Arbuscular mycorrhizal fungi induce the non-mevalonate methylerythritol phosphate pathway of isoprenoid biosynthesis correlated with accumulation of the ‘yellow pigment’ and other apocarotenoids


*For correspondence (fax +49 345 5582 166; e-mail


Plants and certain bacteria use a non-mevalonate alternative route for the biosynthesis of many isoprenoids, including carotenoids. This route has been discovered only recently and has been designated the deoxyxylulose phosphate pathway or methylerythritol phosphate (MEP) pathway. We report here that colonisation of roots from wheat, maize, rice and barley by the arbuscular mycorrhizal fungal symbiont Glomus intraradices involves strong induction of transcript levels of two of the pivotal enzymes of the MEP pathway, 1-deoxy-D-xylulose 5-phosphate synthase (DXS) and 1-deoxy-D-xylulose 5-phosphate reductoisomerase (DXR). This induction is temporarily and spatially correlated with specific and concomitant accumulation of two classes of apocarotenoids, namely glycosylated C13 cyclohexenone derivatives and mycorradicin (C14) conjugates, the latter being a major component of the long-known ‘yellow pigment’. A total of six cyclohexenone derivatives were characterised from mycorrhizal wheat and maize roots. Furthermore, the acyclic structure of mycorradicin described previously only from maize has been identified from mycorrhizal wheat roots after alkaline treatment of an ‘apocarotenoid complex’ of yellow root constituents. We propose a hypothetical scheme for biogenesis of both types of apocarotenoids from a common oxocarotenoid (xanthophyll) precursor. This is the first report demonstrating (i) that the plastidic MEP pathway is active in plant roots and (ii) that it can be induced by a fungus.


Arbuscular mycorrhizas (AMs) are mutualistic symbioses between a small number of fungal species (Zygomycotina) and the roots of most terrestrial plants (for reviews see Gianinazzi-Pearson 1996; Harrison 1997; Harrison 1999; Smith & Read 1997). AMs can determine plant biodiversity and productivity and are thus important components in sustainable agroecosystems ( Van der Heijden et al. 1998 ). They can also confer bioprotection against root pathogens ( Cordier et al. 1998 ). The respective fungi colonize root cortical cells and improve water uptake and acquisition of mineral nutrients while plants supply the fungi with carbohydrates. Arbuscules, highly branched haustoria-like fungal organs invading the cortical cells, and the plant-derived periarbuscular membrane surrounding them are regarded as the key sites for this bi-directional nutrient exchange.

Two distinct groups of apparent apocarotenoids, i.e. cleavage products of oxygenated C40 carotenoids (xanthophylls), accumulate specifically in plant roots colonized by arbuscular mycorrhizal fungi. The first group comprises glycosylated C13 cyclohexenone derivatives identified from cereals ( Maier et al. 1995 ), other grasses ( Maier et al. 1997 ) and solanaceous plants ( Maier et al. 1999a ; Maier et al. 1999b ). Root pathogens and endophytes as well as abiotic stressors fail to elicit the accumulation of these compounds ( Maier et al. 1995 ). The second type of apocarotenoid is an acyclic C14 polyene compound from maize named mycorradicin, which is a major component of the so-called ‘yellow pigment’ of mycorrhizal maize roots ( Klingner et al. 1995 ). Since the first description of this phenomenon ( Jones 1924), yellow-coloured mycorrhizal roots have been found in many plants ( Bothe et al. 1994 ; Klingner et al. 1995 ).

