Mycorrhiza Helper Bacteria stimulate ectomycorrhizal symbiosis of Acacia holosericea with Pisolithus alba


Author for correspondence: Robin Duponnois Tel: +221 849 33 24 Fax: +221 832 16 75 Email:


  •  The influence of two fluorescent pseudomonads strains (HR13 and HR26) on the ectomycorrhizal symbiosis between Pisolithus alba and Acacia holosericea is reported here.
  •  We measured ectomycorrhizal establishment, fungal growth in the soil (by HPLC) and soil microbial biomass (using the fumigation–extraction method) in treatments with or without pseudomonads.
  •  Bacteria inoculated with the fungal symbiont stimulated ectomycorrhizal formation and shoot or root biomass. Only HR13 significantly increased fungal biomass in the soil. The bacteria stimulated fungal growth and production of phenolic compounds. Sequence analysis of the two fluorescent Pseudomonas revealed 99% homologuey between HR13 and P. monteilii, and 98% between HR26 and P. resinovorans.
  •  It is clear that some bacteria (Mycorrhiza Helper Bacteria) can stimulate the establishment of the ectomycorrhizal symbiosis in tropical conditions.


Mycorrhizal fungi are an ubiquitous component of most ecosystems throughout the world and play an important role in soil processes (Smith & Read, 1997). As the fungal symbiosis modifies root functions, its microbial communities differ from those of the uninfected rhizosphere and of the surrounding soil (Katznelson et al., 1962; Ames et al., 1984; Garbaye & Bowen, 1987, 1989; Garbaye, 1991). This microbial compartment is commonly named ‘mycorrhizosphere’ (Linderman, 1988). The microorganisms associated with the mycorrhiza may complement mycorrhizal activities (i.e. N2-fixing bacteria, phosphate-solubilizing bacteria) (Secilia & Bagyaraj, 1987; Toro et al., 1996). However, more specific relationships between the microbial populations of the mycorrhizosphere and the fungal symbionts can affect the establishment of the mycorrhizal symbiosis (Garbaye & Bowen, 1987). Some bacteria can have a negative or positive effect on mycorrhiza formation (Bowen & Theodorou, 1979; Garbaye & Bowen, 1989).

Mycorrhiza Helper Bacteria (MHB) which have been previously defined as telluric bacteria promoting the development of mycorrhizal symbiosis (Garbaye, 1994), have been isolated from different plant–fungal combinations such as AM fungi (herbaceous plants) (Meyer & Linderman, 1986; Paula et al., 1992; Staley et al., 1992; von Alten et al., 1993; Requena et al., 1997) or ectomycorrhizal fungi (trees) (Dunstan et al., 1998; Duponnois & Garbaye, 1991a; Rózycki et al., 1994). The location and survival of MHBs have been assessed (Frey-Klett et al., 1997) as well as their selectivity and specificity towards the fungal symbiont from which they have been isolated (Garbaye & Duponnois, 1992; Duponnois et al., 1993).

However, the effect of MHBs on ectomycorrhizal symbiosis has only been investigated with a few northern hemisphere tree species (Pseudotsuga menziesii, Quercus robur) (Duponnois & Garbaye, 1991a,b; Garbaye et al., 1992) and a limited number of fungi, principally Laccaria bicolor (Duponnois & Garbaye, 1991a,b; Frey-Klett et al., 1997; Frey et al., 1997). To our knowledge, only one fast growing tree species has been studied (Eucalyptus diversicolor) to determine the impact of MHBs on its ectomycorrhizal status (Dunstan et al., 1998).

As Australian Acacias are frequently used in agroforestry plantations to rehabilitate degraded soils in sahelian areas, it is very important to identify factors which could optimize the effects of controlled mycorrhization. This study was undertaken with A. holosericea, an Australian acacia frequently planted in dry tropical areas and an ectomycorrhizal fungus, Pisolithus sp. strain COI 007, previously screened for its stimulating effect on A. holosericea plant growth (Duponnois et al., 2000). The aims of this study were to explore the interactions between the establishment of this ectomycorrhizal symbiosis and two bacterial strains belonging to the fluorescent pseudomonad group, isolated from a soil highly colonized by ectomycorrhizal fungi.

