Hyperosmotic stress may result in osmotic volume loss from the body to the environment in animals that cannot control the water permeability of their integument. Euryhaline animals (which have a wide tolerance range of environmental salinities) have generally evolved the ability to counteract cell volume shrinkage by accumulating inorganic and organic osmolytes within their cells to balance internal and external osmolalities. Molluscs use very different combinations of amino acids and amino acid derivatives to achieve this goal. Theodoxus fluviatilis is a neritid gastropod that is distributed not only in limnic habitats in Europe but also in brackish waters (e.g. along the shoreline of the Baltic Sea). Animals from brackish sites survive better in high salinities than animals from freshwater locations. The results of the present study indicate that these differences in salinity tolerance cannot be explained by differences in the general ability to accumulate amino acids as organic osmolytes. Although there may be differences in the metabolic pathways involved in osmolyte accumulation in foot muscle tissue, the two groups of animals accumulate amino acid mixtures equally well when stepwise acclimated to their respective maximum tolerable salinity for extended periods. Among these amino acids, alanine and proline, as well as the osmolyte urea, hold a special importance for cell volume preservation in T. fluviatilis under hyperosmotic stress. It is possible that the accumulation of various amino acids during hyperosmotic stress occurs via hydrolysis of storage proteins, while alanine and proline are probably newly synthesised under conditions of hyperosmotic stress in the animals.
Animals facing unstable environments have evolved a range of genetic adaptations to deal with extreme environmental conditions (Rowińsky and Rogell, 2017). Within this genetic framework, animals respond to rapid changes of conditions by gene regulatory or post-transcriptional mechanisms leading to altered protein expression, or by post-translational means such as protein modification, changes in enzyme activities or alterations in metabolism (Lockwood and Somero, 2011). The ability of individuals to modify their phenotype as a response to environmental change or stress is termed phenotypic plasticity (Woods, 2014).
Aquatic animals with an integument that is not entirely water impermeable (Oglesby, 1981) undergo osmotic water gain when exposed to external salinities lower than that of their own body fluids (hypoosmotic stress). This results in volume expansion of the internal body fluids that has to be limited by excretion of a hypotonic urine through the kidneys (Dantzler, 2016) or other excretory organs (Kirschner, 1967). Concomitantly, such animals reduce their internal osmolyte load by degrading or excreting inorganic ions or organic osmolytes to reduce the driving force for osmotic water influx (Pierce, 1982). If such animals are exposed to external salinities higher than that of their own body fluids, the animals lose water to the environment and their body fluid volume decreases. To avoid detrimental shrinkages in extracellular and intracellular fluid volumes, animals accumulate inorganic ions and organic osmolytes (amino acids and their derivatives, polyols, sugars, methylamines, methylsulfonium compounds or urea) in their internal fluid compartments to reduce the driving force for osmotic water loss (Burg, 1995; Yancey et al., 1982; Yancey, 2005; Burg and Ferraris, 2008). These substances have to be taken up, be newly synthesised or mobilised from high molecular mass precursors (Wehner et al., 2003). In the case of animals that accumulate amino acids as organic osmolytes under conditions of hyperosmotic stress in their body fluids, it is still unclear whether these small molecules are taken up from the external medium, acutely synthesised, generated by degradation of storage proteins or accumulated through a combination of these processes (Manahan et al., 1983; Deaton, 1987; Gilles and Péqueux, 1981).
Euryhaline molluscs mainly accumulate amino acids as organic osmolytes in their cells under hyperosmotic stress (Shumway et al., 1977; Pierce and Amende, 1981). While bivalves have been thoroughly investigated in terms of organic osmolyte accumulation under hyperosmotic stress, there is only a limited number of studies in oligohaline gastropods (Deaton, 2009; Taylor and Andrews, 1988; Symanowski and Hildebrandt, 2010) with highly differing results. Although Littorina littorea is an intertidal gastropod species experiencing harsh changes in environmental salinity, it does not utilise intracellular accumulation of free amino acids to avoid cell volume changes under hyperosmotic stress (Taylor and Andrews, 1988). In contrast, the euryhaline snail Theodoxus fluviatilis, which occurs in freshwater lakes in central Europe as well as in brackish water along the coastline of the Baltic Sea, accumulates substantial amounts of free amino acids and urea (measured as the sum of ninhydrin-positive substances, NPS) in foot muscle tissue upon being transferred to higher salinities (Symanowski and Hildebrandt, 2010). Snails collected from freshwater sites (FW) and those collected from brackish water sites (BW) showed clear differences in their respective salinity tolerance (Zettler et al., 2004; Bandel, 2001; Hubendick, 1947; Wiesenthal et al., 2018). FW individuals were not able to cope with salinities as high as those tolerated by BW snails. Even when given the time to acclimatise to gradually increasing salinity of their environmental medium over several days, FW snails barely survived in salinities up to 21‰. Snails collected from BW sites, however, survived salinities up to 28‰ when gradually acclimated (Wiesenthal et al., 2018).
