Contrasted resistance of stone-dwelling Geodermatophilaceae species to stresses known to give rise to reactive oxygen species

Authors


Correspondence: Maher Gtari, Laboratoire Microorganismes et Biomolécules Actives, Université de Tunis Elmanar (FST) et Université de Carthage (ISSTE), 2092 Tunis, Tunisia. Tel.: +216 70 860 553; fax: +216 70 860 553; e-mail: maher.gtari@fst.rnu.tn

Abstract

Stones in arid environments are inhabited by actinobacteria of the family Geodermatophilaceae like the genera Blastococcus and Modestobacter frequently isolated from altered calcarenites. Their habitat requires adaptation to light-induced and other stresses that generate reactive oxygen species. Here, we show that representative members of the species Blastococcus saxobsidens,Geodermatophilus obscurus, and Modestobacter multiseptatus are differentially adapted to stresses associated with arid environments. Whereas B. saxobsidens was found to be sensitive to gamma radiation (D10 = 900 Gy; 10% survival at 900 Gy), M. multiseptatus was moderately (D10 = 6000 Gy) and G. obscurus was highly tolerant (D10 = 9000 Gy). A difference in resistance to high-frequency (λ value = 254 nm) UV was shown by B. saxobsidens,M. multiseptatus, and G. obscurus, being sensitive, tolerant, and highly tolerant (D10 of 6, 900, and > 3500 kJ m−2, respectively). Tolerance to desiccation, mitomycin C and hydrogen peroxide correlated with the ionizing radiation and UV resistance profiles of the three species and were correlated with the pigments synthesized. Resistance to heavy metals/metalloids did not follow the same pattern, with resistance to Ag2+ and Pb2+ being similar for B. saxobsidens,M. multiseptatus, and G. obscurus, whereas resistance to inline image, Cr2+, or Cu2+ was greater for B. saxobsidens than for the other two species. The stress resistance profiles of M. multiseptatus and B. saxobsidens were reflected in different calcarenite colonization patterns. While M. multiseptatus was predominantly isolated from the first two millimeters of stone surface, B. saxobsidens was predominantly isolated from the deeper part of the stone where it is better protected from sun irradiation, suggesting that the response to light- and desiccation-induced oxidative stress is an important driver for niche colonization in the stone biotope.

Introduction

Among the most common bacteria associated with stone biodeterioration, Geodermatophilaceae represent a distinctive family often found on calcareous and carbonatic stones such as marble, limestone, and calcarenite (Eppard et al., 1996; Urzì & Realini, 1998; Urzì et al., 2001, 2004; Brusetti et al., 2008). The actinobacterial family Geodermatophilaceae (Normand, 2006) includes the genera Geodermatophilus (Luedemann, 1968), Blastococcus (Ahrens & Moll, 1970), and Modestobacter (Mevs et al., 2000), having a Gram-positive cell wall, chemoorganotrophic, mesophilic cells with a strongly pigmented rudimentary mycelium. Some of the biodeterioration capability of the Geodermatophilaceae is also because of their ability to develop as endoliths, growing into stones to a depth of a few millimeters, as observed in a limestone tunnel at a Mayan archeological site in Mexico (Ortega-Morales et al., 2005).

To enable stone colonization, bacterial cells should be able to resist a number of environmental stresses such as solar radiation, desiccation, salt, and metal resistance, many of which result in the production of reactive oxygen species (ROS) generated by the radiolysis of water molecules or the Fenton reaction (Imlay et al., 1988; Schiavano et al., 1990; Potts, 1994; Stohs & Bagchi, 1995; Cabiscol et al., 2000; Daly et al., 2007; Ashraf, 2009; Daly, 2009; Pan et al., 2009). Indeed, ionizing radiation resistance is correlated with cytosolic protection from ionizing-radiation-induced ROS (Daly et al., 2007; Daly, 2009) and consequently to other ROS-related traits like desiccation tolerance (Dose et al., 1991, 1992; Mattimore & Battista, 1996; Battista et al., 1999; Billi et al., 2000; Tanaka et al., 2004; Rainey et al., 2005; Fredrickson et al., 2008). Ionizing-radiation-resistant bacteria (IRRB) isolated from radioactive environments are generally much more tolerant to exposure to oxidative stress (and ionizing radiation) than their counterparts from environments subjected only to background levels of radiation (Bagwell et al., 2008).

