UV-Resistant Actinobacteria from High-Altitude Andean Lakes: Isolation, Characterization and Antagonistic Activities^†

Site description and sampling

Table 1 indicates all the sampling sites where the isolates were obtained. They included the following lakes: Laguna Diamante (Catamarca, Argentina), Antofalla (Catamarca, Argentina), Laguna Santa María (Salta, Argentina), Laguna Socompa (Salta, Argentina), Tolar Grande (Salta, Argentina) and Salina Grande (Jujuy, Argentina). All of them belong to the HAALs, spreading out at up to 6000 m altitude, comprising ecosystems along the Altiplano–Puna Plateau up to the High Andes. As the ozone column is naturally thinner over the tropics and the UV intensity increases with the altitude, the Puna–High Andes region is consequently exposed to an exceptionally high solar irradiation including harmful UV levels; climatologists and biologists exploring the Puna–High Andes region reported astonishing solar levels in different sampling points or climatological stations for ground measurements 29, 30 reaching 10 W m⁻² in some lakes 1-4. Among these lakes, Laguna Socompa is placed at the base of the still active volcano Socompa. It is located at an altitude of 3570 m, and Farías et al. (2013) reported for the first time in Argentina stromatolites forming at the shore of this lake. These authors also studied the physicochemical characteristics of the water surrounding stromatolites and found that during the summer field campaign (February), the lake water at the stromatolites site was relatively warm (20–24°C), alkaline, rich in dissolved ions such as Na⁺, K⁺, Mg²⁺, Ca²⁺, Cl₂ and SO₄⁻² (total conductivity up to 130 mS cm⁻¹), rich in organic carbon, nitrate, phosphate, silicate and iron, and contained a strikingly high amount of arsenic. Other extreme lake included in this study was Laguna Diamante, located inside the crater of Galan Volcano and placed at 4589 m (Catamarca, Argentina), which presents a unique set of conditions, that is high pH (9–11), high arsenic concentrations (up to 230 mg L⁻¹), high salinity (270 g L⁻¹, 217 mS cm⁻¹), high UV radiation (84 Wm⁻² of UVA–UVB at noon). Gaylussite crystals covered by a red biofilm were collected from this lake and used for our isolation procedures 31.

Three HAAL samples were collected in February 2010 and October 2011, during austral summer and spring, respectively. The samples were obtained from soil, sediment, modern stromatolites and water from multiple lakes (Table 1, Fig. 1). All the samples were collected in sterile flasks and stored between 5 and 10°C until further processing at the laboratory 6.

Some strains previously isolated from Lagunas Azul, Vilama, Aparejos, Chaxas, Huaca-Huasi and Salina Grande (identified with A, V, Ap, Ch, H and LC in the strain code, respectively) and now belonging to the LIMLA-PROIMI Extremophilic Strain Collection were also used in this study. The isolates from Lagunas Socompa, Antofalla Verde and Rojo, Grande, Diamante, Tolar, Negra, Verde, Cabe and Pocitos (Act for all lakes and St for the ones isolated from stromatolites in the strain code) are reported in this study (Tables 2 and 3). Permission for sample collection in the Salta's lakes was granted by the Ministerio de Ambiente y Producción Sustentable, Salta, Argentina (number 000388; 17–09–2010).

Culture media and isolation procedure

Soil, sediment, modern stromatolites, biofilm, gaylussite crystals (associated with red biofilm 31, flamingo feces and water samples (1 g and 1 mL, respectively)) from lakes located in the HAALs were aseptically diluted in sterile 0.9% NaCl solution (9 mL) and incubated in an orbital shaker at 30°C at 150 rpm for 30 min. The suspensions were allowed to settle, and serial dilutions up to 10⁻⁶ were prepared. One hundred microliters of serial dilutions was spread over the surface of isolation agar media on Petri dishes. The following selective media were used to isolate Gram-positive bacteria 32: (i) GYM agar (10 g of malt extract, 4 g of yeast extract, 4 g of glucose, 2 g of Ca₂CO₃, 1000 mL of distilled water, 15 g agar, pH 7.5); (ii) ISP-2 32 agar (10 g of malt extract, 4 g of yeast extract, 4 g of glucose, 1000 mL of distilled water, 15 g agar, pH 7.2); (iii) stromatolite agar (STA, 100 g dry stromatolite, 6 g yeast extract, 15 g agar, 1000 mL of distilled water). The pH of all media was adjusted to 7.0–7.5. All media were supplemented with nalidixic acid (10 μg mL⁻¹) and cycloheximide (10 μg mL⁻¹) to inhibit the growth of Gram-negative bacteria and fungi, respectively. The isolation plates were incubated at 30°C for 2 weeks. Colonies of different size, color and morphology were streaked individually several times until single colonies of a single type were observed. Stock cultures were prepared for each strain by transferring biomass from each of the purified isolates into vials containing 1.5 mL 20% (w/v) sterile glycerol solution and stored at −20°C.

