INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Research studies of inhibitors of bacteria, fungi, and cultured cells have accumulated, and many effective drugs have appeared for clinical use. However, the number of reports of antimicroalgal compounds is relatively small, although these organisms cause such problems as blooms by cyanobacteria, red tide and the production of marine toxins by dinoflagellates, and biofilm formation on marine structures by diatoms. In particular, some bloom-forming cyanobacteria are known to produce toxic metabolites such as anatoxins (8) and microcystins (3). The mechanism by which the cyanobacteria form blooms remains to be investigated, but the threat and damage to human and animal life and to industry are serious. In the present study, we screened strains for antimicroalgal compounds against one cyanobacterium and three eukaryotic microalgae. A culture broth of each of 2,594 marine bacterial strains was examined in this screening, and among them, 37 strains were found to produce anticyanobacterial substances in the culture on marine broth 2216 (MB). Interestingly, no inhibitory activity toward the three eukaryotic microalgae tested was apparent in the broths of any of these bacterial strains.

We purified the anticyanobacterial compound from C-979, which showed the highest anticyanobacterial activity. The chemical structure was determined for the compound, and the bioactivity profile was examined. Furthermore, we examined if the above-mentioned 37 strains produce the same anticyanobacterial compound that C-979 produces.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Microalgae and cultivation medium. All the microalgal strains and the media used in this study are listed in Table 1. F/2, K + ESM, MC, SOT, and CT media have been described elsewhere by Guillard and Ryther (15), Miyashita et al. (21), Watanabe (32), Ogawa and Terui (23), and Watanabe and Ichimura (34), respectively. CSi (35) is a modified C medium (17) supplemented with silicate.

Collection of marine bacteria. Sampling of marine bacteria was conducted in July 1994 at Yap (Micronesia) and Palau (Belau), in July 1995 at Palau, and in October 1995 at Okinawa (the southwest islands of Japan). The bacteria were collected in a shallow sea area from the surface of fresh benthos (mainly macroalgae) and from sea sands by using selective media reported elsewhere (T. Takadera, M. Nishijima, K. Yoshikawa, M. Araki, and H. Sano, Poster Session 2nd Asia-Pacific Mar. Biotechnol. Conf. 3rd Asia-Pacific Conf. Algal Biotechnol., p. 97, 1997). The halophilicity of the collected bacteria was confirmed by inoculating each bacterium separately onto marine agar 2216 (Difco Laboratories, Detroit, Mich.), which contained peptone (0.5%), yeast extract (0.1%), NaCl (1.9%), iron(II)-citrate (0.01%), and some other inorganic salts and onto a salt-free agar plate of Bacto Peptone (Difco; 0.5%), Bacto Yeast Extract (Difco; 0.1%), and iron(II)-citrate (0.01%), before incubation at room temperature. The bacterium which could grow only on marine agar was judged halophilic and is called the marine bacterium in this study. The bacterial cells were preserved in 75% artificial seawater (Tropic Marin; Aquarientechnik, Wartenberg, Germany) supplemented with 10% glycerol at 80°C until needed for use. A total of 2,594 isolates, including 135 isolates from Yap, 1,043 from Palau, and 1,416 from Okinawa, were obtained as marine bacteria.

Cultivation of the marine bacterium and preparation of the screening sample. A loopful of cells of the individual marine bacterium was inoculated into 3 ml of MB (Difco) and cultured at 30°C for 48 h on a reciprocal shaker. The cultured broth was lyophilized and suspended in 1 ml of 60% ethanol. The supernatant of this suspension was diluted with methanol (1:3) and used for the antimicroalgal screening.

Bioassay for marine microalgae and freshwater green algae. A sample solution of 10 µl was put into each well of a 96-well tissue culture plate (Becton Dickinson & Co., Franklin Lakes, N.J.) in duplicate and air dried. A microalgal culture, 200 µl, was poured into each well. The culture plate was incubated at 22°C under fluorescent light at 25 microeinsteins m² s¹ with a light-dark period of 12 h each for 3 to 5 days. 3-(3',4'-Dichlorophenyl)-1,1-dimethylurea (DCMU; Nacalai Tesque, Inc., Kyoto, Japan) at 10 µg/ml was used as a positive control. The activity was evaluated by visual observation of the well, except for Brachiomonas submarina and Prorocentrum micans, whose sensitivity to the sample was determined by their motility under an optical microscope. 2-Heptyl-4-hydroxyquinoline N-oxide (HQNO) was purchased from Sigma Chemical Co. (St. Louis, Mo.), and common L-form amino acids were obtained from Wako Pure Chemical Industries, Ltd. (Osaka, Japan).

