Author for correspondence: Pascale M. A. Seddas Tel: +33 (0)3 80 69 35 64 Fax: +33 (0)3 80 69 37 53 Email: firstname.lastname@example.org
• Mechanisms of root penetration by arbuscular mycorrhizal (AM) fungi are unknown and investigations are hampered by the lack of transformation systems for these unculturable obligate biotrophs. Early steps of host infection by hemibiotrophic fungal phytopathogens, sharing common features with those of AM fungal colonization, depend on the transcription factor STE12.
• Using degenerated primers and rapid amplification of cDNA ends, we isolated the full-length cDNA of an STE12-like gene, GintSTE, from Glomus intraradices and profiled GintSTE expression by real-time and in situ RT-PCR. GintSTE activity and function were investigated by heterologous complementation of a yeast ste12Δ mutant and a Colletotrichum lindemuthianum clste12Δ mutant.
• Sequence data indicate that GintSTE is similar to STE12 from hemibiotrophic plant pathogens, especially Colletotrichum spp. Introduction of GintSTE into a noninvasive mutant of C. lindemuthianum restored fungal infectivity of plant tissues. GintSTE expression was specifically localized in extraradicular fungal structures and was up-regulated when G. intraradices penetrated roots of wild-type Medicago truncatula as compared with an incompatible mutant.
• Results suggest a possible role for GintSTE in early steps of root penetration by AM fungi, and that pathogenic and symbiotic fungi may share common regulatory mechanisms for invasion of plant tissues.
Interactions between symbiotic Glomeromycota fungi and plant roots are relatively nonspecific, permanent and mutualistic (Schüssler et al., 2001; Strack et al., 2003). The resulting association, arbuscular mycorrhiza (AM), is the most widespread root symbiosis in nature (Wang & Qiu, 2006), and may have played a major role in land colonization by plants (Redecker et al., 2000; Krings et al., 2007). Establishment of the AM symbiosis comprises several well-defined stages (Gianinazzi-Pearson, 1996; Harrison, 1998; Hause & Fester, 2005), which include spore germination and development of presymbiotic hyphae, appressoria formation upon contact with a host root, penetration of the outer root cell layer and differentiation of highly branched haustoria (arbuscules) within cortical cells where nutrient exchange between the symbionts occurs.
The development of this mutualistically beneficial relationship must result from coordinated genetic programmes in both partners and be driven, at each stage, by reciprocal signalling events. Whilst an extensive list of genes has been reported to be modulated within the established symbiosis (Balestrini & Lanfranco, 2006; Küster et al., 2007), few data are available about cellular and molecular events essential to fungal morphogenesis and root penetration during the early stages of AM interactions. Only a limited set of genes have been identified in plants which could be implicated in the perception of fungal signals necessary for the activation of signal transduction pathways (Kosuta et al., 2003, 2008; Weidmann et al., 2004; Sanchez et al., 2005). In M. truncatula, these are DMI1, DMI2/MtSYM2 and DMI3/MtSYM13 (Endre et al., 2002; Anéet al., 2004; Lévy et al., 2004; Mitra et al., 2004).
Several fungal proteins potentially implicated in the perception of the root epidermis, such as calcium signalling proteins, and a 14-3-3-like regulator that is also induced during appressorium formation in the rice pathogen Magnaporthe grisea (Takano et al., 2003), have been identified in the AM fungus Glomus mosseae (Breuninger & Requena, 2004). However, the function of the majority of AM fungal genes isolated so far could only be deduced from sequence comparison and transcriptional analysis. Their functional characterization is hampered by the fact that mutant screening and stable transformation systems do not exist for Glomeromycota (Forbes et al., 1998; Harrier & Millam, 2001; Helber & Requena, 2008). Heterologous studies in yeast have been used to investigate gene function (Harrison & van Buuren, 1995; Lanfranco et al., 2005), but yeast complementation with AM fungal genes can sometimes be problematic (Aono et al., 2004). Transformable hemibiotrophic fungal pathogens (Colletotrichum spp., M. grisea), whose development shares several common features with the AM fungal lifecycle (spore, appressoria or haustoria-like structure formation, Perfect et al., 1999), could constitute alternative candidates for the characterization of AM fungal genes related to root colonization and interactions with plant tissues (Hahn & Mendgen, 2001). In addition, hemibiotrophic fungal pathogens and Glomeromycota symbionts can trigger close genetic responses in their host plants (Güimil et al., 2005).
We have isolated the first STE12-like gene (GintSTE) from an AM fungus (G. intraradices) and begun its functional characterization. Transcriptional activation of GintSTE coincides with fungal penetration of roots by G. intraradices. Its widespread occurrence in Glomeromycota and its capacity to complement the penetration defect of a C. lindamuthianum clste12Δ mutant has led us to hypothesize that GintSTE could play a role in AM fungal penetration of host tissues, and that STE12 could be a common genetic feature regulating penetration of plant tissues by both symbiotic and pathogenic fungi.
