Formulation development of the deleterious rhizobacterium Pseudomonas trivialis X33d for biocontrol of brome (Bromus diandrus) in durum wheat

Authors


Correspondence

Thouraya Souissi, Institut National Agronomique de Tunisie, 1082 Tunis-Mahrajène, Tunisia. E-mail: tsouissi@alinto.com

Abstract

Aim

To develop an appropriate formulation of the deleterious rhizobacterium Pseudomonas trivialis X33d and to evaluate its effectiveness to reduce brome growth.

Methods and Results

Two formulations of Ps. trivialis X33d, a semolina-kaolin granular formulation (Pesta) and talc-kaolin powder, were prepared and their effectiveness in reducing brome growth was evaluated. Both brome suppression and cell viability of X33d were higher in Pesta granular formulation than in talc-kaolin powder one. The impact of storage temperature and the addition of adjuvants (sucrose and oil) to the granular formulation of X33d were assessed in order to improve the shelf life of the formulation. The longest viability was found in formulated product supplemented with adjuvants and stored at 4°C. The effect of Pesta granules supplemented with adjuvants and stored for 6 months at 4°C on brome and wheat growth under controlled and greenhouse conditions was evaluated. The X33d formulation in Pesta increased the growth of wheat and reduced brome growth.

Conclusion

Our results indicate that Ps. trivialis X33d formulated in Pesta has potential as a bioherbicide to control brome.

Significance and Impact of the Study

Because of the impracticality of applying bacterial cell suspension on a large scale, the use of Pesta granules of X33d against brome could help in achieving a sustainable agriculture application of a bioherbicide.

Introduction

In Tunisia, great brome (Bromus diandrus Roth., syn.Bromus rigidus Roth. subsp. gussonii Parl.) is widely distributed in cereal crops resulting in yield losses that can reach up to 80% in heavily infested wheat-growing areas (Souissi et al. 2000, 2001). Control methods commonly used to suppress brome growth in wheat crops are essentially chemical and cultural. However, these methods are prohibitively expensive, and there are no selective herbicides for the control of brome in wheat (Mazzola et al. 1995; Souissi et al. 2000). In addition, excessive use of chemical herbicides has resulted in the development of herbicide resistance in many weed biotypes (Heap 2012).

In order to overcome these limitations, efforts have been made to develop alternative and more effective measures to manage weeds. Biological control of weeds with living micro-organisms was reported to be a good alternative to chemical treatment. Deleterious rhizobacteria (DRB) are among the micro-organisms that have been reported to have a potential as biocontrol agents in controlling weeds (Elliott and Lynch 1985; Kremer 1987; Schippers et al. 1987; Kremer et al. 1990; Kennedy et al. 1991; Kremer and Kennedy 1996; Kremer 2000; Kennedy et al. 2001; Flores-Vargas and O'Hara 2006; Li and Kremer 2006; Kennedy and Stubbs 2007; Banowetz et al. 2008; Mejri et al. 2010). However, successful application of live bacteria in a field setting depends upon many factors including environmental conditions and soil survival. Therefore, there is a need to formulate these biocontrol agents to enhance their field potential and facilitate their storage and application. Many weed biological control agents have been formulated into liquid, solid and powder substrates (Green et al. 1998) because of the low bacterial survival in liquid inoculants (Singleton et al. 2002; Tittabutr et al. 2007; Albareda et al. 2008) and impractical use of cell suspensions for large-scale application due to the difficult handling, transport and storage of the inoculum (Rabindran and Vidhyasekaran 1996; Vidhyasekaran et al. 1997a). A formulated bioherbicide can be defined as a mixture of the active ingredient (the biological agent) within a carrier or solvent that delivers the active ingredient to the target weed, and the adjuvants that improve the survival and effectiveness of the product in adverse environmental conditions (Boyette et al. 1991; Hynes and Boyetchko 2006; Chutia et al. 2007; Ash 2010). Among the different possible formulations, the dry solid or powder formulations provide several advantages. Bacterial cells immobilized in dry carrier are protected from the external environmental factors and their survival and efficacy are preserved in adverse environmental conditions (Boyette et al. 1991; Shabana et al. 2003; Kinay and Yildiz 2008). Dry carriers also allow efficient and easy delivering of bacteria to the target weed (Sabaratnam and Traiquair 2002). For instance, the wheat-gluten matrix known as Pesta has been used to formulate granular biocontrol agents. This matrix is adaptable to many different micro-organisms and ingredients (Daigle et al. 1997) and is nontoxic, cost-effective and easy to store and use (Elzein et al. 2004). Pesta formulations have been widely used to deliver mycoherbicides such as Colletotrichum truncatum against hemp sesbania (Sesbania exaltata) (Connick et al. 1991, 1996) and Fusarium oxysporum against sunflower broomrape (Orobanche cumana) (Shabana et al. 2003) and Striga spp. (Elzein et al. 2004). Pesta formulations for bacteria with bioherbicide activities have also been used with Pseudomonas fluorescens BRG100 against green foxtail (Setaria viridis) (Daigle et al. 2002), P. fluorescens strain G2-11 against velvetleaf (Abutilon theophrasti) (Zdor et al. 2005) and P. fluorescens LS102 and LS174 against leafy spurge (Euphorbia esula) (Brinkman et al. 1999). Talc powder dry formulations have been considered as a simple and cost- and time-effective technique. It has been mainly used to formulate bacteria (Hofte et al. 1991; Vidhyasekaran et al. 1997a) for large-scale field applications in different crops for the management of soilborne plant pathogens (Rabindran and Vidhyasekaran 1996; Vidhyasekaran et al. 1997a,b).