Materials and Methods


The ectomycorrhizal fungus, strain COI007 has been named as Pisolithus alba by Francis Martin (INRA Nancy, France) on the basis of rDNA ITS phylogeny. It was isolated from a sporocarp collected under a monospecific forest plantation of A. mangium in southern Senegal and routinely maintained on MMN agar (Marx, 1969) at 25°C. Mycelial inoculum was grown in glass jars filled with an autoclaved (120°C, 20 min) mixture of vermiculite and peat moss (4 : 1; v:v). This substrate was previously moistened with liquid medium. Then 10 fungal plugs were put aseptically into each glass jar. The substrate was completely colonized by the fungus after 8 wk at 30°C in the dark.

Isolation of bacterial strains and sequence analysis

The fluorescent pseudomonads, called HR13 and HR26, were isolated, respectively, from the rhizosphere and from the adherent soil around nodules of 3-month-old A.mangium plants growing in a sandy soil in a plantation at Cassankil (Casamance, South of Senegal) where sporocarps of Pisolithus sp. were detected. The bacteria were randomly chosen from a number of fluorescent pseudomonads which were isolated. They were cultured on King’s B medium (King et al., 1954) and cryopreserved at −80°C in glycerol 60%/Tryptic Soy Broth medium (1 : 1; v:v).

Small, subunit ribosomal DNA (16S rDNA) was amplified using the primers fD1(5′-AGAGTTTGATCCTGGCTCAG-3′) and rD1 (5′-AAGCTTAAGGAGGTCAT CCAGCC-3′) (Weisburg et al., 1991). Products of amplification were purified with a QIAquick PCR purification kit (Qiagen, Hilden, Germany) before sequencing. The sequencing reaction was performed by Genome Express (Grenoble, France), by PCR amplification in a final volume of 20 µl using 100 ng of PCR products, 5 pmoles of primer and 8 µl of BigDyeTerminators premix according to Applied Biosystems protocol. After heating to 96°C for 3 min, the reaction was cycled as follow: 30 cycles of 30 s at 96°C, 30 s at 55°C, and 4 min at 60°C (9600 thermal cycler, Perkin Elmer Applied Biosystems, Foster City, CA, USA). Excess BigDyeTerminators were removed in exclusion columns. The samples were dried in a vacuum centrifuge and dissolved with 1.6 µl of deionized formamide EDTA pH 8.0 (5/1). The samples were loaded onto an Applied Biosystems 373XL sequencer and run for 12 h on a 4.5% denaturing acrylamide gel. Sequences were aligned using CLUSTAL X software (Thompson et al., 1997). Phylogenetic distances were calculated according to the neighbour-joining method (Saitou & Nei, 1987).

Influence of bacterial isolates on the fungal growth in axenic conditions

The fungal strain COI007 was grown in Petri dishes on MMN agar at 25°C for 2 wk. Agar plugs (diameter: 4 mm; 4 mm thick) were taken from the margin of the fungal colonies. Bacterial strains were cultured in Petri dishes on 0.3% Tryptic Soy Broth (TSB, Difco Laboratories, Detroit, MI, USA) agar at 25°C for 3 d. Sterile magnesium sulphate solution (5 ml) (0.1 M) was poured on the bacterial culture and mixed with a bent glass rod to obtain a homogenous bacterial suspension. The control treatment was prepared as previously described from a Petri dish containing the same agar medium but without bacteria.

One set of experiments employed direct liquid contact between the mycelium and the bacteria. Fungal plugs were dipped in the bacterial suspensions or in the control solution for 1–2 min and transfered surface side up into empty Petri dishes (Duponnois & Garbaye, 1990).