Bioassay for freshwater cyanobacteria. A bioassay for the freshwater cyanobacteria was conducted according to the chlorophyll absorbance method described by Uchida et al. (31). Algal cells, which had been incubated at 20°C in 100 ml of CT medium for 14 days, collected by centrifugation at 1,000 × g for 10 min, and then washed with 10 ml of CT medium, were suspended in an appropriate volume of CT medium, and their optical density at 665 nm was adjusted to 3.0. Samples of each dissolved in 20 µl of methanol were put into a 96-well tissue culture plate in quadruplicate, and the solvent was removed in an air atmosphere. The algal suspension (100 µl) and 100 µl of CT medium were added to the wells, and the culture solution was incubated under fluorescent lighting (24 h/day) at 20°C for 6 days. Colistin sulfate (Wako), 40 µg/ml, was used as a positive control in this assay. The optical density of the incubation mixture at 665 nm was measured with a microplate reader (Tosoh MPR-A4iII; Tokyo, Japan), and the mean value for the optical density was calculated. The cyanobacterium was judged sensitive (designated by plus signs in Tables 3 and 4) when the density had been reduced to 85% or less and as tolerant (designated by a minus sign) when the density had been increased to 115% or more. When the density change was within 15%, the result was designated +w.

Taxonomical studies of C-979. Conventional taxonomical features of the marine bacterium C-979 were determined according to the procedures described by Cowan (7), except that all the media used were supplemented with 3% NaCl or prepared in 75% artificial seawater, because strain C-979 required NaCl for growth. The change in pH value during the test of glucose utilization (O/F test) was monitored by phenol red, and G+C content of bacterial DNA was determined by high-performance liquid chromatography (HPLC) with a catalytic reaction using nuclease P1 (Seikagaku Co., Tokyo, Japan). A morphological study was carried out mainly by transmission electron microscopy. The isoprenoid quinone composition of strain C-979 was analyzed according to the method of Nishijima et al. (22).

Fermentation of C-979 and purification of the anticyanobacterial compound. C-979 was cultured in a 1-liter Erlenmeyer flask containing 200 ml of MB (1). The cells were removed by centrifugation, and the supernatant fluid (2.4 liters) was washed twice with equal volumes of ethyl acetate. The aqueous layer was concentrated to dryness in vacuo. The resulting residue, resolved in 2 liters of water, was passed through a column of active charcoal (500 ml; Nacalai Tesque; HCl treated). The unabsorbed fraction was acidified by hydrochloric acid and then dialyzed against distilled water at ambient temperature, using a cellulose ester membrane with a molecular weight cutoff of 100 (Spectra/Por CE; Spectrum Medical Industries, Inc., Houston, Tex.). The desalted solution was applied to 50 ml of Dowex 50W (Dow Chemical Co., Midland, Mich.) resin, eluting with 5.6% aqueous ammonia. The concentrated eluate with the anticyanobacterial activity was further separated on a DEAE-2SW HPLC column (7.8 mm [inside diameter] by 300 mm; Tosoh), eluting with 20 mM ammonium acetate.

Determination of anticyanobacterial compound. A mass spectral analysis was conducted with a JEOL JMS-SX102 mass spectrometer (MS). The ¹H-nuclear magnetic resonance (NMR) spectrum was measured with a Varian Unity 500 NMR spectrometer, while the infrared (IR) spectrum was obtained from a KBr pellet with a JASCO FT-IR 7000 spectrophotometer (Japan Spectrophotometry Co., Tokyo, Japan). The configuration of the -carbon on -cyanoalanine (CNAla) produced by C-979 was determined according to the advanced Marfey method described by Harada and colleagues (13). Authentic -cyano-L-alanine (L-CNAla) was obtained from Sigma, and 1-fluoro-2,4-dinitrophenyl-5-L-leucineamide (L-FDLA) and -D-leucineamide (D-FDLA) were synthesized by the method described by Marfey (20). The FDLA derivative of CNAla was analyzed in a TSKgel octyldecyl silane 80Ts column (4.6 mm [inside diameter] by 150 mm; Tosoh) at 40°C with a linear gradient of acetonitrile (30 to 70% in 40 min) containing 10 mM trifluoroacetic acid at a flow rate of 1.0 ml/min.

NMR measurement of the -cyanoalanine productivity. The supernatant fluid (1 ml) of the C-979 culture in 200 ml of medium in a 1-liter Erlenmeyer flask was washed once with ethyl acetate, and the aqueous layer was lyophilized. The residue was resolved in 1 ml of D₂O and analyzed by ¹H-NMR. Sarcosine (Wako) was used as the internal standard. The peaks selected for estimating the concentration were those of N-methyl protons in sarcosine (ca. 2.7 ppm) and of the methylene next to the -carbon in CNAla (ca. 3.2 ppm).

