role of a phenazine antibiotic from pseudomonas ... · bacteria and pathogens in the rhizosphere is...

Vol. 170, No. 8JOURNAL OF BACTERIOLOGY, Aug. 1988, P. 3499-35080021-9193/88/083499-10$02.00/0Copyright © 1988, American Society for Microbiology

Role of a Phenazine Antibiotic from Pseudomonas fluorescens inBiological Control of Gaeumannomyces graminis var. tritici

LINDA S. THOMASHOW* AND DAVID M. WELLERAgricultural Research Service, U.S. Department of Agriculture, Root Disease and Biological Control Research Unit,

Washington State University, Pullman, Washington 99164-6430

Received 18 February 1988/Accepted 9 May 1988

Pseudomonasfluorescens 2-79 (NRRL B-15132) and its rifampin-resistant derivative 2-79RN10 are suppres-sive to take-all, a major root disease of wheat caused by Gaeumannomyces graminis var. tritici. Strain 2-79produces the antibiotic phenazine-l-carboxylate, which is active in vitro against G. graminis var. tritici andother fungal root pathogens. Mutants defective in phenazine synthesis (Phz-) were generated by TnS insertionand then compared with the parental strain to determine the importance of the antibiotic in take-allsuppression on wheat roots. Six independent, prototrophic Phz- mutants were noninhibitory to G. graminisvar. tritici in vitro and provided significantly less control of take-all than strain 2-79 on wheat seedlings.Antibiotic synthesis, fungal inhibition in vitro, and suppression of take-all on wheat were coordinately restoredin two mutants complemented with cloned DNA from a 2-79 genomic library. These mutants contained TnSinsertions in adjacent EcoRI fragments in the 2-79 genome, and the restriction maps of the region flanking theinsertions and the complementary DNA were colinear. These results indicate that sequences required forphenazine production were present in the cloned DNA and support the importance of the phenazine antibioticin disease suppression in the rhizosphere.

Certain strains of fluorescent pseudomonads, when ap-plied to planting material or soil, can provide biologicalcontrol of root pathogens (1, 7, 13, 42, 44, 45, 49). Thesebeneficial pseudomonads colonize the root system and pro-tect it against pathogens, thereby improving plant growthand yield (7, 13, 44, 45, 48). Both fluorescent siderophoresactive in iron chelation (reviewed in references 1, 30, 36, and42) and antibiotics (12, 18, 21, 22, 24) produced by thepseudomonads have been implicated as important mecha-nisms mediating pathogen suppression. Substantial evidencehas accumulated to support the role of siderophores indisease suppression (2, 3, 27, 28, 41, 45, 51, 58). However,the importance of antibiotics in these interactions betweenbacteria and pathogens in the rhizosphere is much lessclearly established, partly because antibiotics never havebeen detected directly in natural rhizosphere soil (56, 57).

Support for the role of antibiotics in biological control hascome mainly from studies that have shown correlationsbetween bacterial inhibition of pathogens in vitro and dis-ease suppression in the soil. The evidence has sometimesbeen strengthened by additional biochemical or genetic data.For example, purified antibiotics identified as pyrrolnitrin(21), pyoluteorin (22), and geldanamycin (40) were used tocorrelate the activity spectrum of each antibiotic in vitrowith control of specific target pathogens by the respectiveantibiotic-producing strains and to demonstrate that theantibiotics themselves were sufficient to limit disease. Theamount of geldanamycin produced by Streptomyces hygro-scopicus in sterile soil was sufficient to be detected directly,although it is unclear whether this would have been the casein natural soils where antibiotic stability and persistencegenerally are much lower (reviewed in reference 56). Studiesinvolving mutants defective in antibiotic production alsohave supported the role of antibiotics in disease suppression

* Corresponding author.

(12), especially when the chemical nature of the antibiotic isnot known.Pseudomonas fluorescens 2-79 (NRRL B-15132), isolated

originally from the rhizosphere of wheat, has been shown(50, 52) to suppress take-all, a major root and crown diseaseof wheat and barley (47) caused by the fungal pathogenGaeumannomyces graminis var. tritici. At least part of thissuppressiveness is correlated with fungal inhibition in vitro,since mutants of strain 2-79 defective in inhibition are alsoreduced in suppressiveness on plants (13; D. M. Weller,W. J. Howie, and R. J. Cook, Phytopathology, in press). Anantibiotic isolated from strain 2-79, phenazine-1-car-boxylate, is active in vitro against G. graminis var. tritici andseveral other wheat root pathogens at concentrations as lowas 1 ,ug/ml (6, 17). The purpose of this study was todetermine the importance of the phenazine antibiotic insuppression of G. graminis var. tritici by 2-79 on wheatroots. Transposon mutagenesis was used to generate mu-tants defective in phenazine synthesis, and the wild-type andmutant derivatives then were compared for their ability tosuppress take-all.

MATERIALS AND METHODS

Organisms, plasmids, and culture conditions. The bacterialstrains and plasmids used in this study are shown in Table 1.P. fluorescens 2-79 and its rifampin-resistant derivative2-79RN10 have been described previously (50, 52). P. fluo-rescens strains were grown at 20 or 28°C on nutrient brothyeast extract (NBY) medium (15) or on the defined mediumof Kanner et al. (KM medium) (26). Escherichia coli strainswere grown at 37°C on Luria-Bertani (LB) broth or plates(33) containing 5 g of NaCl per liter. E. coli JM109 wasgrown and maintained on AB minimal medium (43) supple-mented with thiamine (1 mg/liter). Unless otherwise noted,antibiotics were used when appropriate at the followingconcentrations (micrograms per milliliter): rifampin, 100;

3499

on April 23, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/

3500 THOMASHOW AND WELLER

TABLE 1. Bacterial strains and plasmids

Strain or plasmid Relevant characteristicsa Source or reference

P. fluorescens2-79 (NRRL B-15132) Phz+ RifP This laboratory (50, 52)2-79 RN1o Phz+ Rif' (spontaneous) This laboratory2-79-B46 2-79RN1O::Tn5 Phz- This study2-79-782 2-79RN1O::Tn5 Phz- This study2-79-99 2-79RN1O::Tn5 Phz- This study2-79-2510 2-79RN1O::TnS Phz- This study2-79-21455 2-79RN1O::Tn5 Phz- This study2-79-42355 2-79RN1O::Tn5 Phz- This study2-79-10147 2-79RN1O::Tn Phz reduced This study2-79-611 2-79RN10 Phz- This study2-79-B46R 2-79-B46 Phz+ recombinant This study2-79-782R 2-79-782 Phz+ recombinant This study

E. coliWA803 met thi supE hsdR hsdM 46JM109 recAl endAl gyrA96 thi hsdRI7 supE44 relAI A-X (lac-proAB) (F' 59

traD36 proAB lacIqZAAM5)DHSa F- endAl hsdR17 supE44 thi-l gyrA96 relAl A(argF-lacZYA)U169 Bethesda Research Laboratories

