2005 fems lett bacteriocins

8/13/2019 2005 FEMS Lett Bacteriocins

1/8

Antagonistic activity among2,4-diacetylphloroglucinol-producing fluorescent Pseudomonas spp.

Shamil Validov a, Olga Mavrodi b, Leonardo De La Fuente b, Alexander Boronin a,David Weller c, Linda Thomashow c, Dmitri Mavrodi b,*

a Skryabin Institute of Biochemistry and Physiology of Microorganisms of the Russian Academy of Sciences, Pushchino, Russiab Department of Plant Pathology, Washington State University, 362 Johnson Hall, Pullman, WA 99164-6430, USA

c USDA Agricultural Research Service, Root Disease and Biological Control Research Unit, Washington State University, Pullman,

WA 99164-6430, USA

Received 15 September 2004; received in revised form 28 October 2004; accepted 4 November 2004

First published online 18 November 2004

Edited by Y. Okon

Abstract

Strains of fluorescent Pseudomonas spp. that produce 2,4-diacetylphloroglucinol (2,4-DAPG) differ in their ability to colonize

roots. In this study, we screened 47 2,4-DAPG-producing strains representing17 distinct genotypes for antagonistic activity associ-

ated with the production of bacteriocins. Upon induction, over 70% of the strains inhibited the growth of other isolates in vitro.

Greenhouse assays indicated that populations of sensitive strains in wheat rhizosphere soil declined more rapidly in the presence

of antagonists than when introduced alone. Antagonism can influence the ability of biocontrol agents to establish and maintain

effective population densities in situ.

2004 Federation of European Microbiological Societies. Published by Elsevier B.V. All rights reserved.

Keywords: Antagonism; Rhizosphere; Pseudomonas; Biological control; Bacteriocin

1. Introduction

Fluorescent Pseudomonas spp. that produce 2,4-

diacetylphloroglucinol (2,4-DAPG) can provide biolog-

ical control of soilborne pathogens on a wide range of

crops, and they have a key role in the suppressivenessof some soils to plant pathogens [1,2]. Although most

strains studied to date are members of the species P.

fluorescens, they nevertheless are phenotypically and

genotypically diverse [3,4]. Among the major differ-

ences associated with biological control are the capac-

ity to produce pyoluteorin and pyrrolnitrin and the

ability to colonize and persist in the rhizosphere. Colo-

nization and persistence are critical because a success-

ful biocontrol agent must establish and maintain a

minimum threshold population density in order to pro-

tect the host. We recently used genomic subtraction toexplore differences between the genomes of two closely

related 2,4-DAPG-producing strains differing in their

ability to colonize the rhizosphere of wheat [5]. Among

DNA fragments present in the superior colonizer Q8r1-

96 but not in the less rhizosphere-competent strain Q2-

87 was SSH6, a clone with similarity to colicin M from

Escherichia coli. This finding prompted us to explore

antagonistic activity among related 2,4-DAPG-produc-

ing isolates.

0378-1097/$22.00 2004 Federation of European Microbiological Societies. Published by Elsevier B.V. All rights reserved.

doi:10.1016/j.femsle.2004.11.013

* Corresponding author. Tel.: +1 509 335 3269; fax: +1 509 335

7674.

E-mail address: [email protected](D. Mavrodi).

www.fems-microbiology.org

FEMS Microbiology Letters 242 (2005) 249256
mailto:[email protected]:[email protected]


2/8

Antagonism between microorganisms is mediated by

antibiotics, lytic enzymes, exotoxins and bacteriocins,

the last of which represent the most abundant class of

bacterial defensive factors. Bacteriocins often are de-

fined as proteinaceous compounds that are produced

by certain strains and kill closely related strains or spe-

cies[6]. Bacteriocin production by Gram-negative bacte-ria has been studied most extensively in enterobacteria

and P. aeruginosa. However, bacteriocin-producing iso-

lates occur frequently among plant-associated fluores-

cent Pseudomonas spp., and it has been proposed that

such strains may have a competitive advantage in the

rhizosphere[7,8]. Many of the findings to date are based

on studies in vitro, and the impact of bacteriocins on the

survival of fluorescentPseudomonasspp. in situ remains

largely unexplored.

In this study, we assessed antagonistic activity among

47 different 2,4-DAPG-producing strains representing

17 different genotypes distinguished by rep-PCR [3].

