Regulation of Compatible Solute Accumulation in Larvae of the Mosquito Culex Tarsalis: Osmolarity VERSUS SALINITY

In a companion paper (Patrick and Bradley, 2000), we determined some of the physiological traits that are associated with increased salinity tolerance in larvae of mosquitoes of the genus Culex. Larvae of C. tarsalis were able to regulate body volume by drinking and to attenuate increases in hemolymph Na⁺ concentration during acute hyperosmotic shock. During acclimation to high salinity (640 mosmol kg⁻¹ sea water), this species accumulated high levels of the disaccharide trehalose and the amino acid proline in the hemolymph, thereby increasing the osmolality of the body fluids to conform to that of the environment. In addition, we discovered that proline was serving as both an intra- and extracellular compatible solute in this species; this is the first demonstration of the same compatible solute being accumulated both intra- and extracellularly in an animal.

The utilization of trehalose and proline as osmolytes raises questions about how levels of these solutes are regulated during the osmoconforming response. Trehalose is considered to be a storage compound found only in the hemolymph and synthesized in the fat body from two glucose molecules (Clements, 1992). Proline synthesis is thought to occur primarily in the fat body, with the main pathway of its production stemming from intermediates of the tricarboxylic acid pathway (e.g. α- ketoglutarate). This amino acid is utilized in protein synthesis and is an important flight fuel in adult insects, in which it functions to augment the Krebs cycle (Hochachka and Somero, 1984; Chapman, 1998). Proline and trehalose are solutes that are used in energy metabolism, have a low molecular mass and possess chemical properties that are not disruptive of enzymatic structure or function when these solutes are accumulated in large quantities. Therefore, it is not surprising that a diversity of organisms from most of the taxonomic kingdoms utilize trehalose or proline as compatible solutes (for a review, see Somero and Yancey, 1997). However in most multicellular organisms, both osmolytes are accumulated intracellularly, whereas C. tarsalis larvae employ them as extracellular or, in the case of proline, as both intra- and extracellular, osmotic effectors.

We characterized the regulation of these two compatible solutes in the larvae of C. tarsalis. In this study, we examine the response of hemolymph proline and trehalose concentrations to changes in the osmolality of the external medium as controlled by the addition of sea salts (predominantly NaCl) or the polyol sorbitol. In doing so, we wished to determine (i) whether the accumulation of proline and trehalose is being cued by changes in salinity or osmolality and (ii) whether the larvae are responding to changes in their environment or to more proximal signals, such as changes in the hemolymph status.

Rearing and holding procedures for Culex tarsalis follow that described by Patrick and Bradley (2000).

Larvae to be used experimentally were reared in Irvine tapwater until instar 3–4. At this stage, batches of C. tarsalis larvae were transferred to smaller trays (18 cm×13 cm×6 cm) containing either 300 ml of 100 mosmol kg⁻¹ seawater medium (prepared by dilution of a concentrated seawater solution made from tapwater and Instant Ocean Salts, Aquarium Systems) or 300 ml of 100 mosmol kg⁻¹ sorbitol medium (prepared from tapwater and sorbitol; Sigma, catalog no. S-1876). A small amount of ground rabbit chow was added to each tray. The osmolality of each experimental medium was increased by 100 mosmol kg⁻¹ every other day until the final osmolality of 600 mosmol kg⁻¹ was reached. The osmotic concentration of the medium was measured using a Wescor vapor pressure osmometer. Separate batches of control larvae were maintained in tapwater, with water additions or changes occurring at the same time as for the experimental groups.

The osmolality of each experimental medium was measured after the larvae had been acclimated for 2 days and was approximately 636±18 mosmol kg⁻¹ (mean ± S.E.M., N=4). This approximately 40 mosmol kg⁻¹ increase in the osmolality of the medium is attributed to evaporation. Hemolymph sampling and Malpighian tubule removal procedures were as described by Patrick and Bradley (2000). Hemolymph samples were assayed for osmotic, ion, free amino acid and trehalose concentrations. Intracellular fluid from the Malpighian tubules was assayed for free amino acid concentrations.