Isoprenoids are assembled from the C5 monomer isopentenyl diphosphate (IPP). Until recently IPP biosynthesis in plants was thought to proceed exclusively through the cytosolic acetate–mevalonate pathway. There is now extensive evidence for a second, independent route to IPP called the glyceraldehyde 3-phosphate (GAP)/pyruvate pathway, deoxyxylulose phosphate pathway, Rohmer pathway or methylerythritol phosphate (MEP) pathway ( Arigoni et al. 1997 ; Eisenreich et al. 1998 ; Lichtenthaler 1999; Rohmer et al. 1993 ). (The latter designation was agreed upon at the 4th European Symposium on Plant Isoprenoids 1999 in Barcelona and is therefore adopted here.) The MEP pathway is operative in bacteria, green algae and higher plants but has not been described for fungi ( Disch & Rohmer 1998; Eisenreich et al. 1998 ). It resides in the plastid compartment of plants and provides certain classes of isoprenoids, including carotenoids ( Arigoni et al. 1997 ; Fellermeier et al. 1999 ). cDNAs encoding two enzymes central to this pathway, 1-deoxy- d-xylulose 5-phosphate synthase (DXS) and 1-deoxy- d-xylulose 5-phosphate reductoisomerase (DXR), have recently been cloned from plants ( Bouvier et al. 1998 ; Lange & Croteau 1999; Lange et al. 1998 ; Schwender et al. 1999 ). In addition, the evolutionary highly conserved CLA1 gene of Arabidopsis thaliana, whose disruption results in carotenoid deficiency, altered chloroplasts and an albino phenotype ( Mandel et al. 1996 ) is a DXS homolog. Virtually nothing is known yet about developmental or environmental regulation of plant DXS and DXR genes. Very recently, a CTP-dependent enzymatic step downstream of DXR has been identified, which is encoded in the ygbP gene of Escherichia coli. The product of this reaction, 4-diphosphocytidyl-2-C-methylerythritol, was shown to be efficiently incorporated into carotenoids by isolated chromoplasts of Capsicum annuum ( Rohdich et al. 1999 ).

We present here a first account of MEP pathway regulation by mycorrhizal fungi, based on the finding of strong induction of transcript levels of DXS and DXR genes. The correlation of this phenomenon with the appearance of yellow roots and with the accumulation of apocarotenoids leads us to suggest a role for formation and cleavage of carotenoids in AM physiology.


Occurrence of distinct classes of apocarotenoids in yellow mycorrhizal roots of wheat and maize

Expanded clay inoculum of the mycorrhizal fungus Glomus intraradices, modified from previous experiments and containing masses of hyphae, was employed in the present study of isoprenoid metabolic activity in mycorrhizal roots of various cereals. Strongly colonized mycorrhizal wheat roots exhibited a distinct yellow-orange colour and maize roots turned intensely yellow upon mycorrhization ( Fig. 1a,c). HPLC analysis of extracts from these roots showed various cyclohexenone derivatives (numbered peaks in Fig. 1b,d), some of which have been identified previously ( Maier et al. 1995 ; Maier et al. 1997 ). In addition, a broad peak of yellow constituents was observed (marked with an asterisk in Fig. 1b,d). Upon alkaline treatment of these constituents several distinct signals could be resolved. The major yellow component from wheat roots was identified as mycorradicin (10,10′-diapocarotene-10,10′-dioic acid), described previously only for mycorrhizal maize roots ( Klingner et al. 1995 ), by comparison with authentic mycorradicin of HPLC elution profiles, online UV spectroscopical data, LC-MS fragmentation patterns and NMR (data not shown). In short, the negative ion electrospray MS of mycorradicin displayed an [M-H] ion at m/z 247 and fragments at m/z 203 and 159 derived from successive losses of CO2 units from the [M-H]–. Detailed data will be published elsewhere. Some other structurally related components are currently being investigated which are likely to form an ‘apocarotenoid complex’ including other constituents disassembled only upon alkaline treatments.

Figure 1.

Yellow colouration and HPLC analysis for isoprenoid compounds of mycorrhizal roots 6 weeks after sowing.

(a) Control and mycorrhizal wheat plants.

(b) HPLC analysis of wheat root extracts detected at 244 nm. Cyclohexenone derivatives [peaks 1–6] were recognised by their characteristic UV absorption spectra ( Maier et al. 1995 ; Maier et al. 1997 ; Maier et al. 1998 ). Peaks 1–4 were tentatively identified by comparison of retention times with reference compounds as nicoblumin [1], 13-hydroxyblumenin [2], and blumenin [3,4] ( Maier et al. 1995 ; Maier et al. 1997 ; Maier et al. 1999a ; R, glucuronosylglucose), the latter presumably accumulating as R,S-isomers (V. Wray, W. Maier and D. Strack, unpublished results). Minor peaks appearing in controls represent other unknown compounds. A broad peak of yellow constituents, the mycorradicin-containing ‘apocarotenoid complex’ (*) from which mycorradicin [*] was isolated and identified, is marked (R, unknown moieties).

