Rapid adaptation of phytoplankters to geothermal waters is achieved by single mutations: were extreme environments ‘Noah's Arks’ for photosynthesizers during the Neoproterozoic ‘snowball Earth’?

Authors

  • Eduardo Costas,

    1. Genética (Producción Animal), Facultad de Veterinaria, Universidad Complutense, E–28040 Madrid, Spain;
    Search for more papers by this author
    • *

      The three authors contributed equally to this work.

  • Antonio Flores-Moya,

    1. Biología Vegetal (Botánica), Facultad de Ciencias, Universidad de Málaga, Campus de Teatinos s/n, E–29071 Málaga, Spain
    Search for more papers by this author
    • *

      The three authors contributed equally to this work.

  • Victoria López-Rodas

    1. Genética (Producción Animal), Facultad de Veterinaria, Universidad Complutense, E–28040 Madrid, Spain;
    Search for more papers by this author
    • *

      The three authors contributed equally to this work.


Author for correspondence:
A. Flores-Moya
Tel:+34 952131951
Fax:+34 952131944
Email: floresa@uma.es

Summary

  • • Geothermal waters often support remarkable communities of microalgae and cyanobacteria apparently living at the extreme limits of their tolerance. Little is known about the mechanisms allowing adaptation of mesophilic phytoplankters to such extreme conditions, but recent studies are challenging many preconceived notions about this. The aim of this study was to analyse mechanisms allowing adaptation of mesophilic microalgae and cyanobacteria to stressful geothermal waters.
  • • To distinguish between the pre-selective or post-selective origin of adaptation processes allowing the proliferation of mesophilic phytoplankters in geothermal waters, several Luria–Delbrück fluctuation analysis were performed with the microalga Dictyosphaerium chlorelloides and the cyanobacterium Microcystis aeruginosa, both isolated from nonextreme waters. Geothermal waters from seven places in Italy and five icebound places at Los Andes in Argentina were used as selective agents.
  • • Physiological adaptation was achieved in the least toxic waters. In contrast, rapid genetic adaptation was observed in waters ostensibly lethal for the experimental organisms. This adaptation was achieved as consequence of single mutations at one locus.
  • • It was hypothesized that a similar mechanism of rapid genetic adaptation could explain the survival of photosynthetic life during the Neoproterozoic ‘snowball Earth,’ where geothermal refuges such as those studied could have been ‘Noah's Arks’ for microalgae and cyanobacteria.

Introduction

Extreme aquatic environments often support remarkable communities of phytoplankters living at the extreme limits of their tolerance (Seckbach & Oren, 2007). Survival and growth of such microorganisms in habitats characterized by extreme values of pH, toxic mineral concentrations, temperature, salinity and other stress factors, is an interesting topic from both biochemical and physiological points of view (Fogg, 2001). Since algae and cyanobacteria are the principal primary producers of aquatic ecosystems (Kirk, 1994; Falkowski & Raven, 1997), the capability of phytoplankters to proliferate in these extreme natural environments is very relevant for understanding these intriguing communities.

Little is known about the mechanisms allowing adaptation of phytoplankton to such extreme conditions. Phytoplankton organisms can survive in adverse environments as a result of physiological adaptation (i.e. acclimatization) supported by modifications of gene expression (Bradshaw & Hardwick, 1989; Fogg, 2001). However, when the values of some environmental factors exceed the physiological limits, survival depends exclusively on adaptive evolution, by the occurrence of mutations that confer resistance (Sniegowski & Lenski, 1995; Belfiore & Anderson, 2001). It is accepted that short-term or fluctuating stress is best met by physiological adaptation, while continuous or predictable stress can be met by genetically determined response systems (Bradshaw & Hardwick, 1989; Davison & Pearson, 1996). Moreover, it is usually assumed that genetic adaptation to extreme environments is achieved very slowly. However, recent studies are changing many preconceived notions about the adaptation of microalgae to extreme environments. As an example, eukaryotic microalgae (rather than extremophile bacteria) contributed the highest fraction of the biomass (at least 60%) in the Rio Tinto (Amaral Zettler et al., 2002), an extremely acidic (pH 1.7–2.5) environment with a very high heavy metal content (Gónzalez-Toril et al., 2003). Molecular studies show that these microalgae are closely related to neutrophilic species rather than acidophilic lineages. For this reason, it was proposed that adaptation from neutral to extreme environments must occur rapidly (Amaral Zettler et al., 2002). Moreover, it was suggested that microalgae resistant to Rio Tinto waters arose randomly by rare spontaneous mutations and, as a result, algal populations were able to rapidly adapt to Rio Tinto water by means of selection of resistant mutants growing in nonextreme conditions (Costas et al., 2007). It also has been proposed that rare pre-selective mutants can be sufficient to ensure the adaptation of mesophilic algae to other extreme natural habitats. Flores-Moya et al. (2005) demonstrated that algae inhabiting the extreme acidic, sulphurous water from La Hedionda Spa (southern Spain) could originate by selection of pre-selective mutants of mesophilic algal lineages inhabiting nonextreme environments. Similarly, microalgal adaptation to the stressful acidic, metal-rich mine waters from Mynydd Parys (North Wales, UK) and Aguas Agrias stream (southwest Spain) is also caused by selection of preselective mutants of mesophilic algal lineages that arose previous to toxic water exposure (López-Rodas et al., 2008a,b).

