| Collection Mémoires et thèses électroniques | |
| AccueilÀ proposNous joindre |
| Chapitre IV. Contrasts in shell and soft tissue physiology of dreissenid mussels coincident with shifts in species dominance. | ||
|---|---|---|
 |  | |
Table des matières
La croissance est fréquemment utilisée pour évaluer l’effet d’une suite de facteurs environnementaux sur l’état physiologique de l’organisme. Dans les mollusques bivalves, la croissance est souvent mesurée par l’augmentation dans les dimensions de la coquille, laquelle est pratique mais ne peut pas estimer précisément la croissance somatique du tissu mou comme plasticité phénotypique et l’énergied’allocation flexiblepeut se confondre avec cette mesure de croissance. Cette étude compare la variabilité dans la croissance de la coquille, le rapport ARN/ADN (un indicateur de métabolisme cellulaire) et changements dans la biomasse du tissu des moules de type Dreissena (Dreissena polymorpha Pallas and D. bugensis Androsuv) dans des environnements contrastants. Les transplants réciproques de D. polymorpha entre les environnements de rivières et de l’estuaire indiquaient que la croissance de la coquille répondait différemment que la croissance du tissu mou. Une des mesures de croissance seules de la croissance de la coquille montrait un composant consistent spécifique de la population source: la croissance étant plus pauvre dans l’estuaire, meilleur en rivière, et intermédiaire dans les transplants. Au contraire, le rapport ARN/ADN était plus haut dans l’estuaire et égal ailleurs, suggérant que la croissance de la coquille ne peut jamais refléter précisément des mesures à court terme de la croissance de la moule. En environnement de rivière, les moules quagga croissaient plus rapidement que les moules zébrées, ce qui pouvait refléter le remplacement de D. polymorpha par D. bugensis le long du système Grands Lacs-St Laurent. Elles n’ont pas cependant la même performance que celles transplantées en environnement de l’estuaire suggérant que leur influence serait limitée aux portions strictement d’eau douce du système.
Differences in growth rates arefrequently used to evaluate the effect of abiotic factors on the limits of species distribution and dominance. In bivalve molluscs, growth of both shell and somatic tissue are often inferred by measured changes in shell alone. For mussels, shell growth may not always reflectgrowth in soft tissue accurately because of confounding factors such as physiological asynchrony between shell and tissue, flexible energy allocation or population differentiation. This study compares the relationship between shell growth, RNA/DNA ratio (an indicator of cellular metabolism and thus an index of short-term tissue condition), and changes in soft tissue mass of Dreissena polymorpha Pallas and D. bugensis Androsuv) from contrasting source populations . Reciprocal transplantation of D. polymorpha between riverine and estuarine environments indicated that summer shell growth was consistently lower for the estuarine source individuals and better for the riverine source individuals, independent of which environment that they were transplanted to.In contrast, the RNA/DNA ratiowas higher in the estuary control treatment relative to other D. polymorpha treatments. There were no overall differences in RNA/DNA ratios between D. polymorpha and D. bugensis from the same site, though summer shell growth of the smaller quagga mussels was almost two times greater than D. polymorpha. D. bugensis did not grow as well when transplanted to an estuarine environment, suggesting that their distribution will be limited to freshwater habitats. Thus this experiment shows that shell growth of dreissenid mussels does not always accurately reflecttissue growth for either species and that physiological performance of D. polymorpha is specific to source-destination treatment combinations.
Growth is a physiological trait integrating both internal and external factors. As such it is plastic, responding to a variety of factors, including food availability, food quality,ontogeny, or allocation of energy to seasonal activities like reproduction. Many authors have theorized that the success of certain invasive species may be due to a high degree of plasticity for growth in novel, heterogeneous environments (Ricciardi and Rasmussen 1998, Kolar and Lodge 2001, Sakai et al. 2001). Zebra and quagga mussels (Dreissena polymorpha and D. bugensis, respectively), both recent invaders of North America, are fast-growing freshwater mussels that have quickly colonized a broad range of freshwater environments as they rapidly spread across the eastern half of the continent (Thorp and Covich 2001, McMahon 2002). Unfortunately for our need to understand the constraints on aquatic invasive species, there are only a few direct examinations of the physiological adaptability of this genus (but see Wright et al. 1996, Wilcox and Dietz 1998 and references in McMahon 2002).
