, in addition to those previously reported for seasonal and annual scales. More broadly, our data demonstrate that large numbers of parasite cells were released from both live and dead oysters, suggesting that live oysters should not be discounted as significant contributors to P. marinus abundance in the water column (Fig. 4).
Salinity has consistently been identified as an important driver of P. marinus distribution and dynamics (Mackin, Reference Mackin1951; Ray, Reference Ray1954; Soniat et al., Reference Soniat, Klinck, Powell and Hofmann2005). In the field, naive oysters placed in regions with high salinity developed higher intensity infections than those placed in regions with low salinity (Paynter and Burreson, Reference Paynter and Burreson1991), either due to higher abundances of P. marinus in the water column leading to increased infection pressure, higher pathogenicity and/or a higher reproduction rate within the oyster. When oysters are inoculated with a standard number of P. marinus cells and placed in a closed system, higher salinity increases the intensity of infections over the course of several weeks, suggesting that the main effect of salinity is on P. marinus reproduction within the oyster host (Fisher et al., Reference Fisher, Gauthier and Winstead1992; Chu et al., Reference Chu, La Peyre and Burreson1993; Ragone Calvo and Burreson, Reference Ragone Calvo and Burreson1993). Our result that oysters in the high salinity treatment had higher intensity infections after 4 months corroborates this finding – while we did not use a fully closed system, water in the holding tank was changed infrequently (biweekly) and we therefore assume that the vast majority (if not all) of new infections were acquired from neighbouring oysters rather than the incoming water. Together, these findings suggest that the main effects of salinity on P. marinus–C. virginica interactions are on the increased abundance of the parasite inside the oyster rather than on increased infectivity or behaviour in the water column. Whether this effect is due to a reduction of oyster defence/immunity or an increase in parasite reproduction rate due to higher parasite fitness should be further investigated. Both of our experiments demonstrated that when infection intensity is accounted for, salinity does not impact the rate at which P. marinus cells are released over the course of several hours. Therefore, increased abundance of P. marinus in high salinity water is likely due to an increase in the intensity of infections resulting in higher numbers of P. marinus cells released into the water column. While in vitro low salinity (7 ppt) has been shown to reduce the viability of P. marinus (La Peyre et al., Reference La Peyre, Casas and La Peyre2006), we did not see an impact of the low salinity treatment on the number of P. marinus cells released in either experiment. This could be because our low salinity treatment was not low enough to reduce viability, or because those cells released in the low salinity treatment would not have been viable if they remained at that salinity for a longer period of time rather than being put into RFTM media after 6 h.
Our observed diurnal cycle of P. marinus release was correlated with both water temperature and solar exposure, factors that could influence both host and parasite processes. Oysters have been shown to display both daily and seasonal cycles of feeding activity, but this is the first study to look for daily cycles of P. marinus release into the water column. Work on Crassostrea gigas has confirmed the presence of an endogenous circadian rhythm that can be influenced by season (Mat et al., Reference Mat, Massabuau, Ciret and Tran2012), light and tidal cycle (Tran et al., Reference Tran, Nadau, Durrieu, Ciret, Parisot and Massabuau2011). This circadian rhythm could be driving oyster feeding and defecation and therefore the release of P. marinus cells observed here. Temperature is also a strong driver of oyster behaviour and interacts with these other influences on circadian rhythm. Early experiments by Galtsoff on C. virginica from the Chesapeake Bay found the rate of water intake is dependent on temperature, with oysters remaining closed below temperatures of 5°C (Galtsoff, Reference Galtsoff1928). These results have been replicated by Comeau et al. (Reference Comeau, Mayrand and Mallet2012) who found that C. virginica from the Gulf of St. Lawrence increase the number of valve openings (and therefore water filtration) as water temperatures warm in the spring. Interestingly, however, once oysters entered an ‘active’ period in the spring, temperature no longer exerted an influence on opening, with oysters staying open for periods of time ranging from 1 min to 13.8 days. During the active period, oysters did exhibit a diurnal cycle independent of temperature, with both length of time and degree of valve openness peaking at 18:00 h (Comeau et al., Reference Comeau, Mayrand and Mallet2012). Time of day alone could therefore explain our finding that maximum release of Perkinsus cells occurred between 12:00 and 18:00 h irrespective of temperature. However, our oysters and P. marinus also experienced a much greater fluctuation in temperatures over a 24 h period than the oysters studied by Comeau et al., and these more drastic changes in temperature likely influenced P. marinus release in addition to time of day.
Experiment #2 allowed us to test the effects of temperature on P. marinus release regardless of time of day, demonstrating that temperature alone increases the rate of P. marinus release. Temperature is known to be an important driver of P. marinus reproduction and infectivity both in vitro (Chu and Greene, Reference Chu and Greene1989a; Chu and Volety, Reference Chu and Volety1997) and in vivo (Fisher et al., Reference Fisher, Gauthier and Winstead1992; Calvo et al., Reference Calvo, Dungan, Roberson and Burreson2003), and could therefore alter pathogen behaviour or reproduction leading to increases in release. By counting the number of P. marinus cells released from a given oyster in both ambient and heated environments, we controlled for the significant individual variation observed in Experiment #1 and the strong influence of infection intensity on the number of cells released. This experimental design allowed us to get reliable estimates of the increase in the number of cells released due to increased temperature. The water temperature fluctuations in Experiment #1 were consistent with temperatures experienced by intertidal oysters in shallow embayments of the Chesapeake Bay during summer months (Maryland Department of Natural Resources). In the wild, both mean temperature and temperature variability will be highly dependent on location within the estuarine environment and relation to tidal height (i.e. subtidal vs intertidal and amount of time spent out of water). Previous work has found that intertidal oysters have higher prevalence and intensity of Perkinsus infections than subtidal oysters (Malek and Byers, Reference Malek and Byers2017), with a positive correlation between intensity and air temperature (Malek and Byers, Reference Malek and Byers2018), indicating that the higher temperatures experienced when out of the water contribute to intensity. Some of the spatial heterogeneity of Perkinsus abundance in estuaries is likely driven by a combination of spatial and temporal variability in temperature (Malek and Breitburg, Reference Malek and Breitburg2016). In order to accurately capture the maximum P. marinus abundance in the water column at a given location, water sampling should occur in the afternoon close to the hottest part of the day.