The rearing and holding procedure were the same as described above. Batches of C. tarsalis larvae were then transferred from freshwater holding trays to smaller trays (18 cm×13 cm×6 cm) containing 300 ml of 100 mosmol kg⁻¹ seawater medium (treatment A) or 300 ml of 100 mosmol kg⁻¹ sorbitol medium (treatment B). The osmolality of each experimental medium was increased by 100 mosmol kg⁻¹ every other day until the final osmolality of 450 mosmol kg⁻¹ was reached. After 2 days, larvae were sampled for hemolymph (as described in Patrick and Bradley, 2000) from each treatment group. The remaining larvae in each treatment (A and B) were then transferred to a medium of 450 mosmol kg⁻¹ salinity plus 150 mosmol kg⁻¹ sorbitol, giving a total osmolality of 600 mosmol kg⁻¹. After 2 days in this mixed medium, larval hemolymph was sampled. Separate batches of larvae were maintained in tapwater, with water changes and hemolymph sampling occurring at the same time as for the experimental groups.

Note that all experiments were replicated on separate batches of larvae and on a different date.

To examine the relationship between hemolymph Na⁺ concentrations and hemolymph proline concentrations as external salinity was varied, we held larvae in different salinities for a minimum of 2 days. Data for freshwater larvae and for 34 % and 64 % seawater-acclimated larvae are from the results of the companion paper (Patrick and Bradley, 2000); data for freshwater larvae and for 10 %, 20 %, 45 % and 50 % seawater larvae were collected in the present study. The larvae were sampled as described above, and hemolymph Na⁺ and proline concentrations were determined as described below.

Analytical methods are described in greater detail in the companion paper (Patrick and Bradley, 2000). Hemolymph osmolality was measured using a Clifton freezing point depression nanoliter osmometer. Hemolymph sodium, potassium, calcium and magnesium concentrations were determined using atomic absorption spectrophotometry (Varian, AA275 series). Hemolymph Cl⁻ concentrations were determined spectrophotometrically using a colorometric assay (Gonzalez et al., 1998). Hemolymph trehalose concentrations were measured using a modification of the protocol of Parrou and François (1997) in which trehalose, a disaccharide, is hydrolyzed to its two glucose units and assayed colorimetrically. Levels of free amino acids were assayed in both extracellular (hemolymph) and intracellular (lysed Malpighian tubules) fluid samples. The samples were analyzed for free amino acids at the UCLA Protein Microsequencing Facility using standard Pico-Tag chemistry (Cohen and Strydom, 1988).

We measured hemolymph sorbitol concentrations to determine whether the larvae take up exogenous sorbitol when they are held in the sorbitol medium or whether sorbitol is an osmolyte synthesized by C. tarsalis. Batches of larvae were placed in fresh water and in 200, 400 or 600 mosmol kg⁻¹ sorbitol or 400 mosmol kg⁻¹ sea water. After 2 days, hemolymph was collected from the larvae as described above. Sorbitol was assayed by quantifying the conversion of sorbitol to fructose via sorbitol dehydrogenase (Sigma, catalog no. S1128). In a microplate well, 10 μl of sorbitol standard (0.2–2 mmol l⁻¹) or hemolymph sample (400-fold dilution) was added to a solution of 90 μl of Tris buffer, pH 9, and

1.6 mmol l⁻¹ β-NAD⁺. The initial absorbance of each sample well was measured on a microplate reader at 340 nm. Sorbitol dehydrogenase solution (1 μl; 6.6 units μl⁻¹) was added to each well, and final absorbance readings were taken 2 h later.

All data are reported as means ± S.E.M. Comparisons between groups were performed using analysis of variance (ANOVA) (overall P⩽0.05) with multiple comparisons (Scheffe’s test) if ANOVA proved significant.