In many invertebrates, measuring individual growth in relation to these external constraints can be complicated by two contrasting compartments of overall animal growth: soft tissue and skeletal mass. In mussels, growth of soft tissue is commonly estimated from an allometric relationship between shell length and soft tissue or total mass (Smit et al. 1992, MacIssac 1994, Jokela and Mutikainen 1995, Jokela 1996, Young et al. 1996, Chase and Bailey 1999a, Stoeckman and Garton 2001). However, because many bivalves can vary the allocation of energy between shell and soft tissue growth depending on environmental conditions or life history needs (Hilbish 1986, Berard et al. 1992, Jokela and Mutikainen 1995, Jokela 1996, Stoeckmann and Garton 2001), shell and somatic growth can be asynchronous and thus be an incomplete or misleading proxy for either soft tissue or total mussel growth. For example, Dreissena typically show seasonal declines in soft tissue mass of up to 40% in association with summer metabolic demands and spawning events without parallel change in shell dimensions or mass (Stoeckman and Garton 1997) very similar to the seasonal asynchrony seen in Mytilus (Gosling, 1992).
The first goal of this study was to compare measures of tissue and shell growth of Dreissena spp . including a newer approach that reflects rates of cellular synthesis, RNA/DNA ratios. The RNA/DNA ratio has been shown to be a sensitive indicator of soft tissue condition and metabolism in a several marine molluscs (Dahlhoff et al. 2002, Okumura et al. 2002, Dahlhoff 2004), and it could potentially provide a direct index of metabolic rates in dreissenid mussels. The premise of this technique is that the quantity of RNA in a cell varies with protein synthesis in contrast to the quantity of DNA in a cell, which is fixed and insensitive to either environmental or nutritional conditions (Dahlhoff 2004). Because increases in the cellular levels of protein precede overall accumulation of mass, the RNA/DNA ratio varies more rapidly than whole animal mass in relation to changes in environmental or nutritional conditions. For example, the response time of this ratio to environmental conditions and food regimes in marine molluscs ranges from hours to days (Chicharo et al. 2001 , Okumura et al. 2002), supporting the idea that RNA/DNA ratio reflects short-term metabolic activity of soft tissue. By comparing this ratio with changes in shell length and soft tissue biomass, it may be possible to get a more complete and accurate estimate of variability in individual physiology and performance (Dahlhoff et al. 2002, Dahlhoff 2004).
To provide the contrasting conditions that can highlight differences in shell and tissue physiology, we chose two divergent environments within the St. Lawrence River: a site within the oligotrophic upper river and a second in the turbulent and brackish (0 – 2 PSU) fluvial estuary. While Dreissena is present at both locations the turbulent flow, periodic exposure to salinity, and elevated suspended sediment common to the fluvial estuary are potentially very stressful for the mussels (Alexander et al. 1994, Summers et al. 1996, Dietz et al. 1996, Wright et al. 1996, Thorp et al. 1998, Wilcox and Dietz 1998). Environment is not the only factor that controls growth, and selection can lead to substantial genetic differentiation in even a single generation (Schmidt and Rand 1999, Veliz et al. 2004). Optimally to distinguish genotypic divergence from phenotypic plasticity, a breeding study carried through the F2generation is typically used. Unfortunately, the planktonic larval stage of Dreissena is difficult to culture and rear in the lab, making this type of breeding study problematic. An alternative approach is the use of experiments in which animals from two different source environments are exchanged, with a sub-set maintained in each environment as a control. This reciprocal transplant design can then be used to distinguish reversible phenotypic plasticity from non-reversible change. By reciprocally transplanting mussels between populations in the fluvial estuary and riverine sections of the St. Lawrence River, the physiological flexibility and rapid adjustment ability of Dreissena can be assessed. This soft tissue to shell mass comparison of transplants also permits assessment of whether change in shell length is actually a robust indicator of change in either soft-tissue or total mass. In addition, we can assess the utility of RNA/DNA ratios in providing information complementary to direct measurements of soft tissue growth alone.