The present study was conducted to: (1) elucidate whether differences in the ability to accumulate free amino acids in the tissues explain these differences in tolerance to hyperosmotic stress, (2) identify those amino acids that contribute the most to intracellular osmotic adjustments, and (3) determine whether accumulated amino acids are newly synthesised or generated by hydrolysis of storage proteins.
MATERIALS AND METHODS
Animal collection and transfer experiments
Theodoxus fluviatilis (Linneaus 1758) were collected from three FW and two BW sites in northern Germany. The FW lakes Schmaler Luzin and Carwitzer See are part of the Feldberger Seenlandschaft and lie about 100 km north of Berlin. The BW sites are on the coastline of the Baltic Sea near Greifswald and on the island of Hiddensee. Collections took place between May and September 2016 and snails were subjected to 15 day transfer experiments in the laboratory at room temperature (21°C). FW animals were transferred from their original medium (salinity of 0.5‰, 1 day) to their maximum salinity of 21‰ for 3 days using a step by step regime (3.7‰ for 3 days, 6.9‰ for 3 days, 10‰ for 5 days). BW animals were also transferred from their original medium [salinities of 7.5‰ (site S5) or 9‰ (site S6), 1 day] to their maximum salinity of 28‰ for 3 days using a stepwise regime (12‰ for 3 days, 18‰ for 3 days, 24‰ for 5 days). Control animals were held at their original salinities throughout the experiment. Detailed information on collection, storage and the transfer experiments has previously been provided (Wiesenthal et al., 2018).
The snails were quickly cooled down to 4°C for 10–15 min before the foot muscle was dissected. The isolated foot muscle of each individual was blotted dry, weighed using a precision scale (Quintix, Sartorius, Göttingen, Germany) to the nearest 0.001 g and immediately placed in liquid nitrogen. Each of the frozen tissue samples was homogenised on ice (T-8 Ultraturrax, IKA, Staufen, Germany) in 300 µl deionised water and centrifuged at 16,000 g for 4 min at 4°C (Heraeus Fresco 21, Thermo Scientific, Waltham, MA, USA). The supernatant was transferred to a new reaction tube. Proteins in the supernatant were precipitated by adding 60 µl of an ice-cold 0.575 mol l−1 solution of 5–sulfosalicylic acid dihydrate (Roth, Karlsruhe, Germany) to the homogenate. The mixture was vortexed (REAX 2000, Heidolph, Schwabach, Germany) and centrifuged at 16,000 g for 4 min at 4°C, leaving the extracted free amino acids (FAAs), amino acid derivatives and urea from both intracellular and extracellular fluids in the supernatant. The supernatant was frozen at −20°C in aliquots of 50 µl.
Quantification of the FAAs and other NPS, including urea
For the amino acid analyses, 50 µl of the extracted osmolyte samples was diluted with 50 µl of lithium loading buffer (Laborservice Onken, Gründau, Germany) and filtered by centrifugation at 7000 g for 3 min at 4°C using micro-centrifuge filter tubes equipped with a 0.2 µl nylon membrane (Laborservice Onken). The filtrates were transferred to HPLC sample vials (Macherey & Nagel, Düren, Germany). A 1:20 dilution of the standard amino acid mixture provided by the manufacturer (Laborservice Onken) was further diluted with lithium buffer in order to obtain 100 µl of a 1:200 dilution. This standard mixture was freshly prepared for every analysis series of biological samples and run at the beginning and at the end of each series of HPLC analyses to check for differences in signal intensity potentially related to the replacement of HPLC solvents during a series.
Analyses were performed on a Biochrom 30+ amino acid analyser (Laborservice Onken) using the original buffer kits (Laborservice Onken). Samples and standards were automatically applied to the machine by an autosampler that kept the samples at 4°C before injection. Post-column ninhydrin-derivatised substances were detected using a SPD-20AV Prominence HPLC UV-Vis detector (Shimadzu, Columbia, MD, USA). Editing of the chromatograms was done using OpenLAB software (Agilent Technologies, Waldbronn, Germany). Peak areas were normalised against internal standards (hydroxylysine), and relative substance amounts in the samples (mmol kg−1 tissue fresh mass) were calculated using the analysis data of the amino acid standard mixtures and Microsoft Excel.
Quantification of the potential compatible osmolytes trimethylamine N-oxide, glycine betaine, sarcosine, glycerol, myo-inositol and glycerophosphorylcholine
Only samples of snails from collection site S5 were prepared. This was done as described above (‘Sample preparation’) with the addition of 43 µl of a 1 mol l−1 sodium bicarbonate solution (Roth, Karlsruhe, Germany) to each 50 µl aliquot in order to neutralise the pH value. These steps were carried out as preparation for the quantification of compatible osmolytes with a sample size of 3.