Based on inhibition studies in vitro, we utilized five of

these strains to evaluate the impact of antagonism on

competition in the rhizosphere of wheat.

2. Materials and methods

2.1. Bacterial strains and plasmids

Most of the strains used in the study were described

previously [3,4,9]. P. fluorescens EPS808 and EPS817

were a kind gift from Dr. E. Montesinos (University

of Girona, Girona, Spain). P. fluorescens PHL1C2 waskindly provided by Dr. P. Lemanceau (Universite de

Bourgogne, Dijon, France). Other strains and plasmids

are listed in Tables 1 and 2. Bacterial cultures were

grown in LuriaBertani (LB) medium[10], Tryptic Soy

Broth (TSB) (Difco Laboratories, USA), MMP minimal

medium [11] or one-third strength Kings medium B

(KMB) [12] at 28 C. Antibiotics were used at the fol-

lowing concentrations: kanamycin, 25lg/ml; rifampicin,

100 lg/ml; cycloheximide, 100 lg/ml; chloramphenicol,

13 lg/ml; ampicillin, 40 lg/ml; kanamycin, 25 lg/ml,

and gentamycin, 12 lg/ml. Strains tagged with mini-

Tn7 were spontaneous rifampicin-resistant derivatives

[9].

2.2. DNA manipulations and analyses

Standard techniques were used for plasmid DNA iso-

lation, agarose gel electrophoresis, and transformation

[10]. PCR amplification was carried out with Plati-

numTaq DNA polymerase (Life Technologies, USA)

according to the manufacturers recommendations.

DNA was sequenced using an ABI PRISM BigDye Ter-

minator Cycle Sequencing Ready Reaction Kit (Applied

Biosystems, USA), and sequence data were analyzed

with OMIGA 2.0 software (Accelrys, USA). The oligo-

nucleotide primer glnSmod was designed with Oligo 6.0

software (Molecular Biology Insights, USA).

2.3. Antagonism in vitro

Strains ofPseudomonasspp. were screened for bacte-

riocin production as described by Govan[13].Each tes-

ter strain was spread across TSB agar in a band

approximately 2.5 cm wide. After 18 h at 28 C, the

plate was irradiated at 254 nm for 1 min and incubated

at 28 C for another 3.5 h. Bacterial growth was scraped

from the agar with a cotton swab soaked in chloroform

and the plates were exposed to chloroform vapor to kill

all remaining bacteria. Samples (15 ll) from overnightcultures of indicator strains in TSB were then streaked

perpendicular to the tester band (see Fig. 1). Alterna-

tively, tester cultures grown at 28 C in 10 ml of LB

broth were collected by centrifugation and suspended

in 1 ml of LB broth amended with 1 lg ml1 of mitomy-

cin C to induce bacteriocin production. The cultures

were shaken for 5 hr and lysed with 100 ll of chloro-

Table 1

Bacterial strains and plasmids used in this study

Strain Descriptiona Reference or source

P. fluorescens

Pf-5Km Pf-5 tagged with mini-Tn7-gfp1. Phl+ Rifr Kmr This study

Q2-87Km Q2-87 tagged with mini-Tn7-gfp1. Phl+ Rifr Kmr This study

Q8r1-96Km Q8r1-96 tagged with mini-Tn7-gfp1. Phl+ Rifr Kmr This study

Q8r1-96Gm Q8r1-96 tagged with mini-Tn7-gfp2. Phl+ Rifr Gmr This study

FTAD1R36Gm FTAD1R36 tagged with mini-Tn7-gfp2. Phl+ Rifr Gmr This study

FTAD1R34Gm FTAD1R34 tagged with mini-Tn7-gfp2. Phl+ Rifr Gmr This study

Plasmids

pBK- mini-Tn7-gfp1 pUC19- derive d delivery vector for mini-Tn7-gfp1. Kmr cat bla gfp Mob+ [18]

pBK- mini-Tn7-gfp2 pUC19- derive d delivery vector for mini-Tn7-gfp2. Gmr cat bla gfp Mob+ [18]

pUX-BF13 Donor of Tn7 transposase. tnsABCDER6K bla Mob+ [16]

a Phl+, the strain produces 2,4-diacetylphloroglucinol; Rifr, rifampin resistance; Kmr, kanamycin resistance; Gmr, gentamycin resistance;cat,

chloramphenicol acetyltransferase;bla, b-lactamase.