4.80dlacZAM15 -

PlasmidspGS9 Cmr Kmr repl5A TraN Tn5 46pLAFR3 IncP Tcr cos+ rlx+ 38pRK2013 IncP Kmr TraRK2+ ArepRK2 repEI 16pSUP104 Cmr Tcr 39pCU101 Cmr repi5A TraN 46pRZ102 ColEl TnS 25pIC19R Apr 34pPHZ49-1 pLAFR3 containing P. fluorescens 2-79RN10 genomic DNA, Tcr This studypPHZ3-384 pLAFR3 containing P. fluorescens 2-79RNIO genomic DNA, Tcr This studypPHZ3-304 pLAFR3 containing P. fluorescens 2-79RN1O genomic DNA, Tcr This studypPHZ102-55 pLAFR3 containing P. fluorescens 2-79RNIO genomic DNA, Tcr This studypPHZ49-6 pSUP104 containing the 7.8-kb BamHI fragnent from pPHZ49-1, Tcr This studya Apr, Cmr, Kmr, Rif, and Tcr indicate resistance to ampicillin, chloramphenicol, kanamycin, rifampin, and tetracycline, respectively. Phz-, Deficient in

phenazine production.

cycloheximide, 100; nalidixic acid, 100; tetracycline, 25;ampicillin, 100; chloramphenicol, 30; kanamycin, 25 or 100for E. coli or P. fluorescens, respectively.The isolate of G. graminis var. tritici used in this study

originated from a single ascospore and was maintained at 4°Con diluted potato dextrose agar (PDA) prepared as describedpreviously (15). The oat-kernel inoculum of the fungus usedto infest soil was prepared as described previously (55).

Fungal inhibition assay. Inhibition of G. graminis var.

tritici by bacterial strains in vitro was assayed on KM agarplates prepared in diluted potato broth (40 g of diced potatoper liter; 15) instead of H20 (KM-PDA). Samples (5 ,ul,containing approximately 5 x 106 cells) from overnightcultures of strain 2-79RN10 or its derivatives (Table 1) in KMbroth were spotted 1 cm from the edge of petri plates andallowed to soak into the agar. A 0.5-cm plug from the leadingedge of a culture of G. graminis var. tritici grown for 5 daysat 28°C on KM-PDA was placed in the center of the plate.Plates were incubated at 28°C and scored after 4 or 5 days bymeasuring the distance between the edges of the bacterialcolony and the fungal mycelium. Inhibition was expressedrelative to a control strain [2-79RN10 or 2-79RN1O(pSUP104)]spotted on the same plate.Root colonization assay. Surface-sterilized wheat seeds

(cultivar Fielder or Daws) were pregerminated for 20 h andmixed with bacteria scraped from King medium B (KMB)(15) agar plates and suspended in 0.5% methylceliulose (50)to give a final concentration of approximately 1 x 108 to 2 x

108 CFU per seed. The control was treated with onlymethylcellulose. Conical plastic tubes (20.7 by 4 cm [topdiameter]; Ray Leach Cone-tainer, Canby, Ore.) were filledwith about 100 g of natural or steamed (for 1 h) soil (Thatunaor Palouse silt loam) adjusted to a matric potential of about-0.5 bars, which is optimal for root colonization by 2-79RN10 (23). One seed was sown in each tube and coveredwith 1 cm of sterile vermiculite. The rack holding the tubeswas then enclosed in a clear plastic bag to retard moistureloss and incubated at 15 to 18°C for 11 to 13 days in a growthchamber with a 12-h photoperiod. During that time the plantswere not watered, but the matric potential of the soildecreased only slightly. The plants then were removed fromthe tubes and shaken gently to remove all but the mosttightly adhering rhizosphere soil. The two longest seminalroots of each plant were sectioned, and the two root seg-ments from the 1- to 3-, 3- to 5-, or 5- to 7-cm sections belowthe seed were combined and sonicated together for 20 s in 10ml of 1 M phosphate buffer (37) to remove bacteria from theroots. The washings were diluted and plated on KMB agarcontaining rifampin and cycloheximide. Each bacterial seedtreatment was replicated at least seven times per assay, witha single seedling serving as a replicate. A randomizedcomplete block design was used for each experiment.

Take-afl suppression assay. The tube assay (52) was used todetermine the ability of the bacteria to protect wheat againsttake-all. Conical plastic tubes (16.5 by 2.5 cm [top diameter])were filled with sterile vermiculite to a depth of 6.5 cm,

J. BACTERIOL.


http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/

ROLE OF A PHENAZINE ANTIBIOTIC 3501

followed by 5 g of soil (Shano silt loam) that had beensteamed (for 1 h), air dried, and amended with 0.45% (wt/wt)G. graminis var. tritici as 0.25- to 0.5-mm particles ofpulverized, colonized oat kernels (55). Two seeds treatedwith bacteria (50) in 0.5% methylcellulose (1 x 108 to 2 x 108CFU per seed) were sown per tube, covered with 1 cm ofsterile vermiculite, and watered. The control was treatedwith only methylcellulose. Racks of tubes were enclosed inplastic for 4 days and then uncovered and incubated asdescribed above. Each tube was given 5 ml of dilute Hoag-land solution (52) twice weekly. After 3 to 4 weeks theseedlings were washed and evaluated for severity of take-allon a scale of 0 through 8, where 0 indicates no detectablesymptoms and 8 indicates dead or nearly dead. This scale isa modification of one reported earlier (52). In each experi-ment, treatments were replicated four or five times with eachreplicate consisting of 10 tubes (20 seedlings). A randomizedcomplete block design was always used.Transposon TnS mutagenesis of strain 2-79RN10 and mutant

characterization. The suicide plasmid pGS9 in E. coli WA803(46) was used to generate TnS insertions in strain 2-79RN10.E. coli WA803(pGS9) was sensitive to the phenazine antibi-otic at concentrations normally present in agar plates onwhich 2-79RN10 was grown (L. S. Thomashow, unpublisheddata). A phenazine-resistant spontaneous mutant ofWA803(pGS9) was selected by spreading an overnight cul-ture on NBY plates on which 2-79RN10 previously had beengrown on sterile nitrocellulose filters for 48 h. This phena-zine-resistant donor strain was grown overnight with aera-tion in LB broth to a density of 1 x 109 to 5 x 109 CFU/ml.Recipient cultures were grown at 20°C to the stationaryphase (1 x 109 to 5 x 109 CFU/ml) by inoculating NBY brothfrom a glycerol stock maintained at -20°C and shaking theculture for 16 to 18 h. Approximately 109 donor cells weresedimented, suspended in 50 ,il of H20, transferred to anitrocellulose filter on a fresh, prewarmed LB agar plate, andincubated for 1.5 h at 37°C. Approximately 109 cells of2-79RN10 were sedimented in an Eppendorf tube, washedbriefly in TE buffer (10 mM Tris hydrochloride [pH 8.0], 1mM EDTA), suspended in 50 Rl of H20, and transferred tothe filters containing the donor cells. The plates were incu-bated for 24 h at 28°C. Transconjugants were selected onNBY plates containing rifampin and kanamycin at 50 and 100pug/ml, respectively. The age of the recipient culture wascritical: significantly higher frequencies of transconjugantswere obtained with stationary-phase rather than logarithmi-cally growing cells, but when the cells were in the late-stationary phase substantial lysis occurred during the TEwash. Conjugation frequency was not increased by washingat 0°C or by including 1 M NaCl or 0.4 M sucrose in the TEbuffer, was about twofold lower when recipient cells werewashed in 1 mM EDTA alone, and was lower by a factor of101 to 102 when recipients were washed in H20, 10 mM Trishydrochloride (pH 8.0) or 25 mM MgCl2 or CaCl2 (Tho-mashow, unpublished data).