Bioassay for other organisms. Antibacterial activity was tested by the paper disk method, using 50 µg of L-CNAla on an 8-mm-diameter disk (Advantec Toyo Co., Tokyo, Japan). The effect of L-CNAla against the macroalga Ulva conglobata (spores and fronds) was examined in PES medium (26) by the method described by Hattori et al. (16). The bioassay with barnacle cyprid and brine shrimp was conducted with Balanus amphitrite (10 bodies) and Artemia salina (10 bodies) added to 5 ml of seawater containing L-CNAla. The solution was incubated together in triplicate under the conditions described by Kon-ya et al. (18).

Assessment of CNAla in the culture of marine bacteria. Activated charcoal (0.5 g) was added to the tube with the cultured broth of a marine bacterium, which had been incubated for 48 h on 10 ml of MB and mixed well, before the tube was allowed to stand. The supernatant (4 ml) was transferred to a new test tube. Dowex 50W ion-exchange resin (1 g) was added, the suspension was mixed well, and the liquid fraction was discarded. The resin was washed three times with 5 ml of distilled water and then poured into 4 ml of a 5.6% ammonium solution. The solution containing the compounds released from the resin was concentrated to dryness with a centrifugal evaporator. The residue was resuspended in 50 µl of methanol and was applied to a thin-layer chromatograph (TLC) (silica gel 60F₂₅₄; Merck 1.05554; Darmstadt, Germany) with a mobile phase of ethanol-28% ammonium solution-water (18/1/4). Under these developing conditions, the relative flow of authentic L-CNAla was about 0.75, and CNAla on the TLC plate could be specifically detected by its blue-green color after being treated with a methanolic solution of ninhydrin (Wako) (4).

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Screening for antimicroalgal metabolites in the culture broth of marine bacteria. A growing body of research on natural products from organisms, especially microorganisms in the ocean, has appeared recently (11). Among these organisms, we are focusing on marine bacteria, which are defined as those bacteria which need sodium chloride for survival, because these bacteria can be relatively easily cultivated in the laboratory and because marine organisms which cannot survive without sodium chloride are considered to have unique metabolites which do not occur in their terrestrial counterparts. A total of 2,594 marine bacteria cultured on MB were tested with the antimicroalgal assay against the cyanobacterium Oscillatoria amphibia NIES-361, the unicellular green alga B. submarina NIES-375, the dinoflagellate P. micans NIES-12, and the diatom Skeletonema costatum NIES-16. As a result, the 37 cultured broths of the marine bacteria showed anticyanobacterial activity toward O. amphibia. Interestingly, none of these samples showed the antialgal activity against the other three microalgae. Among them, the cultured broth of C-979 showed the highest anticyanobacterial activity and was selected for further investigation.

Bacterial characteristics of C-979. Marine bacterium C-979 was separated from the surface of a fresh macroalga, Caulerpa serrulata var. serrulata f. lata, which had been collected from shallow water in the sea off Palau in July 1994. This strain was subjected to standard biological and physiological tests. It was a gram-negative facultative anaerobic rod with a polar flagellum, and lateral flagella were observed when it grew on solid media. The O/F test showed fermentation, and the catalase and oxidase tests were positive. The G+C content of its DNA was 46.0 mol%, and the major isoprenoid quinone was ubiquinone-8. According to the literature (2), these results indicate that this strain could be classified as Vibrio sp. Strain C-979 has been deposited at the National Institute of Bioscience and Human Technology (Agency of Industrial Science and Technology, Ministry of International Trade and Industry, Tsukuba-shi, Japan) with the accession no. FERM P-16360.

Determination of anticyanobacterial compound produced by C-979. A well-grown seed culture (20 ml) of strain C-979 in MB was used to inoculate a 180-ml medium in a 1-liter Erlenmeyer flask and was incubated at 30°C on a rotary shaker (110 rpm) for 24 h. The bioactive compound was found in the supernatant fraction, not in the cells. The bioassay-guided purification described in Materials and Methods afforded 16.4 mg of pure bioactive compound from 2.4 liters of the supernatant.

Productivity of -cyano-L-alanine by C-979. The initial production of L-CNAla by C-979 was estimated by using ¹H-NMR spectroscopy. Compared with the amount of protons in the internal standard, the concentration of CNAla in the cultured broth was 2.3 mM, that is, 0.26 g of CNAla per liter was initially produced by C-979 on MB. Thus, the yield from our purification was calculated to be 2.6%.