We have previously described the isolation, identification and physiological characterization of the deleterious rhizobacterium, Pseudomonas trivialis X33d, a promising biocontrol agent against great brome in wheat (Mejri et al. 2010). The main objectives of this work were to develop an appropriate formulation of the strain X33d and to assess its efficacy in inhibiting great brome growth under controlled and greenhouse conditions. In addition, the storage conditions (i.e. temperature and addition of adjuvants) required in order to improve the survival of X33d in the formulation were standardized.

Materials and methods

Bacterial inoculant

The rhizobacterium Pseudomonas trivialis strain X33d used in this study was previously isolated from the rhizosphere of durum wheat (Triticum durum) growing in Northern Tunisia. Pseudomonas trivialis X33d is a gram-negative, rod-shaped fluorescent bacterium. The bacterial strain is cytochrome oxidase positive and is able to oxidize glucose and to synthesize a high amount of IAA.

A single colony of X33d from a freshly growing culture was inoculated in 200 ml of ¼ strength KB broth (KB, Dhingra and Sinclair 1995) and incubated at 25°C on a rotary shaker at 100 rev min−1 for 48 h. Cells were pelleted by centrifugation (4000 g, 10 min) and resuspended in 0·1 mol l−1 MgSO4·7H2O. Bacterial density of the suspension was determined by serial dilution plating on ¼ strength KB agar (Camper 1986).

Preparation of different formulations of Pseudomonas trivialis X33d

Bacterial inoculum was prepared in two different formulations, namely a semolina-kaolin-based granular formulation known as ‘Pesta’ granular formulation and talc-kaolin powder formulation. ‘Pesta’ granular formulation was prepared as described by Daigle et al. (2002) with slight modifications. Semolina and kaolin were sterilized separately for 15 min at 120°C. Upon cooling, 800 g of semolina flour, 200 g of kaolin and 400 ml of the bacterial suspension containing 1010 CFU ml−1 were mixed with gloved hands until forming cohesive dough. Then, the dough was kneaded, folded repeatedly and flattened several times using a rolling pin in order to obtain sheets having 1 mm thickness. The dough sheet was laid on a screen to air dry. The dried dough sheets were then broken into pieces, ground with a mortar and pestle and sieved through a 2-mm screen in order to obtain a uniform size. All granules were packed in small plastic bags and stored either at 4°C or at room temperature (22 ± 2°C).

