Metabolomics-Based Chemotaxonomic Classification of Streptomyces spp. and Its Correlation with Antibacterial Activity

Abstract

Secondary metabolite-based chemotaxonomic classification of Streptomyces (8 species, 14 strains) was performed using ultraperformance liquid chromatography-quadrupole-time-offlight-mass spectrometry with multivariate statistical analysis. Most strains were generally well separated by grouping under each species. In particular, S. rimosus was discriminated from the remaining sevens pecies (S. coelicolor, S. griseus, S. indigoferus, S. peucetius, S. rubrolavendulae, S. scabiei, and S. virginiae) in partial least squares discriminant analysis, and oxytetracycline and rimocidin were identified as S. rimosus-specific metabolites. S. rimosus also showed high antibacterial activity against Xanthomonas oryzae pv. oryzae, the pathogen responsible for rice bacterial blight. This study demonstrated that metabolite-based chemotaxonomic classification is an effective tool for distinguishing Streptomyces spp. and for determining their species-specific metabolites.

Introduction

Streptomyces is a group of gram-positive filamentous bacteria that grow in soil and decaying vegetation. It is the largest genus of Actinobacteria and includes over 500 species [61]. Streptomycetes have historically been examined for their ability to biologically control plant disease and animal pathogens [22,23, 41]. Among various species of Streptomycetes, S. somaliensis and S. sudanensis cause actinomycosis in animals [40, 49], and S. scabiei causes common scab in potatoes [13, 35]. Streptomyces is a particularly well-known agent against leaf blight caused by Xanthomonas oryzae pv. oryzae. The genus Xanthomonas, comprising gram-negative plant pathogenic bacteria, causes a variety of diseases in plants such as citrus canker, walnut blight, rice leaf blight, and bacterial spot [20, 33, 38, 55]. Secondary metabolites of Streptomyces are associated with the production of antibiotics and toxic substances in animals and plants. For example, antibiotics such as alnumycin, fistupyrone, and vinylamycin have been obtained from Streptomyces spp. [6, 30, 31]. S. coelicolor and S. hygroscopicus are known to produce antibiotic metabolites such as perimycin and herbimycin [8, 43]. In addition, particular species of Streptomyces such as S. scabiei, S. ipomoeae, and S. somaliensis are harmful to humans and animals. Therefore, these species are crucial to clearly classify and differentiate the various species. Traditional taxonomic procedures based on phenotypic traits and DNA relatedness involve classification of new microorganisms mostly using DNA sequencing [27, 37]. Bacterial identification has been generally based on their morphological, physiological, and chemical characteristics [59]. However, solely using one taxonomic method is not sufficient for bacterial classification. Classification of bacterial species in the genus Streptomyces has typically been supported by other types of investigation, such as chemical, enzymatic, serological, and hybridization studies, and RNA/DNA sequence analysis [12,57]. Despite the wide variety of investigations, taxonomic approaches to identifying Streptomyces species remain difficult because of the various morphological, cultural, physiological, and biochemical characteristics that are observed at the interand intraspecies levels [1]. Recently, several studies using a chromatographic approach to classify bacteria have been reported [21, 56, 25]. The classification of microorganisms based on metabolite profiling has emerged as a useful chemotaxonomic tool that provides detailed information about the differences and similarities between species [7, 32, 51, 58]. In addition, liquid chromatography mass spectrometry (LC-MS) has been used to identify active antimicrobial substances in target microorganisms [10, 17].

In this study, the secondary metabolites of Streptomyces were profiled to classify Streptomyces strains based on cluster-specific properties by using ultraperformance liquid chromatography-quadrupole-time-of-flight-mass spectrometry (UPLC-Q-TOF-MS). We also evaluated the growth inhibition of Xanthomonas oryzae pv. oryzae by Streptomyces species extracts and tentatively identified antibacterial metabolites.