Transconjugants were replicated onto KM agar to test forauxotrophy. Specific nutritional requirements were identi-fied as described previously (19). Fluorescence was deter-mined on KMB agar, and phenazine antibiotic productionwas scored visually (the antibiotic is pigmented; see below)on either homemade PDA (200 g of potato, 15 g of dextrose,and 15 g of agar per liter) prepared as described previously(15), commercial PDA (Difco Laboratories, Detroit, Mich.),or KM-PDA.Measurement of phenazine production. Phenazine antibi-

otic production in vitro was assayed in broth cultures grown

for 24 h at 20°C in KM broth or NBY broth supplementedwith glucose to a final concentration of 2%. Samples ofcultures were extracted twice with equal volumes of ben-zene. Absorbance of the pooled extracts was determinedspectrophotometrically at 349.5 nm, and readings were con-verted to micrograms of phenazine-1-carboxylate from astandard curve prepared with the purified antibiotic (17).Antibiotic concentrations were expressed relative to thetotal protein of the producing cultures, assayed by theBradford method (5). Phenazine antibiotic production onplates was indicated by colony pigmentation and the pres-ence of a dark zone in the medium surrounding producingcolonies examined under longwave (365-nm) UV irradiation.As little as 0.25 to 0.5 ,ug of purified antibiotic (17) spotted onplates was detectable by this method.Recombinant DNA techniques. Plasmid DNA was isolated

from E. coli and P. fluorescens as described by Birnboim andDoly (4) or by a scaled-up modification that included equi-librium centrifugation in cesium chloride-ethidium bromide.Genomic DNA was prepared from cells lysed as describedby Currier and Nester (14). Restriction enzyme digestionsand ligation reactions were performed as recommended bythe suppliers (New England BioLabs, Inc., or BethesdaResearch Laboratories, Inc.). Standard procedures wereused for preparation and transformation of E. coli, electro-phoresis, restriction mapping, nick translation, Southerntransfer, and nitrocellulose filter hybridization (33).

Library construction. Total DNA from strain 2-79RN10was partially digested with EcoRI and size fractionated byagarose gel electrophoresis. Fragments 25 to 35 kilobases(kb) in size were isolated from gels and ligated into thecosmid vector pLAFR3 (38), which had been linearizedpreviously with EcoRI. Recombinant molecules were pack-aged in vitro with Packagene as recommended by the sup-plier (Promega Biotec Co., Madison, Wis.) and transducedinto E. coli JM109. Insert frequency was estimated from thenumber of white Tcr transductants on LB agar containingthe chromogenic substrate 5-bromo-4-chloro-3-indolyl-P-D-galactoside and isopropyl-p-D-thiogalactopyranoside (JM109transductants containing cosmids without inserts are bluedue to a complementation of P-galactosidase activity). Trans-ductants with recombinant plasmids were stored at -20°C inmicrotiter dishes.Complementation of mutants. Cosmid clones were mobi-

lized into phenazine-nonproducing mutants of 2-79RN10 intriparental matings with pRK2013 (16) as the helper plasmid.Matings were performed either on filters (see above) withdonors individually or in pools of five to seven or by amodification of the replicator technique described by Ma-rugg et al. (35). Briefly, individual donors were transferredfrom LB agar plates by a replicator with metal prongs toplates of LB agar preseeded with a mixture of approximately107 HB1O1(pRK2013) and 108 pseudomonad recipient cells.After 48 h transconjugants were replicated to KM-PDA andscored visually for phenazine production. In both cases therecipient cells were washed briefly in TE buffer beforemating.

Construction of pPHZ49-6. Because strain 2-79RN10 hasnatural chloramphenicol resistance, the following series ofmanipulations was carried out to clone the 7.8-kb BamHIfragment from pPHZ49-1 into the Cmr region of pSUP104such that the Tcr marker in the vector was preserved.Cleavage of pPHZ49-1 with BamHI resulted in 2-79 DNAfragments of 12.0, 7.8, and 4.2 kb. These were ligated intothe BamHI site in pIC19R (34) and transformed into E. coliDHSa. Plasmid DNA isolated from a clone containing the

VOL. 170, 1988


http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/


- ^ _DIDFIG. 1. Fungal inhibition assay. Plates were prepared as described in the text. On each plate the bacteria at the left and right were

2-79RN10. (A) Tn5 mutants: top, 2-79-B46; bottom, 2-79-782. (B) Complemented mutants: top, 2-79-B46 (pPHZ49-1); bottom, 2-79-782(pPHZ49-1). (C) Recombinationally restored strains: top, 2-79-B46R; bottom, 2-79-782R.

7.8-kb BamHI fragment of pPHZ49-1 in pIC19R was par-tially digested with EcoRl. Fragments of approximately 7.8kb were recovered from an agarose gel, ligated into theEcoRI site of pSUP104 (39), and transformed into E. coliDH5a. Transconjugants selected as Tcr were screened forAps (absence of pIC19R sequences) and Cm' (insertion intothe EcoRI site). Plasmid DNA isolated from Tcr Aps Cmstransformants was examined by restriction enzyme analysis.One clone, pPHZ49-6, contained the 7.8-kb BamHI fragment(flanked by EcoRI sites from the pIC19R polylinker) insertedinto the EcoRI site of pSUP104. pPHZ49-6 can be mobilizedinto strain 2-79RN10, and transconjugants can be selected asTcr.

RESULTS

Isolation and characterization of phenazine mutants. Mu-tants defective in phenazine antibiotic production (Phz-)were obtained by TnS mutagenesis. Kmr transconjugants ofP. fluorescens 2-79RN10 were recovered at a frequency of 3X 10-7 per initial recipient. Of these, 0.3% of 4,360 trans-conjugants from nine separate matings were auxotrophs witha variety of different nutritional requirements. Spontaneousreversion to prototrophy, coincident with restoration ofsensitivity to kanamycin, occurred at frequencies of lessthan 108, indicating that the auxotrophic phenotype waslinked to the TnS insertion and that TnS insertions in2-79RN10 were relatively stable. An additional 0.25% of TnSmutants were nonfluorescent on KMB, indicating loss ofability to produce a fluorescent siderophore.The Kmr transconjugants were screened visually on agar

for the ability to produce phenazine antibiotic. Coloniesproducing the phenazine contained dark blackish-green"birdshot" aggregates and deposited yellow crystals in theunderlying agar on PDA, the medium initially used to detectantibiotic synthesis. However, the presence, color, andamount of phenazine produced varied greatly with the age ofthe cultures and media, the thickness of the agar, the sourceof potato extract (commercial or homemade), the distance ofthe colony from the edge of the plate, and other undefinedconditions. The antibiotic was produced consistently andcould be screened reliably on KM medium, on which colo-nies producing the phenazine appeared yellow after 4 days.On plates of KM supplemented with potato extract (KM-PDA), finely dispersed dark green granules were present

within the colonies after 6 to 7 days. The antibiotic also wasdetectable in amounts as small as 0.25 to 0.5 ,ug in the agarsurrounding colonies examined under longwave UV irradia-tion.Fewer than 0.3% of transconjugants examined were al-

tered in phenazine production. The majority of these stillproduced the antibiotic but in reduced amounts as estimatedfrom the intensity and diameter of the UV-absorbing shadowzone surrounding colonies on KM agar. Six prototrophicTnS mutants unable to produce detectable phenazine wereisolated from seven independent matings, resulting in morethan 6,000 transconjugants. These mutants failed to inhibitgrowth of G. graminis var. tritici in vitro (Fig. 1, Table 2).No mutants were found with apparent qualitative changes inpigmentation that might indicate the accumulation of otherphenazines, possibly biosynthetic intermediates.