DNA extraction, 16S rDNA amplification and electrophoresis analysis

The 51 selected isolates were subjected to 16S rDNA sequence analysis for precise genera identification. To prepare cultures for isolation of genomic DNA from the strains, a single colony was transferred to a sterile flask with 10 mL of LB. The flasks were incubated at 30°C under stirring at 150 rpm until dense growth was observed (2–5 days). Cell material from 1.5 mL of these cultures was harvested by centrifugation, and genomic DNA was extracted using the DNAeasy Kit (Qiagen) according to the supplied protocol and stored at −20°C prior to use. The 16S rRNA gene was amplified from genomic DNA by PCR with the bacterial universal primers F27 and R1492. PCR products were run in 0.8% agarose gel, stained with SYBR Safe (Invitrogen) and visualized using a Gel DocTM XR+ with Image LabTM software (Bio-Rad). Purification of DNA from gel slabs was performed using QIAquick Gel Extraction Kit (Qiagen), and DNA sequencing was performed by the dideoxy chain termination method with an ABI Prism 3730XL DNA analyzer, using the ABI Prism BigDye Terminator Cycle Sequencing Ready Reaction Kit (PE Biosystems) according to the manufacturer's protocol.

The obtained DNA sequences were compared with the data accessible through NCBI (National Centre for Biotechnology Information) using BLASTN. Sequences were aligned with most closely homologous Actinobacteria 16S rDNA sequences from the GenBank, and a phylogenetic tree was constructed using the Molecular Evolutionary Genetics Analysis (MEGA) software version 3.1 33. The bootstrap values (1000 replicates) were obtained using the neighbor-joining method provided by the software. DNA sequences were deposited to GenBank under accession numbers KJ187409 through KJ187466.

Phylogenetic analysis

The isolates were identified using the EzTaxon server (http://www.ezbiocloud.net/eztaxon), on the basis of 16S rRNA sequence data. The most closely related type strains were also obtained from this server. GenBank accession numbers are shown in Table 2.

Sequences from 16S rDNA were aligned online with Silva Incremental Aligner (SINA) to the rRNA gene databases provided by the SILVA ribosomal RNA project. Based on this alignment, phylogenetic trees were constructed with Fasttree 2.1.7 with the maximum-likelihood method using Jukes–Cantor evolution model. Trees were drawn in MEGA6.

Amplification of the 16S–23S rDNA spacer region (ITS) fingerprinting

PCR amplification of 16S–23S ribosomal DNA (rDNA) ITS fingerprinting was performed as previously described 34, 35. For DNA isolation directly from colony material, a single colony was picked from a fresh culture on GYM medium and resuspended in 100 μl of sterile distilled water in a 0.5 mL tube. The mixture was vortexed for 15 s and then incubated at 99°C for 10 min. After centrifugation (5000 g, 4°C, 5 min), the supernatant containing the released DNA was used as the source of target DNA for PCR amplification. The internal transcribed spacer (ITS) region was amplified using primers S-D-Bact-1494-a-S-20 (5′-GTCGTAACAAGGTAGCCGTA-3′) and L-D-Bact-0035-a-A-15 (5′-CAAGGCATCCACCGT-3′). The PCR protocol used consisted of an initial denaturation step of 4 min at 94°C, followed by 30 cycles of 1 min denaturation at 94°C, 7 min annealing at 55°C and 2 min extension at 72°C, and a final extension of 7 min at 72°C and was maintained at 4°C until tested. Ramps between the denaturation step and hybridization is 50% compared to the default and from the hybridization step, and the extension is 95% regarding default. PCR products were run in 2% agarose gel, stained with SYBR Safe (Invitrogen) and visualized using a Gel DocTM XR+ with Image LabTM software (Bio-Rad).