The aim of this study was to determine whether mesophilic cyanobacteria and microalgae are capable of rapid adaptation to extreme environments by single mutations. For this purpose, cosmopolitan, mesophilic phytoplankton species isolated from nonextreme waters were selected; specifically, Dictyosphaerium chlorelloides was isolated from a pristine, slightly alkaline (pH 8.0) mountain lake from Sierra Nevada (southern Spain) while the cyanobacterium Microcystis aeruginosa was isolated from a pristine pond with nonacidic waters (pH 8.1) in Doñana National Park (south-west Spain). In both species, we analysed adaptation to seven different geothermal systems from Italy (including warm ponds, hot springs, seltzer springs and extremely acid hot springs) and five icebound geothermal waters at Los Andes in Argentina (including warm ponds, hot springs, seltzer hot springs and geysers). We studied whether the mesophilic chlorophyceans and cyanobacteria could adapt to survive and grow in these geothermal waters, by using an experimental design known as fluctuation analysis (Luria & Delbrück, 1943). This experimental approach allowed us to differentiate between physiological and genetic adaptation. It was found that in the majority of the extreme geothermal waters, adaptation was achieved by the photosynthetic mesophiles through very rapid genetic adaptation caused by single mutations. In addition, we hypothesize that a similar mechanism of rapid genetic adaptation could explain the survival of photosynthetic life during the ‘snowball Earth,’ where geothermal refuges, such as the ones studied in Italy and Los Andes, could be ‘Noah's Arks’ for cyanobacteria and microalgae.

Based on geologic and paleomagnetic evidence, the ‘snowball Earth’ hypothesis proposes that a series of global glaciations occurred during the Neoproterozoic era, between c. 740 million years ago (Mya) and 580 Mya, with the ice line reaching the Equator and with a sea-ice cover > 100 m in thickness in tropical latitudes (Kirschvink, 1992; Hoffman et al., 1998). Although Hyde et al. (2000) suggested a climatic scenario in which a partly frozen Earth had ice-free oceans at the Equator (‘slushball Earth’), it seems that only the ‘snowball Earth’ hypothesis can explain all the geological and paleomagnetic data of Neoproterozoic glacial deposits (Schrag & Hoffman, 2001). The ‘snowball Earth’ model implies a drastic survival pressure on the photosynthetic biota because liquid water and sunlight were not available simultaneously, for millions of years, in any place on the Earth. In fact, it is supposed that primary production collapsed (Schrag & Hoffman, 2001), as revealed by negative carbon isotope anomalies in carbonate rocks (Kaufman et al., 1997; Hoffman et al., 1998; Rothman et al., 2003). However, photosynthetic prokaryotic (cyanobacteria) and most eukaryotic phyla (including green, red and chromophytic algae) evolved before the late Neoproterozoic glaciations (Knoll & Bauld, 1989; Knoll, 1992) and they must have survived these extreme environmental conditions.

Recent evidence suggests that global Neoproterozoic glaciations contributed only modestly to the major extinction of autotrophic and heterotrophic eukaryotes (Corsetti et al., 2006). Schrag & Hoffman (2001) proposed that the survival of photosynthesizers during such extended glaciations was achieved in ice-free refuges associated with volcanic activity, such as hot springs and thermal ponds (geothermal areas). A highly adapted community of photosynthetic microbes can develop in geothermal waters – cyanobacteria if the pH is > 4.8, as well as a few genera of eukaryotic ‘cyanidia’ (Brock, 1973; Ward et al., 1998; Castenholz, 2000; Donachie et al., 2002; Sand, 2003; Ciniglia et al., 2005; Walker et al., 2005; Gaylarde et al., 2006; Jing et al., 2006; Lehr et al., 2007; Pinto et al., 2007). However, these stressful environments are lethal for mesophilic cyanobacteria and algae because of extreme values of pH and high temperatures, as well as high concentrations of dissolved heavy metals (Brock, 1978; Dando et al., 1998; Webster & Nordstrom, 2003; Tyrovola et al., 2006). Therefore, if the proposal of Schrag & Hoffman (2001) about the survival of photosynthesizers in ice-free volcanic refuges during the ‘snowball Earth’ is true, it could be hypothesized that mesophilic cyanobacteria and algae have the potential to develop adaptations allowing their survival and growth in these refuges.

Materials and Methods

Sampling sites, in situ analysis of geothermal waters, and algal identification

Seven different geothermal waters from Italy, and five different icebound geothermal waters at Los Andes, Neuquén, Argentina, were studied (descriptions of sampling points are given in Table 1). The values of conductivity, temperature and pH at the sampling sites were determined using a YSI 6820-C-M probe (Yellow Springs, OH, USA). Cyanobacteria and microalgae were identified in fresh samples (directly after collection in each site) using a McArthur portable microscope (Kirk Technology, Cambridge, UK). In addition, 3 l of water was collected from each location, filtered (0.22 µm, Stericup; Millipore Co., Billerica, MA, USA), kept in a closed bottle excluding any air, and stored at 4°C in darkness until the laboratory experiments (toxicity tests and fluctuation analysis; see below) were performed.