As a final element of this study, we compared the responses of the zebra mussel (D. polymorpha)and the more recently introduced quagga mussel (D. bugensis). The quagga mussel appears to be displacing the zebra mussel across much of the Great Lakes basin (Mills et al. 1996, Ricciardi and Whorisky 2004, Wilson et al. 2006), but this species is still rare in the lower St. Lawrence River and fluvial estuary (Mingelbier, personal communication; Casper, unpublished data). We conducted experiments with both species to evaluate whether there are significant physiological differences between sympatric populations of zebra and quaggamussels that might help explain these observed patterns of species dominance. We hypothesized that the growth rate of quagga mussels would be greater than that of zebra mussels, consistent with the continuing displacement of the latter by the former in many of the habitats where they occur sympatrically.
The St. Lawrence is the second-largest river in North Americawith a 574,000 km2watershed, a basin that holds ~ 18% of the world’s freshwater (23,000+ km3), a 655 km3mainstem and mean annual discharge of 12,000 m3/s into a 130 km long fluvial estuary at Quebec City, (St. Lawrence Center 1996, Thorp et al. 2005). Two sub-populations of zebra mussels were reciprocally transplanted from the upstream fluvial mainstem and the downstream freshwater estuarine zone, sites separated by >600 km. These source populations exist in stable but contrasting physical-chemical water masses: Robinson Bay near Massena, New York (44o59’ 19’’ and 74o49’ 45’’), represents a riverine environment dominated by the oligotrophic outflow of Lake Ontario, whereas the second site at Saint-Francois on the eastern tip of Île d’Orléans, Québec (46o59’ 47’’ and 69o48’ 21”), represented the head of the tidally brackish (0-2 PSU) and highly turbid (>100 NTU), freshwater fluvial estuary (Figure 4.1, henceforth R and FE sites, respectively). All mussels were removed from rocks by hand and held at 15oC in the laboratory for a period of 24-72 h during which they were individually numbered with small plastic bee tags (glued to the shell) and measured for maximum shell length to the nearest 0.01 mm using digital calipers. To obtain an estimate of measurement error for shell length, caliper measurements were repeated 10 times on 6 different individuals ranging in size from 14 to 26 mm in length. Our mean (SD) estimate of measurement error was 0.075 (0.048) mm – less than 20 % of the maximum change in shell length observed during the experiment. The size distribution of the zebra mussels varied slightly with source population:FE source zebra mussels had 30% greater initial soft-tissue DW than R source zebra mussels for an equivalent shell length [the estimated mean initial length of R source zebra mussels was 25 mm (0.4 SE, 20-30 mm range) with an estimated soft tissue DW of (0.002*length) - 0.026 = 42 mg (r2= 0.85), and for FE source zebra mussels the mean initial length was 27 mm (0.4 SE, 20-33 mm range) with an estimated soft tissue DW of (0.003*length) - 0.025 = 55 mg (r2= 0.71).Quagga mussels from the R source population were markedly smaller than co-occurring zebra mussels, ranging from 14 to 21 mm for a mean of 18 (0.3).Unfortunately quagga mussels were not found in sufficient numbers at the fluvial estuarine location to be reciprocally transplanted. All mussels were then placed in a pair of20x25cmenvelope shaped cages made of 1-cm2plastic mesh (Vexar™) and anchored on cinder blocks. For this field experiment, approximately 200 adult zebra mussels were collected from hard substrata at least 2-m deep at both sites; 100 were returned to two cages in the native environment as controls and 100 were transplanted into two cages at the other site as treatments. While this design does not allow us to attribute any differences in organismal growth to a specific environmental parameter, our primary goal was to assess whether there was any divergence in shell and tissue physiological responses from contrasting source environments (i.e., site differences). The experimental unit for statistical purposes is thus the individual mussel and the two cages represent simply a hedge against disturbance or loss rather than replicates in any statistical sense. We also transplanted quagga mussels (D. bugensis) from the riverine site to the estuarine site in a similar way, but the reciprocal transplant was not feasible as quagga mussels could not be found at the estuarine site. Both control and transplanted quagga mussels were placed in the same cages as used for the zebra mussels. The experiment began in mid-July 2002 and ended 58 days later. Because growth and mortality could be strongly influenced by temperature, we measured water temperature every 30 min during the experiment using data loggers (Onset StowawaysTM) placed inside the experimental cages.
