Sensitivity of Haloquadratum and Salinibacter to antibiotics and other inhibitors: implications for the assessment of the contribution of Archaea and Bacteria to heterotrophic activities in hypersaline environments
Rahel Elevi Bardavid,
Department of Plant and Environmental Sciences, The Institute of Life Sciences, and The Moshe Shilo Minerva Center for Marine Biogeochemistry, The Hebrew University of Jerusalem, Jerusalem, Israel
Correspondence: Aharon Oren, Department of Plant and Environmental Sciences, The Institute of Life Sciences, The Hebrew University of Jerusalem, Jerusalem 91904, Israel. Tel.: +972 2 6584951; fax: +972 2 6528008; e-mail: email@example.com
Antibiotics and bile salts have been used to differentiate between heterotrophic activity of halophilic Archaea and Bacteria in saltern ponds. In NaCl-saturated brines of crystallizer ponds, most activity was attributed to Archaea. Following the recent isolation of Haloquadratum, the dominant archaeon in the salterns (reported to be sensitive to chloramphenicol and erythromycin), and the discovery of Salinibacter, a representative of the Bacteria, in the same ecosystem, reevaluation of the earlier data is required. The authors measured amino acid incorporation by Haloquadratum and Salinibacter suspended in crystallizer brine to investigate the suitability of antibiotics and bile salts to distinguish between archaeal and bacterial activities. The amino acid uptake rate per cell in Salinibacter was two orders of magnitude lower than that of Haloquadratum under the same conditions. Salinibacter was inhibited by chloramphenicol, erythromycin, and deoxycholate, but not by taurocholate. Erythromycin did not inhibit incorporation by Haloquadratum, but moderate inhibition was found by chloramphenicol at 10–50 μg mL−1. Deoxycholate was highly inhibitory, but only partial inhibition was obtained in the presence of 25 μg mL−1 taurocholate. Inhibition by chloramphenicol and taurocholate increased with increasing salt concentration. Erythromycin and taurocholate proved most valuable to differentiate between archaeal and bacterial activities in saltern brines.
Hypersaline environments are inhabited by representatives of the two domains of prokaryotes: Archaea and Bacteria. Archaea of the family Halobacteriaceae are typically found at the highest salinities, including in salt-saturated lakes such as the Dead Sea and the northern part of Great Salt Lake, UT (Oren, 1994, 2002a). Within the domain Bacteria, many halophiles can be found as well. Most are moderately halophilic, thriving best at salt concentrations up to 100–200 g L−1, but some grow up to salt saturation (Ventosa et al., 1998; Oren, 2002a).
Solar salterns for the production of NaCl by evaporation of seawater consist of a series of ponds with increasing salinity, up to saturation in the crystallizer ponds in which halite precipitates. A succession of life forms is found in such salterns, from organisms adapted to seawater salinity in the first evaporation ponds to extreme halophiles requiring salt concentrations of 150–200 g L−1 and above (Pedrós-Alióet al., 2000; Oren, 2002a). Halophilic Archaea dominate in the highest salinity ponds, and the red color of the crystallizer brines is at least to a large extent due to the bacterioruberin carotenoids of the Halobacteriaceae and possibly to their retinal pigments as well.
In the past, attempts have been made to assess the relative importance of Archaea and Bacteria along the salt gradient in saltern systems in Israel and in Spain, based on the use of inhibitors allegedly specific for either group. The use of low concentrations of bile salts (taurocholate, deoxycholate) to differentiate between members of the Halobacteriaceae and other types of halophiles was already suggested in the 1950s, when it was discovered that Halobacterium cells are lysed by low concentrations of bile salts (Dussault, 1956). This finding was long forgotten, but the phenomenon was rediscovered in the 1980s (Kamekura et al., 1988; Kamekura & Seno, 1991). It was subsequently used as the basis for attempts to differentiate between bacterial and archaeal heterotrophic activities along the salt gradient in salterns (Oren, 1990a, 1991, 2002b; Pedrós-Alióet al., 2000; Gasol et al., 2004). Different antibiotics have also been used in such studies, including chloramphenicol and erythromycin, to inhibit bacterial protein synthesis and anisomycin to inhibit the process in Archaea (Oren, 1990b, 1991; Pedrós-Alióet al., 2000; Gasol et al., 2004); aphidicolin was used as an inhibitor of thymidine incorporation in Archaea (Oren, 1990c). The conclusion of these studies was that below c. 250 g L−1 total dissolved salts, all heterotrophic activity could be attributed to representatives of the bacterial domain, and that at the higher salinities Archaea took over.