When C. tarsalis larvae were held in the 640 mosmol kg⁻¹ seawater medium for 2 days, hemolymph Na⁺ and Cl⁻ concentrations increased significantly by 1.7-fold and 2.8-fold respectively (P<0.0001) (Fig. 1A). There was no change in hemolymph K⁺ concentration. Small but significant decreases in hemolymph concentrations of both Ca²⁺ (P<0.0001) and Mg²⁺ (P<0.0024) were found. In contrast, when larvae were acclimated to 640 mosmol kg⁻¹ sorbitol medium, hemolymph Na⁺ and Cl⁻ levels decreased significantly (P<0.001) (Fig. 1B). Hemolymph K⁺ and Ca²⁺ levels were the same as in seawater medium, while hemolymph Mg²⁺ levels decreased by a small but significant amount (P<0.0012).

Hemolymph proline levels increased almost 50-fold, to 70 mmol l⁻¹, from freshwater values when larvae were held in 640 mosmol kg⁻¹ seawater medium (P<0.0001) (Fig. 2A). Hemolymph concentrations of alanine, threonine, arginine, histidine and glutamate also increased significantly, but all less than 10 mmol l⁻¹ (P<0.001). Hemolymph serine and glycine concentrations showed no significant change (Fig. 2A). In contrast, hemolymph proline concentration increased only slightly (P<0.0001) to approximately 8 mmol l⁻¹ when larvae were held in 640 mosmol kg⁻¹ sorbitol (Fig. 2B). Hemolymph alanine, serine and histidine concentrations increased significantly, and concentrations of the other amino acids remained unchanged (P<0.01) (Fig. 2B). Intracellular proline levels changed in concert with hemolymph proline levels during high-salinity and high-sorbitol treatments. Levels of the remaining intracellular amino acids did not vary in an osmotically significant manner (data not shown).

Hemolymph trehalose levels increased significantly, over twofold, in larvae transferred from fresh water to the 640 mosmol kg⁻¹ seawater medium (P<0.0011) (Fig. 3). Similar results were obtained in larvae transferred to high- sorbitol medium (P<0.0001). There was no significant difference between the hemolymph trehalose responses in the high-salinity and high-sorbitol treatment groups (Fig. 3).

In both the 640 mosmol kg⁻¹ sorbitol and seawater treatments, hemolymph osmolality approximated the osmolality of the external medium (sorbitol group 627±13 mosmol kg⁻¹, N=7; seawater group 635±8 mosmol kg⁻¹, N=8)

This experiment involved two treatment groups (A, B) in which the larvae experienced similar increases in external osmolality, but the external solute used differed. In treatment A, external osmolality was increased using sea water, in treatment B using sorbitol. After 2 days, some animals were removed from both media to characterize hemolymph composition. Both treatment groups of larvae were then placed in an identical final medium composed of 450 mosmol kg⁻¹ sea water plus 150 mosmol kg⁻¹ sorbitol. This experimental design allowed us to investigate the signal for proline and trehalose accumulation.

In treatment A (Fig. 4A), when larvae were held in 450 mosmol kg⁻¹ sea water for 2 days, hemolymph levels of both Na⁺ and Cl⁻ increased significantly compared with freshwater conditions (increases of 1.4- and twofold, respectively, P<0.0002) (Fig. 4A). Following transfer from 450 mosmol kg⁻¹ sea water to the final mixed medium (450 mosmol kg⁻¹ sea water plus 150 mosmol kg⁻¹ sorbitol), hemolymph concentrations of both ions decreased significantly from the values observed in 450 mosmol kg⁻¹ sea water (P<0.0003) but were not significantly different from values obtained from larvae acclimated to fresh water (Fig. 4A). Hemolymph K⁺ levels did not vary throughout the treatment regime (Fig. 4A). Hemolymph Ca²⁺ levels decreased when larvae were transferred from 450 mosmol kg⁻¹ sea water to the final seawater+sorbitol mixed medium and were significantly lower than those in larvae acclimated to fresh water (P<0.0095) or 450 mosmol kg⁻¹ sea water (P<0.0038). Hemolymph Mg²⁺ concentrations decreased significantly (P<0.0162) in 450 mosmol kg⁻¹ sea water and remained at this level in the final mixed medium (P<0.0029) (Fig. 4A).

