Journal of Applied Microbiology ISSN 1364-5072

ORIGINAL ARTICLE

Growth-inhibiting effects of concentrations of fusaric acid

on the growth of Bacillus mojavensis and other biocontrol
Bacillus species
C.W. Bacon1, D.M. Hinton1 and A. Hinton Jr2
1 Toxicology & Mycotoxin Research Unit, ARS, USDA, Russell Research Center, Athens, GA 30604, USA
2 Poultry Processing & Meat Quality Research Unit, ARS, USDA, Russell Research Center, Athens, GA 30604, USA

Keywords
Bacillus amyloliquefaciens, Bacillus
atrophaeus, Bacillus licheniformis, Bacillus
mojavensis, Bacillus subtilis, Bacillus
vallismortis, biological control, bacterial
endophyte, fumonisin, fusaric acid, Fusarium
verticillioides, mycotoxin, Paenibacillus
lentimorbus, Paenibacillus popilliae, Zea mays.
Correspondence
C.W. Bacon, USDA, ARS, Russell Research
Center, P.O. Box 5677, 950 College Station
Road, Athens, GA 30604, USA.
E-mail: cbacon@saa.ars.usda.gov
2005/0853: received 22 July 2004, revised 30
April 2005 and accepted 1 May 2005
doi:10.1111/j.1365-2672.2005.02770.x
Abstract
Aims: To determine the effects of concentrations of fusaric acid on the growth
of several strains of the biocontrol bacterial endophyte Bacillus mojavensis and
other species within the Bacillus subtilis group, as well as the genetic relation-
ships within this small group of Gram-positive bacteria, and their antagonisms
to Fusarium verticillioides, which produce fusaric acid.
Methods and Results: The growth of 50 Bacillus strains and species were tested
at two concentrations of fusaric acid determined in maize infected by an isolate
of F. verticillioides. Molecular characterizations of the strains and species of
bacteria were determined with an automated ribotyper. The growth of bacteria
measured under both concentrations with an automated turbidometer, Bio-
screen, indicated that fusaric acid was toxic to most strains of the bacterial
endophyte B. mojavensis. However, the effects of these two concentrations on
other Bacillus species varied in that fusaric acid was either bacteriocidal or bac-
teriostatic to most species.
Conclusions: These data indicate that the concentrations of fusaric acid are
inhibitory to the growth of most Bacillus species, some of which are used as
biocontrol agents. This suggests that the endophytic and saprophytic states of
F. verticillioides and other Fusarium species cannot be controlled by fusaric-
acid-sensitive Bacillus species.
Significance and Impact of Study: Mycotoxic Fusarium species, such as F. vert-
icillioides, are competitive because all produce fusaric acid, which is inhibitory
to biocontrol bacteria, and mutants tolerant to fusaric acid must be developed
in order to be effective on biocontrol bacteria.

the bacterium Bacillus mojavensis was shown to be endo-

Introduction
Endophytic and epiphytic species of Bacillus have been
shown to have potential for reducing pathogenicity and
in planta growth and mycotoxin accumulation by Fusa-
rium species (Hallmann et al. 1997; Chanway 1998; Sturz
et al. 2000, Bacon et al. 2001). An important endophytic
species is Fusarium verticillioides, which produces the
fumonisin mycotoxins in maize, and is associated with
equine leucoencephalomalacia, pulmonary oedema syn-
drome in swine, liver cancer in rats and human oesopha-
geal cancer (Riley et al. 1993; Marasas 2001). Recently,
phytic and capable of serving as a biocontrol for the end-
ophytic state of F. verticillioides and for reducing
fumonisin accumulation in maize (Bacon and Hinton
2001, 2002). However, this species and 11 other Fusarium
species known to date have been shown to produce fusa-
ric acid (5-butylpicolinic acid) (Backmann 1956; Luz
et al. 1990; Bacon et al. 1996; Capasso et al. 1996), a
toxin implicated in the wilt of tomato (Yabuta et al.
1937; Gaumann 1957) and diseases of other plants (Yab-
uta et al. 1937; Tamari and Kaji 1954; Gaumann 1957;
Davis 1969; Drysdale 1984). Fusaric acid, a common