Biological activity of -cyano-L-alanine against microalgae. The biological activity of L-CNAla and some other antibiotics against four microalgal strains is summarized in Table 3. L-CNAla inhibited the growth of only the cyanobacterium O. amphibia, but DCMU and HQNO were toxic to both the cyanobacterium and the eukaryotic microalgae, although B. submarina was found to be tolerant of HQNO at a concentration of 50 µg/ml or less. Biological activity of L-CNAla against several microalgae was checked, with the results being indicated in Table 4. Some cyanobacteria were sensitive to L-CNAla at a concentration of 0.4 to 25 µg/ml, but others were resistant at a concentration of 50 µg/ml. The sensitive cyanobacteria included O. amphibia NIES-361, Synechococcus sp. strain CSIRO 94, Entophysalis deusta CCAP 1479/7, Microcystis aeruginosa NIES-298, Microcystis viridis NIES-102, and M. viridis NIES-103. On the other hand, no eukaryotic microalgae tested were sensitive to L-CNAla.

Biological activity of -cyano-L-alanine against other environmental organisms. The MIC for all the tested pathogenic bacteria, Bacillus subtilis 10707, Enterococcus hirae ATCC 10541, Staphylococcus aureus ATCC 6538P, Escherichia coli ATCC 26, Klebsiella pneumoniae ATCC 10031, Proteus vulgaris ATCC 6897, Pseudomonas aeruginosa BinH #1, and Shigella sonnei ATCC 9290, and for the yeast Candida albicans ATCC 10231, was >83.3 µg/ml. As a result of the susceptibility test by the paper disk method, no growth inhibition of the terrestrial bacteria B. subtilis IFO3134, S. aureus IFO12732, and E. coli IFO15034 nor of the marine bacteria Salinivibrio costicola ATCC 33508 and Pseudoalteromonas haloplanktis subsp. haloplanktis ATCC 14393 was apparent at the L-CNAla concentration of 50 µg/disk. The green macroalga, U. conglobata, was tested by adding 50 µg of L-CNAla per ml to the PES medium in which the spores or fronds were separately incubated. After 5 days, neither spores nor fronds had been killed by L-CNAla. The attachment ratio (73%) of the spores did not seem particularly low compared to that of the negative control (93%). On the other hand, the germination of the attached spores was completely inhibited, while the spores in the dish of the negative control germinated at the ratio of 81%. L-CNAla, up to 50 µg/ml, was not toxic either to A. salina or to the cyprid of B. amphitrite during the 48-h incubation in seawater. The barnacle cyprids settled and metamorphosed similarly with or without L-CNAla.

Production of -cyanoalanine by marine bacteria. To examine the production of CNAla by marine bacteria, the 37 isolates which had produced anticyanobacterial substances from among the 2,594 original isolates were cultured in a volume of 10 ml. CNAla in the broth was detected on a TLC plate treated with ninhydrin. As a result, 36 strains out of 37 were found to produce CNAla.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

Anticyanobacterial activity against O. amphibia NIES-361 was observed not in the ethyl acetate extract but in the aqueous layer in the supernatant of the culture broth of C-979 on MB. The bioactive compound in the aqueous layer was too hydrophilic to be extracted into the organic layer of the solvent systems of dichloromethane-methanol-broth (5:6:4 [vol/vol/vol], ethyl acetate-isopropanol-broth (8:3:10), or n-butanol-acetic acid-broth (4:1:5). The compound was not absorbed by activated charcoal, nor was it adsorbed to Sephadex LH-20 (Pharmacia, Uppsala, Sweden) or TSKgel Toyopearl HW40 (Tosoh), although these are commonly used for separating a compound with high polarity (29). As C-979 was cultivated on MB that contained several inorganic salts at relatively high concentrations, the bioactive compound could not be desalted because neither the required compound nor the salts were extractable by an organic solvent and were not adsorbable to the adsorbent used. To purify the active compound, it was necessary to remove a large amount of the salts from the broth.

The supernatant was treated with ethyl acetate, and then the resulting aqueous layer was passed through the column of activated charcoal. The unabsorbed fraction contained the desired bioactive compound and a large amount of sodium acetate. It was very difficult to separate both compounds by the usual technique. The separation could be achieved by dialysis with a cellulose ester membrane when hydrochloric acid was added to the inner solution. L-CNAla and some other organic compounds could not pass through the membrane, and almost all the salts together with sodium acetate were excluded from the dialysate. Further treatment of the dialysate with Dowex-50W ion-exchange resin and reversed-phase HPLC afforded pure L-CNAla.