250 S. Validov et al. / FEMS Microbiology Letters 242 (2005) 249256


3/8

form. After mixing and centrifugation, the aqueous

phase was spotted on LB agar and overlayed with soft

agar containing a single indicator strain. Lysates also

were tested after trypsin digestion, freeze-thawing, and

Microcon column filtration as described by Riley et al.

[14] in order to distinguish between the production of

low-molecular weight (S-pyocin-like) and phage tail-like

(R- and F-pyocin-like) bacteriocins and bacteriophages.

Plates for all assays in vitro were incubated at 28 C andscored for for the presence (+) or absence () of inhibi-

tion at 12 and 24 h. Each tester strain was replicated

once per experiment and the entire collection was

screened twice. The data were combined and a strain

was considered to be an antagonist only if the results

of both assays agreed. The combined results were used

to produce a two-dimensional rectangular matrix of bin-

ary codes that was analyzed with MVSP 3.12 software

(Kovach Computing Services, UK) using a simple

matching coefficient that considers both the presence

and the absence of antagonistic activity.

2.4. Transposon tagging

Strains of P. fluorescens were tagged by electropora-

tion [15] with 300 ng each of pBK-mini-Tn7 gfp 1 or

pBK-mini-Tn7 gfp 2 (Table 1) and the helper plasmid

pUX-BF13 [16]. The tagged clones were isolated on

LB agar amended with kanamycin or gentamycin. Tn7

preferentially inserts in a single orientation into a spe-

cific neutral intergenic site, att Tn7, present in many

eubacterial genomes[17]. To confirm the site of transpo-

son insertion in our strains, we amplified the region

flanking the transposon by PCR with the oligonucleo-

tide primers cat [18] and glnSmod (5 0-AAY CTS GCS

AAG TCG GTS AC-30), which are specific to the

mini-Tn7-borne chloramphenicol acetyltransferase gene

and the 3 0-end of the glutamine synthetase gene, respec-

tively. The cycling included a 2-min initial denaturation

at 94 C, followed by 6 cycles of 94 C for 30 s, 52 C for

15 s, and 72 C for 30 s, followed by 29 more cycles with

Table 2

In vitro antagonism among 2,4-diacetylphloroglucinol-producingPseudomonas spp.

Genotypea No. of surveyed strains No. of strain-antagonists

within the genotype

No. of strains that are

inhibited by the genotype

Breadth of killingb

A 3 2 3 2 (11.8)

B 2 0 0 0

C 2 2 13 3 (17.6)

D 11 9 19 7 (41.2)E 4 4 11 8 (47.1)

F 2 2 3 2 (11.8)

G 1 1 3 3 (17.6)

H 1 0 0 0

I 1 1 16 5 (29.4)

J 2 2 3 1 (5.9)

K 3 1 5 4 (23.5)

L 3 1 4 1 (5.9)

M 3 2 12 4 (23.5)

N 1 1 1 1 (5.9)

O 2 2 10 6 (35.3)

P 4 4 18 13 (76.5)

Q 2 2 12 3 (17.6)

a

Genotypes were defined previously by BOX-PCR genomic fingerprinting[3,4].b No. of genotypes that are inhibited by the surveyed genotype. Values in parenthesis indicate percentages, where 100% corresponds to 17 different

genotypes represented in the collection.

Fig. 1. An example of a plate from the inhibition assay. The test

strain, P. fluorescens Q8r1-96, was streaked across the surface of

tryptic soy agar to a width of approximately 2.5 cm (plate D). After

incubation for 18 h at 28 C the bacteria were scraped from the agar

and the plates were exposed to chloroform vapor to kill the remaining

bacteria. Fifteen microliters of an overnight culture of each indicator

strain in TSB were then applied across the original inoculum (plates A,

B, and C). The plates were incubated for an additional 24 h at 28 Cand examined for zones of inhibition (summarized inTable 2). On this

figure, four strains inhibited by Q8r1-96 are: FTAD1R34 (1), HT5-1

(2), Q2-87 (3), and ATCC49054 (4).

S. Validov et al. / FEMS Microbiology Letters 242 (2005) 249256 251


4/8

the annealing temperature of 60 C and a final extension

at 72 C for 5 min. The 0.65-kb amplified fragments

were purified with a QIAEX II gel extraction kit (Qia-

gen, USA) and sequenced with the primers cat and

glnSmod.

