Introduction

Actinobacteria are the largest source of biologically active compounds that are widely used in medicine, veterinary, agriculture, and others (Demain and Adrio 2008). Around 55 % of all antibacterial compounds are coming from the species from the genus Streptomyces and 11 % from other actinobacteria (Strohl 1997). The search for new antibiotics constantly continues in order to combat new challenges that are rising due to appearance of new pathogens and development of antibiotic resistance by existing pathogenic bacteria. This search involves isolation of new producing strains, including new actinobacteria. Nowadays, the necessary prerequisite for successful screening program for antibacterial producers is going for new loci (Monciardini et al. 2014). Classic sources of actinobacteria are exhausted, and discovery of new biologically active compounds is rather inefficient. Thus, the screening programs are often oriented to expand the isolation sources to new geographical/geological locations with extreme and/or unusual conditions or symbiotic or pathogenic interactions (Alvin et al. 2014; Nikapitiya 2012). It is a fact that bacterial strains obtained from unusual and extreme ecosystems and their biological communities tend to produce new biologically active compounds (Jenke-Kodama and Dittmann 2009). Successful examples of such programs are isolation of new actinobacteria from wasps and their nests (Kroiss et al. 2010; Madden et al. 2013), leaf-cutting ants (Currie et al. 2003), and other invertebrates, as well as from plants (Alvin et al. 2014; Mushegian et al. 2011), thermal springs (Valverde et al. 2012), radioactive wastes (Bagwell et al. 2008; Mao et al. 2007), and others. Microbial communities of marine sponges have been also studied as a promising source of pharmaceutical and biotechnical important chemicals for many years (Hentschel et al. 2012; Monciardini et al. 2014; Nikapitiya 2012).

The peculiarities and biodiversity of ancient Lake Baikal ecosystem are well known and described by Kozhova and Izmest’eva (1998), Timoshkin et al. (2001), and others. In view of the high degree of the endemicity fauna of Lake Baikal is the promising object for screening for new actinobacteria strains. Amphipods (Amphipoda, Crustacea), gastropods (Gastropoda, Mollusca), freshwater sponges (Porifera), planarians (Planariidae, Plathelminthes), and caddisfly (Trichoptera) are the largest groups of macroinvertebrates with the highest level of biodiversity and distribution (Kozhova and Izmest’eva 1998; Timoshkin et al. 2001). All tested organisms are inhabitants of the lake’s littoral community, experiencing respective environmental and anthropogenic impacts. Baikal amphipod and gastropod groups are widely studied organisms in terms of ecology, taxonomy, and ecophysiology. Sponges are well studied with respect to their filtering activity and microbial consortia. Similarly, planaria have been well documented as parasitic symbionts of fishes and amphipods, and caddisfly larvae have been considered an important indicator for the biomonitoring of ancient ecosystems (Timoshkin et al. 2001).

Organisms inhabit Lake Baikal from the water’s edge to maximum depths (1642 m). In terms of endemic peculiarities and stable environmental conditions, the majority occupy highly specialized ecological niches, including food spectrum, thermal regime, biotic interactions, and others. The greatest proportion of species richness (60–70 %) is concentrated in the littoral zone and represented mostly by benthic invertebrates (Timoshkin et al. 2001). They can accumulate high numbers of microorganisms and their spores due to their active water mass filtering capacities and/or active moving along the bottom of the lake. Dwellers and scavengers of Lake Baikal are of particular interest for actinobacteria isolation purposes.