(c) Control and mycorrhizal maize plants.

(d) HPLC analysis of root extracts from maize plants. The detection of cyclohexenone derivatives and indication of yellow mycorradicin-containing ‘apocarotenoid complex’ is as in (b).

Previous attempts to isolate cyclohexenone derivatives from mycorrhizal maize roots had been unsuccessful ( Maier et al. 1997 ). However, the present analysis of strongly colonized maize roots of the dwarf-1 variety gave typical peaks of cyclohexenone derivatives with a characteristic UV-absorption maximum at 244 nm. HPLC characteristics were identical to several wheat cyclohexenone derivatives detected along with the mycorradicin-containing yellow ‘apocarotenoid complex’ ( Fig. 1d). These results, together with new data from tobacco and tomato ( Maier et al. 1999a ; Maier et al. 1999b ), add to the view that both mycorradicins and cyclohexenone derivatives are widespread, if not universal, products of stimulated isoprenoid metabolism in mycorrhizal plant roots.

AM-induced accumulation of transcripts coding for DXS and DXR, enzymes from the MEP pathway involved in carotenoid biosynthesis

We next initiated studies to elucidate the biosynthetic pathway of these isoprenoids. First results have recently been obtained by feeding of 13C-labeled glucose to mycorrhizal barley followed by 13C-NMR spectroscopy of the major cyclohexenone derivative blumenin. These results indicated the formation of cyclohexenone derivatives via the MEP pathway ( Maier et al. 1998 ), which was in agreement with the postulated carotenoid origin of cyclohexenone derivatives but did not exclude supply of metabolites from the shoot. In order to decide whether the accumulation of the cyclohexenone derivatives results from an increased cleavage rate of existing carotenoids or involves de novo synthesis of carotenoids and activation of their biosynthetic machinery in roots, changes in transcript levels of DXS and DXR genes were investigated. The rice EST clones S11559 (GenBank accession number D46713) and S11168(GenBank accession number D46469) are predicted to encode monocot DXS and DXR enzymes. This conclusion is based on (i) very high scores of 88% identity (92% similarity) of a 594 amino acid sequence deduced from S11559 to the CLA1 encoded DXS protein as determined earlier ( Campos et al. 1997 ) and (ii) on an even higher score of 89% and 93% identity (93% and 96% similarity) of a S11168 deduced 165 amino acid sequence stretch to the N-terminal part of mature DXR proteins from Mentha piperita ( Lange & Croteau 1999) and Arabidopsis thaliana ( Schwender et al. 1999 ), respectively. To obtain the latter confirmative data the available S11168 sequence information (GenBank accession number D46469) has been extended by our own analysis (M.H. Walter and D. Strack, unpublished results).

Using the rice DXS and DXR probes on wheat root RNA samples, collected over a period of 8 weeks, RNAs of mycorrhizal roots showed strong hybridisation signals with both probes at the expected sizes of approximately 2.9 kb and 1.8 kb, respectively, while signals in controls were very faint or below detection limits ( Fig. 2a). In agreement with the different sequence similarity scores, highly stringent conditions could be used for hybridisation with the DXR probe, whereas slightly reduced stringency was required for the DXS experiments. Additional data from this experiment ( Fig. 2b,c) will be dealt with below.

Figure 2.

Time course of appearance of DXS and DXR transcripts, apocarotenoids and mycorrhizal structures in wheat roots.

(a) Total RNA gel blots of control and mycorrhizal roots collected over a period of 8 weeks, hybridised with (i) a 0.6 kb PstI fragment encoding DXS from rice EST S11559 ( Campos et al. 1997 ; GenBank accession number D46713) detecting a 2.9 kb transcript under slightly reduced stringency conditions (top); (ii) a 1.8 kb NotI/SalI fragment from rice EST S11168 (GenBank accession number D46469) encoding DXR detecting a 1.8 kb transcript under high stringency conditions (below); and (iii) a 18S ribosomal probe (bottom, used as control for equal loading). Numbers 1–8 indicate weeks after inoculation.