Another set of experiments was carried out without liquid contact using two-compartment dishes. The fungal plugs were laid on the dry bottom of one compartment while the other was filled with 0.3% TSB agar medium inoculated or not (control treatment) by the bacterial strains. Gaseous diffusion from one side to the other was possible.

In both sets of experiments, three dishes, each with five mycelial plugs, were prepared for each treatment. Then the dishes, sealed with tape to prevent drying during incubation, were placed in an incubator at 25°C for 1 wk. Fungal growth measurements were made through the lid using a stereomicroscope (magnification × 40) and the mean radial growth in two perpendicular directions was calculated. The data were statistically compared to the control treatments without bacteria with the Student ‘t’ test (P < 0.05).

To quantify the fungal biomass of each plug the ergosterol content was estimated according to the modified method of Grant & West (1986). Approx. five plugs per treatment were carefully taken and placed into Eppendorf tubes (1 plug per tube) filled with 0.5 ml MgSO4 0.1 M. The plugs were crushed, centrifuged (15 000 g, 20 min) and the pellets were resuspended into 10 ml methanol, 2.5 ml ethanol and 1 g of KOH solution. The mixture was boiled for 30 min at 70°C using a condenser. After cooling, 5 ml of distilled water were added and then the solution was filtered through a Whatman paper. The ergosterol was extracted by 1 min hand shaking with 2 × 10 ml n-hexane. The water from hexane extracts was eliminated by addition of ammonium sulphate, followed by rotary evaporation to dryness. Ergosterol was resuspended with 300 µl methanol-HPLC and its amount was determined by HPLC with an UV-detector at 282 nm using a 150-mm Biorad RP column (4.6 mm inner diameter) packed with Bio-Sil C18 HL 90–3S (3 µm particle size) and a 3-cm Biorad guard column. The mobile phase was 100% Methanol and a flow rate was 1 ml min−1. There were two measurements per sample. The data were statistically compared to the control treatment without bacteria with the Student ‘t’ test (P < 0.05).

Five other plugs were sampled to measure their total water-soluble phenol contents using the Folin-Ciocalteu reagent (Box, 1983) and the results were expressed as mg gallic acid equivalents plug−1. The data were statistically compared with the control treatment without bacteria with the Student ‘t’ test (P < 0.05).

The rest of the plugs were transferred into 16 ml closed bottles (one per bottle) to measure their basal respiration rate. The bottles were flushed with ambient air, and incubated at 37°C for 3 h. Air samples were analysed every hour for CO2 content using direct injection into a GC Analytical Instruments SRA (MTI P200, Microsensor Technology Inc., Fremont, CA, USA) equiped with a TCD detector and a Molecular Sieve 5 A Plot column, with Helium as carrier gas. Each time, the data were statistically compared to the control treatments without bacteria with the Student ‘t’ test (P < 0.05).

Inoculation experiment

Soil was collected in a stand of A. holosericea (Sangalkam, 50 km, East of Dakar, Senegal), crushed, passed through a 2-mm sieve and autoclaved for 40 min at 120°C. After autoclaving, its physical and chemical characteristics were as follow: pH (H2O) 5.3; clay (%) 3.6; fine silt (%) 0.0; coarse silt (%) 0.8; fine sand (%) 55.5; coarse sand (%) 39.4; carbon (%) 0.17; nitrogen (%) 0.02; C/N 8.5; Total P (ppm) 39 and olsen P (ppm) 4.8. The soil was mixed with fungal inoculum (10 : 1; v:v) and packed in 1 dm3 polythene pots. The treatments without fungus received an autoclaved mixture of moistened vermiculite peat moss at the same rate.

Seeds of A. holosericea (provenance Bel Air) were surface sterilized in 95% sulphuric acid for 60 min, rinsed with sterile distilled water and germinated on 1% agar. After 3 d of incubation at 25°C in the dark, one pregerminated seed was planted per pot.

