Seasonal control of particle clearance by isolated gills from the clam Mercenaria mercenaria Free

Most bivalve molluscs are filter feeders, so their gills are equipped with a suite of effectors, including mucocytes, several ciliary tracts and a complex musculature, that work together to serve two functions, feeding and respiration. For nearly a decade, we have been studying the lateral cilia and the branchial musculature of the heterodont clam Mercenaria mercenaria (Veneridae), focusing especially on their innervation,pharmacology and remarkable seasonality(Gainey et al., 1999; Gainey et al., 2003; Gainey and Greenberg, 2003; Gainey and Greenberg, 2005). Recently, I discovered that isolated quarters of a gill not only clear colloidal graphite from suspension but also transport the cleared particles in a manner analogous to that in feeding – a complex behavior approaching that of intact gills. The rate at which isolated gill preparations clear graphite can be measured and so the pharmacology and seasonality of clearance can be determined. We can then ask whether the seasonal control of clearance,an integrated function, can be predicted from information about two critical components of the system – the lateral cilia and branchial musculature.

The eulamellibranch gills of Mercenaria are plicate(Kellogg, 1892). That is,within each demibranch, adjacent filaments are connected to each other by interfilament tissue junctions, while the ascending and descending limbs of some of the filaments are also connected by interlamellar tissue junctions(the septa), an arrangement that produces the plicae and the water tubes(Fig. 1).

Particle clearance and transport by the gills, i.e. feeding, results from the coordinated activity of three distinct groups of cilia: (1) the lateral cilia, which create the water currents(Purchon, 1968; Morton, 1983); (2) the latero-frontal cirri, which capture particles(Beninger et al., 1997; Silverman et al., 1996; Silverman et al., 1999)– although the exact role of these cirri in particle capture is controversial (Beninger, 2000; Riisgard and Larsen, 2000; Silverman et al., 2000; Ward et al., 2000) – and(3) the frontal cilia, which move particles to the food grooves(Purchon, 1968; Morton, 1983; Murakami, 1989).

Clearance by the gill, however, is determined not only by the coordinated activity of the cilia but also by the gill's geometry, which includes the spacing of the filaments, the shape of the plicae and the cross-sectional area of the water tubes. These geometric properties are controlled by the branchial muscles: contraction of the longitudinal muscles narrows the spacing between adjacent filaments whereas contraction of the water tube muscles not only narrows the spacing but constricts the water tubes and changes the shape of the plicae (Gainey et al.,2003). In an intact clam, of course, the clearance rate is also related to the degree of valve gape and the cross-sectional area of the siphons (Jorgensen and Riisgard,1988; Riisgard et al.,2003).

In Mercenaria, the lateral cilia are stimulated by 5-hydroxytryptamine (5-HT) and inhibited by dopamine (DA), while the frontal cilia are inhibited, but not completely stopped, by 5-HT(Gainey et al., 1999). The branchial musculature, by contrast, is stimulated by 5-HT, DA and acetylcholine (ACh) and relaxed by ACh(Gainey et al., 2003). Moreover, nitric oxide (NO) potentiates the stimulatory effects of 5-HT, DA and ACh on the gill muscles from November through June but has no effect from July through October (the off-season). Indeed, the entire NO signaling cascade is absent from the branchial musculature during the off-season(Gainey and Greenberg,2003).

Here, I report on the measurement of clearance rates by isolated pieces of gill from Mercenaria mercenaria and the control of clearance by 5-HT and DA. I presumed that the effects of 5-HT and DA on clearance should reflect their effects on the lateral cilia and musculature. Thus, I hypothesized that at low doses 5-HT should increase clearance rates by stimulating the lateral cilia and should decrease rates at higher doses by contracting the gill musculature, thus reducing the space between the filaments. In addition, the effects of 5-HT should be more pronounced during the winter than the summer due to the seasonal expression of the NO signaling cascade. DA should stop clearance, not only by inhibiting the lateral cilia but also by contracting the gill musculature. Finally, because the size and shape of the water tubes are an indicator of the pumping activity of eulamellibranch gills, I made morphometric measurements of the cross-sectional area of water tubes and also the interfilament space in response to 5-HT. Preliminary results of these data were presented to the Society of Integrative and Comparative Biology(Gainey and Greenberg,2007).