(b) Quantification from mycorrhizal samples of the predominant cyclic apocarotenoid glycoside blumenin ( Maier et al. 1995 ; light columns, 100% = 5.3 μmol g−1 dry weight) and the mycorradicin-containing yellow ‘apocarotenoid complex’ (dark columns, % of maximum). Values represent the average of duplicates. The amounts of metabolites in controls were below detection limit.

(c) Approximate rate of colonisation and ratios of different mycorrhizal structures (hyphae, arbuscules, vesicles and spores). The total height of the columns represents the percentage frequency of colonisation of analysed root segments. The average values from two independent samples each are given. The density of structures per segment (not shown) also increases during AM development, thus adding to the observed effects.

In order to prove that this induction is a general phenomenon and not specific to wheat, roots from maize, rice (the system homologous to the probe) and barley plants were analysed using identical conditions. A transcript species of the same size as the DXS of wheat (2.9 kb) was also strongly induced in mycorrhizal maize roots, whereas levels of a slightly smaller transcript (approximately 2.8 kb) were enhanced during the mycorrhizal state only to a minor extent ( Fig. 3a). Signals of low intensities were also obtained in controls of rice and barley contrasted again by strong hybridisation to RNA samples from mycorrhizal roots. Use of the DXR probe gave similar results ( Fig. 3a). However, only in the case of maize DXS transcripts could two bands, which seem to be differentially regulated, be separated. Separation of white or faintly coloured and intensely yellow root pieces from a mycorrhizal maize root system, followed by RNA blot analysis, indicated a strict correlation of the levels of the 2.9 kb DXS transcript species, but not the 2.8 kb species, with the development of the yellow colour ( Fig. 3b). Such a differential regulation of transcript species could not be shown for DXR under the conditions used.

Figure 3.

Steady state levels of DXS and DXR transcripts in control and mycorrhizal roots of various cereals.

(a) Rice DXS and DXR probes (see Fig. 2) hybridised to total RNA from control (C) and mycorrhizal (M) maize, rice and barley roots (5, 8 and 6 weeks after sowing, respectively).

(b) DXS and DXR probes hybridised to total RNA isolated from white or faintly coloured roots (wr) or intensely yellow roots (yr) from within a mycorrhizal root system of maize. The DXS probe detected two transcript species (approximately 2.9 kb and 2.8 kb) which are apparently differentially regulated.

Correlation of apocarotenoid accumulation, DXS and DXR transcript abundance and arbuscule frequency

To analyse a potential link between the induction of DXS and DXR transcript levels, the appearance of carotenoid cleavage products and the development of mycorrhizal structures, additional data were collected from the time course experiment with wheat ( Fig. 2b,c). The continued rise in DXS and DXR transcript levels during the course of experimentation was accompanied by a strong concomitant accumulation of the two distinct AM-specific apocarotenoids ( Fig. 2b). The variation in metabolite levels and in their ratios in this and in earlier experiments ( Maier et al. 1995 ; Maier et al. 1997 ) and also in the present study (not shown) was always quite high, which is probably due to different rates of recovery of compounds, particularly mycorradicin, from an ‘apocarotenoid complex’ under the non-hydrolytic conditions used here. This is also reflected in the poor resolution after HPLC analysis of yellow constituents ( Fig. 1b,d). For cyclohexenone derivatives, blumenin, the most prominent representative of this class of compounds in gramineous plants ( Maier et al. 1997 ), was quantified.

Upon microscopical inspection of roots stained by trypan blue, a typical, but accelerated, development of AM structures compared to previous experiments ( Maier et al. 1995 ; Maier et al. 1997 ) was observed ( Fig. 2c). A high frequency 2 weeks after inoculation of arbuscules ( Fig. 2c) coincided with a sharp early increase in the DXS and DXR transcript levels ( Fig. 2a). Subsequent symbiotic stages were characterised by continued synthesis and degradation of arbuscules and by the appearance of vesicles and spores.