After retrieval, surviving mussels were randomly sub-sampled for one of the three complementary indicators of mussel condition: change in shell length, soft tissue mass, and RNA/DNA ratio.Shell growth is given as the total change in the maximum shell length of all surviving mussels after 58 days standardized to a 25.7 mm mussel for each treatment group. Soft tissue was separated from the shell for 20 – 30 individuals, and soft tissue dry weight (DW) was measured after drying for 48 hours at 60oC. For comparative purposes, initial DW for the standard sized mussel was also estimated from the regression of shell length against DW of mussels collected at the time experiments were begun. This value was then subtracted from the final DW for the same standard size of mussel to give the total change in the dry weight of soft tissue biomass.
A sub-sample of 10 individuals from each treatment was frozen at –80oC for RNA/DNA analysis immediately after retrieval. The soft tissue of the mussel 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 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 then separated from the protein pellet by treatment with hot perchloric acid and cooled to room temperature.
Separate ANCOVAs were conducted for each species due to differences in experimental design and initial sizes. For D. polymorpha the model was response variable = initial length * source population * transplant destination. For D. bugensis the model wasresponse variable = initial length * transplant destination. Initial length was used as a covariate in all models as both growth rate and mortality can both vary ontogenetically. Conversely, temperature was not included in the model statement because there was little difference between the two sites. Data were checked for normality and homoscedasticity first by visually examining for outliers and then by checking for randomness using plots of residuals from preliminary ANCOVAs.
Mortality of manipulated mussels varied depending on source, destination, and species even though there was little difference in water temperature among sites during the experiment (Figure 4.2). Mortality of R site zebra mussels was slightly greater at the FE site relative to their source location (15 vs. 9%). Mortality of FE site zebra mussels was 40% at their source location compared to a total lack of mortality at the R site. Mortality of quagga mussels was more than 3 times higher at their natal R site than at the FE site (15 % vs. 4%).
Shell growth D. polymorpha was low during the experiment; of the four transplant combinations only R to R mussels increased in shell length while the others either did not change or decreased (Figure 4.3). In ANCOVA terms shell growth was significantly affected by both source and destination without interaction. D. polymorpha from the R site did better than mussels from the FE site and mussels growing in the R site doing better than those growing in the FE site (Table 4.1, Figure 4.3). Growth of zebra mussels transplanted from the R to the FE site was lower in comparison with R control mussels, but did not decline to the level of FE control or transplanted mussels (Figure 4.3, Table 4.1). Consistent with this effect of site, estuarine source zebra mussels transplanted to the R site did not experience any reduction in shell length (shell growth was not different from zero) in contrast to the FE controls, which decreased in size. Likewise, R control D. bugensis had positive shell growth, but when transplanted to the FE site the shell growth of these animals became negative (Figure 4.3, Table 4.2). At the R site, quagga mussel growth rate was over 3-fold greater than that of coexisting zebra mussels, though this difference is confounded by the aforementioned difference in size between the two groups.
