| Collection Mémoires et thèses électroniques | |
| AccueilÀ proposNous joindre |
| Chapitre III. Life at the edge: Physiological constraints on freshwater mussels (Dreissena polymorpha Pallas and D. bugensis Andrusov) in a fluvial estuary | ||
|---|---|---|
 |  | |
Table des matières
Basée sur la faible densité et la distribution restreinte des moules zébrées (Dreissena polymorpha) que nous avons observées dans l’estuaire fluvial, nous avons émis l’hypothèse que la mortalité, la condition et la croissance des moules zébrées dépendraient de la salinité. Nous avons testé cette hypothèse en utilisant la simulation de régime de salinité variante au lieu d’une exposition constante. L’échantillonnage de terrain montrait que les moules quagga (D. bugensis) sont absentes, alors nous avons été plus loin dans l’hypothèse que cette espèce serait plus négativement affectée quand elle est testée en laboratoire. Contrairement à nos attentes, nous avons trouvé que les moules expérimentant une exposition périodique à 2 PSU de salinité étaient en meilleure condition et avaient un meilleau taux de croissance que celles exposées à des conditions continues d’environnements de 0 à 2 PSU. Alors qu’il n’y avait pas de différences interspécifiques dans la croissance ou la biomasse dans nos 36 jours d’expérience, nous avons trouvé que la salinité fluctuante affectaient le contenu organique du tissu mou et les concentrations en glycogène des moules zébrées plus que chez les moules quagga.
Based on the low density and restricted distribution of zebramussels (Dreissena polymorpha) that we observed in the fluvial estuary, we hypothesized that mortality, condition, and growth of zebra mussels would depend on salinity. We tested this hypothesis using a diurnally fluctuating salinity regime as well as of exposure to constant conditions. As field sampling showed that quagga mussels (D. bugensis) were consistently absent, we further hypothesized that this species would be more negatively impacted than zebra mussels. Contrary to our expectations, we found that mussels experiencing periodic salinity of 2 PSU were in better condition and had higher growth rates than those exposed to either continuous 0 or 2 PSU environments. While there were no interspecies differences in growth or biomass in our 36-day experiment, we did find that fluctuating salinity affected the soft tissue organic content and glycogen concentrations of zebra mussels more than quagga mussels.
Estuaries are characterized by a variety of fluctuating abiotic factors, chief among which are temperature, salinity, and turbidity as well as strong turbulence and advection. For the animals that live in these systems, particularly sessile benthic species, the osmotic challenges presented by fluctuating salinity are particularly strong. Despite these harsh features, large macro-tidal estuarine transition zones (ETZ) have high pelagic primary and secondary productivity (Vincent et al. 1996, Winkler et al. 2003) even though they are often dominated by only a few species (Waide et al. 1999, Levin et al. 2001). Thus even small changes in the taxonomic composition have the potential to generate large functional consequences if the species are strong interactors. Zebra and quagga mussels (Dreissena polymorpha Pallas and D. bugensis Andrusov) commonly inhabit brackish deltas and estuaries in high numbers in their Ponto-Caspian home range (Strayer and Smith 1993). This is a concern because as invasive dreissenid mussels displace native North American benthic fauna, they can potentially affect a wide range of seston characteristics including dissolved material (Roditi et al. 2000, Baines et al. 2005), bacteria (Silverman et al. 1996, 1997), particulate matter (Roditi et al. 1996, Baker et al. 1998), and small-bodied zooplankton (Thorp and Casper 2002, 2003, Wong et al. 2003). Field studies have linked dramatic declines in riverine and estuarine plankton to the arrival of dreissenid mussels via both direct and indirect interactions (Caraco et al. 1997, Bastviken et al. 1998, Welker and Walz 1998, Strayer et al. 1999, Caraco et al. 2000).