Specimens of Mercenaria mercenaria L. from various locations along the northeast Atlantic coast (mostly Cape Cod) were purchased from Hannaford Supermarkets in Portland, ME, USA. The clams were held at 10°C on a 12 h:12 h light:dark cycle in 30‰ natural seawater from Casco Bay, ME,USA. Before being used in experiments, the clams were held for at least 3 days but not more than 2 weeks. During this period, the clams were fed `Coral and Clam Diet' (Reed Mariculture, Inc., Campbell, CA, USA) every other day.

The two gills were dissected away from the body wall distal to the visceral ganglion, and the branchial nerves removed. Each gill was split in two along its dorsal cut edge to produce four demibranchs per clam. Each demibranch was then cut into anterior and posterior halves, and a loop of thread was tied to the free (uncut) end of the ventral margin. Finally, the isolated pieces of gill were placed in a dish of artificial seawater [ASW; 1000 mosmol kg^–1, 10°C; made according to the MBL (Woods Hole) recipe(http://www.mbl.edu/BiologicalBulletin/COMPENDIUM/Comp-ASW.html)].

Aquadag (Acheson Colloids, Port Huron, MI, USA), a paste of graphite particles with a median particle diameter of 0.8 μm, was made up as a 4%colloidal suspension in distilled water; 4.7 ml of this stock was then diluted with ASW to a final volume of 1 liter and stored at 10°C. This suspension was aerated for at least 5 min before each clearance experiment both to mix and aerate it.

To measure the clearance of colloidal graphite, pieces of gill were hung from aluminum hangers by a loop of thread in glass vials (4 cm high × 2 cm diameter) containing 10 ml of the ASW/Aquadag suspension. Samples (0.5 ml)were removed with a pipettor at the beginning of the experiment, and every 3 min thereafter for 15 min; the pipettor tip was placed 2 cm into the vial adjacent to the piece of gill. The relative concentration of graphite in the solution was determined with a spectrophotometer set to a wavelength of 500 nm. Four or six pieces of gill were used in each experiment. Two experiments,each with four vials lacking gills, were performed to measure the rate of spontaneous settling of the graphite. Over the 15 min duration of the experiments, no settling in these vials could be measured.

Clearance rates were calculated by regressing the ln(fractional initial absorbance) vs time (min). The rate constants – the slopes of the regression lines – were estimated with either SAS JMP v.5.1 (SAS Institute, Cary, NC, USA) or Systat v.11 (Systat Software, Inc., San Jose, CA,USA). Rate constants with a 1-tailed P value of >0.05 were considered statistically insignificant and entered as 0. Mass-specific clearance rates (ml min^–1 g^–1 dry mass) were calculated by multiplying the rate constants by 8.5 ml (the average volume of fluid in the vials) and dividing by the dry mass of the gills, which was obtained after the tissues had been dried overnight at 60°C.

For these rate determinations, the gills were dissected and the loop of thread attached. The gills were then allowed to recover overnight. Before these measurements, the gills were examined with a microscope to ensure that the lateral cilia were active, the ostia open and the water tubes inflated.

Before exposure to 5-HT, DEANO (diethylamine/nitric oxide complex) or 5-HT+ l-NAME (nitro-l-arginine methyl ester), clearance rates were first determined on untreated pieces of gill to measure the basal clearance rate; these gills had been dissected 1–2 h before use. Before this initial trial, five random patches on each piece of gill were inspected with a microscope to determine if the lateral cilia were active. Pieces of gill with more than two patches of active lateral cilia were rejected; the rationale being that only gills clearing graphite at the basal rate would show maximal stimulation in response to an agonist.

After determining the basal rates, the gills were transferred, on their hangers, to a second vial of ASW containing the appropriate agent for the following times: 5-HT for 5 min, DEANO for 15 min and l-NAME for 20 min. After the allotted time, the gills were transferred to a third vial containing ASW/Aquadag and the appropriate agents, and water samples were removed as in the initial run. One or two pieces of gill were left untreated for the second run and served as untreated controls.