Exposure to heavy metals, other potential stressing factors in the stone biotope, can result in ROS generation and the related resistance, but may be counteracted by other mechanisms such as transport outside the cell (Nies, 1995), adsorption on exocellular structures such as melanin (Fogarty & Tobin, 1996), or enzymatic reduction to less toxic forms (Spain et al., 2007; Poopal & Laxman, 2009).

The capacity to resist the above-mentioned stresses should be an important feature of microorganisms such as the Geodermatophilaceae for surviving under the harsh conditions associated with stones, particularly in arid ecosystems, for shaping their ecology in the stone biotope and for driving the colonization of different niches in the stone at different depths in the stone (Gorbushina, 2007).

Indeed, it was reported that there was a predominance of the IRRB of the genus Deinococcus along with Geodermatophilus recovered from an arid soil after exposure to the highest doses of ionizing radiation tested, ranging from 17 to 30 kGy (Rainey et al., 2005). Together with Deinococcus and Bacillus, Geodermatophilaceae have also been detected from atmospheric samples from the Sahara and snow deposits in Europe (Chuvochina et al., 2011), illustrating their ability to be transported over long distances, and resist radiation and desiccation stresses during travel in the high atmosphere. Geodermatophilaceae have been predominantly recovered from extreme environments characterized by dry conditions such as those that exist in Antarctic or hot desert soils (Mevs et al., 2000; Essoussi et al., 2010), rocks, and monument surfaces (Eppard et al., 1996; Urzì & Realini, 1998; Urzì et al., 2001, 2004; Brusetti et al., 2008). Members of the Geodermatophilaceae have been found to have enzymes adapted to relatively high temperature such as their esterases (Essoussi et al., 2010; Jaouani et al., 2012).

The aim of the present study was to investigate the resistance profiles of the three representative genera of the Geodermatophilaceae, Blastococcus, Geodermatophilus, and Modestobacter, to different environmental stresses including ROS-generating gamma and short-wavelength UV irradiation, desiccation, exposure to mitomycin C, and hydrogen peroxide but also salt, heavy metals, and metalloids. We have found distinct resistance in the three genera that are reflected in different colonization patterns of calcarenite by Blastococcus and Modestobacter, suggesting that response to light-induced oxidative stress is an important driver of niche colonization in the stone biotope.

Materials and methods

Resistance to ionizing radiations

Geodermatophilus obscurus ssp. obscurus (strain ATCC 25078 = DSM 43160 = IFO (now NBRC) 13315 = JCM 3152 = NRRL B-3577 = VKM Ac-658) (Luedemann, 1968) was kindly provided by Dr. David Labeda (NRRL bacterial collection, USA), and Blastococcus saxobsidens (strain DD2) and Modestobacter multiseptatus (strain BC501) were isolated from calcarenite walls in Sardinia and from Carrara marble cave, respectively (Urzì et al., 2001). Strains were routinely grown in the complex Luedemann medium (Luedemann, 1969), comprising yeast extract, amine extract, glucose, soluble starch, and CaCO3. Nonsporulating cultures were obtained as described by Ishiguro & Wolfe (1970) by adding tryptose (Difco, Detroit, USA). Nonsporulating exponentially growing cultures (3 days old) of the three strains were washed twice, homogenized by forcing repeatedly through 0.7 × 30 mm sterile syringe with 21G needles, and resuspended in 1 mL of 0.9% NaCl, for subsequent experiments.

One milliliter of cell suspension of each strain was introduced into separate sterile 1.5-mL Eppendorf tubes and subsequently exposed to 0.5–12 kGy of gamma irradiation, at a dose rate of 126.12 Gy min−1, in a 60Co irradiator. The experiments were independently performed in two different laboratories CNSTN, Sidi Thabet, Ariana, Tunisia, and CEA Cadarache, France. Afterward, the cell suspensions were cooled on crushed ice, serially diluted in 0.9% NaCl, plated on solid Luedemann medium, and incubated at 30 °C. After a 2-week period, CFUs were counted and the percent survival was calculated based on the CFUs of a nonirradiated control.