Both bacterial strains (HR13 and HR26) were grown in 0.3% TSB liquid medium. After 3 d culture at 25°C on a rotary shaker, the bacterial cultures were centrifuged (2400 g, 40 min) and the pellets were resuspended in 0.1 M MgSO4. The final concentration of each bacterial inoculums was more than 108 CFU per ml.

Immediately after planting, five ml bacterial suspension were injected into each pot. The treatments without bacteria received 5 mL of 0.1 M magnesium sulphate. Plants were watered twice a week with tap water without fertiliser. The pots were arranged in a randomized complete block design with 10 replicates per treatment. The seedlings were screened from the rain and grown under natural light during the hot season in Senegal (daylength approximately 12 h, mean day temperature 30°C).

Four months after inoculations, the plants were harvested. The oven dried (1 wk at 65°C) weight of the shoot was measured. Root nodules induced by indigenous rhizobia were counted and weighed on a oven dry basis (1 wk at 65°C). Sub-samples of roots with rhizosphere and mycorrhizosphere soil (100 mg f. wt) were randomly collected from each root system. Then the root systems were washed, cut into short pieces, mixed and the percentage of ectomycorrhizal short roots (number of ectomycorrhizal short roots/total number of short roots) was determined on a random sample of at least 100 short roots under a stereomicroscope (magnification × 40).

The soil of each pot was carefully mixed and 20 g of moist soil were collected to determine the microbial biomass using the fumigation-extraction method (Amato & Ladd, 1988). The results were expressed as µg C g−1 of dry soil. A further 2 g of soil were sampled to measure the content of ergosterol (Grant & West, 1986). The NH4 and NO3 contents were determined according to the method of Bremner (1965). Enumeration of CFU were performed on plate count agar media (TSA medium for the total number of bacteria and King’s B medium for the fluorescent pseudomonads) from 1 g samples collected in each pot as previously described (Duponnois & Bâ, 1998). Each subsample of mycorrhizal and nonmycorrhizal roots (100 mg), was placed into sterile Eppendorf tubes and crushed in 1 ml of MgSO4 0.1 M. Then, 100 µl of each suspension was spread on TSA or King’s B plates to enumerate the total number of bacteria and of fluorescent pseudomonads, respectively. The roots were oven dried (1 wk at 65°C) and weighed. Mycorrhizal dependancy was determined by expressing the difference between the total dry weight of the mycorrhizal plant and the total dry weight of the nonmycorrhizal plants as a percentage of the total dry weight of the mycorrhizal plant.

The data were analysed with a one-way ANOVA and the mean values were compared using Student’s t-test (P < 0.05). Bacterial populations were expressed as CFU per gram of soil or per gram of fresh matter and log transformed. For percentage mycorrhizal infection, data were transformed by Arcsinvx.


Sequence analysis

The 16S rDNA genes (1492 and 1491 bp fragments) were sequenced for HR13 and HR26, respectively. The sequence data have been deposited in the EMBL nucleotide sequence database under the accession numbers AY032725 (HR13) and AY032726 (HR26). These sequences were compared with 16S rDNA of 57 reference strains of ‘authentic’Pseudomonas species as described by Anzai et al. (2000). Ninety nine per cent of homologuey was found between HR13 and the sequence (AF064458) of type strain CIP 104883T of P. monteilii, whereas HR26 exhibited 98% of homologuey with type strain ATCC 14235T of P. resinovorans (AB021373). Analysis of the phylogenetic tree derived from these sequences (Fig. 1) indicated that HR13 belongs to the ‘P. putida group’ and that HR26 to the ‘P. aeruginosa group’.

Figure 1.

Phylogenetic concensus tree based on the alignment of 1491 bp and 1492 bp, respectively, for HR13 and HR26 of the 16S rDNA gene. Strain and accession numbers are shown in the tree. The stability of the grouping was verified by boostrap analysis (500 replicates) using CLUSTAL X software.