Direct comparison of the final DW among all four treatments could not be made due to differences in initial soft-tissue mass of the two source populations. However all zebra mussels lost soft-tissue mass during the course of the experiment, though zebra mussels from the R site had an overall smaller loss of mass (65%) relative to the FE site zebra mussels. (It must be kept in mind, however, that in terms of percentages, this difference is much less given that zebra mussels from the R site initially only had 70% of the mass of those from the FE site for an equivalent size mussel.) For R site zebra mussels there was no significant difference between those kept at the R site and those transplanted to the FE site, whereas zebra mussels from the FE site lost more mass when transferred to the R site compared to the controls (Figure 4.3).FE site zebra mussels had 30% greater initial soft-tissue DW than R source zebra mussels for an equivalent shell length. Relative to zebra mussels, changes in soft-tissue DW for the riverine quagga mussels were far lower (<20%) in spite of their much smaller size. There was a significant treatment effect with quagga mussels transplanted to the FE site losing weight while quagga mussels retained at the R site showing no change.The RNA/DNA ratio of the zebra mussels did not parallel the trends seen in shell growth (Figure 4.4, Table 4.1) with the highest values for the FE controls. The RNA/DNA ratio ofquagga musselstransplanted to the FE site was not significantly different from controls (Figure 4.4, Table 4.2), and shell length never explained a significant portion of the variation in RNA/DNA ratio for either species (Tables 4.1 and 4.2).
Examining a combination of shell and tissue characteristics also allows us to see that shell growth alone does not always reflect changes in overall mass as it is commonly presented; in fact the shell and tissue characteristics can be at odds (e.g. lowest shell growth but higher metabolism for estuarine source animals at the estuarine site). With its stable regional-scale environmental heterogeneity (St. Lawrence Centre 1996, Thorp et al. 2005) and contrasting physiological constraints for Dreissena spp. (Karatayev et al. 1998), the St. Lawrence river-estuary is an ideal system for exploringtheflexibility/differentiation in physiological performance among spatially separated populations of these highly invasive mussels. In addition, our experiment also contrasted the physiology of the two sympatric species of Dreissena, the zebra mussel (D. polymorpha) and the quagga mussel (D. bugensis) in environments differing in terms of salinity, turbidity, food availability, and hydrodynamic conditions (St. Lawrence Centre 1996, Thorp et al. 2005), any one of which could pose a physiological challenge (Karatayev et al. 1998). Our working hypothesis was that the fluvial estuary would generally be a poor environment for both species because of elevated turbidity, exposure to brackish conditions, and tidally induced turbulence. We thus expected naïve source populations of both mussels transplanted into this fluvial estuary as well as those who are residents of that environment to have lower rates of shell growth,relatively lower levels soft tissue mass, and lower metabolic activity. For our measure of shell growth in both species, our results support these predictions with poorer performance at the fluvial estuary site (both species) and poorer performance by mussels native to this site (zebra mussel). In contrast, measures of soft tissue did not uniformly indicate that the FE site was necessarily a poorer environment. Although zebra mussels from this site did lose the most soft-tissue mass, they were also initially larger on a size-specific basis. As the experiment was initiated after the usual time of spawning for this species in this system (St. Lawrence Centre 1996), we have assumed here that these changes are all due to changes in somatic tissue rather then gonadal tissues. This suggests that while destination conditions in the fluvial estuary did affect performance, there was also an effect of the source population. RNA/DNA did not parallel changes in shell or tissue traits. Unexpectedly, both the final tissue mass and RNA/DNA ratio were higher for the estuarine source zebra mussels at the estuarine site, although these mussels were initially larger and lost more mass during the course of the experiment than the other source-destination-species combinations. Another counter-intuitive result was that the large decrease in soft-tissue mass of the estuarine mussels transplanted to the riverine site was not reflected in their RNA/DNA ratio. One explanation for this apparent inconsistency may be due to the different temporal scales over which these measures integrate the performance of the mussels. One of the most important considerations when interpreting measures of condition and performance is the temporal scale over which they integrate. RNA/DNA ratios are generally assumed to respond rapidly to changes in the environment (typically within days or lessChicharo et al. 2001, Dahlhoff et al. 2002, Okumura et al. 2002, Dahlhoff 2004) whereas tissue mass in Dreissena has strong patterns of seasonal variation (Stoeckman and Garton 1997, Chase and Bailey 1999b) and shell accumulation would reflect the entire lifespan. Thus, in the context of this study, this combination of measures was not redundant. Summer is an energetically demanding period of maximum reproductive output during a period potential heat and food stress often resulting in significant weight loss among native freshwater mussels in general (Dillon 2000, Thorp and Covich 2001) as well as dreissenids (Stoeckman and Garton 1997). Lower cellular metabolism at the riverine site could be seen as a paradox, elevated RNA/DNA in the fluvial estuary despite declining shell and tissue condition. However previous work has suggested that in environments with fluctuating salinity, an animal exposed to a salinity regime outside its optimal range of osmotic function may show reduced cellular activity (Garton and Stickle 1980). In addition, the possibility that the cellular levels of RNA can vary with daily fluctuations in the endogenous or exogenous environment (Chicharo et al. 2001) also supports the idea that mass and RNA/DNA ratio are not responding on the same time-scale. Thus we speculate that exposure to salinity in the fluvial estuary is preventing the reduction in RNA/DNA ratio seen in the other source-destination combinations. Laboratory experiments are needed to examine more precisely this physiological response, in order to confirm this result and interpretation for naïve zebra mussels.