Phenotypic trait analysis: morphology

After purification, fresh colonies incubated at 30°C on LB agar for 48 h were prepared for phenotypic trait analysis. Morphological characteristics of the 183 isolates were examined by light microscopy (400–1000X NIKON 80i eclipse, Japan) with Nikon Digital Sight camera DS-5M-L1. Gram stain status was determined with standard Gram stain according to the supplied protocol (Britania). Colony aspect was described as Bergey's Manual of Determinative Bacteriology 36 on ISP-2 agar media. For scanning electron microscopy, selected strains were grown on ISP-2 agar. The agar blocks were cut from the growth medium and fixed (glutaraldehyde 4.25% in phosphate buffer 0.1 m pH 7.4) for 3 h at 4°C followed by dehydration in a graded acetone series. The critic drying point was obtained by exchanging the acetone through liquid CO₂. The samples were covered by gold and visualized using a Zeiss Supra 55VP (Carl Zeiss NTS GmbH, Germany) scanning electron microscope (CIME-CCT-CONICET).

Phenotypic trait analysis: UV resistance and polyextremophilic profile

Polyextremophilic assays were carried out with 38 new isolated strains and 13 selected strains from the LIMLA collection (all Gram-positive strains). The drop method 10 was used for the screening of all parameters using LB agar media. Cell cultures were grown overnight, after measuring optical density at 600 nm (OD₆₀₀) to achieve a dilution of 0.5 at OD₆₀₀. In the case of strains with mycelial growth (Streptomyces), spore suspensions were produced according to Albarracin et al. (2010). Subsequently, the cell/spore suspensions were subjected to serial dilutions. Petri dishes were divided in 24 squares, so 24 strains (10 μL aliquots of 10⁻¹ dilution) were then loaded onto LB agar plates.

To evaluate UVB resistance (280–320 nm), LB agar plates were exposed to 2.0–3.0 W m⁻² UVB radiation at different times (0, 30, 60, 90, 120, 180 and 240 min) using UVB lamps (09815-06, Cole Parmer Instruments Company), with an emission maximum at 312 nm. The radiation intensity reaching agar surface in disposable Petri plates was measured under the plastic lid using a UVB/UVA radiometer (09811-56, Cole Parmer Instrument Company) to 312 nm with an amplitude band of 300–325 nm, and the plated were then incubated in the dark to prevent photoreactivation for 72 h at 30°C. Escherichia coli DH5α and Streptomyces coelicolor and Exiguobacterium sp. S17 and Acinetobacter sp. Ver3 strains were used as negative and positives controls, respectively, for in situ UVB resistance assay. Controls of unexposed samples were run simultaneously in darkness. Microbial growth was recorded with three positive signs (+++) when it was similar to the growth in the control, with two positive signs (++) when it was slightly different from the growth in the control, with one positive sign (+) when the growth was too low to allow colony counting, with a negative sign (−) when it was no growth at all. For plotting the data in a graphic, three, two or one sign was taken as 3, 2 or 1 unit of growth for each dilution and then summed up together to reach a unique value. Negative signs were counted as null unit of growth.