Cultures of each strain were washed once and resuspended in an equal volume of 0.9% NaCl, and 1-mL aliquots were plated onto Luedemann medium plates and subsequently exposed in triplicate to a 30 W, 254 nm UV bactericidal source in a laminar flow hood for various time periods. The plates were then incubated at 30 °C and the CFUs counted after a 2-week incubation period. The percent survival was calculated based on the CFUs of a nonirradiated control.

Tolerance of Geodermatophilaceae

Strains DSM 43160, DD2, and BC501 were surveyed for tolerance to desiccation, mitomycin C, hydrogen peroxide, sodium chloride, and heavy metals/metalloids.

To assess the desiccation tolerance, cultures of each strain were washed once and resuspended in an equal volume of 0.9% NaCl, and 1-mL aliquots were placed onto glass cover-slips (1 inch × 1 inch). The cover-slips were then placed in sterile Petri dishes inside a vacuum desiccator containing silica gel. The percentage survival for each strain was determined at 20, 40, 60, and 80 days after desiccation. At each time-point, one cover-slip containing each strain was removed, and one milliliter 0.9% NaCl was added to the cover-slips to rehydrate the cells. Cells were then 10-fold serially diluted, plated onto Luedemann medium, and incubated at 30 °C to determine the percent survival.

To assess the tolerance to mitomycin C, the induced DNA damage antibiotic, cultures' suspension was incubated at 30 °C with mitomycin C at a final concentration of 2 mg mL−1 for 5, 10, 20, 40, 60, 80, and 100 min, centrifuged, and washed in 0.9% NaCl. Tenfold dilutions were then plated onto Luedemann medium, and CFUs were enumerated after a 2-week incubation period at 30 °C.

To assess the tolerance to the hydrogen peroxide, cell suspensions were mixed to an equal volume of 3, 5, 10, 20, and 30% (v/v) H2O2, incubated in the dark for 10 min, and then spread onto fresh Luedemann agar medium at 30 °C. Recovery and growth were evaluated based on the CFUs number of a nontreated control after a 2-week incubation period at 30 °C.

To assess the tolerance to sodium chloride, cell suspensions were grown at 30 °C in Luedemann medium containing various concentrations of NaCl. Growth of colonies from spreading cells was evaluated based on the CFUs number of a nontreated control after a 2-week incubation period at 30 °C.

The sensitivity of the strains to heavy metals was determined by a growth inhibition plate assay as described by Richards et al. (2002). Minimum inhibitory concentration (MIC) of tested metals was determined on CB medium (Ishiguro & Wolfe, 1970) consisting of 0.2% casamino acids (Difco), 0.2% glucose, 1.0% (v/v), and 1.0% (v/v) trace mineral solution at pH 7.0. Growth was monitored after 1, 2, 3, and 4 weeks at 30 °C.

Correspondence analysis of the physiological data

Physiological data (Table 1) were coded as 0 when absent or 1 when present, or as in the case of stressors as the dose resulting in 10% survival (LD10), and in the case of metals/metalloids as the MIC. This yielded a matrix that was then subjected to a correspondence analysis according to Benzécri (1973) as implemented in the r statistics program (Team, 2007) using the FactoMineR package (Lê et al., 2008).

Table 1. Ecological features and multistress profiles of Geodermatophilus obscurus DSM 43160, Modestobacter multiseptatus BC501, and Blastococcus saxobsidens DD2
HabitatGeodermatophilus obscurus DSM 43160Modestobacter multiseptatus BC501Blastococcus saxobsidens DD2References
Deteriorated rocks, sandRock surfaces, regolithInternal parts of the rocksThis study
  1. a

    MICs were expressed in mM of heavy metal/metalloid.

