Antibacterial potential of Actinobacteria from a Limestone Mining Site in Meghalaya, India

This work attempts to assess the antimicrobial potential of actinobacteria isolated from limestone mining sites which hitherto, is an under-explored niche for exploring novel bioactive metabolites. Actinobacteria were selectively isolated from Mawsmai, Meghalaya, India, a limestone mining area, using different pretreatment methods. Forty-seven isolates were obtained, which were identified based on their morphological, biochemical and chemotaxonomical characteristics. Streptomyces was the dominant cultivable genera which constituted 76% of the isolates cultivated. All the isolates were screened for antimicrobial activity against three Gram-negative viz. Escherichia coli, Pseudomonas aeruginosa and Klebsiella pneumoniae, and three Gram-positive bacteria viz. Staphylococcus aureus, Bacillus subtilis and Micrococcus luteus and besides, two candidal species viz. Candida albicans and C. tropicalis. 19% of the total isolates showed antibacterial activity against at least one of the test bacterial strains used. The identity of the four bioactive isolates viz. LD-21, LD-29, LD-34 and LD-39 was confirmed as Streptomyces sp. on the basis of their 16S rDNA sequence and 16S rRNA secondary structure analysis. These isolates showed antibacterial activity against at least two Gram-positive bacteria and all the four harbored at least one of the three biosynthetic gene clusters viz. type-I and type-II polyketide synthases and non-ribosomal peptide synthetase which are related to synthesis of bioactive metabolites.

are potential antibiotics producers [8][9][10][11][12][13][14] . The state of Meghalaya located in the north-eastern part of India falls under the Indo-Burma Biodiversity Hotspot 15 . Therefore, mining for antibiotics producing actinobacteria from under-explored ecological niche of limestone mines of diversityrich state of Meghalaya enhances the possibility of discovery of novel metabolites. This paper, therefore, reports on diversity of actinobacteria having antibacterial activities from limestone mining areas located at Mawsmai, Meghalaya, India.

MATERIALS AND METHODS Sample collection
The sampling site was from a limestone mining area at Mawsmai near Sohra in Meghalaya, India (N25 o 15.364' E91 o 43.885'). Soil samples from around the mining area were aseptically collected in sterilized polythene bags and brought to the lab and stored at 4±1 o C until further use.

Determination of soil physico-chemical characteristics
Soil and air temperatures (in o C) were measured in three different locations using a soil thermometer and an ordinary thermometer respectively. Moisture content (%) of the soil was determined gravimetrically 16 . pH was determined using a digital pH meter 17 . Organic carbon was determined following the method of Anderson and Ingram (1993) 17 . Total nitrogen was determined by the Kjeldahl Method 18 . Available Phosphorus was determined by the molybdate blue method 16 . Potassium was determined by the flame photometer method 18 .

Soil pretreatment and isolation of actinobacteria
The collected soil samples were air-dried at room temperature for one week and mixed equally into one composite sample. The air-dried samples were then pretreated by (i) dry heating at 120 o C for 1 hour, (ii) wet heating at 70 o C for 15 minutes in a water bath, (iii) 1.5% phenol at 25±1 o C for 30 minutes, (iv) 0.2% humic acid at 25±1 o C for 30 minutes and (v) combination of i & iii. The unpretreated air-dried soil samples were also used as inoculum for isolation. 1g pretreated and unpretreated soil samples were suspended in 9ml sterilized distilled water and serially diluted. All the undiluted and serially diluted soil sample suspensions were used as inocula for isolation.
Five selective media were used for isolation viz. Actinomycete Isolation Agar (AIA), Bennett's Agar (BA), Starch Casein Agar (SCA), Streptomyces Agar (SA) and Humic acid Vitamin Agar (HVA) 19 . The above media were amended with Nystatin (50µg ml -1 ) and Rifampicin (20µg ml -1 ). 100µl of the soil sample suspensions were inoculated on the above selective media and incubated at 28±1 o C for up to four weeks. Based on colony morphology, colonies were picked from the culture plates, sub-cultured and maintained in Bennett's agar medium. The recovery of actinobacterial colonies from different pretreated samples and media were recorded as actinobacterial population number in cfu g -1 soil. All observations were recorded in triplicates and expressed as sample mean ± standard error. Multiple comparisons of sample means were done using ANOVA with Duncan's test at p ≤ 0.05. Statistical analysis was performed with XLSTAT software.

