Arbuscular mycorrhiza in mini-mycorrhizotrons: first contact of Medicago truncatula roots with Glomus intraradices induces chalcone synthase


Author for correspondence: Andres Wiemken Tel: +41 (0)61 267 23 11 Fax: +41 (0)61 267 23 30


  •  The association of plants with arbuscular mycorrhizal (AM) fungi is widespread in nature, but little is known about molecular aspects of this symbiosis. Particularly during the early stages of the AM symbiosis, it is difficult to monitor growth of the two partners, to dissect gene expression patterns and to correlate them with plant, fungal or symbiosis development.
  •  A new system, the ‘mini-mycorrhizotron’, was established to cultivate seedlings of Medicago truncatula in mycorrhizal symbiosis with Glomus intraradices under gnotobiotic conditions. This system allows natural growth of the symbiotic partners and permits the continuous noninvasive observation of the development of plant and fungus under a microscope.
  •  The mini-mycorrhizotron was used to determine the stage of induction of a mycorrhiza-related gene detected by differential display-reverse transcription-PCR, namely a novel chalcone synthase (Mt-chs1). The gene is induced in roots at the stage of the first fungal contact.
  • • The mini-mycorrhizotron allowed identification and cloning of a symbiosis-related gene, and the correlation between its expression and the developmental stage of the symbiosis was established. This provides a useful tool for molecular and developmental studies of the early stages of AM symbioses.

Mt-chs1 nucleotide sequence can be found at the Genbank data base (accession no. AJ277211). The following sequences were also used: Glycine maxchs (X52097), chs1 (X54644); chs2 (X65635), chs3 (X53958), chs5 (L07647), chs6 (L03352), chs7 (M98871); Pisum sativumchs (X80007), chs1 (D10661), chs2 (X63334), chs3 (D88261), chs4 (D88260), chs5 (D88262), chs7 (D88263); Medicago sativachs1 (L02901), chsI (X68106), chs2 (L02902), chs4–1 (U01018), chs8 (L02904), chs9 (L02905), 12–1 (U01021); Medicago truncatulachs1, this work (AJ277211); Trifolium subterraneumchs1 (M91193), chs2 (M91194), chs3 (L24515), chs4 (L24516), chs5 (L24517), chs6 (M91195); Phaseolus vulgarischs (X06411) and Vigna unguiculatachs (X74821).


Symbiotic interactions between plant roots and arbuscular mycorrhizal (AM) fungi are widespread in nature, but little is known about the molecular mechanisms leading to recognition, establishment and functioning of this symbiosis. A characteristic feature of the fungi forming arbuscular mycorrhizas is that they exhibit a strong biotrophic dependence on their host plants (Gianinazzi-Pearson, 1983). In the absence of the host, their growth is limited to a relatively short time, after which hyphal growth ceases (Bonfante & Perotto, 1995). In vitro cultures of host plants in mycorrhizal association with AM fungi are valuable research tools because the physiology of infected and uninfected plants can be compared without interference from other rhizosphere organisms. It is possible to cultivate AM fungi in Petri dishes hosted by transformed roots, a method suitable for producing high-quality fungal spores (Bécard & Piché, 1992). However, the plant partner is mutilated in this system and it is difficult to extrapolate results obtained from it to symbioses occurring in natural environments. Particularly during early stages of the AM symbiosis, it is difficult to monitor the growth of both partners, to dissect gene expression patterns and to correlate them with plant, fungal or symbiosis development. In several attempts to address these questions, different techniques were evaluated. For instance, sterile spores and plantlets were incubated on agar-coated microscope slides, on paper-coated slides or in Fahraeus slides (Hepper, 1981). The slides were then placed in closed test tubes with liquid medium. The first visible sign of the symbiotic interaction was strong hyphal growth and appressoria formation. These methods allowed the cultivation of mycorrhized plants for up to 120 d, when new, small spores were observed on the external mycelium. The development of the fungus on agar slopes or in Fahraeus slides (Nutman, 1959) could be followed in situ under a microscope. However, each slide could be observed only once, because this compromised the aseptic conditions of the system, destroying it. All these methods are rather labour-intensive and cannot be performed on a large-scale basis. For these reasons, methods are needed to help cultivation of mycorrhizal plants while avoiding contaminations with other organisms, and to facilitate molecular analysis of the early stages of the AM symbiosis.