In larvae for which 450 mosmol kg⁻¹ sorbitol (treatment B) was the intermediate medium, the decline in hemolymph levels of Na⁺ and Cl⁻ (Fig. 4B) was significant relative to values in fresh water (P<0.0032, P<0.031, respectively). In the final mixed medium (450 mosmol kg⁻¹ sea water plus 150 mosmol kg⁻¹ sorbitol), hemolymph Na⁺ and Cl⁻ levels increased significantly from the 450 mosmol kg⁻¹ sorbitol values (P<0.0002). Cl⁻ levels were also significantly higher than freshwater values (P<0.01). There was no significant change in hemolymph K⁺, Ca²⁺ and Mg²⁺ concentrations throughout the second experiment (Fig. 4B).

Proline was the hemolymph amino acid that showed the greatest response during the two treatments (A, B) (Fig. 5). Proline levels increased significantly from 9 to 62 mmol l⁻¹ following transfer from fresh water to 450 mosmol kg⁻¹ sea water (treatment A) (P<0.0001, Fig. 5). After 2 days in the final mixed medium (450 mosmol kg⁻¹ sea water plus 150 mosmol kg⁻¹ sorbitol), hemolymph proline concentration decreased significantly from the 450 mosmol kg⁻¹ seawater value (P<0.0001). It decreased further by day 4 (P<0.0001). Concentrations of other amino acids in the hemolymph showed small fluctuations throughout the experiment (data not shown). Hemolymph alanine concentration increased significantly (P<0.045) in larvae held in 450 mosmol kg⁻¹ sea water and then transferred to the final mixed medium (results not shown). In treatment B, when larvae were transferred from fresh water to 450 mosmol kg⁻¹ sorbitol, hemolymph concentrations of threonine decreased significantly (P<0.018); levels of all other amino acids did not vary (data not shown). Upon transfer to the final mixed medium (450 mosmol kg⁻¹ sea water plus 150 mosmol kg⁻¹ sorbitol), hemolymph concentrations of both proline (Fig. 5) and alanine (not shown) increased significantly compared with values in freshwater larvae (P<0.0098, P<0.016, respectively) and larvae from 450 mosmol kg⁻¹ sorbitol treatments (P<0.013, P<0.002, respectively). After 4 days in the final medium, hemolymph proline levels in larvae from both treatments, regardless of whether they had experienced high intermediate levels of salinity or sorbitol, were not significantly different (P<0.325, unpaired t-test) (Fig. 5).

In the treatment involving 450 mosmol kg⁻¹ sea water (treatment A), hemolymph trehalose concentration increased significantly (P<0.0001) compared with values for larvae in fresh water; they increased further and significantly in the final mixed medium (P<0.0001, Fig. 6A). In the treatment in which 450 mosmol kg⁻¹ sorbitol was used (treatment B), trehalose concentration did not increase significantly in larvae transferred from fresh water to 450 mosmol kg⁻¹ sorbitol. When larvae were transferred from 450 mosmol kg⁻¹ sorbitol to the final medium, hemolymph trehalose concentration increased significantly compared with both the freshwater (P<0.0001) and 450 mosmol kg⁻¹ sorbitol (P<0.013) values (Fig. 6B). The hemolymph trehalose concentration attained in the final holding medium in treatment A, in which 450 mosmol kg⁻¹ sea water was the intermediate medium (Fig. 6A), was significantly greater than in treatment B, in which sorbitol was the intermediate (P<0.0002, unpaired t-test, Fig. 6B).