The zebra mussel, the first of the two dreissenids to be discovered in North America, is being displaced by the quagga mussel throughout the lower Great Lakes and the upper portions of the St. Lawrence River (Mills et al. 1999, Ricciardi andWhorisky2004, Wilson et al. 2006). Currently, although the quagga mussels dominate the upper portions of the St. Lawrence, they decline sharply below Montréal, Québec, and are rare in the fluvial estuary (M. Mingelbier, personnel communication; Casperunpublished data). When sympatric river source zebra and quagga mussels were compared at their source location, the shell growth of quagga mussels was higher while the amount of tissue mass lost during summer is lower. When both species are transplanted to the estuary, there is a negative impact on the soft tissue mass of both species, (zebra mussels lose 25% whereas quagga mussels lose 10-15%) even though the RNA/DNA values of these two species-site combinations are largely unaffected. Thus, of the three approaches to measuring physiological performance, only changes in shell length provided results that were somewhat consistent with our predictions. After adjusting for the size differences between using ANCOVA procedures, our data indicate that in the upper reaches of the river, quagga mussel shell growth is greaterthanany other source-destination-species combination. This finding of a greater shell growth rate for quagga mussels is consistent with observed trends in abundance between these two species in the St. Lawrence (Ricciardi and Whorisky 2004). Controlled laboratory experiments have shown that quagga mussels have higher assimilation efficiencies (Baldwin et al. 2002) and lower respiration rates (Stoeckman and Garton 1997) than zebra mussels thus plausibly explaining the comparatively higher growth rates found in this and other studies (Stoeckman 2003). However, such interpretations are confounded by the differences in size and possibly age between the populations of the two species. Even with ANCOVA procedures the sizes of the two species were only adjusted to within 8 mm. Thus, in addition to possible inherent difference in assimilation, the greater quagga shell growth at the river site could also be due to ontogenetic differences between the species-treatment combinations. At the river site where quagga mussel colonies are overtaking zebra mussels; the quagga mussel shell growth is also greater while its loss of tissue mass is less. Based on these shell/tissue contrasts, we hypothesize that the problem for quagga mussels in the fluvial estuary is related to limitationson shell physiology. Further experimentation using mussels of similar size and, ideally, from the same recruitment cohort is required to confirm this idea. Our interpretation of fixed source- and species-specific differences in physiology must be tempered by the realization that the quagga mussel may alternatively be a victim of either lower fecundity or reduced capacity of the larval stage, either of which might temporarily produce a similar result. Continued tracking of changes in the ratio of quagga to zebra mussels over time is likely to clarify this caveat.