The aim of the present work was the development of a novel culture system which has been designated the ‘mini-mycorrhizotron’. This system allows the continuous, nondestructive monitoring of the growth of the fungus and the host plant. In particular, it is possible to observe the development of the symbiosis at any time over a period of 20–30 d, and to count fungal structures along the root without staining it. Furthermore, any contamination with other rhizosphere organisms is excluded. Continuous observation of each chamber allows the selection of samples at one specific developmental stage and the harvesting of them for further molecular analysis. The mini-mycorrhizotron is therefore a highly suitable tool to investigate the early stage of the interaction, where recognition occurs between the two partners. This paper demonstrates the use of the mini-mycorrhizotron to identify plant genes induced in the AM symbiosis. Using the mini-mycorrhizotron in conjunction with the mRNA differential display technique, a new symbiosis-related chalcone synthase (CHS) cDNA has been isolated from Medicago truncatula induced at the stage of the first fungal contact.

Materials and Methods

Construction of the mini-mycorrhizotrons

The mini-mycorrhizotron was built with two incubation chambers of 500-µl void volume (Sigma, Buchs, Switzerland). Each incubation chamber consists of a thin, transparent plastic slide with a 0.5-mm silicon rubber glued on the perimeter. Two incubation chambers can be perfectly sealed by pressing their silicon profiles together, creating a mini-mycorrhizotron with a void volume of 1 ml. A small hole cut into the silicon rubber sealing allows the hypocotyl to grow through. The mini-mycorrhizotrons were kept upright in an inverted Magenta box (Sigma, Buchs, Switzerland) in a specially designed rack (Fig. 1).

Figure 1.

Construction and use of the mini-mycorrhizotron. (a) Two incubation chambers (25 mm wide × 45 mm high × 3 mm thick) consisting of thin and transparent plastic slides with a 0.5-mm silicon rubber glued on the perimeter with an opening on the top. (b) The chambers are closed together and the mini-mycorrhizotrons are kept upright. The figure shows a 10-d-old Medicago truncatula seedling in a mini-mycorrhizotron sealed with office clamps. (c) Several mini-mycorrhizotrons in a rack.

Glomus intraradices Schenck and Smith was cultivated on carrot root cultures in two-compartment Petri dishes as previously described (Bécard & Piché, 1992). Nonmycorrhizal roots were grown in the same way to produce the control inoculum. To prepare the inoculum solution, 10–12 wk after inoculation the agar of the compartment of the Petri dish where only the fungus grew was mashed and suspended in liquid medium M (Bécard & Fortin, 1988) without sucrose. Similarly, a control solution was prepared from the ‘fungal’ compartment of a dish with nonmycorrhizal roots.

Medicago truncatula cv. Jemalong (line A17) seeds were surface-sterilized with concentrated sulphuric acid and pre-germinated on 0.7% agar for 5 d. Each seedling was then enclosed between two slides, with the hypocotyl protruding through the opening in the rubber sealing of the mini-mycorrhizotron. The number of spores was adjusted to 400 ml−1 by diluting the inoculum solution with liquid medium M without sucrose, and 1 ml of the solution containing AM spores was injected with a sterile syringe in the chamber with the seedling. Ten mycorrhizal and 10 control chambers were prepared per time-point. The mini-mycorrhizotrons were wrapped with aluminium foil, incubated vertically in a Magenta box (Sigma) and transferred to a growth cabinet at 22°C for 16 h in the light (100 µmol m−2 s−1) and 18°C for 8 h in the dark.

