phlD-basedgeneticdiversityanddetectionofgenotypesof2,4-diacetylphloroglucinol-producingPseudomonas£uorescensLeonardo De La Fuente1, Dmitri V. Mavrodi1, Blanca B. Landa2, Linda S. Thomashow3 & David M. Weller3

1Department of Plant Pathology, Washington State University, Pullman, WA, USA; 2Departamento de Agronomıa, Escuela Tecnica Superior de

Ingenieros Agronomos y de Montes, Universidad de Cordoba, Cordoba, Spain; and 3United States Department of Agriculture, Agricultural Research

Service, Root Disease and Biological Control Research Unit, Pullman, WA, USA

Correspondence: David M. Weller, United

States Department of Agriculture,

Agricultural Research Service, Root Disease

and Biological Control Research Unit, P.O. Box

646430, Washington State University,

Pullman, WA 99164-6430, USA. Tel.: 11 509

335 6210; fax: 11 509 335 7674;

e-mail: wellerd@mail.wsu.edu

Received 15 March 2005; revised 26 October

2005; accepted 27 October 2005.

First published online 24 January 2006.

doi:10.1111/j.1574-6941.2006.00074.x

Editor: Kornelia Smalla

Keywords

Pseudomonas; rhizosphere colonization;

2,4-diacetylphloroglucinol; biological control;

phlD; allele-specific PCR.

Abstract

Diversity within a worldwide collection of 2,4-diacetylphloroglucinol-producing

Pseudomonas fluorescens strains was assessed by sequencing the phlD gene.

Phylogenetic analyses based on the phlD sequences of 70 isolates supported the

previous classification into 18 BOX-PCR genotypes (A–Q and T). Exploiting

polymorphisms within the sequence of phlD, we designed and used allele-specific

PCR primers with a PCR-based dilution endpoint assay to quantify the population

sizes of A-, B-, D-, K-, L- and P-genotype strains grown individually or in pairs in

vitro, in the rhizosphere of wheat and in bulk soil. Except for P. fluorescens Q8r1-

96, which strongly inhibited the growth of P. fluorescens Q2-87, inhibition between

pairs of strains grown in vitro did not affect the accuracy of the method. The allele-

specific primer-based technique is a rapid method for studies of the interactions

between genotypes of 2,4-diacetylphloroglucinol producers in natural environ-

ments.

Introduction

Isolates of Pseudomonas fluorescens producing the antibiotic

2,4-diacetylphloroglucinol (2,4-DAPG) have been recovered

from soil, the rhizosphere of many crop species (Thoma-

show & Weller, 1995; Keel et al., 1996; McSpadden Gardener

et al., 2000; Lee & Kim, 2001; Weller et al., 2002; De La

Fuente et al., 2004) and marine environments (Isnansetyo

et al., 2003). These bacteria have been studied intensively

because they are responsible for the suppressiveness of some

soils against take-all disease of wheat caused by Gaeuman-

nomyces graminis var. tritici (Weller et al., 2002; de Souza

et al., 2003b) and black root rot of tobacco caused by

Thielaviopsis basicola (Stutz et al., 1986; Weller et al., 2002).

They are also highly effective biocontrol agents of a wide

variety of plant diseases when applied as seed or soil

treatments (Stutz et al., 1986; Fenton et al., 1992; Keel et al.,

1992; Shanahan et al., 1992; Schnider et al., 1995; Cronin

et al., 1997a, b; Duffy & Defago, 1997; Lee & Kim, 2001;

Raaijmakers & Weller, 2001; de Souza et al., 2003b; De La

Fuente et al., 2004). Disease suppression is associated with

the production of 2,4-DAPG, which has broad antiviral,

antibacterial, antifungal, antihelminthic and phytotoxic

properties (Shanahan et al., 1992; Bangera & Thomashow,

1999; Isnansetyo et al., 2003; Kamei & Isnansetyo, 2003).

2,4-DAPG may either inhibit soil-borne pathogens directly

(Thomashow & Weller, 1995; Bonsall et al., 1997; Cronin

et al., 1997a,b; Raaijmakers & Weller, 1998; Kamei & Isnan-

setyo, 2003; de Souza et al., 2003a) or induce systemic

resistance in the host plant (Iavicoli et al., 2003).

Considerable genetic diversity exists amongst 2,4-DAPG-

producing P. fluorescens isolates, and has been characterized

by amplified ribosomal DNA restriction analysis (ARDRA)

(Keel et al., 1996; McSpadden Gardener et al., 2000;

Picard et al., 2000), random amplified polymorphic DNA

(RAPD) analysis (Mavrodi et al., 2001), whole-cell repeti-

tive sequence-based PCR (rep-PCR) (McSpadden Gardener

et al., 2000; Landa et al., 2002a, 2005), restriction fragment

length polymorphism (RFLP) analysis of phlD, a key gene in

the 2,4-DAPG biosynthetic operon (Mavrodi et al., 2001;

McSpadden Gardener et al., 2001; Ramette et al., 2001;

Wang et al., 2001; Landa et al., 2002a, 2005), sequence

analysis of phlD (Ramette et al., 2001; Mazzola et al., 2004)

and denaturing gradient gel electrophoresis (DGGE)

FEMS Microbiol Ecol 56 (2006) 64–78c� 2006 Federation of European Microbiological SocietiesPublished by Blackwell Publishing Ltd. No claim to original US government works.

(Bergsma-Vlami et al., 2005). Because phlD is highly con-

served (Keel et al., 1996), it is a reliable marker for producers

of 2,4-DAPG (phlD1). We distinguished 18 genotypes (A–Q

and T) by rep-PCR with the BOXA1R primer (BOX-PCR)

in a worldwide collection of over 200 phlD1 isolates

(McSpadden Gardener et al., 2000; Landa et al., 2002a,

2005), and these correlate closely with groupings revealed

by RFLP analysis of phlD (Mavrodi et al., 2001; McSpadden

Gardener et al., 2001; Landa et al., 2002a, 2005). Additional

genotypes have been described recently by Mazzola et al.

(2004) (PfY and PfZ) and by McSpadden Gardener et al.

(2005) (R and S). Most genotypes distinguished by BOX-

PCR produce similar quantities of 2,4-DAPG in vitro

(Mavrodi et al., 2001) and in situ (Raaijmakers & Weller,

2001), have similar substrate utilization profiles (McSpad-

den Gardener et al., 2000; Raaijmakers & Weller, 2001) and

do not differ markedly when compared by classical bacter-

iological tests (McSpadden Gardener et al., 2000; Raaij-

makers & Weller, 2001). However, they differ significantly

in their competitiveness in the rhizosphere, and, at least on

pea and wheat, the genotype of an isolate is predictive of its

ability to establish and maintain population densities suffi-

cient to suppress disease (Landa et al., 2002a, 2003). How-

ever, the basis of the relationship between genotype and

colonization is not known. 2,4-DAPG is not a rhizosphere

competence factor (Raaijmakers & Weller, 2001), and bac-

terial traits and genes previously shown to contribute to root

colonizing ability (Lugtenberg et al., 2001) do not appear to

be unique to certain genotypes. Mavrodi et al. (2002) used

suppression subtractive hybridization to show that D-geno-

type isolates, known for aggressive colonization of wheat

and pea, have sequences not found in other genotypes with

average rhizosphere colonization ability. The role of these

sequences, including one encoding for the synthesis of

bacteriocins (Validov et al., 2005), is still under investiga-

tion.

Interactions amongst 2,4-DAPG-producing genotypes in

situ remain poorly understood, because it is difficult to

simultaneously distinguish and quantify multiple genotypes

in soil and rhizosphere samples. Different techniques have

been described for the quantification of total phlD1 popula-

tions in situ. Raaijmakers et al. (1997) used colony hybridi-

zation to detect and quantify indigenous phlD1 populations

in the wheat rhizosphere. This technique is accurate but

labour intensive, and does not discriminate between 2,4-

DAPG genotypes. McSpadden Gardener et al. (2001) have

developed a quantitative PCR-based dilution endpoint assay

that detects phlD1 population sizes essentially identical to

those obtained by colony hybridization (Landa et al.,

2002b), but is more rapid and the dominant phlD1 geno-

type can be identified by RFLP analysis of the amplified phlD

fragments. Landa et al. (2002a, 2003) used the dilution

endpoint method to demonstrate that certain genotypes

displace others in the same rhizosphere. However, a signifi-

cant limitation of the assay is that, when strains of two or

more genotypes are present, the subordinate strains are

difficult to detect by RFLP analysis (McSpadden Gardener

et al., 2001; Landa et al., 2003).

The objectives of this study were: (1) to assess the

diversity of 2,4-DAPG producers based on phlD sequence

data; and (2) to expand the ability of the dilution endpoint

assay (McSpadden Gardener et al., 2001) to trace the

population dynamics of phlD1 genotypes competing in the

rhizosphere. Using phlD sequences from strains of 20 BOX-

PCR genotypes, we demonstrated that sequence analysis

supports genotypes that have been previously defined using

molecular fingerprinting methods. We then used the se-

quence information to design phlD allele-specific primers

for A-, B-, D-, K-, L- and P-genotype strains. Using these

primers, we were able to monitor two competing genotypes

in mixed cultures in rhizosphere and soil samples even when

one genotype was up to 1000-fold less abundant than the

other.