At the time when the above-mentioned studies were performed, little was known about the nature of the prokaryotes that inhabit the crystallizer ponds of the salterns. The dominant type of organism was known to be a flat, square gas-vacuolate archaeon (Oren, 1994; Guixa-Boixareu et al., 1996; Oren et al., 1996). This organism was only recently brought into culture (Bolhuis et al., 2004; Burns et al., 2004), and has now been described as Haloquadratum walsbyi. The species description (Burns et al., 2007) states that the organism is sensitive to chloramphenicol and erythromycin, two antibiotics used in the past to differentiate between archaeal and bacterial activities, based on the assumption that the Halobacteriaceae are insensitive to these inhibitors. Another recent development was the discovery of the red, extremely halophilic Salinibacter ruber, a representative of the Bacteroidetes phylum of the Bacteria, as a common inhabitant of saltern crystallizer ponds worldwide (Antón et al., 2000, 2002; Oren & Rodríguez-Valera, 2001; Elevi Bardavid et al., 2007). The question should therefore be asked to what extent Salinibacter may contribute to the overall heterotrophic activity in the saltern ponds, and how the different inhibitors used in such studies affect its activity.
Now the two organisms that appear to dominate saltern crystallizer ponds worldwide have been obtained in pure culture, their behavior toward the different antibiotics and other antibacterial substances can be studied, to assess the merit of the different ‘specific’ inhibitors used in ecological studies aimed to obtain an understanding of the prokaryote activities in salterns and in other hypersaline environments. Here, the authors report on studies on the influence of chloramphenicol, erythromycin, and bile salts on Haloquadratum and Salinibacter cells suspended in saltern brines, discuss the impact of the results obtained on the evaluation of earlier field experiments, and suggest improved protocols for such ecological studies in the future.
Materials and methods
Organisms and growth conditions
Salinibacter ruber strain M31T (DSM 13855T) (Antón et al., 2002) was grown at 35 °C in 1-L portions of medium in 2-L Erlenmeyer flasks in an illuminated New Brunswick Innova 44 shaker (100 r.p.m.). The medium composition was (g L−1): NaCl, 195; MgSO4·7H2O, 25; MgCl2·6H2O, 16.3; CaCl2·2H2O, 1.25; KCl, 5.0; NaHCO3, 0.25; NaBr, 0.625; and yeast extract, 1.0, pH 7.0. The medium was sterilized by autoclaving.
Haloquadratum walsbyi strain C23T (JCM 12705T) (Burns et al., 2007) was grown under similar conditions in medium containing (g L−1): NaCl, 240; MgCl2·6H2O, 30; MgSO4·7H2O, 35; KCl, 7, to which 5 mL of 1 M NH4Cl; 2 mL 0.25 M K−PO4 buffer, pH 7; 4.4 mL 25% Na-pyruvate; 10 mL 0.5% peptone (Oxoid)+0.1% yeast extract; 3 mL vitamin solution; and 1 mL trace element solution SL10 (Widdel et al., 1983) were added (L−1). The pH was adjusted to 7 and the medium was sterilized by filtration through 0.2 μm pore-size cellulose acetate filters. The composition of the vitamin solution was (mg L−1): p-aminobenzoate, 13; biotin, 3; nicotinic acid, 33; hemicalcium d-(+)-pantothenate, 17; pyridoxamine·HCl, 50; thiamine·HCl, 33; cyanocobalamin, 17; d,l-6,8-thioctic acid, 10; riboflavin, 10; and folic acid, 4.