A different repeated-measures design was used to measure the inhibitory effects of DA. Rather than a basal rate determination, pieces of gill were first exposed to 10^–5 mol l^–1 5-HT to ensure that they were clearing graphite at their maximal rate. After this maximal rate determination, the pieces of gill were suspended for 5 min in a second vial containing only ASW and DA and were then transferred to a third vial containing ASW/Aquadag and DA for the treatment rate measurement. To retard the oxidation of DA, an ascorbic acid buffer (see Malanga, 1975) was added to the ASW. In these experiments, data are expressed as the ratio of treatment rate/maximal rate. Regression lines were fitted to these data with a logistic function of the form:response=1/[1+exp(β₀+β₁×log DA)], whereβ ₀ and β₁ are intercept and slope parameters,respectively. The concentrations of agonist giving half-maximal responses(EC₅₀) were estimated according to the following formula:EC₅₀=10^(–β₀/β₁). These non-linear regressions were done in Systat v.11.

The clearance rates are highly skewed, having a distribution approximating an exponential. Accordingly, the data were ln-transformed to give a normal distribution (Kolmogorov–Smirnov one-sample test 2-tailed P=0.17, N=367). Because clearance rates of 0 are undefined by this transformation, 1 was added to all basal clearance rates for statistical testing. Not surprisingly, the ratios of the treatment rate to the basal rates were also skewed and were ln-transformed to achieve normality(Kolmogorov–Smirnov one-sample test 2-tailed P=0.67, N=159). All statistical tests were performed on ln-transformed data,but the results are presented untransformed in figures and tables for clarity.

Pieces of gill were dissected and pinned to the bottom of Petri dishes (50 mm diameter), the bottoms of which were coated with Sylgard® (Dow Chemical, Midland, MI, USA); the dishes contained 5 ml of ASW. To measure the cross-sectional area of the water tubes, small pieces of gill (about 1 mm high and 5–10 mm long) were cut perpendicular to the plane of the gill filament so that the cut open ends of the water tubes were visible from above. To measure the interfilament spaces, an entire demibranch was pinned flat upon the bottom of the dish, with the outer face of the demibranch facing up. These gill preparations were allowed to relax for 60 min and were then exposed to increasing concentrations of 5-HT. After the addition of 5-HT, the solution was mixed with a Pasteur pipette, and then 5 min later the gills were photographed with a Spot RT color CCD camera (Diagnostic Images, Sterling Heights, MI, USA) at a magnification of 50× for the measurements of water tube areas or 100× for the measurements of gill filament spacing. After another 5 min, 5-HT was again added to the gills to make the next higher concentration, and the process was repeated. The cross-sectional areas of the water tubes and the interfilament spaces were measured with Spot Advanced image analysis software (v.3.5; Diagnostic Images).

When isolated pieces of gill were hung in a suspension of colloidal graphite in ASW, the graphite was removed from the suspension. The graphite,embedded in mucus, moved down the outer face of the filaments to the ventral food groove. Depending upon the orientation of the piece of gill, the graphite–mucus `thread' either accumulated in the food groove and on the loop of thread (anterior piece of gill) or it fell off the piece of gill and accumulated on the bottom of the vial (posterior piece of gill; Fig. 2).

If gills were allowed to recover overnight, the lateral cilia were active,the ostia were open and the water tubes inflated. The mean mass-specific clearance rate of these untreated gills was 18.42±4.02 ml min^–1 g^–1 dry mass (s.e.m.; N=24). These measurements were made in October.

The mean mass-specific clearance rate of untreated pieces of gill was 14.51±0.79 ml min^–1 g^–1 dry mass(s.e.m.; N=589 pieces of gill from 144 clams). Because the lateral cilia on these pieces were largely inactive, this basal clearance rate is due mostly to the activity of the latero-frontal and frontal cilia. A nested analysis of variance (ANOVA), with individual clam nested within month as an error term, revealed no significant differences in monthly basal clearance rates (P=0.22).

Irrespective of the season, 5-HT had a biphasic effect upon clearance rates: low doses stimulated clearance rates, whereas higher doses decreased the maximal rate. During the winter, the threshold for 5-HT stimulation was 10^–6 mol l^–1; the response peaked at 10^–5 mol l^–1 and then declined, so that doses of 5×10^–5 and 10^–4 mol l^–1 had no net effect upon clearance(Fig. 3). At 10^–5 mol l^–1 5-HT, the mean clearance rate increased by 8.8 to 31.82 ml min^–1 g^–1 dry mass (s.e.m.=7.09; N=10). During the summer, the gills were less responsive to 5-HT and the threshold increased from 10^–6 mol l^–1 to 5×10^–6 mol l^–1(Fig. 3). Although the maximal stimulation was still at 10^–5 mol l^–1 5-HT,the mean clearance rate increased by only 2.3 to 18.49 ml min^–1 g^–1 dry mass (s.e.m.=3.81; N=14). This increase is significantly lower than that during the winter (1-tailed P=0.02; N=24).