Several attempts of characterization of Lake Baikal actinobacterial microbiome have been already published by Zakharova et al. (2013). The same researchers reported that actinobacteria represents around 16 % of the entire microbiome of the lake sediments. Studies of the microbiome of two endemic sponge species also demonstrated high occupancy by the actinobacteria species (Gladkikh et al. 2014). Terkina et al. (2002) reported isolation and initial classification of actinobacteria from water and sediments of the lake. Interestingly, these authors observed quite high enrichments of isolates from the lake sediments in Micromonospora species (up to 59 %), while Streptomyces were predominant in the water samples. The same group reported a high dosage of biologically active strains among Lake Baikal isolates, including those inhibiting the growth of pathogenic bacteria as well as tumor cells (Terkina et al. 2006). These studies clearly demonstrated the huge potential of the ecosystem of Lake Baikal for discovery of new species of bacteria with unique metabolic profiles. Thus, the main goal of our research was to investigate the possibility of isolation of new actinobacteria species from the atypical sources, mainly focusing on endemic invertebrates of Lake Baikal, with the focus on the biological activity of obtained species. Here, for the first time, we report the isolation of cultured actinobacteria from several dominant endemic benthic organisms of Lake Baikal including amphipods, planarians, caddisfly, and their analysis in terms of phylogeny and antibacterial activity. The presented data shows a great potential for the discovery of new producers of biologically active compounds from Lake Baikal opening a perspective for deeper investigations in this direction.

Materials and methods

Sample collection, preparation, and isolation of actinobacteria

The endemic species representing dominant taxa from Lake Baikal invertebrates’ fauna were chosen for actinobacteria isolation: spongia (Baicalospongia bacilifera), amphipoda (Pallasea concelloides, Brandtia sp.), caddisfly Trichoptera sp. (larvae), and turbellaria (Baikalobia variegata). The animals were collected in February 2014 in Listvyanka village (South Baikal, N 51.867936, E 104.829715) using a benthic dragnet. The depth of sampling was 5–7 m. Immediately following sampling, the animals were placed into thermostatic conditions at 6 °C with constant aeration and transferred to the laboratory. The samples were washed with sterile water, fixed, and homogenized manually in 20 % sterile glycerol in ratio 1:10. As a negative control, water from tanks where animals were kept was used.

The isolation and cultivation of actinobacteria strains were performed on the solid nutrient media MS (soy flour, 20 g/L; d-mannitol, 20 g/L; agar, 20 g/L; pH 7.2) and Czapek agar (NaNO3, 2 g/L; starch, 30 g/L; MgSO4 7H2O, 0.5 g/L; KCl, 0.5 g/L; FeSO4 7H2O, 0.01 g/L; K2HPO4, 1 g/L; agar, 20 g/L; pH 7.2) at 28 °C supplemented with antibiotics cycloheximide (50 μg/mL) and phosphomycin (100 μg/mL) (Kieser et al. 2000). Homogenates were diluted 1:10, 1:100, and 1:1000 in sterile 1 % saline solution, and dilutions were plated on MS media in 3 replicates. Plates were incubated for 14 days at 28 °C and checked every 24 h for actinomycetes colonies appearance. Actinomycetes were recognized based on colony morphology: solid density of colonies, growth inside of the agar media, and steady border of colonies (Kieser et al. 2000). Colonies with a leathery texture with or without aerial hyphae were transferred from the primary plates onto new MS media. An attempt was made to obtain in pure culture all colonies observed on the primary plates. In this way, 25 isolates were obtained. Active and rare strains were deposited in Russian Collection of Agricultural Microorganisms (RCAM), St. Petersburg. Information about the strain deposition is available on www.arriam.spb.ru (accession numbers: 03463-03476).

16S rRNA gene sequencing and analysis

Strains were grown in 10 mL of TSB media at 28 °C for 3 days, and total DNA was isolated as described (Kieser et al. 2000). Amplification of the 16S ribosomal RNA (rRNA) gene was carried out with primers: 8F (AGA GTT TGA TYM TGG CTC AG) and 1510R (TAC GGY TAC CTT GTT ACG ACT T). Obtained fragments were gel purified using QIAquick Gel Extraction Kit (Qiagen, Venlo, Netherlands) and sequenced with the amplification primers 8F and 1510R generating almost the entire gene sequence (1365~1401 bp). The forward and reverse sequences were assembled with SeqMan DNAStar software (Lasergen, Houston, USA). The entire gene sequences were aligned with the bacterial 16S rRNA gene sequences from the RDP-II database (Wang et al. 2007). Ninety-one rDNA sequences were aligned using MAFFT v7.017 (gap open penalty 1.53, offset value 0.123, scoring matrix 200PAM/k = 2, algorithm: auto) (Katoh 2013). A dendrogram was built using Geneious (Kearse et al. 2012) (Tamura-Nei genetic distance model, neighbor-joining method, Escherichia coli as an outgroup, bootstrap value 1000, consensus tree with 50 % support threshold). Sequences were deposited in the GenBank (accession numbers: KP749314-KP749323, KP749325-KP749338, and KT005998).