In separate experiments, two Teku-Multiflor plates with 150 slots (Wyss Samen und Pflanzen AG, Basel, Switzerland) were filled with an autoclaved mixture of 3.5 parts of coarse-grained sand (Birs Sand, Industrielle Werke, Basel, Switzerland), one part quartz sand (Quartz d’Alsace, Kaltenhouse, France), 0.5 parts of loam (Botanical Garden, Basel, Switzerland) and five parts of Terragreen (Maagtechnik, Zürich, Switzerland). The plates were planted with one 5-d-old Medicago seedling per slot and each seedling was inoculated with 1 ml of inoculum solution and cultivated in a growth cabinet (same conditions as before). At harvest (0, 5, 10, 15 and 20 d post-inoculation), the plants were removed from the substrate and washed in water, and the majority of the roots frozen in liquid nitrogen. The rest of the roots was stained with trypan blue, fungal structures were counted and root lengths were measured (Giovannetti & Mosse, 1980).

In some experiments, seedlings enclosed in mini-mycorrhizotrons were inoculated with Sinorhizobium meliloti (strain 1021) or a spore suspension of Fusarium solani f. sp. phaseoli, Fusariumsolani f. sp. pisi or Rhizoctonia solani. Bacteria were grown overnight in succinate minimal medium (Schmidt et al., 1992), and pelleted by centrifugation (10 min, 2500 g). The bacterial pellet was resuspended in 50-ml medium M without sucrose and used as inoculum. Fungi were grown on V-8 agar plates (200 ml V-8 juice from Campbell Soup Company, Camden, NJ, USA; 2 g CaCO3; 15 g agar from United States Biological, Swampscott, MA, USA; and 800 ml of water). Spores were harvested from 2- to 4-wk-old cultures as follows: sterile water was added on to the agar surface and the spores were scraped off with a spatula. The density of the spore suspension was determined and diluted to 300 spores ml−1 with medium M without sucrose.

Microscopical observations

The development of the symbiosis could be monitored efficiently on a regular basis and without disturbing the interaction. Minimycorrhizotrons with plants and fungi were observed daily under a microscope (Zeiss Axioplan, Zeiss AG, Oberkochen, Germany) at 200× magnification; the growth of fungi and roots could be easily monitored. Every 5 d, germinated spores and hyphal branching points lying in the optical field around the root, as well as root–hyphal contact sites and appressoria, were counted. The lengths of the main and the lateral roots were measured by laying the chamber on millimetre paper.

After opening of the minimycorrhizotrons, three root pieces of length approx. 3 mm were excised near the apex, the middle and the upper region, stained in trypan blue and microscopically checked for internal fungal structures under the microscope. The rest of the root was frozen in liquid nitrogen for RNA analysis. Ten control samples were harvested after 15 culture days and completely stained in trypan blue to monitor internal fungal development.

Some 3-mm root fragments and control roots were analysed by environmental scanning electron microscopy (ESEM, Philips Electron Optics, Eindhoven, the Netherlands). This method allows the sample environment to be varied through adjustment of a range of pressures, temperatures and gas compositions. Root samples could thus be examined in their natural states (i.e. in a wet atmosphere and with water partial pressures between 613 and 667 Pa) without further modification or preparation (Danilatos, 1988; Philips Electron Optics, 1996).

