BioControl

DOI 10.1007/s10526-017-9851-7

A multi-criteria approach for the selection of efficient

biocontrol agents against Botrytis cinerea on tomato
in Algeria
Yousra Bouaoud . Claire Troulet . Abdelhamid Foughalia . Odile Berge .
Kamel Aissat . Marc Bardin

Received: 5 April 2017 / Accepted: 6 November 2017

Ó International Organization for Biological Control (IOBC) 2017

Abstract In order to find biocontrol agents that are Keywords Gray mold  Biocontrol  Efficacy 
both efficient against Botrytis cinerea Pers. and Pseudomonas spp.  Solanum lycopersicum
adapted to tomato growing conditions in Algeria,
121 bacterial strains were collected from tomato plants
and nearby soils in two Bejaia greenhouses. A total of
37 strains were selected based on their ability to grow Introduction
on agar medium and on their different level of B.
cinerea mycelial growth inhibition in dual-culture Botrytis cinerea Pers. the causal agent of gray mold is
tests. These strains were identified at the species level one of the most damaging fungal plant pathogen
and those that corresponded to potential pathogens for worldwide (Dean et al. 2012) with significant eco-
humans or mammals were discarded. Among the nomic impact for multiple cash crops including tomato
remaining 25 candidates, three strains were selected (Elad 2016). In Algeria, gray mold is considered as
among the Pseudomonas genus for their significant one of the main tomato diseases (Aissat et al. 2008;
protective efficacy against B. cinerea on tomato, their Adjebli et al. 2015). In the Bejaia region, most of the
ability to grow at 15–25 °C and their inability to grow tomato production occurs in unheated greenhouses.
at 37 °C. These three strains significantly reduced the Under these conditions, the fungus can infect and
development of necrotic lesion and the sporulation of cause damages to tomato leaves, stem and fruit (Aissat
B. cinerea in a dose-dependent manner. This study et al. 2008). Leaf symptoms are frequent but result in
constitutes a first step towards the biological control of minor to mild damages to host health. In contrast, the
B. cinerea in tomato greenhouses in Algeria. expansion of necrotic lesions on stems, due to routine
practices such as leaf pruning, may typically result in
plant death (Aissat et al. 2008). Infections of tomato
pruning wounds by B. cinerea result in stem cankers
Handling Editor: Fouad Daayf. and can cause substantial yield losses (O’Neill et al.
1997; Decognet et al. 2010). Hence, stem cankers must
Y. Bouaoud  A. Foughalia  K. Aissat
Laboratoire d’écologie Microbienne, Faculté des Sciences be avoided to mitigate damages to host health and
de la Nature et de la Vie, Université Abderrahmane Mira, reduce crop losses (Aissat et al. 2008).
Bejaia 06000, Algeria Globally, chemical compounds constitute the main
control method to reduce gray mold incidence on
Y. Bouaoud  C. Troulet  O. Berge  M. Bardin (&)
Plant Pathology, INRA, 84140 Montfavet, France crops (Fillinger and Walker 2016). The use of
e-mail: marc.bardin@inra.fr chemical pesticides, including those used against B.

123
 Y. Bouaoud et al.

cinerea, has been increasing dramatically since 2000 greenhouse in France (BC1; Plougastel 1989) and in
in Algeria (1013 and 2005 tons of active ingredients in Algeria (ALG66; Bejaia 2007). They were maintained
2000 and 2005, respectively; http://www.fao.org/ as spore suspensions in glycerol phosphate buffer at
faostat/fr/#home). Alternative control methods such - 20 °C. To prepare inoculum for further assays, an
as biological control agents against Botrytis-incited aliquot of each spore suspension was used to grow
diseases have gained increasing attention worldwide these strains on potato dextrose agar medium (PDA,
(Nicot et al. 2016) but few microbe-based products are Difco Laboratory Detroit, 39 g l-1) in Petri dish
commercially available in the North-African region. plates. The plates were then incubated in a growth
To control B. cinerea on tomato, three products are chamber (21 °C, photoperiod L:D 14:10,
commercialized in Morocco (http://eservice.onssa. 114 lmol m-2 s-1) for 14 days. Conidia suspension
gov.ma/IndPesticide.aspx), one in Tunisia (http:// was adjusted to 106 spores ml-1 for further use as
www.flehetna.com/fr/annuaire-flehetna/phyto), and described by Decognet et al. (2009).
none in Algeria (Anonymous 2015). Therefore, there
is an urgent need to search for new biological control Collection of bacterial strains and culture
agents that are both efficient against the disease, and conditions
adapted to the crop and the environmental conditions
encountered in this region. One hundred and twenty-one strains were collected in
The selection of a microbiological control agent 2014 in two tomato greenhouses in Bejaia (Algeria). In
should be based on an appropriate screening procedure this region, production of tomato occurs in soil and
against the targeted plant pathogen (Köhl et al. 2011). exclusively in unheated plastic tunnel greenhouses
First, it is best to isolate microorganisms from without any climatic regulation. The greenhouses
appropriate niches in areas with environmental con- were located on a thin strip of land between the sea
ditions similar to those encountered in the targeted shore and the mountains (generally less than 500 m
region of use. Second, the choice of screening methods wide). Climate conditions on this geographical loca-
to evaluate the antagonistic potential of microbes tion are favorable to the development of gray mold
against the targeted pathogen is also critical (Pliego during the tomato growing period (Aissat et al. 2008).
et al. 2011). Twenty-eight strains were isolated directly from
The objective of the present study was to identify tomato stems in a heavily Botrytis-diseased green-
efficient biocontrol agents against B. cinerea and house. Pieces of stems were removed from tomato
adapted to tomato growing conditions in Algerian plants and deposited directly on nutrient agar medium
greenhouses. To this end, bacteria strains were (NA, peptone 5 g l-1, yeast extract 2.5 g l-1, beef
collected in tomato greenhouses in Algeria. Then, a extract 1 g l-1, sodium chloride 5 g l-1, agar
multi-criteria approach including molecular charac- 15 g l-1, pH 7). Ninety-three strains were also
terization of bacteria strains, in vitro and in planta isolated from two soil samples and purified as
bioassays was designed. This approach resulted in the described by Bizuye et al. (2013). From this collection
selection of three strains out of the 121 initially of bacteria, 11 strains from tomato and 26 strains from
isolated. These three strains were then tested for their soil samples were selected for further studies
ability to protect tomato and to inhibit the sporulation (Table 1). These strains were maintained on 20%
of B. cinerea in a dose-dependent manner. Addition- glycerol and stored at - 20 °C. In order to evaluate
ally, the link between in vitro inhibition and in planta their ability to grow at different temperatures, the
protection for a subset of 25 strains was assessed. presence/absence of growth was qualitatively deter-
mined at 15, 22, 25 and 37 °C on Tryptic Soy Agar
medium (TSA: Trypton Caseine Soja, 30 g l-1 and
Materials and methods Agar, 15 g l-1, Fisher Scientific) after seven days. For
further experiments, cell suspension were made by
Fungal strains adding bacterial cells grown on TSA medium at 25 °C
for 48 h to 1 ml distilled sterile water. The final
Two strains of B. cinerea were used in this study. concentration achieved for each bacterial suspension
These strains were collected from tomato plants in a was assessed by dilution plating on TSA medium. The