The potential compatible osmolytes glycerol, myo-inositol, sarcosine and glycerophosphorylcholine (GPC) were quantified in the thawed samples (n=3) using the Enzytec Fluid Glycerol Assay Kit (ID-No: 5360; Thermo Fisher Scientific, Vantaa, Finland), the VitaFast Inositol Assay Kit (ID-No: 1009; Institut für Produktqualität GmbH, Berlin, Germany), the Sarcosine Colorimetric/Fluorometric Assay Kit (ID-No: K636-100; BioVision Inc., Milpitas, CA, USA) and the Glycerophosphorylcholine Assay Kit (ID-No: K433-100; BioVision Inc.) respectively. The assays were carried out according to the manufacturers' instructions with a downscaling for the small volume of the sample tissue. Undiluted samples as well as a 1:100 sample dilution were measured to ensure that the standard curve covered the detected amount in the sample. The amount of glycine betaine and trimethylamine N-oxide (TMAO) was determined as described in Valadez-Bustos et al. (2016) and Wekell and Barnett (1991). The chemicals and solutions for these assays were acquired from Sigma-Aldrich Chemie GmbH (Munich, Germany) and Roth. Again, the instructions were downscaled to match the small volume of sample available and measured in the undiluted sample as well as in a 1:100 dilution.
The statistical analysis was carried out with the free software R 3.3.2 (http://www.R-project.org/). The result for each individual component (NPS) was tested for normal distribution with the Shapiro–Wilk test and for homogeneity of variance with the Fligner–Killeen test. When testing for potential increases in the NPS between the control groups and stressed groups for each site, either the Welch t-test or the Kolmogorov–Smirnov test (KS test) was used, depending on the outcome of the Shapiro–Wilk test (normally distributed data: Welch t-test; non-normally distributed data: KS test). Differences between control and stressful conditions among sites as well as differences between sites among control groups and among stressed groups were tested for with the Kruskal–Wallis test and the post hoc test after Dunn (1964) with a P-value adjustment (Benjamini and Hochberg, 1995) (R package: ‘PMCMR’). A Kruskal–Wallis test was additionally used to exclusively compare the FW sites under both control and high-salinity conditions. Graphs were generated with R 3.3.2 (R package: ‘tiff’, ‘raster’, ‘lattice’ and ‘RColorBrewer’).
Specimens of T. fluviatilis were collected at different locations with different environmental salinities ranging from 0.5‰ (FW sites) to 9‰ (BW sites). The total amount of amino acids in the foot muscle of these animals under their original salinity conditions clearly correlated with the external salinity (Fig. 1). This shows that animals of this species use organic osmolytes not only to prevent cell volume changes during acute osmotic stress but also to balance the osmotic concentration of the body fluids with that of the external medium during normal salinity conditions.
The total amino acid accumulation upon stepwise acclimation of the animals to their maximum tolerable salinity (21‰ for animals from FW sites S1–S3; 28‰ for animals from BW sites S5 and S6) was very similar in all cases (Fig. 2). The amounts of all amino acids were up to 27-fold higher in stressed animals than in control animals. Such an organic osmolyte accumulation accounts for approximately 21–27% (osmolality of accumulated FAA/osmolality of the environment) of all osmolytes in the foot muscle tissue under the respective conditions.
To analyse the relative contributions of individual amino acids and amino acid derivatives to the observed changes in overall organic osmolytes upon hyperosmotic stress in the animals, we measured the methionine (Met), tryptophan (Trp), taurine (Tau), glycine (Gly), alanine (Ala), isoleucine (Ile), leucine (Leu), arginine (Arg), lysine (Lys), proline (Pro), threonine (Thr), serine (Ser), β-alanine (β-Ala), histidine (His), phenylalanine (Phe), glutamate (Glu), valine (Val), tyrosine (Tyr), cysteine (Cys), γ-aminobutyric acid (GABA), hydroxy-proline (OH-Pro), β-amino-isobutyric acid (BAIBA), α-amino adipic acid (AAAA), α-aminobutyric acid (AABA) and cystathionine (Cysth) content in each of the samples (Table S1). The results indicated that the contributions of individual amino acids to these changes under osmotic stress in the animals were very different and ranged from 2-fold (Leu in BW animals; Fig. 3A) to 650-fold (proline in FW animals; Fig. 3B, Table S1). It was very obvious that the increase in amino acid content was much more pronounced in osmotically stressed FW animals than in stressed BW animals (Fig. 3). This appears to correlate with the magnitude of change in external osmolality (42-fold in FW animals; 3-fold in BW animals) to which these animals were exposed to in order to impose the maximum tolerable osmotic stress.