Materials andmethods

Bacteria and culture conditions

Table 1 shows the bacterial strains and phlD gene sequences

used in this study. Included were 79 phlD1 strains represent-

ing 18 of the 22 described BOX-PCR genotypes (McSpadden

Gardener et al., 2000; Landa et al., 2002a, 2005) and 70 phlD

sequences from isolates of 20 genotypes. Because all phlD1

fluorescent pseudomonads belong to ARDRA groups 1, 2 or

3 (Keel et al., 1996; McSpadden Gardener et al., 2000), and

all such isolates that have been identified to species are

Pseudomonas fluorescens (Shanahan et al., 1992; Keel et al.,

1996; Lee & Kim, 2001; Raaijmakers & Weller, 2001; Ramette

et al., 2001; Weller et al., 2002), we refer to all strains used in

this study as P. fluorescens. Except for Phl1C2, provided by

P. Lemanceau (Universite de Bourgogne, Dijon, France), all

of the strains have been described previously (Table 1).

Unless indicated (Table 1), genotypes were determined

previously by BOX-PCR and/or RFLP analysis of phlD

(McSpadden Gardener et al., 2000; Mavrodi et al., 2001;

Landa et al., 2002a, 2005). Because both approaches provide

similar results (Mavrodi et al., 2001; McSpadden Gardener

et al., 2001; Landa et al., 2002a, 2005), we commonly use the

term BOX-PCR genotype to refer to genotypes determined

by either approach.

All strains were cultured at 28 1C on one-third strength

King’s medium B (1/3� KMB) agar (King et al., 1954).

Stock cultures were stored in 1/3�KMB broth plus 20%

glycerol at � 80 1C. When appropriate, kanamycin (25

mg mL�1) and gentamicin (12mg mL�1) were added to the

medium.

FEMS Microbiol Ecol 56 (2006) 64–78 c� 2006 Federation of European Microbiological SocietiesPublished by Blackwell Publishing Ltd. No claim to original US government works.

65Diversity and detection of genotypes of Pseudomonas fluorescens

Table 1. Bacterial strains and isolates of 2,4-diacetylphloroglucinol (2,4-DAPG)-producing Pseudomonas fluorescens and phlD gene sequences used in

this study

Strain

Source�

(crop, country) Genotypew Referencez

phlD sequence

Accession number Reference‰

Pf-5 Cotton, USA A Howell & Stipanovic

(1979)

AF214457 McSpadden

Gardener et al. (2000)

CHA0 Tobacco,

Switzerland

A Stutz et al. (1986) AJ278806 Ramette et al. (2001)

MVP1-3 Pea, USA A Landa et al. (2002a) AY928628 This study

PGNR1 Tobacco, Ghana A Keel et al. (1996) AY928629 This study

PINR2 Tobacco, Italy A Keel et al. (1996) AY928630 This study

UP61 Birdsfoot trefoil,

Uruguay

A Bagnasco et al. (1998) AF214108 De La Fuente et al. (2004)

Q2-87 Wheat, USA B Vincent et al. (1991) PFU41818 Bangera &

Thomashow (1999)

Q2-87Kmz Wheat, USA B Validov et al. (2005) NA�� –

Q1-87 Wheat, USA B Keel et al. (1996) AY928631 This study

Q4-87, Q9-87, Q12-87, Q13-87 Wheat, USA B Keel et al. (1996) NA –

Q2-1 Wheat, USA B McSpadden Gardener

et al. (2000)

AY928632 This study

STAD384-97 Wheat, USA C Mavrodi et al. (2001) AY928633 This study

STAD376-97 Wheat, USA C Mavrodi et al. (2001) AY928634 This study

P2112 Red pepper, Korea C Lee & Kim (2001) AY928635 This study

Q8r1-96 Wheat, USA D Raaijmakers &

Weller (1998)

AF207693 McSpadden

Gardener et al. (2000)

Q8r1-96Gmz Wheat, USA D Validov et al. (2005) NA –

L5.1-96 Wheat, USA D Landa et al. (2003) AY928636 This study

MVP1-6 Pea, USA D Landa et al. (2002a) AY928637 This study

FTAD1R34 Wheat, USA D Mavrodi et al. (2001) AY928638 This study

ATCC49054 Potato, Russia D Smirnov &

Kiprianova (1990)

AY928639 This study

FFL1R9, Q128-87, Q2-5,

Q8r10-96, Q8V8, QT1-5,

OC4-1

Wheat, USA D Mavrodi et al. (2001) NA –

11–18 Maize, France Dww Picard & Bosco (2003) AF396848 Picard & Bosco (2003)

4MA30, MVP4-54, MVW1-2,

MVW2-2, 2RP17, 1WSU3,

5WSU2

Wheat or pea, USA D Landa et al. (2002a) NA –

Q2-2 Wheat, USA E McSpadden

Gardener et al. (2000)

AY928640 This study

Q37-87 Wheat, USA E Keel et al. (1996) AY928641 This study

QT1-6 Wheat, USA E Keel et al. (1996) NA –

5MR2 Pea, USA E Landa et al. (2002a) AY928642 This study

JMP6 Wheat, Netherlands F McSpadden

Gardener et al. (2000)

AY928643 This study

JMP7 Wheat, Netherlands F McSpadden

Gardener et al. (2000)

AY928644 This study

FFL1R18 Wheat, USA G McSpadden

Gardener et al. (2000)

AY928645 This study

CV1-1 Wheat, USA H McSpadden

Gardener et al. (2000)

AY928646 This study

FTAD1R36 Wheat, USA I McSpadden

Gardener et al. (2000)

AY928647 This study

FFL1R22 Wheat, USA J McSpadden

Gardener et al. (2000)

AY928648 This study

FFL1R8 Wheat, USA J McSpadden

Gardener et al. (2000)

AY928649 This study

F113 Sugar beet, Ireland K Shanahan et al. (1992) AJ278811 Ramette et al. (2001)

1M1-96 Wheat, USA L Raaijmakers &

Weller (2001)

AF207692 McSpadden

Gardener et al. (2000)

W4-4 Wheat, USA L AY928650 This study

FEMS Microbiol Ecol 56 (2006) 64–78c� 2006 Federation of European Microbiological SocietiesPublished by Blackwell Publishing Ltd. No claim to original US government works.

66 L. De La Fuente et al.

Table 1. Continued.

Strain

Source�

(crop, country) Genotypew Referencez

phlD sequence

Accession number Reference‰

McSpadden

Gardener et al. (2000)

1MA1 Pea, USA L Landa et al. (2002a) AY928651 This study

Phl1C2 Tomato, France Mww Uz AY928652 This study

D27B1 Wheat, Netherlands M McSpadden

Gardener et al. (2000)

AY928653 This study

PILH1 Tomato, Italy M Keel et al. (1996) AJ278810 Ramette et al. (2001)

HT5-1 Wheat, USA N McSpadden

Gardener et al. (2000)

AY928654 This study

7MA12 Pea, USA O Landa et al. (2002a) AY548378 This study

7MA20 Pea, USA O Landa et al. (2002a) AY928660 This study

7MA15 Pea, USA P Landa et al. (2002a) AY928655 This study

6WSU4 Pea, USA P Landa et al. (2002a) AY928656 This study

4MA6 Pea, USA P Landa et al. (2002a) AY928657 This study

6MA17 Pea, USA P Landa et al. (2002a) AY928658 This study

MVP1-4 Pea, USA P Landa et al. (2002a) AY548379 This study

MVP1-4Kmz Pea, USA P Validov et al. (2005) NA –

MVW1-1 Wheat, USA P Landa et al. (2002a) AY548380 This study

7MA16, 8MA21,

MVP1-5, MVW1-3, 8WSU5

Pea or wheat, USA P Landa et al. (2002a) NA –

MVW4-2 Wheat, USA Q Landa et al. (2002a) AY548381 This study

MVW4-3 Wheat, USA Q Landa et al. (2002a) AY928659 This study

FFL2R4 Wheat, USA T Landa et al. (2005) AY928626 This study

FFL2R5 Wheat, USA T Landa et al. (2005) AY928622 This study

FFL2R54 Wheat, USA T Landa et al. (2005) AY928624 This study

FFL2R55 Wheat, USA T Landa et al. (2005) AY928623 This study

FFL2R61 Wheat, USA T Landa et al. (2005) AY928625 This study

FFL3R49 Wheat, USA T Landa et al. (2005) AY928627 This study

42–36 Maize, France NDzz –z AF396857 Picard & Bosco (2003)

42–27 Maize, France ND – AF396856 Picard & Bosco (2003)

39–8 Maize, France ND – AF396855 Picard & Bosco (2003)

37–27 Maize, France Dww – AF396854 Picard & Bosco (2003)

22–27 Maize, France ND – AF396853 Picard & Bosco (2003)

19–41 Maize, France ND – AF396852 Picard & Bosco (2003)

19–30 Maize, France ND – AF396851 Picard & Bosco (2003)

19–7 Maize, France ND – AF396850 Picard & Bosco (2003)

18–33 Maize, France ND – AF396849 Picard & Bosco (2003)

7–37 Maize, France ND – AF396847 Picard & Bosco (2003)

