Introduction
The search for life on Mars and the icy worlds centres on the presence of liquid water with chemical and environmental conditions suitable for living systems on Earth. Special Regions have been defined for Mars that extend to the known chemical and physical extremes of growth for microorganisms (Rummel et al. Reference Rummel2014). On Mars, liquid water near the surface is expected to be salty, since freezing-point depression keeps brines liquid at lower temperatures (Chevrier & Altheide Reference Chevrier and Altheide2008). Dry deliquescing salts can absorb moisture from the atmosphere and may form heavy brines under appropriate humidity and temperature conditions, although eutectic brines of known salts on Mars may be too cold to be habitable (Chevrier & Altheide Reference Chevrier and Altheide2008; Davila et al. Reference Davila2010). Melted permafrost may form transient ground waters that are expected to be high in salts, with the greatest potential for liquid water in gullies and hillside streaks (McEwen et al. Reference McEwen, Ojha, Dundas, Mattson, Byrne, Wray, Cull, Murchie, Thomas and Gulick2011). While it is more difficult to estimate the concentration of salts in the oceans of the icy worlds, heavy brines are likely present (McCord et al. Reference McCord, Hansen and Hibbitts2001; Muñoz-Iglesias et al. Reference Muñoz-Iglesias, Prieto-Ballesteros and Bonales2014; Hand & Carlson Reference Hand and Carlson2015; Hsu et al. Reference Hsu2015).
Hypersaline regions on Earth are primarily rich in NaCl. This is not the case for Mars, and perhaps the icy worlds, although sodium and chloride salts are also present (McCord et al. Reference McCord, Hansen and Hibbitts2001; Osterloo et al. Reference Osterloo, Hamilton, Bandfield, Glotch, Baldridge, Christensen, Tornabene and Anderson2008; Muñoz-Iglesias et al. Reference Muñoz-Iglesias, Prieto-Ballesteros and Bonales2014). Sulphate salts of calcium, iron and magnesium tend to predominate in near-surface Mars regolith (Clark et al. Reference Clark2005; Gendrin et al. Reference Gendrin2005). Calcium, iron and magnesium salts of perchlorate and chlorate also are significant on Mars (Chevrier et al. Reference Chevrier, Hanley and Altheide2009; Gough et al. Reference Gough, Chevrier, Baustian, Wise and Tolbert2011; Clark & Kounaves Reference Clark and Kounaves2015). While the main solutes on Mars and the icy worlds are ionic salts, habitable regions to-date have been defined in part using microbial tolerances obtained with non-ionic solutes. The lowest water activity at which microbial (fungal) growth has been observed was reached in a medium supersaturated with sucrose and fructose (Pitt & Christian Reference Pitt and Christian1968). It is not clear if this demonstration is relevant to ionic solutions or to bacteria.
Specific solute effects elicit cellular responses that are particular to the solute used (Scott Reference Scott1953; Beuchat Reference Beuchat1981; Chirife Reference Chirife1994; Chirife et al. Reference Chirife, González and Resnik1995; Williams & Hallsworth Reference Williams and Hallsworth2009; Cray et al. Reference Cray, Russell, Timson, Singhal and Hallsworth2013). Solute concentrations and low water activities alone do not explain the inhibition of microbial growth in highly osmotic media. Where solutes fall on the Hofmeister series (effect of ions on protein precipitation) suggest either a disruptive (chaotropic) or stabilizing (kosmotropic) effect on macromolecules (Hofmeister Reference Hofmeister1888; Lewith Reference Lewith1888). An agar gelation assay has been suggested as a measure of chaotropicity (Cray et al. Reference Cray, Russell, Timson, Singhal and Hallsworth2013); however, the mechanisms by which solutes affect cellular functions seem to be more complex than can be described by a single physicochemical assay. The kosmotropic solutes glucose, MgSO4, NaCl and sucrose have been most widely studied, along with the chaotrope fructose, and glycerol, which behaves differently depending on concentration (Scott Reference Scott1953; Harris Reference Harris, Parr, Gardner and Elliott1981; Belamri et al. Reference Belamri, Mekkaoui and Tantaoui-Elaraki1991; Chirife Reference Chirife1994; Williams & Hallsworth Reference Williams and Hallsworth2009; Lievens et al. Reference Lievens, Hallsworth, Pozo, Belgacem, Stevenson, Willems and Jacquemyn2015). Growth at record-low a w was observed with xerophilic (arid-tolerant) yeasts in high sugars or in mixtures of glycerol with other solutes (Pitt & Christian Reference Pitt and Christian1968; Grant Reference Grant2004).