Morphology
HyphaeRudimentaryMevs et al. (2000), Urzì et al. (2004), Normand (2006)
SporesMotileMotileMotileMevs et al. (2000), Urzì et al. (2004), Normand (2006)
Sporangia+Mevs et al. (2000), Urzì et al. (2004), Normand (2006)
Buds++Normand (2006)
Colony color in Luedemann mediumBlackPink initially, black laterOrangeNormand (2006)
PhysiologyNormand (2006)
Requirement for O2AerobicAerobicAerobic/microaerophilicNormand (2006)
Temperature range (°C)18–370–283–40Normand (2006)
Temperature optimum (°C)24–2819–2125Normand (2006)
Tolerance to 3% (w/v) NaCl++Normand (2006), This study
Resistance to
γ-rays (D10, Gy)90006000900This study
UV254 (D10, kJm−2)35009006This study
Desiccation (60days)~ 20% viability~ 10% viability0% viabilityThis study
Mitomycin C (40min)100% viability~ 100% viability~ 0% viabilityThis study
H2O2 (lethal dose, %)30305This study
Heavy metals/metalloids (MICs)aThis study
Ag2+0.10.10.1
inline image501085
Co2+0.30.50.5
Cr3+0.5220
Cu2+113
Ni2+0.511
Pb2+303030

Stone sampling

Twelve stone samples from an ancient wall in Cagliari (Sardinia, Italy) named C1 in Table 1 in the study by Urzì et al. (2001) were collected as described previously (Urzì et al., 2001). Of these samples, three (CI1, CI2, and CI3) were taken at a depth of about 2 cm, within the same microsite area sampled by Urzì et al. (2001). Three other samples (CO1, CO2, and CO3) were taken from the region between the stone surface and 2 mm below it at a distance of 2 cm around the microsite area C1. Three samples (As, Bs, and Ds) were taken at 1 cm of distance from each microsite. Finally, three other samples (Ad, Bd, and Dd) were collected approximately 2 cm below samples As, Bs, and Ds. Each sample consisted of about 1 g of aseptically collected calcarenite powder from scraped surfaces was taken under sterile conditions for each sample.

The sampling site is subjected to a typical Mediterranean climate. It is a wall oriented in a 110°E/SE direction, at an elevation of 30 m above sea level, close to the sea front (estimated distance from the sea of about 600 m), and it is exposed permanently to the dominant northward sea winds. The stone material is a calcarenite, called locally ‘Pietra Cantone’, characterized by high hygroscopicity and high susceptibility to deterioration.

Bacterial counts, isolation, and identification of Geodermatophilaceae

Each stone sample was suspended (ratio 1 : 10 w/v) in physiological saline solution (0.85% NaCl). The suspension was stirred for 60 min using a vortex mixer. One milliliter of each suspension and its 10-fold dilutions were spread onto Luedemann medium (Luedemann, 1969) supplemented with cycloheximide (100 μg mL−1) to avoid fungal growth. Incubation was carried out at 30 °C. Enumeration and observation of the CFUs were performed after 7 and 15 days. Bacterial colonies were described morphologically. Those resembling the morphology of members of Geodermatophilaceae were randomly chosen and restreaked four times on Luedemann medium. A collection of 173 Geodermatophilaceae-like strains was thus obtained.

To extract DNA, bacterial isolates were grown at 30 °C for 2 days on Luedemann medium. Well-isolated colonies were washed three times in sterile physiological saline. Total genomic DNA was extracted using a CTAB–SDS lysis protocol (Ausubel et al., 1994). Genomic DNA was checked and carefully quantified both in an agarose gel with a λ-DNA marker (Sigma, Milan, Italy) and with SmartSpec 3000 spectrophotometer (Bio-Rad, Milan, Italy). The almost entire 16S rRNA gene was amplified with primers 16S-27F and 16S-1494R (Lane, 1991). The reaction mixture contained 1× PCR buffer (Invitrogen, Milan, Italy), 1.5 mM MgCl2, 0.12 mM dNTPs, 0.3 μM of each primer, 1.2 U of Taq DNA polymerase (Invitrogen), and 50 ng of genomic DNA in a final volume of 50 μL. DNA was denatured at 94 °C for 4 min and then subjected to 5 initial cycles of 94 °C for 1 min, 50 °C for 45 s, and 72 °C for 1 min, and then 30 cycles of 92 °C for 45 s, 55 °C for 45 s, and 72 °C for 1 min. A final extension at 72 °C for 10 min was included. ARDRA profiles were determined as described previously (Essoussi et al., 2012) using the following restriction enzymes: RsaI, AluI, HaeIII, HhaI, and HpaII (Amersham Pharmacia Biotech, Milan, Italy), and the resulting haplotypes were compared with the ARDRA fingerprinting patterns obtained previously (Urzì et al., 2001). One 16S rRNA gene for each ARDRA haplotype was sequenced between primers 16S-27F and 16S-1494R as described previously (Brusetti et al., 2008). Sequences were checked using the check_chimera program to determine the presence of any hybrid sequences (Maidak et al., 2001). Phylogenetic affiliations were preliminarily obtained using blastn (Altschul et al., 1997) and confirmed with the Naïve Bayesian rRNA Classifier provided by the Ribosomal Database Project website (Maidak et al., 2001). Nucleotide sequences were multi-aligned with ClustalX v.1.83 (Thompson et al., 1997). Alignments were checked manually, and poorly aligned or divergent regions were eliminated using GBlocks v.0.91b (Castresana, 2000) with a minimum block length of five and allowed gap positions equal to half. Phylogenetic trees were constructed with Treecon 1.3b (Van de Peer, University of Konstanz, Germany), using the neighbor-joining method (Saitou & Nei, 1987) and a Jukes & Cantor (1969) distance matrix. A bootstrap analysis (Felsenstein, 1985) with 1000 replicates was performed to estimate the reproducibility of trees. The partial nucleotide sequences of the 16S rRNA genes were deposited in the EMBL nucleotide sequence database (GenBank/EMBL/DDBJ) under the accession numbers FR865876 to FR865898.