Table 1.  Physico-chemical characteristics and description of the geothermal waters sampled to study algal adaptation, algal community inhabiting in the geothermal waters, and inhibitory effect of these waters on Malthusian fitness (m) and effective quantum yield from photosystem II (ΦPSII) of Dictyosphaerium chlorelloides (Dc; Chlorophyceae) and Microcystis aeruginosa (Ma; Cyanobacteria), respectively
Sampling locationDescriptionConductivity (mS)Temperature (°C)pHAlgal communityInhibition of m (% control)Inhibition of ΦPSII (% control)
DcMaDcMa
Geothermal waters from Italy
Bagno VignoniWarm pond with fumaroles3.738.66.0Dense biofilms with high diversity of cyanobacteria and microalgae 17 19 14  15
Amiana MarniHot spring with sulphide7.743.86.7Biofilms with low diversity of cyanobacteria and microalgae100100100100
PienzzaWarm pond with fumaroles7.427.72.5Low diversity of chlorophyceans100100100100
Infierno SujoSeltzer spring with benzene6.021.02.9Low diversity of chlorophyceans100100100100
Sufione PisciarellyAcid hot spring and fumaroles2.430.1 (in the adjacent pond)2.0Not detected100100100100
FangaryWarm pond with fumaroles8.332.15.7Low diversity of cyanobacteria and chlorophyceans100100100100
PuzzolyHot spring with fumaroles6.730.1 (in the adjacent pond)3.2Low diversity of chlorophyceans100100100100
Geothermal waters from Argentina
Doña SaraSeltzer hot spring forming a pond6.559.96.4High diversity of cyanobacteria and microalgae 55 99 64 70
Aguas CalientesWarm pond with fumaroles3.450.45.6Mats of benthic microalgae100100100100
Las PapasWarm pond with sulphur8.236.66.5Floating mats of microalgae100100100100
Los TachosGeyser forming a pond6.346.17.0Microalgae and cyanobacteria100100100100
La MaquinitaHot spring forming a pond1.237.14.0Not detected100100100100

Experimental organisms and culture conditions

A wild-type strain of the chlorophycean D. chlorelloides (Naumann) Komárek and Perman and of the cyanobacterium M. aeruginosa (Kützing) Lemmermann (both strains from the Algal Culture Collection of the Veterinary Faculty, Complutense University, Madrid, Spain) were grown in 100-ml cell culture flasks (Greiner, Bio-One Inc., Longwood, NJ, USA) with 20 ml BG-11 medium (Sigma-Aldrich Chemie, Taufkirchen, Germany), at 22°C under continuous light of 60 µmol m−2 s−1 over the waveband 400–700 nm. Although D. chlorelloides usually forms two- or four-celled (rarely, 16-celled) colonies, and is capable of sexual reproduction in nature, this strain was exclusively propagated by asexual reproduction, and it was represented by single-celled individuals. Cultures were axenically maintained in mid-log exponential growth (Cooper, 1991) by serial transfers of subcultures to fresh medium, and only cultures without detectable bacteria were used in the experiments. The absence of bacteria in the cultures was confirmed periodically (once every week) by epifluorescence microscopy after staining with acridine orange. Before the experiments, the cultures were cloned by isolating a single cell, to avoid including any previous spontaneous mutants accumulated in the culture.

Toxicity test: effect of geothermal waters on Malthusian fitness and effective quantum yield

The changes in effective quantum yield from photosystem II (ΦPSII), and Malthusian fitness (m), were measured when the wild-type strains of D. chlorelloides and M. aeruginosa were cultured in the geothermal water samples. Samples (5 × 105 cells) from mid-log exponentially growing cultures of both species were placed in culture tubes (Sarsted Co., Nümbrecht, Germany) containing 5 ml of the geothermal water. Controls were cultured in BG-11 medium.

The effective quantum yield (ΦPSII) of the strains was measured in triplicates of each geothermal water sample and controls using a ToxY-PAM fluorimeter (Walz, Effeltrich, Germany) after 48 h exposure. Effective quantum yield was calculated as follows:

image(Eqn 1)

(inline image and Ft are the maximum and the steady-state fluorescence of light-adapted cells, respectively; Schreiber et al., 1986).

Malthusian fitness values were also calculated in three replicates in each geothermal water sample as well as in three controls, using the equation of Crow & Kimura (1970):

m = Loge (Nt/N0)/t, (Eqn 2)

(t = 5 d; N0 and Nt are the cell numbers at the start and at the end of the experiment, respectively). Cell number in experiments and controls was counted using a Beckman (Brea, CA, USA) Z2 particle counter.