123
A multi-criteria approach for the selection of efficient biocontrol agents

Table 1 Molecular identification, growth at various tempera- measured by dual-culture method. Tentative phylogenetic
tures and antagonistic properties of the 37 bacterial strains affiliation is based on 16S rDNA or gyrB gene sequences
isolated in 2014 in Bejaia (Algeria) against B. cinerea
Bacterial Type of Bacterial species Accession Growth Mycelial growth inhibition (%)c
code sample number temperature on
TSA medium
(°C)b
GenBank 15 22 25 37 ALG66 BC1

T16 Soil Bacillus methylotrophicus KY307916 - ? ? ? 55.6 ± 1.7 cdefg 58.0 ± 0.6 bcd
T18 Soil Bacillus methylotrophicus KY307917 - ? ? ? 64.7 ± 2.2 a 60.1 ± 0.6 b
T21 Soil Bacillus cereusd KY307918 ? ? ? ? 34.8 ± 0.6 i 53.2 ± 0.6 cde
T23 Soil Bacillus methylotrophicus KY307919 - ? ? ? 62.3 ± 0.6 ab 64.1 ± 1.8 ab
T28 Soil Bacillus methylotrophicus KY307920 - ? ? ? 52.7 ± 0.6 defg 50.6 ± 1.5 e
T29 Soil Bacillus methylotrophicus KY307921 - ? ? ? 64.4 ± 1.4 ab 68.5 ± 0.9 a
T31 Soil nda ? ? ? ? 12.2 ± 0.1 m 17.7 ± 1.0 gh
T33 Soil Pseudomonas azotoformans KY307922 ? ? ? - 0.0 ± 0.0 o 2.1 ± 0.3 kl
T34 Soil Bacillus methylotrophicus KY307923 - ? ? ? 59.1 ± 1.4 abcd 61.3 ± 0.9 b
T57 Soil Bacillus aerophilus KY307924 - ? ? ? 23.9 ± 1.1 k 19.0 ± 0.1 g
T75 Soil Bacillus flexus KY307925 ? ? ? ? 20.0 ± 0.0 kl 0.0 ± 0.0 l
T80 Soil Stenotrophomonas maltophiliaa,d KY379237 ? ? ? ? 4.6 ± 1.0 no 17.9 ± 3.4 gh
T83 Soil Pseudomonas azotoformans KY307926 ? ? ? - 2.7 ± 1.2 o 0.0 ± 0.0 l
T90 Soil Bacillus methylotrophicus KY307927 - ? ? ? 63.6 ± 1.7 ab 64.5 ± 1.2 ab
M4 Soil nda ? ? ? ? 10.0 ± 0.0 mn 19.6 ± 0.0 g
M6 Soil nda ? ? ? ? 0.0 ± 0.0 o 0.0 ± 0.0 l
M7 Soil Pseudomonas azotoformans KY307907 ? ? ? - 0.0 ± 0.0 o 4.5 ± 0.5 jkl
M9 Soil Bacillus cereusd KY307908 ? ? ? ? 60.9 ± 1.7 abc 58.8 ± 0.9 bd
M24 Soil Bacillus thuringiensis KY307909 ? ? ? ? 52.0 ± 0.0 efg 51.6 ± 0.0 ce
M25 Soil Bacillus cereusd KY307910 ? ? ? ? 49.5 ± 0.9 gh 49.0 ± 1.5 e
M32 Soil Stenotrophomonas maltophiliaa,d KY379236 ? ? ? ? 0.0 ± 0.0 o 1.5 ± 0.8 kl
M36 Soil Bacillus pseudomycoides KY307911 ? ? ? ? 18.8 ± 0.7 kl 7.8 ± 0.8 ijk
M42 Soil Bacillus aerophilus KY307912 ? ? ? ? 5.1 ± 0.1 no 11.5 ± 0.7 hi
M43 Soil Bacillus thuringiensis KY307913 ? ? ? ? 57.5 ± 2.5 bcdef 57.7 ± 1.2 bd
M46 Soil Bacillus thuringiensis KY307914 ? ? ? ? 58.2 ± 1.4 abcde 50.0 ± 1.2 e
M65 Soil Bacillus aryabhattai KY307915 ? ? ? ? 1.3 ± 0.0 o 0.0 ± 0.0 kl
CT3 Tomato stem Pseudomonas aeruginosad KY307901 ? ? ? ? 51.2 ± 0.9 fg 57.9 ± 0.6 bd
CT6 Tomato stem Bacillus thuringiensis KY307902 ? ? ? ? 4.0 ± 1.6 no 5.1 ± 1.1 jkl
CT8 Tomato stem Brevibacterium linens KY307903 ? ? ? ? 5.8 ± 2.0 no 17.2 ± 1.3 gh
CT9 Tomato stem Pseudomonas extremorientalis KY307904 ? ? ? - 17.0 ± 0.8 l 16.0 ± 1.5 gh
CT20 Tomato stem Leucobacter chromiireducens KY307905 - ? ? - 29.4 ± 2.5 j 17.3 ± 1.1 gh
CT22 Tomato stem Pseudomonas helmanticensis KY307906 ? ? ? - 45.0 ± 1.0 h 42.3 ± 1.4 f
CS1 Tomato stem Brachybacterium KY307896 ? ? ? ? 0.0 ± 0.0 o 9.3 ± 0.0 ij
paraconglomeratum
CS2 Tomato stem Bacillus cereusd KY307897 ? ? ? ? 10.5 ± 2.0 mn 1.4 ± 0.5 kl
CS7 Tomato stem Planomicrobium glaciei KY307898 ? ? ? ? 0.0 ± 0.0 o 15.3 ± 1.7 gh