Mussel filtration rate is a functional attribute that connects benthic suspension feeders like Dreissena to the pelagic zone, but this rate can be strongly influenced by both organismal and population level factors, such as responses to osmotic stress and population size structure, respectively. Thus, to evaluate the potential extent of Dreissena’s filtration on the estuarine environment, we need a better understanding of its physiological constraints. In the ETZ of the St. Lawrence River turbidities approach the upper maximum of their tolerance in previous studies (Alexander et al. 1994, Summers et al. 1996), and the salinity and current speeds commonly observed in this system have also been shown to be stressful for Dreissena under laboratory conditions (Dietz et al. 1996, Wilcox and Dietz 1998, Ackerman 1999). Laboratory experiments have also shown that at continuous exposure to salinities above 2 – 4 PSU, adults are impaired and that significant mortality begins around 8 PSU (Mackie and Kilgour 1992, Kilgour et al. 1994). Dreissena has two abilities that could allow it to resist short-term exposure to brackish water: one behavioral and the other physiological. Behaviorally mussels can close their shells for hours to days, an ability that has allowed them to tolerate emersion and transportation hundreds of kilometers overland (Johnson et al. 2001). Physiologically, adults acclimate and reproduce under continuous exposure to brackish water of 6-8 PSU if the ratio of Na to K allows an osmotically balanced internal environment (Fong et al. 1995, Dietz et al. 1996). However, in contrast to the continuous salinities used in previous experiments, the St. Lawrence River-Estuary transition zone (ETZ) fluctuates both on daily and seasonal time-scales (St. Lawrence Centre 1996), exposing mussels in this estuary to continually variable salinity regimes. This distinction is particularly important because the daily fluctuation in salinity is more rapid than that of the laboratory experiments (Wilcox and Dietz 1996) that have shown salinity tolerance in Dreissena.
To evaluate whether Dreissena populations could exist in the St. Lawrence ETZ at abundances sufficient to lead to the kind of pelagic impacts seen in other large-river estuaries, we combined a field sampling presence in relation to tidal salinity with a laboratory assessment of the physiological response to tidally fluctuating salinity. We hypothesized that in addition to the effects of peak salinity, exposure to fluctuating levels would negatively affect Dreissena as much or more than constant exposure. The alternative would be that Dreissena are able to compensate physiologically. Further, to assess whether inter-specific differences in physiology can explain the difference in the distribution of the two species of Dreissena currently found in North America, the zebra and quagga mussels (D. polymorpha and D. bugensis), we included both species in the salinity exposure experiment.
The spatial distribution of mussels in the St. Lawrence ETZ was determined along 11 north/south transects located between the upstream limit of the ETZ on the eastern end of Île d’Orléans at the downstream limit Montmagny on the southern shore of the estuary (Figure 3.1). Triplicate samples were collected using a Van Veen dredge during the latter half of July in 2000, 2001 and 2002. Since this type of dredge can be inefficient when sampling hard packed clay or sand, both common substrates in this section of the estuary, samples were not used unless the dredge was more than 10 % full. Mussels collected with the dredge were held on ice 24 - 36 h then taken to the laboratory for measurement of biomass and maximal length. Each sample was visually classified for dominant sediment type (silt, clay, sand, gravel, and rock) and checked for evidence of mussel presence (e.g. byssal attachments). Shell length was measured to the nearest 0.01 mm using digital calipers.
To characterize the performance of mussels in response to salinity, we conducted a 36-d laboratory experiment at 18oC with both species (D. polymorpha and D. bugensis)in three salinity regimes. These three treatments were continuous exposure to fresh water (0 PSU), continuous exposure to brackish water (2 PSU), and exposure to water that varied between fresh and brackish water once every 24 hours (0-2 PSU). The variable exposure treatment was designed to simulate a single tidal intrusion of salinity every 24 hours. This manipulation was performed using a dual-pump system that fed 5.5 L of 28 PSU water into each replicate aquaria at 8 ml min-1for 6 h, then switched over to a second pump that fed 0 PSU water into the same aquaria at 8 ml min-1for a subsequent 6 h. The overall effect was a gradual rise to peak salinity followed by a gradual decline to 0 PSU, then approximately 12 hours at 0 PSU, repeated daily. 