The current study examines the growth of salinotolerant (able to grow at high-salt concentrations) bacteria in high concentrations of sugars and salts. We compare growth rates based on the degree of saturation and the water activity of the growth media. Our study does not work at the extremes of tolerance to high osmolarity or seek to expand the definition of Special Regions. Instead, we examine the differential responses observed with solutes of varying qualities, and mixtures of these solutes, on a collection of salinotolerant bacteria. We introduce two new terms for describing sugar tolerance. Sucretolerance is the term used here for growth tolerance to high concentrations of sugar. Sucrephilic organisms would be those that cannot grow without high concentrations of sugar.
Materials and methods
Organisms
Previous work in hypersaline environments provided the isolate collections that were screened in the current study. The Great Salt Plains (GSP) of Oklahoma collection of aerobic heterotrophic bacteria includes 93 halotolerant (able to grow in high concentrations of NaCl) isolates from a wet terrestrial environment saturated with NaCl (Caton et al. Reference Caton, Witte, Ngyuen, Buchheim, Buchheim and Schneegurt2004). A similar collection of 64 bacterial isolates was obtained from lake margins and sediments of Hot Lake (HL), WA, an environment saturated in MgSO4 (epsomite) (Kilmer et al. Reference Kilmer, Eberl, Cunderla, Chen, Clark and Schneegurt2014). A group of 12 HL isolates were chosen for further study based on salt tolerance and taxonomic spread (Table 1).
Table 1. Hot Lake (HL) bacterial isolates used in the current study
Media and growth measurements
Artificial nectar (AN) media initially were developed to mimic flower nectar, a natural high-sugar environment. AN series media contained at least 0.5% (w/v) of each fructose, glucose and sucrose. This was then supplemented with sucrose to produce a series of media with a designation indicating % sucrose (w/v), such as AN50, which contains 50% sucrose. AN media also were supplemented with amino acid sources (0.5% (w/v) of both yeast extract and tryptone). Sugar solutions and amino acid solutions were autoclaved separately to prevent browning, presumably due to Maillard reaction products. Media made this way were lightly coloured and caramelization was not substantial. A low-nutrient oligoAN medium was supplemented with only 0.05% (w/v) of both yeast extract and tryptone. The AN medium was further supplemented with salts to produce the ANS media series that contained (per litre) 2.0 g NH4SO4, 2.0 g KCl, 0.36 g CaCl2•2 H2O, 0.16 g MgCl2, 1 mg FeCl3•6 H2O and 0.5 ml trace minerals (per litre; 2 mg CoCl2, 10 mg CuSO4•6 H2O, 200 mg MnCl2•6 H2O, 100 mg NaMO4•2 H2O, 100 mg ZnSO4•7 H2O). Additional salts or sugars were added to ANS as described in the ‘Results’ section. A filter-sterilized antifungal cocktail was added to AN series media from a 100× stock (per 100 ml; 1.0 g cycloheximide, 1.0 g carbendazim and 0.16 g nystatin). Salt Plains (SP) artificial seawater media have been described previously (Caton et al. Reference Caton, Witte, Ngyuen, Buchheim, Buchheim and Schneegurt2004), with 5–30% NaCl added to produce SP5 through SP30 media. Isolates were grown as duplicate shake-tubes in 3 ml of media in a moist box on a rotating platform (150 rpm). Culture density was determined visually or by absorbance at 600 nm on a Genesys 10S spectrophotometer (ThermoFisher) using a medium blank and a threshhold of 0.1 OD units for positive growth.
Water activity and agar gelation
The water activity of each medium was measured using an AqualLab Series 3 water activity meter (Decagon Devices, Inc., Pullman, WA). The instrument was calibrated with standard NaCl solutions and run at room temperature. Agar gelation assays were performed as previously described (Hallsworth et al. Reference Hallsworth, Heim and Timmis2003; Crisler et al. Reference Crisler, Newville, Chen, Clark and Schneegurt2012). Measurements were done in triplicate, with SDs being small enough to fit within the symbols on the curves presented. Chaotropic effect can be calculated using a conversion factor of 4.15 kJ g−1 °C−1 for a 1.5% agar gel (Cornillon et al. Reference Cornillon, Andrieu, Duplan and Laurent1995).
Statistical analyses
Paired t-tests performed within SAS (v9.2) were used to compare the results of sugar and salt tolerance tests for the 44 HL isolates reported here. Statistical comparisons of the selected 12 HL isolates in five different media (fructose, glucose, NaCl, sorbose and sucrose) also were performed using paired t-tests.