Morphological characterization
The selected colonies were grown in ISP3 (oat meal agar) medium by coverslip culture method 20 and incubated at 28±1 o C for 7 days. The coverslip cultures were then placed on a glass slide and observed under the microscope (Olympus CX21i) at 400X magnification for presence of aerial and substrate mycelia, spore chains and other reproductive structures.

B i o c h e m i c a l a n d c h e m o t a x o n o m i c characterization
Biochemical tests viz., casein, xanthine, hypoxanthine and tyrosine hydrolysis were performed following the methods of Berd (1973) 21 . Chemotaxonomic analyses viz. cell wall chemotype of 2,6-diaminopimelic acid (DAP) and whole cell sugars were performed by thinlayer chromatography following the methods of Hasegawa et al. (1983) 22 . The scheme for preliminary identification of actinobacteria was followed according to Berd (1973) 21 and Staneck and Roberts (1974) 23 .

Phylogenetic analysis of 16S rDNA sequence
Genomic DNA of the selected bioactive isolates were extracted by enzymatic method 24 followed by PCR of 16S rDNA. The primers used for amplification were 27F (5'-AGAGTTTGA-TCMTGGCTCAG-3') and A3R (5'-CCAGCCCCACCTT-CGAC-3') 25 . PCR reactions were carried out in an Eco96 TM thermal cycler (Himedia, India) in 25µl reaction mixture containing PCR buffer (Promega) with 1.5mM MgCl 2 , 0.2mM of each dNTP (Promega), 0.5µM of each primer (IDT), 0.625U of Taq DNA polymerase (Promega) and 5µl of extracted DNA as template DNA. Amplification parameter was an initial denaturation of 95 o C for 5 minutes, 35 cycles of 95 o C for 45 seconds, 56 o C for 60 seconds, 72 o C for 90 seconds and final extension of 72 o C for 7 minutes. The amplified products were analyzed by agarose gel electrophoresis (with 1.2% w/v agarose added with 1µl/10ml v/v LabSafe TM Nucleic Acid Stain, G-Biosciences). A 100bp DNA ladder was used for size reference. The amplified products were purified and sequenced at Xcelris Labs Ltd., Ahmedabad, India. The partial 16S rDNA sequences obtained were matched with type strains from EzBioCloud database 26 for determining sequence similarities. Phylogenetic relationships of the selected bioactive isolates along with thirty-two related type strains were determined using MEGA version 7 27 and phylogenetic tree was constructed using the Maximum Likelihood method based on the Tamura 3-parameter model 28 . A discrete Gamma distribution was used to model evolutionary rate differences among sites (+G). The rate variation model allowed for some sites to be evolutionarily invariable (+I). Bootstrap analysis 29 was done with 1000 replicates. The partial 16S rDNA sequences were deposited in GenBank. Phylogenetic analysis of predicted 16S rRNA secondary structure The use of 16S rRNA secondary structures for phylogenetic analysis of actinobacteria was understood from Ghosh et al. (2019) 30 . The 16S rDNA sequences obtained were converted into 16S rRNA sequences and subsequently the 16S rRNA secondary structures with their minimum free energy (MFE) were predicted using the RNAfold server 31 at the ViennaRNA Web Services (http://rna.tbi.univie.ac.at/). The obtained 16S rRNA secondary structures of the selected bioactive isolates along with their MFEs were then compared with the related type strains from EzBioCloud database (which were used for phylogenetic analysis based on 16S rDNA sequence) to confirm their identity. Consensus 16S rRNA secondary structures were also generated using the RNAalifold server 32 at the ViennaRNA Web Services between the bioactive isolates and the related type strains with the closest MFEs to analyze for conserved regions. The selected 16SrDNA sequences with the closest MFEs were aligned using ClustalX 2.1 (http://www.clustal.org/ download/2.1/) and sequence overhangs trimmed before using the aligned file for generating consensus 16S rRNA secondary structure.