On untreated pieces of gill, the lateral cilia were inactive, but both the latero-frontal cirri and frontal cilia were active. The mean interfilament space on these pieces was 20.4 μm (Fig. 4). After exposure to 10^–5 mol l^–1 5-HT, the lateral cilia became active and there was a 23%decrease in the interfilament space. Exposure to 5×10^–5mol l^–1 5-HT stopped the lateral cilia and there was a 76%decrease in the interfilament space. Moreover, the gill filaments were buckled and abutting each other in patches on these gills. ANOVA revealed that the mean interfilament spaces are significantly different (P=0.001); a post-hoc Tukey Kramer test revealed that all of the means are different from each other.

The mean cross-sectional area of the water tubes of untreated gills was 68 139±21 629 μm² (s.e.m.; N=2 measurements, each on six pieces of gill). This value increased by 210% at 10^–5mol l^–1 5-HT and then decreased back to the original size at 5×10^–5 mol l^–1 5-HT(Fig. 5). ANOVA revealed that the means are significantly different (P=0.038); a post-hocTukey Kramer test revealed that the mean percent original area at 10^–5 mol l^–1 5-HT is significantly greater than that at 5×10^–5 mol l^–1 5-HT.

During the winter, NO (generated by 10^–6 mol l^–1 DEANO) tripled the basal clearance rate, but during the summer, NO only doubled this rate. The clearance rates of untreated controls increased by a factor of 1.2. ANOVA revealed that the means of rate ratios(i.e. treatment rate/basal rate) are significantly different(P=0.049; N=74); post-hoc analysis of these data with Dunnett's test revealed that the rate ratios of the gills during the winter are significantly greater than the untreated controls(Fig. 6). Thus, NO stimulated clearance during the winter but had no significant effect during the summer.

During the winter, pretreating pieces of gill with 10^–5mol l^–1l-NAME inhibited the effects of 10^–5 mol l^–1 5-HT (1-tailed P=0.03; N=17; Fig. 7). In addition, ANOVA revealed that the clearance rates of pieces of gills exposed to l-NAME and 5-HT during the winter were equal to the rates of pieces of gill exposed only to 5-HT during the summer (P=0.98; N=113). During the summer, l-NAME had no effect upon the clearance rates of pieces of gill treated with 10^–5 mol l^–1 5-HT (1-tailed P=0.29; N=18).

To study the supposed inhibitory effect of DA on clearance, pieces of gill were first stimulated with 10^–5 mol l^–15-HT, producing a maximal clearance rate. Then the pieces of gills were exposed to DA. The dose–response curve of the pieces of gill during the winter (Fig. 8, dashed line)lies to the left of that for the summer(Fig. 8, solid line). These regression lines are significantly different (P<0.001; Fig. 8). The EC₅₀for the winter is 7.7×10^–8 mol l^–1 DA(95% CI=7.6×10^–8 to 2.3× 10^–7mol l^–1) whereas that for the summer is 9.3×10^–6 mol l^–1 (95%CI=8.4×10^–6 to 1.0×10^–5 mol l^–1).

During both seasons, the mean clearance rates at higher concentrations of DA are statistically greater than 0 (1-tailed P<0.001 for both seasons). Microscopic inspection of these gills revealed that graphite clearance was the result of the activity of the frontal cilia. The gill filaments in these gills were abutting each other in many places and neither the lateral cilia nor the latero-frontal cirri were beating.

Pieces of Mercenaria mercenaria gill clear colloidal graphite, so the control of clearance rates can be studied in a preparation much less complex than an intact clam. Clearance rates are modulated by 5-HT and DA,both of which are present endogenously in the gills of Mercenaria(Gainey et al., 2003). The responses to both agents are markedly seasonal, thus the gills are more sensitive to 5-HT and DA during the winter than during the summer. Additionally, the seasonal effects of 5-HT are modulated by an NO signaling cascade, which is present during the winter but not the summer. Finally, the results show that the rate of graphite clearance is, in part, the result of the interaction between the activity of the lateral cilia and the gill musculature. Thus, the behavior of the gill is predictable from the behavior of its components.