mRNA differential display

mRNA was extracted from the frozen samples using the Plant RNeasy Mini Kit (Qiagen Inc., Chatworth, NJ, USA). Samples were further treated with DNase I using the MessageClean kit according to the manufacturer’s recommendations (GenHunter, Brookline, MA, USA) and processed by differential display-reverse transcription-PCR (Liang et al., 1993). Reverse transcription (RT) reactions were performed using A-anchored oligo-dT(11) primers and 0.2 µg DNase I-treated mRNA. The resulting cDNA preparations were used as templates in PCR amplification with the same oligo-dT as used for the RT and in combination with an arbitrary primer (sequence 5′AAGCTTAGTTATC3′) in the presence of (α-33P)dATP. The thermo-cycler conditions (Genius, Techne, Princeton, NJ, USA) were 94°C for 30 s, 40°C for 2 min, 72° for 1 min for 30 cycles and finally 72°C for 5 min. PCR products were separated on a 7-M urea : 6% acrylamide sequencing gel. The gel was dried and exposed to an X-ray film (Biomax, Kodak, Switzerland) for 24 h. The differentially appearing partial cDNA fragments were thymidine-adenosine (TA)-cloned in vector pGemT (Promega, Madison, WI, USA), screened for false positives (Vögeli-Lange et al., 1996) and sequenced (ABI Prism 310 Genetic Analyser, Perkin Elmer Corp., Norwalk, CT, USA). Fragment-specific primers with sequence forward 5′CGAAAAGATGAATGCAAC3′ and reverse 5′CCAAAACCAAATAACACAC3′ were designed and used for RT-PCR. RT of 2 µg RNA was performed with the Reverse Transcription System (Promega) in a total volume of 25 µl. Oligo-(dT)15 primers were added to the mRNA and incubated at 70°C for 5 min, then the tubes were chilled on ice for 5 min and the avian myeloblaitosis vints reverse transcriptase, the reaction buffer and nucleotides were added. The reaction was incubated for 60 min at 45°C. Usually, 1 µl of the reaction mixture was used for PCR. All PCRs were performed as previously described in 10-µl volumes. To show constitutive expression, RT-PCR with ubiquitin primers (forward 5′ATGCAGATYTTGTGAAGAC3′ and reverse 5′ACCACCACGRAGAC-GGAG3′) and 5.8-s mRNA primers (forward 5′GAATGACTCTCGGCAACGGATAT-CTAGGCTC3′ and reverse 5′GTGACACCCAGGCAGACGTGCCCTCAACC3′) were used.

Cloning and characterization

Genomic plant DNA was prepared with the DNeasy Plant Mini Kit (Qiagen). A full-length clone of the CHS coding region was obtained by touch-down PCR using genomic DNA as template. The thermo-cycler conditions were 94°C for 30 s, five cycles at 52°C for 1 min, five cycles at 49°C, five cycles at 46°C, then 10 cycles at 44°C and 10 cycles at 42°C, and finally 72°C for 1 min.

The forward primer (5′CTGCAGCCATGGTIAGYGTDKMHGARATYMG3′) was designed based on nucleic acid degeneracy as determined from an alignment of CHS sequences of Glycine max and Medicago sativa. As a reverse primer the fragment-specific oligonucleotide was used. The amplified band was eluted from the agarose gel (Geneclean Spin Kit, BIO 101, Vista, CA, USA), cloned and sequenced. Sequence analysis was carried out using the GCG software (Genetics Computer Group, Madison, WI, USA) and the BLAST network services of the National Centre for Biotechnology Information (National Library of Medicine, Bethesda, MD, USA). The phylogenetic tree was elaborated with the Lasergene program of DNASTAR (Madison, WI, USA).


Early stages of AM development

Most AM fungal spores germinated within 10 d post-inoculation in the mini-mycorrhizotrons, and, at each time-point chosen, a characteristic developmental stage was observed which was distinct and different from the subsequent stages. The AM fungus grew as on the Petri dish root-organ cultures and a typical developmental pattern could be recognized: hyphae grew straightforwardly until they came close to a root, then they suddenly branched before producing appressoria, as described by others (Giovannetti et al., 1993).

The roots also developed normally, and no significant difference in total root length, or number and length of lateral roots was noticed between inoculated and noninoculated plants. In addition, no differences in the plant growth pattern and the size were observed between plants grown in the mini-mycorrhizotron or in the Multiflor plates (data not shown).

Fig. 2 shows the development of G. intraradices in a time-course experiment. At the beginning, single and aggregated spores from the inoculum were present. Within 5 d post-inoculation (dpi), spore germination occurred and hyphae were formed. Between 5 and 10 dpi, intense hyphal branching occurred and the first contact sites were detected. Between 10 and 15 dpi the number of branching and contact sites increased drastically and the first appressoria appeared. After 20 dpi, fungal mycelium spread in the chamber and numerous contact sites and appressoria were present. In addition, the first arbuscules were visible after staining with trypan blue. Some mini-mycorrhizotrons were incubated for a longer time and at day 50 arbuscules, vesicles and newly formed spores were clearly visible. Internal root colonization reached in this case 80% of the root length (data not shown). The fungal structures formed in the mini-mycorrhizotrons were morphologically indistinguishable from the ones observed in pot cultures of M. truncatula and G. intraradices.