2.5. Rhizosphere colonization and competition

Rhizosphere colonization assays were performed with

wheat (Triticum aestivum L.) cv. Penawawa in non-ster-

ile soil as described by Landa et al. [19]. Quincy virgin

soil was inoculated with bacteria in a 1% suspension

of methylcellulose to give 1 104 CFU g1 of soil

(fresh weight). Mixed inoculation treatments contained

a 1:1 mixture of competing strains (0.5 104 CFU g1

of soil for each strain). Plants were incubated in a

growth chamber for three successive 3-week cycles at

15 1 C with a 12-h photoperiod. After each cycle,

the population size of the introduced strains was deter-

mined on the roots of six randomly selected plants. Each

treatment was replicated three times and the entire

experiment was repeated twice.

Population densities of introduced strains were enu-

merated by PCR, using a modified dilution endpoint as-

say[3] with an extra step in which bacteria were selected

in kanamycin- or gentamycin-amended media to distin-

guish between strains in mixed inoculation treatments

and to check for cross-contamination of single intro-

duced strains. Bacterial growth was assessed after 72

h, with an OD600P 0.1 considered as positive.

2.6. Statistical analyses

All treatments in the growth chamber were arranged

in a randomized complete block design. Because rhizo-

bacteria on roots are lognormally distributed [20], pop-

ulation data were converted to log CFU g1 (fresh

weight) of soil or root to satisfy the normality assump-

tion of analysis of variance (ANOVA). Data were ana-

lyzed using STATISTIX (version 8.0, Analytical

Software, USA). Differences in population densities

among treatments were determined by ANOVA, and

mean comparisons among treatments were performed

by Fishers protected least significance difference

(LSD) test at P= 0.05.

3. Results

3.1. Distribution of antagonistic activity amongphlD+

Pseudomonas spp.

Inhibition studies in vitro revealed that antagonistic

activity is widely distributed amongphlD Pseudomonas

spp. (Table 2). Of 47 tested strains, only 11 failed to

antagonize another strain. For most genotypes the antag-

onistic activity also positively correlated with the

breadth of killing, i.e., the percentage of the 17 geno-

types inhibited. Cluster analyses indicated that the P, D

and E genotypes harbored the most active antagonists

(Fig. 2(a)). The results reflect to some extent the unbal-

anced nature of our strain collection, in which certain

genotypes are represented by multiple strains while oth-ers are represented only once. Nevertheless, the single I-

genotype strain FTAD1R36 was among the most active

antagonists, inhibiting 16 strains from four different

genotypes (Table 2). Similarly, the P genotype, repre-

sented by four strains, inhibited almost twice as many

other genotypes as did the D genotype, represented by

11 isolates. Other highly active strains included MVW4-

3 (Q genotype), PILH1 (M genotype), and STAD376

(C genotype). At the other extreme, members of the B

and H genotypes failed to antagonize any of the tested

strains, nor did strains CHA0 (A genotype), Q2-5 (D

genotype), ATCC49054 (D genotype), EPS808 (K geno-

type), EPS817 (K genotype), PHL1C2 (M genotype),

W4-4 (L genotype), and 1M1-96 (L genotype) (Fig. 2(b)).

Cluster analyses based on patterns of sensitivity re-

vealed that I, G, and J genotype strains were among

the most resistant while strains of the B, D, E, and P

genotypes were on average more susceptible to inhibi-

tion by other phlD+ strains (data not shown).

3.2. Competition between tagged strains in the wheat

rhizosphere

Based on their behavior in vitro, strains Pf-5, Q2-87,

Q8r1-96, FTAD1R34, and FTAD1R36 were utilized toassess the impact of antagonistic activity on strain sur-

vival in the rhizosphere of wheat. Neither Pf-5 (A geno-

type) nor Q2-87 (B genotype) antagonized other strains

in vitro, and Q2-87 but not Pf-5 was antagonized by

Q8r1-96 (D genotype). Q8r1-96 also antagonized

FTAD1R34 (D genotype) and, along with strain Pf-5,

was antagonized by FTAD1R36 (I genotype). These

strains were tagged either with kanamycin or gentamy-

cin resistance genes [18] by using disarmed mini-Tn7

transposons that inserted into the genome immediately

downstream of glnS. Growth studies at 27 C in one-

third strength KMB and MMP revealed no differences

in the growth kinetics of the tagged derivatives and their

respective wild-type parents (data not shown).