ª 2005 The Society for Applied Microbiology, Journal of Applied Microbiology 100 (2006) 185–194
No claim to original US government works 185
Fusaric acid inhibits Bacillus species
contaminant of feedstuffs (Smith and Sousadias 1993;
Porter et al. 1995), is also mildly toxic to mice (Yabuta
et al. 1937; Porter et al. 1995), and has several important
pharmacological properties (Porter et al. 1995).
In addition to its toxicity to animals, fusaric acid is
toxic to several micro-organisms, as well as plant physio-
logical processes (Maloy and Nunn 1981; Luz et al. 1990;
Marrèet al. 1993). It has been shown to inhibit growth
and metabolism of several strains of soil-borne species of
Pseudomonas fluorescens, especially biocontrol isolates. In
this species, fusaric acid was shown specifically to sup-
press the production of the antibiotic 2,4-diacetylphlorog-
lucinol, which serves to control the growth of most fungi
and is responsible for the biocontrol activity of soil-borne
Paenibacillus fluorescens strains (Notz et al. 2002). Addi-
tional studies have demonstrated the widespread but
varying sensitivity of Paenibacillus fluorescens, Bacillus spe-
cies and Paenibacillus macerans to fusaric acid (Schnider-
Keel et al. 2000; Landa et al. 2002; Notz et al. 2002).
Bacillus mojavensis is closely related to Bacillus subtilis,
from which it and several other species were recently
erected (Nakamura 1989; Roberts et al. 1994, 1996;
Nakamura et al. 1999; Takami et al. 2000). The finding
that B. mojavensis and other closely related species within
the B. subtilis group can only be distinguished by differ-
ences in their DNA sequences, and that there are very lit-
tle phenotypic differences among the group (Roberts
et al. 1994, 1996; Nakamura et al. 1999) suggest the uni-
versal sensitivities of Bacillus species to fusaric acid, which
prompted this study. The universal production of fusaric
acid by F. verticillioides (Bacon et al. 1996) and other Fus-
arium species suggests that biocontrol of these fungi by B.
mojavensis and other Bacillus species might be reduced
considerably, provided toxic concentrations of fusaric acid
are encountered. However, in vitro tests for antagonism
to F. verticillioides by Bacillus species and the sensitivities
of the Bacillus species to fusaric acid are unknown. Thus,
the first objective was to examine the effects of fusaric
acid on the growth of the available strains of the endo-
phytic species B. mojavensis, to distinguish and determine
if there are any relationships at the molecular level to the
desert origins of the strains. The second objective of this
study was to determine if the closely related species of
Bacillus, some of which are intended for use in various
biocontrol strategies, are affected by fusaric acid.
Materials and methods
Bacterial strains and culture conditions
The patented biocontrol B. mojavensis strain RRC 101
(ATCC 55732) and the other bacterial species used in this
study and their sources are presented in Table 1. The bac-
ª 2005 The Society for Applied Microbiology, Journal of Applied Microbiology 100 (2006) 185–194
186 No claim to original US government works
C.W. Bacon et al.
teria were stored as silica gel cultures at )20
C, and cul-
tures were maintained on nutrient agar (Difco
Laboratories). The medium used to study the effects of
fusaric acid on the growth of the bacteria was J-broth
(Gordon et al. 1973), distilled water: tryptone 5.0 g l)1,
yeast extract 15.0 g l)1, K2HPO 3.0 g l)1, glucose
2.0 g l)1, pH 7.5. Inoculum of each strain was prepared
from overnight cultures grown in J-broth.
Growth rate experiments were performed with an auto-
mated turbidometer, the Microbiology Bioscreen C
Reader (Labsystems, Helsinki, Finland), in 100-well sterile
microplates. Each well contained 190 ll of J-broth, to
which was added 10 ll of bacterial inoculum (103 CFU
per ml) in J-broth. The cultures were incubated at 30
C
with constant shaking and the OD600 measured at 30-min
intervals over the incubation period of 48 h. The concen-
trations of fusaric acid (Sigma Chemical) used to measure
the in vitro toxicity, 100 and 200 lg ml)1, were based on
data of the patented isolate of B. mojavensis (Bacon et al.
2004), which showed growth was inhibited by 30–90% of
the controls. The concentrations of fusaric acid for other
toxicity tests were measured at the average in planta con-
centrations of fusaric acid determined in this work.
Growth curves and several other plot parameters were
generated and analysed by the software package of the
Bioscreen C Reader (research express, version 1.00).
The specific rates of growth were expressed as a percent-
age of reduction over controls that contained no fusaric
acid. Sensitivity of fusaric acid was rated at the
100-lg ml)1 concentration as tolerant, moderately toler-
ant or intolerant depending on the percentage of fusaric
acid inhibition of 0–40, 41–80 and 81–100, respectively.
Strains that were not sensitive at this concentration were
rated insensitive. All experiments were performed in trip-
licate, and were repeated independently.
Fungal culture and antagonism assay
The strains of F. verticillioides, RRC 408 and MRC 826,
used to infect the maize cultivar P3167, are two of several
strains of this fungus that infect maize and remain as a
symptomless endophyte on this cultivar and produce the
fumonisin mycotoxins and fusaric acid (3Æ00 ppm) on
autoclaved corn and in planta (Bacon and Hinton 1996b;
Bacon et al. 2001). Another strain used was a UV mutant,
RRC 28-5, developed from RRC 408, which produces
only trace amounts of fusaric acid (0Æ09 ppm) on auto-
claved corn and is referred to as fusaric acid-less mutant
in this study. The fungi were grown on potato dextrose
agar (PDA) medium for 7–14 days at room temperature.
The assay for growth antagonism of strains and species
of bacteria to cultures of F. verticillioides MRC 826 was
determined with an agar diffusion assay on Petri plates of
C.W. Bacon et al. Fusaric acid inhibits Bacillus species

Table 1 The origin of Bacillus species, strains, and relatives used in this study

Species Strain number* Source

Bacillus mojavensis ATCC 51516 Soil, Mojave Desert, CA, USA

ATCC 51517 Gobi Desert, People’s Republic of China, NRRL B-14708
NRRL B-14698T F.M. Cohan, soil, Mojave Desert, CA, USA
NRRL B-14699 F.M. Cohan, soil, Mojave Desert, CA, USA
NRRL B-14700 F.M. Cohan, soil, Mojave, CA, USA
NRRL B-14701 F.M. Cohan, soil, Mojave, CA, USA
NRRL B-14702 F.M. Cohan, soil, Mojave, CA, USA
NRRL B-14703 M.S. Roberts, soil, Gobi Desert, People’s Republic of China
NRRL B-14704 F.M. Cohan, soil, Gobi Desert, People’s Republic of China
NRRL B-14705 F.M. Cohan, Gobi Desert soil, People’s Republic of China
NRRL B-14706 M.S. Roberts, Gobi Desert soil, People’s Republic of China
NRRL B-14707 F.M. Cohan, Gobi Desert soil, People’s Republic of China
NRRL B-14708 F.M. Cohan, Gobi Desert soil, People’s Republic of China
NRRL B-14709 M.S. Roberts, Gobi Desert soil, People’s Republic of China
NRRL B-14710 F.M. Cohan, Gobi Desert soil, People’s Republic of China
NRRL B-14711 F.M. Cohan, Gobi Desert soil, People’s Republic of China
NRRL B-14712 F.M. Cohan, Gobi Desert soil, People’s Republic of China
NRRL B-14713 M.S. Roberts, Gobi Desert soil, People’s Republic of China
NRRL B-14714 M.S. Roberts, Sahara Desert soil, Nefta, Tunisia
NRRL B-14715 M.S. Roberts, Sahara Desert soil, Nefta, Tunisia
NRRL B-14716 M.S. Roberts, Sahara Desert soil, Nefta, Tunisia
NRRL B-14718 F.M. Cohan, Sahara Desert soil, Nefta, Tunisia
NRRL B-14719 M.S. Roberts, Sahara Desert soil, Nefta, Tunisia
NRRL B-14817 K.E. Duncan, soil, Tumamoc Hill, AZ, USA
NRRL B-14818 K.E. Duncan, soil, Tumamoc Hill, AZ, USA
NRRL B-14824 M.S. Roberts, soil, Sahara Desert, Nefta, Tunisia
RRC 101 (ATCC 55732) Maize kernels, northern Italy
RRC 112 Rifampicin mutant of RRC 101
Bacillus subtilis ATCC 6051T Type for the species, Marbury strain, H.J. Conn
BD170 D. Pubnau (transformation host)
RRC 111 BD170 transformed with B. mojavensis RRC 101DNA
RRC 113 UV mutant 12 of RRC 111
RRC 114 UV mutant 50 of RRC 111
ATCC 6633 N.R. Smith
ATCC 33608 D. Dubnau (BD170)
ATCC 49343 K.F. Anderson (S.N. McDermott)
ATCC 55422 Sandy soil, Japan
ATCC 55614 Soil, fungal antagonist (patented)
B. subtilis ATCC 55675 Transport enhancer (patented)
Bacillus amyloliquefaciens NRRL B-14393T Soil, J. Fukomoto as strain F
Bacillus atrophaeus NRRL NRS-213T Colorado soil, N.R. Smith
Bacillus licheniformis NRRL NRS-1264T Soil, R.E. Gordon
Paenibacillus popilliae NRRL B-2309T Anti-Japanese beetle spore dust
Paenibacillus lentimorbus NRRL B-2522T Diseased Japanese beetle grub
Bacillus vallismortis NRRL B-14890 Sand dune with mesquite tree, Death Valley
NRRL B-14891 Alluvial fan dune with holly bush, Death Valley
NRRL B-14892 Sand dune with mesquite tree
NRRL B-14893 Alluvial fan with creosote bush, Death Valley
NRRL B-14894 Alluvial fan arroyo with holly bush, Death Valley