Figure 2.

Development of mycorrhizal symbiosis in mini-mycorrhizotrons. (a) The mini-mycorrhizotrons were scored at various times after the start of the experiments. The numbers of germinated spores and hyphal branching points lying in the optical field around the root, hyphal contact sites with roots and appressoria were determined. The data from three time-course experiments were pooled (bars: standard error of the mean, n = 30). Key: light grey, germinated spores; mid-grey, branching points; dark grey, contact sites; black, appressoria. (b) Pictures that represent typical developmental stages, taken without any stain through closed mini-mycorrhizotrons. From left to right: a spore sitting near the root; a germinating spore, which is juvenile as indicated by its whitish colour; a hyphal branching point; a spore with germ tube after contact with the root; and an appressorium; the small round structure is an air bubble trapped between the root and the fungus (200× magnification).

Mini-mycorrhizotrons allowed study of the distribution of contact sites along a root without fixation or staining procedures using ESEM in a wet environment (Fig. 3). Roots appeared nondamaged and root hairs, spores and hyphae were fully turgescent. The germ tubes seemed either to penetrate between epidermal cells or to grow on the root surface parallel to the root axis and to penetrate at a more remote location.

Figure 3.

Environmental scanning electron microscopy (ESEM). The surface of root pieces directly taken from a mini-mycorrhizotron was observed by ESEM using wet mode. (a) Control root with turgescent root hairs. (b) Germinated spores attached to a root. At a water partial pressure of 719.9 Pa the spore on the right is still fully turgescent, whereas the one on the left is already starting to collapse. (c) Detail of (b): a spore with a germ tube penetrating between two epidermal cells. A swelling of the fungal germ tube can be observed, which can be considered to be an appressorium. sp: spores; rh: rhizodermis; gt: germ tube.

Mini-mycorrhizotrons were also successfully used to establish the symbiosis between Sinorhizobiummeliloti and M. truncatula. After 15 dpi up to three nodule-primordia per seedling were visible (data not shown). Furthermore, M. truncatula seedlings were inoculated with different pathogenic fungi in the mini-mycorrhizotrons. After 5 dpi the pathogenic mycelium was dense and many appressoria were visible. Fusarium solani f. sp. phaseoli caused more intense browning of the roots than Fusarium solanif. sp. pisi or R.solani, which did not cause visible symptoms.

Induction of a chalcone synthase gene at the stage of initial fungal root contact

By differential display of mRNA, we detected a prominent upregulated band at the time-point of the initial branching of the fungus when the first contact sites with the root were present. The expression level in noninoculated control plants was low during the whole time of the experiment (Fig. 4). The corresponding gene was cloned and sequenced. The gene contained a coding region of 1170 bp, corresponding to 389 amino acids and an intron of 159 bp inserted at bp 180. Comparison with the Genbank data base showed that the gene encodes a CHS with 94% identity to M. sativa CHS chs2 (Junghans et al., 1993) and therefore it was designated Mt-chs1. Like in other CHSs the region corresponding to the active site from amino acid 143–192 in Mt-chs1 is well conserved and contains the cysteine residue that probably binds the 4-coumaryl-CoA group (Lanz et al., 1991).

Figure 4.

Differential display analysis of mRNAs expressed early during the arbuscular mycorrhizal (AM) symbiosis. Each line represents a time-point (days post-inoculation, dpi) from a time-course experiment with (right part) or without (left part) AM fungus. Arbitrary primers were tested in combination with an oligo-(dT)11-A and an oligo-(dT)11-G primer. In all lines the overall banding pattern between the mycorrhizal and the nonmycorrhizal sample combinations was reproducible. An arrow points to the differentially appearing band later characterized as Mt-chs1.

