Present address: Yoon-Suk Kang, Division of Environmental Science and Ecological Engineering, Korea University, Anam-Dong 5 Ga, Seoul 136-713, Korea.
Editor: Gilbert Shama
Correspondence: Moonjae Cho, Department of Biochemistry, Cheju National University, Jeju 690-756, Korea. Tel.: +82 64 754 3837; fax: +82 64 725 2593; e-mail: email@example.com
Of the 316 actinomycetes strains isolated from various habitats, Streptomyces sp. strain JJ45 showed the strongest antibiotic activity against the plant pathogenic bacteria Xanthomonas campestris pv. campestris and was thus chosen for further study. The 16S rRNA gene sequence (1500 bp) and rpoB gene partial sequence (306 bp) of Streptomyces strains JJ45A and JJ45B were determined. The respective strain JJ45B sequences exhibited 96.8% identity with the Streptococcus gelaticus 16S rRNA gene sequence and 98.4% identity with the Streptococcus vinaceus ATCC 27478 rpoB partial sequence. The fermentation broth of the JJ45B strain was extracted to find an inhibitor of bacterial growth. The distilled water extract showed the highest activity against pathogenic bacteria. The active molecule was isolated by column chromatography on polyacrylamide or silica gel, thin-layer chromatography, and HPLC. It showed growth inhibition activity only toward phytopathogenic Xanthomonas sp. The structure of the compound was identified as α-l-sorbofuranose (3→2)-β-d-altrofuranose based on the interpretation of the nuclear magnetic resonance spectra.
The genus Xanthomonas, one of the most omnipresent groups of Gram-negative plant pathogenic bacteria, causes a variety of diseases in multiple plants (Leyns et al., 1984). Xanthomonas campestris pv. campestris is the causal agent of black rot disease in crucifers (Brassica spp.) and tomatoes (Lycopersicon spp.), and X. campestris pv. vesicatoria causes bacterial leaf spot in peppers (Capsicum spp.). These diseases are characterized by V-shaped chlorotic to necrotic lesions at the margins of the leaves and blackened vascular tissues. Full-leaf yellowing and wilting also occur as the disease advances (Hayward, 1993). Heavy application of sprays containing copper compounds has been used to control these diseases for the last several decades. However, copper-resistant strains and their mechanism of resistance via a copper resistance plasmid have been reported (Bender et al., 1990; Cooksey et al., 1990; Voloudaskis et al., 2005). Search for new antibiotics effective against these pathogenic bacteria is an important area of antibiotic research.
Filamentous soil bacteria belonging to the genus Streptomyces are the best sources for bioactive natural products used as pharmaceuticals; many antibiotics, including 60% of the antibiotics developed for agricultural uses, have been isolated from Streptomyces species (Miyadoh, 1993). The identification of Streptomyces and other related genera at a genus level can be performed easily by a numerical classification method based on morphological and biochemical characteristics (Williams et al., 1983). In addition, 16S rRNA gene sequence data have proven invaluable in streptomycetes systematics, enabling the identification of several new streptomycetes (Mehling et al., 1995). However, 16S RNA gene analysis alone can be misleading because of an intraspecific variation (Clayton et al., 1995). Recently, a new method based on RNA polymerase β-subunit (rpoB) partial sequences has been applied to identify phylogenetic relationships (Kim et al., 2004; Mun et al., 2007).
Previously, we collected 316 actinomycete strains from Jeju Island, screened them against six plant pathogenic fungi, and found 10 strains with antifungal activity (Kim et al., 2001). Using the same library, we screened for antibacterial activity against X. campestris pv. vesicatoria. In the present study, we report the isolation of a new strain of Streptomyces (JJ45) with antibacterial activity against X. campestris. The purification, identification, and structural profile of the active compound from Streptomyces sp. strain JJ45 are described.
Materials and methods
Three hundred and sixteen actinomycete strains were isolated from diverse soil samples from Jeju Island in Korea. Bennett's agar (10 g of glucose, 1 g of yeast extract, 2 g of Bacto peptone, and 1 g of beef extract per liter of distilled water) was used for the selective isolation and preservation of actinomycetes. Indicator microorganisms (X. campestris pv. vesicatoria, X. campestris pv. campestris, and X. campestris pv. oryzae) were grown overnight in Luria–Bertani (LB) medium at 27 °C, and a modified disk-agar plate diffusion method (Okeke et al., 2001) was used to determine the antibacterial activities of the actinomycetes against X. campestris pv. vesicatoria. Briefly, sterile filter paper disks (8 mm) were soaked with 20 μL of each actinomycetes culture medium, dried at 37 °C, and applied to the surface of an LB agar plate seeded with 0.9% water agar containing the culture broth of the indicator bacteria. The plates were incubated for 15 h at 28 °C, and the antibacterial activity was evaluated by measuring the diameter of the clear inhibition zone. The strains with the highest antimicrobial activity against the indicator bacteria were selected and named Streptomyces sp. JJ45A and JJ45B.