Saltern brine sample
Saltern crystallizer brine was collected from pond 304 of the salterns of the Israel Salt Company in Eilat (Oren, 1990a) on 12 June 2006. The brine had a density of 1.235 g mL−1 at 23 °C and a total dissolved salt content of 368 g L−1. The brine was cleared by centrifugation (15 min, 7700 g), followed by filtration through 0.2 μm pore-size membrane filters.
Late-exponential growth phase cells were collected by low-speed centrifugation (2700 g, 10 min at room temperature), and the pellets were gently suspended in Eilat saltern brine cleared by centrifugation and filtration as described above or in such brine diluted with distilled water to 90% or 80% of its salinity as indicated. In one experiment, the Salinibacter culture was placed in SnakeSkinR-pleated dialysis tubing (10 000 molecular weight cut-off; Pierce) and dialyzed four times at room temperature for 20 h against a 10-fold volume of Eilat crystallizer brine. Haloquadratum suspensions were prepared to an OD600 nm of 0.04, equivalent to c. 8 × 107 cells mL−1 as counted microscopically using a Petroff–Hauser counting chamber and a microscope equipped with phase-contrast optics. Salinibacter suspensions contained around 6 × 108 cells mL−1 (OD600 nm=0.29).
Cell suspensions (2 mL in glass test tubes) were preincubated in a water bath at 35 °C for 1 h. Then, inhibitors (bile salts, antibiotics) were added as indicated and preincubation was continued for 20 min. The tubes were then supplemented with 5 μL of l-[U-14C]-amino acid mixture (Amersham CFB 104; 50 μCi mL−1; 52 mCi milliatom−1). At zero time and after different periods of incubation (15 min–4 h), 0.25 mL portions were filtered on glass fiber filters (Whatman GF/C). Filters were washed with 2 × 5 mL cold 10% trichloroacetic acid, dried, and counted in a scintillation counter (Beckman LS 2800) with 5 mL of Packard Ultima Gold scintillation cocktail. Amino acid incorporation rates were calculated, assuming that the labeled amino acids in the mixture had an average molecular weight of 130 and a carbon content of 45%, and that no significant amino acid pool was present in the cell suspensions prior to addition of the label, which would have caused a decrease in specific activity (Oren, 1990b).
The following inhibitors were used in the experiments: Na-taurocholate (Sigma; added from a 10 mg mL−1 solution), Na-deoxycholate (BDH, added from a 10 mg mL−1 solution), chloramphenicol (Sigma, added from a 2 mg mL−1 solution), and erythromycin (Fluka; added from a 1 mg mL−1 solution). Equivalent amounts of distilled water were added to control experiments to keep the salt concentration equal in all systems.
Haloquadratum walsbyi cells suspended in saltern crystallizer brine incorporated labeled amino acids at a rate of 10.3±5.0 nmol h−1 1010 cells−1 (Fig. 1, left panel). In earlier experiments with Eilat saltern brines dominated by Haloquadratum (1.5−4 × 1010cells L−1), amino acid incorporation rates between 5 and 60 nmol L−1 h−1 (Oren, 1990a, b) were measured, i.e., values in the same order of magnitude as those obtained with cultured cells. When the brine was diluted with distilled water, the incorporation rates increased sharply, 10% and 20% dilutions supporting rates two and 3.5 times of those measured in undiluted brine. Variability between experiments was rather large, probably due to differences in the age and physiological state of the cells; however, the same trend was found in each individual experiment with a single batch of cells. The results thus show that the salinity of the saltern crystallizer brine is supra-optimal for Haloquadratum.
Similar amino acid uptake experiments performed with S. ruber cells suspended in saltern brine and its dilutions (Fig. 1, right panel) showed incorporation rates per cell about two orders of magnitude lower than those for Haloquadratum walsbyi. If these rates correctly represent its potential to incorporate amino acids, Salinibacter, which is also present in much smaller numbers, cannot be expected to compete with Haloquadratum for amino acids in the salterns. To exclude the possibility that the low rates measured in these experiments were due to damage to the cells from the gentle centrifugation and resuspension procedure, a batch of Salinibacter culture was equilibrated with saltern brine by extensive dialysis. Amino acid incorporation rates in this suspension did not significantly differ from those obtained in cells collected by centrifugation. Also, microscopically, no damage to the cells could be observed.