6–28 Maize, France ND – AF396846 Picard & Bosco (2003)

3–1 Maize, France ND – AF396845 Picard & Bosco (2003)

Q65c-80 Wheat, USA ND – AJ278807 Ramette et al. (2001)

PITR2 Wheat, Italy ND – AJ278809 Ramette et al. (2001)

CM10A2 Cucumber,

Switzerland

ND – AJ278808 Ramette et al. (2001)

HP72 Bentgrass, Japan ND – AB125213 –

HR3-A13 Wheat, USA PfY – AY391780 Mazzola et al. (2004)

PR3-A52 Wheat, USA PfZ – AY391779 Mazzola et al. (2004)

2–79‰‰ Wheat, USA – Weller & Cook (1983) – –

�Source of Pseudomonas fluorescens isolated from the rhizosphere of various crops.wGenotype as determined previously by BOX-PCR (McSpadden Gardener et al., 2000; Landa et al., 2002a) and phlD restriction fragment length

polymorphism (RFLP) analysis (McSpadden Gardener et al., 2001; Mavrodi et al., 2001; Landa et al., 2002a).zReferences are shown only for strains used for specificity screening of the allele-specific primers. U, unpublished; –, not tested.‰References in which phlD sequences were published. –, unpublished phlD sequence.zAntibiotic-resistant derivatives marked with mini-Tn7gfp1 or mini-Tn7gfp2 that confer resistance to kanamycin (Q2-87Km, MVP1-4Km) or gentamicin

(Q8r1-96Gm), respectively.��NA, phlD sequence not available.wwGenotype determined by phlD RFLP and BOX-PCR (B.B. Landa and D.M. Weller, unpublished results, 2002), and confirmed in this study by allele-specific PCR.zzND, classification by BOX-PCR or phlD RFLP analysis was not determined.‰‰Pseudomonas fluorescens 2-79 does not contain phlD and was used as a negative control in PCRs.

FEMS Microbiol Ecol 56 (2006) 64–78 c� 2006 Federation of European Microbiological SocietiesPublished by Blackwell Publishing Ltd. No claim to original US government works.

67Diversity and detection of genotypes of Pseudomonas fluorescens

Amplificationandsequencingof phlDgenefragments

A 745 bp phlD gene fragment from each strain was amplified

by PCR using the ‘generic’ primers Phl2a and Phl2b (Raaij-

makers et al., 1997). The amplification products were gel

purified with a QIAEX II Gel Extraction Kit (Qiagen,

Valencia, CA) and sequenced using an ABI PRISM BigDye

Terminator Cycle Sequencing Ready Reaction Kit (Applied

Biosystems, Foster City, CA). Cycle sequencing was per-

formed with primers B2BF and BPR4 (McSpadden Garden-

er et al., 2001) and Phl2a and Phl2b (Raaijmakers et al.,

1997). The phlD DNA sequences were assembled using the

Lasergene package (DNA Star Inc., Madison, WI). Consen-

sus sequence quality was checked using the translating BLAST

function BLASTX (Altschul et al., 1997) (National Center for

Biotechnology Information, NCBI). GenBank accession

numbers for the resulting phlD sequences are shown in

Table 1.

Phylogenetic analysis

Thirty-three phlD sequences (Table 1) were identified in the

non-redundant GenBank dataset (NCBI) using the www-

based NCBI nucleotide BLAST search engine (Altschul et al.,

1997) with default parameters and the 2,4-DAPG biosyn-

thetic locus of P. fluorescens Q2-87 as the query sequence.

Sequences longer than 600 bp, together with the 37 phlD

sequences obtained in this study (Table 1), were trimmed to

an identical length of 599 bp and aligned using the CLUSTALW

service at the European Bioinformatics Institute (URL

http://www.ebi.ac.uk/clustalw). Phylogenetic analyses were

performed with PHYLIP 3.62 (Felsenstein, 1989) using the

neighbour-joining method and, for comparison purposes,

with the maximum parsimony method. For both methods,

both DNA and protein sequences were used. The confidence

of the topology of the neighbour-joining phylogenetic tree

for the phlD gene was assessed by bootstrap analysis with

1000 resamplings. Phylogenetic trees were visualized with

TreeView (Win32) (URL http://taxonomy.zoology.gla.ac.uk/

rod/rod.html).

Allele-specific PCR

phlD allele-specific primers for detection of the six most

studied BOX-PCR genotypes were designed on the basis of

phlD sequences from: P. fluorescens UP61, CHA0 and PF-5

(genotype A); Q8r1-96 and 11-18 (genotype D); F113

(genotype K); 1M1-96 (genotype L); PILH1 (genotype M);

7MA12 (genotype O); MVP1-4 and MVW1-1 (genotype P);

and MVW4-2 (genotype Q) (Table 1). The sequences were

aligned using CLUSTALW, and the alignment was used to

design allele-specific PCR primers (Table 2) with the aid of

OLIGO V.6.65 software (Molecular Biology Insights Inc.,

Cascade, CO) according to the following criteria: (1)

oligonucleotide length of 18–25 bp; (2) melting temperature

(Tm) of Z55 1C with differences of � 13.5 1C between

primer pairs; (3) five to seven nucleotides at the 30 end

including at least one allele-specific nucleotide; (4) 30

terminal nucleotide different from A; (5) no predicted

hairpin loops, duplexes or primer-dimers; and (6) amplifi-

cation product of Z300 bp. An in silico test of primer

specificity was conducted by comparing the primer se-

quences with the non-redundant GenBank dataset with

parameters set for the identification of short, nearly exact

matches.

Conditions for allele-specific amplification of phlD initi-

ally were determined using purified genomic DNA from

strains of the A, B, D, K, L and P genotypes as template.

PCRs contained 1�Taq reaction buffer (Promega Corp.,

Madison, WI), 1.5 mM MgCl2, 200 mM of each dATP, dCTP,

dGTP, dTTP (Perkin-Elmer, Norwalk, CT), 20 pmol of each

primer and 1.2 U of Taq DNA polymerase (Promega Corp.).

Table 2. PCR primers used in this study

Primer Sequence Tm ( 1C)a Positionb Reference

A-align TGT TGC CGG CTT CCG ACA GGC TGC 86.6 1944–1967 This study

A-gen-exten GAC ATT CGC ATG GTG ATC GTC ACT 76.0 2545–2571cc This study

B-genlong CTT GAA TCC CGG CGC CTT ATC G 77.0 2239–2260 This study

D-gen-2 GCA GGA AGT AAG ACC CGG TA 64.2 2212–2231 This study

K-gen ACC ACC GAC GAC ATT CGA 64.0 2563–2580c This study

L1-exten CTC GCT ATC AGG CAG GAA GTA AGA A 79.5 2200–2224 This study

P-gen CGA GCC AAT GAT TTT GCT 64.5 2403–2421c This study

B2BF ACC CAC CGC AGC ATC GTT TAT GAG C 65.6 2640–2664c McSpadden

Gardener etal., 2001

BPR4 CCG CCG GTA TGG AAG ATG AAA AAG TC 63.4 2036–2061 McSpadden

Gardener etal., 2001

aMelting temperature based on the ‘nearest-neighbour’ thermodynamic method as predicted by the OLIGO v. 6.65 primer analysis software.bPosition refers to the sequence of the phl operon from Pseudomonas fluorescencens Q2-87 (Accession number PFU41818).cPrimer corresponds to complementary DNA strand.

FEMS Microbiol Ecol 56 (2006) 64–78c� 2006 Federation of European Microbiological SocietiesPublished by Blackwell Publishing Ltd. No claim to original US government works.

68 L. De La Fuente et al.

Amplification was performed with a PTC-200 thermal cycler

(MJ Research, Watertown, MA). The cycling programme

consisted of initial melting at 95 1C for 180 s, followed by 35

cycles of melting at 94 1C for 60 s, annealing at 60 1C for 60 s

and extension at 72 1C for 60 s, and a final extension at 72 1C

for 300 s. The annealing temperature and MgCl2 and

dimethylsulphoxide (DMSO) concentrations subsequently

were adjusted as shown in Table 3. Specificity was evaluated

by whole-cell PCR from 1-day-old single colonies of

70 strains (Table 1) representative of all 18 genotypes.

Freeze–thawed cell suspensions (2.5mL) were used as tem-

plates when samples from serial dilution plates or broth

cultures were analysed. PCR amplification products

were separated in 0.8% agarose gels containing ethidium

bromide (0.5mg mL�1).

Detectionofgenotypes inmixedbroth culturesbyallele-specific PCR

Strains were grown individually and in pairwise combina-

tions in vitro using a protocol similar to that of Landa et al.