*NRRL, Northern Regional Research Laboratory culture collection, Peoria, IL, USA; ATCC, American Type Culture Collection, Rockville, MD, USA;
RRC, Russell Research culture collection, Athens, GA, USA.

All B. mojavensis, B. vallismortis and B. amyloliquefaciens strains are indicated here according to Roberts et al. (1994, 1996) and Nakamura (1989).

nutrient agar (Difco Inc.) as the base agar. This assay was here as the diffusible inhibitor. Bacterial inocula and F.
based on the production of an inhibitor as the major verticillioides plug were each placed on the opposite edge
characteristic of the patented isolate, which is referred to of a Petri dish. The fungal inocula were added to the

ª 2005 The Society for Applied Microbiology, Journal of Applied Microbiology 100 (2006) 185–194
No claim to original US government works 187
Fusaric acid inhibits Bacillus species
plates 5 days after the bacterial inoculum, and plates were
incubated at 25–27
C. The widths of cleared zones of ant-
agonism (distances between the bacterial and fungal
growth) were measured after 21 days and classified as
previously described (Bacon and Hinton 2002): ),
<0Æ10 mm; +, <3 mm; ++, 3–9 mm; +++, >9–18 mm;
++++, > 18–20 mm; +++++, >20 mm.
Ribotyping and fatty acid analysis
Molecular characterization and identification of B. moja-
vensis and Bacillus species were performed with an auto-
mated ribotyping instrument (RiboPrinterTM Microbial
Characterization System, Qualicon/Dupont, Wilmington,
DE, USA), which used DNA fragments cut with the
restriction enzyme EcoRI prior to hybridization to a
cloned chemiluminescently labelled rDNA probe. Bacterial
strains were cultivated overnight in J-broth, harvested by
centrifugation (10 000 g) for 10 min and re-suspended in
phosphate buffered saline, proceeding according to the
microbial characterization system ribotyping protocol.
Ribotype banding profiles were developed and numerical
analyses and cluster analyses were carried out using
appropriate software available to the program.
Confirmation of the diversity of bacteria by the ribo-
typing procedure was performed with the fatty acid
methyl esters (FAME) procedure. The fatty acid profiles
of B. mojavensis isolates, along with B. subtilis strains,
were performed using the MIDI Sherlock Microbial
Identification System (MIS) (MIDI Inc., Newark, DE,
USA). Strains were transferred from nutrient agar to
Difco tryptic soy broth (TSBA) (Becton, Dickson, and
Co., Sparks, MD, USA) supplemented with 1.5% of
Difco Agar (Becton, Dickson, and Co.) and incubated
at 28
C for 24 h. Following three consecutive transfer
and incubation periods on TSBA, each culture was har-
vested, and fatty acids were extracted from the cellular
membranes using a four-step extraction procedure. The
first step consisted of saponification by heating the cells
at 100
C for 30 min in a methanolic sodium hydroxide
solution composed of 15% (w/v) NaOH, dissolved in
50% (v/v) methanol. The second step consisted of
methylating the saponified cellular fatty acids in a solu-
tion of 3Æ25 N hydrochloric acid in methanol (46%
v/v) at 80
C for 10 min. During the third step, FAME
analyses were then extracted from the aqueous phase
following a 10-min liquid–liquid extraction with a 1 : 1
(v/v) solution of hexane and methyl-tert-butyl ether.
The resulting FAME extracts were washed for 5 min in
a sodium hydroxide solution (1Æ08%, w/v). During the
fourth step, the FAME extracts were transferred to
sample vials for identification and quantification by an
Agilent 6890 gas chromatograph (GC) and computer
ª 2005 The Society for Applied Microbiology, Journal of Applied Microbiology 100 (2006) 185–194
188 No claim to original US government works
C.W. Bacon et al.
ChemStation (Agilent Technologies, Palo Alto, CA,
USA) that was operated by mis, version 4.5 software.
The mis software identified and quantified FAMEs of
bacterial isolates and compared the FAME profiles with
the TSBA 40 MIS computer library to which the type
strain B. mojavensis, NRRL B-14698 (ATCC 51516), the
biocontrol strain RRC 101 and other known isolates of
the B. subtilis group were added. Fatty acid analyses were
determined on the 24 strains of B. mojavensis, the closely
related strains including two strains of B. subtilis, as well
as on the type strains of Bacillus amyloliquefaciens, Bacil-
lus licheniformis, and Bacillus atrophaeus. Isolates were
identified as B. mojavensis when their FAME profiles
matched the profile of the type and the degree of the
match between the unknown isolates, and the data for
the isolate entered in the MIS library were indicated by
the Similarity Index value assigned to unknowns. Similar-
ity Index values may range from 0Æ00 (no match) to 1Æ00
(identical match).
Results
Band pattern analysis using the restriction enzyme EcoRI
indicated that all B. mojavensis isolates grouped into two
large clusters with one major conserved region similar to
B. subtilis, but that these clusters were distinct and differ-
ent from the other closely related reference species
(Fig. 1). Ribotyping after EcoRI restriction showed that all
strains shared a faint band that was characteristic of the
genus, with the exception of B. lichensformis. All strains
of B. mojavensis shared a band, although faint, which was
absent in the other species. However, this band was
shared by B. subtilis and B. atrophaeus but was greatly
reduced in the mutant strains of B. mojavensis RRC 111
UV12. The ribotyping procedure reduced the 14 strains
of B. mojavensis into six ribotype clusters and separated
this species from the other closely related species within
the B. subtilis group (Fig. 1). The data indicated that this
procedure, although not intended, is apparently useful for
strain typing of this species within the B. subtilis group.
The ribotype profiles did not distinguish the strains of B.
mojavensis into any pattern characteristic of and unique
to the strains isolated from the four major desert regions
of the world (Fig. 1).
The patented biocontrol endophytic strain, RRC101, its
mutants 111 and UV12, all formed a unique subcluster
consisting of 13 bands that were distinctly different from
the other isolates, except for strain NRRL B-14818 which
also clustered within this subgroup indicating a close gen-
etic relationship to RRC101. However, strain RRC101 ori-
ginated from corn kernels obtained from northern Italy,
whereas NRRL B-14818 originated from a desert soil sam-
ple from Arizona, USA. Strains in this subcluster showed
C.W. Bacon et al. Fusaric acid inhibits Bacillus species