Adaptation to geothermal waters: fluctuation analysis from sensitivity to resistance

A modified fluctuation analysis (Costas et al., 2001; López-Rodas et al., 2001) was carried out to study the adaptation of the experimental strains to geothermal waters. The modification involves the use of liquid culture medium rather than plating on solid medium (Luria & Delbrück, 1943).

Briefly, for each geothermal water analysed, two different sets of experimental cultures were prepared (Fig. 1). In the set 1 experiments, from 70 to 96 (depending on the strain and geothermal water; see Table 2) culture tubes were inoculated with approx. 102 cells of D. chlorelloides or M. aeruginosa (N0; a number small enough to reasonably ensure the absence of pre-existing mutants in the strain). Cultures were grown in 5 ml BG-11 medium at 22°C until c. 105 cells (Nt). Then, cultures were centrifuged to form a pellet of cells in the tube, the medium was decanted and 5 ml of one of the geothermal waters at 30°C was added to each tube. For the set 2 control, from 25 to 30 aliquots (depending on the species and geothermal water; see Table 2) of c. 105 cells of D. chlorelloides or M. aeruginosa from the same parental populations growing in BG-11 medium at 22°C, were separately transferred to culture tubes containing 5 ml of geothermal water at 30°C. Cultures were observed for 75 d (thus ensuring that one mutant cell could generate enough progeny to be detected, yet not reach the stationary phase) and the resistant cells in each culture were counted.

Figure 1.

Schematic diagram of possible results obtained in the experiment (modified from the classic Luria and Delbrück fluctuation analysis). Set 1, different cultures of Dictyosphaerium chlorelloides and Microcystis aeruginosa (each started from a small inoculum, N0 = 102 cells) were propagated under nonselective conditions (i.e. BG-11 medium) until a very high cell density was reached (Nt ≈ 105 cells), and then transferred to the selective agent (i.e. the different geothermal waters). If resistant cells arose during the exposure to geothermal waters (physiological adaptation), the number of resistant cells in all the cultures must be similar (Set 1A). If resistant cells arose by rare mutations occurring in the period of the propagation of cultures (i.e. before exposure to geothermal waters) the number of resistant cells in all the cultures must be different (Set 1B). In the figure, one mutational event occurred late in the propagation of culture 1 (therefore, the density of geothermal water-resistant cells found is low) and early in the propagation of culture 3 (thus, density of geothermal-resistant cells found is higher than in culture 1); no mutational events occurred in culture 2. Set 2, Different replicates from the same parental culture sampling the variance of the parental population are used as an experimental control. In this case, the number of resistant cells in all the cultures must be similar.

Table 2.  Fluctuation analysis to study adaptation of Dictyosphaerium chlorelloides (Chlorophyceae) and Microcystis aeruginosa (Cyanobacteria) to different geothermal waters from Italy and Argentina
Geothermal waters from ItalyBagno VignoniAmiana MarmiPienzaInfierno SujoSufione PisciarellyFangaryPuzzoli
Set 1Set 2Set 1Set 2Set 1Set 2Set 1Set 2Set 1Set 2Set 1Set 2Set 1Set 2
Dictyosphaerium chlorelloides
Number of replicate cultures74287030702870307030   70307030
Number of cultures containing the following no. of resistant cells:
 0 0 033 059 062 07030   19027 0
 < 104 0 0 5 0 228 730 0 0   1703630
 104–2.5 × 104 0 021 0 1 0 1 0 0 0   150 7 0
 > 2.5 × 10474281130 8 0 0 0 0 0   1930 0 0
Variance : mean ratio (of the number of resistant cells per replicate)  0.9 1.260* 0.831* 1.060* 1.1> 100*0.972* 0.9
Adaptation processPhysiologicalGeneticGeneticGeneticNoneGeneticGenetic
Microcystis aeruginosa
Number of replicate cultures75307030702870307030   70307030
Number of cultures containing the following no. of resistant cells:
 0 0 054 0702870307030   6307030
 < 104 0 0 9 0 0 0 0 0 0 0    430 0 0
 104–2.5 × 104 0 0 730 0 0 0 0 0 0    30 0 0
 > 2.5 × 1047530 0 0 0 0 0 0 0 0    00 0 0
Variance/mean ratio (of the number of resistant cells per replicate) 1.2 1.191* 1.1    9*1.2
Adaptation processPhysiologicalGeneticNoneNoneNoneGeneticNone
Icebound geothermal waters from ArgentinaDoña SaraAguas CalientesLas PapasLos TachosLa Maquinita
Set 1Set 2Set 1Set 2Set 1Set 2Set 1Set 2Set 1Set2
  • The characteristics of the different geothermal waters are shown in Table 1.

  • *

    Variance/mean > 1; P < 0.001, using χ2 as a test of goodness of fit.