Screening of isolates for antimicrobial activity
Isolates were screened for antimicrobial activity against  Bennett's broth supplemented with 0.05% humic acid, pH 7.5 33 in 15ml tubes and incubated at 28±2 o C for 15 days. The broth cultures were then filtered using an ordinary filter paper followed by sterilization using membrane filters (PVDF 0.2µm). Uninoculated media were used as negativecontrols while amikacin (30µg), vancomycin (30µg) and amphotericin-B (20µg) discs were used as positive-controls. Antimicrobial screening was performed by agar-well diffusion method 34 . Antimicrobial activity was assessed as zone of inhibition (mm) by subtracting the diameter of the agar-well from the diameter of inhibition. All the culture filtrates were prepared and tested in triplicates.

Detection of biosynthetic gene clusters
PCR was performed on the genomic DNA of the selected bioactive isolates for checking the presence of three biosynthetic gene clusters (BGCs) related to synthesis of bioactive metabolites viz. type I polyketide synthase (PKS-I), type II polyketide synthase (PKS-II) and Non-ribosomal peptide synthetase (NRPS) 8,35 . Three pairs of primers viz. K1F

Soil physico-chemical characteristics
Soil and air temperatures of the study site at the time of sample collection (July) were 21 o C and 19 o C respectively akin to a sub-tropical climate as Mawsmai, Meghalaya, India falls under such climatic zone 36 . Moisture content of the soil was 16.64±3.85%. The average soil pH was slightly alkaline (7.12±0.9). Soil organic carbon was 2.46±0.05%. Total soil nitrogen, available phosphorus and potassium were 0.08±0.01%, 3.39±0.2µg g -1 and 1.35±0.65µg g -1 respectively. Moisture content of the soil was higher than that reported by Lamare and Singh (2017) 37 who conducted study on changes in soil quality in a limestone mining area in East Jaintia Hills from the same state of Meghalaya, India. They reported soil moisture content which ranged from 5.55% to 6.86%. The soil moisture content from our study site was higher because samples were collected on a rainy day. Soil moisture content from limestone mining sites is usually low due to presence of a mixture of granular, aggregated and blocky soil structure lowering the water holding capacity and also as a result of low organic matter 37 . Lamare and Singh (2017) also reported slightly alkaline nature of limestone soil 37 and such soils are generally characterized by alkaline pH 38 . Soil organic carbon from our study site was also higher than that reported by Lamare and Singh (2017) who recorded 0.23% to 0.47% of soil organic carbon from intensively mined areas and the reason was probably our soil samples was a composite of samples collected from intensively mined and sparsely mined areas. Total soil nitrogen, though slightly higher than that reported by Lamare and    Singh (2017) 37 , was low due to loss of vegetation cover due to mining activities accompanied with loss of nitrogen-fixing microorganisms 37 . Available phosphorus and potassium were comparatively much lesser to that reported by Lamare and Singh (2017) 37 . Other workers have also reported low levels of phosphorus and potassium from limestone mining areas 39,40 . Lamare and Singh (2017) however, reported no significant changes in phosphorus and potassium levels of limestonemined and unmined areas 37 . Generally, soil organic carbon, nitrogen, available phosphorus and potassium are relatively lower in concentration in limestone mining areas 41 .

Isolation of actinobacteria
Highest actinobacterial population number was recorded from un-pretreated and humic acid pretreated soil samples i.e. respectively 5.6±2.13x10 2 and 4.0±1.56x10 2 cfu g -1 soil. Statistically insignificant difference was recorded in population number of these two methods at p<0.05 (Fig.1). The differences in population numbers between wet heat pretreated and humic acid pretreated samples were also statistically insignificant at p<0.05 (Fig.1). The three pretreatment methods were significantly better than the other pre-treatment methods. Minimum number of Actinobacteria was recovered using dry heating and phenol pretreatment methods. This may be because most actinobacteria are heat killed when samples are exposed to 100 o C for a duration of one hour 42 and phenol has been reported to reduce the growth of actinobacteria 43 . Among the selective media used, humic acid vitamin agar (HVA) and Streptomyces agar (SA) statistically gave the highest actinobacterial population number at 3.78±1.95x10 2 and 3.11±1.88x10 2 cfu g -1 soil respectively and were significantly better than the other three media used at p<0.05 (Fig.1). Humic acid present in HVA serves the sole carbon and nitrogen sources and has been observed to activate spore germination in many actinobacterial species 19 . SA on the other hand is one of the routinely used media for selective isolation and maintenance of actinobacteria. The un-pretreated samples recorded highest actinobacterial population when inoculated on HVA (10.67±0.88x10 2 cfu g -1 soil) and SA media (10.00±0.58x10 2 cfu g -1 soil). The difference was, however, statistically insignificant at p<0.05 (Fig.1). The actinobacterial population number recorded from humic acid pretreated sample inoculated on HVA medium (9.00±0.58x10 2 cfu g -1 soil) was also not significantly different from the unpretreated sample inoculated on SA medium (10.00±0.58x10 2 cfu g -1 soil).