(2003) in order to determine the detection limit of each

genotype by phlD-directed allele-specific PCR. Analyses

included P. fluorescens Pf-5, Q2-87, Q8r1-96, F113, 1M1-96

and MVP1-4 alone; P. fluorescens Q8r1-96 plus Pf-5, Q2-87,

F113, 1M1-96 or MVP1-4; and P. fluorescens F113 plus

MVP1-4. All strains have previously been characterized with

regard to their rhizosphere colonization (Landa et al., 2002a,

2003), and strain pairs were chosen based on competition

studies performed in the rhizosphere in this study and

others (L. De La Fuente and D.M. Weller, unpublished data,

2003). Overnight cultures of each strain were harvested by

centrifugation in a microfuge (10 000 g, 5 min), suspended

in 1/3�KMB to an absorbance at 600 nm (OD600 nm) of 0.1

(c. 1� 108 CFU mL�1) using a SAFIRE microplate reader

(Tecan Systems, Inc., San Jose, CA) and added (250mL per

well) to 96-well microtitre plates. For pairs of strains, 125 mL

of appropriate tenfold serial dilutions was added to the wells

to obtain mixture ratios of 1 : 1, 1 : 10, 10 : 1, 1 : 100 and

100 : 1. Samples were replicated three times and the entire

experiment was conducted twice. The plates were incubated

at room temperature for 72 h and bacterial growth was

monitored at 24 h intervals by measuring OD600 nm. Samples

(20mL) were taken at 0, 24, 48 and 72 h from each well, and

freeze–thawed prior to analysis by PCR with primers specific

for each genotype in the culture.

Inhibition betweengenotypes inmixedbrothcultures

Viable population sizes were determined as plate counts in

order to establish whether the failure to detect certain

genotypes by PCR when they were grown in combination

resulted from the inhibition of one strain by another or from

ineffective PCR amplification with the allele-specific pri-

mers. The experiment was performed with derivatives of

strains Q2-87, Q8r1-96 and MVP1-4 marked with mini-

Tn7gfp1 or mini-Tn7gfp2 (Koch et al., 2001), which confer

resistance to kanamycin (Q2-87Km, MVP1-4Km) or genta-

micin (Q8r1-96Gm), respectively (Validov et al., 2005). The

use of tagged derivatives allowed us to differentiate between

the competing strains by plating on selective media contain-

ing the corresponding antibiotics. Cultures (4 mL) of in-

dividual strains or mixtures (Q8r1-96Gm plus Q2-87Km or

MVP1-4Km) at ratios of 1 : 1, 1 : 10 and 10 : 1 in 1/3� KMB

broth were incubated on a rotary shaker (200 r.p.m.) at

28 1C for 72 h, and viable cell numbers were scored at 24-h

intervals by plating on 1/3� KMB agar amended with either

kanamycin or gentamicin. Three repetitions were used for

each measurement, and the experiment was repeated three

times.

Genotypedetection byallele-specific PCR in soiland rhizospheresamples

Materials from a previous study (Landa et al., 2003) were

employed to demonstrate that allele-specific primers can

Table 3. PCR conditions for the detection of different genotypes of 2,4-diacetylphloroglucinol (2,4-DAPG)-producing Pseudomonas species using

allele-specific primers

Target BOX-PCR genotype

Allele-specific primers

DTm ( 1C)a Tann ( 1C)b Optimal reaction conditions Amplicon size (bp)Upper Lower

A A-align A-gen-exten 10.6 62.2 1.5 mM MgCl2 628

B B-genlong B2BF 2.5 60.0 1.5 mM MgCl2 421

D D-gen-2 B2BF 1.4 66.0 1.5 mM MgCl2 428

K BPR4 K-gen 13.5 63.0 1.5 mM MgCl214% DMSOc 544

L L1-exten B2BF 0.0 66.0 1.0 mM MgCl2 460

P BPR4 P-gen 13.0 65.0 1.5 mM MgCl2 385

aDifference between melting temperatures (Tm) within a given primer pair based on the ‘nearest-neighbour’ thermodynamic method as predicted by

the OLIGO primer analysis software.bOptimal annealing temperature found experimentally.cDMSO, dimethylsulphoxide.

FEMS Microbiol Ecol 56 (2006) 64–78 c� 2006 Federation of European Microbiological SocietiesPublished by Blackwell Publishing Ltd. No claim to original US government works.

69Diversity and detection of genotypes of Pseudomonas fluorescens

quantify 2,4-DAPG producers of multiple genotypes in soil

and on roots. These materials were selected because they

allowed a direct comparison of population sizes detected by

the original PCR-based dilution endpoint assay (McSpad-

den Gardener et al., 2001) and our new approach with allele-

specific primers. In the original experiment (Landa et al.,

2003), a Shano silt loam was inoculated with rifampicin-

resistant derivatives of Q8r1-96 and F113, individually

or as a 1 : 1 mixture, each at c. log 4 CFU per gram of soil.

Three treatments were then established. The first consisted

of pots filled with soil inoculated with the bacteria and sown

with wheat (Triticum aestivum L. cv. Penawawa) seeds. After

3 weeks (referred to as one ‘cycle’), the plants were

harvested, roots with adhering rhizosphere soil (designated

RRS) were assayed for the introduced bacteria, and the soil

and remaining roots were returned to the pots and sown

again to begin the next cycle. Wheat was grown for a

total of five 3-week cycles and the roots were assayed for

the introduced bacteria after each cycle (see below).

The second treatment was identical to the first, except that

the roots were not incorporated into the soil after each

cycle, and the soil without roots (designated non-rhizo-

sphere soil, NRS) was assayed for the introduced bacteria.

The third treatment consisted of pots filled with inoculated

soil and watered weekly, but in which no plants were

sown. At 3-week intervals, the soil was removed, shaken

and returned to the pots, and the samples (designated bulk

soil, BS) were processed as described previously (Landa

et al., 2003).

Densities of Q8r1-96 and F113 in all three sets of samples

were determined by Landa et al. (2003) using the PCR-based

dilution endpoint technique (McSpadden Gardener et al.,

2001) on six randomly selected roots (repetitions) from each

treatment. Briefly, root samples were vortexed in water,

three-fold serial dilutions were made in microtitre dishes,

and aliquots from each well were transferred to another dish

containing 1/3� KMB1111 broth amended with chloram-

phenicol, ampicillin, cycloheximide and rifampicin. Growth

in the last turbid well (terminal dilution culture, TDC),

detected by measuring OD600 nm, is considered to have

originated from a single cell. The presence of phlD in the

TDC was confirmed by PCR with ‘generic’ primers B2BF/

BPR4, and the most abundant genotype was identified by

phlD RFLP analysis. The population sizes were calculated by

applying the formula [Vi� 200� 3n/fresh weight root (g)],

where Vi is the volume of water used to suspend the root

system and n is the number of dilutions corresponding to

the TDC. McSpadden Gardener et al. (2001) discussed the

mathematical basis of the assay. Population sizes of phlD1

isolates quantified by the dilution endpoint assay were

identical to those determined by dilution plating (Landa

et al., 2002b). Thus, we continue to express population sizes

as CFU for convenience.

As reported by Landa et al. (2003), amplification pro-

ducts from treatments in which strain Q8r1-96 or F113 had

been introduced individually produced the phlD RFLP

profiles expected for each strain. However, in mixed treat-

ments of Q8r1-96 plus F113, only the F113 profile was

detected because the density of Q8r1-96 was lower than that

of F113. Here, we used the allele-specific primers and the

original broth cultures that had been assayed previously

(Landa et al., 2003), and stored at � 80 1C for over 18

months, to determine the population sizes of both F113 and

Q8r1-96. Validation of the utility of the primers would

require that the allele-specific primers would detect the

same population sizes of Q8r1-96 and F113 as the original

dilution endpoint assay.

Dataanalysis

Population data from the soil studies were converted to log

CFU per gram of soil or root fresh weight to satisfy the

assumptions of analysis of variance (ANOVA). The coloniza-

tion of the rhizosphere or soil by each strain introduced

individually, or as a mixture, was determined by two

methods. The area under the colonization progress curve

(AUCPC) (Landa et al., 2002a) was calculated by using the

trapezoidal integration method (Campbell & Madden,

1990). The mean colonization was calculated as the mean

population density across all cycles (Landa et al., 2002a).

Differences in population densities between treatments were

determined by ANOVA and mean comparisons between treat-

ments were performed by Fischer’s protected least signifi-

cant difference (LSD) test at P = 0.05 using STATISTIX 8.0

software (Analytical Software, St. Paul, MN). For soil or

rhizosphere samples in which no PCR signal was detected,

the population density was set at the detection limit of the

allele-specific PCR assay (log 3.26 CFU per gram of sample).

Results

phlD sequences and phylogenetic analyses

The 37 2,4-DAPG-producing Pseudomonas fluorescens

strains from which phlD was sequenced in this study are of

diverse host crop and geographical origins (Table 1). These

sequences, together with those from other studies (McSpad-

den-Gardener et al., 2000; Mavrodi et al., 2001; Landa et al.,

2002a, 2005; Mazzola et al., 2004), were used for phyloge-

netic analysis of 20 of the 22 BOX-PCR genotypes described

to date (A–Q, T, PfY, PfZ). When available, phlD sequences

from at least three strains from each genotype were in-

cluded, although genotypes G, H, I, K and N currently are

represented by a single strain. Comparison of sequences

from the 20 BOX-PCR genotypes revealed identities ranging

from 76.0 to 99.7%, confirming previous observations (Keel

FEMS Microbiol Ecol 56 (2006) 64–78c� 2006 Federation of European Microbiological SocietiesPublished by Blackwell Publishing Ltd. No claim to original US government works.