The talc-kaolin-based powder formulation was prepared as described by Vidhyasekaran and Muthuamilan (1995). One kg of talc (Mg3Si4O10(OH)2) was mixed with 10 g of carboxymethylcellulose (CMC). To adjust the pH of the mixture to neutral, 10 g of sample was suspended in 100 ml of deionized water for 10 min. The pH of the supernatant is then measured, and based on its value, the pH of the mixture is adjusted by adding calcium carbonate (CaCO3). Then, the mixture was autoclaved for 30 min on each of two consecutive days. Four hundred millilitres of 1010 CFU ml−1 bacterial suspension was added to the mixture and mixed well with gloved hands under aseptic conditions. The product was dried overnight under sterile conditions, packed in two plastic bags and stored separately at 4°C and room temperature.

Evaluation of the phytotoxic activity of formulated Pseudomonas trivialis X33d under controlled conditions

A pot experiment was performed to select the most efficient formulation of Ps. trivialis X33d in reducing the growth of great brome. Seeds of great brome were surface-sterilized in 3% sodium hypochlorite for 2 min and rinsed five times for 4 min in sterile distilled water. The seeds were pregerminated on moist sterile filter paper and incubated in the dark at 22°C for 48–72 h.

Two pregerminated brome seeds were placed in 6-cm-diameter pot filled with 200 g of sterilized mixture of field soil (chalky-clay, 1·4% of organic matter, pH 7·8), sand and peat (1 : 1 : 1 v/v/v). Then, 2 g of each fresh formulation or 10 ml of 2-day-old bacterial suspension containing 108 CFU ml−1 was incorporated to the soil in each pot. Just after inoculation, the viability of X33d in both ‘Pesta’ granular and talc powder formulations stored at room temperature was checked weekly for 2 months. Viable bacteria were counted by the dilution plate method on King's medium B. Three Petri plates for each dilution and five replications for each carrier sample were maintained.

Control pots did not receive any bacterial treatment. Plants were grown in a growth chamber at 22°C, 70% relative humidity and a light intensity of 35 000 Lux with an 18/6h light/dark photoperiod and watered to saturation when required. The experimental design was completely randomized with five replications. The experiment was conducted in duplicate, and the results shown are from the pooled data from both the experiments.

Three weeks after inoculation, brome plants from each pot were harvested. Roots were separated from shoots and washed in order to eliminate the adhering soil. Then, roots and shoots were dried overnight at 70°C, and dry weights of each plant were recorded.

Effect of the temperature on the shelf life of the ‘Pesta’ granular formulation of Pseudomonas trivialis X33d

The viability of the formulated granules of X33d stored at 4°C and room temperature was followed monthly over 6 months. Viable bacteria were counted by dissolving 1 g of granular formulation in 9 ml of MgSO4 buffer, shaking for 10 min and plating on ¼ strength KB agar after serial dilution. The colony forming units (CFU g−1) were counted after 48 h of incubation at 25°C. Five replicate samples were used, and three Petri plates for each dilution were maintained. The final values (log CFU g−1) of viable and culturable bacteria were the average of five readings.

Effect of the addition of the adjuvants on the survival of ‘Pesta’ granular formulated Pseudomonas trivialis X33d stored at 4°C

To evaluate whether adjuvants (sucrose and unrefined corn oil) could increase the viability of ‘Pesta’ granular formulated Ps. trivialis X33d during storage, an experiment with five replicates was performed by adding oil and sucrose to the standard granulated formulation. The semolina-kaolin-based granular formulation containing sucrose and oil as adjuvants was prepared by supplementing 200 g sucrose and/or 100 ml of unrefined corn oil to 600 g semolina, 200 g kaolin and 400 ml of bacterial suspension. The formulation was prepared following the procedure described above. The granules were then packed in sterile plastic bags and stored under refrigeration at 4°C. The viability of granular formulated Ps. trivialis X33d without adjuvants (-S-O), with oil alone (-S+O), with sucrose alone (+S-O) and with the two adjuvants (+S+O) was assessed monthly for 6 months. Five replicates for each treatment and three Petri plates for each dilution were maintained. The final values (log CFU g−1) of viable and culturable bacteria were the average of five readings.