123
 Y. Bouaoud et al.

Table 1 continued

Bacterial Type of Bacterial species Accession Growth Mycelial growth inhibition (%)c
code sample number temperature on
TSA medium
(°C)b
GenBank 15 22 25 37 ALG66 BC1

CS9 Tomato stem Bacillus cereusd KY307899 ? ? ? ? 17.7 ± 1.0 l 20.7 ± 3.5 g
CS17 Tomato stem Staphylococcus sp.d KY307900 ? ? ? ? 0.0 ± 0.0 o 0.0 ± 0.0 l
a
Strains M32 and T80 (Stenotrophomonas maltophilia) were identified based on gyrB gene sequence; nd not determined. DNA
sequences for strains M4, M6 and T31 were not suitable for analysis despite several attempts to sequence the genes 16 s rDNA, gyrB
and rpoD
b
?/- presence or absence of growth on TSA medium after seven days of incubation
c
Percent mycelial growth inhibition of the strains ALG66 and BC1 of B. cinerea relative to the control without bacteria, measured by
dual culture assay on Petri plates ± SE. For each strain of B. cinerea, data were analyzed with an ANOVA (F = 237.1, df = 36,74,
P \ 0.0001 for ALG66 and F = 217.3, df = 36,74, P \ 0.0001 for BC1). Values of mycelial growth inhibition of each B. cinerea
strain with the same letter are not significantly different relative to the Newman–Keuls multiple range test (P [ 0.05)
d
Bacterial strains identified as species known as pathogens of human or mammals based on scientific literature

maximum cell concentrations obtained were between inhibition due to volatiles compared to the control was
107 and 109 CFU ml-1, depending on the bacterial calculated. This test was conducted with three repli-
strain. cates for each bacterial strain.

In vitro inhibition of Botrytis cinerea Molecular identification of the bacteria

The antagonistic activity of each of the 121 bacterial Genomic DNA of 37 bacterial strains was extracted
strains against each of the two B. cinerea strains was using Qiagen Kit and PCR amplification of their 16S
assessed on PDA medium by dual-culture assay, as rDNA gene was carried out as described by Berge
described by Petatan-Sagahon et al. (2011). Petri dish et al. (2002) with the universal forward primer fd1 (50 -
plates were incubated in a growth chamber (21 °C, AGAGTTTGATCCTGGCTCAG-30 ) and the univer-
photoperiod L:D 14:10) for three days. This test was sal reverse primer S17 (50 -GTTACCTTGTTAC-
conducted with three replicates for each bacterial GACTT-30 ). PCR amplification of the two
strain. Growth was measured based on the radius of housekeeping genes gyrB and rpoD was performed
mycelium produced (in mm). Bacteria antagonistic for strains for which clean 16S rDNA gene sequence
activity was calculated as the percent of mycelial could not be obtained, as described by Morris et al.
growth inhibition compared to the control. (2008). PCR products were sent for sequencing
(GenoScreen Lille, France). The sequences were
In vitro volatile effect compared with reference sequences obtained from
EzTaxon-e and NCBI databases using BLAST anal-
Antagonism due to bacteria volatile compounds was ysis, and strains were identified on the basis of the best
evaluated for 25 bacterial strains, according to the match in the database.
protocol described by Jamalizadeh et al. (2008).
Sealed Petri dish plates were incubated in controlled Induction of HR on tobacco plants
conditions for two or three days at 21 °C (photoperiod
L:D 14:10). The diameter of fungal colony (mm) in the The ability of the bacterial strain to induce hypersen-
presence of volatiles produced by each bacteria strain sitive response (HR) on tobacco leaves was deter-
was measured. The percentage of mycelial growth mined. Positive HR induced on tobacco is an indicator

123
A multi-criteria approach for the selection of efficient biocontrol agents

of Gram-negative bacteria pathogenicity on plants Estimation of development and sporulation of B.