Population densities of single strains or differentially-

marked strain pairs did not differ significantly (P= 0.05)

immediately after inoculation (cycle 0) and ranged from

log 3.0 to 3.5 CFU g1 of bulk soil (Fig. 3). However,

after one cycle, large differences in the population sizes

of antagonist and sensitive strains were observed on

roots in soil that had been co-inoculated with a mix of

Q8r1-96 and Q2-87 or FTAD1R36 and Pf-5. Population

densities of the antagonists in these mixtures generally

were unaffected by the presence of the sensitive strains,



5/8

while populations of the sensitive strains (Q2-87 and Pf-

5) dropped below the detection limit of approximately

log 3.26 CFU g1 of root after only one cycle and never

recovered (Fig. 3(a) and (e)). No such rapid decline was

observed in single inoculations of Q2-87 or Pf-5, and for

Q2-87, the difference between population densities in

single and mixed inoculations after one cycle was over

log 3.0 CFU g1 of root.

Simple Matching Coefficient

O

Q

F

C

A

B

H

N

L

J

IM

G

K

D

P

E

0.4 0.5 0.6 0.7 0.8 0.9 1

Simple Matching Coefficient

EPS808(K)Q2-87(B)

ATCC49054(D)1M1-96(L)CHA0(A)PHL1C2(M)Q2-1(B)EPS817(K)CV1-1(H)Q2-5(D)W4-4(L)FFL1R22(J)MVP1-3(A)FFL1R9(D)HT5-1(N)Q128-87(D)Q8r1-96(D)5MR2(E)FFL1R8(J)FTAD1R34(D)MVW1-2(D)D27B1(M)MVW4-2(Q)Q37-87(E)Pf-5(A)MVP1-6(D)JMP6(F)JMP7(F)FFL1R18(G)MARV1(L)7MA12(O)

L5-1(D)QT1-5(D)4MA6(P)STAD384(C)MVW1-1(P)6WSU4(P)F113(K)QT1-6(E)Q2-2(E)7MA20(O)MVP1-4r(P)OC4-1(D)MVW4-3(Q)FTAD1R36(I)PILH1(M)STAD376(C)

0.7 0.75 0.8 0.85 0.9 0.95 1

(a)

(b)

Fig. 2. Cluster analyses of antagonistic phenotypes within a representative collection ofphlD-positive Pseudomonasspp. Patterns of inhibition were

clustered by rep-PCR genotype (a) or by individual bacterial strain (b). The UPGMA tree was generated with MVSP 3.12 from a similarity matrix

generated using the simple matching coefficient.



6/8

In paired inoculations, strain Q8r1-96 outcompeted

the sensitive strains (i.e., Q2-87 and FTAD1R34), as

well as the supposed antagonist FTAD1R36 (Fig. 3(a),

(c) and (d)). It also displaced Pf-5 (Fig. 3(b)), the least

rhizosphere-competent strain in this study, despite the

fact that these two strains were mutually non-antagonis-

tic in vitro. In all experiments, Q8r1-96 was largely unaf-

fected by the presence of a competitor and maintained

population densities equal to or greater than log 6.0

CFU g1 of root in both single and mixed inoculations.

4. Discussion

DAPG-producing fluorescent Pseudomonas spp. of

the D genotype are among the most aggressive root-

colonizing biocontrol agents studied to date. On the

other hand, D-genotype strains are virtually indistin-

guishable from members of most other genotypes when

grown in vitro, suggesting that relatively few differences

at the genetic level might be sufficient to confer the col-

onization phenotype typical of premier PGPR [3]. As

one approach to identify novel genes associated withthis trait, we isolated DNA fragments present in the

D-genotype strain Q8r1-96 but not in an average colo-

nizer, Q2-87 (B genotype) [5]. One subtracted gene,

SSH6, exhibited similarity to the bacteriocin colicin

M and later was found to form part of a 17-kb pyo-

cin-like locus in the genome of Q8r1-96 (data not

shown).

Bacteriocins of pseudomonads are best characterized

in P. aeruginosa, where they traditionally were called

pyocins [21]. As in other eubacteria, these bacteriocins

include both low- and high-molecular weight species.