Ribotypes Strains Species Origin

% similarity
40 50 60 70 80 100

NRS1264 Bacillus licheniformis

NRRLB-14824 Bacillus mojavensis S

NRRLB-14700 Bacillus mojavensis M

ATCC51517 Bacillus mojavensis G

NRRLB-14698 Bacillus mojavensis M

NRRLB-14707 Bacillus mojavensis G

NRRLB-14711 Bacillus mojavensis G

NRRLB-14709 Bacillus mojavensis G

ATCC51516 Bacillus mojavensis M

RRC111,UV12 Bacillus mojavensis I

RRC111 Bacillus mojavensis I

NRRLB-14818 Bacillus mojavensis AZ

RRC101 Bacillus mojavensis I

NRRLB-14714 Bacillus mojavensis S

NRRLB-14817 Bacillus mojavensis S

ATCC6051 Bacillus subtilis

BD170 Bacillus subtilis

ATCC6051-2 Bacillus subtilis

NRRLB-14393 Bacillus amyloliquefaciens

NRS-213 Bacillus atrophaeus

Figure 1 Phylogenetic dendrogram and ribotype patterns obtained with EcoRI restriction enzyme of 16S rDNA sequences of 14 randomly selec-
ted Bacillus mojavensis strains and other type species of the Bacillus subtilis group. Dendrogram was calculated with Dice coefficient using UPGMA
clustering methods. Origin of strains: S, Sahara Desert; M, Mojave Desert; G, Gobi Desert; I, southern Italy; Az, Arizona.

the most fragmented polymorphisms and subspecies vari-
ation. While we were successful in distinguishing several
selected strains from B. subtilis using the fatty acids analy-
sis, this proved difficult when we analysed the other spe-
cies within the B. subtilis group (data not shown).
However, the utility of analysing for specific fatty acids
for species of this group was indicated as the desired pro-
cedure (Roberts et al. 1994), which was not utilized in
our approach but which our data indicate should not be
the definitive analysis of this diverse but relatively small
group of species.
The diversity in strain variation is reflected in the phe-
notypes of strains of B. mojavensis, which varied in their
ability, both quantitatively and qualitatively, to produce
an agar-diffusible inhibitory substance, as well as toler-
ances to the lower concentration of fusaric acid (Table 2).
Specific rates of growth of all strains were measured
under the two concentrations. Bacillus mojavensis NRRL
B-14708 was the only strain tolerant to fusaric acid at the
lower concentration. All strains were sensitive to the
higher concentration of fusaric acid; therefore, strains
were rated according to their sensitivities at the lower
concentration. The largest percentage of fusaric-acid-tol-
erant strains was from the Gobi Desert strains, and strain
NRRL B-14708 showed natural resistance at the low con-
centration, i.e. not sensitive. Only one strain from the
Mojave Desert was rated moderately tolerant, NRRL B-
14699. The specific rates of growth data indicated that
the remaining species of the B. subtilis group tested were
also sensitive to the concentrations of fusaric acid
(Tables 3 and 4). Bacillus subtilis ATCC 55614 was
the most tolerant strain of this species (Table 3), while

ª 2005 The Society for Applied Microbiology, Journal of Applied Microbiology 100 (2006) 185–194
No claim to original US government works 189
Fusaric acid inhibits Bacillus species C.W. Bacon et al.