The two amino acids that accounted for most of the total change in organic osmolyte content in the foot muscle of stressed animals were alanine and proline (Fig. 3B), in terms of both fold-change and absolute amount (Fig. 4), but urea seemed to be quantitatively important as well (Fig. 5C). Alanine increased 10- to 20-fold in BW animals exposed to an external salinity of 28‰, and 80- to 130-fold in FW animals exposed to a salinity of 21‰, reaching a mean of 47 and 88 mmol kg−1 fresh mass, respectively. Proline accumulated 60- to 80-fold in BW animals exposed to 28‰ salinity (43 mmol kg−1 fresh mass), and 500- to 640-fold in freshwater animals exposed to 21‰ salinity (17 mmol kg−1 fresh mass). The tryptophan content also appeared to increase in the foot muscle (44-fold), but this was due to large variations in individual measurements at very low levels (<1 mmol kg−1 fresh mass) and was, therefore, not considered as relevant. Other amino acids, however, also contributed to the organic osmolyte accumulation in foot muscle tissue under hyperosmotic stress (Fig. 3A), but to lesser degrees compared with proline or alanine (Figs 3B, 4B, 5A, 6A). As shown in Fig. 6B,C, amino acids like taurine or glycine were generally present at higher levels in the foot muscle of osmotically stressed animals compared with the respective controls, but their contribution to the total change in organic osmolytes was relatively small. Moreover, the pattern of changes in individual amino acids in the foot muscle differed between FW and BW animals. For instance, much larger quantities of taurine were accumulated in BW animals (S5, S6) than in FW animals (S1–S3) under hyperosmotic stress (Fig. 6B). For glycine, in contrast, very similar quantities were accumulated in stressed animals irrespective of their origin (Fig. 6C). Overall, glycine, taurine as well as the amino acid derivatives GABA, BAIBA, AAAA and AABA were present at low concentrations in the foot muscle tissue in control animals (0.001–3.8 mmol kg−1 fresh mass). During hyperosmotic stress, these substances increased by 6- to 46-fold (maximum total of 10 mmol kg−1 fresh mass), which was not considered to be a very relevant contribution to the total organic osmolyte content. β-Alanine was also detected in rather small amounts in control animals and hyperosmotically stressed FW snails, but accumulated to 16 mmol kg−1 fresh mass in stressed BW snails, which corresponded to an approximately 9-fold increase (Figs 3B and 7A). This means that the amount of β-alanine amounted to roughly 3% of the quantity of urea that was accumulated under hyperosmotic stress in BW snails (Fig. 5C).
As these differences in the accumulation of individual amino acids and amino acid-derived molecules in cells may be explained by different potential mechanisms (entry and exit through amino acid transporters, amino acid biosynthesis and degradation or protein biosynthesis and degradation), we tried to find markers among the amino acids and amino acid derivatives that may help to disentangle the underlying pathways of amino acid accumulation during hyperosmotic stress. The patterns of changes in amino acid quantities in the foot muscle tissue of those amino acids that are generally considered to be essential (Met, Leu, Ile, His, Phe, etc.) were very similar to those observed for the non-essential amino acids (Gly, Ser, Arg, Glu, etc.) (Fig. 3A). Thus, a comparison of accumulated essential and non-essential amino acids did not allow any conclusions with respect to the accumulation mechanism. Because – in animal cells – the irreversible hydroxylation of proline only takes place when this amino acid is integrated in protein strands (mainly collagen; Gorres and Raines, 2010), the occurrence of free hydroxy-proline indicates that cells actively turn over their protein content. Additionally, the accumulation of hydroxy-proline under stress may indicate that hydrolysis of storage proteins is accelerated. Also, transamination and deamination of amino acids are frequent reactions in the synthesis and degradation of amino acids (Bröer and Bröer, 2017). They may result in free ammonia that is partially transformed to urea in limnic snails (Haggag and Fouad, 1968; Horne and Boonkoom, 1970). Differences in the urea content of foot muscle tissue may therefore be used as an indicator for different turnover rates in amino acid metabolism.
Based on these assumptions, we analysed the proline and the hydroxy-proline content of each of the snail samples. As shown in Fig. 5, the proline content in the foot muscle was low in FW animals as well as in BW animals under control conditions but increased substantially in both groups under osmotic stress conditions (Fig. 5A). The hydroxy-proline content in the foot muscle tissue, however, already clearly differed between the groups under control conditions (Fig. 5B). It was very low (<0.1 mmol kg−1 fresh mass) in FW animals, but substantially higher in BW animals in their natural medium (approximately 1 mmol kg−1 fresh mass). During osmotic stress, hydroxy-proline increased considerably in FW animals, but did not change significantly in BW animals (Fig. 5B). While the pattern of urea quantities in the foot muscle tissue (FW animals approximately 3 mmol kg−1 fresh mass; BW animals approximately 300 mmol kg−1 fresh mass) was similar to that of hydroxy-proline in the control animals (Fig. 5), the data showed that there was some urea accumulation under hyperosmotic stress in the FW animals (approximately 50 mmol kg−1 fresh mass), but a much more substantial increase in osmotically stressed BW animals, where the amount of urea virtually doubled (to approximately 600 mmol kg−1 fresh mass) (Fig. 5C).