Streptomyces sp. JJ45A and JJ45B were cultured in production medium (1.0% glucose, 0.3% yeast extract, 0.4% soy bean meal, 1.0% corn steep liquor, 0.05% K2HPO4, and 0.01% MgSO4·7H2O, pH 7.0) on a rotary shaker at 160 r.p.m. and 28 °C. The packed cell volume method was used to measure cell growth, and the cells and medium were tested for antibacterial activity at 8-h intervals.
Preparation of DNA
DNA was prepared using a bead beater–phenol extraction method (Kim et al., 2004). Briefly, a loopful of each isolate culture was suspended in 200 μL of TE buffer (10 mM Tris-HCl, 1 mM EDTA, and 100 mM NaCl, pH 8.0), placed in a 2.0-mL screw-cap microcentrifuge tube with 100 μL (packed volume) of glass beads (diameter, 0.1 mm; Biospec Products, Bartlesville, OK) and 100 μL of phenol : chloroform : isoamyl alcohol (25 : 24 : 1) (Sigma-Aldrich, St. Louis, MO). To disrupt the bacteria, the tube was oscillated on a Mini-BeadBeater (Biospec Products) for 1 min. The phases were separated by centrifugation (12 000 g, 15 min), and the aqueous phase was transferred to a clean tube. The DNA was precipitated by adding 10 μL of 3 M sodium acetate and 250 μL of ice-cold ethanol and incubating the mixture at −20 °C for 10 min. The DNA pellet was washed with 70% ethanol, dissolved in 60 μL of TE buffer, and used as a PCR template.
Amplification of 16S rRNA gene and rpoB by PCR
The 16S rRNA gene was amplified using the following primers: 27F (forward), 5′-AGAGTTTGATCATGGCTCAG-3′; and 1492R (reverse), 5′-AAGGAGGTGATCCARCCGCA-3′ (Kim et al., 2004; Mun et al., 2007). The primers for amplification of a partial sequence of rpoB were forward, 5′-TCGACCACTTCGGCAACCGC-3′; and reverse, 5′-TCGATCGGGCACATGCGGCC-3′ (Kim et al., 2004). For the PCR, template DNA (50 ng) and 20 pmol of each primer were added to a PCR mixture tube (AccuPower PCR PreMix; Bioneer, Daejeon, Korea) containing 1 U of Taq DNA polymerase, 2.5 μL of each deoxynucleoside triphosphate, 50 mM Tris-HCl (pH 8.3), 40 mM KCl, 1.5 mM MgCl2, and gel loading dye to a final volume of 20 μL. The reaction was subjected to 30 cycles of amplification (30 s at 94 °C, 45 s at 60 °C, and 45 s at 72 °C), followed by a 5-min extension at 72 °C (model 9600 thermocycler; Perkin-Elmer Cetus). The PCR products were electrophoresed in 1.2% agarose gels and purified with a Qiaex II gel extraction kit (Qiagen, Hilden, Germany).
16S rRNA gene and rpoB nucleotide sequencing
The sequences of the purified 16S rRNA gene and rpoB PCR products were directly determined using the forward and reverse PCR primers with an Applied Biosystems model 373A automatic sequencer and a BigDye Terminator Cycle sequencing kit (Perkin-Elmer Applied Biosystems, Warrington, UK). To analyze the phylogenetic relationships of the novel actinobacteria strains, the 16S rRNA gene sequences and rpoB partial sequences of the novel strains were aligned with those of reference strains (Kim et al., 2004; Mun et al., 2007), using clustalw (Saitou & Nei, 1987). Phylogenetic trees were constructed by the neighbor-joining method using mega4.0 software (Thompson et al., 1997). This method is widely used for phylogenetic analysis. The branch supporting values were evaluated using 1000 bootstrap replications.
Fermentations were carried out in a 7-L jar fermentor (Korea Fermentation Co., Inchon, Korea) containing 5 L of the same medium used for culturing, at 28 °C and with 10% dissolved oxygen. The pH, packed cell volume (%), and antibacterial activity were monitored at 6-h intervals. After 36 h of fermentation, the culture medium (5 L) was mixed with 300 g of celite and separated from the precipitate by filtration.