(Sundh et al. 2011). Eight weeks old tobacco plants cinerea on detached tomato stem
(Nicotiana tabacum cv. Xanthi) were grown in a
greenhouse in individual pots containing a commercial To test the effect of bacteria on the development of
potting mixture (TS03 404, Klasmann-Deilmann, stem canker and on the sporulation ability of B.
Germany). A bacterial suspension (200 ll at 108 cinerea, a bioassay was performed on detached tomato
CFU ml-1 depending on bacterial strain) was injected stem. Three bacterial strains (CT22, CT9 and T33)
within the leaves with a syringe (Morris et al. 2010). were tested at various concentrations each against the
Pseudomonas syringae strain CC94 was used as strains BC1 and ALG66 of B. cinerea. The same
positive control and sterile distilled water as negative inoculation procedure was performed as described
control. After 24–48 h, presence or absence of necrosis above for the whole tomato plant test. After 2 h of
was recorded. incubation in a growth chamber (21 °C, photoperiod
L:D 14:10, RH above 90%), two 4-cm stem pieces
Whole tomato plant test were excised around the inoculated zones from each
tomato plant. Each stem sample was weighed (grams
Twenty-five bacterial strains were evaluated for their fresh weight) and then placed individually in the
protective efficacy against the two strains of B. cinerea 55-mm diameter cap of a screw-cap tube. The caps
on whole tomato plants. Tomato plants (Solanum were placed in transparent polystyrene boxes over
lycopersicum cv. Monalbo, INRA Avignon, France) moistened filter paper, covered with lids, and placed in
were grown in a greenhouse in individual pots a growth chamber (21 °C, photoperiod L:D 14:10,
containing a commercial potting mixture and watered 114 lmol m-2 s-1). For each concentration of a given
daily with a nutrient solution (100–200 ml day-1 of bacteria, six replicate samples were inoculated (two
NPK 16–10–25 supplemented with micronutrients) per plant 9 three plants) and the whole experiment
during 8–10 weeks. Tests were performed by inocu- was repeated three times. The length of stem lesion (in
lating successively B. cinerea and bacteria on pruning mm) was measured between the 3rd and the 7th day
wounds of tomato plants as described by Bardin et al. after inoculation. The amount of spores produced by
(2008). The control was assessed by inoculating B. B. cinerea after ten days of incubation was counted
cinerea without the bacterial suspension. In order to under the microscope using Malassez hemocytometer
test the potential pathogenicity of the bacteria on according to the method described by Abro et al.
tomato plants, each strain was also inoculated alone on (2013). Spore production was expressed as the number
pruning wounds. The length (mm) of resulting stem of spores per gram of fresh weight and the percentage
lesions was monitored daily. Disease development on of sporulation inhibition compared to the control was
stems was assessed by computing the area under the calculated.
disease progress curve (AUDPC) between four and
seven days after inoculation as described by Decognet Statistical analysis
et al. (2009). Three plants per bacterial strain, each
with three pruning wounds, were inoculated. A As normality test indicated that B. cinerea mycelial
percentage of tomato protection generated by the growth inhibition data for the 121 strains of bacteria
bacteria, relative to the control without bacteria, was were not normally distributed (Shapiro–Wilks nor-
computed for each strain. mality test, P \ 0.0001), we performed a non-para-
Additionally, the effect of the concentration of metric analysis of variance (Kruskal–Wallis test) to
bacteria on the efficacy of protection of tomato against evaluate a bacterial strain effect on the mycelial
B. cinerea ALG66 was assessed for the three strains growth inhibition of B. cinerea. To test the effect of
CT22, CT9 and T33. Three concentrations of each the bacteria on the development of B. cinerea in Petri
bacteria suspension (the maximal dose and two plates (radius or diameter in mm) or on the plants
dilutions at 10-1 and 10-2 of the maximal dose) were (AUDPC or sporulation), data were subjected to
tested. Three independent repetitions of this test were analysis of variance (ANOVA) followed by compar-
performed. ison of means using the Newman–Keuls multiple
range test (NKT), when appropriate. The relationship

123
 Y. Bouaoud et al.

between the different parameters calculated (% of grow at 15 °C suggesting a large range of growth
in vitro growth inhibition, % of protection on tomato temperature (Table 1). The strains did not cause any
plants) was assessed by computing the non-parametric symptoms on tomato, and the five Gram-negative
Spearman rank coefficient. Statistical analyses were strains did not induce HR on tobacco, suggesting an
performed using Statistica (Statsoft Inc., Tulsa, USA). absence of pathogenicity on plant.
Data are reported in the text, tables and figures as mean
values ± SE. In vitro volatile effect

Mycelial growth of B. cinerea on PDA medium was

Results affected by volatiles produced by some of the 25
bacterial strains tested (Fig. 2). For each B. cinerea
In vitro effect of the 121 bacterial strains strain, a significant bacterial strain effect was observed
(ANOVA done on mycelial growth diameter,
Dual-culture assay were performed on PDA medium F = 49.0, df = 25,130, P \ 0.0001 for BC1 and
in order to evaluate the ability of each antagonist F = 14.6, df = 25,55, P \ 0.0001 for ALG66). Nine-
candidate to inhibit B. cinerea mycelial growth. For teen bacterial strains significantly reduced the myce-
each B. cinerea strain, a significant bacterial strain lial growth of BC1 strain, compared to the control
effect was observed (Kruskal–Wallis test done on without bacteria (NKT applied to the mycelial growth
mycelial growth inhibition, v2 = 342.0, df = 120, diameter data, P \ 0.05) and sixteen strains signifi-
P \ 0.0001 for BC1 and v2 = 347.0, df = 120, cantly reduced ALG66 strain mycelial growth of
P \ 0.0001 for ALG66). A highly significant correla- (NKT applied to the mycelial growth diameter data,
tion was detected between the mycelial growth P \ 0.05). A significant correlation was detected
inhibition of BC1 and the mycelial growth inhibition between the percentage of mycelial growth inhibition
of ALG66 on PDA medium (Fig. 1; Spearman’s of BC1 and the percentage of mycelial growth
correlation R = 0.76, P \ 0.05). Among the 121 inhibition of ALG66 in presence of the 25 bacterial
strains collected, 37 were retained for further inves- strains (Spearman’s correlation R = 0.67, P \ 0.05).
tigations based on their ability to grow on TSA and NA
media at 25 °C (data not shown) and because of their Protection efficacy on tomato plant of the 25
different in vitro antagonistic activities against B. bacterial strains
cinerea (Table 1).
Protection of tomato against B. cinerea is highly
Identification and characterization of the 37 variable among the bacterial strains tested (Fig. 3) and
bacterial strains a significant strain effect was detected (ANOVA on
AUDPC values, F = 17.5, df = 25,223, P \ 0.0001
DNA sequences of 16S rDNA or gyrB genes allowed for BC1 and F = 15.0, df = 25,286, P \ 0.0001 for
the phylogenetic affiliation of 34 strains (Table 1). ALG66). The length of stem lesion due to BC1 and
Sequences have been deposited in Genbank (https:// ALG66 was significantly reduced in the presence of 12
www.ncbi.nlm.nih.gov/). The strains M4, M6 and T31 and seven bacteria strains, respectively, compared to
were excluded for subsequent experiments because the control without bacteria (NKT applied to the
clean sequences of 16S rDNA, gyrB and rpoD genes AUDPC values, P \ 0.05). Three strains, i.e. CT9,
could not obtained. A literature search was carried out CT22 and T33 resulted in almost a total protection of
on the 34 identified strains to determine those that are tomato and were selected for further studies.
potentially pathogens for humans and mammals. Nine
of them were discarded because of potential risks Relationship between the different screening
(Table 1). Among the remaining 25 strains, 19 were methods
able to grow at 37 °C but, in the absence of evidence of
human or mammals pathogenicity, they were con- In order to evaluate if in vitro mycelial growth
served for further analysis. All the strains were able to inhibition of B. cinerea is a good indicator of the
grow at 22 and 25 °C, and 16 of them were able to efficacy of bacterial strains as biological control agent