Dictyosphaerium chlorelloides
Number of replicate cultures7025   70257025   70257025
Number of cultures containing the following number of resistant cells:
 07025   23 038 0   48 07025
 < 104 0 0   27 023 0    725 0 0
 104–2.5 × 104 0 0   1425 425    7 0 0  0
 > 2.5 × 104 0 0    6 0 5 0    7 0 0 0
Variance : mean ratio (of the number of resistant cells per replicate) 1.2 1.1> 100* 0.837* 1.3> 100* 0.9
Adaptation processPhysiologicalGeneticGeneticGeneticNone
Microcystis aeruginosa
Number of replicate cultures9625   96259625   96259625
Number of cultures containing the following number of resistant cells:
 036 0   96259625   47  09625
 < 10441 0    0 0 0 0   39 0 0 0
 104–2.5 × 1041125    0 0 0 0    625 0 0
 > 2.5 × 104 8 0    0 0 0 0    4 0 0 0
Variance : mean (of the number of resistant cells per replicate)   29* 1.1
Adaptation processGeneticNoneNoneGeneticNone

Two different results can be found in the set 1 experiment when conducting a fluctuation analysis, each result being interpreted as the independent consequence of two different phenomena of adaptation. In the first case, if resistant cells arose during the exposure to the selective agent (i.e. by physiological adaptation), the variance in the number of cells per culture would be low because every cell is likely to have the same chance of developing resistance (Fig. 1, set 1A). Consequently, inter-culture (tube-to-tube) variation would be consistent with the Poisson model (i.e. variance : mean ratio approx. 1). By contrast, if resistant cells arose before the exposure to the selective agent (i.e. genetic adaptation by rare spontaneous mutation occurring during the time in which the cultures grew to Nt from N0 cells before the exposure to geothermal water), a high variation in the interculture number of resistant cells per culture would be found (Fig. 1, set 1B). Consequently, the tube-to-tube variation would not be consistent with the Poisson model (i.e. variance : mean ratio > 1). Obviously, another result – 0 resistant cells in each culture – could also be found, indicating that neither selection on spontaneous mutations that occur before geothermal water exposure, nor specific adaptation during the exposure to the geothermal water, took place.

The set 2 cultures are the experimental controls of the fluctuation analysis (Fig. 1). Variance is expected to be low, because set 2 samples the variance of the parental population. If the variance : mean ratio of set 1 is significantly greater than the variance : mean ratio of set 2 (fluctuation), this confirms that resistant cells arose by rare mutations that occurred before exposure to the geothermal water. If a similar variance : mean ratio between set 1 and set 2 is found, it confirms that resistant cells arose during the exposure to the geothermal water.

The fluctuation analysis also allows estimation of the rate of appearance of resistant cells. The proportion of cultures of set 1 showing nonresistant cells after geothermal water exposure (i.e. the first term of the Poisson distribution, named the P0 estimator; Luria & Delbrück, 1943) was the parameter used to calculate the mutation rate (µ) as:

µ = −LogeP0/(Nt − N0) (Eqn 3)

Mutation-selection equilibrium

If the mutation from a normal wild-type geothermal water-sensitive allele to a geothermal water-resistant allele is recurrent, and the resistant allele is detrimental to fitness in the absence of geothermal water, then new mutants arise in each generation, but most of these mutants are eliminated sooner or later by natural selection, if not by chance. Thus, at any given time there will be a certain number of resistant cells that are not yet eliminated. According to Kimura & Maruyama (1966), the average number of such mutants will be determined by the balance between µ and the rate of selective elimination (s):

q = µ/(µ + s), (Eqn 4)

where q is the frequency of the geothermal water-resistant allele and s is the coefficient of selection calculated as:

image(Eqn 5)

where inline image and inline image are the Malthusian fitness of geothermal water-resistant and geothermal water-sensitive cells measured in nonselective conditions (i.e. BG-11 medium), respectively.

Results

The seven geothermal waters from Italy were collected from different sites (Table 1) in an attempt to analyse the effects of different geothermal waters on phytoplankton. Consequently they showed substantial differences in pH (from 2.0 to 6.7), conductivity (from 2.4 to 8.3 mS) and temperature (from 21.0 to 43.8°C) (Table 1). Usually the phytoplankton flora in such geothermal waters is very poor, and indeed few species of eukaryotic microalgae, at low densities (< 500 total cells ml−1), were detected in most of the geothermal waters analysed. Geothermal waters from Pienzza and Infierno Sujo contained eukaryotic microalgae (mainly chlorophyceans), but cyanobacteria were not detected (Table 1). Neither microalgae nor cyanobacteria were found in Sufione Pisciarelli (Table 1). By contrast, dense and diverse communities of microalgae and cyanobacteria inhabited the warm pond of Bagno Vignoni (Table 1). Most of geothermal waters analysed were lethal for the wild-type strain of D. chlorelloides and M. aeruginosa in laboratory experiments. The extremely toxic effect of these geothermal waters produced total inhibition of growth and photosynthetic performance (ΦPSII) in cultures of both species (Table 1). Only water from Bagno Vignoni caused scant inhibition of growth and photosynthesis of laboratory cultures (Table 1).