Evaluation of the stored ‘Pesta’ granular formulation of Pseudomonas trivialis X33d supplemented with adjuvants on great brome and durum wheat under controlled conditions

‘Pesta’ granular formulation supplemented with sucrose and oil and stored for 6 months at 4°C was tested in pots under controlled conditions to evaluate its effect on the growth of great brome and durum wheat. Two hundred grams of the sterile substrate consisting of field soil, sand and peat (1 : 1 : 1 v/v/v) was placed in each pot. The experiment was conducted in a completely randomized design with five replicates according to the following plant treatments: noninoculated great brome or durum wheat (Br, Wh), great brome or durum wheat inoculated with the granular formulation without adjuvants (Br-S-O, Wh-S-O) and great brome or durum wheat inoculated with the granular formulation with adjuvants (Br+S+O, Wh+S+O). The granular formulation of Ps. trivialis X33d was applied on great brome and durum wheat as previously described. The experiment was conducted in a growth chamber under controlled conditions (22°C, 70% relative humidity, light intensity of 35 000 Lux with 18/6-h light/dark photoperiod). Three weeks after inoculation, shoot length and number of leaves of the two plants species in each pot were measured. Root and shoot dry weights were measured by placing the roots and shoots separately into small, preweighed brown paper bags, drying them at 70°C for 24 h and then weighing them.

Evaluation of the phytotoxic activity of the stored ‘Pesta’ granular formulated Pseudomonas trivialis X33d supplemented with adjuvants under greenhouse conditions

The efficacy of the ‘Pesta’ granules of the strain X33d supplemented with adjuvants and stored for 6 months at 4°C was evaluated under greenhouse conditions. The experiment was performed in buckets filled with 12 kg of nonsterilized mixture of soil, sand and peat (1 : 1 : 1, v/v/v). Five seeds of great brome or durum wheat were placed in each bucket. Ten grams of the stored granules was incorporated into 1 kg of substrate. Granules containing no bacteria were used as controls. The experimental design was a completely randomized bloc with five replications.

Plants were harvested at maturity, and the root and shoot dry weights as well as the total number of seeds per plant of great brome and durum wheat were determined. Data regarding yield parameters of wheat such as the weight of thousand grains and grain yield were also determined.

Statistical analysis

Data were submitted to analysis of variance and to Fisher's least significant difference test (≤ 0·05) using a Statview statistics package.

Results

Selection of the most efficient formulation of Pseudomonas trivialis X33d

In order to assess the conservation of bioherbicide activity of Ps. trivialis X33d following the formulation procedure itself, and select the bacterial treatment with the highest suppression activity against brome, the biomass of brome plants inoculated with a bacterial cell suspension or with the two formulations was measured. The bacterial suspension, by itself, as well as the two formulations was effective in suppressing brome growth (Fig. 1). The bioherbicide activity shown by the cell suspension (root – 39%, shoot – 18·5%) and the ‘Pesta’ granular formulation of Ps. trivialis X33d were comparable (root – 33% and shoot – 15%) (Fig. 1). Brome dry biomass was also reduced in plants treated with the talc powder formulated strain X33d by 29·5% (root) and 8·5% (shoot), respectively (Fig. 1), compared to uninoculated plants. The efficacy of the two formulations in inhibiting brome growth varied according to the carrier used (Fig. 1).

Figure 1.

Efficacy of different formulations of Pseudomonas trivialis X33d on root (a) and shoot (b) dry weight of great brome in pot under controlled conditions. Values are the means of combined data of two experiments. Bars represent ±SD; means designated with the same letter are not significantly different (P = 0·05) according to Fisher's least significant difference test.