We have provided in this paper the first molecular evidence for AM-specific regulation of the MEP pathway of IPP biosynthesis, which is, to our knowledge, the first report on any microbe-related activation of this pathway. This induction seems to be restricted to the mutualistic symbiotic interaction analysed in our work since the cyclohexenone apocarotenoid metabolites were shown to not accumulate in response to root pathogens, endophytes or abiotic stressors ( Maier et al. 1995 ). Furthermore, a yellow pigment was not observed in pathogen-infected roots ( Bothe et al. 1994 ). Experiments are underway to confirm this point at the transcript level. In pathogenic or parasitic interactions the classical acetate-mevalonate pathway is activated, e.g. in solanaceous plants, leading to accumulation of sesquiterpenoid phytoalexins ( Chappell 1995; Westwood et al. 1998 ). Present information on DXS gene and transcript regulation is limited to aerial plant parts ( Mandel et al. 1996 ) and specialized monoterpene essential oil production systems such as the peppermint leaf trichome gland cells ( Lange et al. 1998 ). No information exists on DXR regulation.

The possible complexity of plant DXS and DXR gene families, which might comprise unique symbiosis-regulated members as shown for other gene families ( Miller et al. 1998 ), is unknown. Preliminary results on the possible existence of specific transcript species in yellow mycorrhizal maize roots ( Fig. 3b) need to be confirmed and substantiated by sequence data. The 2.9 kb DXS transcript species appeared only after development of the yellow root colour and might represent an isoform which is regulated in the corresponding later stages of mycorrhization. In contrast, the 2.8 kb DXS and the DXR transcripts were detectable at low to intermediate levels in controls but were upregulated early in mycorrhization, i.e. in the white roots of the mycorrhizal maize root system. These newly developing white roots were largely devoid of the yellow colour and of arbuscules but already contained mycorrhizal structures (hyphae, appressoria).

1-Deoxy- d-xylulose can also serve as a precursor of the enzymic co-factors thiamin and pyridoxol, but 2-C-methyl- d-erythritol 4-phosphate is an intermediate committed to the MEP pathway. Whereas the magnitude of DXS transcript accumulation already makes a mere involvement in biosynthesis of these co-factors very unlikely, the strong induction of transcript levels of the committed enzymic step catalysed by DXR in the system homologous to the probe under high stringency conditions convincingly demonstrates activation of the MEP pathway by mycorrhizal fungi.

Cellular alterations in cortical cells invaded by fungal arbuscules include fragmentation of the vacuole, migration of the nucleus to the centre of the cell and an increase in the number of mitochondria, plastids, dictyosomes and endoplasmatic reticulum ( Gianinazzi-Pearson 1996; Harrison 1999). The AM-induced activation of the plastid-located MEP pathway may be related to these observations and to earlier reports on unusual plastids in mycorrhizal roots, which resemble chromoplasts and always occur close to fungal arbuscules ( Dexheimer et al. 1990 ; Scannerini & Bonfante-Fasolo 1977). DXS and DXR transcript accumulation might be a marker for such a transition in root plastid development to parallel chromoplast formation for carotenoid biosynthesis specifically involved in AM physiology.

Based on their concomitant accumulation and apparent nature as carotenoid cleavage products, we propose a biogenetic scheme for both cyclohexenone derivatives and mycorradicin ( Fig. 4). It is postulated that both classes of compounds originate from a common oxygenated C40 carotenoid (xanthophyll) such as lutein or zeaxanthin, which are known to be synthesized through the MEP pathway ( Arigoni et al. 1997 ). Oxidative cleavage of such carotenoids may yield a C13 cyclohexenone derivative and a C27 intermediate, an azafrin-like structure ( Kuhn & Brockmann 1935). The intermediate may then be further metabolized to a second C13 cyclohexenone derivative and a C14 polyene compound (mycorradicin), the latter of which corresponds to the central portion of the original carotenoid precursor, as proposed previously ( Bothe et al. 1994 ). Both types of apocarotenoids seem to be part of a complex multi-component system, since alkaline treatment not only releases mycorradicin but also additional C13 cyclohexenone derivatives (T. Fester and D. Strack, unpublished results). This may explain the considerable variation in the amount of apocarotenoids isolated from mycorrhizal roots ( Fig. 2b) and the lack of a 2 : 1 ratio for cyclic versus acyclic cleavage products as expected from the biogenetic scheme. Furthermore, additional complexity of C13 cyclohexenone derivatives is becoming obvious through new data from mycorrhizal tobacco and tomato ( Maier et al. 1999b )

Figure 4.