Results
Screening for sucretolerance
Growth tolerances to 50% sucrose for bacterial isolates from the GSP and HL are given in Table 2. The a w for media with 50% sucrose is 0.91, just slightly lower than the a w for media with 10% NaCl (Table 3). All of the GSP and HL isolates have been shown previously to grow in media containing 10% NaCl (Caton et al. Reference Caton, Witte, Ngyuen, Buchheim, Buchheim and Schneegurt2004; Kilmer et al. Reference Kilmer, Eberl, Cunderla, Chen, Clark and Schneegurt2014). However, only a fraction of these isolates grew in medium containing 50% sucrose, with 85 and 89% of isolates from HL and the GSP, respectively, failing to grow in AN50. Adding a basal salts mixture to AN50 enhanced growth, but 76 and 82% of isolates from HL and the GSP, respectively, still did not grow in ANS50. The oligotrophic medium, oligoAN50, had one-tenth the amino acid content of AN50 and it did not support the growth of any GSP isolates, and only two Hot Lake isolates grew (Bacillus licheniformis and Corynebacterium variabilis).
Table 2. Screening of salinotolerant bacterial isolates for growth tolerance to media supplemented with 50% sucrose
Table 3. Water activities of AN media supplemented with various concentrations of sugars or salts
These observations were extended for 44 HL isolates (Table 4). Screening for growth tolerances was performed in media containing a range of sucrose and NaCl concentrations. All isolates grew at 5 and 10% NaCl, with only three failing to grow at 20% NaCl (a w of 0.85). Nine of these HL isolates grew at 30% NaCl with an a w of 0.76. In contrast, only 52% of these isolates grew at 30% sucrose, which has an a w of 0.96. At the highest concentration, 70% sucrose (a w of 0.90), only seven isolates grew. The responses to sugars and salts by this collection of isolates showed statistically significant differences (P = 0.001).
Table 4. Growth of select Hot Lake isolates at various concentrations of NaCl or sucrose
Growth in other sugars
A group of 12 HL isolates (Table 1) was selected for more detailed comparisons of growth tolerances to high concentrations of sugars and salts. The highest solute concentrations that were tolerated by these isolates are given in Table 5 and the a w for each medium is given in Table 3. An agar gelation assay was used to determine the chaotropicities of the sugar solutes included in this study (Fig. 1). Fructose and sorbose were chaotropic by this measure at concentrations above 20%. Glucose and sucrose were kosmotropic at lower concentrations and appeared to be more chaotropic at the highest concentrations examined.
Fig. 1. The chaotropicities of different sugar solutions were determined by their effects on the gelation temperature of agar.
Table 5. Maximum growth tolerances of select Hot Lake isolates to various sugars and salts
All twelve isolates grew in 30% sucrose, with five growing in a supersaturated ANS70 medium. Note that HL 12, 14 and 82 are all Halomonas, with HL 12 and 14 clustering closely on the phylogenetic tree, yet their tolerances to growth in high sucrose differ substantially. While HL 12 grows in ANS70, HL 14 and 82 did not grow in ANS40 or at higher sucrose concentrations. Planococcus isolates HL 20 and 91 showed a similarly disparate response, with HL 91 growing in ANS70 but HL 20 limited to ANS30. In contrast, each pair of Bacillus, Marinococcus and Nesterenkonia isolates showed approximately equal growth tolerances to sucrose.
Growth tolerances to high concentrations of glucose and fructose were determined for this group of 12 HL isolates (Table 5). Overall tolerance to these sugars was lower than for sucrose, despite similar water activities. Several of the isolates failed to grow in ANS30 with glucose or fructose, the lowest sugar concentration tested. Only HL 12 grew above 40% glucose or fructose (a w 0.94–0.95). As with sucrose, HL 12 outperformed the other Halomonas isolates, HL 14 and 82. Planococcus isolate HL 91 again outperformed HL 20, but to what degree is not clear since sugar concentrations below 30% were not tested. Tolerances to glucose and fructose differed in some cases, the most dramatic differences seen with HL 12 and 64. Greater differences were seen between sucrose tolerance and the other two sugars, with certain isolates (HL 11, 68 and 91) growing at 70% sucrose, but growing up to only 30% glucose or fructose. All of the differences between responses to various sugars were statistically significant (P < 0.0001–0.045).
Separate tests showed that fructose, glucose and sucrose could all be used for growth as a sole carbon source by these 12 HL isolates (data not shown). These tests also showed that sorbose and cellobiose (β-1,4-glucose) could not be used as sole carbon sources for growth. The a w for media supplemented with sorbose were similar to those of the other monosaccharides (Table 3). Sorbose was more inhibitory to growth than the other sugars tested (Table 5). Most isolates did not grow in a medium supplemented with 20% sorbose. Only HL 11 grew above 20% sorbose. Cellobiose was not as soluble as the other sugars and 10%, the highest concentration obtained, did not substantially reduce a w (>0.99). Four isolates did not tolerate even 1% cellobiose and seven grew at 10% cellobiose.