 

Materials and Methods

Chemicals and Reagents

Methanol, water, and acetonitrile were purchased from Fisher Scientific (Pittsburgh, PA, USA). Yeast extract, malt extract, and nutrient broth (NB) were purchased from Becton Dickinson (Franklin Lakes, NJ, USA). Agar powder, calcium carbonate, and glucose were purchased from Junsei Chemical (Tokyo, Japan). Glucose/soybean meal/sodium chloride (GSS) broth was purchased from MB Cell (Seoul, Korea). 2,4-Diacetylphloroglucinol (DAPG) was purchased from Toronto Research Chemicals (Ontario, Canada) as a positive control. Formic acid and oxytetracycline hydrochloride and standard compounds were obtained from Sigma-Aldrich. (St. Louis, MO, USA)

Strains

The Streptomyces strains used in this study consisted of 8 species and 14 strains, which are S. coelicolor (KACC 20100), S. griseus (KACC 20084 and KACC 20731), S. indigoferus (MJM 8645), S. peucetius (KCTC 9038 and KCTC 9199), S. rimosus (KACC 20082 and KACC 21078), S. rubrolavendulae (MJM 4426), S. scabiei (KACC 20135, KACC 20200 and KACC 20227), and S. virginiae (KCTC 1747 and KACC 14680).These strains were obtained from the Korean Agricultural Culture Collection (KACC, Korea) and the Korean Collection for Type Cultures (KCTC, Korea). S. indigoferus MJM8645 and S. rubrolavendulae MJM4426, which were isolated from soil and identified by 16S rRNA sequence alignment, were obtained from Bioscience and Bioinformatics, Myongji University, Korea.

Culture Condition

All Streptomyces strains were cultured on yeast malt agar (YMA; 4 g yeast extract, 10 g malt extract, 4 g glucose, and 20 g agar per liter) plates for 3 days at 28℃. One piece of agar (6 mm) containing Streptomyces was transferred into 50 ml of yeast malt broth (YMB) in a 250 ml Erlenmeyer flask and maintained at 28℃ for 3 days with shaking. A 4 ml aliquot of this culture was added to 80 ml of GSS broth in a 250 ml baffled flask and incubated for 10 days at 28℃ for high production of metabolites.

For the antibacterial assay, Xanthomonas oryzae pv. oryzae (Xoo) was used. Optimal growth of Xoo is reportedly on yeast extract glucose carbonate (YGC) agar containing 5 g yeast extract, 10 g glucose, 30 g calcium carbonate, and 15 g agar per liter. Xoo was grown on YGC agar at 28℃ for 2 days and identified on the basis of its morphology. One bacterial colony was transferred into 5 ml of NB in a 13 ml tube and further incubated at 28℃ with shaking at 200 rpm for 24 h.

Extraction Conditions

The culture broth of Streptomyces was centrifuged at 4℃ and 5,000 rpm for 10 min, and the supernatant removed. The remaining pellets were washed 3 times with 20 ml of phosphate-buffered saline. Twenty milliliters of methanol was added to the washed pellets, which were then sonicated for 20 min and extracted for 9 h at 28℃ and 200 rpm in a shaker incubator. The extract solution was evaporated using a speed-vacuum machine (Biotron, Seoul, Korea)

UPLC-Q-TOF-MS Analysis

A UPLC Acquity System (Waters, Milford, MA, USA) equipped with a binary solvent delivery system, a UV detector, and an autosampler was combined with a Waters Q-TOF Premier MS system (Micromass MS Technologies, Manchester, UK). An Acquity UPLC BEH C18 column (100 × 2.1 mm i.d. × 1.7 μm particle size; Waters) was used and the column oven temperature was set at 37°C. The mobile phase consisted of water (A) and acetonitrile (B) with 0.1% formic acid (v/v) at a flow rate of 0.3 ml/min. The initial condition was 5% solvent B for 1 min followed by a linear gradient over 10 min, ending at 100% solvent B. The full-scan mass spectral range was 100–1,000 m/z. MS was performed with the following conditions: ion source temperature at 100℃; desolvation gas flow at 300 l/h; cone gas flow at 0 l/h; cone voltage at 40 V; and capillary voltage at 2.5 V.