Because large amounts of urea can have denaturing effects on proteins, the quantities of TMAO, glycine betaine, sarcosine, glycerol and GPC were measured as these substances are considered to be compatible osmolytes that may alleviate negative impacts of high urea concentrations on protein conformation (Withers and Guppy, 1996; Yancey, 2005). All of these substances were below detection limits in FW as well as in BW animals and did not show noticeable increases under hyperosmotic stress. Myo-inositol, however, slightly increased in individuals exposed to hyperosmotic stress (Fig. 7B). Under control conditions, a mean±s.d. of 23.5±24.3 mmol kg−1 fresh mass was observed. Under hyperosmotic conditions, an increase by a factor of 1.8 to 42.5±17.9 mmol kg−1 fresh mass was observed in BW snails. This amounted to roughly 7% of the urea level in BW snails under hyperosmotic conditions.
Although the BW and FW groups of the studied T. fluviatilis show the ability to accumulate organic osmolytes as a response to hyperosmotic stress equally well, they differ in the pathways of acquiring these organic osmolytes. The main constituents of the increased amounts of organic osmolytes are alanine, proline and urea, which seem to be most important for an initial coping with high environmental salinity conditions.
It has been known for a long time that a number of euryhaline aquatic molluscs – including bivalves and gastropods – are hyper-regulators in very dilute salinities, but are basically osmoconformers at higher osmotic concentrations of the environmental medium (Deaton, 2009). As molluscs generally cannot control the water permeability of their integument very well, they undergo rapid changes in fluid volume in extracellular and intracellular compartments when exposed to environmental media with osmotic concentrations that are either hyperosmotic or hypoosmotic with respect to the body fluids of the animals. Such passive responses are very similar in marine bivalves (Gainey, 1987; Hosoi et al., 2003) and limnic/brackish water gastropods (Symanowski and Hildebrandt, 2010). Euryhaline species survive such substantial changes in body fluid volume and are able to prevent extreme swelling (under hypoosmotic stress) or shrinkage (under hyperosmotic stress) in cell volume by, respectively, rapidly releasing organic osmolytes from their cells or accumulating them to adjust the intracellular osmolality to the external conditions (Pierce, 1982; Yancey, 2005).
The European neritid T. fluviatilis is generally considered to be a widely distributed limnic snail, but it also occurs in BW along the shorelines of the Black Sea and the Baltic Sea (Bunje, 2005; Bunje and Lindberg, 2007). Because this species does not have any pelagic larval stages, each population is quite stationary, which results in a patchy distribution of populations. The adult animals within one population are very similar in terms of shell size and patterning (Wiesenthal et al., 2018). However, across several different populations, the variability of these parameters seems to be high (Neumann, 1960; Kangas and Skoog, 1978; Zettler et al., 2004; Symanowski and Hildebrandt, 2010; Wiesenthal et al., 2018), although some mitochondrial marker genes have very similar sequences (Bunje, 2005; Bunje and Lindberg, 2007). Thus, to date it is unclear whether the differences in shell size and patterning may be explained by genetic variation (local adaptation) or by phenotypic or developmental plasticity (Glöer and Pešić, 2015). It has previously been reported (Symanowski and Hildebrandt, 2010) that the salinity tolerance of animals from FW locations in northern Germany is less well developed than that of BW animals. Additionally, we have learned that reaction norms (survival in different salinities) may shift in animals of both groups by a stepwise acclimation of the animals to increasing or decreasing salinity. And yet, the FW animals were still not able to survive salinities higher than 21‰, whilst the BW animals did well in salinities of 28‰ after such a stepwise acclimation (Wiesenthal et al., 2018). This indicates that different genetic limitations of plasticity may exist in FW and BW groups of animals, pointing towards local adaptation. Our main aim in this study was to elucidate whether the ability or the mode of limiting passive cell volume changes by intracellular organic osmolyte accumulation may be one of these factors.
A general observation in euryhaline molluscs under hyperosmotic stress is that the patterns of the individual free amino acids recruited for intracellular volume adjustments are diverse (Deaton, 2009). Quantitatively relevant osmolytes in different species are taurine, glycine, alanine and proline (Yancey et al., 1982; Hosoi et al., 2003; Yancey, 2005). Our results in T. fluviatilis show that alanine (mainly in the FW animals; Fig. 6A) and proline (mainly in the BW animals; Fig. 5A) as well as urea (Fig. 5C) are the most relevant organic osmolytes in foot muscle tissue after a 2 week transfer regime and 72 h exposure to a final hyperosmotic condition. The high levels of urea and its importance as an organic osmolyte were surprising, because despite it being known to be present at high concentrations in elasmobranch body fluids and to accumulate in terrestrial snails during aestivation and desiccation, such an accumulation has not been reported in aquatic snails (Tam et al., 2003; Hazon et al., 2003; Horne and Barnes, 1970; Arad, 2001; Rees and Hand, 1993; Hiong et al., 2005).