Metabolites analysis

Isolated strains were inoculated in 10 mL of TSB, grown for 2 days at 28 °C, and 2 mL of pre-culture were used to inoculate the 50 mL of production media. Nutrient media NL-19 (soy flour, 20 g/L; d-mannitol, 20 g/L; pH 7.2) and SG (glucose, 20 g/L; soy peptone, 10 g/L; CaCO3, 2 g/L; CoCl2, 0.001 g/L; pH 7.2) were used as production media. Strains were cultivated on both media at 28 °C for 4 days. Metabolites from cultural liquid were extracted with ethyl acetate (Sigma, St. Louis, USA). The compounds from biomass and solid MS media were extracted with acetone:methanol mixture (ratio 1:1). The obtained extracts were evaporated and dissolved in methanol. Samples were initially analyzed on low-resolution ion trap LC-MS amaZone system (Bruker, Billerica, USA). Samples were separated on an Ultimate 3000 HPLC system (Dionex, Sunnyvale, USA) using C18 column (Affymetrix, Santa Clara, USA) and linear gradient of acetonitrile against 0.1 % ammonium formate solution in water over time of 20 min with flow rate 0.5 mL/min. Detection was performed in both negative and positive modes. After an initial assessment, samples were analyzed by the ultra-high resolution mass spectrometry using LC-QTOF system maXis II (Bruker, Billerica, USA). HPLC conditions used were the same as for low-resolution LC-MS analysis. The mass detection was performed in both positive and negative modes with the detection range set to 160–2500.

Data were collected and analyzed by Brucker Compass Data Analysis software, version 4.1 (Bruker, Billerica, USA). The screening for known compounds was performed using Dictionary of Natural Products database version 6.1 (CRC Press, Boca Raton, USA), using the following parameters for search: accurate molecular mass, absorption spectra, and source of compounds isolation (Whittle et al. 2003). Compounds were considered to be similar when the difference in accurate mass was less than 0.02 and absorption spectrum was identical.

Biological activity assay of extracts from isolated strains

Antimicrobial activities of extracted metabolites were assayed by disk diffusion method (Ruangpan and Tendencia 2004). For this, 40 μL of each extract was loaded on 6-mm-diameter paper disk. Antibacterial activities of metabolites produced by strains grown on MS solid media were assayed by agar diffusion method. For this, 8-mm-diameter agar block of each culture was placed on the test cultures. Test cultures of Bacillus subtilis ATCC 6633, Staphylococcus carnosus ATCC 51365, Pseudomonas putida KT 2440, E. coli ATCC 25922 (TolC), E. coli K12, and Saccharomyces cerevisiae BY4742 were plated from the liquid cultures on LB broth or YPD (yeast) plates, dried for 20 min prior disk/agar blocks were applied. Several extracts that showed activity against S. cerevisiae were tested against Candida albicans DSM1665. The test cultures were obtained from Leibniz-Institut DSMZ-German Collection of Microorganisms and Cell Cultures (Braunschweig, Germany). The zones of inhibition were measured manually with accuracy ±1 mm.

Results and discussion

Isolation and phylogenetic characterization of actinobacteria from Lake Baikal endemic invertebrates

Several reports on characterization of microbial communities of water and sediments of Lake Baikal have been published earlier (Terkina et al. 2002, 2006). These publications revealed high enrichments of the lake microbiome with the actinobacteria species. Attempts on characterization of the microbiome of Baikal freshwater sponges have been reported as well (Jenke-Kodama and Dittmann 2009; Kaluzhnaya et al. 2011). We assumed that microbial specimens associated with the endemic fauna of Lake Baikal should preserve the high degree of endemicity as well. Isolated evolution of the lake biota further increases the chances that the secondary metabolism genes of the bacterial community also evolved separately from the general pool significantly raising the possibility to find novel chemical scaffolds with new biological activities. The macroinvertebrates of Lake Baikal such as amphipods, sponges, turbellaria, and caddisfly are characterized by high endemism (100, 80, 90, and 30 % of representatives, respectively) (Russinek et al. 2012). Due to this, several endemic species from these dominant taxa of littoral and sublittoral zones were used as a source of new actinobacteria strains. Five species were collected in two replicates and used as the sources for actinobacteria isolation (Table 1).