Extraction and purification of the active compound
Figure 1 shows the scheme used to extract and isolate the active compound from Streptomyces sp. JJ45. The cell-free culture medium was mixed with the same volume of ethanol and centrifuged at 2800 g to remove any biomass. The resulting supernatant was concentrated on an evaporator (N-1000, EYELA). The dry extract was dissolved in distilled water (dH2O), shaken slowly, and extracted sequentially with dH2O, n-hexane, and ethyl acetate. The antibacterial activities of the three extracts were assayed, and the dH2O extract showed the highest activity. The dH2O extract was fractionated by gel filtration chromatography on a Bio-gel P2 column (610 × 16 mm; Bio-Rad) in dH2O (0.5 mL min−1). The active fractions were pooled and analyzed by thin-layer chromatography (TLC) on silica gel plates (SiO2, Merck) with acetonitrile (ACN) : methanol (MeOH) : dH2O (6 : 1 : 3, v/v/v). The plates were observed under UV light at 310 nm. For large preparations, chromatography was accomplished with a silica gel column (Kiselgel 60, 370 × 11 mm, 230–400 mesh; Merck), using the same ACN : MeOH : dH2O solvent system. After elution, an inhibition test and TLC analysis were performed, and the active fractions were pooled and concentrated by evaporation. The samples were further purified by preparative HPLC using a SPECTRA system (UV1000, P2000) with a Phenomex PRIDIGY ODS column (250 × 10 mm) and ACN : dH2O (4 : 1, v/v) as the mobile phase; elution was at 3 mL min−1, and the absorbance of the eluant was monitored at 254 nm. The active fractions were concentrated using a freeze drier (Coldvac 50; Hanil R&D, Korea).
Identification of the compound
All nuclear magnetic resonance (NMR) experiments were performed on a Bruker Avance 400 spectrophotometer (9.4 T, Karlsruhe, Germany) at 298 K. For 1H NMR analysis, 16 transients were acquired with a 1-s relaxation delay using a 90° pulse of 10.2 μs, and a spectral width of 6000 Hz with 32 K data points. The 13C NMR and DEPT spectra were obtained with a spectral width of 24 000 Hz using a 90° pulse of 10.3 μs and 64 K data points. All two-dimensional spectra were acquired with a 2 K × 256 data matrix. The delays for long-range coupling by heteronuclear multiple-bond correlation were 70 and 35 ms, and the mixing time for NOESY was 1 s. Before Fourier transformation, 2 K zero-filling and sine-squared bell window functions were applied using xwin-nmr (Bruker) (Shin et al., 2006).
Results and discussion
Isolation and selection of antibiotic-producing streptomycetes
Among the 316 streptomycetes strains isolated from various habitats on Jeju Island (Kim et al., 2001), two strains, named JJ45A and JJ45B, formed surrounding clear zones when grown on agar plates with X. campestris pv. vesicatoria. The JJ45A strain easily formed white spores and produced brown pigments, whereas the JJ45B strain did not. The temperature range for the growth of the two strains was 25–30 °C, with optimal growth at 28 °C. Their growth characteristics strongly suggested that these two strains belong to the genus Streptomyces. Figure 2 shows the time course for the growth and production of bactericidal secondary metabolites in a flask culture. For strain JJ45A, the cells took nearly 48 h to begin growing and reached the stationary phase after 48 h of cultivation. However, the antibacterial activity was the highest after 24 h of incubation and then decreased gradually. The packed cell volume of the JJ45B strain increased rapidly, and the antibacterial activity showed a slight increase after 24 h of cultivation. For both strains, the antibacterial activity in the medium did not increase after 24 h. During flask culture, the pH varied between 5.5 and 7.9 for strain JJ45A and between 6.5 and 7.6 for strain JJ45B (data not shown). The JJ45B strain demonstrated the most pronounced clear zones (Fig. 3a), and this isolate was selected for scaled-up fermentation in order to purify the antibacterial compound.
16S rRNA gene and rpoB gene phylogenetic analysis
After PCR amplification from strains JJ45A and JJ45B, the 16S rRNA gene and partial rpoB gene were sequenced, and the sequences were aligned using the clustalw multiple alignment program (Saitou & Nei, 1987). The sequence analysis showed that the 16S rRNA gene and the partial rpoB sequences of JJ45A and JJ45B were identical. The phylogenetic relationships were determined between the novel actinobacterium and previously reported reference strains (Kim et al., 2004; Mun et al., 2007). The 16S rRNA gene sequence of JJ45 had 96.8% sequence identity with the 16S rRNA gene sequence from Streptococcus gelaticus (data not shown), and the rpoB partial sequence of JJ45 showed 98.4% identity with the rpoB partial sequence from Streptococcus vinaceus ATCC 27478 (Fig. 4.). This suggests that JJ45 may be a novel species belonging to the genus Streptomyces.
