Identification and Characteristics of a Novel Burkholderia Strain with Broad-Spectrum Antimicrobial Activity

Abstract

A Burkholderia strain isolated from soil is capable of inhibiting the growth of bacteria, plant-pathogenic fungi, pathogenic yeasts, and protozoa. Inhibition does not involve cell contact or the presence of living cells, suggesting that at least a substantial portion of the antimicrobial activity is due to the excretion of extracellular compounds.

There is growing awareness of the need for development of new antimicrobial agents for the treatment of human, animal, and plant diseases. One possible source of novel antimicrobial agents is predator bacteria. Predator bacteria are a group of nonobligate bacteria that prey on microorganisms, including other bacteria, fungi, yeasts, and protozoa (2). These bacteria inhibit the growth of bacteria, fungi, yeasts, protozoa, and even other predator bacteria yet are able to grow without prey microorganisms (2). One predator bacterium, strain 679-2, that inhibited the growth of a wide variety of microorganisms in laboratory tests (3, 4) was also shown to be resistant to copper, survive inoculation into soil (3), and control fungal diseases of alfalfa and tomato (5). Unfortunately, strain 679-2 segregated colonial variants lacking antimicrobial activities at a frequency of 15% (3, 4), preventing isolation of the compound(s) responsible for that strain's antimicrobial activity. Because many predator bacteria were resistant to copper and were antagonistic to other microorganisms (2), copper-resistant bacteria were isolated from soil and tested for antimicrobial activity. Herein we report the isolation, characteristics, and polyphasic taxonomic study (14) of a novel Burkholderia strain whose antimicrobial activity is broad, extracellular, and stable.

Strain 2.2 N was isolated from a Hagerstown silty clay loam (pH 6.2) collected on the campus of The Pennsylvania State University. Dilutions of a soil suspension were plated on copper agar (0.25% [wt/vol] heart infusion agar [Difco, Detroit, Mich.] and 0.01% [wt/vol] CuCl₂ · 2H₂O [pH 6.5]). Colonies appearing on copper agar after 1 to 7 days incubation at 28°C were selected, and one, strain 2.2 N, was found to have antimicrobial activity against Micrococcus luteus, Saccharomyces cerevisiae, and Aspergillus niger. Strain 2.2 N was grown in 0.25-strength Tryptic Soy Broth (BBL Microbiology Systems, Cockeysville, Md.) containing 0.2% (wt/vol) sucrose (TSB+S) or on TSB+S containing 1.5% (wt/vol) agar at 30°C with aeration. Cell-free culture fractions were recovered by centrifugation (5,000 × g for 30 min at 4°C) of cultures grown in TSB+S broth medium for 48 h at 30°C in baffled flasks at 120 rpm. The cell-free, spent culture medium was decanted and sterilized by passage through a 0.22-μm-pore-size filter or pasteurized by heating at 80°C for 10 min. Antimicrobial activity of cultures or culture fractions was measured by a zone-of-inhibition assay. Micrococcus luteus, Mycobacterium smegmatis, Saccharomyces cerevisiae, Cryptococcus neoformans, and Candida albicans were grown from a single colony in 0.1-strength brain heart infusion broth (Difco) as previously described (3, 4). Cultures were incubated for 18 h at 30°C and could be refrigerated and used for 2 weeks. Spore suspensions of A. niger, Botrytis cinerea, and Septoria nodorum were prepared from lawns of sporulating colonies on 0.1-strength brain heart infusion broth containing 1.5% (wt/vol) agar (A. niger) or 20% (vol/vol) V8 Juice, 0.25% (wt/vol) CaCO₃, and 1.5% (wt/vol) agar (B. cinerea and S. nodorum). To induce spore formation, 0.1-strength media were employed. Following incubation at 25°C and spore formation, spores were harvested in water, washed by centrifugation, and suspended to a turbidity equal to a McFarland standard of 0.5. A 0.1-ml portion of each test culture was added to 3 ml of melted and cooled (45°C) top agar (0.1-strength brain heart infusion broth [Difco] containing 0.7% [wt/vol] agar [BBL Microbiology Systems]). After mixing, the cell suspension was poured over the surface of agar medium made up of 0.1-strength brain heart infusion broth containing 1.5% (wt/vol) agar. For measurement of activity against B. cinerea and S. nodorum, the agar medium consisted of 20% (vol/vol) V8 Juice, 0.25% CaCO₃, and 1.5% (wt/vol) agar. After 3 h drying at room temperature, 10 μl of a culture or a culture fraction was spotted on the target organism lawn and the spots were allowed to dry. Plates were incubated at 30°C and examined every day for the appearance of zones of inhibition of microbial growth. Activity was measured as the diameter (in millimeters) of the zones of inhibition. Antibiotic activity against Tetrahymena pyriformis was detected by adding cells of strain 2.2 N to T. pyriformis strain ATCC 30202 grown in medium containing 0.5% Proteose Peptone, 0.5% tryptone, and 0.02% K₂HPO₄. Samples were removed to measure strain 2.2 N colony counts on TSB+S agar medium and T. pyriformis cells by direct microscopic counts.