Results

Resistance of Geodermatophilaceae to radiations and oxidative stresses

Cultures of G. obscurus, B. saxobsidens, and M. multiseptatus were assayed for sensitivity to gamma irradiation. Geodermatophilus obscurus and M. multiseptatus demonstrated no loss of viability at doses below 3000 Gy, whereas B. saxobsidens exhibited significantly lower levels of survival over the entire dose range tested. In Fig. 1a, the ionizing radiation resistance of B. saxobsidens is compared with those of G. obscurus and M. multiseptatus. The shoulder in the resistance curve, characteristic of G. obscurus and M. multiseptatus, is completely absent in the resistance curve of B. Saxobsidens. At doses above 1000 Gy, B. saxobsidens displayed progressive loss of survival.

Figure 1.

Survival curves for Geodermatophilus obscurus DSM 43160 (filled squares), Blastococcus saxobsidens DD2 (open squares), and Modestobacter multiseptatus BC501 (filled circles) following exposure to gamma irradiation (a), UV irradiation (b), desiccation (c), mitomycin C (d), and hydrogen peroxide (e). The horizontal black lines in the graphs indicate the 10% survival threshold.

A survival of 10% of the original B. saxobsidens population was reached after < 30 s of UV irradiation at 6 kJ m−2. No CFUs were detected after 4 min (Fig. 1b). The two other strains were more resistant, because 10% survival was reached after slightly more than 60 min in the case of M. multiseptatus and a little bit more than 120 min in the case of G. obscurus, which corresponded to 900 and 3500 kJ m−2, respectively.

The survival curves following exposure to desiccation for G. obscurus, B. saxobsidens, and M. multiseptatus are given in Fig. 1c. After 60 days in a desiccator, G. obscurus and M. multiseptatus remained viable, exhibiting approximately 20% and 10% survival, respectively; whereas B. saxobsidens was not viable (Fig. 1c).

Cultures of G. obscurus and M. multiseptatus could grow without any significant loss of viability after exposure for 40 min to 2 mg mL−1 of mitomycin C (Fig. 1d). For the same final mitomycin C concentration, B. saxobsidens cultures displayed almost no survival (Fig. 1d). These observations, suggesting resistance to oxidative stresses, were supported by the resistance in the presence of H2O2 (Fig. 1e). Whereas nearly 1% of H2O2-treated G. obscurus and M. multiseptatus cultures survived to 10% of H2O2, B. saxobsidens did not survive exposure to 5% H2O2 (Fig. 1e).

Resistance of Geodermatophilaceae to salts and heavy metals

The three species of the Geodermatophilaceae showed different resistance to salinity (Table 1). While M. multiseptatus and B. saxobsidens could grow in the presence of 3% w/v NaCl, G. obscurus cultures did not.

In general, B. saxobsidens showed the highest resistance to heavy metals and metalloids being far more resistant than M. multiseptatus, and G. obscurus to arsenate, chromium, and copper (Table 1). Except for arsenate, M. multiseptatus was found to be slightly more resistant to heavy metals than G. obscurus.