Dreissenid mussels exert a tremendous influence on both the pelagic and benthic communities in aquatic systems based on a combination of their population densities, demographics, and filtration rate (Welker and Walz 1998, Ietswaart et al. 1999, Strayer et al. 1999, Thorp and Casper 2002, 2003). However this influence is strongly influenced by the size structure, growth rate, and physiological condition. Therefore accurate measurements of the growth rates corrected for localized differences between populations in those functional characteristics may have important ramifications for evaluating and predicting the impacts (MacIsaac 1996, Ricciardi and Rasmussen 1998). Dreissenidshellgrowth has been measured in a number of river and lake systems and exhibits a wide range of variation (Table 4.3), likely a result of variation in a suite of environmental factors including temperature. This study shows that these responses to the abiotic template may not be plastic, but fixed. The temperature-dependent growth of Dreissena (MacIsaac 1994, Thorp et al. 1998) suggests that northern populations should have some of the lowest growth rates in the literature.The shell growth rates reported here fall at the low end of these ranges. Only one other study comes from a large river system with similar geographic characteristics (the Rhine River) where growth rates appear substantially greater than in the St. Lawrence River. However, in spite of the similarities in latitude and size, the Rhine River study differs in two important ways. First, the Rhine growth study included a 4-month spring period with an inherently greater potential for growth due to a spring bloom of lipid-rich diatoms (Smit et al. 1992). Second, the Rhine River also has much higher levels of dissolved organic carbon (Smit et al. 1992, Thorp et al. 2005) available for use by dreissenids (Roditi et al. 2000, Baines et al. 2005). The only other comparable large river study is from the mid-latitude Ohio River, which reported substantially greater shell growth than the St. Lawrence animals (Thorp et al. 1998). In this case however, a smaller initial mussel size combined with the longer growing season and the richer food resources found in this warmer river may explain the greater annual growth. Of the studies we reviewed, the St. Lawrence populations of zebraand quagga mussels tend to conform more closely to those from oligotrophic lentic systems within the Great Lakes basin than other large rivers (Table 4.3).
Unlike quagga mussels , zebra mussels are present in both river and estuary environments. Despite the potential for negative impacts from tidally driven currents and salinity, the St. Lawrence Estuary has higher levels of bacteria and phytoplankton food resources than the riverine zone (Vincent et al. 1996). Because Dreissena’s food use ranges from dissolved carbon and the bacteria to small plankton (Roditi 2000, Thorp and Casper 2002, Baines et al. 2005), any individuals able to tolerate estuarine conditions would also have access to a richer food supply. The higher tissue mass of estuarine animals at the estuarine site and the lack of convergence in growth and condition of zebra mussels transplanted to the estuary with estuarine control mussels (i.e., no significant effect of either destination alone or source-destination interaction) supports the idea that non-reversible physiological changes (termed ecotypes) instead of that of physiological plasticity. When estuary source individuals are moved to the river site, growth does increase but not to the level of the local river source animals.
Our results alsosuggest rapid differentiation could be occurring within a seemingly contiguous population in the St. Lawrence River. Because the Great Lakes-St. Lawrence River corridor was first colonized in the early 1990s, these populations are at least 10 – 15 generations old. This is enough time in other animals’ examples for microevolutionary mechanisms to lead to divergence provided there is either restricted larval dispersal or auto-recruitment in local environments (Schluter 2000). Our evidence for divergence among local populations, though limited to analysis of adults, is reinforced by the presence of mechanisms restricting dispersal of planktonic larvae (believed the dominant type of gene flow for Dreissena) seen in other large river systems (Stoeckel et al. 1997, Schneider et al. 2003). In addition to the restrictions on planktonic larvae there are also a number of hydrological mechanisms that can lead to local retention and isolation of animals in large rivers (Reckendorfer et al. 1999, Reynolds 2000, Speirs and Gurney 2001Carr et al. 2004, Casper and Thorp 2007). There are three possible post-settlement mechanisms that can lead to this pattern (sensu Schluter 2000); (1) development of ecotypes through irreversible physiological change during development; (2) single generation selection for specific genotypes in divergent environments; and (3) sequential selection within a closed populations leading to local adaptation.Either working separately or in concert, pre- and post-settlement mechanisms provide a physical template capable of restricting gene flow between populations. Our results imply physiological differentiation between the two populations and sites that could either be due to plasticity in performance, albeit with a possible genetic component (sensu Schluter 2000).