Table 2 Inhibitory activity of fusaric acid on the specific rate of Table 3 Inhibitory activity of fusaric acid on the specific rate of
growth of Bacillus mojavensis strains growth of Bacillus subtilis strains and their antagonism to Fusarium
verticillioides
% Fusaric acid
inhibition* % Fusaric acid inhibition*

100 200 Strains 100 lg ml)1 200 lg ml)1 Antagonism

Strains Origin lg ml)1 lg ml)1 Antagonism

ATCC 6051 74a 95a )
ATCC 51516 Mojave 68a 92a ) BD 170 84a 93a )
ATCC 51517 Gobi 61a 88a ++ RRC 111 37b 94a +++
NRRL B-14698T Mojave 92b 93a + RRC 113 57a 77a )
NRRL B-14699 Mojave 21c 91a ) RRC 114 73a 78a ++
NRRL B-14700 Mojave 76a 83a ) ATCC 6633 84a 66b )
NRRL B-14701 Mojave 67a 83a +++ BD170 60a 96a )
NRRL B-14702 Mojave 47b 78a ) ATCC 49343 29c 66b ++
NRRL B-14703 Gobi 91a 94a ) ATCC 55422 53a 69b +++++
NRRL B-14704 Gobi 27c 97a ) ATCC 55614 25c 73a +++++
NRRL B-14705 Gobi 49b 79a ) ATCC 55675 39b 84a )
NRRL B-14706 Gobi 26c 94a )
NRRL B-14707 Gobi 62b 87a ++ *Inhibitory activity expressed as a percentage of the control values.
NRRL B-14708 Gobi 0d 93a ) Values represent the means of six replicate cultures. Means within a
NRRL B-14709 Gobi 68a 92a + column followed by a different letter were significantly different at
NRRL B-14710 Gobi 87b 92a + P £ 0.05 according to Fisher’s protected least significant difference
NRRL B-14711 Gobi 70a 92a ) test.
NRRL B-14712 Gobi 19c 90a )
Zones of antagonism were measured after 21 days following inocula-
NRRL B-14713 Gobi 61a 94a ) tion of plates with F. verticillioides MRC 826; ), <0.10 mm (not
NRRL B-14714 Sahara 64a 83a + antagonistic); +, <3 mm; ++, 3–9 mm; +++, >9–18 mm; ++++, >18–
NRRL B-14715 Sahara 82b 72b ) 20 mm; +++++, >20 mm.
NRRL B-14716 Sahara 77a 89a ++
NRRL B-14718 Sahara 33c 81a ++
NRRL B-14719 Sahara 63a 86a ++
Table 4 Inhibition by fusaric acid to the specific rates of growth of
NRRL B-14817 Sahara 62a 91a +
Bacillus species and their antagonism to Fusarium verticillioides
NRRL B-14818 Sahara 55a 86a +++
NRRL B-14824 Sahara 69a 88a )
% Fusaric acid inhibition*
RRC 101 Italy 59a 83a +++
(ATCC 55732) Species 100 lg ml)1 200 lg ml)1 Antagonism

RRC 112 Mutant of 81b 94b ++
RRC101 Bacillus amyloliquefaciensT 28a 60a C++
Bacillus atrophaeusT 33a 71a ++++
*Percentage of fusaric acid inhibition was determined within a 48-h Bacillus licheniformisT 74b 82a —
incubation period at 30
C. Values represent the means of six replicate Paenibacillus 61b 69a +
cultures. Means within a column followed by a different letter were popilliae B-2309T
significantly different at P £ 0.05 according to Fisher’s protected least Paenibacillus 81b 75a ++
significant difference test. lentimorbus B-2522T

Antagonism determined after 21 days following co-inoculations of Bacillus vallismortis 67b 75a ++
fungi with bacteria on nutrient agar plates; ), <0.10 mm (not antago- B-14890
nistic); +, <3 mm; ++, 3–9 mm; +++, >9–18 mm; ++++, >18– B. vallismortis B-14891 64b 69a ++
20 mm; +++++, >20 mm. B. vallismortis B-14892 0c 13b )
B. vallismortis B-14893 80b 72a ++
B. vallismortis B-14894 57b 70a )
Bacillus vallismortis NRRL B-14892 was the most tolerant
of the remaining species (Table 4). Indeed, this strain was *Inhibitory activity expressed as a percentage of the control values.
tolerant of the higher concentration of fusaric acid, indi- Values represent the means of six replicate cultures. Means within a
cating resistance. However, this strain was not antagonis- column followed by a different letter were significantly different at
tic to F. verticillioides. P £ 0.05 according to Fisher’s protected least significant difference
test.
The antagonistic responses varied according to the

Zones of antagonism were measured after 21 days following inocula-
strain of B. mojavensis as indicated by the zones of inhibi- tion of plates with F. verticillioides MRC 826; ), <0.10 mm (not
tion. Only 44% of the strains produced this diffusible antagonistic); +, <3 mm; ++, 3–9 mm; +++, >9–18 mm; ++++, >18–
substance on the nutrient agar culture medium. However, 20 mm; +++++, >20 mm; C++, contact kill with necrotic hyphae and
all strains were antagonistic because those strains that did an antagonism zone.