Table 1 Actinobacteria strains isolated from macroinvertebrates of Lake Baikal
Full size table

Initial screening provided 25 independent isolates based on morphological features (Fig. S1). As can be seen from Table 1, the majority of actinomycetes were obtained from the samples originating from net-spinning caddisfly Trichoptera sp. larvae and amphipods Brandtia sp. Surprisingly, only two strains were isolated from the two sponge’s samples. The previous studies have shown that freshwater sponges are a great reservoir of actinobacteria representing 12–26 % of their microbial community (Costa et al. 2013; Gladkikh et al. 2014; Kaluzhnaya et al. 2012; Kaluzhnaya and Itskovich 2014). However, published reports are based on cultivation independent approaches, rather than on direct isolation of bacteria, thus encountering all microbiome including uncultured species. In order to fully uncover the potential for actinobacteria isolation from sponges, the media composition would have to be adjusted accordingly as well as analysis of samples collected in different seasons, especially during peaks of phytoplankton vegetation.

Phylogenic analysis based on 16S RNA showed that the majority of them belongs to the genus Streptomyces (15 out of 25) (Fig. S2; Table 1). This is typical since this genus is usually dominant in all types of samples. Furthermore, Streptomyces species were previously found to be dominant actinobacteria in water and sediments of Lake Baikal (Terkina et al. 2002). The second most “abundant” group of isolates belonged to genus Nocardia (6 out of 25)—5 isolates out of 6 from amphipods and 1 from caddisfly larvae. Nocardia are typically isolated from soil samples. However, these bacteria could be also found in diverse environments such as air, water, plants, and rotten materials (Faghri et al. 2014). The large portion of Nocardia isolates was reported from the termite nests where they represent 8 % of all actinobacteria (Sujada et al. 2014). Recently, Nocardia soli was isolated from Baikal freshwater sponges by applying a new procedure for bacteria cultivation (Jung et al. 2014). Terkina and co-authors did not observe any Nocardia species in both water and sediment samples from Lake Baikal (Terkina et al. 2002). This makes us believe that the observed enrichment of amphipod’s samples with the Nocardia reflects the specialized rather than random association/interaction between actinobacteria and hosts. Only two isolates were obtained from the sponge B. bacilifera (Table 1). Typical actinobacteria strains associated with the marine sponges belong to the Micromonospora and Salinispora genera (Abdelmohsen et al. 2014; Vicente et al. 2013). However, the only Micromonospora strain obtained in this work originated from the Turbellaria worms while the sponge samples gave Streptomyces and Pseudonocardia species. Three representatives of the less abundant and studied genera of actinomycetes, including Agromyces and Aeromicrobium, were also found to be affiliated with the Baikal invertebrates (Fig. 1, S1; Table 1). Whereas Agromyces species are usually isolated from soil samples and marine sediments (Gledhill and Casida 1969), Aeromicrobium are more abundant and were typically recovered from soil, marine sediments, water, and even from air (Tang et al. 2008). Both strains Aeromicrobium sp. IB2014/016-9 and Agromyces sp. IB2014/011-11 16S rRNA are forming a tight clade with several representatives of the respective genera (Fig. 1). This finding indicates the diversity of species of actinomycetes that could be isolated with the used approach.