Determination of the size of the genetic component of site differences in growth would normally require a common garden experimental design with the backcrossing of the F1and F2generations, something logistically difficult because of the complex life cycle of the mussel. Despite this limitation, the reciprocal transplant at a minimum can indicate whether or not a trait or suite of traits is plastic or fixed with the potential of fixed to a geographic location. While two populations displaying non-reversible geographic differentiation are not necessarily showing local genetic adaptation (i.e. the source-destination interaction was not significant for either parameter), the significant source effects indicate that these differences are not simply due to the environment influence of the destination alone. By also using both shell and soft tissue growth measurements in the context of the reciprocal transplant manipulation, evidence for the specificity of physiological constraint of the effect of genetic and environmental factors on phenotype can be examined (i.e. does it impact the general biology of the animal or just limit a particular aspect).
This experiment indicates that shellgrowth in dreissenid mussels can be a population specific-trait with a significant geographically fixed component within a contiguous population. Additionally, examining the responses of shell and tissue physiology in relation to spatial heterogeneity shows that using shell growth can be an inaccurate and misleading proxy for changes in tissue mass. Specifically in this study, shell growth by itself suggests the fluvial estuary site is a poor growth environment while the soft tissue-specific measures indicate that elevated cellular activity and greater biomass. Thus we conclude that if D. polymorpha can tolerate or adjust to conditions in the fluvial estuary, then they benefit by maintaining higher tissue mass loss over the summer than mussels associated with the riverine environments, at least in the St. Lawrence system.
Table 4.1. ANCOVA of the shell and body growth, final body mass, and RNA/DNA ratio of Dreissena polymorpha with an adjusted length of 25.7 mm-1 (model factors are mussel source location, transplant destination, and source-destination interaction, with initial shell length as covariable, * = p < 0.05).
|
|
Table 4.2. ANCOVA of shell and soft tissue growth, soft tissue % organic content, and RNA/DNA ratio for Dreissena bugensis with an adjusted shell length of 17.6 + 0.23 (standard error) from the river site (models are transplant destination with initial shell length as a covariate, * = p < 0.05).
|
|
Table 4.3. Comparison of growthrates based on shell length for Dreissena polymorpha and D. bugensis from the St. Lawrence River with populations from other systems (SG = shell growth, z = D. polymorpha, Q = D. bugensis, LE = Lake Erie source population, NA = not available, IL = initial length).
|
|
Figure 4.1. Map of the St. Lawrence River and fluvial estuary.The two source populations of Dreissena spp. are indicated by an R (the river site near Massena, NY) and an FE (the fluvial estuary site at the eastern tip of Ile d’Orléans, QC).
Figure 4.2. Comparison of temperature profiles between the transplant locations during the 58-day experiment.The river site is at Massena NY and the estuary site is at the eastern tip of Île d’Orléans, QC.
Figure 4.3. Change in shell length, change in tissue mass, and final tissue mass (LS means and standard errors) ofstandardized to 26 mm)and D. bugensis (standardized to 18 mm) during the course of the 58 day summer transplant experiment. Capital letters indicate Tukey-Kramer post-hoc differences among D. bugensis transplants.Lower case letters indicate Tukey-Kramer post-hoc differences among D. bugensis transplants.Note that there were no quagga mussels collected at the FE site. R = river site near Massena, NY; FE = fluvial estuary site at the eastern tip of Île d’Orléans, QC.
Figure 4.4. Ratio of RNA to DNA for Dreissena polymorpha (standardized to 26 mm)and D. bugensis (standardized to 18 mm). Capital letters indicate Tukey-Kramer post-hoc differences among D. bugensis transplants.Lower case letters indicate Tukey-Kramer post-hoc differences among D. bugensis transplants.Note that there were no quagga mussels collected at the FE site. R = river site near Massena, NY; FE = fluvial estuary site at the eastern tip of Île d’Orléans, QC.
© Andrew F Casper, 2007
| | | |
| Chapitre III. Life at the edge: Physiological constraints on freshwater mussels (Dreissena polymorpha Pallas and D. bugensis Andrusov) in a fluvial estuary |  | Chapitre V. Significant inter-annual variation in seasonal-spatial differences in growth among populations of Dreissena polymorpha (Pallas) and D. bugensis (Andrusov) |