The five different geothermal ponds from Los Andes remain icebound except for a few days in the summer, but do not freeze during the winter, maintaining temperatures ≥ 30°C. Eukaryotic and prokaryotic photosynthesizers are often represented in these ponds (Table 1). Doña Sara maintains a high diversity of cyanobacteria and microalgae, and dense algal populations also inhabit Aguas Calientes, Las Papas and Los Tachos. Only in La Maquinita were we unable to detect photosynthesizers (Table 1). Despite this, water from Aguas Calientes, Las Papas, Los Tachos and La Maquinita were lethal for our two strains of mesophilic photosynthesizers: m and ΦPSII of the wild-type strains of D. chlorelloides and M. aeruginosa were totally inhibited by these four geothermal waters (Table 1). Water from Doña Sara was also toxic, but not lethal (Table 1).

When conducting the fluctuation analysis, different results were obtained (Table 2). In most of the geothermal waters from Italy (Amiana Marmi, Pienza, Infierno Sujo, Fangary and Puzzoly for D. chlorelloides, and Amiana Marmi and Fangary for M. aeruginosa) and Los Andes (Aguas Calientes, Las Papas and Los Tachos for D. chlorelloides, and Doña Sara and Los Tachos for M. aeruginosa) the cell density was drastically reduced in each experimental culture of sets 1 and 2 in both species owing to destruction of sensitive cells. However, after further incubation for several weeks, some cultures increased in density again, apparently owing to growth of geothermal water-resistant variants. In the case of set 1, some cultures recovered after 75 d of geothermal water exposure (Table 1). By contrast, every set 2 culture recovered, and geothermal water-resistant cells were detected in all cultures (Table 2). In addition, low fluctuation was observed in set 2 (variance : mean ratio = 1, consistent with Poisson variability; P < 0.05, using χ2 as a test of goodness of fit), which indicated that the high fluctuation found in set 1 cultures should be caused by processes other than sampling error (Table 2). As in set 1 cultures, the variance significantly exceeded the mean (variance : mean ratio > 1; P < 0.001 using χ2 as a test of goodness of fit), so it could be inferred that geothermal water-resistant cells arose by rare, pre-selective spontaneous mutations rather than by specific physiological adaptation appearing in response to geothermal waters. By contrast, both strains used in the study proliferated in Bagno Vignoni as result of physiological adaptation, and D. chlorelloides also acclimatized to Doña Sara (variance : mean ratio = 1, consistent with Poisson variability; P < 0.05, using χ2 as a test of goodness of fit, in both set 1 and set 2 waters; Table 2).

When the species diversities observed in situ in the different geothermal waters (Table 1) were compared with the results of laboratory fluctuation analysis (Table 2), a correlation could be observed. Bagno Vignoni is the location with the highest diversity and abundance of cyanobacteria and microalgae; D. chlorelloides and M. aeruginosa easily proliferated in Bagno Vignoni by means of physiological adaptation. High microalgal diversity was also observed in Doña Sara and D. chlorelloides also proliferated without difficulty in Doña Sara by physiological adaptation. By contrast, algae and cyanobacteria were detected in neither Sufione Pisciarelli nor in La Maquinita, and our experiment found that neither D. chlorelloides nor M. aeruginosa were able to adapt to the water from these sites. A few species of microalgae (but not cyanobacteria), at low densities, were detected in Pienzza, Infierno Sujo, Puzzoly, Aguas Calientes and Las Papas; laboratory fluctuation analysis showed that rare spontaneous pre-selective mutations allowed adaptation of D. chlorelloides to these water samples, but M. aeruginosa was unable to proliferate. Cyanobacteria were detected in Amiana Marmi, Fangary, Doña Sara and Los Tachos; fluctuation analysis showed that spontaneous resistance-mutants of M. aeruginosa proliferated in these water samples.

The estimated mutation rates (µ) from sensitivity to resistance to the different geothermal waters ranged from 1.4 × 10−6 to 1.5 × 10−5 mutants per cell per division in D. chlorelloides, and from 1.1 × 10−6 to 1.1 × 10−5 in M. aeruginosa (Table 3). Isolated D. chlorelloides and M. aeruginosa geothermal water-resistant mutants growing in the absence of the selective agent (i.e. in BG-11 medium) showed fitness values lower than those found in the wild-type strains (data not shown). The relative values of fitness of resistant mutants and sensitive wild types were used to compute the coefficient of selection (s) of geothermal-resistant mutants and by using the values of µ and s, the frequency (q) of resistant alleles as the consequence of the balance between mutation and selection was calculated (Table 3). A frequency of 1.5–20 resistant mutants per 106 cells in D. chlorelloides, and 1.4–12 resistant mutants per 106 cells in M. aeruginosa, could be maintained in the absence of the selective agent as the consequence of the balance between recurrent mutation and selection (Table 3).