24 continuous flow-through tanks were divided into 8 replicates of the 3 salinity treatments, but because of the configuration of the tanks and plumbing, treatments were grouped into separate banks of 8. Thus, due to logistic constraints, we were unable to randomly intersperse replicates among experimental units, one limitation of the experimental design (Hurlbert 1984). Each replicate tank within a treatment contained 18-22 individuals of each mussel species in separate 10 x 10 x 10 cm plastic mesh cages (1 cm2openings). Mussels of both species with no previous exposure to brackish water were collected from the oligotrophic Great Lakes water mass near Massena NY (44o59’ 19’’ and 74o49’ 45’’), individually marked (“bee tags” attached by epoxy glue), measured (initial length) and then randomly assigned to one of eight replicate tanks within each salinity treatment for a total of approximately 170 D. polymorpha and 180 D. bugensis per treatment. Analysis prior to the experiment showed that there was a slight difference in initial mussel size between the two species (F1,1051= 35.29, p = <0.05), with the mean length of zebra mussels and quagga mussels being 20.7 and 19.4 mm-1respectively, but no difference among the treatments within a species (F1,1051= 1.51, p = 0.22). Because size differences can confound the interpretation of growth data, we selected the same size mussels (20±1 mm) for all species/salinity treatment combinations. Salinity was monitored using conductivity measurements at both a fixed time on every third day and at multiple intervals during a single cycle of the variable salinity treatment to ensure that minimum and maximum daily salinities remained constant. At no time did we find a significant variation in conductivity either within a single tank or between tanks from the same treatment during the experiment (F = 0.071, p = 0.89). To limit the potentially confounding effect of differences in food quality and quantity between the riverine and estuarine environments, the animals were not fed during the experiment period. While starvation itself is a stress, it was applied equally to all the treatments. Moreover, an initial pilot experiment indicated that both species could survive more than 100 days of starvation at 15oC before mortality occurs (Casper, unpublished data).
During the 36-d experiment, we used six measures of performance and condition to evaluate the impact of exposure to salinity: survival, shell growth, shell mass, tissue mass, tissue glycogen content, and RNA/DNA ratio. To estimate survival, we examined mussels every 3-d for the duration of the experiment. A mussel was considered dead and was removed if shells were not closed when disturbed lightly with a probe. All other measurements were made at the end of the experiment. The mean of each response variable per tank was calculated (n = 10-15 individuals per tank) and analyzed using 2-factor ANOVA (species x salinity exposure). Mussels from each tank were measured first for shell growth and then separated into 4 replicate subsets for mass and glycogen content analyses. The mass of organic carbon (mg OC) in shell and tissue was determined separately for 10 – 15 individuals from each treatment-species combination by drying to a constant weight at 600C loss of mass after ignition at 550oC for 4 h. The glycogen content of a different subset of 10-15 individuals was estimated using a modified acid hydrolysis-phenol method (sensu Naimo et al. 1998, Patterson et al. 1999, Okumura et al. 2002). Finally, we used RNA/DNA ratio as a biochemical measure of cellular-level responses. Used initially as a growth indicator in fish, the RNA/DNA ratio has also been successfully applied to marine gastropods and bivalves (Okumura et al. 2002, Dahlhoff 2004), suggesting that it may also be used on freshwater molluscs. To determine the RNA/DNA ratio, the soft tissue was dissected from the shell and homogenized in distilled water (5 times the volume/wet weight of the tissue). The extraction and purification procedure for nucleic acids was adapted from Buckley and Bulow (1987). In this procedure, the homogenized mussel is first treated with cold perchloric acid (HClO4) to remove free nucleotides, amino acids, and other low molecular weight molecules. These homogenization and cold perchloric acid steps also precipitate the RNA, DNA, and proteins into a pellet for further analysis. The RNA fraction of duplicate samples of each individual mussel is then solubilized in a KOH solution. DNA is separated from the protein pellet by treatment with hot perchloric acid and cooled to room temperature. RNA and DNA concentrations are then obtained by spectrophotometry (absorbance at 260 nm).
Visual analysis of the dredged sediments collected between 2000 and 2002 confirms previous findings (d’Anglejan and Brisbois 1978) that the surface sediments in the ETZ are dominated by a mixture of sand and clay (Figure 3.2). The log-likelihood test of independence (Zar 1984) shows there was no significant difference in the distribution of sediment types between the north and south channels (G = 8.754, p = <0.05) despite differences in depth and water circulation patterns (d’Anglejan and Brisbois 1978). Of all mussels collected from the ETZ over the 3-yr period, only one quagga mussel was found, leading us to conclude that, to this point, the ETZ is dominated by zebra mussels. The size frequency of the collected zebra mussels was initially unimodal, with most animals in the 18 – 20 mm length category in 2000 (Figure 3.3). In next two years the number of animals recovered declined by half, their mean size increased by 2 mm, but there was little apparent recruitment until a cohort of 10 mm animals appeared in 2002. Even though both quagga and zebra mussels are known to inhabit soft sediments in the Great Lakes (Berkman et al. 1998 , Haltuch et al. 2000), more than 90 % of the adult mussels collected were from cobble or larger sized rocks resting on the sediment surface, not on the sediments themselves (Figure 3.2). Despite the equal distribution of all sediment types between the north and south channels, adult mussels were limited to the southern sites (G4= 81.868, p = <0.05) and even where they were collected, live mussels occurred in only half of the samples (Figure 3.2).