A phylogenetic comparison (Fig. 5) revealed that there are at least five subgroups of CHS in different legumes. Pisum sativum chs4 and Trifolium subterraneumchs3 form the most distant group that diverged earlier in the phylogeny. The second branch contains CHS of T. subterraneum and P. sativum, and of the more distant G. max. All other G. max CHSs are part of a separate branch. Also, the remaining CHSs of P. sativum and T. subterraneum are clustered. The CHSs of Medicago are separated into two main groups with the first containing chs1, 9 and 4–1, and the second chs8, I, 12–1, 2 and Mt-chs1.

Figure 5.

Phylogenetic tree of selected legume chalcone synthases. The tree was constructed using the Clustal method (Higgins & Sharp, 1988, 1989). The accession numbers for the sequences used were: Glycine maxchs (X52097), chs1 (X54644); chs2 (X65635), chs3 (X53958), chs5 (L07647), chs6 (L03352), chs7 (M98871); Pisum sativumchs (X80007), chs1 (D10661), chs2 (X63334), chs3 (D88261), chs4 (D88260), chs5 (D88262), chs7 (D88263); Medicago sativachs1 (L02901), chsI (X68106), chs2 (L02902), chs4–1 (U01018), chs8 (L02904), chs9 (L02905), 12–1 (U01021); Medicago truncatulachs1, this work (AJ277211); Trifolium subterraneumchs1 (M91193), chs2 (M91194), chs3 (L24515), chs4 (L24516), chs5 (L24517), chs6 (M91195); Phaseolus vulgarischs (X06411) and Vigna unguiculatachs (X74821).

Temporal expression pattern of Mt-chs1 before appressoria formation

Using a pair of specific primers designed to amplify the 3′-end of the cDNA, the expression pattern of Mt-chs1 was studied in more detail. RT-PCR analysis of nine independently repeated experiments revealed an induction of Mt-chs1 after the first root contact of the fungus consistent with the one detected by the differential display (Fig. 6). Since different results on the temporal induction of CHS by AM fungi are reported in the literature (Volpin et al., 1994, 1995; Blee & Anderson, 1996; Mohr et al., 1998), the experiment was repeated nine times. In all experiments the fungus induced Mt-chs1 (Fig. 6a, rows 1–9). However, some variation was detected in the noninoculated plants. In experiment 1, Mt-chs1 is expressed constitutively at a low level. A different pattern was observed in experiments 2, 3, 5 and 7: in particular, in experiment 5 Mt-chs1 is slightly present at 5 dpi, then induced between 10 and 15 dpi to disappear again towards 20 dpi. An opposite pattern was observed in experiment 9: the transcript was present at 5 and 20 dpi, but not at 10 and 15 dpi. Furthermore, in experiments 4, 6 and 8 no expression of Mt-chs1 was observed at all (Fig. 6a, rows 1–9). However, the transcripts used as constitutive controls always appeared in the same intensity as in Fig. 6(a) rows 10 and 11. The induction of Mt-chs1 by AM fungi was also observed in plants cultivated in Multiflor plates (Fig. 6b, row 1).

Figure 6.

Expression pattern of Medicago truncatula chs1. (a) The expression patterns of the gene in plants grown with or without the arbuscular mycorrhizal (AM) fungus were compared. Rows 1–9: replicate experiments with the induction of Mt-chs1 by the mycorrhizal fungus. At each time-point five roots were pooled, and mRNA was extracted, reverse transcribed (RT) and analysed by PCR. The amplified fragment was 170 bp long. Row 10: RT-PCR using primers for ubiquitin (150 bp product), and row 11: using primers corresponding to 5.8-s rRNA (120 bp product) as controls for constitutive expression. These controls refer to the first replicate. The controls of the following replicates (data not shown) were identical to those shown. (b) Row 1: expression of Mt-chs1 in a time course with wild-type plants grown in Multiflor plates; row 2: control using ubiquitin primers. Row 3: expression of Mt-chs1 in a time course with dmi1 mutant plants (Nod, Myc−1) in mini-mycorrhizotrons; row 4: expression of ubiquitin in the same experiment. Row 5: plants inoculated with Fusarium solani f. sp. phaseoli (F. ph), Fusarium solani f. sp. pisi (F. pi) or Rhizoctonia solani (R. s.), or noninoculated (n. i.) and grown for 0 or 5 d in mini-mycorrhizotrons. Row 6: expression of ubiquitin as constitutive control. Row 7: PCR with Mt-chs1 primers on genomic DNA of Medicago truncatula (M. t.) or Glomus intraradices (G. i.) to asses plant origin of the template. Primers for other unknown differentially expressed cDNAs produced a signal only with G. intraradices DNA (data not shown).