Hypothetical scheme for biogenesis of C13 cyclohexenone derivatives and C14 mycorradicin.

Proposed biosynthesis of a common xanthophyll precursor (e.g. lutein) through the MEP pathway, represented by two of its pivotal enzymes, DXS and DXR, followed by oxidative cleavage of the carotenoid backbone to yield one C14 and two C13 cleavage products (R, see Fig. 1).

A similar co-occurrence of cyclic C13 and acyclic C14 apocarotenoids with unknown ratios has been reported for fruits, e.g. starfruit (Averrhoa carambola L.), in which these compounds contribute to the flavour as volatiles or semi-volatiles ( Lutz & Winterhalter 1992; Winterhalter & Schreier 1995). Cyclic C15 products, closely related to abscisic acid precursors, have been identified in fruits as well. They might further be metabolized into C13 compounds ( Winterhalter & Schreier 1995). A cDNA encoding a dioxygenase directly involved in abscisic acid (C15) biosynthesis through oxidative cleavage of the 11,12 (11′,12′) double bonds of C40 carotenoids has been cloned utilizing the maize viviparous mutant VP14 ( Tan et al. 1997 ). Related enzymes with regioselectivity directed towards the 9,10 (9′,10′) double bond giving direct rise to C13 derivatives might also exist. A VP14-related cDNA has recently been isolated from a subtraction library of an myc mutant devoid of arbuscules and wild-type mycorrhizal pea roots (P. Franken, personal communication), thus lending additional support to the carotenoid cleavage hypothesis.

Whether carotenoids (xanthophylls) themselves are functional or are instantly cleaved is an open question. It has been shown that exogenous application of the apparent cleavage product blumenin strongly inhibits early fungal colonisation and arbuscule formation, implicating that cyclohexenone derivatives might act in the plant control of fungal spread ( Fester et al. 1999 ). Current experiments deal with identification and localisation of still hypothetical xanthophylls as precursors in mycorrhizal roots. Xanthophylls can be rigidifying components of plant thylakoid membranes owing to their dipolar nature ( Havaux 1998). A similar function in the periarbuscular membrane would be conceivable. Individual arbuscules undergo rapid turnover with an average lifetime of 4–10 days after which they completely disappear from the infected root cell ( Alexander et al. 1988 ). Accumulation of both mycorradicin and cyclohexenone derivatives might thus be the result of arbuscule senescence. Such an origin would not exclude the control function suggested above. Alternatively, the cleavage products themselves could be integrated in or associated to membranes or plant/fungus interfaces. For example, the glycoside esters of unusual (apo)carotenoids of thermophilic bacteria are bound to fatty acids and are speculated to play a role in membrane stabilisation ( Burgess et al. 1999 ).

In summary, this paper opens up new perspectives to direct further research (i) into novel functions of root plastids; (ii) into AM-mediated regulation of carotenoid biosynthesis and degradation; and (iii) into the mechanisms of arbuscule formation and degradation.

Experimental procedures

Plant cultivation and fungal root colonisation

Triticum aestivum L. cv. Caprimus, Zea mays L. cv. dwarf-1 ( Schneider et al. 1992 ), Oryza sativa L. cv. Koshihikari and Hordeum vulgare L. cv. Salome plants were grown in expanded clay in plastic pots in a growth chamber under a 16 h 400 μm−2 sec−1 light regime at 18°C night/23°C day. Fertilization was performed once a week with Long Ashton solution as described previously ( Maier et al. 1995 ) except that phosphate was reduced to 10% of the normal levels. Fungal inoculum [Glomus intraradices Schenck & Smith, isolate 49 (courtesy of Dr von Alten, Hannover, Germany)] consisted of fungal propagules in expanded clay, which was highly enriched in hyphae by repeated growth (up to four times) of wheat and maize plants in the same substrate for periods of up to 6 weeks. This procedure afforded a ‘high density fresh hyphae inoculum’ which resulted in a more rapid and synchronous colonization of host roots (typically 50–60% of root segments colonised after 2–3 weeks) compared to previous methods ( Maier et al. 1995 ; Maier et al. 1997 ). Rates of colonization and ratios of the various mycorrhizal structures were evaluated microscopically after staining roots with trypan blue as described by Maier et al. (1995) .