As the environmental origin of the strains was typically saline (Table 4), we tested halotolerance and halo-alkalotolerance by growing the strains in salt-supplemented media (NaCl and Na₂CO₃). The bacteria were grown in liquid medium until OD₆₀₀: 0.6 and then diluted 10⁻¹ with sterile physiological saline solution. In the case of strains with mycelial growth (Streptomyces), spore suspensions were produced according to Albarracin et al. (2010) and then diluted 10⁻¹ with sterile physiological saline solution. Ten microliters of these dilutions was inoculated to LB agar media supplemented with different concentrations of Na₂CO₃ 2.5 and 5.0% and NaCl 1 to 30% w/v. Plates were incubated at 30°C for 72 h. To determine the tolerance of arsenic [As(V) and As(III)], LB agar media containing different concentrations of arsenic, As(III), 2.5, 5, 7.5 and 10 mm and As(V), 50, 100, 150 and 200 mm, were spread with 10 μL aliquots of 10⁻¹ dilution of each strain. The plates were incubated at 30°C during 72 h. For further characterization of the strains, the temperature range at 4, 8, 15, 20, 25, 30, 37, 45, 50 and 65°C and pH range from 4 to 12 for growth were studied in LB agar media. All polyextremophilic assays were conducted in triplicate. Control LB agar plates were inoculated with the selected strains without the stress factor (salts, pH, temperature and arsenic). Microbial growth was recorded and plotted in graphics as indicated for UV assays 13.

Antagonistic assays: secondary metabolites production screening

The capacity of the active isolates to produce secondary metabolites was tested. Isolates were suspended in distilled water and inoculated in the surface of modified Bennett's agar media (1 g of beef extract, 1 g of yeast extract, 10 g of glucose, 2 g of peptone, 10 g of NaCl, 15 g of agar, 1000 mL of distilled water) by streaking the producer across the middle of the plate. This was performed to allow the culture to be established on the agar surface and let secondary metabolites to spread out prior to inoculation of the plates with other strains. After 7 days of incubation at 30°C, six species of test bacteria were streaked perpendicular to the producing organism; this includes the same 51 chosen strains and another 21 strains (from the 183 isolates) of Bacillus. Staphylococcus aureus, Enterococcus faecalis, Acinetobacter sp. Ver3, Exiguobacterium sp. S17, Escherichia coli and Rhodotorula representatives were additionally used as controls. The strains were incubated at 30°C for 3 days 37. The production of cytotoxic compounds is expressed by the inhibition zones between the different isolates. Microbial inhibition score was recorded with three positive signs (+++) when growth was totally inhibited, with two positive signs (++) when it was partially inhibited, with one positive sign (+) when the growth was little inhibited, with a negative sign (−) when there was no inhibition at all. For plotting the data in a graphic, we followed the same procedure described for polyextremophilic assays 37.

Isolation of Gram-positive strains from diverse HAAL's niches

Selective isolation techniques were employed to specifically obtain actinobacteria from different niches of the HAALs (Fig. 1 and Table 1). A total of 183 Gram-positive isolates were acquired from all samples with shapes varying from rods to hyphae. Putative 38 actinobacterial strains plus 13 actinobacteria from the LIMLA collection were selected for further studies (Fig. 2). Tables 2 and 3 summarize the source and type of sample, and the phenotype of the selected isolates.

Molecular taxonomical and morphological characterization of selected HAAL actinobacteria

Sequencing of the 16S rRNA gene produced partial sequences of ca. 1400 bp in length, which allowed identify all isolates as actinobacteria. Accordingly, they were assigned to the genera Streptomyces, Micrococcus, Microbacterium, Nesterenkonia, Kocuria, Rhodococcus, Arthrobacter, Micromonospora, Blastococcus, Brevibacterium and Citrococcus, with sequence identity above 98% to type strains. A tree was constructed using the isolates and the most similar type strain to each (Fig. 3). Some genera presented more variation; for example, the two Rhodococcus strains were more similar to distinct type strains than to each other. The same situation arises in the Microbacterium isolates. However, the three Nesterenkonia isolates were closely related and showed the highest similarity with the same type strain. For instance, several strains belonging to Streptomyces were also very closely related; in fact, the isolates Act 8 and Act 25 had exactly the same sequence but depicted clear phenotypic differences. This could be also a limitation of the databases regarding certain genera, but it is not the case for Streptomyces. A “rare” strain is found between the isolates: Act 15 is the strain with less similarity to a type strain, being 95.69% similar to Streptomyces pratensis ch24T (JQ806215). For details, see Figure S1.