70 L. De La Fuente et al.

et al., 1996; McSpadden Gardener et al., 2001; Ramette et al.,

2001) that phlD is highly conserved amongst 2,4-DAPG-

producing P. fluorescens. A neighbour-joining tree (Fig. 1)

comprising phlD sequences from 70 strains showed that the

previous classifications based on BOX-PCR and phlD RFLP

(McSpadden-Gardener et al., 2000, 2001; Mavrodi et al.,

2001; Landa et al., 2002a, 2005) correlated well with clusters

obtained by the alignment of 599 bp of the phlD gene

sequence. The same tree topology was obtained by inferring

the phylogeny with the neighbour-joining method applied

to protein sequence data and with maximum parsimony

applied to both DNA and protein sequences.

Allele-specific primerdesign

Thirteen phlD sequences from strains of BOX-PCR genotypes

A, B, D, K, L, M, O, P and Q were aligned to detect differences

that could be exploited to design allele-specific primers.

Fig. 1. Unrooted neighbour-joining tree based on alignment of partial phlD DNA sequences from 70 Pseudomonas fluorescens strains. Strains are

representative of 20 of the 22 BOX-PCR 2,4-diacetylphloroglucinol (2,4-DAPG)-producing genotypes (letters in parentheses) identified to date. No

letters after strain names correspond to non-determined genotype classifications by BOX-PCR or phlD restriction fragment length polymorphism (RFLP)

analysis. Bootstrap values from 1000 resamplings are indicated for each branch as percentages, and values below 50% were omitted. Scale bar

indicates the number of nucleotide substitutions per site.

FEMS Microbiol Ecol 56 (2006) 64–78 c� 2006 Federation of European Microbiological SocietiesPublished by Blackwell Publishing Ltd. No claim to original US government works.

71Diversity and detection of genotypes of Pseudomonas fluorescens

Despite the overall sequence similarity, sufficient differences

were identified and used to generate primer sequences that

had 100% identity only with the targeted genotype-specific

phlD alleles (Tables 2 and 3). The six pairs of allele-specific

primers amplified phlD in all representatives of the target

genotypes in the 70 strains (Table 1) screened. Two instances

of nonspecific amplification were also observed: genotype K-

specific primers amplified phlD in strains of genotypes C and

M, and genotype L-specific primers amplified phlD in strains

of genotype Q (data not shown).

Sensitivityofdetection invitro

The sensitivity of detection of allele-specific PCR was assessed

in broth cultures containing paired strains of different

genotypes. Strains of genotypes K (F113) and P (MVP1-4)

were detected by their respective primers at 24, 48 and 72 h

whether grown individually or together and regardless of the

ratio at which they were inoculated. Similarly, strains of

genotypes A (Pf-5), B (Q2-87), K, L (1M1-96) and P were

detected at 24, 48 and 72 h whether grown individually or

after co-inoculation with strain Q8r1-96 (genotype D) at

ratios of 1 : 1 or 10 : 1. The detection limit in such mixtures

was comparable with that when the strains were grown

individually, c. log 5.7 cells mL�1 for the B-, D-, K- and P-

genotype strains and about 10-fold higher for the A- and L-

genotype strains. The OD600 nm values for cultures of indivi-

dual strains and strain mixtures did not differ significantly

over the course of the 72-h experiment (data not shown).

When Q8r1-96 (genotype D) was inoculated at a ratio of

10 : 1 with other strains, both it and strains of the A, K and P

genotypes were detected at 24, 48 and 72 h. In contrast, the

B-genotype strain Q2-87 was detected only at t = 0 and the

L-genotype strain 1M1-96 was not detected at any sampling

time. When Q8r1-96 (genotype D) was introduced in 100-

fold excess, neither Q2-87 (genotype B) nor 1M1-96 (geno-

type L) was detected at any time. However, at a ratio of

1 : 100 (Q8r1-96 to any other strain), Q8r1-96 (genotype D)

was always detected.

Impactof competition in broth culturesonviablecell populations

Strains were paired on the basis of an expectation of

significant growth inhibition [Q8r1-96Gm (genotype D)

and Q2-87Km (genotype B)] or little growth inhibition

[Q8r1-96Gm (genotype D) and MVP1-4Km (genotype P)].

P. fluorescens Q8r1-96Gm (genotype D) inhibited the

growth of Q2-87Km (genotype B) at ratios of 1 : 1 and

10 : 1. Plating on selective media revealed that, when Q8r1-

96Gm (genotype D) was the more abundant strain in the

inoculum (ratio of 10 : 1) (Fig. 2), the density of Q2-87Km

(genotype B) after 24 h was nearly nine orders of magnitude

less than when it was cultured alone. However, when the

strains were inoculated at a ratio of 1 : 1, populations of Q2-

87Km (genotype B) decreased by only log 2–3 CFU mL�1.

No significant inhibition occurred when Q2-87Km (geno-

type B) was the more abundant component of the inoculum

[Q8r1-96Gm (genotype D) and Q2-87Km (genotype B) at a

ratio of 1 : 10]. The results were consistent when the experi-

ment was repeated.

A similar but less severe inhibition of MVP1-4Km

(genotype P) by Q8r1-96Gm (genotype D) was observed

when these two strains were grown together. The effect was

most pronounced when the ratio of Q8r1-96Gm (genotype

D) to MVP1-4Km (genotype P) was 10 : 1, which caused the

density of MVP1-4Km (genotype P) to decrease by about

log 4 CFU mL�1 when compared with its population when

grown alone. The population density of MVP1-4Km (geno-

type P) stabilized close to log 6 CFU mL�1.

Detectionof2,4-DAPG-producingPseudomonasfluorescens in the soil and rhizospherebyallele-specificPCR

Broth cultures from dilutions of root washings obtained in a

previous study (Landa et al., 2003) were analysed by PCR

with primers specific for D- and K-genotype strains. First, we

determined whether these primers were as efficient as B2BF

and BPR4, the generic phlD primers employed in the

standard PCR-based dilution endpoint assay (McSpadden

Gardener et al., 2001). Population densities of F113 (geno-

type K) (calculated on the basis of results with the K-

genotype primers), introduced alone or in combination with

Q8r1-96 (genotype D), in the rhizosphere (RRS, Fig. 3d),

bulk soil subjected to wheat cropping (NRS, Fig. 3e) and bulk

soil without cropping (BS, Fig. 3f) were identical to those

reported originally by Landa et al. (2003) using B2BF/BPR4

and phlD RFLP analysis. Similarly, the population densities of

Q8r1-96 (genotype D) (calculated on the basis of results with

the D-genotype primers) introduced alone in the RRS, NRS

and BS treatments (Figs 3a–c) were identical to those

reported earlier (Landa et al., 2003). The detection limit with

the D- and K-genotype-specific primers was log 3.26 CFU per

gram of sample, equivalent to that for the generic phlD

primers B2BF and BPR4 (McSpadden Gardener et al., 2001;

Landa et al., 2002b). These results support our observations

from in vitro studies, showing that the genotype-specific

primers are as effective as the generic primers.

Next, we determined whether the D-genotype-specific

primers could detect Q8r1-96 (genotype D) when it was not

the dominant genotype in a mixture. In mixed treatments in

which F113 (genotype K) was the dominant strain, Landa

et al. (2003) could not detect Q8r1-96 (genotype D);

however, we consistently detected Q8r1-96 in the RRS, NRS

and BS samples (Figs 3a–c; Table 4). In samples from all

three treatments, the population density of Q8r1-96

FEMS Microbiol Ecol 56 (2006) 64–78c� 2006 Federation of European Microbiological SocietiesPublished by Blackwell Publishing Ltd. No claim to original US government works.

72 L. De La Fuente et al.

(genotype D) was significantly reduced by about 10- to

1000-fold in the presence of F113 (genotype K). Further-

more, the AUCPC for Q8r1-96 introduced in combination

with F113 was significantly smaller (Po 0.05) than when it

was introduced alone (Table 4). The negative effects of F113

(genotype K) on the density of Q8r1-96 (genotype D) were

greater in RRS samples than in NRS and BS samples (Fig. 3).

In contrast, population densities of F113 were not affected

by the presence of Q8r1-96 (Figs 3a–c; Table 4) in samples

from any of the three treatments, and were identical to those

reported by Landa et al. (2003). These results further

validate the utility of the genotype-specific primers in

detecting and quantifying subordinate genotypes in mixed

populations.