Table 3.  Mutation rate (µ, mutants per cell per generation), coefficient of selection against resistant mutant (s) and frequency of the geothermal water-resistance allele (q) in Dictyosphaerium chlorelloides (Chlorophyceae) and Microcystis aeruginosa (Cyanobacteria), during genetic adaptation to different geothermal waters from Italy and Argentina (see Table 2 for the fluctuation analysis)
 Dictyosphaerium chlorelloidesMicrocystis aeruginosa
µsqµsq
Italy
Amiana Marmi8.5 × 10−60.731.2 × 10−52.7 × 10−60.922.9 × 10−6
Pienzza2.0 × 10−60.852.4 × 10−6
Infierno Sujo1.4 × 10−60.911.5 × 10−6
Fangary1.5 × 10−50.732.0 × 10−51.1 × 10−60.781.4 × 10−6
Puzzoly1.1 × 10−50.911.2 × 10−5
Argentina
Aguas Calientes1.3 × 10−50.891.5 × 10−5
Las Papas6.9 × 10−60.917.6 × 10−6
Los Tachos4.2 × 10−60.854.9 × 10−67.9 × 10−60.869.2 × 10−6
Doña Sara1.1 × 10−50.931.2 × 10−5

Discussion

There is growing scientific interest in how inhabitants of geothermal waters adapt to living in some of the most extreme conditions on Earth (Sand, 2003). Geothermal areas are often hazardous for mesophilic photosynthesizers because of the presence of dissolved mineral components such as arsenic and mercury, hydrogen sulphide, and highly acidic or very alkaline conditions (Webster & Nordstrom, 2003; Tyrovola et al., 2006; Lehr et al., 2007). Indeed, we unable to detect cyanobacteria or microalgae in the geothermal waters of Sufione Pisciarelly (Italy) and La Maquinita (Los Andes), and most of geothermal waters analysed were lethal for nearly all cells of our strains of D. chlorelloides and M. aeruginosa. For example, six geothermal waters (from a total of seven) analysed from Italy and four icebound geothermal waters (from a total of five) analysed from Los Andes, totally inhibited growth and photosynthetic quantum yield of D. chlorelloides and M. aeruginosa. Consequently, the survival of mesophilic cyanobacteria and microalgae in most geothermal waters could only be achieved by some kind of genetic adaptation.

Apparently, adaptation to geothermal waters is not easy. The eukaryotic microalga D. chlorelloides was unable to adapt to waters from Sufione Pisciarelly and La Maquinita, whereas the cyanobacterium M. aeruginosa did not adapt to water samples from Pienza, Infierno Sujo, Sufione Pisciarelly, Puzzoli, Aguas Calientes, Las Papas and La Maquinita (7 geothermal waters from a total of 12). Since adaptation to extreme environments (such as geothermal waters) seems to be difficult, the classic point of view assumes that genetic adaptation at such extreme conditions is a gradual process. By contrast, here we propose an alternative explanation for adaptation of cyanobacteria and microalgae to geothermal waters. When D. chlorelloides and M. aeruginosa were cultured in different geothermal waters, usually cultures show that all the sensitive cells are destroyed by the toxic effect of such waters. However, after further incubation for 75 d, some cultures recovered, owing to the growth of cells that were resistant to the toxic effect of geothermal waters. The key to understanding adaptation of mesophilic cyanobacteria and microalgae to the extremely adverse conditions of the geothermal waters is to analyse the rare variants that proliferate after the massive destruction of the sensitive cells by this selective agent. Fluctuation analysis is an appropriate procedure to discriminate between geothermal water-resistant cells arising by rare spontaneous mutations occurring randomly during replication of organisms before exposure to this selective agent and geothermal water-resistance arising through specifically acquired adaptation induced by geothermal waters (Sniegowski & Lenski, 1995; revised by Sniegowski, 2005).

Genetic adaptation by rare spontaneous mutation seems to be the usual mechanism allowing adaptation of microalgae and cyanobacteria to geothermal waters. The large fluctuation in number of resistant cells detected in the set 1 experiments compared with the insignificant fluctuation in set 2 controls (observed in 8 geothermal water samples from a total of 12 with D. chlorelloides as well as in 4 geothermal water samples from a total of 12 in M. aeruginosa), unequivocally demonstrates that these resistant cells arose by rare spontaneous single mutations (which occur before geothermal waters exposure) and not through direct and specific adaptation in response to geothermal waters. Consequently, mesophiles can adapt to such geothermal waters much more rapidly by single mutations than if the ability to survive required multiple mutations. Recent evidence suggests that mutation at one locus can enable adaptation of mesophile cyanobacteria and microalgae to other hostile natural environments (Flores-Moya et al., 2005; Costas et al., 2007; López-Rodas et al., 2008a,b) as well as to sudden anthropogenic chemical contamination (Costas et al., 2001; López-Rodas et al., 2001; García-Villada et al., 2002, 2004; López-Rodas et al., 2007). This evidence suggests that the different species of mesophilic algae and cyanobacteria follow the same pattern. Consequently, it could be hypothesized that mesophilic populations would be likely to survive extreme environmental changes.