To test whether the induction of Mt-chs1 is related to plant defence, its expression was analysed in the dmi1 Myc−1 mutant of M. truncatula (Catoira et al., 2000) (Fig. 6b, row 2). Dmi1 is the first out of four alleles identified at the locus named dmi1.Dmi1 cannot be penetrated by the fungus in the early stages of the interaction, although a higher number of appressoria are detected, and no functional AM symbiosis is formed. Concomitantly with the halt of the colonization, RT-PCR analysis revealed a strong induction of Mt-chs1 in the mutant, suggesting that induction may be related to a defence response.

To study whether the induction of Mt-chs1 is mycorrhiza-specific its expression was compared with M. truncatula seedlings inoculated with fungal pathogens (Fig. 6b, row 4). At both times, the intensity of the bands in plants inoculated with F. solani f. sp. pisi and R. solani was not different from the intensity of the bands in the corresponding noninoculated control, showing that these fungal pathogens do not induce Mt-chs1. Neither pathogen induced disease symptoms. In contrast, the pathogens F. solani f. sp. phaseoli caused disease symptoms and the mRNA of Mt-chs1 was clearly induced at 5 dpi.


Early developmental stages of AM

In this work a new system to cultivate mycorrhizal seedlings under gnotobiotic conditions was established. Mini-mycorrhizotrons allow natural growth of the symbiotic partners during the early stages, because, in contrast to root cultures, in mini-mycorrhizotrons the plants are photosynthetically active and it is not necessary to add additional sources of carbohydrates. Also, the chambers permit the continuous and noninvasive observation of the development of plant and fungus under a microscope, without disturbing the system. In particular, it was possible to observe the development of the symbiosis at any time over a period of at least 20 d, and to count fungal structures along the root without staining it, confirming that the establishment of the AM symbiosis in the chambers was highly reproducible. Furthermore, roots grown in mini-mycorrhizotrons can be directly observed by ESEM, avoiding washing and fixation procedures that can detach fungal structures and create artefacts. Observation of each chamber allows the selection of samples at one specific developmental stage and harvesting of them for further molecular analysis, allowing the establishment of correlations between plant gene expression and mycorrhiza development. Any interaction with unwanted organisms can be excluded because of the aseptic growth conditions. Moreover, in addition to the interaction with mycorrhizal fungi, plants also interact normally with rhizobia or pathogenic fungi when grown in the mini-mycorrhizotrons, allowing comparison of the reactions of the plant to the different organisms.

Some previous attempts to establish a similar system were made by different authors (reviewed in Hepper, 1981); however, all of these approaches were extremely labour-intensive and visual observation of the symbiosis often led to the loss of the aseptic conditions.

Induction of a defence response before appressoria formation

Chalcone synthase (EC is the enzyme catalysing the first step committed to the flavonoid biosynthesis. Flavonoids act as antimicrobial compounds as medicarpin in M. sativa (Harrison & Dixon, 1994), but also play a role as signalling molecules which can stimulate AM fungi spore germination in vitro (Tsai & Philipps, 1991). The activation of the flavonoid pathway by CHS is well studied in other plants, such as P. sativum, Phaseolus vulgaris, Nicotiana tabacum, Petunia hybrida, Petroselinum crispum and many others (Block et al., 1990; van der Meer et al., 1992; An et al., 1993; Bate et al., 1994; Blee & Anderson, 1996). Legumes contain a family of between three and 12 CHS genes, as reported for M. sativa, G. max and P. sativum (Junghans et al., 1993). However, it is not known how many genes code for CHS in M. truncatula. The CHSs share 82–90% sequence identity among the genera and are highly conserved (McKhann & Hirsch, 1994).