ª 2005 The Society for Applied Microbiology, Journal of Applied Microbiology 100 (2006) 185–194
190 No claim to original US government works
C.W. Bacon et al.
not produce a zone of antagonism caused necrotic hyphae
upon contact with the bacterial colony (data not shown).
This suggests either that there are two active inhibitory
compounds or that the same inhibitory substance is not
membrane permeable in all strains. Several other Bacillus
species were antagonistic to the fungus (Tables 3 and 4).
Indeed, the largest antagonistic responses were those pro-
duced by two strains of B. subtilis, ATCC 55614 and
55422, followed by B. atrophaeus NRS-213. Strain ATCC
55614 was also very antagonistic to F. verticillioides. How-
ever, one species did not produce an agar-diffusible sub-
stance, but was inhibitory as evidenced by the production
of necrotic hyphae upon contact with the fungal colony
(data not shown).
Discussion
We have established that all the strains and species of
Bacillus tested in this study were sensitive to fusaric acid.
Several Bacillus species used in this study have been recom-
mended for use as biocontrol agents; however, our data
suggest that if fusaric acid is in their environment, their
effectiveness may be severely limited. Thus, the numbers of
bacteria that are sensitive to fusaric acid, and probably ant-
agonized by species of Fusarium, are larger than previously
considered. The Fusarium species are a highly successful
group of fungi that produce a variety of secondary meta-
bolites, some of which figure into an organism’s competi-
tion for colonization of plants, especially in homologous
niches such as soil, phyllosphere and endophytic spaces. In
addition to its arsenal of defensive secondary metabolites
(Thrane 2001), the Fusarium species produce fusaric acid,
which historically is known as a wilt toxin.
However, the toxicity of fusaric acid is wide-acting; it
is biologically active in micro-organisms, animals and
plants. This is due to its activity as a lipophilic chelating
agent (Malini 1966; Maloy and Nunn 1981). Specifically,
in plants this toxin has been shown to act as a weak un-
dissociated acid while in the apoplasm; however, when it
passes through the cell membrane and accumulates in the
symplasm, it is dissociated and the extent of dissociation
is Ph dependent (Marrèet al. 1993). In the symplasm it is
known to inhibit mitochondrial respiration (Paquin and
Waygood 1957; Arias 1985), decrease ATP concentration
and destroy membrane integrity (D’Alton and Etherton
1984; Arias 1985). In general, the activity of this toxin in
plants proceeds first with a nonspecific phase that
involves cell membrane penetrations, followed by a sec-
ond phase that involves specific cell targets such as mito-
chondria (Marrèet al. 1993). Thus, when testing for
activity at the cellular level, it is necessary to test for tox-
icity with values that are anticipated to occur in the sym-
plasm (Davis 1969; Bacon et al. 2004).
ª 2005 The Society for Applied Microbiology, Journal of Applied Microbiology 100 (2006) 185–194
No claim to original US government works
Fusaric acid inhibits Bacillus species
The infection of hosts and biosynthesis of fusaric acid
by isolates of Fusarium species in the absence of disease
symptoms (Bacon and Hinton 1996a) suggest that toxic
concentrations of fusaric acid are not produced in the
Fusarium-infected plants, fusaric acid is not produced, or
modern-day plant cultivars are resistant to the in planta
concentrations that are produced. Because all isolates of
F. verticillioides examined to date produce fusaric acid
(Bacon et al. 1996), and because most of these isolates
maintain a symptomless endophytic biotrophic or hemi-
biotrophic association with maize, the in planta interac-
tion of fusaric acid might be multifunctional and
therefore complex. Further, the biosynthesis of fusaric
acid by soil-borne nonpathogenic isolates of other Fusa-
rium species suggests a role in the interaction with other
soil-borne organisms. The values of fusaric acid used in
this study are similar and relevant to those expected to
occur in the symplasm. Indeed, the amounts of fusaric
acid isolated from natural products are well within the
range used in this study (Smith and Sousadias 1993; Por-
ter et al. 1995). Because fusaric acid readily chelates sev-
eral ions, which decreases it toxicity (Backmann 1956;
Malini 1966; Bochner et al. 1980; Duffy and Défago
1997), the interaction of specific metal ions with the final
expression of toxicity should also be considered.
Recently, studies have shown that fusaric-acid-produ-
cing strains of Fusarium oxysporum altered the biocontrol
activity of a strain Paenobacillus fluorescens by repressing
the expression of genes involved in the production of the
inhibitor 2,4-diacetylphloroglucinol, rendering this bacter-
ium ineffective in controlling the growth of the fungus
(Schnider-Keel et al. 2000; Notz et al. 2002). The biocon-
trol B. mojavensis RRC101 also produces an inhibitory
substance in vitro, although it has been neither chemically
identified nor established as being produced in planta.
The use of B. mojavensis as an endophytic biocontrol has
decreased the accumulation of the fumonisin mycotoxins
and endophytic infection of maize by F. verticillioides
under greenhouse conditions (Bacon et al. 2001). Biocon-
trol of this fungus by B. mojavensis is high when tested
against nonproducing fusaric acid mutants of F. verticil-
lioides (Bacon et al. 2004).
The eight Bacillus species used in this study are very clo-
sely related but are easily distinguished by differences in
their DNA sequences, and cannot be distinguished by
phenotypic characterization in general (Roberts et al. 1994,
1996; Nakamura et al. 1999). Significant differences in the
levels of ten fatty acids were reported by Roberts et al.
(1994) 1996) for distinguishing the species used in this
study. We found better separation using the ribotyper,
which is in agreement with Roberts et al. (1994) 1996) and
Nakamura et al. (1999). All strains of B. mojavensis were
rated as endophytic (Bacon and Hinton 2002), although
191
Fusaric acid inhibits Bacillus species
not all were tested for in planta protection against infec-
tion by F. verticillioides. Because all were endophytic the
relationship between ribotype patterns, geographic origin
and the endophytic trait cannot be explored. The 14 ran-
domly selected strains used for ribotyping were isolated
from different sampling locations of the great deserts, but
many showed subspecies variation in the ribotypes, as well
as subspecies variation fragment polymorphisms, inde-
pendent of geographic origin. An analysis of the remaining
13 strains would not alter the geographic pattern observed.
Thus, there was no correlation between specific desert ori-
gin, ribotype pattern and fusaric acid sensitivity. Further,
the ribotype groups in the area of subspecies variation
indicated a considerable overlap of genetically diverse
strains and fragment polymorphisms within the world’s
distribution, suggesting a high degree of genetic plasticity,
which indicates that these isolates are not clones.
In addition to the Bacillus species studied in this work,
Landa et al. (2002) determined that B. circulans and
B. megaterium were susceptible to graded concentrations
of fusaric acid, of which 100 and 200 lg/ml showed the
most consistent inhibition as did the present study on B.
mojavensis. Several other Pseudomonas and Paenibacillus
species have now been identified as being sensitive to fus-
aric acid, and it has been shown that two species of Bacil-
lus were the most sensitive of the three genera used in
their study (Landa et al. 2002). Our research indicates
that there is marked variation within the genus Bacillus
but in general there is sensitivity with the additional bio-
control species used in this study. This and other studies
indicate the widespread occurrence of fusaric acid sensi-
tivity not only within the Bacillus group but also among
bacteria in general.
The cluster of ribotypes produced with the patented
isolate, RRC 101, was distinctly different from the other
clades of B. mojavensis and formed its own cluster. This
was especially true of the ribotype bands located within
the subspecies variation of the pattern. It is interesting
that the ribotypes of the isolate NRRL B-14818 were iden-
tical to that obtained by the biocontrol isolates and
mutants derived from it. While the two strains have
distinctly different geographical origins (Table 1), it is
possible that the origins of strains RRC 101 and NRRL
B-14818 were actually similar, suggesting that these
strains may well be clones. The origin of RRC 101 was
maize kernels isolated from seed samples from northern
Italy. However, these data possibly suggest that the origin
of seed stock was from the western USA. Both strains