Isolation and quantification of isoprenoid compounds by HPLC

Freeze-dried roots were homogenized by grinding with solid CO2 in a mortar and extracted three times with CH2Cl2, once with methanol and twice with 80% aqueous methanol. The combined CH2Cl2-extracts and the combined methanolic extracts were evaporated at 40°C in vacuum to dryness, the residues redissolved in THF (for CH2Cl2-extracts) and 80% aqueous methanol, respectively. Cyclohexenone derivatives and the yellow pigment-containing ‘apocarotenoid complex’ were present in the methanolic extracts which were centrifuged and subjected to HPLC analyses.

The HPLC system (Waters 600, Milford, MA, USA), the Nucleosil C18 column (Macherey-Nagel, Düren, Germany) and quantification of cyclohexenone derivatives by using ABA as the external standard were as described previously ( Maier et al. 1995 ). In order to optimize the separation of the compounds of interest we used a two-step linear gradient. The first step achieved the separation of cyclohexenone derivatives running from 5 to 20% solvent B (acetonitrile) in solvent A (1.5% phosphoric acid in water) within 29 min at a flow rate of 1 ml min−1 and the following second step eluted the yellow pigment-containing ‘apocarotenoid complex’ by reaching 80% solvent B in solvent A within 18 min. The amount of the ‘apocarotenoid complex’ was estimated as the percentage of the maximal peak areas (100%) reached from extracts of mycorrhizal roots.

Identification of mycorradicin

Mycorradicin was identified by LC-MS analysis of the ‘apocarotenoid complex’ of mycorrhizal wheat roots after treatment with 0.3 m aqueous KOH for 30 min at 30°C. The negative ion electrospray (ES) mass spectra were obtained from a Finnigan MAT TSQ 7000 instrument (electrospray voltage 3.5 kV, CID offset voltage 10 V; heated capillary temperature 220°C; nitrogen as sheath gas) coupled with a Micro-Tech Ultra-Plus MicroLC system equipped with a RP18 column (4 μm; 1 × 100 mm, Ultrasep). For the HPLC, a linear gradient system was used from water-acetonitrile 4 : 1–1 : 9 (containing 0.2% acetic acid) within 15 min; flow rate 70 μl min−1. The MS fragmentation of mycorradicin with Rt 13.6 gave the following signals [m/z (rel. int.)]: 247 ([M-H], 100), 203 ([M-H-CO2], 32), 159 ([M-H-2CO2], 5). Furthermore, UV spectrum and HPLC elution profile were identical to the mycorradicin, provided by H. Bothe (Köln, Germany). Mycorradicin was a major component of the ‘apocarotenoid complex’, along with some other yellow components exhibiting similar UV-spectra but different HPLC retention times (detection at 400 nm).

RNA gel blot analysis

Total RNA was analysed using the NorthernMax system of AMBION® according to the manufacturer (Ambion Inc., Austin, TX, USA). Hybridisations were performed in 1 m NaCl, 1% (w/v) SDS, 10% (w/v) dextransulfate, 100 μg ml−1 salmon sperm carrier DNA at 65°C (high stringency) or 63°C (slightly reduced stringency) overnight. Final washes were in 0.1× SSC, 0.5% SDS (high stringency) or 2× SSC, 0.5% SDS (slightly reduced stringency) at 60°C. A 32P-labeled PstI fragment from the central portion of the rice S11559 cDNA (position 586–1192 of GenBank accession number D46713) or a 1800 bp NotI/SalI fragment from the S11168 cDNA (GenBank accession number D46469) were used as probes.


We thank T. Sasaki (MAAF DNA Bank, Japan) for providing the S11559 and S11168 cDNAs; J. Schmidt and V. Wray for performing LC-MS and NMR analyses, respectively; H. Bothe for providing authentic mycorradicin; M. Kaldorf for communicating the mycorradicin purification procedure; H. von Alten for providing the Glomus intraradices isolate 49; and M. Rauscher for critically reading the manuscript. The skilful technical assistance of K. Manke is gratefully acknowledged. This work was supported by the Deutsche Forschungsgemeinschaft and the Fonds der chemischen Industrie.


  1. GenBank accession numbers D46713and D46469.