A correspondence analysis (Fig. 2) permitted grouping of data of resistance to metals/metalloids on the right of the plot associated with the B. saxobsidens, while data of resistance to oxidative stresses were associated with the other two taxa, M. multiseptatus and G. obscurus. The first axis on which B. saxobsidens was distinct from the other two strains represented 70% of the total variance, while the second axis accounted for 30%.

Figure 2.

Correspondence analysis of physiological properties of Geodermatophilus obscurus DSM 43160 (Go), Blastococcus saxobsidens DD2 (Bs), and Modestobacter multiseptatus BC501 (Mm). Resistance was coded for each stressor as the LD10, in the case of metals/metalloids as the MIC, and in the case of physiological characters as ‘0’ or ‘1’.

Distribution of Modestobacter and Blastococcus in calcarenite (‘Pietra Cantone’) depth

To verify whether the differences in the resistance patterns observed between B. saxobsidens and M. multiseptatus could be associated with the distribution in the calcarenite depth, we analyzed the distribution of the two genera near the surface (first 2 mm of the material) and at a depth of about 2 cm of an altered calcarenite from bricks of the ancient wall of the city of Cagliari, Sardinia, Italy (Fig. 3). We examined this kind of material because previous studies showed that the calcarenite from different sites in the wall was colonized by B. saxobsidens and M. multiseptatus, while G. obscurus was very rare or not present at all (Urzì et al., 2001; Brusetti et al., 2008).

Figure 3.

Distribution of Modestobacter and Blastococcus in the calcarenite ‘Pietra Cantone’ of an ancient wall in Cagliari. (a) General image of the sampling sites of the wall. (b) Magnification of the sampling site C indicated by the arrow in the (a) as an example of the sampled material. Samples were taken both from the surface (Cag, CO1, CO2, and CO3) and at the depth of 2 cm (CI1, CI2, and CI3) taken after the removal of Cag sample, visible as a lighter triangle. (c) Histogram of the relative percentage of strains of Modestobacter (black) and Blastococcus sp. (white) isolated from each sampling site. At the top of each column is reported the number of the strains isolated, the first number refers to Modestobacter and second number to Blastococcus.

Twelve stone samples were collected, six from the surface of the stone and six from about 2 cm below surface after removing the surface layer with a sterile spatula (Fig. 3a). The positions of the sampling points are shown in Fig. 3b. Bacterial counts on Luedemann medium showed levels of culturable bacteria between 4.04 and 6.86 log CFU g−1 d.w. after 7 days of incubation, while after 14 days, the levels increased to 4.28 and 7.11 log CFU g−1 d.w. Geodermatophilaceae-like colonies ranged between 3.30 and 6.86 log CFU g−1 d.w. after 7 days and between 4.53 and 7.04 log CFU g−1 d.w. after 14 days, except for samples Bs, Bd, CI3, and CO1 from which no Geodermatophilaceae-like colonies grew.

Collectively, a total of 173 strains were randomly isolated from all samples and initially identified through strains grouping by ARDRA analyses, followed by sequencing the almost complete 16S rRNA gene of at least one strain per each of the 17 resulting ARDRA haplotypes. None of the haplotypes could be affiliated to G. obscurus, but all could be affiliated to Modestobacter and Blastococcus genera. In total, 14 haplotypes (120 strains) were affiliated to M. multiseptatus and three (53 strains) to B. saxobsidens (Fig. 4). The distribution of the two genera within the samples that allowed the isolation of the Geodermatophilaceae showed that most of the Modestobacter isolates were recovered from samples collected from the surface of the calcarenite, while Blastococcus isolates were mainly recovered from the 2-cm-deep sample material (Fig. 3c).

Figure 4.

16S rRNA gene–based phylogenetic affiliation of the Geodermatophilaceae isolated from the calcarenite microsites shown in Fig. 4. In the tree are presented the phylogenetic positions of the strains for which almost the entire 16S rRNA gene sequence has been obtained, along with reference strains of the Geodermatophilaceae.