The plant protective activity of strain 2.2 N against plant-pathogenic fungi was assessed by spraying three plants (5 to 10 mm tall) for each trial with fungal spore suspensions. After drying, the infected plants were sprayed with an undiluted culture of strain 2.2 N grown in TSB+S broth until thoroughly wet (approximately 2 ml). Untreated infected and uninfected plants served as controls for the four or eight independent trials (see Table 3). After incubation in humidity chambers for a time suitable for manifestation of disease, all plants were evaluated by visually estimating the level of disease control. Untreated, infected plants were given a rating of 0% disease control, and uninfected plants were assigned a rating of 100% disease control. No phytotoxicity was observed on any plant sprayed only with cultures of strain 2.2 N.

TABLE 3.

Protection of plants infected with fungi by cultures of strain 2.2 N

Plant	Fungus	No. of trials	% Disease control
Tomato	Phytophthora infestans	8	70 ± 19
Tomato	Alternaria solani	8	64 ± 28
Grape	Plasmopora viticola	8	60 ± 42
Pepper	Botrytis cinerea	8	100 ± 0
Wheat	Septoria nodorum	8	96 ± 7
Wheat	Puccinia recondita	8	63 ± 42
Banana	Mycosphaerella fijiensis	4	100 ± 0
Peanut	Cerospora arachidocola	4	91 ± 13
Peanut	Cercosporidium personatum	4	91 ± 22

The biochemical and substrate utilization patterns of strain 2.2 N were assessed using the API-NFT, API-ZYM, and API-CH50 tests (bioMerieux Vitek, Inc., Hazelwood, Mo.). Susceptibility to fusaric acid was assessed as described previously (13). Antibiotic susceptibility was measured using the Sceptor Pseudomonas/Resistant MIC Panel (Becton-Dickinson, Sparks, Nev.) following the manufacturer's directions. The cellular fatty acid profile for strain 2.2 N was generated by MIDI, Inc. (Newark, Del.). Nucleic acids were isolated (11), and the 16S rRNA gene was amplified by PCR, employing a pair of universal 16S rRNA primers: 27f (forward) and 1522r (reverse) (8). The DNA sequenced was a PCR product of genomic DNA generated by primers 27f and 1522r, and the sequence reported covered positions 27 through 1522 of the 16S rRNA gene (98%) using the Escherichia coli 16S rRNA numbering system. Two sets of primers were used to sequence the PCR product (i.e., set 1 [27f and 907r] and set 2 [704f and 1522r]), leading to a sequence overlap in the center of the gene. The 16S rRNA gene was sequenced three times in both the forward and reverse directions employing an automated DNA sequencer (University of Virginia Biomolecular Research Facility, Charlottesville, Va.). In addition, the PCR product was also cloned into pBluescript (Stratagene, La Jolla, Calif.) and sequenced using the T7 and T3 primers both by automated DNA sequencing and by using Sequenase following the manufacturer's directions (U.S. Biochemical, Inc., Cleveland, Ohio). The two different approaches were employed to reduce the frequency of sequence ambiguities. The 16S rRNA sequence obtained was aligned with related 16S rRNA sequences by using the Vector NT1 version 5.2 AlignX Program and ClustalW analysis (InforMax, Bethesda, Md.), and the dendrogram (Fig. 1) was drawn using Treeview version 1.6.1 freeware.

FIG. 1.

Relationship of Burkholderia strain 2.2 N 16S rRNA sequence to Burkholderia and Ralstonia species. Two Pseudomonas species and E. coli are shown also. The bar represents the number of nucleotide substitutions per base.

Cells of strain 2.2 N are gram negative and motile, approximately 0.5 to 1.0 μm wide by 1.5 to 3.0 μm long. Colonies on TSB+S agar medium were circular, 1 to 2 mm in diameter, amber colored, and beehive shaped after 48 h of incubation at 30°C. No fluorescent pigments were formed on TSB+S agar medium or on King's medium A or B (9). No growth factors were required. Approximately 1% of colonies on TSB+S agar were 2 mm in diameter, circular, and translucent. The predominant amber colony type gave rise to both amber colonies and translucent colonies. The translucent colonies gave rise to only other translucent colonies. Strain 2.2 N grew at 25 to 37°C but poorly at 45°C.