LC-ESI-MS/MS Analysis

Liquid chromatography-electrospray ionization-tandem mass spectrometry (LC-ESI-MS/MS) was performed using a Varian 500-MS ion trap mass spectrometer (Varian, Palo Alto, CA, USA), which consisted of an LC binary pump (Varian 212), a photodiode array detector (Prostar 335), and an autosampler (Prostar 410). The LC system was equipped with a Varian PurSuit XRs C18 column (100 × 2.1 mm i.d. × 3 μm particle size). Mobile phases consisted of water (A) and acetonitrile (B) with 0.1% formic acid (v/v). The initial mobile phase consisted of 90% A and 10% B, which was maintained for 2 min, followed by an increase to 90% B over 25 min, which was maintained at 90% B for 5 min, and then rapidly decreased to 10% B, which was maintained for 5 min. The flow rate was set to 0.2 ml/min and 10 μl of sample was injected. The full-scan mass spectral range was 100–1,000 m/z. The running parameters for analyzing the samples were as follows: spray needle voltage, 5 kV; capillary voltage, 80 V; drying temperature, 300°C; drying gas (nitrogen) pressure, 20 psi; nebulizer gas (air) pressure, 40 psi. Tandem mass spectrometry analysis was carried out using scan-type turbo data-dependent scanning (DDS) under the same conditions. LC-ESI-MS/MS data were analyzed using the MS workstation software (ver. 6.9; Varian, USA).

Data Processing and Multivariate Statistical Analysis

The UPLC-Q-TOF-MS data were acquired with MassLynx software (ver. 4.1; Waters), and raw files were converted to a NetCDF file (*.cdf). After conversion, peak extraction, retentiontime correction, and alignment were performed using the metAlign software package (http://www.metalign.nl). After analysis, the resulting data file (*.csv) was then transferred to a Microsoft Excel data sheet for sequential multivariate analysis. Multivariate statistical analysis was performed using the SIMCA-P+ ver. 12.0 (Umetrics, Umea, Sweden). Principal component analysis, partial least squares discriminant analysis (PLS-DA), and hierarchical clustering analysis (HCA) were employed to determine the distribution of Streptomyces strains. Numerical value of R2X and Q2X showed statistical significance and explained variation in the PLS model. R2X was a fraction of the Sum of Squares (SS) of the entire X explained by the current component. Variables grouped by PLS-DA and HCA, which are significantly different metabolites separating Streptomyces strains, was determined on the basis of the variable importance of the projection (VIP) value (VIP > 0.7) and p-value (p < 0.05). The p-values of different metabolite-based cluster groups were determined using STATISTICA (ver. 7.0; StatSoft, Tulsa, OK, USA).

Phylogenetic Analysis

The 16S ribosomal RNA (rRNA) sequences of 14 strains of Streptomyces spp. were obtained from the GenBank database (http://www.ncbi.nlm.nih.gov/genbank/). The 16S rRNA sequences were aligned and analyzed by MEGA ver. 4.0 (The Biodesign Institute, Tempe, AZ, USA), and a phylogenetic tree for the datasets was inferred from the neighbor-joining method.

Antibacterial Activity and Isolation of Active Compound

The modified broth microdilution method [5,26] was used to screen Streptomyces for antibacterial activity. The optical density (OD) value of Xoo culture medium was adjusted to 0.5 at UV 600 nm by a GENESYS 6 UV-Vis Spectrophotometer (Thermo Scientific, Waltham, MA, USA) and diluted 1:20 (v/v) in NB medium, and then the diluted inoculum (190 μl) was seeded in each well of a 96-well plate. For screening of antibacterial activity, each well was treated with 10 μl of crude Streptomyces extract (1,000 μg/ml). The plates were incubated at 28℃ for 24 h, and the OD of the bacteria was measured at 570 nm using an EL808 Microplate Reader (BioTek, Winooski, VT, USA).