ITS fingerprinting method is a molecular tool based on the sequence and length heterogeneity of the bacterial rRNA operon 16S–23S intergenic spacer and provides a high phylogenetic resolution. It can discriminate bacterial isolates up to the subspecies level. To manage the large set of isolates in our collection, ITS-PCR fingerprinting was applied as a complementary screening method. Among the 183 isolates, the 51 chosen strains proved to have 27 distinct groups (G1–G27) (Fig. 3). Each group was composed by strains belonging to the same genus, although some genera were represented by more than one group. The most common pattern was ITS-G3 constituted by 11 Streptomyces strains from Laguna Socompa. The second most frequent pattern was ITS-G18, formed by six Micrococcus strains isolated from Laguna Socompa and Chaxas. Supporting Information Table S1 and Fig. 3 summarize the ITS groups.

The distribution of the isolated actinobacteria with respect to the sampled lake (Fig. 4A), the niche (Fig. 4B) and the genera (Fig. 4C) was likewise analyzed. Laguna Socompa was the lake in which the major quantity of strains was isolated from, followed by L. Chaxas, L. Diamante and L. Azul. In contrast, L. Antofalla, Tolar Grande, Santa María, Salar Pocitos, L. Verde and L. Negra did not yield any actinobacterium in spite of the different media and procedures employed. The niche of isolation was also important for obtaining actinobacteria: Soil samples were by far the ones producing the major quantity of isolates which is known to be a common habitat for actinobacteria (Albarracin et al., 2010), followed by water, feces and stromatolites. Most isolates were identified as belonging to the Streptomyces genus, followed by Micrococcus and Microbacterium. Strains belonging to Nesterenkonia, Kocuria, Rhodococcus, Arthrobacter, Micromonospora, Blastococcus, Micromonospora and Citrococcus were less abundant in our samples.

The morphology of representatives of each genus was observed using SEM (Fig. 4D). For Streptomyces, it was observed the typical hyphal network and for Microbacterium the short rods occurring singly or in irregular clusters. The Nesterenkonia representative showed their coccoid cells grouped in pairs, tetrads and irregular clusters, while Kocuria cells grouped in tetrads. The Arthrobacter strain presented coccoid shape, a morphotype commonly found in this genus when cells are in the stationary phase or when they are exposed to desiccation and starvation 38.

UV resistance and polyextremophilic profile

To determinate the ability of the selected 51 actinobacterial strains to tolerate high doses of UV radiation and other extreme conditions common at their original environment, we carried out physiological assays against different extreme conditions as explained in the above-method section. The extreme conditions tested cover the range of original conditions assumed to be present by previous analyses of our group (1–8) and also those considered opposite, performed as controls and to assess complete physiological ranges. For instance, high pH is common at the lake samples, but we also tested low pH. The polyextremophilic profiles of each strain are then summarized (Table 4).

Data obtained from the physiological assays are shown in Fig. 5 as percentages of the strains belonging to each genus that had a good growth after exposed to a given stress factor. Note that the percentages were based on the number of strains per genera, that is 25 strains for Streptomyces, nine for Micrococcus, four for Microbacterium, three for Nesterenkonia and Kocuria, two for Rhodococcus and only one strain from the rest of genera. For the graphic, two temperature ranges were considered as low (4–30°C) and high (37–65°C). In this case, just few strains were able to grow in the low range of temperature. Although representatives of Streptomyces, Micrococcus, Nesterenkonia, Kocuria and Brevibacterium were able to grow at 4°C, this was not optimum. In turn, a good number of strains grown in the range of high temperature and belonged to Micrococcus, Streptomyces, Micromonospora, Arthrobacter, Brevibacterium, Rhodococcus and Blastococcus genera. In comparison, control strains including Streptomyces coelicolor showed poor growth at higher temperatures (data not shown).

Regarding UVB resistance, all 51 actinobacterial isolates showed a high tolerance to the exposure from 2.0 to 3.0 W m⁻² UVB radiation at different times corresponding to 100% of strains per genera. Only two Streptomyces strains (Act 11 and Act 23), both isolated from the soil of Laguna Socompa, did not tolerate the maximum time of irradiation, being able to tolerate only 180 and 120 min of exposure to UVB radiation, respectively, resulting in this case to a value of 92% for Streptomyces. Streptomyces coelicolor, a common laboratory horse strain in actinobacteria research, could develop well after 90 min of exposure but was inhibited after 120 min of exposure. In turn, Escherichia coli was quite sensible to UVB; that is, the inhibition of growth was evident after 60 min, while after 240 min, the strain was not able to develop.