123
A multi-criteria approach for the selection of efficient biocontrol agents

80

Mycelial Growth Inhibition

70
60

(% BC1)
50
40
30
20
10
0
0 10 20 30 40 50 60 70
Mycelial Growth Inhibition (% ALG66)

Fig. 1 Relationship between the mycelial growth inhibition of PDA medium during three days at 22 °C. Each point corre-
B. cinerea BC1 strain and ALG66 strain caused by the 121 sponds to the mean of three replicates. The 37 strains selected
bacterial strains isolated in this study (Spearman’s rank for further analysis were indicated in grey (filled grey circle)
coefficient R = 0.76, P \ 0.0001). Tests were realized on

100
Volatiles inhibition (%)

80
60
40
20
0
-20
CT22

CT20
M24

M43

M46

M36

M42
M65
CT9

CT6

CT8
CS7

CS1
T33
T83
T18

T16

T23

T34
T90

T28
T29

T57

T75
M7

Bacterial isolates

Fig. 2 Effect of volatiles from the 25 bacteria isolates on B. a significant difference in the mycelial growth of BC1 and
cinerea BC1 (black bars) and ALG66 (white bars) mycelia ALG66 respectively, compared to the control without bacteria
growth on PDA medium, measured as percent of mycelia (NKT applied to the mycelial growth values, P \ 0.05). The
growth inhibition. Each value represents the mean of three percentages of volatiles inhibition are ranked in decreasing
replicates. The vertical bar associated to each histogram values against BC1 strain
correspond to SE. Symbols * and ? above histograms indicate

on tomato, relationships between protection of tomato two in vitro tests, dual-culture and volatiles
and mycelial growth inhibitions were studied. A (R = 0.30, P = 0.15 for BC1 and R = 0.39,
significant correlation was observed between the P = 0.05 for ALG66).
inhibition of the mycelial growth of B. cinerea due
to bacteria volatiles and the protection of tomato plant Dose effect of bacteria on the lesion development
against the two B. cinerea strains (Spearman’s corre- of B. cinerea on tomato
lation R = 0.45, P = 0.03 for BC1 and R = 0.46,
P = 0.02 for ALG66). No significant correlation was Two types of bioassays were performed to assess the
observed between the inhibition of B. cinerea mycelial efficacy of different doses of the strains CT22, CT9
growth in dual-culture assays and the protection of and T33 to protect tomato against B. cinerea. In the
tomato plant (R = 0.07, P = 0.75 for BC1 and detached tomato stem bioassay, a high level of
R = 0.17, P = 0.41 for ALG66). More surprisingly, protection ([ 81%) was observed for each strain at
no significant correlation was observed between the doses superior to 107 CFU ml-1 (Fig. 4). For doses

123
 Y. Bouaoud et al.

100
80

Protection (%)
60
40
20
0
-20
-40
-60
CT22

CT20
M24
M65
M46

M42

M43

M36
CT9

CT8

CT6
CS1

CS7
T33

T57

T28

T90
T29

T75

T23
T16
T18
T83

T34
M7

Bacterial strains

Fig. 3 Effect of the 25 bacterial isolates on the protection of significant difference of the development of stem lesion due to
tomato (in %) against B. cinerea strains BC1 (black bars) and BC1 and ALG66 respectively, compared to the control without
ALG66 (white bars). Each value represents the mean of nine bacteria (NKT applied to the AUDPC values, P \ 0.05). The
replicates. The vertical bars associated to histograms correspond percentages of protection are ranked in decreasing values
to the SE. Symbols * and ? above histograms indicate a against BC1 strain

lower than 107 CFU ml-1, results were variable, i.e. Effect of bacteria on the sporulation of B. cinerea
protection rate ranged from 34% to 100% depending on tomato
on the B. cinerea strain and the bacteria strain. A two-
way ANOVA (strain of B. cinerea 9 dose of bacteria) Sporulation of BC1 or ALG66 fungal strain was
revealed an absence of B. cinerea strain effect completely or partially inhibited after ten days of
(F = 0.0, df = 1,122, P = 0.93 and F = 0.1, incubation at 21 °C in presence of one of the bacterial
df = 1,120, P = 0.74, for CT22 and CT9, respec- strain CT9, CT22 or T33 (Fig. 6). Inhibition of
tively) and a significant dose effect (F = 3.3, sporulation ranged from 18 to 100% depending on
df = 9,122, P = 0.001 and F = 8.9, df = 11,120, the strain of bacteria, the dose of bacteria and the strain
P \ 0.0001 for CT22 and CT9, respectively) on the of B. cinerea. For each independent test, a significant
protection of tomato, with no interaction effect effect of the dose of bacteria on the sporulation of B.
(F = 0.3, df = 9,122, P = 0.97 and F = 0.2, df = cinerea was systematically observed (ANOVA on
11,120, P = 0.99, for CT22 and CT9, respectively). number of spores produced giving a P \ 0.05 for each
For T33, a significant interaction was detected (two- of the three independent tests for each of the three
way ANOVA, F = 3.3, df = 11,120, P = 0.0006). A strains of bacteria). The number of spores produced by
one-way ANOVA realized for each B. cinerea strain B. cinerea in presence of bacteria was always signif-
revealed a significant dose effect (ANOVA, F = 15.9, icantly lower than that of the control, except for T33 at
df = 11,60, P \ 0.0001 and F = 7.3, df = 11,60, 3.6 106 CFU ml-1 against BC1 (Fig. 6) for which the
P \ 0.0001 for BC1 and ALG66, respectively). protection rate measured at Seven days after inocula-
In the whole-tomato plant bioassay, a significant tion was of 34%.
correlation was observed between the dose of bacteria
and the percentage of protection against B. cinerea
ALG66 for each of the three bacteria tested (Fig. 5). A Discussion
high protection rate ([ 75%) was observed for the
strains CT22 and T33 at all the doses tested. Three bacteria strains belonging to the species Pseu-
domonas extremorientalis (CT9), P. helmanticensis
(CT22) and P. azotoformans (T33) have shown an
efficient and significant ability to protect tomato plants
against B. cinerea. These bacteria, isolated within
greenhouses heavily contaminated by B. cinerea,