Experiment-wide mussel mortality ranged from 0 to 8% after 36 days of exposure, but there were no significant differences between either species or treatments (F3,20= 0.963, p = 0.39). There was no shell growth during the experiment, but shell length significantly decreased for mussels in the freshwater treatment. No differences were observed between the two species (Table 3.1 and Figure 3.4). No significant differences in shell mass (F5,67= 1.09, p = 0.41), shell organic carbon content (F5,67= 2.01, p = 0.06), or soft-tissue mass (F5,67= 0.69, p = 0.63) were observed among any of the species-treatment combinations (Figures 3.4 and 3.5). In contrast both tissue organic carbon and glycogen content were influenced by salinity treatment, though only organic carbon content declined due to treatment alone (Table 3.2 and 3.3). Glycogen content in zebra mussels declined with increasing exposure while conversely levels in quagga mussels were elevated in the constant salinity treatment (Figure 3.5). RNA/DNA ratio also showed a species by treatment interaction with zebra mussel values higher in both salinity treatments while for quagga mussels the ratio was highest in the variable salinity but lowest in the constant salinity (Table 3.4 and Figure 3.6).
Although the median mussel size (18-20 mm) in the ETZ is large enough to have a filtration effect on plankton in experimental situations (Diggins 2001, Wong et al. 2003), their scarcity in our sampling grid shows that their numbers are not currently comparable with those observed to have in situ impacts on water column properties (Strayer et al. 1999, Thorp and Casper 2002, 2003). In light of this, we suggest that, at their current numbers, filter-feeding by the ETZ populations of adult dreissenid mussels are unlikely to have a significant impact on the plankton community although two other unexplored alternatives need to be evaluated before a Dreissena -mediated limitation of the ETZ plankton scenario can be ruled out entirely. First, this invasion began in the mid-1980s and may still be on-going. Often, there is a time lag before the full colonization potential of invasive species is seen. Second, the influence of the mussels on estuarine plankton may be spatially disconnected, more an indirect result of high mussel densities upriver and less directly by mussels living within the ETZ. Our results indicated that there was low availability of suitable substrata in either the north or south channels (sand and silt dominated). In this regard the sediments of ETZ resemble that of profundal Great Lakes zones where dreissenid mussels, quagga mussels in particular, have been found living on soft substrates (Mills et al. 1993, Roe and MacIsaac 1997, Berkman et al. 2000, Haltuch et al. 2000). However a high density, long-term dreissenid colonization in the ETZ is still unlikely because of tidally driven turbulence and episodic disturbance from winter storms and ice scour that prevail in this ecosystem (St. Lawrence Centre 1996).
Based on our experimental results, we conclude that Dreissena is physiologically able to tolerate tidally brackish conditions and thus are unlikely to be excluded from the ETZ by salinity alone. However,field verification showed that Dreissena distribution extended further downstream in the southern channel than the northern of the St. Lawrence ETZ despite uniformity in available substrate. This pattern conforms to the established tidal intrusion of the salt wedge in this system (further inland on the north shore, d’Anglejan and Brisbois 1978, Kranck 1979). Alternatively, if Dreissena is unable to colonize the soft sediments of the ETZ, then their influence is more likely to be associated with high numbers of mussels in upstream reaches affecting river seston in transit to the estuary (St. Lawrence Centre 1996, Thorp et al. 2005). The relatively small numbers of sessile adult Dreissena in the ETZ also appear to be inconsistent with the large numbers of their planktonic larvae that are dispersed into the estuary (Winkler et al. 2003, Barnard et al. 2003). If the annual supply of potential recruits to the ETZ is not limited and if neither recruitment nor preferred habitat (hard, rocky surfaces) are limiting, then annual recruitment could be expected to be relatively steady. Instead, based on the size frequency data, recruitment is intermittent with only one discernable cohort over the 3 years sampled. From this we can conclude that another factor(s) may be limiting the mussels. Possible factors include disturbance of the rocky substrate from storms and ice scour or physiological constraints on either adults or larvae.The fact that a congeneric species, the quagga mussel D. bugensis, which has similar dispersal mechanisms and substratum requirements and is equally or more abundant upriver (Ricciardi and Whoriskey 2004) but absent from the ETZ, suggests that neither larval supply nor substrate limitation are at the root of the difference.