Incorporation of amino acids by Haloquadratum in crystallizer brine and its dilutions as described above was completely abolished in the presence of 10 μg mL−1 Na-deoxycholate. In undiluted brine, also Na-taurocholate (25 and 50 μg mL−1) caused complete inhibition. When the cells were suspended in 90% saltern brine, significant activity was measured in the presence of 25 μg mL−1 taurocholate (about 30% of the control rate), and in brine diluted to 80% of its original salinity, some activity was even detected in the presence of 50 μg mL−1 taurocholate (Fig. 2). These results show that the degree of inhibition by the inhibitor is salinity-dependent.
The species description of Haloquadratum walsbyi states that the organism is sensitive to erythromycin (Burns et al., 2007). This was unexpected as other members of the Halobacteriaceae are insensitive to the antibiotic, and significant rates of leucine incorporation were found in Spanish crystallizer ponds in the presence of erythromycin (Pedrós-Alióet al., 2000). In the present study the sensitivity of Haloquadratum walsbyi C23T to erythromycin was not confirmed: no inhibition was observed when the organism was grown in standard medium as described in ‘Materials and methods’, supplemented with 25 or 50 μg mL−1 erythromycin, and incorporation of labeled amino acids by cell suspensions in saltern brine proceeded at the same rate in the presence of 40 μg mL−1 erythromycin as in its absence.
Haloquadratum walsbyi was also reported to be sensitive to chloramphenicol (Burns et al., 2007), an antibiotic used earlier in attempts to differentiate between archaeal and bacterial amino acid incorporation in salterns (Oren, 1990b, 1991). Amino acid incorporation proceeded in the presence of 20 and 40 μg mL−1 chloramphenicol, concentrations used in earlier experiments with field samples. However, the rates were significantly lower than those in the absence of the antibiotic. Similar to what was observed for Na-taurocholate, the degree of inhibition depended not only on the concentration of the inhibitor but also on the salt concentration of the brine in which the cells were suspended (Fig. 3).
Salinibacter was not inhibited by Na-taurocholate at concentrations up to 50 μg mL−1. However, in the presence of 10 and 25 μg mL−1 Na-deoxycholate, amino acid incorporation was reduced to 58% and 18% of the control rate, respectively (Fig. 4). In view of the very low rates of uptake in undiluted brine, this experiment was performed in brine diluted to 80% of its original salinity. Following 3 days of incubation with 25 μg mL−1 Na-deoxycholate, cells appeared less dark in the phase contrast-microscope than did the control cells, suggesting structural damage. Amino acid incorporation by Salinibacter cells in 80% saltern brine was inhibited more than 95% by chloramphenicol (20 μg mL−1) and by erythromycin (20 and 40 μg mL−1).
Oremland & Capone (1988) opened their review on ‘Use of “specific” inhibitors in biogeochemistry and microbial ecology’ with the statement that ‘Inhibitors are like old sports cars: They are fun to play around with, but you should never trust them!’, and they wrote that ‘all work with inhibitors is inherently suspect’. This is surely true when antibiotics and other inhibitors are applied to field samples in microbial ecology studies, as microbial ecosystems are generally complex, and in most cases their main components are even unknown.