The low molecular-weight bacteriocins, or S-pyocins,

are protease-sensitive enzymes that kill other strains

through their DNase, RNase, or membrane pore-form-

ing activity. The high molecular-weight forms comprise

two categories, R- and F-type pyocins, that closely

resemble non-flexible contractile and flexuous non-con-

tractile bacteriophage tails, respectively. In P. aerugin-

osa, the R- and F-type pyocins are related to two

different lineages of bacteriophages. The high molecu-

lar-weight pyocins are thought to kill sensitive cells

through depolarization of the cytoplasmic membrane.

Although bacteriocins generally are associated with bac-

terial competitiveness in the environment, this associa-

tion represents a relatively unexplored topic and weare aware of only one other recent study in which

plant-associated Pseudomonas spp. were screened for

bacteriocinogenic activity[7].

Most of the strains we tested were antagonistic to

other fluorescent Pseudomonas spp. in vitro after induc-

tion. It is unlikely that this activity was due to antibiot-

ics because these strains previously were evaluated for

the production of antibiotics other than 2,4-DAPG [4].

Although we considered antagonism to be a single phe-

notypic trait in this work, it could well be due to multi-

ple low-molecular weight and/or phage tail-like

bacteriocins acting in concert [8,21]. In fact, analyses

of recently sequenced microbial genomes indicate that

isolates ofP. fluorescens have the potential to produce

numerous bacteriocins [8]. To distinguish between low-

molecular weight and phage tail-like bacteriocins and

bacteriophages, we subjected mitomycin C-induced ly-

sates to trypsin digestion, freeze-thawing, and ultrafiltra-

tion. In most cases, antagonistic activity was associated

with a high molecular weight, protease-resistant frac-

tion, implicating phage tail-like bacteriocins and bacte-

riophages, and indeed, electron microscopy revealed

pyocin-like particles in mitomycin C-induced lysates of

some strains.

FTAD1R36

FTAD1R34

Pf-5

Q2-87

FTAD1R36(A) + Pf-5(S)

Q8r1-96

1

2

3

4

5

6

7

8

9

0

1

2

3

4

5

6

7

8

9

0

9

1

2

3

4

5

6

7

8

Cycle1 2 3

Cycle1 2 3

(e)

(c)

(a) (b)

(d)

FTAD1R36(A) + Q8r1-96(S)Q8r1-96(A) + FTAD1R34(S)

Pf-5(N) + Q8r1-96(N)Q8r1-96(A) + Q2-87(S)

population

density(logC

FU

g-1of

root)

Rhizo

sphere

Fig. 3. Population dynamics of selected pairs of mini-Tn7-tagged

strains on roots of wheat cv. Penawawa grown in non-sterile Quincy

virgin soil for three successive cycles of 3 weeks each. For each pair of

competing strains the letters in parentheses indicate antagonistic (A),

sensitive (S), or neutral (N) phenotypes observed in vitro. The

competing strain pairs were: Q8r1-96Gm and Q2-87Km (panel a);

Q8r1-96Gm and Pf-5Km (panel b); Q8r1-96Km and FTAD1R34Gm

(panel c); Q8r1-96Km and FTAD1R36Gm (panel d); and

FTAD1R36Gm and Pf-5Km (panel e). Open symbols correspond totreatments inoculated with a single strain at a density of approximately

1 104 CFU g1 soil. Filled symbols indicate treatments that were

inoculated by a 1:1 mixtureof strains,each at a densityof approximately

0.5 104 CFU g1 soil. The control treatment consisted of soilamended

with a 1% methylcellulose suspension. Mean and standard deviations

for one experiment are presented. Dotted lines represent cycles where

population densities of strainsQ2-87 (panela) and Pf-5 (panels b ande)

fell in mixed inoculations below the detection limit of assay.



7/8

Bacteriocins generally are considered to be intraspe-

cies competitiveness factors that may help producing

strains to invade a new ecological niche and/or repel

other strains from an already occupied niche [6,22].

However, despite the abundance of data from studies

in vitro, the ecological role(s) of bacteriocins in nature

remains obscure. We therefore were particularly inter-ested in the impact of antagonism on the competitive-

ness of our isolates in the rhizosphere of wheat, a

niche to which they should be well-adapted.