While the preferred rocky substrate is distributed across the northern and southern St. Lawrence ETZ, Dreissena are limited to the southern channel implying substrate is not the limiting factor. Topographically the St. Lawrence ETZ is divided into a narrow, deep northern channel and a broad, shallow southern channel by an archipelago of islands extending downstream from of Île d’Orléans, strongly controlling the tidal intrusion of salinity (d’Anglejan and Brisbois 1978, Kranck 1979, Vincent and Dodson 1999). The northward push of Coriolis circulation forces, combined with natural tendency of the freshwater discharge to flow into the southern channel, has the net effect of promoting a greater tidal intrusion of salinity up the northern channel typically as far as the eastern tip of Île d’Orléans. This means that the same salinity conditions found at of Île d’Orléans rarely reach past the city of Montmagny on the south shore of the ETZ, 20 km further downstream (d’Anglejan and Brisbois 1978, Kranck 1979). Unlike the availability of substrata, salinity has a pronounced north-south channel asymmetry. Elevated turbidity and rapid changes in salinity are also significant physiological challenges for dreissenids (Mackie and Kilgour 1992, Alexander et al. 1994, Kilgour et al. 1994, Summers et al. 1996). Thus the asymmetric tidal intrusion of salinity mirrors the mussel distribution suggesting that it may be an important variable for dreissenids. We focused on laboratory testing on whether tidally fluctuating salinity was a limiting factor due to our observations on the spatial pattern of this freshwater mussel’s distribution in the ETZ which appeared to be bounded by intrusions of low salinity water.
Several authors have looked at the effect of salinity on zebra mussel survival and physiology in the laboratory, but in each case assays were conducted with constant salinity treatments at fixed levels. In contrast, the tidal nature of the St. Lawrence system means that sessile invertebrates in situ may experience daily fluctuations in salinity due to tidal intrusion of the salt wedge. In principle, one would expect that physiological responses to fluctuating salinity would be intermediate relative to the conditions of constant fresh water and constant brackish water, but this was not the case in this study. The lack of growth in the fresh water treatment is counter-intuitive for a fresh water animal, but the reduced oxygen consumption (and presumably cellular catabolism) and activity of some stenohaline estuarine organisms exposed to higher and lower salinity (Kinne 1971, Findley et al. 1978, Garton and Stickle 1980) suggests a potential explanation. The brackish conditions of the experiment (2 PSU) were well within the tolerance range of adult Dreissena (Dietz et al. 1996, Wilcox and Dietz 1998), but continuous exposure to even this low osmotic stress did reduce organic carbon component of the tissue of both species suggesting that the mussels were using energy reserves. Glycogen levels for zebra mussels were indeed lower, coinciding with elevated cellular activity as seen with the RNA/DNA ratio. However, this same pattern was not seen in quagga mussels. Instead, quagga mussel glycogen levels were highest and RNA/DNA ratios lowest in the continuous brackish treatment. We interpret this to mean that zebra mussels continue to function metabolically in the face of increasing salinity exposure while quagga mussels do not. Because our experiment was run under starvation conditions, we cannot attribute these responses to salinity alone. Whereas glycogen is considered the principal short-term energy storage material for most bivalves (Thorp and Covich 2001, Smolders et al. 2004), zebra mussels could be using an alternative energy source when the glycogen was depleted such as lipids or proteins. The higher RNA/DNA ratio in zebra mussels exposed to any salinity, despite declining organic and glycogen content, lends support the possibility of an alterative energy storage material. However, more detailed biochemical studies will be needed to address the question of the order of priority among carbohydrate, protein, and lipid metabolism in the face of physiological challenges such as salinity.