Purification of the antibacterial compound
The packed cell volume obtained by fermentation was double that obtained by flask culture, but the antimicrobial activity showed no difference between the two culture methods (data not shown). We obtained 5 L of culture solution of Streptomyces sp. JJ45B using scaled-up fermentation. The antimicrobial compound was purified by first extracting the culture filtrate with EtOH. The supernatant was vacuum-evaporated to dryness, taken up in dH2O, and extracted with dH2O and organic solvents. The dH2O extract was the most active and was chromatographed on polyacrylamide or silica gel. An inhibition test with the dry extract (initially 5.4 g) allowed the purification of the fractions with antibacterial activity. After the silica gel chromatography, the active fractions were pooled, concentrated, and analyzed by TLC, which revealed three bands with Rf values of 0.9, 0.7, and 0.53 under UV310 light. An inhibition test showed that only the Rf 0.9 band possessed antibacterial activity. The concentrated active fractions (340 mg) were further fractionated by preparative HPLC using an ODS column. We obtained two peaks, A and B, with retention times between 3 and 7 min and tested these for inhibitory activity against X. campestris pv. campestris. Antimicrobial activity was only detected in the peak A fraction (Fig. 3b). The final dry extract (103 mg) was isolated as a white powder, which was soluble in water, but insoluble in alcohols or ethyl acetate.
This final extract was tested for antibacterial activity against several phytopathogenic bacteria such as X. campestris pv. vesicatoria, X. campestris pv. campestris, X. campestris pv. oryzae, X. axonopodis, Pseudomonas syringae pv. syringae, and Erwinia amylovora. The compound inhibited the growth of only Xantomonas sp. (data not shown) and thus has a very narrow spectrum of action, as does the bacteriocin glycinecin A produced by X. campestris pv. glycines 8ra (Heu et al., 2001).
Structural elucidation of the active compound
Matrix-assisted laser desorption/ionization time-of-flight analysis of the compound from peak A revealed a mixture of two compounds with molecular masses of 381.0 and 550.9. The NMR data suggested that both compounds consisted of sugars, and five saccharides were observed. Two of these saccharides were connected, based on the HMQC-TOCSY experiments (Supporting Information, Fig. S1). A total of 12 peaks suggested that the compound was a disaccharide consisting of two hexoses. Further analysis of two-dimensional NMR data (Figs S2–S4) showed that there were two groups of peaks: one at 61.4/3.62, 62.4/3.76, 74.1/4.00, 75.9/3.39, 76.5/4.15, and 103.7 and the other at 60.2/3.75, 69.3/3.41, 74.2/3.18, 75.8/3.43, 81.4/3.83, and 95.9/4.57. The first group of peaks matched well with α-l-sorbofuranose, and the second group matched β-d-altrofuranose. The molecular mass of 381 was produced by 180+180−18+39 (K). Therefore, the small compound was considered a disaccharide. To elucidate the linkage between the two saccharides, NOESY experiments were performed (Fig. S5). An NOE peak was observed between the 3-H of sorbofuranose and the 2-H of altrofuranose, showing that the structure of the small compound was α-l-sorbofuranose (3→2)-β-d-altrofuranose (Fig. 5). We partially separated the two compounds using size exclusion gel chromatography (Bio-gel P2, 15 × 2300 mm; Fig. 1), and the smaller compound showed stronger bactericidal activity.
A large number of antibiotics are glycosides. In numerous cases, the glycosidic residues are crucial for their activity, but sometimes glycosylation only improves the pharmacokinetic parameters. Recent developments in molecular glycobiology have improved our understanding of aglycone vs. glycoside activities and have made it possible to develop more active or more effective glycodrugs (Křen & Rezanka, 2008). Recently, the development of carbohydrate-based antibiotics has been suggested as a new approach to the problem of antibiotic resistance (Ritter & Wong, 2001). Dimeric aminoglycosides are possibly the best candidates for carbohydrate-based antibiotics (Agnelli et al., 2004). However, to our knowledge, the present work is the first description of a naturally occurring simple disaccharide with bactericidal activity. The mechanism of action of α-l-sorbofuranose (3→2)-β-d-altrofuranose against Xanthomonas spp. is a topic for future research, but a clue to its mechanism may be found in the existence of sorbose and altrose in microorganisms. l-Sorbose is the precursor of 2-keto-l-gluconic acid, which is the precursor of vitamin C (Sugisawa et al., 2005). Although l-altrose is present in many bacterial cell wall components or precursors (Senchenkova et al., 1995; Liu & Tanner, 2006), d-altrose is found rarely. Judging by its narrow spectrum of action, α-l-sorbofuranose (3→2)-β-d-altrofuranose may disturb a pathway involving sorbose or altrose, perhaps a pathway unique to Xanthomonas sp.
This work was supported by a grant from the Academic Research Fund of the Cheju National University Development Foundation.