The accumulation of amino acids as major organic osmolytes in animals under hyperosmotic stress (Fig. 2) may occur through hydrolytic degradation of storage proteins. This conclusion is supported by the observation that even essential amino acids (Met, Ile, Leu, Lys, Thr, His, Phe, Val) were accumulated 5- to 11-fold in FW animals or 2- to 9-fold in BW animals under hyperosmotic stress (Fig. 3A). Uneven representation of amino acid residues in storage proteins may explain the observation that the abundance of some amino acids (e.g. Met) changed more than that of others (e.g. Leu). Another potential explanation is that some of the FAAs present in cells upon protein hydrolysis were subject to metabolic conversion whilst others were largely unaffected. The rate of increase of organic osmolytes also points towards metabolic processes of turnover or synthesis rather than the uptake of amino acids (Bishop et al., 1994).
Non-essential amino acids, however, may also be newly synthesised during hyperosmotic stress. This may be a potential explanation for the observed accumulation of alanine (Fig. 6A). Changes in alanine were larger than those of all other amino acids (except proline), and definitely larger than would be expected from hydrolysis of a standard protein in which alanine would account for approximately 5% of all amino acids. From our data, we cannot draw any conclusions on the mode of alanine accumulation (accelerated synthesis, decelerated metabolic conversion) in T. fluviatilis under hyperosmotic stress. The mode may be similar to those used in other molluscs, as alanine accumulation is a widely distributed phenomenon in euryhaline molluscs during hyperosmotic stress (Deaton, 2009; Bishop et al., 1994; Kube et al., 2006) along with the accumulation of taurine, β-alanine and glycine. However, in contrast to these other marine mollusc species, the last three mentioned organic osmolytes are not really relevant for overall adjustments of the internal osmolality in T. fluviatilis. Proline, in contrast, accumulated to a very high degree in the foot muscle tissue of animals under hyperosmotic stress, which raises the question of its source. A potential mechanism is the hydrolysis of storage proteins with high proline content, e.g. collagens (Li and Wu, 2018), and proline uptake into the cells (Bröer and Bröer, 2017). Proline-rich repeats (PRRs) of proteins may contain up to 50% proline residues (Williamson, 1994). Assuming that such a proline-rich protein is overall composed of 17% proline residues [as in the collagen alpha-2(IX) chain-like protein of Biomphalaria glabrata, UniProt A0A2C9KD15] and the other amino acid residues would be evenly represented, the quotient of amino acid representation in the fully hydrolysed protein would theoretically be approximately 3 between proline and any other amino acid. The comparison of amino acid amounts in tissues of stressed animals showed that these quotients were between 3 and 25 (Pro versus Leu: 10, Ile: 25, Arg: 3, Gly: 5, Ser: 3, Thr: 15, Val: 15) in FW or between 6 and 80 (Pro versus Leu: 20, Ile: 80, Arg: 8, Gly: 6, Ser: 8, Thr: 16, Val: 50) in BW animals. Even when considering that the molecular masses of the different amino acids are somewhat different, and that actual proteins are not ideally composed of even portions of all 20 amino acids, and under the assumption that these quotients may be secondarily affected by metabolic conversion after mobilisation from storage proteins, the data imply that proline accumulation in the foot muscle of snails under hyperosmotic stress cannot be solely explained by hydrolysis of proline-rich storage proteins.
In animal and plant cells, proline levels under osmotic stress are also controlled by proline synthesis mediated by the glutamate pathway (Szabados and Savouré, 2010). Proline is synthesised from glutamate via delta-1-pyrroline-5-carboxylate (D-1-P5C) through two sequential reduction reactions catalysed by the D-1-P5C synthase and the P5C reductase. The genes encoding these enzymes have been identified in bivalves and were transcriptionally activated during osmotic stress (Meng et al., 2013). We have identified transcript sequences of T. fluviatilis homologues of D-1-P5C synthase and D-1-P5C reductase in a transcriptome database (Wiesenthal, 2018) of these snails (GenBank accession number: MK316364; MK316365; MK316366), which suggests that proline may indeed be rapidly synthesised from glutamate during hyperosmotic stress. The sequences associated with potential homologues of the enzyme converting proline to glutamate, proline dehydrogenase and D-1-P5C dehydrogenase, have also been identified in the T. fluviatilis transcript database (GenBank accession number: MK316367; MK316368; MK316369). Proline synthesis from glutamate is supported by the observation that glutamate levels do not change much in BW animals under osmotic stress, whilst glutamate levels undergo similar changes to other amino acids in tissues of stressed FW animals (Fig. 3A). Proline synthesis and accumulation in foot muscle cells may also be supported by another metabolic pathway leading from arginine to D-1-P5C via ornithine (Li and Wu, 2018). This pathway has also been identified as relevant for proline accumulation in the ribbed mussel (Geukensia demissa) (Bishop et al., 1994). Urea is generated as a by-product of arginine conversion to ornithine. Because urea accumulation under hyperosmotic stress was observed in both FW and BW animals (Fig. 5C) to a certain extent, it is likely that this pathway of proline synthesis may be generally used for supporting proline synthesis and proline accumulation.