Discussion

Ultraviolet radiation, together with desiccation, heat-cooling cycles, and fluctuating concentration of elements together constitute a fierce challenge for microbes living exposed on inhospitable substrates such as the desert topsoil whence G. obscurus was initially isolated or on stone surfaces whence the other two bacterial taxa, M. multiseptatus and B. saxobsidens, are routinely isolated (Urzì et al., 2001). Conversely, gamma irradiation is not expected to be prevalent in these desert soils and sun-exposed stone surfaces. Gamma irradiation adaptation would thus constitute a spandrel or an exaptation (Gould, 1997), which is why we sought to characterize the physiological response to several stresses common in dry biotopes and see whether they were all correlated. This would indicate a common evolutionary strategy. Moreover, the type of adaptation to these stresses should also be reflected in the colonization pattern of the stone material and would constitute a driver for the ability to thrive on the stone material.

Studies on ROS-generating stresses, including gamma irradiation, in Eubacteria have been focused mainly on species of the genera Deinococcus, particularly D. radiodurans (for review, see Dose et al., 1991, 1992; Cox & Battista, 2005; Daly et al., 2007; Makarova et al., 2007; Daly, 2009; de Groot et al., 2009 and references therein). Besides its extreme resistance to ionizing radiation and other stresses (Cox & Battista, 2005), D. radiodurans belongs to the distinct bacterial clade DeinococcusThermus that has remarkably contrasted phenotypes relative to Thermus thermophilus, thus representing a valuable tool for in vitro and in silico comparative experiments (Omelchenko et al., 2005). The first issue examined in the present study was the possible existence of other bacterial clades that might extend and enrich such studies. We found that like D. radiodurans (Battista et al., 1999), when exponential-phase nonsporulating cultures of G. obscurus are exposed to a 5-kGy dose of gamma irradiation, these cells survive with almost no loss of viability (Fig. 1). In addition, as with D. radiodurans (Battista et al., 1999; Blasius et al., 2008), the ability of G. obscurus to survive the accumulation of DNA DSBs following exposure to a 10-kGy dose of gamma irradiation is not only because of the protection of DNA, because individual cells suffer massive DNA damage (result not shown), but because of other mechanisms that might have analogies to those of D. radiodurans (Cox & Battista, 2005; Blasius et al., 2008; Daly, 2009).

Of the other two species we studied that are related to G. obscurus, which belong to the two other genera in the family Geodermatophilaceae (Normand, 2006), M. multiseptatus was also ionizing radiation resistant, but very surprisingly, B. saxobsidens was ionizing radiation sensitive. To our knowledge, this is the first study that demonstrates the presence of a clade similar to the DeinococcusThermus clade in terms of the presence of resistant/sensitive traits to ionizing radiation. This finding is in agreement with previous data, indicating that the DeinococcusThermus and the GeodermatophilusModestobacterBlastococcus clades both belong to the Terrabacteria group known for its resistance to environmental hazards (Battistuzzi & Hedges, 2009). Yet, compared with the DeinococcusThermus clade, the GeodermatophilusModestobacterBlastococcus clade has morphological features that may help in the adaptation to dry biotopes including the formation of spores (Table 1).

The two clades are also comparable in that both contain besides mesophiles, the thermophiles Thermus and Acidothermus, both found in Yellowstone hot springs. Thermus is mostly known for the production of the thermostable Taq enzyme, whereas Acidothermus was isolated from a 55 °C pond in a search for thermostable cellulases (Mohagheghi et al., 1986) and its genome was shown to be a recent adaptation to high temperature (Barabote et al., 2009). This adaptation has been shown to have occurred within about 250 MY (Normand et al., 1996), as estimated from the 5% distance between 16S rRNA gene sequences using the rate of 50 MY/% proposed by Ochman & Wilson (1987). The three Geodermatophilaceae species with 3–4% divergence in their 16S genes would also have appeared at about the same time (200 MY), whereas the Thermus-Deinococcus clade with 24% distance in their 16S genes would be much more ancient.