The MIC (minimum inhibitory concentration) values of S. rimosus (KACC 20082 and KACC 21078), which showed high antibacterial activity, were measured to confirm the antibacterial activity compared with the positive control, DAPG [19,44]. S. rimosus extract was serially diluted 2-fold to obtain a final sample concentration ranging from 2 to 500 μg/ml, and incubated for 24 h at 28℃; the OD was measured before and after incubation. The MIC of oxytetracycline, which was tentatively considered an active compound, was also determined. The MIC was defined as the lowest concentration of test sample resulting in 90% growth inhibition. All experiments were carried out in triplicate. The MIC value was calculated using the following formula:

where C24 = control (not treated), incubated for 24 h; C0 = control (not treated) at 0 h; T24 = treated, incubated for 24 h; and T0 = treated at 0 h.

We performed preparative high-performance liquid chromatography (prep-HPLC) with a YMC-Pack Pro C18 reversed-phase column (250 × 4.6 mm i.d. × 5 μm particle size) to identify active compounds. All fractions were assayed for antibacterial activity by MIC and active fractions were analyzed by UPLC-Q-TOF-MS and LC-ESIMS/MS. Putative active compounds were identified on the basis of comparative analysis of the mass spectrum, molecular weight, molecular formula, and UV absorbance data, relative to those of standard compounds.

 

Results

Comparison of Metabolite-Based Taxonomy and Phylogenetic Taxonomy Studies of Streptomyces

16S rRNA sequences of the 8 Streptomyces species (14 strains) were determined and compared by MEGA4.0 software (Fig. 1A). The phylogenetic tree obtained from the 14 Streptomyces strains was mainly divided into two characteristic branches. The first branch consisted of S. rimosus, S. coelicolor, and S. rubrolavendulae (bootstrap value 97%), and they were divided into two groups: S. rimosus group, bootstrap value 100%; and S. coelicolor and S. rubrolavendulae group, bootstrap value 87%. The second branch included S. indigoferus, S. scabiei, S. griseus, S. peucetius, and S. virginiae. They were divided in two parts, S. indigoferus and S. scabiei (bootstrap value 84%), and S. griseus, S. peucetius, and S. virginiae (bootstrap value 90%). Each part was again further subdivided. In general, most strains were well separated by grouping under each species.

Fig. 1.Taxonomy studies of the 14 Streptomyces strains. (A) Neighbor-joining phylogenetic tree based on 16S rRNA gene sequences. The numbers at the branching points are the percentages of occurrence in 1,000 bootstrapped trees. The bar indicates a distance of 0.005 substitutions per site. (B) Metabolite-based hierarchical clustering analysis (HCA) of Streptomyces strains analyzed by UPLC-Q-TOF-MS.

The HCA dendrogram constructed by secondary metabolites analyzed by UPLC-Q-TOF-MS is shown in Fig. 1B. The dendrogram was divided into two branches: S. rimosus and other seven species. Other species were divided into two characteristic branches: S. virginiae and other species (S. indigoferus, S. coelicolor, S. griseus, S. scabiei, S. peucetius, and S. rubrolavendulae). These six species were again divided into two sub-branches. The first sub-branch includes S. indigoferus, S. coelicolor, S. griseus, and S. scabiei. The second sub-branch includes S. peucetius and S. rubrolavendulae. Both dendrograms were typically well separated for each species.

Secondary Metabolite-Based Separation of Streptomyces and Identification of Species-Specific Metabolites