Growth in salts
HL is an environment rich in MgSO4 and all of the isolates from this area were salinotolerant to both NaCl and MgSO4 (Kilmer et al. Reference Kilmer, Eberl, Cunderla, Chen, Clark and Schneegurt2014). Each of the 12 selected HL isolates grew in a medium with 10% NaCl and all but three grew at 20% NaCl (a w 0.85). Two HL isolates grew at nearly saturated NaCl. Different halotolerances among isolates in the same phylogenetic clade has been noted previously in the HL and GSP collections (Caton et al. Reference Caton, Witte, Ngyuen, Buchheim, Buchheim and Schneegurt2004; Litzner et al. Reference Litzner, Caton and Schneegurt2006; Kilmer et al. Reference Kilmer, Eberl, Cunderla, Chen, Clark and Schneegurt2014), and different sucretolerances were noted among isolates from the same clade in the current report. Tolerances to MgSO4 were generally high among the HL isolates, with half growing at near saturation. Note that MgSO4 does not reduce a w to the same extent as NaCl (Table 3), since MgSO4 does not effectively dissociate in water.
There was not a clear correspondence between sucretolerance and halotolerance for these isolates. All of the differential responses to NaCl and various sugars were statistically significant (P < 0.016). The comparisons of growth tolerances are best expressed based on the degree of saturation, expressed as a fraction of the corresponding saturating solute concentrations (Fig. 2). For certain isolates (HL 11 and 54) sucretolerance and halotolerance were both relatively high. For other isolates (HL 80 and 82), while overall tolerance was lower, tolerances to sugar and salt were similar. In contrast, there were isolates (HL 12, 55, 68 and 91) where sucretolerance was far greater than halotolerance based on the degree of saturation. Some isolates (HL 20, 64 and 76) showed the opposite relationship, having higher halotolerance than sucretolerance.
Fig. 2. Maximum growth tolerances of salinotolerant bacterial isolates to NaCl (solid bars) and sucrose (open bars) expressed as degree of saturation.
Tolerance to sugar and salt mixtures
A series of ANS media were prepared with varying mixtures of sucrose and NaCl at different ratios. Sucrose concentrations ranged from 10 to 50% and NaCl concentrations ranged from 5 to 15%. The media prepared exhibited a range of a w from 0.97 to 0.75. Growth tolerances for the 12 selected HL isolates were measured in each medium (Table 6). Notice that in several cases, isolates grew with certain mixtures, even when the a w was lower than the a w of other mixtures, which did not support growth. For instance, HL 11, 12, and 54 grew in media with 30% sucrose and 15% NaCl (a w 0.83) but did not grow in media with 40% sucrose and 10% NaCl (a w 0.87). All three grew when the solute was ≤60% sucrose or ≤20% NaCl. HL 64 and 82 performed particularly poorly when exposed to mixtures of solutes, while HL 20, 76 and 80 grew more poorly as the sucrose content increased in the presence of NaCl. None of the isolates grew in the presence of ≥40% sucrose with >5% NaCl. As with the individual solutes, growth responses to mixtures of solutes differed among Halomonas isolates (HL 12, 14 and 82) but were more closely matched within Marinococcus (HL 11 and 54) and Bacillus isolates (HL 55 and 68).
Table 6. Growth tolerances of select Hot Lake isolates to various mixtures of sucrose and NaCl
Discussion
The concept of specific solute effects is not new, having first been recognized by Scott (Reference Scott1953) and later popularized by reviews in the food science literature (Beuchat Reference Beuchat1981; Chirife & Buera Reference Chirife and Buera1996). More recently this concept has been reintroduced as part of a discussion of chaotropic and kosmotropic solutes (Williams & Hallsworth Reference Williams and Hallsworth2009; Cray et al. Reference Cray, Russell, Timson, Singhal and Hallsworth2013). Simply stated, individual solutes exert specific effects on the growth of cells that are not entirely dependent on water activity. For instance, hydrophobic solutes often inhibit growth at higher water activities than hydrophilic solutes. Compounds also may be toxic when taken up by a cell. The Hofmeister series and extrapolations from agar gelation assays may have some value as indicators of specific solute effects, but the mechanisms working on living cells seem more complicated than the processes that underlie these individual physicochemical assays. From another angle, there also have been challenges to the accuracy of water activity measurements themselves. Non-equilibrium states may be maintained through interplay of mobility transformations and water dynamics in substrates that can transition from rubbery states to glass or crystalline states (Chirife & Buera Reference Chirife and Buera1994, Reference Chirife and Buera1996).