In this study, CHS expression was used to monitor the response of M. truncatula to G. intraradices in mini-mycorrhizotrons during the ealry stages of the interaction. One gene encoding CHS in M. truncatula, Mt-chs1, was identified by differential display comparing mRNA of seedlings grown with and without the AM fungus. RT-PCR analysis was used to confirm that Mt-chs1 is induced in plants grown in mini-mycorrhizotrons after 5–10 dpi, when the fungus for the first time contacts the root surface. An enhanced expression of CHS in M. truncatula mycorrhized with Glomus versiforme was described earlier (Harrison & Dixon, 1993). As a hybridization probe the M. sativachs2 gene, which is 94% similar to Mt-chs1, was used and therefore the same gene as described here may have been detected. Moreover, CHS expression was detected by in situ hybridization in the same system at later stages in arbuscule-containing cells (Harrison & Dixon, 1994). The less similar (89%) heterologus probe Ms-chs1 used in this case may have also detected Mt-chs1, indicating that this gene, or a homologue, is continuously up to the arbuscular stage of the symbiosis.

Our experiments showed that Mt-chs1 is expressed at a low level also in noninoculated plants. This could have led some authors to conclude that the transcription of CHS, if at all, is only slightly enhanced (Blee & Anderson, 1996; Mohr et al., 1998). In particular, the presence of the highest levels of CHS transcripts in young elongating roots (McKhann & Hirsch, 1994), localized mainly in the cortical cells (Harrison & Dixon, 1994), could explain the presence of background expression in the noninoculated plants cultivated in the mini-mycorrhizotrons. Alternatively, the low and fluctuating level of CHS in the noninoculated plants may be due to an as yet unidentified environmental factor. Since the expression pattern of Mt-chs1 was similar in plants grown in Multiflor plates and in mini-mycorrhizotrons, it can be excluded that the gene expression pattern is an artefact.

The enhanced expression level of CHS mRNA after treatment of M. truncatula with F. solani f. sp. phaesoli, which also produces symptoms, and the inability to induce the transcript by F. solani f. sp. pisi, which does not produce symptoms, provides evidence that induction of Mt-chs1 is also part of the defence pathway. In fact, F. solani f. sp. phaesoli also induced the expression of the pathogen-related M. truncatula chitinases classes I, II, III-1 and IV, whereas the F. solani f. sp. pisi induced only the class IV chitinase (Salzer et al., 2000). Similarly, F. solani f. sp. pisi could not induce a defence response in the closely related P. sativum (Mohr et al., 1998). Both this and the lack of induction of Mt-chs1 by R. solani indicate that Mt-chs2 is only induced in compatible interactions. Along this line, Mt-chs1 was strongly induced in the mutant dmi1. This mutant exhibits a massive defence response when brought into contact with AM fungi, an event that might prevent a successful fungal penetration, as shown in pea mutants (Gianinazzi-Pearson et al., 1996).

Our study describes the development and validation of a novel culture system for mycorrhizal symbiosis. Mini-mycorrhizotrons represent an attractive tool for molecular and developmental studies of the early stages of symbioses in a gnotobiotic environment. The mini-mycorrhizotrons allowed identification and cloning of a symbiosis-related gene, and the correlation between its expression and the developmental stage of the symbiosis could be established. This tool might be useful to identify further early symbiosis genes.


We thank Dr G. Bécard (Université Paul Sabatier, Toulouse, France) for the dual in vitro culture system, Dr T. Huguet (INRA-CNRS, Castanet-Tolosan, France) for M. truncatula A17 seeds, Dr D. R. Cook (Texas A & M University, College Station, TX, USA) for providing us with the dmi1 mutant, and Daniel Mathys (REM-Labour, University of Basel, Basel, Switzerland) for the ESEM pictures. This work was supported by the Swiss National Science Foundation.