123
A multi-criteria approach for the selection of efficient biocontrol agents

Fig. 4 Protection of tomato CT22

against B. cinerea exhibited 100
by the three bacteria CT22,

Protection (%)
CT9 and T33. Tests were 80
realized on detached tomato 60
stem at different
concentrations of bacteria 40
(dose of bacteria in
20
CFU ml-1 presented in a
Log scale) against two 0
strains of B. cinerea BC1 1E+05 1E+06 1E+07 1E+08 1E+09 1E+10
(filled triangle) and ALG66 Dose of bacteria (CFU ml-1)
(open circle). Each point
corresponds to the mean of
six replicates. The vertical CT9
bars associated to each point 100
Protection (%)

correspond to the SE
80
60
40
20
0
1E+05 1E+06 1E+07 1E+08 1E+09 1E+10
Dose of bacteria (CFU ml-1)

T33
100
Protection (%)

80
60
40
20
0
1E+05 1E+06 1E+07 1E+08 1E+09 1E+10
Dose of bacteria (CFU ml-1)

either on tomato plant (CT9, CT22) or in the soil helmanticensis was identified for the first time in
(T33), were able to grow at temperatures usually forest soil in Spain (Ramirez-Bahena et al. 2014) and
encountered in the tomato greenhouses of Bejaia and has been shown as capable of partially suppressing the
favorable for the development of B. cinerea (Jarvis mycelial growth of the plant pathogenic fungi Fusar-
1992; Adjebli 2014). This suggests that these strains ium solani and Phoma herbarum (Fan et al. 2015).
are potential efficient biocontrol agents to manage B. Dual-culture assays on agar plates suggested that this
cinerea under tomato growing conditions in Algeria, bacteria was able to produce metabolites with toxic
i.e. within unheated tunnel greenhouses located on a properties towards a wide range of fungi, including B.
thin strip of land bordered by the sea to the north and cinerea as shown in the present study with CT22.
the mountains to the south (Aissat et al. 2008). Pseudomonas azotoformans is involved in potato plant
To our knowledge, this study is the first that growth promotion (Andreote et al. 2009) and is able to
revealed the ability of these three species of Pseu- reduce the severity of anthracnose caused by Col-
domonas to protect tomato against B. cinerea and to letotrichum orbiculare on cucumber plants after root
reduce its sporulation ability. Pseudomonas treatment (Sang et al. 2014). In our study, we have

123
 Y. Bouaoud et al.

Fig. 5 Relationship 100 CT22

between the protection of

Protection (%)
tomato plant against B. 80
cinerea strain ALG66 (in
%) and the concentration of 60
bacteria (dose of bacteria in
CFU ml-1 presented in a 40
Log scale). The vertical bars R = 0.89
20 P = 0.04
associated to each point
correspond to the SE. The 0
non-parametric Spearman’s 1E+06 1E+07 1E+08 1E+09 1E+10
rank coefficient and the
associated P value were Dose of bacteria (CFU ml-1)
assessed for each bacteria
100 CT9
Protection (%)

80
60
40
R= 0.79
20 P = 0.04

0
1E+06 1E+07 1E+08 1E+09 1E+10
Dose of bacteria (CFU ml-1)

100
T33
Protection (%)

80
60
40
R = 0.83
20
P = 0.04
0
1E+06 1E+07 1E+08 1E+09 1E+10
Dose of bacteria (CFU ml-1)

shown that P. azotoformans T33 is able to protect bacteria against B. cinerea on tomato. These metabo-
tomato against B. cinerea with no direct in vitro lites will be characterized in subsequent studies for
inhibition, suggesting that a different mechanism than these three bacterial strains.
antibiosis underlies the protective ability of this strain. Another striking result obtained in this study is that
Pseudomonas extremorientalis is able to reduce the a low dose of the selected bacteria allows a partial
development of Fusarium solani on cucumber and the protection of tomato against B. cinerea and a signif-
in vitro growth of Pythium ultimum but it has shown no icant reduction in the production of secondary inocu-
effect against B. cinerea (Egamberdieva et al. 2011). lum. The reduction of sporulation ability by B. cinerea
In our study, a reduced in vitro effect towards B. could be an important target for disease management
cinerea was observed with the strain CT9 suggesting strategies, in particular in the greenhouse where
that different strains of P. extremorientalis may have secondary inoculum produced on the crop has a
different antagonistic properties. The three strains are crucial role in the spread of gray mold epidemics
all capable of producing volatiles having effective (Decognet et al. 2009; Bardin et al. 2014). This fungus
anti-Botrytis properties suggesting a possible role of can produce a large quantity of spores on diseased
these compounds in the protective efficacy of these tissue (Nicot et al. 1996) that have shown to be

123
A multi-criteria approach for the selection of efficient biocontrol agents

CT22

Number of spores per

BC1
14
a'
12

gram (×105)
10 a ALG66
8
A
6 b'
b
4 B
2 c c' C
0
control 10 10 control 1.9 10 1.9 10 control 1.9 10 1.9 10
Dose of bacteria (CFU ml-1)
CT9
Number of spores per

BC1
14 a'
gram (×105)

12 ALG66
10 a
8 b' A
6 b
4
2 c b' B B
0
control 1.7 10 1.7 10 control 3.5 10 3.5 10 control 3.5 10 3.5 10
Dose of bacteria (CFU ml-1)

T33
Number of spores per

BC1
14 a'
12
gram (×105)

a' ALG66
10 a
8 A
6 b
4 b b' B
2 c B
0
control 4.5 10 4.5 10 4.5 10 control 3.6 10 3.6 10 control 3.6 10 3.6 10
Dose of bacteria (CFU ml-1)