The metabolites of the eight Streptomyces species (14 strains) were analyzed by UPLC-Q-TOF-MS, and each species was separated by PLS-DA (Fig. 2). The PLS1 and PLS3 scores were found to be 11.0% and 7.2% of the total variation, respectively, as observed in the PLS-DA model (R2X = 0.567, R2Y = 0.985, Q2Y = 0.972). In particular, S. rimosus and other strains were significantly classified as PLS1. Metabolites selected by VIP > 0.7 and p-value < 0.05 were determined as the potential variables that separated each species in the PLS-DA model. Selected metabolites were tentatively identified by molecular weight, elemental composition, and mDa using UPLC-Q-TOF-MS software, and MSn fragmentation and UV λmax(nm) analyzed by LC–ESI-MS/MS were used to support the identification. Twenty-one metabolites such as oxytetracycline (3), altamycin A (4), 2-hydroxystaurosporinone (5), USF-19A (7), staurosporine (9), rimocidin (13), strevertene A (14), AN-201 III (20), and 13 unidentified metabolites were detected (Table 1). The structures of the tentatively identified metabolites are shown in Fig. 3. As shown in Table 2, 2-hydroxystaurosporinone (5), staurosporine (9), and three unidentified metabolites (1, 8, 17) were found in S. rubrolavendulae, whereas oxytetracycline (3), altamycin (4), rimocidin (13), and four unidentified metabolites (2, 15, 16, 19) were detected in S. rimosus. Furthermore, three unidentified metabolites (10, 11, 12) were S. scabiei-specific metabolites, and strevertene A (14) and unidentified metabolites (21) were from S. virginiae.

Fig. 2.PLS-DA score plot derived from the UPLC-Q-TOF-MS dataset of Streptomyces spp. Identical species are represented by the same color. S. coelicolor: 20100; S. griseus: 1747, 14680; S. indigoferus: 8645; S. peucetius: 9038, 9199; S. rimosus: ▲20082, △21078; S. rubrolavendulae: 4426; S. scabiei: 20200, 20227, 20135; S. virginiae: 20731 20084.

Table 1.aIdentified metabolites based on variable importance projection (VIP) analysis with a cut-off value of 0.7 and a p-value < 0.05; bRetention time; cMW, Molecular weight; dMF, Molecular formula; emDa stands for error in milliDaltons; fN.I., Not identified; g[M+FA-H]-; hNegative mode; i[M+Na]+

Fig. 3.Structures of significantly different metabolites among Streptomyces spp. The numbers of the compounds in parentheses are the same as those referred to in Tables 1 and 2.

Table 2.The production of each metabolite was calculated by the peak area from an UPLC-Q-TOF-MS chromatogram. +: >0.5 count, ++: >10 count, +++: >100 count, -: <0.5 count or not detected.

Antibacterial Activity of Streptomyces spp. and Their Active Compounds

Crude extracts of the 14 Streptomyces strains were tested for antibacterial activity against Xoo. As shown in Table 3, S. rimosus (KACC 20082 and KACC 21078) and S. scabiei (KACC 20200) displayed the highest relative antibacterial activity, whereas the other strains showed less activity or no activity. To confirm these antibacterial activities, we measured the MIC and IC50 values for S. rimosus KACC 20082 and KACC 21078 (Table 4). The IC50 value was calculated as the concentration that caused 50% growth inhibition of Xoo. The IC50 values of S. rimosus strains KACC 20082 and KACC 21078 were 4.09 μg/ml and 3.00 μg/ml, respectively, whereas that of the positive control DAPG was 6.37 μg/ml.

Table 3.aKACC, Korean Agricultural Culture Collection. KCTC, Korean Collection for Type Culture. MJM, Collection of Myoungji University. bInhibition % of Xoo growth: -, <20; +, >50; ++, >90. Xoo: Xanthomonas oryzae pv. oryzae.

Table 4.aMinimum Inhibitory Concentration. bInhibitory concentration at 50%. MIC and IC50 are expressed as average values from three independent experiments. All values are expressed as the mean ± SD.

To isolate the active compounds from S. rimosus, we performed prep-HPLC analysis. The 12–14 (min) fractions by prep-HPLC exhibited the greatest suppression of Xoo growth (Fig. S1) at more than 100% inhibition. These active fractions were further analyzed by UPLC-Q-TOF-MS and LC-ESI-MS/MS to identify the active compounds. As a result, the active compound was identified as an oxytetracycline by its molecular weight, elemental composition, MSn fragmentation, and UV λmax (nm) as compared with a standard compound. The MIC and IC50 of oxytetracycline were 0.08 μg/ml and 0.04 μg/ml, respectively (Table 4).