The current report demonstrates specific solute effects for sugars and salts that reflect more than water activity alone. Responses by individual bacterial isolates to sucrose and NaCl differed in a way not predicted solely by water activity. Specific solute effects also were observed between sugars. Tolerance to fructose and glucose was similar, but sucrose tolerance often differed among individual isolates. Glucose and sucrose are kosmotropic and fructose is chaotropic by the agar gelation assay, so chaotropicity is not a clear indicator in this instance. Specific solute effects can lead to differential responses to osmotic stress. Tests with various humectants on the growth of Escherichia coli found different responses to mannitol, sorbitol and sucrose only when environmental conditions were non-optimal (Buchanan & Bagi Reference Buchanan and Bagi1994). Survival and growth of Chrysosporium followed this pattern, glucose > sorbitol > glycerol > NaCl (Beuchat & Pitt Reference Beuchat and Pitt1990). Other factors that may lead to specific solute effects include viscosity, oxygen solubility, oxygen diffusivity, dielectric effects, molecular crowding, ionic effects and cellular uptake. It is interesting to note that the non-metabolizable sugar sorbose showed the greatest negative effects on growth of the organic solutes tested here. Sorbose seems to have chaotropic properties similar to fructose, but cellular responses are very different.
Previous studies using very high solute concentrations have mainly focused on organisms that are the most tolerant to growth at low water activities. The definition of Special Regions on Mars depends on the results of experiments designed to determine the extremes of microbial tolerance to the harsh chemical and physical conditions near the surface of Mars (Rummel et al. Reference Rummel2014). While this is valuable for both life detection missions and planetary protection efforts, organisms that can grow at the lowest extremes of water activity are rare in the environment. It is more likely that spacecraft will be contaminated by organisms commonly found in clean rooms or neighbouring environments. In some cases, species in the current report were observed in spacecraft assembly facilities (Caton et al. Reference Caton, Witte, Ngyuen, Buchheim, Buchheim and Schneegurt2004; Kilmer et al. Reference Kilmer, Eberl, Cunderla, Chen, Clark and Schneegurt2014). Our initial work has found that osmotolerant microbes are widespread in common environments and are not limited to extreme environments (Porazka et al. Reference Porazka, Kilmer and Schneegurt2011; Fredsgaard et al. Reference Fredsgaard, Moore, Kurz and Schneegurt2013).
We observed that specific solute effects were expressed differently among the isolates. At similar water activities, individual isolates showed greater tolerances to salts or sugars, and variations in their responses to different salts and sugars. A key point is that different patterns of tolerance were observed among isolates from the same taxonomic cluster. Similar variation was observed between species of stenohaline bacteria grown at high concentrations of NaCl or glycerol (Marshall et al. Reference Marshall, Ohye and Christian1971). Initial work with salinotolerant bacteria directly isolated from a spacecraft assembly facility at Jet Propulsion Laboratory suggests that tolerance to sugars exhibits specific solute effects that differ among the isolates (T.C. Eberl & M.A. Schneegurt, unpublished observations).
The suggestion is that phylogenetic placement is not a firm indicator of tolerance patterns to media high in solutes. Furthermore, the nature of a solute is important for toxicity beyond the effects of lowering water activity. It may not be wise to rely on growth tolerance data from yeasts at the very extremes of sugar concentrations to predict the responses of bacteria to high-salt concentrations. Cellular responses to mixtures of solutes are more difficult to predict, as was observed here with mixtures of sucrose and NaCl. It seems that the relationships can be more complex than a balance of chaotropicities. These considerations should be part of discussions about planetary protection guidelines or when defining Special Regions on Mars and the icy worlds.
Acknowledgements
The authors are grateful for the contributions of Todd Caton, Timothy Eberl, Brian Kilmer, Tammy Kurz, Hieu Nguyen, Christopher Rogers, and Noah Schneegurt. We thank Fadi Aramouni for performing water activity measurements. Preliminary accounts of this work have been presented previously and abstracted (Fredsgaard et al. Reference Fredsgaard, Moore, Kurz and Schneegurt2013, Reference Fredsgaard, Moore and Schneegurt2014). This work was supported by awards from NASA ROSES Planetary Protection Research (09-PPR09-0004 and 14-PPR14-2-0002), Kansas INBRE NIH NIGMS IDeA (P20 GM103418) and The Flossie E. West Memorial Trust Foundation.