Sorbitol is not present in the hemolymph of C. tarsalis larvae held in fresh water (Fig. 7) or in 400 mosmol kg⁻¹ seawater medium (data not shown) but was detected in the hemolymph of larvae held in sorbitol medium. Hemolymph sorbitol levels increased as the external sorbitol concentrations increased, but the former remained below the iso-osmotic line (Fig. 7).

The relationship between hemolymph proline and Na⁺ concentrations is presented in Fig. 8. Hemolymph proline levels increased approximately sixfold between 140 and 155 mmol l⁻¹ hemolymph Na⁺. Increases in proline concentration leveled off at 70 mmol l⁻¹ when hemolymph [Na⁺] reached 180 mmol l⁻¹.

Larvae of the mosquito C. tarsalis have been shown to be strict osmoconformers in media more concentrated than 300 mosmol kg⁻¹ (Garrett and Bradley, 1987). Osmoconformation of the hemolymph with the external medium is achieved by the accumulation of organic osmolytes, including proline and trehalose (Garrett and Bradley, 1987). In a companion paper, we demonstrated that proline is also a major intracellular osmolyte in C. tarsalis acclimated to concentrated media (Patrick and Bradley, 2000). Given the very large increases in both intra- and extracellular proline concentration with increasing salinity, it is clear that the larvae must synthesize proline in response to the increasing salinity of the medium. The same holds true for the large accumulation of hemolymph trehalose. We were interested in elucidating the signals used by the larvae to initiate and control the synthetic processes for both proline and trehalose. The experiments described in the present study initiate the process of determining the mechanisms by which proline and trehalose accumulation is controlled. From these initial experiments, in which external osmolality was varied using different treatments, we have determined that, in the osmoconforming mosquito larva C. tarsalis, the accumulation of the two organic osmolytes, proline and trehalose, is controlled by two fundamentally different signals.

To begin our investigations, we examined the process by which external changes in salinity and osmolality are correlated with internal ionic regulation of the hemolymph. As shown in Fig. 1A, acclimation to highly saline media is associated with an increase in the hemolymph concentrations of both Na⁺ and Cl⁻. These ions continue to be hyporegulated in the hemolymph relative to the external medium, but increases in concentration do occur. This is in sharp contrast to the strict regulation of K⁺ concentration and the small decline observed in Ca²⁺ and Mg²⁺ concentrations. In contrast, if the osmolality of the external medium is elevated by the addition of sorbitol, with no change in external ion levels, the hemolymph concentrations of Na⁺ and Cl⁻ decline, relative to the freshwater condition (Fig. 1B). K⁺ concentration continues to be strictly regulated, indicating that the acclimation to high- sorbitol medium is not inducing a general ionic disturbance. From this first set of experiments, it would seem that external salinity is not the sole determinant of concentrations of the major hemolymph ions (in particular Na⁺ and Cl⁻). This idea is further supported by the second set of experiments in which external osmolality was increased by salinity (treatment A), by sorbitol (treatment B) or by a mixture thereof. In treatment A (Fig. 4A), we observed that Na⁺ and Cl⁻ concentrations, while still hyporegulated, show an increase during acclimation to the intermediate medium of 450 mosmol kg⁻¹, similar to Fig. 1A. However, when larvae were then transferred to the final mixed medium, in which external salt concentration remained at 450 mosmol kg⁻¹ but total osmolality increased to 600 mosmol kg⁻¹, we observed a significant decrease in hemolymph NaCl levels (Fig. 4A). In treatment B (Fig. 4B), hemolymph Na⁺ and Cl⁻ concentrations both decreased significantly in the intermediate sorbitol medium, similar to trends in Fig. 1B. Upon transfer to the final mixed medium, hemolymph Na⁺ and Cl⁻ levels increased. Taken together, these patterns of changes in hemolymph ion levels indicate that environmental salinity is not the only determinant of hemolymph ion levels. Exactly how external osmolality is involved in influencing hemolymph ion regulation remains uncertain, but the finding that external sorbitol is being taken up by the larvae and is serving as an osmoprotectant solute (Fig. 7) is important (see below).