CNAla is biosynthesized by various kinds of organisms. It has been found in several species of legume seeds (27) and seedlings (12). A radioisotope experiment has indicated that the cyanide ion supplied was assimilated to L-CNAla and -L-glutamyl--cyano-L-alanine (12). In higher plants, the toxic cyanide ion has been produced as a byproduct in the biosynthesis of ethylene, a plant growth hormone (25). The cyanide ion, together with L-cysteine, is converted to the nontoxic compound, L-CNAla, by -cyanoalanine synthase (36). The occurrence of L-CNAla in the toxic mushroom Clitocybe acromelalga has been reported elsewhere (14), although the biosynthesis of L-CNAla in this organism is unknown. Many cyanide-resistant bacteria have been found in soil (30). Radioisotope experiments have revealed that cyanide ion was converted to L-CNAla and then to L-asparagine and L-aspartic acid in Bacillus megaterium (6). In this case, L-CNAla was directly formed from O-acetyl-L-serine (enzymatically converted from L-serine) by O-acetylserine sulfhydrylase, and not by -cyanoalanine synthase (5). CNAla has also been found in a reaction mixture containing hydrogen cyanide and a cell extract of E. coli (10).

In our bioassay, the above-mentioned amino acids, L-cysteine, L-serine, L-asparagine, and L-aspartic acid, which can be biologically converted to L-CNAla in some organisms, were not toxic to O. amphibia NIES-361 and Synechococcus sp. strain CSIRO 94 at a concentration of 50 µg/ml, although these cyanobacteria were sensitive to L-CNAla at a concentration of 13 µg/ml or less. These results indicated that anticyanobacterial activity requires the cyano group at the  position of the -amino acid. All CNAla-producing bacteria reported so far have needed a supply of cyanide ion in the culture medium (4, 5, 6, 30). In this study, Vibrio sp. strain C-979 was found to produce a large amount of L-CNAla on MB, which did not contain the cyanide ion (1), and it is thus the first bacterial strain to produce CNAla without a supply of the cyanide ion in the medium, suggesting another metabolic pathway leading to L-CNAla, for example, dehydration of L-asparagine. In Pseudomonas fluorescens NCIMB 11764 (19), Pseudomonas stutzeri AK61 (33), and some pathogenic fungi including Fusarium solani (9), cyanide ion was enzymatically converted to ammonia or formamide and not to CNAla, indicating that these organisms had a different pathway of cyanide metabolism.

CNAla is known to be a simple and physiologically stable amino acid (28). The production of this compound by the marine bacterium C-979 of 0.26 g/liter, estimated by ¹H-NMR spectroscopy, was quite high. CNAla production by marine bacteria has not previously been reported, especially in the absence of the cyanide ion in culture broth. In this study, 36 strains of marine bacteria out of 2,594 tested were found to produce CNAla. Bacterial identification for these 36 isolates is to be conducted, although it is an interesting question if CNAla productivity is limited to vibrios. These 36 strains were collected from the surface of 28 samples, including marine macroalgae, benthos of unknown species, and sea sand, from Palau and Okinawa, Japan. However, CNAla was not apparent in 135 bacterial isolates from Yap. This result suggests a wide distribution of bacteria with the potential to produce CNAla. This has not previously been reported, perhaps partly because CNAla is too hydrophilic and too small to catch the interest of researchers in this field and partly because there has been little research on anticyanobacterial compounds. One exceptional culture broth, that of isolate G-235, showed anticyanobacterial activity as high as that of the 36 other broths, although no CNAla in this broth could be detected by TLC in this study. The active principle has not yet been determined.

The features of L-CNAla bioactivity seem unique: it is nontoxic to various environmental organisms, including bacteria, yeast, eukaryotic microalgae (Table 4), fronds of a macroalgae, and invertebrates, although the germination of spores of the macroalga U. conglobata seems to be inhibited by L-CNAla. On the other hand, it shows inhibitory activity that is specific for some cyanobacteria (Table 4). Potassium ion is known to inhibit the growth of some freshwater cyanobacteria (24), but the toxicity of the ion to other environmental organisms has not been extensively investigated. Few specific inhibitors of organic compounds showing inhibitory activity toward cyanobacteria have been reported so far. The data in Table 3 show that DCMU and HQNO were anticyanobacterial, although both showed some inhibitory action toward even eukaryotic microalgae at a moderate concentration. Our results suggest that the target structure of L-CNAla is specific to cyanobacteria and does not exist in other bacteria.