Moreover, the viability of X33d was higher in ‘Pesta’ granules than in talc powder (Table 1). Two months after inoculation, the cell number decline was greater in the talc powder formulation than in granular one and dropped about 1·5 log units from log10 9·93 CFU g−1 to log10 8·41 CFU g−1. However, the viability of X33d formulated in ‘Pesta’ granules decreased more slowly from log10 9·95 CFU g−1 to log10 8·6 CFU g−1. Based on these results, ‘Pesta’ granular formulation was used for further experiments during this study.

Table 1. Viability of Pseudomonas trivialis X33d in two different formulations at room temperature
TreatmentWeeks after inoculation
0 1234
  1. Means designated with different letter in each column are statistically significant (= 0·05) according to Fisher's least significance difference test. Values are means (log10 CFU g−1 of formulation) ± standard error.

Pesta9·95 ± 0·00a9·68 ± 0·003a9·63 ± 0·00a9·40 ± 0·01a9·34 ± 0·00a
Talc9·93 ± 0·00b9·65 ± 0·00b9·59 ± 0·00b9·34 ± 0·00b9·00 ± 0·00b
 5678 
Pesta8·90 ± 0·00a8·80 ± 0·01a8·75 ± 0·00a8·60 ± 0·00a 
Talc8·82 ± 0·00b8·75 ± 0·00b8·62 ± 0·01b8·41 ± 0·02b 

Survival of ‘Pesta’ granular formulated Pseudomonas trivialis X33d at different temperatures under storage

Survival of stored ‘Pesta’ granular formulated Ps. trivialis X33d depended on the temperature. The viability of Ps. trivialis X33d in ‘Pesta’ formulation was higher at 4°C than at room temperature (Fig. 2). The initial population of Ps. trivialis X33d before storage was log10 9·95 CFU g−1. Thirty days after storage, the viability of the strain X33d in ‘Pesta’ granules at the room temperature and at 4°C was similar (log10 9·34 and 9·36 CFU g−1, respectively). Thereafter, the cell number decline was greater at room temperature than that observed at 4°C (≤ 0·05). After 6 months of storage at room temperature, the viability of ‘Pesta’ granules dropped 2·99 log units to log10 6·96 CFU g−1. However, the loss of the viability of the granules at 4°C was much lower, about 1·5 log units reaching log10 8·46 CFU g−1 at 180 days (Fig. 2).

Figure 2.

Viability of Pesta granular formulated Pseudomonas trivialis X33d under storage at different temperatures. (▲) 4°C; (●) Room temp.

Effect of the addition of the adjuvants on survival of ‘Pesta’ granular formulated Pseudomonas trivialis X33d during storage at 4°C

Survival of Ps. trivialis X33d formulated in ‘Pesta’ granules during the storage at 4°C varied following the addition of the two adjuvants, alone or in combination. Enrichment of formulated ‘Pesta’ granules of Ps. trivialis X33d by both sucrose and oil leads to the highest amount of viable cells (Table 2).

Table 2. Effect of the adjuvants on the viability of Pesta granular formulated Pseudomonas trivialis X33d under storage for 6 months at 4°C
TreatmentDays after storage
0306090120150180
  1. -S-O, without sucrose and oil; -S+O, with oil alone; +S-O, with sucrose alone; +S+O, with sucrose and oil together.

  2. Means designated with the same letter in each column are not statistically significant (= 0·05) according to Fisher's least significance difference test. Values are means (log10 CFU g−1 of formulation) ± standard error.