In hypersaline environments at or approaching salt saturation the situation is far simpler, as the microbial diversity is low. Based on the knowledge of the sensitivity of different types of halophilic microorganisms to inhibitors, studies have been performed in the past to assess the relative contribution of Archaea and Bacteria in such ecosystems using inhibitors ‘specific’ for either group. Chloramphenicol, erythromycin and anisomycin have been used to differentiate between archaeal and bacterial protein synthesis (Oren, 1990b, 1991; Pedrós-Alióet al., 2000; Gasol et al., 2004). Kanamycin and tetracycline proved unsuitable for the purpose, as they failed to inhibit amino acid incorporation in low-salinity saltern ponds in which Archaea cannot be expected to occur in large numbers (Oren, 1990b). Bile salts, which, at low concentrations, cause lysis of most halophilic Archaea, have also been used to differentiate between the groups. At the time of these studies, it was recognized that members of the Halobacteriaceae dominate the high-salinity environments, but the true nature of the most abundant organisms was still unknown. Now, Haloquadratum walsbyi, the dominant archaeon in the saltern crystallizers, has been isolated (Burns et al., 2007), and now it is known that extremely halophilic Bacteria of the genus Salinibacter may also play a role in such environments (Antón et al., 2000, 2002), controlled experiments can be performed to assess to what extent different antibiotics and other inhibitors can indeed be used to obtain relevant information about the functioning of the ecosystem.
Microbiology textbooks commonly state that Archaea are not inhibited by chloramphenicol (e.g. Madigan & Martinko, 2006). This is not necessarily true: growth of many methanogenic Archaea is inhibited already by low concentrations of the antibiotic (Hilpert et al., 1981; Pecher & Böck, 1981). Schmid et al. (1982) claimed that the ribosomes of Halobacterium salinarum may lack binding sites for chloramphenicol. However, it is now known that chloramphenicol does bind to the large subunit of the Haloarcula marismortui ribosome, albeit at a different site compared with the bacterium Deinococcus radiodurans (Hansen et al., 2003). Reports on the sensitivity of the members of the family Halobacteriaceae to chloramphenicol are highly conflicting. This may in part be due to the different methods used for antibiotic sensitivity testing: liquid cultures or agar diffusion tests. Most authors state that halophilic Archaea (Halobacterium, Halococcus, Halorubrum, Haloferax spp.) are resistant to chloramphenicol or are inhibited only at very high (>100 μg mL−1) concentrations (Chow & Mark, 1980; Hilpert et al., 1981; Schmid et al., 1982; Bonelo et al., 1984; Böck & Kandler, 1985); however, a rather high chloramphenicol sensitivity was reported in Halobacterium salinarum by Pecher & Böck (1981) and by Mankin & Garrett (1991). The report that Haloquadratum walsbyi does not grow in liquid culture in the presence of 50 μg mL−1 chloramphenicol (Burns et al., 2007) is highly relevant for the evaluation of earlier field experiments in which chloramphenicol was used to assess the relative importance of Archaea and Bacteria in saltern crystallizer ponds. The results reported here (Fig. 3) showed a significant degree of inhibition of amino acid incorporation rates already at relatively low concentrations of chloramphenicol. This also explains the partial (28–42%) inhibition by 20 μg mL−1 chloramphenicol in Eilat saltern crystallizer brines (Oren, 1990b), an inhibition that also at the time has been attributed to the action of chloramphenicol on the halophilic Archaea present in the brines rather than to the presence of active Bacteria as the activity was completely abolished by anisomycin.
Members of the Halobacteriaceae are generally insensitive to erythromycin or are inhibited only by very high concentrations (>100–150 μg mL−1) (Chow & Mark, 1980; Hilpert et al., 1981; Pecher & Böck, 1981; Bonelo et al., 1984; Böck & Kandler, 1985). The report that Haloquadratum walsbyi does not grow in the presence of 50 μg mL−1 erythromycin (Burns et al., 2007) is therefore surprising. In the experiments reported in the current paper erythromycin did not inhibit growth of Haloquadratum, and no significant inhibition of amino acid incorporation was found by 20 or 40 μg mL−1 erythromycin in cells suspended in saltern brine. Whether the c. 50% inhibition of leucine incorporation reported by Pedrós-Alióet al. (2000) in saltern crystallizer ponds in Spain was due to a partial inhibition of Haloquadratum by the antibiotic or due to the massive presence of other, more sensitive species, cannot be ascertained.