The same diluted drop protocol was performed to evaluate tolerance to Na₂CO₃ and NaCl, using culture media supplemented with different concentrations of Na₂CO₃ (2.5 and 5.0 mm) and NaCl (1–30% w/v). All isolates tolerate up to 5% NaCl (w/v), being moderately halophiles (Table 4). Moreover, 21 strains belonging to Streptomyces, Micrococcus, Nesterenkonia, Kocuria, Arthrobacter, Micromonospora and Blastococcus genera grew on NaCl up to 25% (Table 4). Although HAALs’ salinities varied between 5 and 32%, we found no strain growing at NaCl 30%. Only seven strains were able to grow on LB agar media supplemented with Na₂CO₃ 5.0 mm (Fig. 4), which included the three Nesterenkonia strains Act 18, Act 20 and Act 27 (100% of the strains), and considered facultative alkalophilic salt-tolerant bacteria. In contrast, control actinobacterial strain S. coelicolor was not able to develop well in salt-supplemented media. It was inhibited at concentrations of NaCl 10% to 30% and with concentrations of Na₂CO₃ 2.5 and 5.0 mm.

Growth at different pH values was analyzed in the pH 4–12 range. It was obvious from Fig. 5 that HAALs’ actinobacterial strains were able to tolerate better the high pH than the low pH. Thus, 100% of strains (50 isolates) belonging to the genera Micrococcus, Streptomyces, Nesterenkonia, Micromonospora, Arthrobacter, Brevibacterium, Rhodococcus, Microbacterium, Kocuria, Citricoccus and Blastococcus were able to develop at pH 12; the only exception was Kocuria Act 48 (Table 4). In turn, only 100% of the strains belonging to Microcomonospora and Citrococcus could grow well at low pH. Also, some isolates belonging to the genera Micrococcus (44%), Streptomyces (4%), Rhodococcus (50%) and Kocuria (66%) could grow at this low pH. For the control, Streptomyces coelicolor, the range of growth was between pH 6 and pH 10.

Thirty-seven strains tolerated 7.5 mm As (III), while only 13 (Microbacterium, Micromonospora, Blastococcus, Micrococcus, Brevibacterium, Rhodococcus and Citrococcus) were able to grow at 10 mm As (III). Regarding As (V), 46 grew at the maximum concentration tested (200 mm) including members of the eleven genera. In contrast, Streptomyces coelicolor could grew well in the range of As (III) 2.5 mm to 7.5 mm but was more sensitive for the As (V), showing inhibition of growth at concentrations higher than 100 mm.

Antagonistic activities of polyextremophilic actinobacteria

To preliminarily determine whether the isolates produced any antimicrobial compounds, an antagonistic assay was performed by confronting all isolated strains with each other on agar nutrient media yielding a 51 × 72 isolate array where 3672 antagonism assays were scored. The failure of several species to grow near the producing organism may indicate that it produced an active compound against these competitors.

From the results, it is obvious that all the isolates exhibited antagonistic activities against at least one of the tested bacteria; these also included E. coli, Bacillus, Enterococcus faecalis, Staphylococcus aureus and two yeasts (Rhodotorula sp.). The production of antagonistic compounds was measured and plotted in a graphic (Fig. 6), where three, two or one sign was taken as 3, 2 or 1 unit of inhibition and then summed up together to reach a unique value. Negative signs were counted as null unit as indicated in material and methods 37. All strains with a score above the threshold value 120 were considered as highly productive (Fig. 7A). The data obtained clearly showed that the representatives of Streptomyces, Microbacterium and Micrococcus were potentially the most prominent antimicrobial producers among the tested strains. The origin of isolation (niche) was also important with regard to bioactivity (Fig. 7B). Strains isolated from feces were much more productive than those isolated from soils and water, while strains isolated from the stromatolite were the less active strains.