The effect of DA upon clearance is completely inhibitory, irrespective of the season. This was expected because DA not only inhibits the lateral cilia(Gainey et al., 1999) but also stimulates the gill musculature (Gainey et al., 2003). As was the case for the stimulation of clearance rates with 5-HT, the gills are more sensitive to DA during the winter. The dose–response curves for the effects of DA also reveal another effect– that there is little inhibitory modulation of clearance rates. During summer, clearance rates declined abruptly from maximal to minimal between 5×10^–6 and 10^–5 mol l^–1 DA. The effect is not as pronounced during the winter as the mean clearance rate declined to around 50%, then became minimal. The dose–response curve for the effects of DA on the lateral cilia has a similar shape; the cilia beat maximally and then abruptly stop between 10^–6 and 10^–5 mol l^–1 DA(Gainey et al., 1999).

Irrespective of the season, 5-HT has a biphasic effect upon clearance:concentrations lower than 10^–5 mol l^–1 5-HT stimulate clearance, while higher concentrations inhibit it. The biphasic nature of this response readily follows from the effects of 5-HT on the lateral cilia and the gill musculature. 5-HT stimulates quiescent lateral cilia of Mercenaria with a threshold from 10^–7(Aiello, 1970) to 10^–6 mol l^–1(Gainey et al., 1999) and it also stimulates the gill musculature to contract, again with a threshold of 10^–6 mol l^–1(Gainey et al., 2003; Gainey and Greenberg, 2003). But between 10^–6 and 10^–5 mol l^–1 5-HT, the beating of the lateral cilia is stronger than the contraction of the gill muscles, and microscopic observation of the water tubes shows that their cross-sectional area increases, i.e. the water tubes become inflated when pumping begins. Similar observations on the relationship between the activity of the lateral cilia and inflation of the water tubes have been made in vitro on excised gills of Juxtamusium maldivense (Pectinidae) (Jorgensen,1976) and in vivo, via endoscopy, on the gills of Pyganodon cataracta (Unionidae)(Tankersley, 1996). This behavior is analogous to that of sea anemones that have expelled the water from their coelenteron by contraction of their musculature: reinflation of the coelenteron and extension of the musculature is accomplished by cilia in the siphonoglyphs. At concentrations of 5-HT above 10^–5 mol l^–1, contraction of the gill musculature is strong enough to counteract the beating of the lateral cilia, so the diameter of the water tubes is reduced, as is the interfilament space. Thus, the rate of clearance is reduced due to changes in gill geometry.

The current paper reports that clearance rates in actively pumping gills vary from 5 to 60 ml min^–1 g^–1 mass –a 12-fold range – but the beat of the lateral cilia varies from 7 to 27 Hz – only a ∼4-fold range(Gainey et al., 1999). A limited range in the rate of lateral ciliary beating has also been reported in Mytilus edulis (Catapane et al.,1978). Jorgensen et al.(Jorgensen et al., 1986; Jorgensen et al., 1988),Riisgard and Larsen (Riisgard and Larsen,1995) and Grunbaum et al.(Grunbaum et al., 1998) have shown, in Mytilus edulis, on both experimental and theoretical grounds, that the width of the interfilament space is the primary factor regulating flow through a fillibranch gill. In addition, Silverman and colleagues (Gardiner et al.,1991; Medler and Silverman,1997; Medler and Silverman,2001) proposed that the gill musculature regulates flow in eulamellibranch gills. The striking difference between the rates of lateral ciliary beat and clearance in the Mercenaria gill, when taken together with the changes in interfilament space and water tube cross-sectional area (also reported here), support this hypothesis.