Next, we examined the response of the organic osmolytes to changes in salinity and osmolality. Proline accumulation within the body fluids appeared to be cued by increases in the hemolymph Na⁺ levels. This was first indicated by the almost 50-fold increase in hemolymph proline concentration when larvae were held in 640 mosmol kg⁻¹ seawater medium (Fig. 2A) and the absence of an increase when external osmolality was raised to the same value using sorbitol (Fig. 2B). However, in this experiment, hemolymph Na⁺ and Cl⁻ levels (Fig. 1A) showed an increase similar to that of the external salinity, making it impossible to distinguish whether the signal for proline accumulation (Fig. 2A) was internal (hemolymph NaCl) or external (water NaCl). From the subsequent set of experiments, we were able to deduce that proline accumulation followed changes in hemolymph NaCl concentration. This was evident in treatment A, in which the external salinity remained constant from the 450 mosmol kg⁻¹ seawater medium to the final mixed medium yet the hemolymph Na⁺ and Cl⁻ levels dropped (Fig. 4A). Coinciding with this reduction in hemolymph NaCl concentration (Fig. 4A) was the significant, but gradual, decrease (over 4 days) in hemolymph proline concentrations (Fig. 5). A similar parallel between hemolymph NaCl levels and proline accumulation was found in treatment B, in which both NaCl (Fig. 4B) and proline (Fig. 5) levels increased following transfer from the intermediate sorbitol medium to the final mixed medium. These results, in conjunction with those presented in the companion paper (Patrick and Bradley, 2000), provide evidence that, in the larva of C. tarsalis, hemolymph NaCl concentrations are involved in the regulation of both intra- and extracellular proline levels.

How does an increase in hemolymph NaCl concentration induce accumulation of both intra- and extracellular proline in C. tarsalis? Perhaps the mechanism for salt induction of proline accumulation involves an interaction between inorganic salts and the enzymes required for both the synthesis and catabolism of proline. Presumably this is an intracellular phenomenon, and in insects the fat body has generally been acknowledged as the primary site of proline regulation (Chapman, 1998). With regard to catabolism, the activities of enzymes involved in proline oxidation (e.g. proline dehydrogenase) could be inhibited in a manner characterized for many cellular enzymes (Somero and Yancey, 1997), thereby maintaining high proline levels. It is not clear how this extracellular increase in Na⁺ and Cl⁻ concentrations would interact with intracellular metabolic control mechanisms. It is conceivable that the increase in extracellular NaCl levels that the C. tarsalis larvae experienced in high-salinity medium (Figs 1A, 4A) resulted in an increase in intracellular ion concentrations, causing such an inhibitory effect on proline catabolism.

The role of proline synthesis in the osmoconforming response has been shown to be vital in several osmotolerant microrganisms (Measures, 1975; Whatmore et al., 1990), plants (Hare et al., 1998) and marine invertebrates (Burton, 1991a,b). The specific pathway being upregulated in these organisms has been determined to be the synthesis of proline from glutamate. Proline, as the end product, provides negative feedback on the first enzyme in this pathway, γ-glutamyl kinase, and limits proline production. High levels of inorganic salts interfere with this allosteric effect of proline, thereby allowing synthesis to proceed irrespective of free proline levels (Measures, 1975). In Bacillus subtilis, a bacterium that can accumulate up to 1 mol l⁻¹ proline in response to high salinity (Whatmore et al., 1990), and in the euryhaline copepod Tigriopus californicus (Burton, 1991a), it was found that protein synthesis is necessary for the upregulation of proline synthesis. Therefore, proline accumulation in response to hyperosmotic shock involves an augmentation of the activity of pre-existing enzymes and an induction of synthesis through new enzymes.