This does not rule out the possibility that some of the accumulated proline is derived from hydrolysis of proline-rich storage proteins. Some of the proline residues become hydroxylated by prolyl 4–hydroxylase (Gorres and Raines, 2010) while they are integral protein constituents (Li and Wu, 2018). Thus, the occurrence of free hydroxy-proline in body fluids of animals is generally considered as an indication of accelerated protein degradation (Holm and Kjaer, 2010). As shown in Fig. 5B, hydroxy-proline accumulated to a large extent in FW animals during hyperosmotic stress, whilst no statistically significant change in hydroxy-proline abundance was observed in BW animals. This indicates that the proline accumulation observed in the foot muscle tissue of hyperosmotically challenged FW specimens of T. fluviatilis (Fig. 7A) must be largely derived from storage protein degradation, while the higher accumulation of proline in stressed BW snails cannot be explained by this process, as there was no increase of hydroxy-proline (Fig. 5B). Thus, the proline accumulation in BW snails can be explained by synthesis or transamination of this amino acid. Accelerated turnover and transamination via the glutamate or arginine pathway will lead to increased amounts of ammonia that are largely converted to urea (Bishop et al., 1994; Bröer and Bröer, 2017; Haggag and Fouad, 1968; Horne and Boonkoom, 1970). This is in accordance with the observed urea levels in FW and BW snails (Fig. 5C). While a moderate increase in urea levels could be explained by proline synthesis via the arginine and ornithine pathway in snails from both groups, the BW snails showed a much higher accumulation than the FW ones. As individuals of the FW group only showed a small increase in urea, but a substantial one in hydroxy-proline, they seem to rely on the hydrolysis of storage proteins. BW individuals, in contrast, showed a great increase in urea, yet no difference in the hydroxy-proline accumulation under hyperosmotic stress compared with control animals. This indicates that urea is accumulated as a by-product in the process of proline and alanine synthesis (Figs 5 and 6A).
Bishop et al. (1994) showed that in the ribbed mussel (G. demissa), alanine could quickly be synthesised from pyruvate and that the accumulation of this amino acid was most likely based on protein turnover, while proline could originate from synthesis or protein breakdown. This is in accordance with our findings for FW snails that seem to acquire their alanine and proline under hyperosmotic stress through a combination of protein hydrolysis and amino acid synthesis, while BW snails rely more on synthesis alone.
Because urea, in high concentrations of several hundred millimoles per kilogram, has a perturbing effect on protein conformation, it is known from other organisms that certain compatible osmolytes, such as TMAO, glycine betaine, sarcosine, glycerol, inositol or GPC, are accumulated along with urea to counteract its negative effect (Yancey, 2005; Withers and Guppy, 1996). The textbook example for this behaviour is the group of elasmobranchs that display high TMAO concentrations in their body fluids in parallel with high urea concentrations. It has been suggested that the best counteracting function of TMAO occurs at a molar ratio of 2:1 for urea:TMAO (Tam et al., 2003; Hazon et al., 2003; Yancey, 2005). Aestivating amphibians also accumulate urea, though not quite as much as the T. fluviatilis in this study (Withers and Guppy, 1996). Similar to aestivating desert frogs, however, T. fluviatilis did not accumulate any TMAO along with urea under hyperosmotic stress. In fact, none of the measured potential counteracting solutes listed above showed any noticeable increase with rising urea levels. The only exception was myo-inositol, where an increase was detected (Fig. 7B) that was similar to that observed in aestivating Neobatrachus (Withers and Guppy, 1996). In both cases, the ratio of inositol to urea was 1:14. Whether such a ratio is meaningful in terms of stabilising protein structure in gastropods or amphibians is not clear. Khan et al. (2013) have shown that a ratio of 1:2 (myo-inositol:urea) is needed to counteract the perturbing effect of urea on vertebrate proteins. In the same study, however, it became clear that urea concentrations of 500–600 mmol l−1 – as reached in stressed T. fluviatilis (Fig. 5C) – only had minor effects on the thermal stability of these proteins, and that higher urea concentrations (>800 mmol l−1) are needed to shift the denaturing temperature to substantially lower values. Yancey and Somero (1980) stated that β-alanine also has a counteracting effect, even though it is not as strong as that of sarcosine or betaine. When β-alanine levels correspond to about 8.5% of the urea levels, a protein-stabilising effect is observable (Tam et al., 2003). In this study, the measured levels of β-alanine only amounted to about 2.6% of the urea