 

Discussion

In this study, we sought to classify Streptomyces species based on their secondary metabolites. The metabolitebased and 16S rRNA-based dendrograms exhibited the same patterns of species discrimination. In particular, S. rimosus was separated from other species by PLS1 and PLS3 (Fig. 2). To distinguish each species on the basis of species-specific metabolites, the PLS-DA model was used and 21 metabolites such as oxytetracycline (3), altamycin A (4) 2-hydroxystaurosporinone (5), staurosporine (9), rimocidin (13), strevertene A (14), AN-201III (20), and 14 unidentified metabolites were tentatively identified. 2-Hydroxystaurosporinone (5), staurosporine (9), and two unidentified metabolites (1, 17) were detected as S. rubrolavendulae-specific metabolites. 2-Hydroxystaurosporinone and staurosporine are alkaloid metabolites [39,42], and 2-hydroxystaurosporinone (5) is derived from staurosporine [14,29]. Alkaloid groups are produced by many kinds of organisms, including fungi, bacteria, animals, and plants [24]. Several Streptomyces species are known to produce antibiotics and related alkaloid compounds. Staurosporine is a bis-indole alkaloid originally isolated in 1977 from S. staurosporeus. This member of the indolocarbazole alkaloid family has several well-known biological activities, due to its potential as an anticancer drug [48]. Staurosporine has been shown to inhibit protein kinases to control ATP binding to the kinases [60], and it affects cell viability by participating in apoptosis [9]. The S. virginiae-specific metabolites were strevertene A (14) and unidentified compound (21) (Table 2). Strevertene A is a polyketide compound that was isolated from fermentation products of Streptomyces spp. [52], and was detected in one strain of S. virginiae (KACC 14680) in the present study (Table 2). Oxytetracycline (3), altamycin (4), and rimocidin (13) were identified as S. rimosus-specific metabolites. These are tetraene-related compounds that contain a polyketide group and four carbon-carbon double bonds with a large macrolactone ring [46, 54]. Oxytetracycline and rimocidin were isolated from S. rimosus species in a previous study [18, 50]. Polyketides are produced by almost all living organisms and are a diverse array of natural products. They are particularly highly produced in Streptomyces and have various activities such as antibacterial, antifungal, immunosuppressant, antiparasitic, antitumor, and pharmacological properties [3, 53].

The antimicrobial activities of Streptomyces spp. against various microorganisms, including Staphylococcus aureus, Escherichia coli, Klebsiella pneumonia, Pseudomonas aeruginosa, Bacillus cereus, Aspergillus niger, Candida albicans, Aspergillus fumigates, and Aspergillus flavus have been previously reported [43,59]. In the present study, evaluation of the antibacterial activity against Xanthomonas oryzae pv. oryzae of eight Streptomyces species (14 strains) showed that S. rimosus had relatively high antibacterial activity (Table 3). To identify the active antibacterial compounds in S. rimosus, we performed prep-HPLC analysis and active fractions were analyzed by UPLC-Q-TOF-MS. These results confirmed oxytetracycline as an active compound. Oxytetracycline is a broad-spectrum tetracycline group of antibiotics that interferes with the ability of bacteria to produce essential proteins [47]. Various biological activities of oxytetracycline have previously been reported, including antibacterial [45], anti-acne, and anti-inflammatory effects [15]. In particular, oxytetracycline was mainly used for fire blight management of fruit in America [11]. In this study, eight Streptomyces species (14 strains) were classified based on secondary metabolite profiling analyzed by UPLC-Q-TOF-MS. Among them, S. rimosus showed the highest antibacterial activity against Xoo. As a result of activity-guided prep-HPLC, oxytetracycline was identified as an S. rimosus-specific antibacterial metabolite. This research indicates that chemotaxonomic studies are an effective tool to supplement conventional bacterial classification methods as well as for the evaluation of species-specific bioactivity.