Confrontation between bacteria and the fungal strain COI007 in axenic conditions

Radial growth of COI007 was stimulated by HR13 in liquid culture (Fig. 2a) whereas there was no effect with HR26. When liquid contact between the fungus and the bacterium was prevented in the two-compartment Petri dish system, there was no significant effect of either bacterial isolate (Fig. 2a).

Figure 2.

(a) Effect of the bacterial strains HR13 and HR26 on the radial growth of Pisolithus alba strain COI007 after direct (hatched columns) or gaseous (closed columns) interaction. For each interaction type, the columns indexed by the same letter are not significantly different according to the Student ‘t’ test (P < 0.05). (b) Effect of the bacterial strains HR13 and HR26 on the production of phenolic compounds (expressed as gallic acid equivalents (mg) per fungal plug) by P. alba strain COI007 after direct (hatched columns) or gaseous (closed columns) interaction. For each interaction type, the columns indexed by the same letter are not significantly different according to the Student ‘t’ test (P < 0.05). (c) Effect of the bacterial strains HR13 and HR26 on the ergosterol content of P. alba strain COI007 fungal plugs after direct (Hatched columns) or gaseous (closed columns) interaction. For each interaction type, the columns indexed by the same letter are not significantly different according to the Student ‘t’ test (P < 0.05).

In liquid culture, the gallic acid content of the fungal plugs was significantly greater with HR26 than in the control (Fig. 2b). In the two-compartment system, gallic acid increased significantly with both bacterial strains.

The ergosterol content of fungal plugs was not significantly affected by the bacterial strains in liquid culture (Fig. 2c). By contrast, HR26 significantly increased the ergosterol content of the fungal plugs in the two-compartment system (Fig. 2c).

In liquid culture, CO2 production of fungal plugs was not significantly different with HR13 and HR26 (Fig. 3). By contrast, there were significant differences between bacterial treatments and the control, specially with HR26 which involved the highest CO2 production after 120 min incubation in the two-compartment system (Fig. 4).

Figure 3.

Evolution of the CO2 production of Pisolithus alba strain COI007 fungal plugs after direct interaction with HR13 or HR26. Data indexed by the same letter are not significantly different according to the Student ‘t’ test (P < 0.05). Control, squares; HR13, triangles; HR26, circles.

Figure 4.

Evolution of the CO2 production of Pisolithus alba strain COI007 fungal plugs after gaseous interaction with HR13 or HR26. Data indexed by the same letter are not significantly different according to the Student ‘t’ test (P < 0.05). Control, squares; HR13, triangles; HR26, circles.

Inoculation experiment

The Pseudomonas isolates inoculated alone had no significant effect on plant growth (Table 1). Ectomycorrhizal fungal inoculation (COI007) with or without bacteria increased the height, shoot and root biomasses significantly (Table 1). The positive fungal effect on height and shoot biomass was significantly enhanced when HR26 was inoculated together with the fungus. The same bacterial effect was recorded with HR13 on shoot biomass. However, there were no significant effects on mycorrhizal (Table 1).

Table 1.  Effect of the ectomycorrhizal fungus Pisolithus alba COI007 coinoculated or not with fluorescent Pseudomonas isolates on the plant growth, rhizobial symbiosis and ectomycorrhizal establishment
TreatmentsHeight (cm)Shoot biomass (mg d. wt)Root biomass (mg d. wt)Total number of nodules per plantNodule biomass (mg d. wt)Ectomycorrhizal infection (%)Mycorrhizal dependancy (%)
  • 1

    Data in the same column followed by the same letter are not significantly different according to the one-way ANOVA (P < 0.05).