Populations ofquagga mussels are displacing zebra mussels in many locations throughout non-tidal portions of the Great Lakes/St. Lawrence system (Mills et al. 1999, Stoeckmann 2003, Ricciardi and Whorisky 2004, Wilson et al. 2006). However, quagga mussels were extremely rare during our 3-yr sampling period. Although it has been over 10 years since quagga mussels were first detected in the river, zebra mussels were found in the tidal portions of the St. Lawrence within 5 years of the first detection in the Great Lakes (Mellina and Rassmussen 1994). If quagga mussels are able to survive under tidally variable salinity as well as the zebra mussel can, then their physiological advantage in assimilation (Baldwin et al. 2002, Stoeckmann 2003) may allow them to out compete and displace zebra mussels here too. Unfortunately, other than the lower fecundity of quagga mussels, the differences in reproductive and dispersal biology of the two taxa have not been explored sufficiently for further speculation. In terms of our experimental design, unlike the estuarine situations in the Ponto-Caspian region where Dreissena is common (Strayer and Smith 1993), the St. Lawrence ETZ is low in ionic strength and a high-energy system with two peaks per day of turbidity in excess of 100 NTUs and bottom currents ranging from 10 to 130 cm/s-1(d’Anglejan and Brisbois 1978, Kranck 1979, St. Lawrence Centre 1996). By ignoring the interactive affects of multiple stressors, our 36-d salinity exposure experiment may not represent adequately the daily covariation in salinity, temperature, and flow that characterize the St. Lawrence system that dreissenids experience over their life cycle (d’Anglejan and Brisbois 1978, Kranck 1979, St. Lawrence Centre 1996). Despite this drawback, our comparison of organic carbon, glycogen, and RNA/DNA show that the two species differ in their response to salinity. Furthermore, because both sets of animals were drawn from naïve upriver populations that had never been exposed or acclimated to the ETZ conditions, we can also conclude that the differences in response are intrinsic to the species and not an artifact of differential acclimatization or colonization histories. Finally, although we show evidence of physiological differences among adults that correspond to patterns of distribution, final conclusions about environmental constraints and the ultimate distribution of these two species must wait for parallel studies on the effect of estuarine conditions on larval and juvenile stages.
Table 3.1. Results of 2-factor ANOVA comparing change in shell length for Dreissena polymorpha and D. bugensis in relation to salinity treatments (* indicates p < 0.05 while ** indicates p < 0.01).
|
|
Table 3.2. Results of 2-factor ANOVA comparing final organic carbon content of the soft tissue for Dreissena polymorpha and D. bugensis in relation to salinity treatments (* indicates p < 0.05 while ** indicates p < 0.01).
|
|
Table 3.3. Results of 2-factor ANOVA comparing final glycogen content for Dreissena polymorpha and D. bugensis in relation to salinity treatment (* indicates p < 0.05 while ** indicates p < 0.01).
|
|
Table 3.4. Results of 2-factor ANOVA comparing final RNA/DNA ratio for Dreissena polymorpha and D. bugensis in relation to salinity treatment (* indicates p < 0.05 while ** indicates p < 0.01).
|
|
Figure 3.1. Schematic of the benthic sampling plan in the St. Lawrence estuarine transition zone. Points indicate dredge sites, shaded area marks extent of the distribution of adult Dreissena (2000 – 2002).
Figure 3.2. Comparison of the frequency distribution of the presence/absence of Dreissena (top) and sediment type in the north and south channels of the St. Lawrence estuarine transition zone (bottom) for all samples collected between 2000 and 2002.
Figure 3.3. Size distribution of D. polymorpha collected from the St. Lawrence estuarine transition zone (2000-2002).
Figure 3.4. Total change in shell length, average shell mass, and organic carbon content, for Dreissena polymorpha and D. bugensis in relation to experimental salinity exposure (LS mean +/- standard error from 2-factor ANOVA, letters indicate differences in Tukey HSD post-hoc comparisons).
© Andrew F Casper, 2007
| | | |
| Chapitre II. Disparity in the cost of living: Abiotic influences on shell and tissue traits of Dreissena polymorpha (Pallas) and D. bugensis (Andrusov) |  | Chapitre IV. Contrasts in shell and soft tissue physiology of dreissenid mussels coincident with shifts in species dominance. |