Fig. 1
figure 1

Phylogenetic tree of actinobacteria isolates based on the 16S rRNA gene sequences

Full size image

Analysis biological activity of isolated strains

In the agar block diffusion test, six strains were active against both tested Gram-positive B. subtilis and S. carnosus and seven were inhibiting growth of the TolC antibiotic susceptible mutant of E. coli but not E. coli K12 or P. putida (Table 2, S1). Only one strain Streptomyces sp. IB2014/011-6 was able to prevent growth of E. coli K12. Most of the strains showed cross activity against both Gram-positive and E. coli TolC cultures. Two strains (Micromonospora sp. IB2014/08-1 and Streptomyces sp. IB2014/016-6) were active strictly against B. subtilis and S. carnosus, while Streptomyces sp. IB2014/010-2 and Nocardia sp. IB2014/014-5 only inhibited growth of E. coli TolC. Methanol extracts from the MS agar grown strains were also tested for antibacterial activity. However, not in all cases where the agar block test gave positive results the activity was observed in extract (Table 2, S1). This could be explained by the nature of the active substances and their solubility in solvents used in extract preparation. Interestingly, almost all strains in both tests were found to be active against S. cerevisiae (19 out of 25) but not against C. albicans.

Table 2 Antibacterial and antifungal activity of isolated strain when grown on different media
Full size table

Media composition is often a determining factor for production of different secondary metabolites. Two different liquid media (NL19 and SG) were chosen for metabolite production (Kieser et al. 2000). As mentioned above, the metabolites were extracted separately from the cultural fluid and the biomass. Eleven extracts from the biomass and nine from the supernatant of the strains grown in NL19 media were active against Gram-positive bacteria (Table 2, S2). Also, two extracts from biomass of Streptomyces sp. IB2014/016-1 and Streptomyces sp. IB2014/016-6 were active against E. coli TolC, one of which, Streptomyces sp. IB2014/016-6, was able to hinder growth of P. putida and E. coli K12. At the same time, biomass extract of Nocardia sp. IB2014/017-7 was specifically active against P. putida but not any other test cultures tested. The biomass extract of Streptomyces sp. IB2014/016-5 was active against both yeast S. cerevisiae and C. albicans (Table 2, S2). The extracts from cultural fluids of eight strains were inhibiting growth of E. coli TolC, but only three of them (Streptomyces sp. IB2014/010-1, Streptomyces sp. IB2014/011-1, and Streptomyces sp. IB2014/016-6) were also active against E. coli K12, and only one Streptomyces sp. IB2014/010-1 was hitting Pseudomonas growth. One extract cultural liquid of Streptomyces sp. IB2014/016-2 was active against yeasts.

A similar situation was observed in the case of extracts from the strains grown in SG (Table 2, S3). Nine of the biomass extracts and nine of the cultural fluids extracts were active against the Gram-positive test cultures. At the same time, three strains were inhibiting growth of E. coli TolC, two of which also were active against E. coli K12 (Streptomyces sp. IB2014/016-2 and Streptomyces sp. IB2014/016-6). The Streptomyces sp. IB2014/016-6 during growth in both NL19 (biomass extract) and SG media (cultural liquid extract) was found to produce compounds inhibiting P. putida. Three out of four strains capable to inhibit growth of S. cerevisiae were also active against C. albicans (Streptomyces sp. IB2014/01-2, Streptomyces sp. IB2014/016-2, and Streptomyces sp. IB2014/016-5) (Table 2 S3). Streptomyces sp. IB2014/016-6 was active against C. albicans but not S. cerevisiae.

Dereplication of secondary metabolite profiles of biologically active strains

Modern mass spectrometry methods allow performing not only general analysis of secondary metabolites but also combined with the existing databases provide the possibility for identification of individual compounds. Two different LC-MS protocols were used in this study. When low-resolution technique allows estimating the major components of extracts, the high-resolution experiments open a possibility for dereplication of metabolites. The extracts from all strains grown in all tested conditions were subjected to low-resolution LC-MS analysis in order to estimate the number of major compounds that are produced. In all, samples from 3 to 93 peaks on MS and UV chromatograms were observed (Table S4). It is not clear if several peaks in one extract correspond to different individual metabolites or one family of compounds.