Without COI 007
 Control 6.2a1 127.2a396.1ab 0a 0a 0a
 HR13 7.1a 184.2a280.0a 0a 0a 0a
 HR26 7.2a 164.2a360.2a 0a 0a 0a
With COI 007
 Control17.8b 696.7b400.3ab 4.0ab 9.5ab29.7b54.4b
 HR2625.3c 858.3c600.4b 8.2bc20.2b39.0c61.6b

No nodules were observed in the treatments without fungal inoculation (Table 1). By contrast, ectomycorrhizal inoculation induced nodule formation from contaminant rhizobia and this effect was significantly enhanced (number and biomass of nodules) by HR13 (Table 1). The most likely explanation for this contamination was that the tap water used during this experiment contained rhizobia.

Both bacterial strains significantly increased percentage ectomycorrhizal infection (Table 1).

The fungal and bacterial inoculations had no effect on microbial biomass (Table 2). The NO3 and NH4 contents were significantly reduced when the ectomycorrhizal fungus was inoculated (Table 2). This negative effect was not modified by the bacteria excepted for the treatment HR13 + COI007 where the NH4 content was significantly higher than in the treatment COI007 alone (Table 2).

Table 2.  Effect of the ectomycorrhizal fungus Pisolithus alba COI007 coinoculated or not with fluorescent Pseudomonas isolates on the microbial biomass, NO3, NH4, ergosterol contents and on the abundance of bacterial cells in the soil and roots after 3 month culture in glasshouse condition
TreatmentsMicrobial biomass (µg C g−1 of soil)NO3 (µg N g−1 of soil)NH4+ (µg N g−1 of soil)Ergosterol (µg g−1 of soil)Total CFU per g−1 of dry soilTotal CFU per g−1 of dry root
  • 1

    Data in the same column followed by the same letter are not significantly different according to the one-way ANOVA (P < 0.05).

Without COI 007
 Control23.0a131.6c17.3bc0a2.3 106a7.7 106a
 HR1336.2a22.1abc20.4c0a1.5 1010b1.9 1010b
 HR2643.2a30.1bc12.5b0a1.6 1010b2.7 1010b
With COI 007
 Control29.7a15.6a 2.2a0.48b2.8 106a1.3 107a
 HR1332.3a29.4abc23.8c0.64c1.9 106a5.1 107a
 HR2649.0a19.1ab 0.48a0.52b2.2 106a2.1 107a

Ergosterol was detected only in the soil inoculated with fungal isolates (Table 2). The ergosterol content was significantly enhanced by the addition of HR13 compared to the treatment COI007 alone (Table 2).

The abundance of total culturable bacteria (per g of dry soil or root) was significantly higher in the HR13 and HR26 treatments inoculated alone. By contrast, the coinoculation of the bacterial strains and the ectomycorrhizal isolate caused a depression in the number of bacteria compared to the treatments with the bacteria inoculated without fungus (Table 2). No fluorescent Pseudomonas were detected from the soil and roots after 3 months culture in glasshouse conditions.


P. alba (COI007) dramatically increased plant growth: shoot biomass was increased by 448%. This beneficial effect of the ectomycorrhizal symbiosis has previously been demonstrated with another Pisolithus isolate COI024 collected from an A. holosericea plantation in Senegal (+142% in shoot biomass) (Duponnois et al., 2000). This positive effect was increased with HR13 and HR26 coinoculated with COI007 (shoot biomass enhanced, respectively, by 46% and 23.2% above that measured in the control without bacteria). The percentage ectomycorrhizal infection was also significantly promoted by 158.2% and 31.3%, respectively. Both bacterial strains can be considered as MHB. After 3 months the bacterial inoculants could no longer be detected in the soil or rhizosphere plate counts (a PCR-DGGE analysis would confirm the results). A decline in MHBs after 3 months culture has been already described with BBc6, a P. fluorescens isolate, in the ectomycorrhizal association between Douglas fir and Laccaria bicolor (Frey-Klett et al., 1997). This result suggests that the survival of these two bacterial inoculants is probably dependent on host plant (HR13 and HR26 were present in the mycorrhizosphere of 3-month-old A. mangium) and on soil conditions (general soil conditions differ under the two Acacia stands). Moreover, the sterilized soil was rapidly recolonized by other bacteria. After 3 month culture, the bacterial community composition could be different from the native soil in which HR13 and HR26 were isolated and could include bacteria that are antagonistic to the inoculant strains.