-S-O9·63 ± 0·01a9·29 ± 0·02a8·90 ± 0·00a8·84 ± 0·00a8·49 ± 0·05a8·11 ± 0·04a8·02 ± 0·01a
-S+O9·63 ± 0·02a9·55 ± 0·03b9·35 ± 0·03b9·07 ± 0·05b8·94 ± 0·01b8·82 ± 0·02b8·60 ± 0·00b
+S-O9·63 ± 0·01a9·57 ± 0·02b9·37 ± 0·04b9·09 ± 0·05b8·96 ± 0·01b8·878 ± 0·01b8·62 ± 0·01b
+S+O9·65 ± 0·00a9·60 ± 0·01b9·54 ± 0·04c9·36 ± 0·03c9·17 ± 0·09c8·99 ± 0·01c8·90 ± 0·01c

Evaluation of the stored Pseudomonas trivialis X33d formulated in ‘Pesta’ granules on great brome and durum wheat under controlled conditions

The biocontrol performance of the ‘Pesta’ granular formulation of Ps. trivialis X33d with and without adjuvants after 6 months of storage at 4°C on the growth of great brome and durum wheat was assessed under growth chamber conditions. The results showed that shoot length and leaf number of both great brome and durum wheat were unaffected after the treatment with ‘Pesta’ granules of X33d supplemented with adjuvants (+S+O) or not (-S-O) (data not shown). On the other hand, both formulations with and without adjuvants showed a significant impact on great brome growth compared to uninoculated plants. Higher reductions (root – 28·5%, shoot – 13%) were observed on brome plants inoculated with ‘Pesta’ granules supplemented with adjuvants than on plants inoculated with granules without adjuvants (root – 20%, shoot – 7·5%) (Fig. 3).

Figure 3.

Root (a) and Shoot (b) dry weights of great brome and durum wheat plants inoculated with Pesta granules of Pseudomonas trivialis X33d with or without adjuvants. Bars represent ±SD; means designated with the same letter are not significantly different (= 0·05) according to Fisher's least significant difference test.

The growth of durum wheat was increased by the ‘Pesta’ granular formulation of X33d with or without adjuvants (Fig. 3). Enhancement of root (+18%) and shoot (+15%) dry weight was measured in plants inoculated with the formulation +S+O and, to a lesser extent (+10·5% and +10% for root and shoot dry weight, respectively), in plants treated with-S-O formulation.

Evaluation of the phytotoxic activity of the stored ‘Pesta’ granular formulated Pseudomonas trivialis X33d supplemented with adjuvants under greenhouse conditions

The application of ‘Pesta’ granules of the strain X33d supplemented with adjuvants (+S+O) after 6 months of storage at 4°C affected the growth of great brome. Root and shoot dry weights of great brome were decreased by 22·7% and 11%, respectively, compared to uninoculated plants. Moreover, the total number of seeds produced by great brome was reduced by the formulation (- 18% compared to controls) (Table 3).

Table 3. Effect of stored Pesta granular formulation of Pseudomonas trivialis X33d on the growth of great brome and durum wheat under greenhouse conditions
 Root dry weight (g)Shoot dry weight (g)Total number of seedsGrain yield (g per pot)1000-grain weight (g)
  1. ND, not determined.

  2. a

    Indicates significant reduction or increase compared to noninoculated plants according to Fisher's least significant test (P ≤ 0·05). Values are mean of five replicates ± standard error.

Brome
Control4·50 ± 0·366·20 ± 0·16151 ± 2·92NDND
Pesta granules3·40 ± 0·06a5·50 ± 0·20a124 ± 1·07aNDND
Wheat
Control6·75 ± 0·156·00 ± 0·270ND11·43 ± 0·2329·25 ± 0·60
Pesta granules7·75 ± 0·29a6·75 ± 0·15aND13·12 ± 0·08a33·01 ± 0·89a

On the contrary, the ‘Pesta’ granules of X33d formulated with adjuvants (+S+O) increased the plant growth of durum wheat: both the root (+14·8%) and shoot (+12·5%) dry weights increased compared to uninoculated controls (Table 3). Significant increase in the grain yield (12%) and 1000-grain weight (13%) were recorded in wheat plants inoculated with ‘Pesta’ granules of X33d compared to uninoculated plants (Table 3).