Discussion

With the 37 strains sequenced in this study (Table 1), a

comprehensive collection of 70 phlD sequences from 2,4-

DAPG producers of diverse origin is now available. As in

previous studies based on phlD RFLP (Mavrodi et al., 2001;

Ramette et al., 2001), clusters identified with our phlD

sequence-based phylogenetic tree (Fig. 1) did not correlate

with the geographical origin or the host crop from which the

strains were isolated (Table 1). More importantly, however,

the clusters correspond to the BOX-PCR genotype classifi-

cation previously described by our laboratory (McSpadden

Gardener et al., 2000, 2001; Mavrodi et al., 2001; Landa

et al., 2002a, 2005). Therefore, it is reasonable to assume that

strains not previously characterized by BOX-PCR, but

found in the same sequence cluster with well-defined strains,

can be considered members of the same genotype. For

example, strains 11-18, 37-27, 42-27, 22-27, 7-37, 6-28, 18-

33, 39-8, 42-36, 19-41 and 3-1, isolated in France from

maize (Picard & Bosco, 2003), most probably belong to

BOX-PCR genotype D. This prediction is supported by the

fact that strains 11-18 and 37-27 were confirmed to be

genotype D by allele-specific PCR and phlD RFLP analyses

Fig. 2. Inhibition in vitro between genotypes of 2,4-diacetylphloroglucinol (2,4-DAPG)-producing Pseudomonas fluorescens. Strain Q8r1-96Gm

(genotype D) was inoculated into one-third strength King’s medium B (1/3�KMB) broth with either Q2-87Km (genotype B) or MVP1-4Km (genotype P)

at ratios of 1 : 1, 1 : 10 and 10 : l. Cultures were incubated at room temperature for 72 h and the density of each strain was determined every 24 h by

dilution plating. Filled symbols correspond to single inoculation treatments (�, D; ., B or P) and open symbols correspond to mixed treatments (�, D; ,,

B or P). The genotype of each strain in a mixture (D : B, D : P) and the relative amounts of each are shown. The experiment was repeated three times.

Mean values from a representative experiment are shown. Standard deviation bars are not shown because their sizes were smaller than the symbols.

FEMS Microbiol Ecol 56 (2006) 64–78 c� 2006 Federation of European Microbiological SocietiesPublished by Blackwell Publishing Ltd. No claim to original US government works.

73Diversity and detection of genotypes of Pseudomonas fluorescens

(B.B. Landa and D.M. Weller, unpublished results, 2004).

Therefore, to our knowledge, these are the first D-genotype

isolates found in Europe. Previously, D-genotype strains had

only appeared in US collections (McSpadden-Gardener

et al., 2000, 2005; Landa et al., 2002a, 2005); thus, they

appear to be more widely distributed than previously

thought. In addition, according to the neighbour-joining

tree topology, strain PITR2 may belong to the M genotype

(Fig. 1). On the other hand, a strain such as HP72 appears to

define an independent branch of the tree (Fig. 1), and most

probably represents a new genotype.

Ramette et al. (2001) used phlD sequences from seven 2,4-

DAPG-producing strains to describe three clusters, two of

which correspond to the A- and B-genotype strains, respec-

tively, and a third cluster encompassing strains from K, M

and unidentified BOX-PCR genotypes. Recently, Bergsma-

Vlami et al. (2005) have defined six clusters amongst 37 2,4-

DAPG-producing strains using sequence data from a shorter

fragment (325 bp) of the phlD sequence. Their clustering did

not match their genotype classification by DGGE. In our

study, the use of a larger 599-bp fragment from 70 strains

allowed the differentiation of a larger number of clusters

that do match our BOX-PCR classification (McSpadden-

Gardener et al., 2000; Landa et al., 2002a, 2005). Mazzola

et al. (2004) used a phlD sequence-based tree of 25 strains to

identify two new genotypes, PfY and PfZ. Our tree (Fig. 1)

establishes the sequence of phlD as a valid criterion for

classifying 2,4-DAPG producers according to the BOX-PCR

genotype previously defined in our laboratory (McSpadden-

Gardener et al., 2000; Mavrodi et al., 2001; Landa et al.,

2002a, 2005). This is especially important, as it has pre-

viously been shown that the genotype is predictive of the

colonization ability of a strain on at least some host plants

(Raaijmakers & Weller, 2001; Landa et al., 2002a, 2003).

We exploited naturally occurring variation within the

sequence of phlD to develop allele-specific PCR primers for

the detection of A-, B-, D-, K-, L- and P-genotype strains of

2,4-DAPG-producing Pseudomonas. Using these primers, we

significantly improved the technique developed by McSpad-

den Gardener et al. (2001) by making it suitable for fine-

scale studies of interactions between 2,4-DAPG producers in

the rhizosphere, whilst retaining its high-throughput

Fig. 3. Population dynamics of Pseudomonas fluorescens Q8r1-96 and F113 alone and co-inoculated (Q8r1-96 1 F113) in the wheat rhizosphere (RRS)

(a and d) and in soil in which wheat had been planted (NRS) (b and e) or no plants were grown (BS) (c and f). Natural soil was inoculated with each strain

alone or a 1 : 1 mixture and seeded (RRS, NRS) or not (BS) with wheat. Bacterial populations in the rhizosphere were measured during five consecutive

cycles of 3 weeks each. In each graph, rhizosphere and soil populations of Q8r1-96 (�, �) and F113 (}, ~) strains are compared in single (�, ~) and

mixed (�, }) inoculation treatments. The population of each strain was determined by PCR using allele-specific primers. Mean values and standard

deviations are presented. �Populations below the detection limit (log 3.26 CFU g�1 of sample). Values for the mean colonization and area under the

colonization progress curve (AUCPC) are shown in Table 4.

FEMS Microbiol Ecol 56 (2006) 64–78c� 2006 Federation of European Microbiological SocietiesPublished by Blackwell Publishing Ltd. No claim to original US government works.

74 L. De La Fuente et al.

capabilities. With few exceptions, the allele-specific primers

gave highly specific amplification of the target genotypes.

Amongst the exceptions were the K-genotype primers,

which also detected C- and M-genotype strains, and the

L-genotype primer, which amplified phlD from Q-genotype

strains. Both a perfect match at the 30 end of the primer and

the nature of the terminal base are critical for the specificity

of Taq polymerase-based amplification (Green, 2002). Ac-

cording to Ayyadevara et al. (2000), placement of an A at the

30 end of the oligonucleotide often leads to the lowest

possible degree of discrimination in allele-specific PCR.

Unfortunately, we were forced to use A in the terminal

position in both the K- and L-genotype primers because of

the absence of polymorphic sites within phlD, which may

explain their more limited specificity.

This study refers to two different detection limits in the

allele-specific PCR-based dilution endpoint assay. We first

described the detection limit in broth cultures, which was c.

log 5.7 CFU mL�1 for strains of genotypes B, D, K and P, and

log 6.7 CFU mL�1 for strains of genotypes A and L. McSpad-

den Gardener et al. (2001) found a similar detection limit in

vitro (log 5 CFU mL�1) using the generic primers B2BF and

BPR4. It is difficult to assess exactly how much template

DNA corresponds to these large numbers of cells, because

templates are prepared from freeze–thawed broth cultures in

which only a fraction of the total bacterial population

releases DNA. The fact that the bacterial cultures used in

our study were in the stationary phase of growth when the

PCR was performed may also have influenced the detection

limit (Rezzonico et al., 2003).

We further defined a detection limit of log 3.26 CFU g�1

of root by allele-specific PCR of samples from the rhizo-

sphere. This value is identical to that found by Landa et al.

(2003), who analysed the same samples by PCR with generic

phlD primers. Previous comparisons of three different

techniques to quantify 2,4-DAPG producers in the rhizo-

sphere (Landa et al., 2002b) showed that the detection limit

of the original PCR-based dilution endpoint assay was

comparable with that of traditional dilution plating and

colony hybridization (log 4.0 CFU g�1 of root). Recently, a

proposed alternative technique using DGGE (Bergsma-

Vlami et al., 2005) was shown to have a similar detection

limit (log 3.7 CFU g�1 of root). For our purposes, the speed

and robustness of the dilution endpoint assay outweigh the

drawbacks of a detection limit of about 1000 cells g�1 of

sample.

Given the overall similarity between genotypes of 2,4-

DAPG producers (Keel et al., 1996; McSpadden-Gardener

et al., 2000; Mavrodi et al., 2001), we were surprised to

detect significant inhibition in broth cultures of the B-

genotype strain Q2-87 by the D-genotype strain Q8r1-96.

For example, when Q8r1-96 and Q2-87 were combined in a

ratio of 10 : 1, the density of viable cells of Q2-87 decreased

precipitously from log 6 to log 2 CFU mL�1 after 24 h (Fig.

2). This low density may explain why we could not detect

Q2-87 by PCR when Q8r1-96 was the most abundant

Table 4. Population dynamics of Pseudomonas fluorescens strains F113 and Q8r1-96 on roots of wheat and in bulk soil with or without wheat

cropping�

Sample type Bacterial treatment

Mean colonization (log CFU g�1 of sample)w AUCPCz

Q8r1-96 F113 Q8r1-96 F113

Roots and rhizosphere soil (RRS)

Q8r1-96 7.91 – 38.06 a‰ –

F113 – 7.75 – 37.42 a

Q8r1-961F113 6.40 7.93 31.32 b 38.24 a

Non-rhizosphere soil (NRS)

Q8r1-96 5.78 – 28.49 a –

F113 – 6.17 – 30.42 a

Q8r1-961F113 4.71 5.99 22.99 b 29.37 a

Bulk soil (BS)

Q8r1-96 4.77 – 23.76 b –

F113 – 5.20 – 25.77 a

Q8r1-961F113 3.56 5.00 18.02 c 24.88 ab

�The experiment was conducted in a growth chamber using non-sterile soil. Samples were obtained from the wheat rhizosphere (roots and rhizosphere

soil, RRS), and from soil in which wheat had been planted (non-rhizosphere soil, NRS) or no plants were grown (bulk soil, BS). Natural soil was inoculated

with each strain alone or a 1 : 1 mixture of both strains and seeded (RRS, NRS) or not (BS) with wheat. For details, see ‘Materials and methods’.wMean colonization was calculated as the mean population density across all cycles (Landa et al., 2002a). Population data were converted to log CFU per

gram fresh weight of root.zThe area under the colonization progress curve (AUCPC) was calculated using the trapezoidal integration method (Campbell & Madden, 1990).‰Means in the same column followed by the same letter are not significantly different (P = 0.05) according to Fisher’s protected least significant

difference (LSD) test.