Role of the phenazine antibiotic in suppression of take-all.To determine the importance of the phenazine antibiotic insuppression of take-all, 2-79RN10 and Phz- mutant deriva-tives were compared in the tube assay for their ability tocontrol the disease on wheat seedlings. Plants grown fromseed treated with Phz- mutants had significantly moretake-all, as indicated by increased numbers of root lesionsand reduced shoot heights, than plants grown from seedtreated with the parental strain (Fig. 2, Table 2). Althoughdisease severity varied among experiments, the diseaseratings of mutant-treated seedlings were generally interme-diate between those of the parental strain (least disease) andthe methylcellulose control (most disease). These resultssuggest that the phenazine antibiotic is a major determinantbut not the sole factor responsible for suppression of G.graminis var. tritici by strain 2-79RN10.Root colonization by Phz- mutants. Since the reduced

suppressiveness of Phz- mutants could be due to theirinability to establish and/or maintain significant populationson the roots, studies were carried out to evaluate thecompetitiveness of Phz- mutants in the rhizosphere. Com-parable numbers of 2-79RN10 and the Phz- mutants wererecovered from colonized segments along the length of rootsregardless of whether the assays were conducted in steamedor natural soil (Table 3). These results indicate that themutants do not differ significantly from the parental strain inability to colonize wheat roots even when in competitionwith the normal soil microflora (natural soil).

J. BACTERIOL.


http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/


TABLE 2. Effect of P. fluorescens 2-79RN10 phenazine antibiotic on inhibition ofG. graminis var. tritici and suppression of take-alla

Strain Phenazine Relative Root disease in expt:production' inhibitionc 1 2 3 4

2-79RN10 + 1.0 4.89D 3.66D 3.80C 5.95B2-79-10147 + 0 5.11D 4.20C2-79-2510 - 0 4.20C2-79-21455 - 0 5.71C 4.41BC2-79-99 - 0 5.94BC 4.29BC2-79-611 - 0 6.25B 4.54BC2-79-B46 - 0 6.11BC 4.56B 5.99B2-79-782 - 0 6.15BC 4.28BC 5.88B 6.39A2-79-B46R + 0.89 3.75C2-79-782R + 0.93 5.75BControl (no bacteria) 7.19A 4.91A 6.58A 6.44A

a Bacteria were tested in vitro for inhibition of G. graminis var. tritici or applied at 1 x 108 to 2 x 108 CFU per seed and tested for suppression of take-all asdescribed in Materials and Methods.

b Presence of the antibiotic was detected by UV absorbance on KM-PDA medium as described in the text. Symbols represent easily detectable (+), barelydetectable (±), or not detectable (-) UV absorbance in the medium surrounding regions of heavy bacterial growth.

c Inhibition of fungal growth was measured on KM-PDA as described in the text. The values are the means from three plates with the bacteria spotted twiceper plate.

d Tube bioassays were conducted as described in the text. In experiments 1 and 2 treatments were replicated four times, and in experiments 3 and 4 treatmentswere replicated five times, with 10 tubes per replication. Disease was rated on a 0 to 8 scale; 0 indicates no disease and 8 indicates that the plant was nearly orcompletely dead. After a significant F test, a least significant difference analysis was performed. Means in the same column followed by the same capital letterare not significantly different by least significant difference analysis (P = 0.05). Least significant difference values for experiments 1 through 4 were 0.50, 0.35,0.30, and 0.24, respectively.

Physical characterization of TI5 insertions. The site of TnSinsertion in total DNA from each mutant was determined bySouthern blot analysis with 32P-labeled pRZ102 as the hy-bridization probe. Only one EcoRl fragment with homologyto the probe was present in each mutant (Fig. 3A) except for2-79-611, in which homology to TnS was not detected (datanot shown). None of the mutants contained sequenceshomologous to 32P-labeled pCU101 (data not shown), thevector from which pRZ102 was derived (46), indicating thatthe vector itself was not inserted in the mutants. Among thesix Phz- mutants, 2-79-99 and 2-79-21455 both containedTnS in 16.7-kb EcoRI fragments, and the remaining fourmutants had insertions into unique EcoRI fragments rangingfrom 0.6 to 12.2 kb in size (Table 4, Fig. 3A). Plasmids havenot been detected in strain 2-79RN10 (D. S. Heron, L. S.Pierson, and L. S. Thomashow, unpublished data), and it isassumed that these fragments are of chromosomal origin.Additional restriction mapping of the genomic sequencesflanking the insertion sites in 2-79-99 and 2-79-21455 withHindIII, BamHI, BglII, and KpnI suggested that thesemutants contain Tn5 in distinct EcoRI fragments of the sameapparent size (data not shown).Complementation of phenazine antibiotic production. A

total of 1,021 EcoRI cosmid clones containing 2-79RN10genomic DNA averaging 30 kb in size were screened bycomplementation assays to identify clones that could restorethe ability to produce the phenazine antibiotic. Three cos-mids, pPHZ49-1, pPHZ3-304, and pPHZ102-55 comple-mented the pigmented phenotype and restored fungal inhi-bition fully in mutants 2-79-B46 and 2-79-782 and partially inmutants 2-79-99 and 2-79-2510. Representative data for onecosmid, pPHZ49-1, are shown in Table 4. A third plasmid,pPHZ3-384, partially restored phenazine production andfungal inhibition only to 2-79-21455 when the strain wasgrown on agar plates. Relatively little antibiotic was pro-duced in the broth cultures used to measure phenazineproduction (Table 4). Qualitative and quantitative differ-ences in phenazine production as a function of growthconditions, including aeration, have been reported previ-ously for pseudomonads (17, 20, 26). Mutant 2-79-42355

... ......... . .

*~~ ~ ~ ~ ~~~~~~~~~~~~~~~~~~~

FIG. 2. Suppression of take-all by phenazine-producing and-nonproducing bacterial strains. Seedlings are from experiment 3 ofTable 2 and were grown from nontreated seed (A) or seed treatedwith 2-79-B46 (B), 2-79-B46R (C), or 2-79RN10 (D).