Fig. 6 Number of spores produced per gram of tomato stem by each repetition and for each bacterial isolate, values with the
strains BC1 and ALG66 of B. cinerea after ten days of same lowercase letter (a, b, c for repetition 1 and a0 , b0 , c0 for
incubation, in absence (control, plain histogram in black for repetition 2) are not significantly different according to
BC1 and in white for ALG66) or presence of the three bacterial Newman–Keuls multiple range test (P \ 0.05). For ALG66,
isolates CT22, CT9 and T33 at various concentrations (dots for each bacterial isolate, values with the same uppercase letter
within histogram). Each value represents the mean of six are not significantly different according to Newman–Keuls
replicates. The vertical bars associated to histograms correspond multiple range test (P \ 0.05)
to SE. For BC1, two independent repetitions were done. For

partially suppressed with the help of biological control interactive effects of bacteria biocontrol agents and
agents (Köhl et al. 1995). Another point that now plant fertilization on plant health, as already revealed
deserves to be assessed is the viability and the by Abro et al. (2014).
pathogenicity of the secondary inoculum produced In conclusion, this study highlights the ability of the
on plants treated with the bacteria. In a recent study, bacterial strains P. helmanticensis CT22, P. azotofor-
Abro et al. (2013) revealed that the level of N-fertil- mans T33 and P. extremorientalis CT9 to protect tomato
ization of plants have an effect on the pathogenicity of plants against B. cinerea. It also confirms that dual-
the resulting secondary inoculum produced by B. culture tests are not good indicators of the protective
cinerea. Further experiments will be carried out to test efficacy of a biocontrol agent in planta. Testing the
this hypothesis, as well as to assess the potential efficacy of these three strains in tomato crop conditions

123
 Y. Bouaoud et al.

in the Bejaia region greenhouses is now necessary. greenhouses in northern Algeria. J Phytopathol
These trials will be extended to other areas of tomato 163:124–132
Aissat K, Nicot PC, Guechi A, Bardin M, Chibane M (2008)
production in Algeria, where B. cinerea is a problem. Grey mould development in greenhouse tomatoes under
Further studies will also be carried out to assess the drip and furrow irrigation. Agron Sustain Dev 28:403–409
effect of these bacteria on different tomato cultivars and Andreote FD, de Araùjo WL, de Azevedo JL, van Elsas JD, da
on other major tomato diseases in Algeria such as late Rocha UN, van Overbeek LS (2009) Endophytic colo-
nization of potato (Solanum tuberosum L.) by a novel
blight and powdery mildew. Further studies should also competent bacterial endophyte, Pseudomonas putida strain
aim at understanding the mechanisms underlying bac- P9, and its effect on associated bacterial communities.
teria antagonistic activity to optimize the use of these Appl Environ Microbiol 75:3396–3406
biocontrol agents in the field. Even if an effect of the Anonymous (2015) Index des produits phytosanitaires à usage
agricole. Ministère de l’Agriculture, du Développement
volatiles on the mycelial growth of B. cinerea is rural et de la Pêche. Direction de la protection des végétaux
systematically observed, the role of these molecules in et des contrôles techniques. République Algérienne
the protective efficacy of bacteria in planta is not Démocratique et Populaire
established. Other mechanisms may be involved, which Bardin M, Fargues J, Nicot PC (2008) Compatibility between
biopesticides used to control grey mould, powdery mildew
deserves to be explored. Finally, assessment of the and whitefly on tomato. Biol Control 46:476–483
sensitivity of a collection of B. cinerea strains to these Bardin M, Decognet V, Nicot PC (2014) Remarkable predom-
bacteria needs to be explored in order to evaluate the inance of a small number of genotypes in greenhouse
stability and durability of their efficacy (Bardin et al. populations of Botrytis cinerea. Phytopathology
104:859–864
2015) towards the known diversified population of the Bardin M, Ajouz S, Comby M, Lopez-Ferber M, Graillot B,
pathogenic fungus present in this region of Algeria Siegwart M, Nicot PC (2015) Is the efficacy of biological
(Adjebli et al. 2015). control against plant diseases likely to be more durable than
that of chemical pesticides? Front Plant Sci 6:1–14
Acknowledgements A grant for studies was provided by the Berge O, Guinebretiere MH, Achouak W, Normand P, Heulin T
Ministry of Higher Education and Scientific Research of Algeria (2002) Paenibacillus graminis sp. nov. and Paenibacillus
for Yousra Bouaoud. We are grateful to Dr. Philippe Nicot for odorifer sp. nov., isolated from plant roots, soil and food.
extremely useful discussions throughout this research project Int J Syst Evol Microbiol 52:607–616
and to Caroline Guilbaud, Charlotte Chandeysson, Magali Bizuye A, Moges F, Andualem B (2013) Isolation and screening
Eygrier and Abdelwahab Rai, for their valuable help in the of antibiotic producing actinomycetes from soils in Gondar
experiments. Research on plant biology depends heavily on the town, North West Ethiopia. Asian Pac J Trop Dis
cultivation of plants under experimental conditions that are 3:375–381
controlled, monitored and repeatable. For ensuring a perfect Dean R, van Kan JAL, Pretorius ZA, Hammond-Kosack KE, Di
plant production, the contribution of the Plant Production and Pietro A, Spanu PD, Rudd JJ, Dickman M, Kahmann R,
Environmental Management Service of the Plant Pathology Ellis J, Foster GD (2012) The top 10 fungal pathogens in
Research Unit is greatly appreciated. Finally, the authors greatly molecular plant pathology. Mol Plant Pathol 13:414–430
thank Dr. Christelle Lacroix for the improvement of the English Decognet V, Bardin M, Trottin-Caudal Y, Nicot PC (2009)
language and for the constructive comments on an earlier Rapid change in the genetic diversity of Botrytis cinerea
version of this manuscript. populations after the introduction of strains in a tomato
glasshouse. Phytopathology 99:185–193
Decognet V, Ravetti F, Martin C, Nicot PC (2010) Improved
leaf pruning reduces development of stem cankers caused
References by grey mould in greenhouse tomatoes. Agron Sustain Dev
30:465–472
Abro MA, Lecompte F, Bryone F, Nicot PC (2013) Nitrogen Egamberdieva D, Kucharova Z, Davranov K, Berg G, Makarova
fertilization of the host plant influences production and N, Azarova T, Chebotar V, Tikhonovich I, Kamilova F,
pathogenicity of Botrytis cinerea secondary inoculum. Validov SZ, Lugtenberg B (2011) Bacteria able to control
Phytopathology 103:261–267 foot and root rot and to promote growth of cucumber in
Abro MA, Lecompte F, Bardin M, Nicot PC (2014) Nitrogen salinated soils. Biol Fertil Soils 47:197–205
fertilization impacts biocontrol of tomato gray mold. Elad Y (2016) Cultural and integrated control of Botrytis spp. In:
Agron Sustain Dev 34:641–648 Fillinger S, Elad Y (eds) Botrytis the fungus, the pathogen
Adjebli A (2014) Caractérisation phénotypique et génotypique and its management in agricultural systems. Springer,
des souches de Botrytis cinerea, agent causal de la pour- Cham, pp 149–164
riture grise sur tomate. Ph.D. Thesis, University A Mira Fan ZY, Miao CP, Qiao XG, Zheng YK, Chen HH, Chen YW,
Bejaia Xu LH, Zhao LX, Guan HL (2015) Diversity, distribution,
Adjebli A, Leyronas C, Aissat K, Nicot PC (2015) Comparison and antagonistic activities of rhizobacteria of Panax
of Botrytis cinerea populations collected from tomato notoginseng. J Ginseng Res 40:97–104