were equally inhibitory to F. verticillioides.
We paid a considerable amount of attention to the
effects of fusaric acid on B. mojavensis because of its poten-
tial for use as a biocontrol endophyte, a group of organisms
with general and specific disease control applications that
ª 2005 The Society for Applied Microbiology, Journal of Applied Microbiology 100 (2006) 185–194
192 No claim to original US government works
C.W. Bacon et al.
are current and widespread in agriculture (Hallmann et al.
1997; Chanway 1998; Kobasyashi and Palumbo 2000; Sturz
et al. 2000). Other uses are also anticipated, i.e. surrogate
transformation and paratransgenesis, two methods in
which symbiotic micro-organisms are transformed and
inserted into host plants, which leads to an indirect modifi-
cation of the plant’s genome and subsequent phenotype
resulting in novel methods for designing symbionts, such
as endophytic bacteria, for control of specific diseases or
enhanced processes. However, the ubiquitous production
of fusaric acid by endophytic Fusarium species, pathogenic
and nonpathogenic, raises the questions as a possible
impediment to the successful use of endophytic biocontrol
bacteria. A similar concern for epiphytic, rhizospheric and
phyllospheric biocontrol bacteria has been expressed (Toy-
oda et al. 1988), which together with this present study sug-
gests that an important role for fusaric acid is for defence
and competition for species of Fusarium, and that fusaric-
acid-tolerant mutants of endophytic biocontrol bacteria
should be developed.
Acknowledgements
We thank L. Nakamura for his generous support and
supply of bacterial strains used in this study. We also
thank R. Bennett for technical assistance.
References
Arias, J.A. (1985) Secretory organelle and mitochondrial altera-
tions induced by fusaric acid in root cells of Zea mays.
Physiol Plant Pathol 27, 149–158.
Backmann, E. (1956) Der Einfluss von Fusarinsaure auf Die
Wasserpermeabilitat von Pflanzlichen Protoplasten. Phyto-
pathol Z 27, 255–288.
Bacon, C.W. and Hinton, D.M. (1996a) Fusaric acid and
pathogenic interactions of corn and non-corn isolates of
Fusarium moniliforme, a nonobligate pathogen of corn.
Adv Exp Med Biol 392, 175–191.
Bacon, C.W. and Hinton, D.M. (1996b) Symptomless endo-
phytic colonization of maize by Fusarium moniliforme.
Can J Bot 74, 1195–1202.
Bacon, C.W. and Hinton, D.M. (2001) Control of seedling
blight in wheat by Bacillus mojavensis. Phytopathology
91(suppl), S4.
Bacon, C.W. and Hinton, D.M. (2002) Endophytic and biolo-
gical control potential of Bacillus mojavensis and related
species. Biol Control 23, 274–284.
Bacon, C.W., Porter, J.K., Norred, W.P. and Leslie, J.F. (1996)
Production of fusaric acid by Fusarium species. Appl Envi-
ron Microbiol 62, 4039–4043.
Bacon, C.W., Yates, I.E., Hinton, D.M. and Meredith, F.
(2001) Biological control of Fusarium moniliforme in
maize. Environ Health Perspect 109, 325–332.
C.W. Bacon et al.
Bacon, C.W., Hinton, D.M., Porter, J.K., Glenn, A.E. and Kul-
dau, G.A. (2004) Fusaric acid, a Fusarium verticillioides
metabolite, antagonistic to the endophytic biocontrol bac-
terium Bacillus mojavensis. Can J Bot 82, 878–885.
Bochner, B.R., Huang, J.-C., Schieven, G.L. and Ames, B.N.
(1980) Positive selection for loss of tetracycline resistance.
J Bacteriol 143, 926–933.
Capasso, R., Evidente, A., Cutignano, A., Vurro, M., Zonno,
M.C. and Bottalico, A. (1996) Fusaric and 9,10-dehydro-
fusaric acids and their methyl esters from Fusarium nyga-
mai. Phytochemistry 41, 1035–1039.
Chanway, C.P. (1998) Bacterial endophytes: ecological and
practical implications. Sydowia 50, 149–170.
D’Alton, A. and Etherton, B. (1984) Effects of fusaric acid on
tomato root hair membrane potentials and ATP levels.
Plant Physiol 74, 39–42.
Davis, D. (1969) Fusaric acid in selective pathogenicity. Phyto-
pathology 59, 1391–1395.
Drysdale, R.B. (1984) The production and significance in phy-
topathology of toxins produced by species of Fusarium. In
The Applied Mycology of Fusarium ed. Moss, M.O. and
Smith, J.E. pp. 95–105. New York: Cambridge University
Press.
Duffy, B.K. and Défago, G. (1997) Zinc improves biocontrol
of Fusarium crown and root rot of tomato by Peudomonas
fluorescens and represses the production of pathogen
metabolites inhibitory to bacterial antibiotic biosynthesis.
Phytopathology 87, 1250–1257.
Gaumann, E. (1957) Fusaric acid as a wilt toxin. Phytopatho-
logy 47, 342–357.
Gordon, R.E., Hayward, M.D. and Pang, C.H. (1973) The
Genus Bacillus. Agriculture Handbook No. 427, Washing-
ton, D.C.: United States Department of Agriculture.
Hallmann, J., Quadt-Hallmann, A., Mahaffee, W.F. and
Kloepper, J.W. (1997) Bacterial endophytes in agricultural
crops. Can J Microbiol 43, 895–914.
Kobasyashi, D.Y. and Palumbo, J.D. (2000) Bacterial endo-
phytes and their effects on plants and uses in agriculture.
In Microbial Endophytes ed. Bacon, C.W. and White, J.F.
Jr., pp. 199–233. New York: Marcel Dekker, Inc.
Landa, B.B., Cachinero-Dı́az, J.M., Lemanceau, P., Jiménez-
Dı́az, R.M. and Alabouvette, C. (2002) Effect of fusaric
acid and phytoanticipins on growth of rhizobacteria and
Fusarium oxysporum. Can J Microbiol 48, 971–985.
Luz, J.M., Paterson, R.R.M. and Brayford, D. (1990) Fusaric
acid and other metabolite production in Fusarium oxyspo-
rum f. sp. vasinfectum. Lett Appl Microbiol 11, 141–144.
Malini, S. (1966) Heavy metal chelates of fusaric acid: In vitro
spectrophotometry. Phytopathol Z 57, 221–231.
Maloy, S.R. and Nunn, W.D. (1981) Selection for loss of tetra-
cycline resistance by Escherichia coli. J Bacteriol 145,
1110–1112.
Marasas, W.F.O. (2001) Discovery and occurrence of the fu-
monisin. A historical perspective. Environ Health Perspect
109, 239–243.
ª 2005 The Society for Applied Microbiology, Journal of Applied Microbiology 100 (2006) 185–194
No claim to original US government works
Fusaric acid inhibits Bacillus species
Marrè, M.T., Vergani, P. and Albergoni, F.G. (1993) Relation-
ship between fusaric acid uptake and its binding to cell
structures by leaves of Egeria densa and its toxic effects on
membrane permeability and respiration. Physiol Molec
Plant Pathol 42, 141–157.
Nakamura, L.K. (1989) Taxonomic relationships of black pig-
mented Bacillus subtilis strains and a proposal for Bacillus
amyloliquefaciens. Int J Syst Bacteriol 39, 295–300.
Nakamura, L.K., Roberts, M.S. and Cohan, F.M. (1999) Rela-
tionship of Bacillus subtilis clades associated with strains
168 and W23: a proposal for Bacillus subtilis subsp. subtilis
subsp. nov and Bacillus subtilis subsp. spizizenii subsp. nov.
Int J Syst Bacteriol 49, 1211–1215.
Notz, R., Maurhofer, M., Dubach, H., Haas, D. and Défago, G.
(2002) Fusaric acid-producing strains of Fusarium oxyspo-
rum alter 2,4-diacetylphloroglucinol biosynthetic gene
expression in Pseudomonas fluorescens CHA0 in vitro and
in the rhizosphere of wheat. Appl Environ Microbiol 68,
2229–2235.
Paquin, R. and Waygood, E.R. (1957) The effect of Fusarium
toxins on the enzymatic activity of tomato hypocotyl mito-
chondria. Can J Bot 35, 207–218.
Porter, J.K., Bacon, C.W., Wray, E.M. and Hagler, W.M. Jr.
(1995) Fusaric acid in Fusarium moniliforme cultures,
corn, and feeds toxic to livestock and the neurochemical
effects in the brain and pineal gland of rats. Nat Toxins 3,
91–100.
Riley, R.T., Norred, W.P. and Bacon, C.W. 1993. Fungal tox-
ins: recent concerns. Annu Rev Nutr 13, 167–189.
Roberts, M.S., Nakamura, L.K. and Cohan, F.M. (1994) Bacil-
lus mojavensis sp. nov., distinguishable from Bacillus subtil-
is by sexual isolation, divergence in DNA sequence, and
differences in fatty acid composition. Int J Syst Bacteriol
44, 256–264.
Roberts, M.S., Nakamura, L.K. and Cohan, F.M. (1996) Bacil-
lus vallismortis sp. nov., a close relative of Bacillus subtilis,
isolated from soil in Death Valley, California. Int J Syst
Bacteriol 46, 470–475.
Schnider-Keel, U., Seematter, A., Maurhofer, M., Blumer, C.,
Duffy, B.K., Gigot-Bonnefoy, C., Reimmann, C., Notz, R.,
Défago, G., Hass, D. and Keel, C. (2000) Autoinduction of
2,4-diacetylphoroglucinol biosynthesis in the biocontrol
agent Pseudomonas fluorescens CHA0 and repression by the
bacterial metabolites salicylate and pyoluteorin. J Bacteriol
182, 1215–1225.
Smith, T.K. and Sousadias, M.G. (1993) Fusaric acid content
of swine feedstuffs. J Agric Food Chem 41, 2296–2298.
Smith, T.K., McMillan, E.G. and Castillo, J.B. (1997) Effect of
feeding blends of Fusarium mycotoxin-contaminated
grains containing deoxynivalenol and fusaric acid on
growth and feed consumption of immature swine. J Anim
Sci 75, 2184–2191.
Sturz, A.V., Christie, B.R. and Nowak, J. (2000) Bacterial end-
ophytes: potential role in developing sustainable systems of
crop production. Crit Rev Plant Sci 19, 1–30.
193
Fusaric acid inhibits Bacillus species C.W. Bacon et al.