The mutation rates from geothermal water-sensitivity to geothermal water-resistance in D. chlorelloides ranged from 1.4 × 10−6 to 1.5 × 10−5 mutants per cell per division. These values were similar to those found in this species for the resistance to the potent biocide 2,4,6-trinitrotoluene (García-Villada et al., 2002), but higher than those we have described for resistance to the stressful (pH < 2 and high levels of metals) natural environments from Spain's Rio Tinto (Costas et al., 2007) or, in the case of the chlorophycean Spirogyra, to the sulphurous water from La Hedionda Spa (southern Spain) (Flores-Moya et al., 2005). In this sense, mutation rates from geothermal water-sensitivity to geothermal water-resistance in the cyanobacterium M. aeruginosa (from 1.1 × 10−6 to 1.1 × 10−5) were higher than those for the algicide copper sulphate (García-Villada et al., 2004) or the herbicide glyphosate (López-Rodas et al., 2007). It may be that the mutation for geothermal water resistance is more common than other kinds of mutation.

In contrast, physiological acclimatization only occurred in two geothermal water samples from a total of twelve in D. chlorelloides and only one geothermal water sample from a total of twelve in M. aeruginosa (as demonstrated by the insignificant fluctuation in number geothermal water-resistant cells observed in the set 1 experiment, and similar to the insignificant fluctuation in set 2 controls). In addition, adaptation was not achieved by either D. chlorelloides in Suffione Pisciarelly and La Maquinita, or by M. aeruginosa in seven geothermal water samples from a total of twelve samples analysed.

In a few cases, geothermal waters are lethal for photosynthetic microbes. Conversely, cyanobacteria and microalgae seem to be able to adapt to stressful geothermal waters as a result of rare spontaneous mutations and, if the toxicity is not too high, by physiological acclimatization. Taking into account both the relatively high number of resistant mutants and the countless cells comprising phytoplankton populations, apparently enough geothermal water-resistant mutant cells are present in natural populations of these organisms (before geothermal water exposure) as consequence of the balance between recurrent spontaneous mutations and selection.

This capability of mesophilic algae to adapt to geothermal waters by means of mutation at one or a small number of loci could have significant implications for the survival of algae during the Neoproterozoic ‘snowball Earth.’ Since green, red and chromophyte eukaryotic algae evolved before the late Neoproterozoic glaciations (Knoll & Bauld, 1989; Knoll, 1992) they must have survived these extreme environmental conditions. Although primary production collapsed during the ‘snowball Earth’ period (Kaufman et al., 1997; Hoffman et al., 1998; Schrag & Hoffman, 2001; Rothman et al., 2003), just afterwards massive blooms of eukaryotic algae took place (Elie et al., 2007). In addition, microfossils suggest that global Neoproterozoic glaciations did not make a large contribution to the extinction of autotrophic eukaryotes (Corsetti et al., 2006). The survival of photosynthetic life during the Neoproterozoic ‘snowball Earth’ may have been achieved in ice-free refuges of geothermal waters (Schrag & Hoffman, 2001). Usually, icebound geothermal waters show a considerable gradient of temperature (very hot within the spring, and progressively colder, ultimately freezing). Consequently, temperature is not a selective factor in these areas and only selection for resistance to extreme values of pH, toxic and other chemical stressors will occur. In addition there is no evidence that the ‘snowball Earth’ ancestors of modern mesophilic algae were thermophiles. A recent paper supports the concept of continuous, numerous establishment events by mesophilic microorganisms in volcanic areas, in contrast to very rare arrivals of thermophilic strains (Portillo & González, 2008). It could be hypothesized that among these continuously arriving mesophilic microalgae, resistant mutants easily colonized numerous geothermal waters, which could be ‘Noah's Arks’ during the global glaciations of Neoproterozoic Earth. Arriving mesophilic algae were able to colonize extremely acid and toxic environments of Rio Tinto and La Hedionda (Flores-Moya et al., 2005, Costas et al., 2007). Another alternative could be the survival of microbial communities living within and upon ice (Corsetti et al., 2006), if only a thin (< 2 m) ice cover occurred in tropical latitudes (Pollard & Kasting, 2005). Perhaps eukaryotic algae could proliferate in both scenarios, with geothermal waters as the main refuge of freshwater algae and the sea under a thin ice cover as refuge for marine algae. However, after the Neoproterozoic ‘snowball Earth’ period ended (with very high atmospheric CO2) the water would be acid, warm and very rich in nutrients (Nisbet et al., 2007). Consequently, the algae and cyanobacteria that survived in acid geothermal waters would have been pre-adapted to colonize these new environments.

Acknowledgements

This work was financially supported by CGL 2005-01938 BOS, S-0505/AMB/0374 CAM and P05-RNM-00935 grants. Dr Eric C. Henry (Herbarium, Department of Botany and Plant Pathology, Oregon State University) kindly revised the English style and usage. Dr Fernando Hiraldo (Estación Biológica de Doñana, Consejo Superior de Investigaciones Científicas) and Dr Antonio Delgado (Estación Experimental de Zaidín, Consejo Superior de Investigaciones Científicas) suggested the initial ideas. Manuel de la Riva, Sebastián Dimartino and Obdulio Monsalvo helped us during sampling in the field. Eva Salgado, Fernando Marvá and Mónica Rouco contributed to the laboratory work.

Ancillary