123
A multi-criteria approach for the selection of efficient biocontrol agents

Fillinger S, Walker AS (2016) Chemical control and resistance Ramirez-Bahena MH, Cuesta MJ, Florex-Felix JD, Mulas R,
management of Botrytis diseases. In: Fillinger S, Elad Y Rivas R, Castro-Pinto J, Branas J, Mulas D, Gonzalez-
(eds) Botrytis the fungus, the pathogen and its management Andres F, Velazques E, Peix A (2014) Pseudomonas hel-
in agricultural systems. Springer, Cham, pp 189–216 manticensis sp. nov., isolated from forest soil. Inter J Syst
Jamalizadeh M, Etebarian HR, Alizadeh A, Aminiam H (2008) Evol Microbiol 64:2338–2345
Biological control of gray mold on apple fruits by Bacillus Sang MK, Kim EN, Han GD, Kwack MS, Jeun YC, Kim KD
licheniformis (EN74-1). Phytoparasitica 36:23–29 (2014) Priming-mediated systemic resistance in cucumber
Jarvis WR (1992) Managing diseases in greenhouse crops. induced by Pseudomonas azotoformans GC-B19 and
American Phytopathological Society, St Paul Paenibacillus elgii MM-B22 against Colletotrichum
Köhl J, Molhoek WL, van der Plas CH, Fokkema NJ (1995) orbiculare. Phytopathology 104:834–842
Suppression of sporulation of Botrytis spp. as a valid bio- Sundh I, Hökeberg M, Levenfors JJ, Nilsson AI (2011) Safety
control strategy. Eur J Plant Pathol 101:251–259 assessment of biocontrol and plant growth-promoting
Köhl J, Postma J, Nicot P, Ruocco M, Blum B (2011) Stepwise pseudomonads useful in crop production. Ann Appl Biol
screening of microorganisms for commercial use in bio- 159:291–301
logical control of plant pathogenic fungi and bacteria. Biol
Control 57:1–12
Morris CE, Sands DC, Vinatzer BA, Glaux C, Guilbaud C, Yousra Bouaoud is a plant pathologist and she is currently
Buffiere A, Yan S, Dominguez H, Thompson BM (2008) working between Algeria and France on her PhD thesis in the
The life history of the plant pathogen Pseudomonas syr- field of biocontrol. This research is part of her project devoted
ingae is linked to the water cycle. ISME J 2:321–334 to the development of biocontrol agents against fungal diseases
Morris CE, Sands DC, Vanneste JL, Montarry J, Oakley B, of tomato.
Guilbaud C, Glaux C (2010) Inferring the evolutionary
history of the plant pathogen Pseudomonas syringae from Claire Troulet is involved in the evaluation of efficacy of
its biogeography in headwaters of rivers in North America, biocontrol agents against fungal plant pathogens.
Europe, and New Zealand. mBio 1:1–11
Nicot PC, Mermier M, Vaissiere BE, Lagier J (1996) Differ- Abdelhamid Foughalia is currently a PhD student and
ential spore production by Botrytis cinerea on agar medium researcher at the Centre for Scientific and Technical Research
and plant tissue under near-ultraviolet light-absorbing on Arid Regions (Biskra. Algeria). He was involved in the
polyethylene film. Plant Dis 80:555–558 sampling and in the primary screening of microbials.
Nicot PC, Stewart A, Bardin M, Elad Y (2016) Biological
control and biopesticides suppression of Botrytis incited Odile Berge is a bacterial taxonomy expert involved in
diseases. In: Fillinger S, Elad Y (eds) Botrytis the fungus, environmental bacterial diversity studies.
the pathogen and its management in agricultural systems.
Springer, Cham, pp 165–187
Kamel Aissat is conducting his research in the area of
O’Neill TM, Shtienberg D, Elad Y (1997) Effect of some host
epidemiology, biocontrol and integrated diseases control of
and microclimate factors on infection of tomato stems by
vegetable greenhouse crops. He is co-supervisor of Y. Bouaoud
Botrytis cinerea. Plant Dis 81:36–40
PhD.
Petatan-Sagahon I, Anducho-Reyes MA, Silva-Rojas HV,
Arena-Cuenca A, Tellez-Jurado A, Cardenas-Alvarez IO,
Mercado-Flores Y (2011) Isolation of bacteria with anti- Marc Bardin is conducting his research on biocontrol against
fungal activity against the phytopathogenic fungi Steno- plant diseases with a main interest in the evaluation of efficacy
carpella maydis and Stenocarpella macrospora. Int Mol and durability of biocontrol. This study was realized under his
Sci 12:5522–5537 supervision.
Pliego C, Ramos C, De Vicente A, Cazorla FM (2011)
Screening for candidate bacterial biocontrol agents against
soilborne fungal plant pathogens. Plant Soil 340:505–520

123