VOL. 170, 1988


http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/


TABLE 3. Effect of loss of phenazine antibiotic production on colonization of wheat roots in soil by P. fluorescens 2-79 strains

Mean population (log CFU/cm of root)c in expt:

Strain' Phenazine 1 3 4production' 2 (3-5 cm)3-5 cm 5-7 cm 3-5 cm 5-7 cm 1-3 cm 3-5 cm

2-79 + 2.97 2.55 2.50 3.10 1.98 4.81 3.132-79-10147 + 2.45 2.06 2.88 4.76 3.602-79-21455 - 2.96 2.20 2.64 4.23 2.182-79-99 - 2.86 2.57 3.03 4.05 2.282-79-611 - 2.70 2.34 3.07 3.87 2.552-79-782 - 2.31 2.11 2.02 3.51 2.33 4.27 3.212-79-B46 - 2.87 1.83 3.93 2.622-79-42355 - 4.54 3.312-79-2510 - 4.57 3.242-79-782R + 3.71 2.56 4.21 2.972-79-B46R + 3.80 2.18 4.28 3.08

a Bacteria were applied at approximately 1 x 108 to 2 x 108 CFU per seed.b See footnote b of Table 2.Experiment 1 used Palouse silt loam (pH 4.1) and Daws wheat; experiments 2 and 3 used Thatuna silt loam (pH 5.3) and Fielder wheat; experiment 4 used

Thatuna silt loam (steamed) and Fielder wheat. Treatments were replicated 14 times in experiment 1, 12 times in experiment 2, 21 times in experiment 3, and 7times in experiment 4. Sections of seminal roots 1 to 3, 3 to 5, or 5 to 7 cm below the seed were sampled 11 to 13 days after planting. Means were calculated afterlog transformation (base 10) of the population in each replicate. Means (for each segment) within each experiment are not significantly different (P = 0.05) basedon an F test.

conjugated poorly and was not complemented. Among thecomplemented mutants, the plasmids that restored phena-zine production were consistent and reproducible in thedegree of restoration they conferred and could be recoveredwithout apparent structural rearrangement from the comple-mented transconjugants.The cosmid vector pLAFR3 and its derivative plasmids

were not stably maintained in P. fluorescens 2-79RNIO in theabsence of tetracycline selection. Thus, to evaluate theability of the complemented mutants to suppress take-all onwheat, it was necessary to recombine the cosmid-bornesequences required for phenazine production back into thegenomes of the various restored mutants. Restoration of thewild-type phenotype by recombination, concomitant withloss of the TnS Kmr marker, is also indicative of homology

between the complementing cosmid and the TnS-containingsequences, since recombination can occur only when thewild-type allele of the mutated gene has been cloned. Of thevarious mutant-plasmid combinations for which fungal inhi-bition was restored in trans (Table 4), restoration of thepigmented phenotype by homologous recombination oc-curred only in mutants 2-79-B46 and 2-79-782 complementedwith cosmid pPHZ49-1, pPHZ3-304, or pPHZ102-55. Re-stored strains 2-79-B46R and 2-79-782R, obtained by recom-bination with pPHZ49-1, were Tcs and Kms, consistent withthe loss of pLAFR3 and the Tn5, respectively. Both strainsproduced wild-type levels of the phenazine antibiotic, werefully inhibitory to G. graminis var. tritici in vitro (Fig. 1,Table 4), colonized roots efficiently (Table 3), and were aseffective as the parental 2-79RN10 in suppression of take-all

FIG. 3. Southern blot analysis of sequences homologous to Tn5 (A), the 7.8-kb BamHl fragment from pPHZ49-6 (B), or both pro'bes (C)in DNA from P. fluorescens 2-79RN10. EcoRI digests of pPHZ49-1 (lane 2) or total DNA from 2-79RN10 (lane 1), 2-79-B46 (lane 3), 2-79-B46R(lane 4), 2-79-782 (lane 5), 2-79-782R (lane 6), 2-79-99 (lane 7), 2-79-21455 (lane 8), 2-79-42355 (lane 9), or 2-79-2510 (lane 10) wereelectrophoresed for 17 h at 22 V in 0.75% agarose, transferred to nitrocellulose, and hybridized with 12P-labeled pRZ1O2 (A), pPHZ49-6 (B),or both probes (C). Fragment sizes are indicated in kilobases. The two probes hybridized to different fragments in DNA from four of themutants (lanes 7 through 10), indicating that Tn5 was present in fragments other than those homologous to pPHZ49-6. 2-79-42355 (lane 9)contained the TnS5 in an EcoRI fragment of 12 kb (A) that comigrated with the 12.2-kb fragment homologous to pPHZ49-6 (C). Both probeshybridized to the same fragments in 2-79-B46 and 2-79-782 (lanes 3 and 5), but in B46R and 782R (lanes 4 and 6) the fragments homologousto pPHZ49-6 were reduced to the size of those from the parental strain (B, lane 1) and TnS was no longer present (A, lanes 4 and 6).

J. BACTERIOL.


http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/


TABLE 4. Phenazine production and relative inhibition ofG. graminis var. tritici by restored strains and phenazine

mutants containing recombinant plasmids

TnS target PhenazineStrain fragment production Relative

)a (~jg/mg of inhibition'(kb)a cell protein)b

2-79RN10 52 1.0

2-79-B46 0.6 ND 02-79-B46(pPHZ49-1) 29 0.982-79-B46(pPHZ49-6) 29 1.022-79-B46R 52 0.89

2-79-782 12.2 ND 02-79-782(pPHZ49-1) 27 1.012-79-782(pPHZ49-6) 29 0.942-79-782R 42 0.93

2-79-99 16.7 ND 02-79-99(pPHZ49-1) 13 0.362-79-99(pPHZ49-6) 0.48

2-79-2510 3.4 ND 02-79-2510(pPHZ49-1) 4 0.402-79-2510(pPHZ49-6) 0.58

2-79-21455 16.7 ND 02-79-21455(pPHZ3-384) <1 0.46

2-79-42355 6.5 ND 0a Sizes of EcoRI fragments in which TnS was inserted were determined

from Southern blots hybridized with pRZ102.b The amount of phenazine antibiotic produced was determined for cultures

grown and extracted as described in the text. Values are expressed relative tototal cell protein present in washed cell suspensions from the same culturesand represent the means of three determinations. ND, Not detectable.

c Inhibition of fungal growth was measured on KM-PDA plates as describedin the text. The controls were 2-79RNIO for mutant and restored strains and2-79RN1O(pSUP104) for plasmid-containing strains. The values are the meansfrom three plates with the bacteria spotted twice per plate.

(Table 2, Fig. 2). The coordinate restoration of phenazineantibiotic production and disease suppressiveness in thesestrains further supports the importance of the antibiotic as amechanism of take-all suppression.Mutants 2-79-99, 2-79-2510, and 2-79-21455 did not lose

resistance to kanamycin, nor were they restored to thepigmented phenotype by homologous recombination withany of the cosmids that complemented antibiotic productionin trans. These results indicate that the sequences present inthe complementing cosmids do not correspond to thosecontaining TnS in the three mutants.

A K H B RRi . ..