Diversity of actinobacteria
On the basis of micromorphological characteristics, forty-seven isolates were identified as actinobacteria and all of them showed Gram-positive reaction. Based on biochemical and chemotaxonomic characteristics 21,23 , thirty-six isolates belonged to Streptomyces genera and the remaining eleven isolates were found to be non-streptomycetes, out of which nine were identified as Nocardia, one as Actinomadura and one as Rhodococcus (Table 1). From the cultivated isolates, it showed a dominance of Streptomyces genera which comprised 76% of the total number of isolates. Recent reports on actinobacteria from limestone niches have also shown Streptomyces to be the dominant genera 8,13,44,45 . A number of novel actinobacterial strains have also been reported from other limestone niches. Nimaichand et al. (2012) 46 reported Streptomyces manipurensis sp. nov., a  49 reported Geodermatophilus sabuli sp. nov., a g-radiation-resistant actinobacterium isolated Table 3. Minimum free energy (MFE) in kcal mol -1 of the predicted 16S rRNA secondary structures of the bioactive isolates along with thirty-two related type strains from EzBioCloud database. Five outgroups have also been included for comparison. The MFEs have been arranged in ascending order  50 reported Streptomyces boninensis sp. nov., isolated from the soil of a limestone cave in the Ogasawara Islands of Japan. Dominance of Streptomyces from the cultivated isolates of this work signifies the adaptation of this genus to thrive on a wide range of nutrients including the synthetic media used during isolation and the genus can also grow faster than the other slower-growing actinobacterial genera thereby establishing its micro niche much earlier than the others 9 . Antimicrobial activity Antimicrobial screening results from the agar well diffusion assay revealed that 19% of the isolates showed antibacterial activity against at least one of the bacterial strains tested. The screening results also showed stronger antibacterial activity against Gram-positive than Gram-negative bacteria. The isolates, however, did not show anti-candidal activity. Other recent reports on antimicrobial activity of actinobacteria from limestone niches also showed stronger anti-Gram-positive than anti-Gram-negative bacteria. Nimaichand et al. (2015) 8 reported on antimicrobial activity of actinobacteria from limestone deposit sites in Hundung, Manipur, India where eighteen actinobacterial isolates showed antibacterial activity against Gram-positive Bacillus subtilis and five isolates showed antibacterial activity against Gram-negative Escherichia coli. They however, also reported on isolates with anti-fungal activity including anti-candidal activity. Maciejewska et al. (2016) 9 also reported on antibacterial activity of moonmilk actinobacterial isolates from a limestone cave in Belgium where the isolates displayed stronger antibacterial activity against Gram-positive bacteria (94% of the phylotypes) than Gram-negative bacteria (71% of the phylotypes). They also reported on anti-fungal activity of the moonmilk isolates where 94% showed activity against a range of fungal strains including Candida albicans. Four isolates from this study viz. LD-21, LD-29, LD-34 and LD-39 showed clear antibacterial activity against at least two Gram-positive bacteria (Fig.2). Isolate LD-21 showed the highest antibacterial activity against Staphylococcus aureus (26±0.58 mm). Antibacterial activity of culture filtrate of LD-21 is shown in Fig.3 2.33±0.88 mm respectively. All the above four bioactive isolates also showed antibacterial activity against Gram-negative bacterium Pseudomonas aeruginosa but the zone of inhibition of their culture filtrates on the agar well diffusion assay were not as apparent as the positive control (Amikacin disc 30µg) (data not shown) and there were no antibacterial activity against the other Gram-negative bacteria tested. The above four bioactive isolates also possessed at least one of the three biosynthetic gene clusters (BGCs) viz. PKS-I, PKS-II and NRPS (Table 2). Based on the antibacterial activity profile and BGCs, isolates LD-29, LD-34 and LD-39 seemed to have similar metabolic pathways and metabolite clusters. The antimicrobial screening results from this study recovered bioactive isolates with tendency of antimicrobial activity only towards Gram-positive bacteria and failed to recover isolates with broadspectrum antimicrobial activity. Yet this find is still significant considering the present era where more emphasis are being undertaken to develop narrower spectrum antimicrobial drugs (due to emergence of antimicrobial drug resistance) and recent advancements in rapid and more accurate diagnostic techniques 5 . It would therefore be interesting to test the above bioactive isolates against more Gram-positive indicator strains and multi-drug resistant clinical isolates. Optimization of the physico-chemical conditions for growth of the bioactive isolates for maximum production of bioactive metabolites 51 is currently under way in the hope of finding novel bioactive metabolites.

