Haemolymph Concentration and Apparent Permeability in Varying Salinity Conditions of Gammarus Duebeni, Chaetogammarus Marinus and Gammarus Locusta

Fully marine invertebrate organisms inhabit a stable environment where the external medium is comparable with the internal osmotic concentration of their body fluids, causing few osmotic problems for the cells. Organisms tolerating a variable salinity environment (euryhaline) face the problem of having to vary their regulatory responses as the external medium changes. This problem is heightened in small animals with a relatively large surface area to volume ratio, where small changes in volume can cause large changes in body fluid concentrations. The development of a highly impermeable body surface is impractical for an aquatic animal which respires through part or all of its body surface. However, it would be advantageous for fresh and brackish water animals to reduce their permeability to water and hence limit ion and water fluxes, if this could be done without too great a reduction in respiratory efficiency. Animals experiencing salinity changes would benefit from the ability to restrict the passage of water and ions when large osmotic gradients are present between the body fluids and the external medium. This would necessitate a mechanism controlling the permeability of the body surface in relation to the concentration gradient between the haemolymph and external medium.

Several species of euryhaline Crustacea have been demonstrated to change their apparent permeability to water when the external medium concentration is altered. The euryhaline crabs Rhithropanopeus harrisi (Smith, 1967), Carcinus maenas (Smith, 1970), the arctic amphipod Gammarus setosus (Bolt, 1982) and the euryhaline amphipod Gammarus duebeni (Lockwood & Inman, 1973) all display varying degrees of apparent permeability when acclimated to different salinities.

More recent work on G. duebeni (Bolt, Dawson, Inman & Lockwood, 1980) has shown that this species exhibits large, rapid changes in apparent permeability when exposed to sudden or cycling salinity changes, often associated with periods of salinity when the haemolymph is approaching isosmocity with the external medium. A particular feature is that the half-time of exchange of labelled water decreases dramatically in conditions around isosmocity. The purpose of the present paper is to compare these results with the corresponding changes in apparent permeability in less euryhaline species.

G. duebeni is found in a wide range of salinities from fresh water streams (Hynes, 1950) to 60‰–70‰salt water rock pools in Norway (Davenport, 1979). The present work was carried out on populations from Totton Marsh, Southampton, where the animals are found in small pools in a salt marsh. In this region G. duebeni are found in three localities: (i) in drainage creeks, subject to cyclical salinity changes from 1–22‰ (Lockwood & Inman, 1973); (ii) in low salinity water in fresh water drainage channels where they only encounter saline conditions at high water springs; (iii) in small pools at the extreme high water mark where animals are covered at extreme high water springs and are subjected to extremes of salinity due to evaporation and precipitation. During high water springs the animals from these populations intermix, ensuring that there are no ecologically distinct races on the marsh. In contrast, C. marinus and G. locusta, although occurring well into estuaries (Spooner, 1947), are not generally subjected to long periods of low salinity. The G. locusta and C. marinus used in this paper are found co-existing under Fucus sp. in the middle region of a muddy shoreline on Hayling Island, near Portsmouth. These animals were selected as being less euryhaline than G. duebeni, though still being able to tolerate relatively short periods of reduced salinity.

A microprocessor-controlled cycling salinity system was used for the regulation of external medium salinity (Lockwood et al. 1982). The standard cycle used in the present experimental series was 12 h 25 min, varying from l– 31‰ salinity. The temperature was maintained at 15 °C.

Tritiated water was used to determine water fluxes of the amphipods. Measurements of flux in both directions were made on animals in a steady state. However, only influx can be used in experiments in non-steady state systems. These methods are (modified from Lockwood, Inman & Courtenay (1973). Permeability to water of the amphipods was expressed as a half-time of exchange of tritiated water into or out of the animal. Tritiated water samples were counted in a Liquid Scintillation cocktail in a Beckman Series 3000.

The half-time of exchange of influx is calculated by comparing the tritiated water uptake by an animal in a 5-min loading period with the amount taken up when the animal is fully loaded.

Outflux was measured by loading the animals with tritiated water, transferring them to unloading medium and monitoring the loss of tritiated water to the external medium.

These techniques (influx and outflux) compare counts at time t with counts at equilibrium. This enables counts per minute (c.p.m.) to be used directly without the necessity of finding the efficiency of the counter using quench curves.

Haemolymph was collected using a drawn out Pasteur pipette. It was possible routinely to remove 1–5μl of haemolymph from a 100 mg animal. Aliquots of haemolymph (usually 1 μl) were added to 5–10 ml deionized water and analysed using an emission flame spectrophotometer (Pye Unicam SP900 or SP90). Sodium concentrations were determined from a standard curve.

The net fluxes are calculated by the method of Lockwood et al. (1973). All experiments were carried out at 15 °C.