Discussion

Weed biological control by specific micro-organisms is an accepted strategy for weed management (El-Sayed 2005; Caressa et al. 2010). The deleterious rhizobacteria (DRB) belonging to the Pseudomonas group are known to be promising candidates for weed biological control (Kennedy et al. 1991; Kremer and Kennedy 1996; Flores-Vargas and O'Hara 2006). Cell suspension of these DRB has been found to be an effective bioherbicide. However, the use of bacterial cell suspensions in large field scale is hard to manage due to difficulty in handling and storage. Therefore, development of robust formulations, ensuring simple handling, long shelf life and high cost-efficiency, can help to overcome such difficulties. The selection of the appropriate carrier, enabling the dispersion of the biological material to the target plant, is essential for the successful development of a formulation (Nakkeran et al. 2005).

Pseudomonas trivialis X33d is able to reduce the growth (in the terms of dry weight) of the weed Bromus diandrus while increasing the development of durum wheat. Physiological traits possibly involved in this dual behaviour have been discussed in the study by Mejri et al. (2010). In this work, the impact of cell suspension of Ps. trivialis X33d on great brome growth was assessed and compared with the effect induced by ‘Pesta’ granules and talc powder formulations of the bacterial strain. Our results showed that both the bacterial formulations and the cell suspension reduced great brome growth compared to uninoculated controls. The bioherbicide activity of cell suspension and ‘Pesta’ granular formulation of X33d was similar, which may suggest that the phytotoxic activity of X33d was not affected during the formulation process. In addition, our results showed that the cell viability of X33d was higher in ‘Pesta’ granules than that in talc powder formulation. This result shows that cell viability of the biological agent may be affected by the type of the carrier in the formulation.

The active ingredient of a bioherbicide is sensitive to many variables throughout formulation (Connick et al. 1996). Among them, temperature is one of the main factors contributing to the quality of the formulation, which is reflected in the efficacy and the long shelf life of the bioherbicide. Hence, the shelf life of the strain X33d formulated as ‘Pesta’ granules was evaluated during storage at two different temperatures (room temperature and 4°C). According to the literature, ‘Pesta’ granules were considered to be nonviable if the density of bacterial cells in the granules is less than 1 × 105 CFU g−1 (Daigle et al. 2002). Our results demonstrated that the concentration of bacterial cells in ‘Pesta’ granular formulated Ps. trivialis X33d stored for 6 months was higher than 105 CFU g−1, irrespective of the storage temperature. Moreover, bacterial viability in ‘Pesta’ granules stored at 4°C was higher than that measured in granules stored at room temperature. In fact, both cell division and metabolic rate in bacteria are slowed down by storage at low temperatures (4–10°C). In this condition, the depletion of nutrients is reduced and the accumulation of toxic metabolites and the loss of moisture in the carrier are prevented, therefore favouring the long-term storage of the bacterial inoculants (Van Shrevan 1970; Kirsop and Doyle 1991; Trivedi et al. 2005). Consequently, low temperatures of storage have been successfully used for many formulated fungal and bacterial biological control agents (Kirsop and Doyle 1991). For instance, excellent recovery of Colletotrichum truncatem, Alternaria cassiae and A. crassa in ‘Pesta’ granules was observed after 18 months of storage at 4°C (Connick et al. 1991). The survival of Colletotrichum dematium FGCC# 20 in ‘Pesta’ granules after 1 year was found to be higher at 4°C than at room temperature (Singh and Pandey 2010).