Previous research demonstrated that infective P. marinus cells are released through feces and pseudofeces of live oysters (Bushek et al., Reference Bushek, Ford and Chintala2002). However, this release was assumed to be negligible when compared to release by the scavenging and disintegration of infected dead oyster tissue (Calvo et al., Reference Calvo, Dungan, Roberson and Burreson2003; Audemard et al., Reference Audemard, Ragone Calvo, Paynter, Reece and Burreson2006). We found that while more cells are generally released from dead and dying oysters than live oysters, live oysters are capable of releasing tens of thousands of cells over the course of 6 h and should therefore be considered a major contributor to the number of parasites in the water column. Given our finding that live oysters release more cells at higher temperatures, the relative contributions that live and dead oysters make to parasites in the water column may be heavily temperature-dependent. As climate change raises water temperatures, we would expect the number of cells released from live oysters to increase, further threatening oyster populations. In addition, this effect could fundamentally alter large-scale disease processes if in the future the majority of cells were released from live oysters. More work is needed to both understand baseline dynamics and predict future changes. In this experiment, the cause of death was not always obvious as many oysters died before tissue could be collected for RFTM. However, four out of the five non-survivors that were assayed had heavy infections suggesting that the parasite played a role in mortality. While release from dead and dying oysters may result in a spike of P. marinus cells in the water, live oysters have the potential to contribute to a steadier release of cells for a longer period of time. Seven of the survivors released cells at all four timepoints assayed. Furthermore, infective cells were consistently released from live oysters with low intensity. Once a dead oyster has disintegrated or been eaten, it no longer acts as a source for the water-borne parasite population, whereas live oysters may contribute to this population to varying degrees for an entire season. Over the course of several years, live oysters harbouring low-level infections could result in the release of equivalent or larger amounts of P. marinus cells than oysters that quickly develop a high-intensity infection and die in one season. Live infected oysters should therefore be considered as a significant contributor to enzootic P. marinus populations in the water column.
While temperature and salinity are often cited as the two main environmental factors that influence P. marinus abundance and infectivity, our results show that they operate on different components of the P. marinus–C. virginica interaction and at different timescales. Further work is needed to understand when and how temperature and salinity influence the host and parasite individually resulting in the patterns observed here and by others. Controlled and fully crossed experiments are needed to tease out the interactive effects between light, time of day and temperature, which likely influence both the host and parasite in ways not fully captured by our experiments. In addition, future experiments should (1) keep temperature more tightly controlled, (2) examine a wider range of temperatures and (3) alter the length of the experiments to test the effect of duration of temperature changes on the release of parasite cells.
Seven pathogenic species of Perkinsus have been identified to date, infecting a wide variety of hosts (Villalba et al., Reference Villalba, Reece, Camino Ordás, Casas and Figueras2004). While P. marinus is the dominant species in Chesapeake Bay populations of C. virginica, P. chesapeaki has also been found in these oysters, albeit at much lower prevalence and intensity (Reece et al., Reference Reece, Dungan and Burreson2008). We assumed that the Perkinsus cells detected here using RFTM were P. marinus, based on previous work showing this is the dominant parasite in oysters of this region (Gignoux-Wolfsohn unpublished data; Reece et al., Reference Reece, Dungan and Burreson2008). Future work should use molecular methods to both identify the species released from experimental oysters and test if P. chesapeaki and other species exhibit similar rates of release to P. marinus.
In summary, we measured the effects of environmental factors and time of day on the release of P. marinus cells from C. virginica. We found that temperature and time of day had strong effects on the number of parasites released, but that salinity only influenced the abundance of the parasite within its host. Furthermore, we found that live oysters can release numbers of P. marinus cells that are comparable to the amount released from dead and dying oysters. These factors should all be taken into consideration when estimating the population of free-living P. marinus in the water column and variations in this population over both hourly and seasonal timescales.
Supplementary material
The supplementary material for this article can be found at https://doi.org/10.1017/S0031182020002383
Data
All analyses are available on GitHub https://github.com/sagw/DE2019. Raw will be made available on Figshare.
Acknowledgements
We would like to thank Teresa Vaillancourt and Kristina Borst for assistance with Experiment #2, George Smith for experiment construction advice, Shannon Hood for help in procuring infected oysters, and Erik Holum and Simon Roberts for help with automating image analysis and statistics.
Financial support
Funding came from Hunterdon funds to K.M.P.L. and G.M.R. K.M.P.L. was supported as a Robert and Arlene Kogod Secretarial Scholar. M.S.R.N. was supported by an NSF REU. S.A.G.-W. was supported by a Smithsonian Institutional Fellowship.
Conflict of interest
None.
Ethical standards
Not applicable.