In all rhizosphere experiments except for

FTAD1R36+Q8r1-96 (Fig. 3(d)), the antagonist dis-

placed the competing sensitive strain (Fig. 3). The effect

generally was quite pronounced, with the population

density of sensitive strains decreasing by over log 3.0

CFU g1 and, in the case of pairs Q8r1-96+Q2-87 and

FTAD1R36+Pf-5, quickly dropping below the detection

limit. These shifts were consistent from experiment to

experiment and may be indicative of the inhibition of

one strain by another in the plant rhizosphere. The re-

sults also agree with findings of Landa et al. [19], who

examined the interaction between strains Q8r1-96 and

Q2-87 in a de Wit replacement series experiment in

which wheat was grown in soil into which both strains

had been introduced over a range of defined ratios.

The final population size of Q2-87 in mixed inoculations

was lower than predicted even when the strains were at a

ratio of 0.3(Q8r1-96):0.7(Q2-87), consistent with antag-

onism of Q2-87 by Q8r1-96[19].

Surprisingly, P. fluorescens Q8r1-96 displaced every

counterpart in the rhizosphere regardless of whether it

was antagonistic (inhibition of Q2-87 and FTAD1R34),sensitive (sensitivity to FTAD1R36) or neutral (no reac-

tion to Pf-5) to those strains in vitro. Earlier rhizosphere

competition studies[19]also revealed that Q8r1-96 out-

competed MVP1-4r and 1M1-96 but was itself displaced

from the rhizosphere of wheat by another 2,4-DAPG-

producing strain, P. fluorescens F113. All of these

strains are neutral to one another in vitro.

A number of theoretical and in vitro studies predict

that for otherwise equivalent strains, the outcome of

competition between a bacteriocin-producer and a sensi-

tive strain will depend on their initial proportions in the

system[23]. These studies suggest that if a bacteriocin-

producer is introduced at a low level, it will not invade

successfully and its population will eventually shrink

to zero. We speculate that Q8r1-96 may colonize roots

more aggressively than other strains not only because

of its antagonistic activity, but also because it rapidly

establishes basal populations that utilize the available

resources (e.g., root exudates) more efficiently. As a con-

sequence, it grows faster, eventually overcoming even

bacteriocin-producing competitors. Although in our

study we introduced competing strains into soil at a

1:1 ratio, differences in growth rates in the rhizosphere

might have contributed to the decline in the popu-

lation of the antagonist in some treatments (i.e.,

FTAD1R36+Q8r1-96).

Bacteriocin production clearly is not the major deter-

minant of successful rhizosphere colonization by a par-

ticular bacterial strain, but our results suggest that it

may contribute to competition between closely related

strains ofPseudomonas spp., especially when the com-petitors are present at certain favorable ratios.

Acknowledgements

Shamil Validov, Olga Mavrodi and Leonardo De La

Fuente contributed equally to this study. The authors

thank Greg Phillips and Karen Hansen (USDA-ARS

Root Disease and Biocontrol Unit, Pullman, WA) for

help with the bacteriocin screening assays. This work

was supported by the U. S. Department of Agriculture,

National Research Initiative, Competitive Grants Pro-

gram (grant 03-35319-13800).

References

[1] Raaijmakers, J.M. and Weller, D.M. (2001) Exploiting genotypic

diversity of 2,4-diacetylphloroglucinol-producing Pseudomonas

spp.: characterization of superior root-colonizing P. fluorescens

strain Q8r1-96. Appl. Environ. Microbiol. 67, 25452554.

[2] Weller, D.M., Raaijmakers, J.M., Gardener, B.B.M. and Thoma-

show, L.S. (2002) Microbial populations responsible for specific

soil suppressiveness to plant pathogens. Ann. Rev. Phytopathol.

40, 309348.

[3] McSpadden Gardener, B.B., Schroeder, K.L., Kalloger, S.E.,Raaijmakers, J.M., Thomashow, L.S. and Weller, D.M. (2000)

Genotypic and phenotypic diversity of phlD-containing Pseudo-

monas strains isolated from the rhizosphere of wheat. Appl.

Environ. Microbiol. 66, 19391946.

[4] Mavrodi, O.V., McSpadden Gardener, B.B., Mavrodi, D.V.,

Bonsall, R.F., Weller, D.M. and Thomashow, L.S. (2001) Genetic

diversity ofphlD from 2,4-diacetylphloroglucinol-producing fluo-

rescentPseudomonas spp. Phytopathology 91, 3543.