Structural analysis of pPHZ49-1 and corresponding ge-nomic regions in 2-79RNIO and Tn5 mutants. Restrictionmaps were constructed to verify that the cloned sequences inpPHZ49-1 were homologous to the genomic loci bearing Tn5insertions in 2-79-B46 and 2-79-782, as was indicated byrestoration of phenazine production by recombination inthese mutants. The restriction map of the cloned insert inpPHZ49-1 was colinear with a map of genomic DNA from2-79-B46 and 2-79-782 deduced from Southern blot analysesin which 32P-labeled pRZ102 was the hybridization probe(Fig. 4). The alignment of restriction sites in the genomicsequences flanking the sites of TnS insertion in the mutantsindicated that the 0.6- and 12.2-kb fragments into which TnShad inserted in 2-79-B46 and 2-79-782, respectively, werecontiguous in the genomic DNA as well as in the clonedparental DNA. Cosmids pPHZ3-304 and pPHZ102-55 lackedthe leftmost 7.9-kb EcoRI fragment of pPHZ49-1, but con-tained the 0.6- and 12.2-kb fragments, which presumablyaccounts for the ability of these cosmids to restore phena-zine production in 2-79-B46 and 2-79-782 after recombina-tion. There was no homology between the, cloned 2-79RN10sequences in the recombinant plasmids pPHZ49-1 andpPHZ3-384 (data not shown).Mutants 2-79-99 and 2-79-2510, which were only partially

complemented by pPHZ49-1 but which could not be restoredby homologous recombination, contained Tn5 in EcoRIfragments of 16.7 and 3.4 kb, respectively (Fig. 3A). NoEcoRI fragments of these sizes were present in pPHZ49-1,and the restriction maps of the genomic sequences flankingthe TnS insertion sites in the mutants did not align with theplasmid map. Similarly, there was no apparent alignment ofrestriction sites in the maps of pPHZ384-1 and the genomicregion flanking TnS in 2-79-21455 (data not shown). Southernblot analyses in which 32P-labeled pPHZ49-6 (which containsan internal 7.8-kb BamHI fragment cloned from pPHZ49-1)was hybridized with DNA from 2-79-99 and 2-79-2510 alsodemonstrated that the mutated and complementing se-quences were nonhomologous (Fig. 3B). The cloned probehybridized with EcoRI fragments in both mutants of 0.6,10.0, and 12.2 kb, the same sizes as were present in DNAfrom wild-type 2-79RN10 and in pPHZ49-1 (Fig. 3B). Rehy-bridization of the blot with the TnS probe 32P-pRZ102 (Fig.3C) revealed additional bands of 22 and 9.2 kb in 2-79-99 and2-79-2510, respectively, which contained TnS. These blotsalso showed that bands of 6.3 and 18 kb, representing TnSinserted into EcoRI fragments of 0.6 and 12.2 kb in themutants 2-79-B46 and 2-79-782, respectively, were notpresent in the recombinationally restored strains 2-79-B46Rand 2-79-782R (Fig. 3A). Instead, these strains containedfragments of 0.6 and 12.2 kb, indistinguishable from thosepresent in DNA from the parental strain (Fig. 3B), thus

BK L R2-79RNIO

B R HK RR HB B RR BK LK KLR..,B . la , m . .,,,

C2KB

pPHZ49- 1

B RR B Pi-H------ pPHZ49-6

FIG. 4. (A) Location of TnS insertions (arrows) in the 0.6- and 12.2-kb EcoRI fragments of 2-79 DNA which resulted in loss of ability toproduce phenazines in mutants 2-79-B46 and 2-79-782, respectively. (B) Restriction site map of pPHZ49-1. (C) Subcloned BamHI fragmentfrom pPHZ49-1 present in pPHZ49-6. Both pPHZ49-1 and pPHZ49-6 complemented the mutations in 2-79-B46 and 2-79-782. Restrictionendonuclease sites: R, EcoRI; B, BamHI; H, Hindlll; K, KpnI; L, BgII. Thick lines on pPHZ49-1 represent pLAFR3.

VOL. 170, 1988


http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/


substantiating the conclusion that their genomic DNA hadbeen restored by homologous recombination to its originalconfiguration.

DISCUSSION

The results of this study indicate that TnS mutants of P.fluorescens 2-79RN10 unable to produce detectable quanti-ties of the phenazine antibiotic were no longer inhibitory toG. graminis var. tritici in vitro; more importantly, theyprovided significantly less control of take-all on wheat seed-lings (Table 2, Fig. 2). Complementation of the mutants withcloned homologous sequences from a library of wild-type2-79RN10 DNA coordinately restored antibiotic synthesis,fungal inhibition in vitro, and control of take-all on wheatroots to wild-type levels. These results indicate that thesequences containing Tn5 in strains 2-79-B46 and 2-79-782are required for phenazine synthesis and strongly supportthe importance of the phenazine antibiotic in biologicalcontrol of take-all.

Since comparable numbers of mutant and parental bacte-ria were isolated from roots (Table 3), the reduced suppres-siveness of Phz- mutants was not due to their inability toestablish and maintain effective populations in the rhizo-sphere. Studies with other biocontrol agents also havedemonstrated that loss of antibiotic or siderophore produc-tion does not affect the ability of the bacteria to colonizeroots (2, 3, 13, 29, 51), although possible long-term differ-ences in populations have not yet been assessed.

It is presently unclear whether the mutants described hereare impaired specifically in phenazine synthesis per se. Thephenazines are products of the aromatic amino acid biosyn-thetic pathway, with chorismate as the branch point inter-mediate (11, 32). Phenazine production can approach severalhundred milligrams per liter in 24 h (31; L. S. Thomashowand L. S. Pierson, unpublished data), and it is likely thatmutations in intermediary metabolism or aromatic biosyn-thesis could significantly limit chorismate availability forphenazine synthesis. Related pathways can dramaticallyinfluence antibiotic biosynthesis as illustrated in recent stud-ies (18, 24) with glucose dehydrogenase mutants of P.fluorescens HV37a that assimilated glucose by an alternativepathway but no longer produced an antibiotic effectiveagainst Pythium ultimum in vitro. The Phz- strains de-scribed in this study were prototrophs with normal growthrates in minimal media and were unimpaired in their abilityto colonize roots effectively (Table 3), indicating that thevarious introduced mutations did not detectably influencegrowth in vitro or in the rhizosphere. However, until muta-tions specifically within the phenazine pathway are charac-terized, the possibility remains that the mutants described inour study may be defective in activities other than phenazinesynthesis that could contribute to the overall suppres-siveness of strain 2-79RN10. One such factor, the fluorescentsiderophore active in other biocontrol and plant growth-promoting systems, was produced in normal amounts onKMB by all of the mutants. A recent study (Weller et al., inpress) suggests that the fluorescent siderophore produced bystrain 2-79RN10 has a role in biological control of take-all.These results suggest that the siderophore may also accountfor at least part of the residual suppressiveness exhibited byPhz- mutants in the tube assay.An unusual and somewhat surprising result of this study is

that the Phz- mutants 2-79-99 and 2-79-2510 were comple-mented in trans by sequences in pPHZ49-1 lacking detect-able homology with those flanking the sites of TnS insertion.