Identification of the bioactive isolates
Based on morphological, biochemical and chemotaxonomic profiles (Table 1), all the four bioactive isolates were found to belong to Streptomyces genera. Phylogenetic analysis of their 16S rDNA sequences also confirmed to be Streptomyces. Isolate LD-21 was closest to Streptomyces niveus NRRL2466 T (DQ442532) with 99.47% sequence similarity and the other three isolates LD-29, LD-34 and LD-39 were closest to Streptomyces camponoticapitis 2H-TWYE14 T (KP784807) with sequence similarity of 99.26%, 99.41% and 99.48% respectively. The generated phylogenetic tree of the bioactive isolates along with the related type strains is shown in Fig.4. GenBank accession numbers of the partial 16S rDNA sequences of the bioactive isolates LD-21, LD-29, LD-34 and LD-39 were assigned from MK659890 to MK659893 respectively. Therefore, the bioactive isolates are now identified as Streptomyces LD-21 (MK659890), Streptomyces LD-29 (MK659891), Streptomyces LD-34 (MK659892) and Streptomyces LD-39 (MK659893). The minimum free energy (MFE) in kcal mol -1 of the predicted 16S rRNA secondary structures of the bioactive isolates also placed them under Streptomyces genera (Table 3). However, on the basis of MFEs, the Streptomyces species with the closest match have changed. The consensus 16S rRNA secondary structures of the bioactive isolates and their closely-related type strains (based on MFE) also showed high degree of conserved regions (Fig.5) confirming that the bioactive isolates belonged to Streptomyces genera. It was also noted that the consensus 16S rRNA secondary structure of Streptomyces LD-21 looked different from the other three bioactive isolates Streptomyces LD-29, Streptomyces LD-34 and Streptomyces LD-39 while these three were similar in their consensus 16S rRNA secondary structures (Fig. 5) which agrees with the maximum likelihood tree (Fig. 4). Thus, it could be inferred that Streptomyces LD-29, Streptomyces LD-34 and Streptomyces LD-39 were phylogenetically closer to one another and they could even belong to the same strain and probably share very similar metabolic pathways and metabolite clusters as already mentioned which is evident from the antibacterial activity of their culture filtrates (Fig.2) and their BGC profiles (Table 2).

CONClUsiON
In present work, actinobacteria were isolated from a limestone mining site. The soil environment at the sample collection site was having minimal nutritional conditions (oligotrophic). Based on morphological, biochemical and chemotaxonomic characteristics, Streptomyces was found to be the dominant cultivable actinobacterium isolated from this environment. This, however, also suggests the use of more specific methods of isolation to recover the rarer genera of actinobacteria. Four bioactive isolates all identified as Streptomyces showed considerable antibacterial activity against Grampositive bacteria and they were also found to harbor at least one of the three biosynthetic gene clusters viz. PKS-I, PKS-II and NRPS which are related to synthesis of bioactive metabolites. More work is needed to assess their activity over a range of the Gram-positive bacterium including multi-drug resistant strains. Optimization of bioactive metabolites production by the bioactive isolates is under way. This under-explored niche from Meghalaya is important for mining of actinobacteria for obtaining novel bioactive metabolites that might have applications in various fields.