Sequence analysis of both Pseudomonas isolates showed that HR13 and HR26 were authentic species of Pseudomonas which includes three subgroups, P. aeruginosa, P. fluorescens and the P. syringae subgroup (Palleroni, 1984). The Pseudomonas putida group (HR26) was included in the P. fluorescens subgroup (Anzai et al., 2000). To our knowledge, these species of Pseudomonas have not previously been studied for their effects on ectomycorrhizal establishment.

Greater percentage ectomycorrhizal infection was correlated with the total number of nodules per root systems. It is thought that nodule formation and functioning are dependant on mycorrhizal inoculation (Cornet & Diem, 1982). The main explanation is that the improvement of P uptake by the host plant resulting from mycorrhizal symbiosis enhances nodulation and N2 fixation (Cornet & Diem, 1982).

The MHB effect was accompanied by an increase in the ergosterol content in the soil. This shows that the bacterium was acting on fungal growth in the soil and consequently could enhance the effective mycelial surface areas provided by fungal hyphae to explore greater volumes of soil and to overcome mineral nutrient and water depletion zones near active root surface. This supports the hypothesis of Frey-Klett et al. (1997) who suggested that the main mechanism involved in the MHB effect concerned the bacterial influence on the fungal growth. Moreover, it has previously been demonstrated that the ability of the bacteria to stimulate mycelial growth was strongly correlated with their effect on the ectomycorrhizal establishment (Garbaye, 1994).

The cocultures of the bacteria and the fungus showed that the radial growth was stimulated by HR13 and fungal biomass (expressed as ergosterol) by HR26. This suggests that different mechanisms may be involved in the MHB effect. Moreover, increased respiration (expressed by the CO2 production) demonstrates some increase in metabolism in the presence of both bacteria.

Another hypothesis to explain the MHB effect has been suggested by Garbaye (1994). The MHBs could improve the receptivity of the root to the fungus before the first mycorrhizas are formed. It has been demonstrated recently that hypaphorine, the major indolic compound isolated from the ectomycorrhizal fungus Pisolithus tinctorius, controls the elongation of root hairs and consequently, is involved in establishment of mycorrhizal association (Bèguiristain & Lapeyrie, 1997; Ditengou et al., 2000). In our study, both bacterial isolates stimulated phenolic production by the fungus. Further investigations must be done to clarify this bacterial effect on the phenol compounds and, in particular, on hypaphorine production.

In conclusion, it appears that some bacteria can help the establishment of ectomycorrhizal symbiosis in tropical conditions with an Australian Acacia. The main mechanisms involved in this phenomena concern the interactions between the MHB and the fungal symbiont. Further research must be undertaken to identify the compounds responsible for the promotion of the fungal growth and the signal molecules implied in the recognition processes between the host and the fungus. As the bacterial inoculants disappeared after 3 month culture, the bacterial effect must be transient and could involve some changes in the recolonizing indigenous microbial community (included Rhizobium in only the coinoculation treatments). In future experiments, a sequential harvesting could determine these interactions between the inoculant bacteria and the other microbial components of the system (Rhizobium, indigenous microbial communities, arbuscular mycorrhizal fungi).

From a practical point of view, the use of MHBs could facilitate the introduction of controlled mycorrhization in nursery and forestry practices through the soudano-sahelian areas.


The authors are very grateful to Dr J. Garbaye (INRA, France), P. Frey-Klett (INRA, France) and P. Roger (IRD, France) for reviewing the manuscript.