A possible explanation for these examples of phenotypiccompensation is that the complementing sequences encodeenzymes with functions similar or identical to those encodedby the mutated DNA. One point in the biosynthetic pathwayleading to the phenazines where this could occur is theallosteric enzyme 3-deoxy-D-arabino-heptulosonate-7-phos-phate synthase, which catalyzes the first step in the commonaromatic pathway. Pseudomonads of taxonomic group Ib,which includes P. fluorescens (10), have two isozymes ofDAHP synthase (53, 54) that control carbon flow through thepathway and represent its primary point of regulation (8, 9,54). One or both of the genetic loci in the mutants might alsoregulate phenazine production either directly or by modulat-ing the activity of peripheral pathways that affect antibioticsynthesis. A genetic analysis of phenazine production instrain 2-79RN10 involving additional mutants and clonedTnS-containing and wild-type sequences is currently under-way and should help to resolve this question.

ACKNOWLEDGMENTS

This work was supported in part by grant 86-CRCR-1-1945 underthe Competitive Research Grants Program administered by theOffice of Grants and Program Systems of the U.S. Department ofAgriculture.We thank R. J. Cook, L. S. Pierson, D. S. Heron, and B. H.

Ownley for reading the manuscript; L. Forse, M. Gillmore, L.Nguyen, and K. Head for technical assistance; J. S. Sitton for helpin preparing the figures; and S. Mooney for typing the manuscript.

LITERATURE CITED

1. Baker, R. 1985. Biological control of plant pathogens: defini-tions, p. 25-39. In M. A. Hoy and D. C. Herzog, (ed.),Biological control in agricultural IPM systems. Academic Press,Inc., New York.

2. Bakker, P. A. H. M., A. W. Bakker, J. D. Marugg, P. J.Weisbeek, and B. Schippers. 1987. Bioassay for studying the roleof siderophores in potato growth stimulation by Pseudomonasspp. in short potato rotations. Soil Biol. Biochem. 19:443-449.

3. Bakker, P. A. H. M., J. G. Lamers, A. W. Bakker, J. D.Marugg, P. J. Weisbeek, and B. Schippers. 1986. The role ofsiderophores in potato tuber yield increase by Pseudomonasputida in a short rotation of potato. Neth. J. Plant Pathol. 92:249-256.

4. Birnboim, H. C., and J. Doly. 1979. A rapid alkaline extractionprocedure for screening recombinant plasmid DNA. NucleicAcids Res. 6:1513-1519.

5. Bradford, M. 1976. A rapid and sensitive method for quantita-tion of protein utilizing the principle of protein-dye binding.Anal. Biochem. 72:248-254.

6. Brisbane, P. G., L. J. Janik, M. E. Tate, and R. F. 0. Warren.1987. Revised structure for the phenazine antibiotic from Pseu-domonas fluorescens 2-79 (NRRL B-15132). Antimicrob.Agents Chemother. 31:1967-1971.

7. Burr, T. J., and A. Caesar. Beneficial plant bacteria. Crit. Rev.Plant Sci. 2:1-20.

8. Byng, G. S., A. Berry, and R. A. Jensen. 1983. A pair ofregulatory isozymes for 3-deoxy-D-arabino-heptulosonate isconserved within group I pseudomonads. J. Bacteriol. 156:429-433.

9. Byng, G. S., and R. A. Jensen. 1983. Impact of isozymes uponpartitioning of carbon flow and regulation of aromatic aminoacid biosynthesis in prokaryotes. Isozymes 8:115-140.

10. Byng, G. S., J. L. Johnson, R. J. Whitaker, R. L. Gherna, andR. A. Jensen. 1983. The evolutionary pattern of aromatic aminoacid biosynthesis and the emerging phylogeny of pseudomonadbacteria. J. Mol. Evol. 19:272-282.

11. Calhoun, D. H., M. Carson, and R. A. Jensen. 1972. The branchpoint metabolite for pyocyanine biosynthesis in Pseudomonas

J. BACTERIOL.


http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/


aeruginosa. J. Gen. Microbiol. 72:581-583.12. Colyer, P. D., and M. S. Mount. 1984. Bacterization of potatoes

with Pseudomonas putida and its influence on postharvest softrot diseases. Plant Dis. 68:703-706.

13. Cook, R. J., and D. M. Weller. 1987. Management of take-all inconsecutive crops of wheat or barley, p. 41-76. In I. Chet (ed.),Innovative approaches to plant disease control. John Wiley &Sons, Inc., New York.

14. Currier, T. C., and E. W. Nester. 1976. Isolation of covalentlyclosed circular DNA of high molecular weight from bacteria.Anal. Biochem. 76:431-441.

15. Dhingra, 0. D., and J. B. Sinclair. 1985. Basic plant pathologymethods. CRC Press, Inc., Boca Raton, Fla.

16. Ditta, G., S. Stanfield, D. Corbin, and D. R. Helinski. 1980.Broad host range DNA cloning system for gram-negative bac-teria: construction of a gene bank of Rhizobium meliloti. Proc.Natl. Acad. Sci. USA 77:7347-7351.

17. Gurusiddaiah, S., D. M. Weller, A. Sarkar, and R. J. Cook.1986. Characterization of an antibiotic produced by a strain ofPseudomonas fluorescens inhibitory to Gaeumannomyces gra-minis var. tritici and Pythium spp. Antimicrob. Agents Chemo-ther. 29:488-495.

18. Gutterson, N. I., T. J. Layton, J. S. Zeigle, and G. J. Warren.1986. Molecular cloning of genetic determinants for inhibition offungal growth by a fluorescent pseudomonad. J. Bacteriol. 165:696-703.

19. Holliday, R. 1956. A new method for the identification ofbiochemical mutants of microorganisms. Nature (London) 178:987.

20. Holliman, F. G. 1969. Pigments of Pseudomonas species. I.Structure and synthesis of aeruginosin A. J. Chem. Soc. (C):2514-2516.

21. Howell, C. R., and R. D. Stipanovic. 1979. Control of Rhizocto-nia solani on cotton seedlings with Pseudomonas fluorescensand with an antibiotic produced by the bacterium. Phytopathol-ogy 69:480-482.

22. Howell, C. R., and R. D. Stipanovic. 1980. Suppression ofPythium ultimum-induced damping-off of cotton seedlings byPseudomonas fluorescens and its antibiotic, pyoluteorin. Phy-topathology 70:712-715.

23. Howie, W. J., R. J. Cook, and D. M. Weller. 1987. Effects of soilmatric potential and cell motility on wheat root colonization byfluorescent pseudomonads suppressive to take-all. Phytopathol-ogy 77:286-292.

24. James, D. W., Jr., and N. I. Gutterson. 1986. Multiple antibiot-ics produced by Pseudomonas fluorescens HV37a and theirdifferential regulation by glucose. Appi. Environ. Microbiol. 52:1183-1189.

25. Jorgensen, R. A., S. J. Rothstein, and W. S. Reznikoff. 1979. Arestriction enzyme cleavage map of Tn5 and location of a regionencoding neomycin resistance. Mol. Gen. Genet. 177:65-72.

26. Kanner, D., N. N. Gerber, and R. Bartha. 1978. Pattern ofphenazine pigment production by a strain of Pseudomonasaeruginosa. J. Bacteriol. 134:690-692.