The majority of pathogenic Gram-negative bacteria are not controlled by drugs. Thus, strains producing metabolites capable to inhibit growth of E. coli and P. putida in our tests are of particular interest. We conducted complete dereplication of metabolic profiles of Streptomyces sp. IB2014/010-1 and Streptomyces sp. IB2014/016-6 (Table S5, S6). In addition to the characterization on low-resolution mass spectrometry platform, active extracts from these strains were also subjected to high-resolution mass spectrometry analysis. Five compounds out of 103 for strain Streptomyces sp. IB2014/010-1 and 10 out of 90 for strain Streptomyces sp. IB2014/016-6 were preliminary predicted based on characteristics from the Dictionary of Natural Products database using search parameters described in the method section. Only high-molecular detected compounds are given below.

We found that Streptomyces sp. IB2014/010-1 is producing variapeptin, a hexadepsipeptide antibiotic with antibacterial activity of azinothricin family (Table 3; S5; Fig. S2) (Nakagawa et al. 1990a, 1990b). It is produced in small quantities when compared to other major peaks on chromatograms (Fig. S2). Variapeptin was isolated in 1990 from extract of Streptomyces variabilis K2919 and shown to inhibit growth of numerous Gram-positive bacteria, but not Gram-negative or fungus. Later, with the expansion of this family of compounds, they were re-discovered as an anticancer drug of general action (Hale and Cai 1996). Other metabolites produced by Streptomyces sp. IB2014/010-1, including several major peaks, do not show any hits in DNP thus making us believe that the majority of them might be new (Table S5).

Table 3 Compounds identified with DNP from extracts of isolated strain active against Gram-negative bacteria and their features
Full size table

Streptomyces sp. IB2014/016-6 was found to produce three compounds of minalemine family: minalemine A (CRC number: DYN62-X), minalemine B (CRC number: DYN64-Z), and minalemine C (CRC number: DYN66-B) (Table 3; S6; Fig. S3) (Exposito et al. 1998). The detected molecular mass and absorption spectra of the identified metabolites correspond to the described features of minalemines (Exposito et al. 1998, 2001; Whittle et al. 2003). Minalemines are linear peptide secondary metabolites, which were previously identified in extracts from marine tunicates (Fig. S3) (Exposito et al. 1998). No biological activity was described for these metabolites. The core part of minalemines consist of l-leucine and an unusual amino diacid, 3-(N-carboxymethyl)-aminodecanoic acid (Ncma), that is rarely found in secondary metabolites (Duncan et al. 2002; Jin et al. 2010). The core of minalemines is connected with an aminogroup of two different ω-aminoguanidines: agmatine and homoagmatine. The individual minalemines differ in the chain length of the Ncma (10, 11, or 12 carbons A, B, and C, respectively) and in the presence or absence of a sulfamic acid group (D, E, F). Although minalemines were originally isolated from the marine tunicates, we observed presence of three out of six of these compounds (minalemines A-C) in the extract of Streptomyces sp. IB2014/016-6. This makes us believe that minalemines are not direct products of ascidian Didemnun rodriguesi but rather are synthesized by some actinomycetes associated with this organism. We failed to find minalemines D-F in extracts of Streptomyces sp. IB2014/016-6. Most probably, modification leading to conversion of minalemines A-C to D-F (addition of sulfamic acid) is indeed conducted by the D. rodriguesi representing an interesting example of interplay between secondary metabolism of actinomycetes and metabolic pathways/activity of its host.

In conclusion, we have isolated 25 actinobacteria strains from the endemic invertebrate species of Lake Baikal proving this ecosystem to be a promising source of new actinomycetes. Among them, representatives of genus Streptomyces and Nocardia were dominant taxa, depending on a source of isolation. As mentioned above, the high number of Nocardia sp. strains were found in amphipoda Brandtia sp. Several rare and less studied actinobacteria strains were isolated as well, including new representatives of Aeromicrobium and Agromyces genera. A large fraction of the isolated strains was able to inhibit growth of Gram-positive bacteria and budding yeast. At the same time, several strains were found to be capable to prevent the growth of Gram-negative test cultures and pathogenic C. albicans. These strains for sure represent a particular interest for deeper investigation with complete phylogenetic classification and identification and characterization of active compounds. In general, even after the initial estimation, we can state that the majority of secondary metabolites accumulated by the isolated strains are not described in available antibiotic databases. This fact proves that the chosen strategy for isolation of new actinobacteria from the unique ecological niches could be successful in order to obtain new biologically active compounds.