Takami, H., Nakasone, K., Takaki, Y., Maeno, G., Sasaki, R.,
Masui, N., Fuji, F., Hirama, C., Nakamura, Y., Ogasawara,
N., Kuhara, S. and Horikoshi, K. (2000) Complete genome
sequence of the alkaliphilic bacterium Bacillus halodurans
and genomic sequence comparison with Bacillus subtilis.
Nucleic Acids Res 28, 4317–4331.
Tamari, K. and Kaji, J. (1954) Studies on the mechanism of
the growth inhibitory action of fusarinic acid on plants.
J Bacteriol 41, 143–165.
Thrane, U. (2001) Developments in the taxonomy of Fusarium
species based on secondary metabolites. In Fusarium, Paul
E. Nelson Memorial Symposium ed. Summerell, B., Leslie,
J.F., Backhouse, D., Brydan, W.L. and Burgess, L.W. pp.
29–49. St. Paul, MN: American Phytopathological Society
Press.
Toyoda, H., Hashimoto, H., Utsumi, R., Kobayashi, H. and
Ouchi, S. (1988) Detoxification of fusaric acid by a fusaric
acid-resistant mutant of Pseudomonas solanacearum and its
application to biological control of Fusarium wilt of
tomato. Phytopathology 78, 1307–1311.
Yabuta, T., Kambe, K. and Hayashi, T. (1937) Biochemistry of
the bakanae-fungus. I. Fusarinic acid, a new product of
the bakanae fungus. J Agric Soc Jpn 10, 1059–1068.

ª 2005 The Society for Applied Microbiology, Journal of Applied Microbiology 100 (2006) 185–194
194 No claim to original US government works