FEMS Microbiol Ecol 56 (2006) 64–78 c� 2006 Federation of European Microbiological SocietiesPublished by Blackwell Publishing Ltd. No claim to original US government works.

75Diversity and detection of genotypes of Pseudomonas fluorescens

component of the mixture. The inhibition of Q2-87 by

Q8r1-96 also occurs in the wheat rhizosphere. Using a de

Wit replacement series, Landa et al. (2003) demonstrated a

competitive disadvantage for strain Q2-87 or strong antag-

onism by Q8r1-96 against Q2-87 in the wheat rhizosphere.

Q8r1-96 produces bacteriocin-like inhibitors active against

Q2-87 (Validov et al., 2005). Theoretical models of interac-

tions between bacteriocin-producing and sensitive strains

have shown that inhibition depends on the relative strain

concentrations (Durret & Levin, 1997). If both are present at

equal densities, the bacteriocin-producing strain cannot

prevail, but, when it is the more abundant strain, it will

antagonize the sensitive strain. This model is consistent with

the observed patterns of inhibition of Q2-87 by strain Q8r1-

96, and reveals a limitation of the allele-specific and dilution

endpoint PCR assays; the results of competition studies can

be compromised if one genotype inhibits the growth of

another during the growth step in the PCR-based dilution

endpoint assay, when dilutions of root washings are trans-

ferred into 1/3� KMB111. In such cases, the density of the

antagonized strain will be underestimated. During mild

inhibition, as in mixed cultures of Q8r1-96 and MVP1-4

(genotype P) (Fig. 2), both strains remained at detectable

levels even though population sizes of MVP1-4 declined

significantly. No bias of allele-specific PCR amplification

occurred when strains were cultured in mixtures unless

strong specific inhibition occurred. Ultimately, it will be

necessary to test specific strains of all of the genotypes in

combination to identify those that are especially sensitive to

inhibition by others.

The most interesting application of allele-specific PCR is

the detection and quantification of two (this study) or more

(L. De La Fuente, B.B. Landa and D.M. Weller, unpublished

results, 2004) strains of different genotypes of 2,4-DAPG

producers competing in soil or on roots. Allele-specific PCR

has a clear advantage over the well-tested original PCR-

based dilution endpoint assay (McSpadden Gardener et al.,

2001, 2005; Landa et al., 2002a, 2003, 2005), which is limited

by its inability to detect less abundant genotypes when their

densities are three- to six-fold less than the dominant

genotype (McSpadden Gardener et al., 2001; Landa et al.,

2003). This is due to the generic nature of primers B2BF/

BPR4, which amplify all BOX-PCR genotypes (McSpadden

Gardener et al., 2001, 2005; Mazzola et al., 2004; Landa et al.,

2005). In our current study with samples from previous

work (Landa et al., 2003), we demonstrated the utility of our

allele-specific primers by quantifying the density of Q8r1-96

in competition with F113 in three different types of rhizo-

sphere and soil samples (Fig. 3; Table 4). Studies to date

(Weller et al., 2002; de Souza et al., 2003b) suggest that

disease-suppressive and crop monoculture soils contain

multiple genotypes, with one genotype often dominant.

However, how interactions between the genotypes contri-

bute to pathogen or competing genotype suppression, and

whether strains of different genotypes inhabit different

niches in the soil and rhizosphere environment, are ques-

tions that have been difficult to address with previously

available techniques.

Acknowledgements

This project was supported by the National Research In-

itiative of the United States Department of Agriculture

Cooperative State Research, Education and Extension Ser-

vice, grant number 2003-35107-13777. B.B. Landa is a

recipient of a Ramon y Cajal grant from the Ministerio de

Educacion y Ciencia of Spain. We thank Olga Mavrodi for

helpful comments.

References

Altschul SF, Madden TL, Schaffer AA, Zhang J, Zhang Z, Mille W

& Lipman DJ (1997) Gapped BLAST and PSI-BLAST: a new

generation of protein database search programs. Nucleic Acids

Res 25: 3389–3401.

Ayyadevara S, Thaden JJ & Shmookler-Reis RJ (2000)

Discrimination of primer 30-nucleotide mismatch by Taq

DNA polymerase during polymerase chain reaction. Anal

Biochem 284: 11–18.

Bangera MG & Thomashow LS (1999) Identification and

characterization of a gene cluster for synthesis of the

polyketide antibiotic 2,4-diacetylphloroglucinol from

Pseudomonas fluorescens Q2-87. J Bacteriol 181: 3155–3163.

Bagnasco P, De La Fuente L, Gualtieri G, Noya F & Arias A (1998)

Fluorescent Pseudomonas spp. as biocontrol agents against

forage legume root pathogenic fungi. Soil Biol Biochem 30:

1317–1322.

Bergsma-Vlami M, Prins ME, Staatzs M & Raaijmakers JM

(2005) Assessment of genotypic diversity of antibiotic-

producing Pseudomonas species in the rhizosphere by

denaturing gradient gel electrophoresis. Appl Environ

Microbiol 71: 993–1003.

Bonsall RF, Weller DM & Thomashow LS (1997) Quantification

of 2,4-diacetylphloroglucinol produced by fluorescent

Pseudomonas spp. in vitro and in the rhizosphere of wheat.

Appl Environ Microbiol 63: 951–955.

Campbell CL & Madden LV (1990) Introduction to Plant Disease

Epidemiology. John Wiley, New York.

Cronin D, Moenne-Loccoz Y, Fenton A, Dunne C, Dowling DN &

O’Gara F (1997a) Role of 2,4-diacetylphloroglucinol in the

interaction of the biocontrol pseudomonad strain F113 with

the potato cyst nematode Globodera rostochiensis. Appl Environ

Microbiol 63: 1357–1361.

Cronin D, Moenne-Loccoz Y, Fenton A, Dunne C, Dowling DN &

O’Gara F (1997b) Ecological interaction of a biocontrol

Pseudomonas fluorescens strain producing 2,4-

diacetylphloroglucinol with the soft rot potato pathogen

FEMS Microbiol Ecol 56 (2006) 64–78c� 2006 Federation of European Microbiological SocietiesPublished by Blackwell Publishing Ltd. No claim to original US government works.

76 L. De La Fuente et al.

Erwinia carotovora subsp. atroseptica. FEMS Microbiol Ecol 23:

95–106.

Duffy BK & Defago G (1997) Zinc improves biocontrol of

Fusarium crown and root rot of tomato by Pseudomonas

fluorescens and represses the production of pathogen

metabolites inhibitory to bacterial antibiotic biosynthesis.

Phytopathology 87: 1250–1257.

Durret R & Levin S (1997) Allelopathy in spatially distributed

populations. J Theor Biol 185: 165–171.

Felsenstein J (1989) Phylip – phylogeny inference package

(version 3.2). Cladistics 5: 164–166.

Fenton AM, Stephens PM, Crowley J, O’Callaghan M & O’Gara F

(1992) Exploitation of gene(s) involved in 2,4-

diacetylphloroglucinol biosynthesis to confer a new biocontrol

capability to a Pseudomonas strain. Appl Environ Microbiol 58:

3873–3878.

De La Fuente L, Thomashow LS, Weller DM, Bajsa N, Quagliotto

L, Chernin L & Arias A (2004) Pseudomonas fluorescens UP61

isolated from birdsfoot trefoil rhizosphere produces multiple

antibiotics and exerts a broad spectrum of biocontrol activity.

Eur J Plant Pathol 110: 671–681.

Green EK (2002) Allele-specific oligonucleotide PCR. PCR

Mutation Detection Protocols (Theophilus BDM & Rapley R,

eds), pp. 47–50. Humana Press, Totowa, NJ.

Howell CR & Stipanovic RD (1979) Control of Rhizoctonia solani

on cotton seedlings with Pseudomonas fluorescens and with an

antibiotic produced by the bacterium. Phytopathology 69:

480–482.

Iavicoli A, Boutet E, Buchala A & Metraux JP (2003) Induced

systemic resistance in Arabidopsis thaliana in response to root

inoculation with Pseudomonas fluorescens CHA0. Mol

Plant–Microbe Interact 16: 851–858.

Isnansetyo A, Cui L, Hiramatsu K & Kamei Y (2003) Antibacterial

activity of 2,4-diacetylphloroglucinol produced by

Pseudomonas sp. AMSN isolated from a marine alga against

vancomycin-resistant Staphylococcus aureus. Int J Antimicrob

Agents 22: 545–547.

Kamei Y & Isnansetyo A (2003) Lysis of methicillin-resistant

Staphylococcus aureus by 2,4-diacetylphloroglucinol produced

by Pseudomonas sp. AMSN isolated from a marine alga. Int J

Antimicrob Agents 21: 71–74.