The viability of living organisms in granular formulations during storage may be affected by the nutritional amendments added to the formulation (Shabana et al. 2003). It has been suggested that the Pesta production process is amenable to incorporation of many solid and liquid additives, which may reduce the cost and alter certain properties of the final formulation (Shabana et al. 2003). The survival of Ps. trivialis X33d formulated in ‘Pesta’ granules amended with sucrose (+S-O-) or oil (-S+O) fully overlapped and were higher than in granules without adjuvant (-S-O). Moreover, the addition of both sucrose and oil (+S+O) to ‘Pesta’ granules induced a synergistic effect on the density of the strain X33d in ‘Pesta’ granules compared to all the other adjuvants, probably due to the provision of additional nutrients. Sucrose is known to extend bacterial survival throughout storage because it has a putative role as membrane stabilizer during the drying (Caesar and Burr 1991; Leslie et al. 1995; Connick et al. 1996), and oil has been shown to enhance the survival of fungi and nematodes in alginate formulations (Quimby et al. 1994; Caesar-Tonthat et al. 1995). Our results partially corroborate the finding of Zidack and Quimby (2002) who reported that the long-term survival of Pseudomonas syringae pv. tabaci and Pseudomonas syringae pv. tagetis in Stabileze formula differs according to the adjuvant added. While the long-term survival of P.s. pv. tabaci was enhanced by oil alone and sucrose and oil in combination, survival of P.s. pv. tagetis was not enhanced by oil, and oil in the formulation reduced the beneficial effects of sucrose. Moreover, the addition of sucrose to Pesta granules partially counteracted the detrimental effect of high water activity on the shelf life of Colletotrichum truncatum (Connick et al. 1996). Finally, granular ‘Pesta’ of Fusarium oxysporum sp. orthoceras supplemented with yeast extract and sucrose showed improved mycoherbicidal efficacy, stability and shelf life over time (Shabana et al. 2003).

Taking into account the results obtained, we decided to assess, under controlled conditions, the efficacy of Pesta granules of Ps. trivialis X33d with or without the two adjuvants and stored for 6 months at 4°C in suppressing great brome growth, without showing any negative effect on durum wheat. The results obtained showed that brome development was reduced by ‘Pesta’ granules of the strain X33d, this inhibition being stronger in ‘Pesta’ granules supplemented with sucrose and oil. On the contrary, wheat biomass was increased in plants inoculated with the ‘Pesta’ granules of the bacterial strain, this promotion being higher in the presence of both adjuvants. The stronger effects on plant growth (suppression of brome and stimulation of wheat) recorded in plants treated with ‘Pesta’ granules of Ps. trivialis X33d added with sucrose and oil are consistent with the observed high survival, and rhizosphere competence, of the bacterial strain in granules supplemented with both the adjuvant and stored at 4°C.

Because the bioherbicide effects and plant growth promotion could be affected by environmental conditions, a further experiment was performed by growing brome and wheat plants in the greenhouse inoculated or not with ‘Pesta’ granules of X33d with adjutants. The results obtained showed that the bioherbicide activity of this bacterial formulation on brome growth and yield and the plant growth–promoting effect on wheat growth and yield were maintained also in uncontrolled environment conditions. Therefore, the efficacy of the ‘Pesta’ granular formulation with adjuvants of the strain X33d was unaffected by the 6-month storage at 4°C. However, the reduction of great brome biomass, as well as the stimulation of wheat biomass, was more evident under controlled conditions than in natural ones. This is consistent with the impact of environmental conditions on the survival, efficiency and rhizosphere competence of micro-organisms behaving as DRB or plant growth–promoting bacteria (PGPB) (Kremer 2005).

This work has demonstrated the successful use of the deleterious rhizobacterium Ps. trivialis X33d formulated in ‘Pesta’ granules as a bioherbicide against great brome. The use and application of such bioformulations in the field can result in the reduction of application of harmful chemicals, protect the environment and biological resources and be an important component of integrated pest management in sustainable agriculture. This investigation is only the first step in order to obtain a suitable formulation for application in the field after storage. Other research needs to be addressed before large-scale applications. In particular, further work will be performed to (i) assess the survival and the efficacy of ‘Pesta’ formulation of X33d in soils with different chemical and physical characteristics and (ii) analyse the root colonization pattern of the strain X33d formulated in ‘Pesta’ granules or as free cells, in brome and wheat.

Acknowledgements

The authors would like to thank Ahlem Bahrouni, a technician at the weed science laboratory at the National Institute of Agronomy of Tunisia, for her valuable help to carry out the experiments. We thank Professor Dana Berner (USDA-ARS, Maryland) for English proofreading of the manuscript.

Ancillary