The one key difference between the osmotolerant organisms discussed above and the osmoconforming larvae of C. tarsalis is that the former utilize proline only as an intracellular osmotic effector, whereas C. tarsalis accumulates proline both in the body fluid and in the tissues. The fact that C. tarsalis larvae can accumulate up to 70 mmol l⁻¹ proline in both compartments suggests that proline synthesis must be quite active. It is conceivable that the fat body provides both intracellular and hemolymph proline because it forms a flat, broad sheath that shares a large surface area with the hemolymph and it has been implicated as the primary site of proline synthesis is several insect species (Chapman, 1998). In addition, there may also be an osmotically induced proline transport system that would allow the synthesized proline to exit the fat body and enter the hemolymph so as to be distributed throughout the body. Such mechanisms have been proposed to occur in osmotolerant prokaryotes (Csonka and Epstein, 1996; Kempf and Bremer, 1998). Volume-sensitive amino acid channels have been characterized in vertebrate erythrocytes (Chamberlain and Strange, 1989).

The present study of C. tarsalis larvae indicates that changes in hemolymph proline concentration follow changes in hemolymph NaCl concentration, but that there is an asymmetry in the proline response. The overall trend was that the decrease in hemolymph proline levels (Fig. 5) lagged behind the decrease in hemolymph NaCl levels (Fig. 4A), whereas proline accumulation (Fig. 5) proceeded faster in response to increased NaCl levels (Fig. 4B). In a euryhaline copepod, Tigriopus californicus, a similar asymmetry in the time course of accumulation (3 h) and reduction (5–24 h) of intracellular proline and alanine concentrations has been documented (Goolish and Burton, 1989). The difference in the response time between the two organisms may reflect the fact that C. tarsalis employs proline as both an intra- and extracellular osmolyte (Patrick and Bradley, 2000), whereas the copepod only regulates intracellular pools of amino acids.

In contrast to the salt-stimulated proline response in C. tarsalis, trehalose accumulation was cued by an increase in osmolality. Whether larvae were held in a high-salinity or sorbitol medium, hemolymph trehalose increased to the same concentration (Fig. 3). Even when different solutes (sea salts, sorbitol), or a mixture thereof, were used to raise the osmolality of the holding medium, hemolymph trehalose levels were correlated with the increase in environmental osmolality and were independent of the changes in hemolymph ion levels (Fig. 6). The results suggest that the larvae detect changes in external osmolality and adjust their trehalose levels accordingly.

In Escherichia coli, trehalose accumulation responds to a more proximal signal (Kempf and Bremer, 1998). In this bacterium, trehalose synthesis is osmotically regulated via expression of genes encoding the biosynthesis enzymes. This system is activated by the acute rise in cellular K⁺ concentration that occurs via the increased active uptake of K⁺, which itself is activated at the gene level (K⁺ transporter induction). This expression of K⁺ transporters, leading to the induction of trehalose synthesis, is determined partially by the strength of the osmotic shock and subsequent drop in turgor. The synthesis and accumulation of trehalose are very rapid and, within a few hours, trehalose replaces the less compatible solute K⁺ and its counteractant glutamate (Dinnibier et al., 1988).

The two-step synthesis of trehalose from glucose has been detailed in the locust fat body (Candy and Kilby, 1961) and is similar to that described in E. coli. However, the osmoregulatory control of trehalose synthesis has not been assessed in insects and has yet to be characterized in eukaryotes despite the fact that this solute dominates in desiccation-tolerant (Crowe et al., 1984) and freeze-tolerant (Duman et al., 1991) organisms. In addition to the upregulation of trehalose synthesis, trehalose degradation must be reduced so as to avoid establishing a futile cycle. Trehalase, the enzyme that hydrolyzes trehalose to α-D-glucose, is inactivated by inorganic salts (Herzog et al., 1990) and has been detected in the hemolymph of a variety of insects (Periplaneta sp., Locusta sp., Phormia sp.; Chapman, 1998). It is conceivable that the osmolality cue is translated through a pathway involving volume regulation. The issue is complicated by the fact that trehalose, like proline, is probably synthesized intracellularly, presumably within the fat body, but must be transported out into the hemolymph.