The earlier use of bile salts, especially taurocholate, in ecological studies (Oren 1990a, 1991; Pedrós-Alióet al., 2000; Gasol et al., 2004) was based on damage of the cell envelope and cell lysis of noncoccoid members of the Halobacteriaceae by low concentrations of such compounds; other prokaryotes are far less sensitive. The experiments with Haloquadratum reported in the present paper show that its behavior toward bile salts is similar to that of most other members of the family, and that low concentrations of taurocholate (25–50 μg mL−1) completely inhibit its activity in saltern brine at the highest salinities.
Based on the current knowledge of the community structure in saltern ponds and on the experiments presented in this paper, it can be concluded that erythromycin (20–40 μg mL−1), anisomycin (10–20 μg mL−1), and Na-taurocholate (25 μg mL−1) are suitable to differentiate between archaeal and bacterial activities in this ecosystem. Chloramphenicol indeed inhibits most halophilic Bacteria, including the extremely halophilic Salinibacter, but especially at higher concentrations, also slows down amino acid incorporation by Haloquadratum (Fig. 3). Anisomycin, known as a protein synthesis inhibitor of eukaryotic ribosomes, is a very effective inhibitor of protein synthesis in halophilic Archaea (Pecher & Böck, 1981); it was also successfully used to selectively enrich and isolate Salinibacter from the saltern environment (Elevi Bardavid et al., 2007).
The experiments with Haloquadratum and Salinibacter suspended in saltern brine diluted to different degrees (Fig. 1) show that for both organisms the salinity of the crystallizer brines is above their optimum, and that their activity in situ is therefore much lower than their potential activity under optimal salinity conditions. Similarly, an increase in the specific growth rate of the heterotrophic community in the crystallizer ponds of Spanish salterns by moderate degrees of dilution was reported by Gasol et al. (2004). The degree of inhibition of Haloquadratum by chloramphenicol and by taurocholate was found to depend on the salt concentration of the brine in which the cells were suspended (Figs 2 and 3). No obvious explanation for this phenomenon was found. However, differences in sensitivity to antibiotics have been observed in the past in several moderately halophilic Bacteria when grown at different salt concentrations (R.H. Vreeland, pers. commun.). These findings also have important implications for the use of taurocholate and different antibiotics in ecological studies of the lower salinity evaporation ponds. In those ponds, the prokaryote diversity is larger than in the crystallizers, and the nature of the dominant heterotrophs is still largely unknown, so that controlled experiments similar to those described in the present study with Haloquadratum and Salinibacter cannot yet be performed for the lower salinity ponds of saltern ecosystems.
Comparison of amino acid incorporation rates per cell in Haloquadratum and Salinibacter when suspended in saltern brine or in dilutions thereof (Fig. 1) shows the rate in Salinibacter to be two orders of magnitude lower than that of Haloquadratum. The difference was probably not due to damage to the Salinibacter cells during the centrifugation step included in the experimental procedure, as cells equilibrated with saltern brine by dialysis showed similar incorporation rates. It appears that Salinibacter cannot efficiently compete for amino acids in the saltern ecosystem dominated by Haloquadratum. These observations agree with the results of microautoradiography experiments following incubation of brines from Spanish salterns with labeled amino acids: the label was found only in the archaeal cells, in spite of the presence of large numbers of Bacteria, including Salinibacter, and in spite of the fact that Salinibacter in pure culture did incorporate amino acids. One explanation brought forward was that the cells may take up amino acids only at a very slow rate (Rosselló-Mora et al., 2003). The question thus remains open as to what the true growth substrates of Salinibacter in its natural environment might be. Rosselló-Mora et al. (2003) were also unable to show uptake of radiolabeled glycerol by Salinibacter in the saltern brines, although the organism is able to use glycerol as a growth substrate (Sher et al., 2004). Analysis of the genome of S. ruber shows broad degradative abilities and a potentially versatile metabolism (Mongodin et al., 2005). However, to what extent it is able to compete with the other components of the microbial ecosystem of the salterns, especially with Haloquadratum, for the available nutrients is still far from clear.
The authors thank the Israel Salt Company in Eilat, Israel, for allowing access to the salterns. This study was supported by the Israel Science Foundation (grant no. 617/07).