All three species were kept at 15 °C prior to and during the experiments. G. duebeni and G. locusta varied from 50–100 mg, and C. marinus varied from 50—150 mg.

All the animals were fed on Bemax prior to experiments and starved throughout any period of acclimatization and experimentation.

The haemolymph sodium concentration of G. duebeni, G. locusta and C. marinus, after acclimation to different salinities are compared in Fig. 1 using data from Lockwood (1964) and Haywood (1970). Comparison of haemolymph of the three species exposed to at least eight cycles in the salinity system is given in Fig. 2.

These three species demonstrate a variation of responses in animals exposed to both steady-state and cycling conditions. In steady state conditions, G. duebeni maintains its haemolymph concentration strongly hyperionic at low salinities, while C. marinus and G. locusta are less hyperionic (Fig. 1). This trend is echoed when the animals are exposed to the cycling salinity conditions, where again G. duebeni maintains its haemolymph relatively constant (295 ± 15 mm-Na⁺), C. marinus fluctuates from 350mm-Na⁺ to 450mm-Na⁺ and G. locusta varies from 200mm-Na⁺ to 450 mm-Na⁺. G. locusta and C. marinus remain hyperionic or isionic throughout the cycle, while G. duebeni is hypoionic for a period of approximately 4h during maximum salinity in the cycle.

These three species therefore exhibit a gradation of responses which correspond to their range of habitats and relative mortality in the experimental regime. G. duebeni do not appear to be adversely affected by the cycling salinity, even when maintained for periods extending up to 2 months. C. marinus are less able to maintain a constant haemolymph concentration, which is reflected by an initial mortality of up to 15% when exposed to the salinity cycle. G. locusta is the least able to cope with cycling salinity conditions of this degree of severity and specimens died throughout the experiment. Nevertheless, C. marinus and G. locusta appeared to show a considerable range of individual tolerance, some animals of both species surviving for several weeks in the cycling regime.

G. duebeni exposed to the cycling salinity system exhibit distinct increases in apparent permeability which coincide with the two periods of isoionicity at 4 and 8 h into the cycle (Figs 2, 3). The apparent permeability increases to a t_1/2 of exchange of 10 min and 5 min at the isionic points, while the lowest apparent permeability (t_1/2 of 24 min) occurs when G. duebeni is hypotonic to the medium (Bolt et al. 1980).

C. marinus and G. locusta do not significantly vary their apparent permeability to water during the salinity cycle. C. marinus maintains a t_1/2 of 9 ± 2 min throughout the cycle and G. locusta maintains a t_1/2 of 4 ± 1 min.

Apparent permeability and haemolymph sodium concentrations were measured in the three species before and after sudden changes in salinity of the external medium. Animals transferred from 2% sea water to 100% sea water are initially forced hypotonic to the external medium. In G. duebeni the haemolymph concentration then slowly increases over a period of approximately 16 h until the animal becomes isotonic and finally slightly hypertonic to the external medium (Fig. 4). During this period the apparent permeability decreases slightly (increase in t_1/2 of exchange) after the initial transference to 100% sea water. The change is from approximately 18 min to 21 min and the latter value is then monitored up to the isoionic point, where there is a sudden increase in apparent permeability (t_1/2 drops to 5 min). Thus G. duebeni appears to be least permeable during periods of hypotonicity when the osmotic bulk flow is out of the animal.

In contrast, C. marinus and G. locusta acclimated to 10% sea water and transferred to 100% sea water do not exhibit significant changes in apparent permeability associated with hypotonicity. The t_1/2 of exchange in C. marinus changes from 9 ·0 to 6 ·1 min, and the t_1/2 of G. locusta shows a small decrease from 4 ·21 to 3 ·2 min (Table 1). Thus C. marinus and G. locusta do not exhibit large changes in apparent permeability even when they are forced hypotonic to the external medium.

If the half-time of exchange of water, the osmotic concentration of the haemolymph and the osmotic concentration of the medium are all known, it is possible to calculate the net fluxes in the animal. When the haemolymph is hypertonic to the external medium, the bulk flow of water is into the animal and if the haemolymph is hypotonic, then the flow is outwards.

The difference between the three species is further emphasized by calculations of the percentage water uptake or loss of body water over the period of one cycle (= 12h). This is achieved by integrating the area under the curve in Fig. 5 for each animal. G. duebeni has a% body water uptake of +15% during the cycle, C. marinus +31 ·9% and G. locusta +60 ·6% body water per cycle. It is assumed that these osmotic fluxes into the animal are matched by urine flow out of the animal if the volume of the animal is to remain constant.