[5] Mavrodi, D.V., Mavrodi, O.V., McSpadden Gardener, B.B.,

Landa, B.B., Weller, D.M. and Thomashow, L.S. (2002) Identi-

fication of differences in genome content among phlD-positive

Pseudomonas fluorescens strains by using PCR-based subtractive

hybridization. Appl. Environ. Microbiol. 68, 51705176.

[6] Riley, M.A. and Wertz, J.E. (2002) Bacteriocins: evolution,

ecology, and application. Ann. Rev. Microbiol. 56, 117137.[7] Parret, A.H.A. and De Mot, R. (2000) Bacteriocin production by

rhizosphere-colonizing fluorescent Pseudomonas. In: Proceedings

of the 5th International PGPR Workshop. Cordoba, Argentina.

[8] Parret, A.H.A. and De Mot, R. (2002) Bacteria killing their own

kind: novel bacteriocins of Pseudomonas and other c-proteobac-

teria. Trends Microbiol. 10, 107112.

[9] Landa, B.B., de Werd, H.A.E., McSpadden-Gardener, B.B. and

Weller, D.M. (2002) Comparison of three methods for monitoring

populations of different genotypes of 2,4-diacetylphloroglucinol-

producing Pseudomonas fluorescens in the rhizosphere. Phytopa-

thology 92, 129137.

[10] Ausubel, F.M., Brent, R., Kingston, R.E., Moore, D.D.,

Seidman, J.G., Smith, J.A. and Struhl, K. (1995) Short Protocols

in Molecular Biology. John Wiley & Sons, Inc., New York, NY.



8/8

[11] Loper, J.E. (1988) Role of fluorescent siderophore production in

biological control ofPythium ultimum by a Pseudomonas fluores-

cens strain. Phytopathology 78, 166172.

[12] King, E.O., Ward, M.K. and Raney, D. (1954) Two simple media

for the demonstration of pyocyanin and fluorescein. J. Lab. Clin.

Med. 44, 301307.

[13] Govan, J.R.W. (1978) Pyocin typing of Pseudomonas aeruginosa.

Method Microbiol. 10, 6191.

[14] Riley, M.A., Goldstone, C.M., Wertz, J.E. and Gordon, D. (2003)

A phylogenetic approach to assessing the targets of microbial

warfare. J. Evolution. Biol. 16, 690697.

[15] Enderle, P.J. and Farwell, M.A. (1998) Electroporation of freshly

plated Escherichia coli and Pseudomonas aeruginosa cells. Bio-

techniques 25, 954958.

[16] Bao, Y., Lies, D.P., Fu, H. and Roberts, G.P. (1991) An improved

Tn7-based system for the single-copy insertion of cloned genes into

chromosomes of Gram-negative bacteria. Gene 109, 167168.

[17] DeBoy, R.T. andGraig, N.L. (2000)Target site selectionby Tn7: att

Tn7transcription and target activity. J. Bacteriol. 182, 33103313.

[18] Koch, B., Jensen, L.E. and Nybroe, O. (2001) A panel of

Tn7-based vectors for insertion of the gfp marker gene or

for delivery of cloned DNA into Gram-negative bacteria at

a neutral chromosomal site. J. Microbiol. Meth. 45, 187

195.

[19] Landa, B.B., Mavrodi, D.V., Thomashow, L.S. and Weller, D.M.

(2003) Interactions between strains of 2,4-diacetylphloroglucinol-

producing Pseudomonas fluorescens in the rhizosphere of wheat.

Phytopathology 93, 982994.

[20] Loper, J.E., Suslow, T.V. and Schroth, M.N. (1984) Lognormal

distribution of bacterial populations in the rhizosphere. Phyto-

pathology 74, 14541460.

[21] Michel-Briand, Y. and Baysse, C. (2002) The pyocins of Pseudo-

monas aeruginosa. Biochimie 84, 499510.

[22] Riley, M.A. and Gordon, D.M. (1999) The ecological role of

bacteriocins in bacterial competition. Trends Microbiol. 7, 129

133.

[23] Durret, R. and Levin, S. (1997) Allelopathy in spatially distrib-

uted populations. J. Theor. Biol. 185, 165171.


2005 fems lett bacteriocins

Documents