Keel C, Schnider U, Maurhofer M, Voisard C, Burger D, Haas D &

Defago G (1992) Suppression of root diseases by Pseudomonas

fluorescens CHA0: importance of the bacterial secondary

metabolite 2,4-diacetylphloroglucinol. Mol Plant–Microbe

Interact 5: 4–13.

Keel C, Weller DM, Natsch A, Defago G, Cook RJ & Thomashow

LS (1996) Conservation of the 2,4-diacetylphloroglucinol

biosynthesis locus among fluorescent Pseudomonas strains

from diverse geographic locations. Appl Environ Microbiol 62:

552–563.

King EO, Ward MK & Raney DE (1954) Two simple media for the

demonstration of pyocianin and fluorescein. J Lab Clin Med

44: 301–307.

Koch B, Jensen LE & 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 Methods 45: 187–195.

Landa BB, Mavrodi OV, Raaijmakers JM, McSpadden Gardener

BB, Thomashow LS & Weller DM (2002a) Differential ability

of genotypes of 2,4-diacetylphloroglucinol-producing

Pseudomonas fluorescens strains to colonize the roots of pea

plants. Appl Environ Microbiol 68: 3226–3237.

Landa BB, de Werd HAE, McSpadden-Gardener BB & Weller DM

(2002b) Comparison of three methods for monitoring

populations of different genotypes of 2,4-

diacetylphloroglucinol-producing Pseudomonas fluorescens in

the rhizosphere. Phytopathology 92: 129–137.

Landa BB, Mavrodi DV, Thomashow LS & Weller DM (2003)

Interactions between strains of 2,4-diacetylphloroglucinol-

producing Pseudomonas fluorescens in the rhizosphere of

wheat. Phytopathology 93: 982–994.

Landa BB, Mavrodi OV, Schroeder KL, Allende-Molar R & Weller

DM (2005) Enrichment and genotypic diversity of phlD-

containing fluorescent Pseudomonas spp. in two soils after a

century of wheat and flax monoculture. FEMS Microbiol Ecol

in press.

Lee E-T & Kim S-D (2001) An antifungal substance, 2,4-

diacetylphloroglucinol, produced from antagonistic bacterium

Pseudomonas fluorescens 2112 against Phytophthora capsici. Kor

J Appl Microbiol Biotechnol 29: 37–42.

Lugtenberg BJJ, Dekkers L & Bloemberg GV (2001) Molecular

determinants of rhizosphere colonization by Pseudomonas.

Annu Rev Phytopathol 39: 461–490.

Mavrodi OV, McSpadden-Gardener BB, Mavrodi DV, Bonsall RF,

Weller DM & Thomashow LS (2001) Genetic diversity of phlD

from 2,4-diacetylphloroglucinol-producing fluorescent

Pseudomonas spp. Phytopathology 91: 35–43.

Mavrodi DV, Mavrodi OV, McSpadden-Gardener BB, Landa BB,

Weller DM & Thomashow LS (2002) Identification of

differences in genome content among phlD-positive

Pseudomonas fluorescens strains by using PCR-based

subtractive hybridization. Appl Environ Microbiol 68:

5170–5176.

Mazzola M, Funnell DL & Raaijmakers JM (2004) Wheat

cultivar-specific selection of 2,4-diacetylphloroglucinol-

producing fluorescent Pseudomonas species from resident soil

populations. Microbial Ecol 48: 338–348.

McSpadden Gardener BB, Schroeder KL, Kalloger SE,

Raaijmakers JM, Thomashow LS & Weller DM (2000)

Genotypic and phenotypic diversity of phlD-containing

Pseudomonas strains isolated from the rhizosphere of wheat.

Appl Environ Microbiol 66: 1939–1946.

McSpadden Gardener BB, Mavrodi DV, Thomashow LS & Weller

DM (2001) A rapid polymerase chain reaction-based assay

characterizing rhizosphere populations of 2,4-

diacetylphloroglucinol-producing bacteria. Phytopathology 91:

44–54.

FEMS Microbiol Ecol 56 (2006) 64–78 c� 2006 Federation of European Microbiological SocietiesPublished by Blackwell Publishing Ltd. No claim to original US government works.

77Diversity and detection of genotypes of Pseudomonas fluorescens

McSpadden-Gardener B, Gutierrez LJ, Raghavendra J, Edema R &

Lutton E (2005) Distribution and biocontrol potential of

phlD1 pseudomonads in corn and soybean fields.

Phytopathology 95: 715–724.

Picard C & Bosco M (2003) Genetic diversity of phld gene from

2,4-diacetylphloroglucinol-producing Pseudomonas spp.

Strains from the maize rhizosphere. FEMS Microbiol Lett 219:

167–172.

Picard C, Di Cello F, Ventura M, Fani R & Guckert A (2000)

Frequency and biodiversity of 2,4-diacetylphloroglucinol-

producing bacteria isolated from the maize rhizosphere at

different stages of plant growth. Appl Environ Microbiol 66:

948–955.

Raaijmakers JM & Weller DM (1998) Natural plant protection by

2,4-diacetylphloroglucinol-producing Pseudomonas spp. in

take-all decline soils. Mol Plant–Microbe Interact 11: 144–152.

Raaijmakers JM & Weller DM (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:

2545–2554.

Raaijmakers JM, Weller DM & Thomashow LS (1997) Frequency

of antibiotic-producing Pseudomonas spp. in natural

environments. Appl Environ Microbiol 63: 881–887.

Ramette A, Moenne-Locoz Y & Defago G (2001) Polymorphism

of the polyketide synthase gene phlD in biocontrol fluorescent

pseudomonads producing 2,4-diacetylphloroglucoinol and

comparison of phlD with plant polyketide synthases. Mol

Plant–Microbe Interact 14: 639–652.

Rezzonico F, Moenne-Locoz Y & Defago G (2003) Effect of stress

on the ability of a phlA-based quantitative competitive PCR

assay to monitor biocontrol strain Pseudomonas fluorescens

CHA0. Appl Environ Microbiol 69: 686–690.

Schnider U, Keel C, Blumer C, Troxler J, Defago G & Haas D

(1995) Amplification of the housekeeping sigma factor in

Pseudomonas fluorescens CHA0 enhances antibiotic

production and improves biocontrol abilities. J Bacteriol 177:

5387–5392.

Shanahan P, O’Sullivan DJ, Simpson P, Glennon JD & O’Gara F

(1992) Isolation of 2,4-diacetylphloroglucinol from a

fluorescent pseudomonad and investigation of physiological

parameters influencing its production. Appl Environ Microbiol

58: 353–358.

Smirnov VV & Kiprianova EA (1990) Bacteria of Pseudomonas

Genus. Naukova Dumka, Kiev, Ukraine.

de Souza JT, Arnould C, Deulvot C, Lemanceau P, Gianinazzi-

Pearson V & Raaijmakers JM (2003a) Effect of 2,4-

diacetylphloroglucinol on Pythium: cellular responses and

variation in sensitivity among propagules and species.

Phytopathology 93: 966–975.

de Souza JT, Weller DM & Raaijmakers JM (2003b) Frequency,

diversity, and activity of 2,4-diacetylphloroglucinol-producing

fluorescent Pseudomonas spp. in Dutch take-all decline soils.

Phytopathology 93: 54–63.

Stutz EW, Defago G & Kern H (1986) Naturally occurring

fluorescent pseudomonads involved in suppression of black

root rot of tobacco. Phytopathology 76: 181–185.

Thomashow LS & Weller DM (1995) Current concepts in the use

of introduced bacteria for biological disease control:

mechanisms and antifungal metabolites. Plant–Microbe

Interactions, Vol. 1 (Stacey G & Keen N, eds), pp. 187–235.

Chapman & Hall, New York, NY.

Validov S, Mavrodi OV, De La Fuente L, Boronin A, Weller DM,

Thomashow LS & Mavrodi DM (2005) Antagonistic activity

among 2,4-diacetylphloroglucinol-producing fluorescent

Pseudomonas spp. FEMS Microbiol Lett 242: 249–356.

Vincent MN, Harrison LA, Brackin JM, Kovacevich PA, Mukerji

P, Weller DM & Pierson EA (1991) Genetic analysis of the

antifungal activity of a soilborne Pseudomonas aureofaciens

strain. Appl Environ Microbiol 57: 2928–2934.

Wang C, Ramette A, Punjasamarnwong P, Zala M, Natsch A,

Moenne-Loccoz Y & Defago G (2001) Cosmopolitan

distribution of phlD-containing dicotyledonous crop-

associated biocontrol pseudomonads of world wide origin.

FEMS Microbiol Ecol 37: 105–116.

Weller DM & Cook RJ (1983) Suppression of take-all of wheat by

seed treatment with fluorescent pseudomonads.

Phytopathology 73: 463–469.

Weller DM, Raaijmakers JM, McSpadden-Gardener BB &

Thomashow LS (2002) Microbial populations responsible for

specific soil suppressiveness to plant pathogens. Annu Rev

Phytopathol 40: 309–348.

FEMS Microbiol Ecol 56 (2006) 64–78c� 2006 Federation of European Microbiological SocietiesPublished by Blackwell Publishing Ltd. No claim to original US government works.

78 L. De La Fuente et al.