A surprising finding in the present study was that sorbitol was taken up from the external medium and accumulated within the hemolymph in a concentration-dependent manner (Fig. 7). The accumulation of sorbitol is intriguing for several reasons. First, the osmoconforming response of C. tarsalis has presumably evolved to defend the internal milieu. These mosquitoes reside in inland bodies of water in which ionic concentrations and ratios can vary unpredictably (Bradley, 1994). Our demonstration of sorbitol accumulation suggests that at least some external osmolytes can be taken up. Sorbitol, a polyol, is considered to be a compatible solute and is utilized by numerous eurkaryotes in response to hyperosmotic shock (Somero and Yancey, 1997), but it is not common in any aquatic environment. Sorbitol is synthesized in adult mosquitoes (Aedes sollicitans; Van Handel, 1969) and in the silverleaf whitefly (Bemisia argentifolii) in response to cold and heat shock (Wolfe et al., 1998). Sorbitol is not present in the C. tarsalis hemolymph under freshwater or saltwater conditions (Fig. 7). It would seem that, in the presence of a compatible osmolyte, C. tarsalis accumulates sorbitol and suppresses proline production. The bacterium E. coli, when placed in a hyperosmotic minimal medium, will synthesize and accumulate trehalose (Strøm and Kaasen, 1993). If the preferred osmolyte glycine betaine or its precursor choline is added to the medium, however, E. coli will accumulate these solutes and not produce trehalose (Somero and Yancey, 1997). It may be that C. tarsalis, like E. coli, can assess the compatibility of an exogenous solute and then proceed to accumulate it. By preferentially accumulating the exogenous compatible solute rather than Na⁺ and Cl⁻, the signal for proline accumulation would be attenuated. Accumulating an exogenous compatible osmolyte may be energetically cheaper than proline production. A final point regarding sorbitol accumulation is that it accounted for most of the hemolymph osmolality that remained after the osmolality of all quantified ions, trehalose and total free amino acids had been taken into consideration. This contrasts with C. tarsalis larvae that are transferred from fresh water to 640 mosmol kg⁻¹ sea water, in which the unaccounted hemolymph osmolality rises by approximately 60 mosmol kg⁻¹, suggesting the presence of an ‘unknown’ osmolyte or osmolytes. These results suggest that the ‘unknown’ osmotically active component was accumulated in response to increased external salinity, as was proline. If larvae are placed in a high-sorbitol medium, the accumulation of both proline and this ‘unknown’ osmolyte(s) is attenuated.

This study has revealed that two of the osmolytes utilized in the osmoconforming strategy of C. tarsalis are regulated by two fundamentally different signals. This is particularly intriguing because both osmolytes are presumably produced within the fat body of the larvae. Proline accumulation tracks increases in hemolymph NaCl levels, whereas trehalose accumulation was cued by osmotic shifts independent of the external (or internal) chemical composition. The relationship between proline and hemolymph Na⁺ concentration, as depicted in Fig. 8, is not a linear function. Instead there is a threshold at approximately 140 mmol l⁻¹ Na⁺. At higher hemolymph Na⁺ concentrations, proline accumulation increases rapidly. It is possible that this threshold indicates the induction of proline synthesis. We are currently pursuing this issue with the hope of eventually characterizing the regulation of proline and trehalose accumulation in C. tarsalis with the degree of definition now possible in the prokaryotes.

The authors would like to thank Drs A. F. Bennett, A. G. Gibbs and R. K. Josephson and the reviewers for helpful comments and reading the manuscript. We would like to thank Dr V. ‘Hawkeye’ Pierce for providing invaluable help with the sorbitol assay and J. Chaney for advice on rearing the mosquitoes. This research was supported by NSF grant IBN 9723404.