Ionizing radiation resistance, which is correlated with the production of ROS (Daly et al., 2007; Daly, 2009) has also been correlated with tolerance to desiccation (Dose et al., 1991, 1992; Mattimore & Battista, 1996; Battista et al., 1999; Billi et al., 2000; Tanaka et al., 2004; Rainey et al., 2005; Fredrickson et al., 2008), to mitomycin C Schiavano et al., 1990), and to hydrogen peroxide (Imlay et al., 1988; Cabiscol et al., 2000). Blastococcus saxobsidens was relatively more sensitive to the lethal effects of desiccation, mitomycin C, and hydrogen peroxide compared with G. obscurus and M. multiseptatus (Fig. 1). However, although ionizing radiation and salt may both generate ROS stresses (Daly et al., 2007; Ashraf, 2009; Daly, 2009; Pan et al., 2009), resistance to ionizing radiation of G. obscurus, M. multiseptatus, and B. saxobsidens failed to correlate with their tolerance to salt (Fig. 1 and Table 1), which is in agreement with a previous study (Shukla et al., 2007). Moreover, the fact that ROS production is a common result of exposure to gamma irradiation (Daly et al., 2007; Daly, 2009) and to heavy metals (Stohs & Bagchi, 1995) was not reflected in the resistance of G. obscurus, M. multiseptatus, and B. saxobsidens to these two stresses (Fig. 1 and Table 1). For example, levels of arsenate (inline image) tolerance of G. obscurus (50 mM), M. multiseptatus (10 mM), and B. saxobsidens (85 mM) were very different. These tolerance levels did not correlate with their resistance to ionizing radiation of the respective organisms. Considered together, all these data can be used to formulate hypotheses about the evolutionary paths and the ecological niches of G. obscurus, M. multiseptatus, and B. saxobsidens. For example, the higher resistance to arsenate of the ionizing-radiation-sensitive B. saxobsidens than the other two strains points out the possibility of a trade-off between these physiological adaptations.

Our data show an agreement between the resistance phenotype and the capacity of occupying specific ecological niches in the calcarenite stone examined and suggest that the resistance to high-energy UV radiation can be a driver for establishing a species on the stone surface. Apparently, at least for the culturable Geodermatophilaceae, all the examined samples from the stone surface were found to yield radio-resistant Modestobacter isolates, while from the deeper (about 2 cm) fraction of the calcarenite, Blastococcus dominated among the culturable Geodermatophilaceae. According to our data measured for the reference strains, both genera exhibit similar resistance to salt and heavy metals/metalloids, but major differences in the ROS-generating resistance patterns. The sensitivity of Blastococcus to high-energy UV radiation can explain its dominance in the deeper part of the stone that should be more protected from the radiation, while its exposure to salts and heavy metals/metalloids at depth should be higher than or similar to that at the surface. By contrast, Modestobacter, which is better adapted to cope with ROS-generating stresses, is mostly confined on the surface and near-surface of the stone instead of its interior. Other physiological features of Blastococcus/Modestobacter do not fit with the localization pattern observed. For instance, Blastococcus has lower requirements for oxygen than the two other genera of the family Geodermatophilaceae, being able to grow under microaerophilic conditions (Normand, 2006). Another interesting differential feature of Blastococccus is a pigmentation pattern from Modestobacter and Geodermatophilus on Luedemann medium, being capable of producing only orange-colored pigment(s), different from that of the two other genera that are jet black, and in the case of Modestobacter has an initial pink-orange pigmentation of young colonies.

The ROS resistance profiles of the three genera of the Geodermatophilaceae, together with the stone localization patterns observed for Modestobacter and Blastococcus on calcarenite, suggest that the relationship between ROS resistance pattern and capacity of colonizing sun-exposed surfaces should be more extensively investigated. The response to light-induced oxidative stress may be an important driver of niche colonization in the stone biotope. The soon to be published genomes of M. multiseptatus, and B. saxobsidens together with that recently published of G. obscurus (Ivanova et al., 2010) should help understand better how the three strains have acquired their distinct physiological characteristics. An exhaustive gene search for the different determinants known to confer metals resistance and for DNA repair enzymes known to help coping with radiation damages will then become possible.

Acknowledgements

Thanks are expressed to José Vicente from the CEA (Cadarache, France) for providing access to the gamma irradiation facility. This work is partially supported by CMCU (Comité mixte Tuniso-Français pour la coopération inter-Universitaire no10/G0903) and BIODESERT (European Community's Seventh Framework Programme CSA-SA REGPOT-2008-2 under grant agreement no 245746) projects.

Ancillary