27. Kloepper, J. W., J. Leong, M. Teintze, and M. N. Schroth. 1980.Pseudomonas siderophores: a mechanism explaining disease-suppressive soils. Curr. Microbiol. 4:317-320.

28. Kloepper, J. W., J. Leong, M. Teintze, and M. N. Schroth. 1980.Enhanced plant growth by siderophores produced by plantgrowth-promoting rhizobacteria. Nature (London) 286:885-886.

29. Kloepper, J. W., and M. N. Schroth. 1981. Relationship of invitro antibiosis of plant growth-promoting rhizobacteria to plantgrowth and displacement of root microflora. Phytopathology 71:1020-1024.

30. Leong, J. 1986. Siderophores: their biochemistry and possiblerole in biocontrol of plant pathogens. Annu. Rev. Phytopathol.24:187-209.

31. Levitch, M. E. 1970. Regulation of aromatic amino acid biosyn-thesis in phenazine-producing strains of Pseudomonas. J. Bac-teriol. 103:16-19.

32. Longley, R. P., J. E. Halliwell, J. J. R. Campbell, and W. M.Ingledew. 1972. The branchpoint of pyocyanine biosynthesis.Can. J. Microbiol. 18:1357-1363.

33. Maniatis, T., E. F. Fritsch, and J. Sambrook. 1982. Molecularcloning: a laboratory manual. Cold Spring Harbor Laboratory,Cold Spring Harbor, N.Y.

34. Marsh, J. L., M. Erfle, and E. J. Wykes. 1984. The pIC plasmidand phage vectors with versatile cloning sites for recombinantselection by insertional inactivation. Gene 32:481-485.

35. Marugg, J. D., M. van Spanje, W. P. M. Hoekstra, B. Schippers,and P. J. Weisbeek. 1985. Isolation and analysis of genesinvolved in siderophore biosynthesis in plant-growth-stimu-lating Pseudoinonas putida WCS358. J. Bacteriol. 164:563-570.

36. Neilands, J. B., and S. A. Leong. 1986. Siderophores in relationto plant disease. Annu. Rev. Plant Physiol. 37:187-208.

37. Parke, J. L., R. Moen, A. D. Rovira, and G. D. Bowen. 1986.Soil water flow affects the rhizosphere distribution of a seed-borne biological control agent, Pseudomonas fluorescens. SoilBiol. Biochem. 18:583-588.

38. Peet, R. C., P. B. Lindgren, D. K. Willis, and N. I. Panopoulos.1986. Identification and cloning of genes involved in phaseolo-toxin production by Pseudomonas syringae pv. phaseolicola. J.Bacteriol. 166:1096-1105.

39. Priefer, U. B., R. Simon, and A. Puhler. 1985. Extension of thehost range of Escherichia coli vectors by incorporation ofRSF1010 replication and mobilization functions. J. Bacteriol.163:324-330.

40. Rothrock, C. S., and D. Gottlieb. 1984. Role of antibiosis inantagonism of Streptomyces hygroscopicus var. geldanus toRhizoctonia solani in soil. Can. J. Microbiol. 30:1440-1447.

41. Scher, F. M., and R. Baker. 1982. Effect ofPseudomonas putidaand a synthetic iron chelator on induction of soil suppres-siveness to Fusarium wilt pathogens. Phytopathology 72:1567-1573.

42. Schippers, B., A. W. Bakker, and P. A. H. M. Bakker. 1987.Interactions of deleterious and beneficial rhizosphere microor-ganisms and the effect of cropping practices. Annu. Rev.Phytopathol. 25:339-358.

43. Schleif, R. F., and P. C. Wensink. 1981. Practical methods inmolecular biology. Springer-Verlag, New York.

44. Schroth, M. N., and J. G. Hancock. 1981. Selected topics inbiological control. Annu. Rev. Microbiol. 35:453-476.

45. Schroth, M. N., and J. G. Hancock. 1982. Disease-suppressivesoil and root-colonizing bacteria. Science 216:1376-1381.

46. Selvaraj, G., and V. N. Iyer. 1983. Suicide plasmid vehicles forinsertion mutagenesis in Rhizobium meliloti and related bacte-ria. J. Bacteriol. 156:1292-1300.

47. Sprague, R. 1950. Diseases of cereals and grasses in NorthAmerica. Ronald Press, New York.

48. Suslow, T. V. 1982. Role of root-colonizing bacteria in plantgrowth, p. 187-223. In M. S. Mount and G. H. Lacy, (ed.),Phytopathogenic prokaryotes, vol. 1. Academic Press, Inc.,New York.

49. Weller, D. M. 1985. Application of fluorescent pseudomonads tocontrol root diseases, p. 137-140. In C. A. Parker, A. D. Rovira,K. J. Moore, P. T. W. Wong, and J. F. Kollmorgen (ed.),Ecology and management of soilborne plant pathogens. TheAmerican Phytopathological Society, St. Paul, Minn.

50. Weller, D. M., and R. J. Cook. 1983. Suppression of take-all ofwheat by seed treatments with fluorescent pseudomonads.Phytopathology 73:463-469.

51. Weller, D. M., and R. J. Cook. 1985. Suppression of rootdiseases of wheat by fluorescent pseudomonads and mecha-nisms of action, p. 99-107. In T. R. Swinburne (ed.), Iron,siderophores, and plant disease. Plenum Publishing Corp., NewYork.

52. Weller, D. M., B.-X. Zhang, and R. J. Cook. 1985. Applicationof a rapid screening test for selection of bacteria suppressive totake-all of wheat. Plant Dis. 69:710-713.

53. Whitaker, R. J., G. S. Byng, R. L. Gherna, and R. A. Jensen.1981. Comparative allostery of 3-deoxy-D-arabino-heptuloso-nate-7-phosphate synthetase as an indicator of taxonomic relat-edness in pseudomonad genera. J. Bacteriol. 145:752-759.

54. Whitaker, R. J., M. J. Fiske, and R. A. Jensen. 1982. Pseudo-monas aeruginosa possesses two novel regulatory isozymesof 3-deoxy-D-arabino-heptulosonate-7-phosphate synthase. J.

VOL. 170, 1988


http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/


Biol. Chem. 257:12789-12794.55. Wilkinson, H. T., R. J. Cook, and R. J. Alldredge. 1985.

Relation of inoculum size and concentration to infection ofwheat roots by Gaeumannomyces graminis var. tritici. Phyto-pathology 75:98-103.

56. Williams, S. T. 1982. Are antibiotics produced in soil? Pedobio-logia 23:427-435.

57. Williams, S. T., and J. C. Vickers. 1986. The ecology of

antibiotic production. Microb. Ecol. 12:43-52.58. Wong, P. T. W., and R. Baker. 1984. Suppression of wheat

take-all and Ophiobolus patch by fluorescent pseudomonadsfrom a Fusarium-suppressive soil. Soil Biol. Biochem. 16:397-403.

59. Yanisch-Perron, C., J. Viera, and J. Messing. 1985. ImprovedM13 phage vectors and host strains: nucleotide sequences of theM13mpl8 and PUC19 vectors. Gene 33:103-119.

J. BACTERIOL.


http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/

role of a phenazine antibiotic from pseudomonas ... · bacteria and pathogens in the rhizosphere is...

Documents