A more detailed study of the apparent permeability of G. duebeni during a period of large permeability change in the cycling salinity system demonstrates that the change in t_1/2 is extremely rapid (Fig. 6). Apparent permeability and blood sodium were measured over the 6-to 10-h period of the cycle. During this period, the animals were changing from being hypotonic to hypertonic to the external medium, and accompanying this transition there was a rapid change in permeability. This is manifested by a rapid transition from high t_1/2 (low apparent permeability) to a low t_1/2 as the animals gradually approach isionicity from hypoionicity and a gradual return to a high t_1/2 (low apparent permeability) as the degree of hypertonicity increases. The t_1/2 drops rapidly from 16 min to 4 min, subsequently returning gradually to 13 min.

The t_1/2 of exchange appears to be proportional to the gradient between the haemolymph and external medium, when the animals are hypertonic to the medium (Fig. 7) (linear regression gives a correlation coefficient of 0 ·993). During hypotonicity the t_1/2 is not proportional to the gradient.

Examination of the water fluxes in Gammarus duebeni, Chaetogammarus marinus and Gammarus locusta has shown that, of these three species, only G. duebeni demonstrates a major change in apparent permeability as the salinity of the medium is varied. Detailed study of these changes shows that in all cases investigated, the change in apparent permeability is extremely rapid when the animal is in transition from being hypotonic to isotonic. Conversely if the animal is in transition from isotonicity to hypertonicity, the t_1/2 of exchange appears to be correlated directly to the gradient between haemolymph and external medium. These observations appear to hold in cycling salinity conditions (Fig. 6), steady state condition (Fig. 1) and in non steady state experiments (Fig. 4). In animals transferred from 2% sea water to 100% sea water (Fig. 4) the change of apparent permeability from a t_1/2 of 21 min to a t_1/2 of 6 min appears extremely rapid, even though the change in the gradient from haemolymph to external medium is relatively slow. Again, this rapid change in apparent permeability occurs as the animal is approaching isosmocity with the external medium. Unfortunately, experimental techniques preclude the measurement of apparent permeability and blood sodium simultaneously and it is thus impossible to know the ionic concentration and t_1/2 in an individual animal.

These results were obtained using THO as a marker to measure water exchange rates. Until such changes in water permeability have been demonstrated by a different approach, the possibility that the permeability change is an artefact of the experimental method must linger. Smith (1967) accepted this limitation and noted the necessity of using the term ‘apparent’ when discussing permeability changes. However, studies using ^5lCr EDTA to monitor urine flow rates during apparent permeability changes have largely removed this doubt (S. R. L. Bolt, in preparation; Bolt, 1982).

It is thus proposed that the ability of G. duebeni to become less permeable in conditions of high water fluxes guards the animal against physiologically embarrassing conditions. This is envisaged as being especially important when the net flux is out of the animal during periods when the body fluids are hypotonic to the medium. This corresponds with the rapid changes in t_1/2 as the animal goes from being isotonic to hypotonic to the medium and vice versa.

The above conclusions are not applicable in C. marinus and G. locusta, neither of which inhabit such euryhaline conditions as those occupied by G. duebeni. It would be interesting to know whether other highly euryhaline forms adopt similar physiological responses as G. duebeni. The isopod Sphaeroma rugicauda co-exists with Gammarus duebeni in the salt marsh pools in Totton, Southampton (Harris, 1967). These animals are of similar size and experience identical osmotic conditions as G. duebeni. Under experimental conditions, S. rugicauda shows an extremely large individual variation of water permeability due to the animals’ behavioural response of rolling up into a tight ball when disturbed (S. R. L. Bolt, unpublished observations). Further work is needed to overcome this problem and to determine if a crustacean of the Order Isopoda living in the same conditions as G. duebeni has evolved a similar response to osmotic stress as G. duebeni.

The three species, G. duebeni, C. marinus and G. locusta clearly show a variation in response to a varying external salinity which can be correlated to their ecological tolerance to osmotically stressful conditions.

G. duebeni inhabits shallow salt marsh pools where salinity change can be unpredictable and extreme. Furthermore, the salinity in the pools can be extremely low (2 ) for protracted periods of up to 14 days. These osmotically demanding conditions necessitate an effective osmoregulatory mechanism. In contrast C. marinus and G. locusta only experience reduced salinity for relatively short periods due to fresh water run off and precipitation at low tides and there is correspondingly less selection pressure for the presence of a mechanism to cope with protracted low salinities. It is thus proposed that the ability to vary apparent permeability is an important facet of osmoregulation, which is advantageous to a species which successfully survives in osmotically stressful conditions.

This work was supported by a Natural Environment Research Council Grant No. GT4/78/ALS/29. I thank Mr M. E. Dawson with whom much of the early work on water permeability was carried out. I also thank Mr N. W. Jenkinson for technical support with the microprocessor and Professor A. P. M. Lockwood for his helpful advice.