concentration. Therefore, just like for myo-inositol, the role of β-alanine as a compatible osmolyte in T. fluviatilis under hyperosmotic stress remains elusive. Even though taurine was not considered as particularly relevant for the overall osmolyte content, it might contribute to counteracting the perturbing effects of urea (Khan et al., 2013). Despite the increase not being very prominent and the total accumulated amount lying below the predicted necessary ratio of 2:1 (urea:taurine), only BW snails showed an accumulation of taurine under hyperosmotic conditions parallel to an accumulation of urea (Fig. 6B) (Khan et al., 2013). This snail species may either offset the perturbing effect of urea through a mixture of β-alanine, myo-inositol, taurine and other compatible osmolytes, which are yet unknown, or exhibit a lower protein sensitivity towards urea, as has been described in desert frogs (Withers and Guppy, 1996). Which osmolytes besides myo-inositol, β-alanine and taurine could be involved is not yet known. The potential roles of other polyols (mannitol, sorbitol) (Rees and Hand, 1993) remain tested in future studies. An interesting aspect is also the location of the organic osmolyte accumulation in the tissue, because the measurements represent the load in both intracellular and extracellular fluids and do not distinguish between them. While the FAAs, the derivatives and compatible osmolytes are mainly accumulated within the cell, urea is most likely evenly distributed between intracellular and extracellular compartments as it can diffuse through the cell membrane (Gallucci et al., 1971). Therefore, the counteracting ability of these compatible osmolytes will only be effective within the cell, while urea will also have perturbing effects on extracellular proteins. When considering the distribution of organic osmolytes and urea, the mixture of the compatible osmolytes myo-inositol, β-alanine and taurine may be the only intracellular antagonists to urea in BW snails. Nevertheless, further counteracting substances must be present in the extracellular fluids to ensure protein stability.
Taken together, the differences in the upper tolerance limits of environmental salinities that have been previously observed in FW and BW snails (Wiesenthal et al., 2018) cannot be unambiguously explained by differences in the general ability to accumulate amino acids as organic osmolytes. Although there may be differences in the metabolic pathways involved in osmolyte accumulation in foot muscle tissue, the two groups of animals accumulate organic osmolyte mixtures equally well when stepwise acclimated to the maximum salinity they can tolerate over extended periods (Fig. 2). The amino acids alanine and proline as well as the secondary metabolite urea are of special quantitative importance for cell volume preservation in T. fluviatilis under hyperosmotic stress. Alanine may be somewhat more important for FW animals (Fig. 6A), while proline and urea may have a greater significance in BW animals (Fig. 5A,C). It seems that basal accumulation of amino acids during hyperosmotic stress occurs via hydrolysis of storage proteins, but alanine and proline in particular are probably newly synthesised. The relative contribution of protein hydrolysis and alanine/proline synthesis may be different in animals collected at the FW and BW sites.
For future studies, it would be interesting to test for the rate of osmolyte accumulation and to follow the metabolites over the entire time of hyperosmotic stress to further elucidate the pathways of osmolyte generation and metabolism.
Results of measurements of ninhydrin-positive substances as well as some figures and discussion points in this paper are reproduced or modified from the PhD thesis of A.A.W. (University of Greifswald, 2018). We would like to thank all members of the working group of ‘Animal Physiology and Biochemistry’ at the University of Greifswald for their help with the snail collections. A special thanks goes to Alexander Kolb for helping with snail maintenance. We also gratefully thank PD Dr Blindow (Biological Station Hiddensee) for obtaining permission to collect snails on the island. Permission for snail collections was given by Nationalparkamt Vorpommersche Boddenlandschaft, Born, Germany (24–5303.3). We would also like to thank the reviewers for their constructive and helpful comments.
Conceptualization: A.A.W., C.M., J.-P.H.; Methodology: A.A.W., J.-P.H.; Validation: A.A.W.; Formal analysis: A.A.W., C.M., J.-P.H.; Investigation: A.A.W.; Resources: J.-P.H.; Data curation: A.A.W., K.H.; Writing - original draft: A.A.W.; Writing - review & editing: A.A.W., C.M., J.-P.H.; Visualization: A.A.W.; Supervision: C.M., J.-P.H.; Project administration: A.A.W.; Funding acquisition: C.M., J.-P.H.
This research was funded by the Deutsche Forschungsgemeinschaft Research Training Group RESPONSE (DFG GRK2010).
The authors declare no competing or financial interests.