Clearance is regulated in isolated gills of Mercenaria by endogenous transmitters, implying that clearance is a regulated and not an autonomous process. That clearance is regulated in response to environmental variables, such as seston concentration and oxygen tension, has been championed by Bayne and colleagues (Bayne et al., 1993; Bayne,1993; Bayne, 1998; Bayne, 1999; Bayne, 2000; Hawkins et al., 1999; Cranford and Hill, 1999; Cranford, 2001; Widdows, 2001). The opposing view, i.e. that clearance is an autonomous process and is partially dependent upon the physical properties of seawater, has been championed by Jorgensen and Riisgard (Jorgensen et al.,1986; Jorgensen et al.,1988; Jorgensen et al., 1990; Jorgensen, 1990; Jorgensen, 1996; Riisgard, 2001a; Riisgard, 2001b; Riisgard et al., 2003). If allowed to recover overnight from dissection, unstimulated gills – in October – had clearance rates that are statistically indistinguishable from gills in the summer that had been stimulated by 5-HT. This implies that clearance is an autonomous process. But the regulation of clearance over a 12-fold range by 5-HT implies that clearance is a regulated process. Taken together, these two sets of seemingly contradictory observations suggest that the isolated, unstimulated gill clears at the maximal summer rate and that the visceral ganglion – removed in the preparations used for this study– regulates clearance by inhibition of the maximal rate.

Although NO stimulates particle clearance during the winter, it has no such effect during the summer. Furthermore, the nitric oxide synthase (NOS)inhibitor l-NAME reduces the effects of 5-HT during the winter, and this effect is statistically indistinguishable from the effects of 5-HT during the summer, when NO signaling has no effect on clearance. These results are consistent with an earlier study of the immunohistochemical localization of NOS and soluble guanylate cyclase (sGC) in the gills(Gainey and Greenberg, 2003). We reported that, during the autumn, NOS is found in the gill musculature,which penetrates each gill filament, but during the summer, NOS is confined to the base of the gill filaments in the interfilament tissue junctions. Similarly, sGC is localized in the muscles and the gill filament epithelium during the winter but is restricted to the tips of the filaments in the area of the frontal cilia during the summer. On the other hand, when whole gills were loaded with the NO-fluorescent probe DAF-diacetate(4,5-diaminofluorescein-diacetate), microscopic observation showed NO pervading the gills in both seasons (L.F.G., unpublished). However, the absence of sGC from the area of the lateral cilia and muscle is consistent with the lack of an effect of NO on particle clearance during the summer. Although the gills synthesize NO during the summer, this substance has no effect upon particle clearance, because sGC, a downstream effector in the signaling pathway, is absent. NO has no direct effect upon the gill musculature (Gainey and Greenberg,2003) but during the autumn and winter it stimulates quiescent lateral cilia to beat (L.F.G., unpublished). Thus, the effect of NO on the gills during the winter is twofold: it directly stimulates the lateral cilia and potentiates the effects of 5-HT upon the branchial muscles.

The seasonal difference between the clearance rates of gills is striking. One of the characteristics of animals showing this type of seasonality, i.e. seasonal rate compensation, is that animals acclimatized to winter conditions will have a higher rate than those acclimatized to summer conditions, provided that both are tested at the same temperature(Hochachka and Somero, 1973; Prosser, 1973). The 5-HT-induced clearance rates of Mercenaria gills show such a response. Thus, at 10^–5 mol l^–1 5-HT –the concentration producing maximal clearance – winter gills had a mean clearance rate of 31.82 ml min^–1 g^–1 dry mass whereas summer gills had a mean clearance rate of only 18.49 ml min^–1 g^–1 dry mass. Moreover, winter gills treated with the NOS inhibitor l-NAME – again at 10^–5 mol l^–1 5-HT – had a mean clearance rate of 14.01 ml min^–1 g^–1 dry mass, which is statistically equal to the summer rate. Therefore, if NO production during the winter is inhibited, seasonal rate compensation is abolished. I therefore propose that the NO signaling cascade and its seasonal expression are the mechanisms of seasonal clearance rate compensation in Mercenaria gills, again implying that clearance is a regulated, as opposed to an autonomous, process.

I thank the following students who contributed to this project: Danielle Alley, Kelsey Anthony, John Concannon and Danielle Grondin. I thank my friend and colleague Michael J. Greenberg, of the Whitney Laboratory for Marine Bioscience (University of Florida), St Augustine, FL for his editorial assistance and insights during this project. M. Lynn Milstead, also of the Whitney Laboratory, drew and prepared most of Fig. 1. I also appreciate the comments of the two anonymous reviewers.