Strain 2.2 N was oxidase and indol negative and reduced nitrates to nitrites. Strain 2.2 N had the following enzymatic activities: acid and alkaline phosphatases, esterase (C₄), lipase, gelatinase, N-acetyl-β-glucosaminidase, N-acetyl-AS-BI-phosphohydrolase, β-galactosidase, β-glucosidase, and cystine, leucine, and valine arylamidases. The strain lacked arginine dihydrolase and urease activities. The following substrates were utilized for growth or energy: adipate, d-arabinose, l-arabinose, capric acid, cellobiose, citrate, esculin, d-fucose, galactose, gluconate, d-glucose, lactose, d-lyxose, malate, maltose, mannose, phenylacetate, sucrose, trehalose, and d-xylose. Though capable of utilizing glucose for growth, strain 2.2 N failed to ferment glucose or other carbohydrates. Compared to members of the genus Pseudomonas, Burkholderia, or Ralstonia (6, 13, 16), strain 2.2 N had a narrower range of substrates utilized as carbon or energy sources. Especially noteworthy was the inability of strain 2.2 N to utilize or form acids from glycerol, mannitol, inositol, or sorbitol, as has been reported for other Burkholderia spp. (13). In common with Burkholderia spp., strain 2.2 N was resistant to a variety of antibiotics (Table 1) and grew in the presence of 0.1% (wt/vol) fusaric acid, in contrast to Ralstonia spp. that fail to grow in fusaric acid at that concentration (13). The cellular fatty acids of strain 2.2 N distinguish that strain from other members of the genus Burkholderia. The pattern of low C_16:0 (i.e., 18%), high C_16:1 (i.e., 21%), and very high levels of C_18:1 (38%) cellular fatty acids distinguishes strain 2.2 N from other members of the genus Burkholderia (16). The presence, albeit low, of 3′OH (i.e., 3-OH C_16:0 = 3%) or cyclopropanoic fatty acids (CPA) (i.e., C₁₇ CPA = 4%, C₁₉ CPA = 2%, and 2-OH C₁₉ CPA = 1%) distinguishes strain 2.2 N from strains of the genus Ralstonia (15). Great weight should not be placed on absolute percentages of fatty acids, because fatty acid compositions of both Burkholderia and Ralstonia spp. have been shown to vary with culture age (13, 15). A dendrogram showing the relationships between the sequences of the 16S rRNA gene of strain 2.2 N (GenBank accession no. AF226727) and of other Burkholderia spp. is shown in Fig. 1. The 16S rRNA gene sequence of strain 2.2 N is similar to those of members of the genus Burkholderia and less similar to that of Ralstonia solanacearum (Fig. 1).

TABLE 1.

Antibiotic susceptibility of Burkholderia strain 2.2 N

Antibiotic(s)	MIC (μg/ml)
Cefotaxime	64
Ceftizoxime	16
Ceftazimine	8
Ceftriaxone	128
Imipenem	16
Trimethoprim-sulfamethoxazole	4/76a
Amikacin	>64
Cefoperazone	>16
Gentamicin	>16
Tetracycline	>8
Ticarcillin-clavulanic acid	>128/2
Tobramycin	>16

^aThe MIC for trimethoprim is shown before the slash, and the MIC for sulfamethoxazole is shown after the slash. 

Strain 2.2 N has a wide range of antimicrobial activity, and zones of inhibition were demonstrated against all the bacteria, yeasts, and fungi (Table 2). The antimicrobial activity of strain 2.2 N did not require the presence of living cells, because broad antimicrobial activity was demonstrated by both pasteurized and filter-sterilized, cell-free culture fractions of strain 2.2 N grown in TSB+S broth medium (Table 2). Cells of strain 2.2 N killed 98% of T. pyriformis cells after 3 days of incubation at 25°C. Microscopic examination of those infected cultures revealed that the T. pyriformis cells were filled with strain 2.2 N cells after 1 day of incubation and burst by the third day of incubation. Greenhouse measurements of protection of plants against fungal infection demonstrated significant broad-spectrum antifungal activity (Table 3). On the basis of the unique 16S rRNA sequence, cellular fatty acid profile, and cultural and biochemical characteristics of strain 2.2 N, it appears that this strain is a representative of the genus Burkholderia. The characteristics of this new member of the genus Burkholderia expand the range of antimicrobial activities demonstrated by representatives of the genus Burkholderia (1, 7). Further, the range of antimicrobial activities of this single strain is broader than ranges in individual strains of both Burkholderia (1, 7) and Pseudomonas (12) species.

TABLE 2.

Antimicrobial activities of cultures, cell-free culture filtrates, and pasteurized culture fractions of strain 2.2 N

Target microbe	Diameter of zone of inhibition (mean ± SD) (no. of trials)
	Culture	Cell-free filtrate	Pasteurized
Micrococcus luteus	18 ± 3 (400)	12 ± 5 (6)	9 ± 3 (4)
Mycobacterium smegmatis	21 ± 3 (13)	NDa	ND
Saccharomyces cerevisiae	17 ± 3 (519)	10 ± 3 (18)	11 ± 3 (57)
Candida albicans	15 ± 3 (12)	8 ± 2 (6)	10 ± 1 (4)
Cryptococcus neoformans	20 ± 4 (16)	9 ± 3 (5)	11 ± 1 (3)
Aspergillus niger	22 ± 5 (36)	13 ± 5 (42)	12 ± 5 (5)
Botrytis cinerea	37 ± 1 (4)	ND	ND
Septoria nodorum	27 ± 6 (5)	ND	ND

^aND, not done.