Materials and Methods
All the sequences of the primers cited in the following sections are listed in Supporting Information, Table S1.
Fungal and plant materials
Spores of Glomus intraradices BEG141, G. mosseae BEG12, G. geosporum BEG11, G. claroïdeum BEG31, G. caledonium BEG20, G. versiforme BEG47, Gigaspora margarita BEG34 and Acaulospora laevis BEG13 were provided by the International Bank of Glomeromycota (IBG, Dijon, France). Sporocarps of G. versiforme were collected at the pot surface and spores were separated by vigorous shaking in the presence of glass beads. Spores of G. intraradices were extracted by laceration of colonized roots in sterile water using a blender. Spores of other fungi were collected from leek pot cultures by wet sieving. Genomic DNA was extracted from spores collected by hand. Spores of G. intraradices were also used for RNA extraction and spore germination experiments.
Surface-disinfected spores (10 min in 2% chloramine T, 0.02% streptomycin, 0.02% gentamycin and 20 µl Tween 20) of G. intraradices (2000) were incubated at 2% CO2 in the presence of 3 ml water, or in root exudates directly obtained by placing five seedlings of wild-type Medicago truncatula Gaertn. cv. Jemalong line J5 or the mycorrhiza-defective dmi3-1/Mtsym13-1 mutant genotype in sterile water (Seddas et al., in press). After 7 d incubation under constant conditions (420 µmol m−2 s−1, 16 : 8 h day : night, 24 : 19°C, 70% humidity), germinated spores were used for RNA extraction or direct fluorescent in situ RT-PCR experiments.
Three independent experiments were performed with the J5 and dmi3-1/Mtsym13-1 genotypes. Seeds were surface-sterilized, germinated on 0.7% bactoagar and then seedlings were transplanted into a sterilized mix of Terragreen (OilDri-US special, Mettman, Germany) and neutral γ-irradiated clay soil (2 : 1, v/v). To initiate root system development and homogenize plant size, plants were precultured 15 d under constant conditions (see earlier) and supplied 5 ml Long Ashton solution (without phosphate, double quantity of nitrate) twice a week (Hewitt, 1966). Plants were then transplanted into a fresh terragreen/γ-irradiated soil mix (2 : 1, v/v) for uninoculated plants, or into a mix of Terragreen and soil-based inoculum of G. intraradices BEG 141 (2 : 1, v/v) for inoculated plants. Plants were harvested each day from 1 to 9 d after inoculation (dai), roots were washed and immediately stored in liquid nitrogen to extract RNAs, or stained with trypan blue to determine the number of appressoria or the AM fungal developmental stage on randomly sampled root pieces.
The Briosi and Cav. race β strain of Colletotrichum lindemuthianum and clste12Δ (B77), a disruption mutant in which part of the CLSTE12 coding sequence has been replaced by a phleomycin resistance cassette (Wong Sak Hoi et al., 2007), were cultured on Bannerot agar medium at 23°C in the dark.
Colletotrichum lindemuthianum infection assays were carried out on the susceptible cultivar P12S of Phaseolus vulgaris L. (see later discussion). Seeds were germinated on moist filter paper for 36 h in the dark. Seedlings were grown on sand under constant conditions (16 : 8 h day : night, 24 : 19°C, and 80 : 40% humidity) with a nutrient supply.
Yeast strains with a Σ1278b background, JCY100 (MATa leu2Δ::hisG his3Δ::hisG trp1Δ::hisG ura3-52) and the derivative JCY600 (ste12Δ::LEU2) were grown on standard YPD and minimal growth media.
Isolation of an STE12 coding sequence
Two degenerate primers, STF and STR, were designed according to the amino acid residues DALERDL and RLEHLKR conserved among STE12 homologues of C. lindemuthianum, C. lagenarium, C. parasitica, F. graminearum, M. grisea, P. marneffei, N. crassa and A. nidulans, and using fungal preferential codon usage (http://mekentosj.com/). Because Glomeromycota have an exceptionally low GC content (c. 30%, Hosny et al., 1997), AT nucleotides were preferentially chosen, where possible, to generate primers with a limited amount of degeneration (64). These primers were used for PCR amplification from G. intraradices BEG141 cDNAs with a proofreading DNA polymerase (Platinum®Taq DNA polymerase High Fidelity; Invitrogen Corporation, Carlsbad, CA, USA). Products were cloned and sequenced. Clones contained an identical 837 pb insert that was homologous to STE12 from filamentous fungi. The complete sequence of GintSTE cDNA was obtained by 5′ and 3′ rapid amplification of cDNA ends (3′ 5′ RACE) using the Generacer kit (Invitrogen) according to the manufacturer's instructions. Gene-specific primers Gi5rev1 and Gi3rev1 were used in conjunction with GeneRacer primers for PCR amplification. The full-length cDNA sequence was analysed using the UTRscan online program (http://www.ba.itb.cnr.it/BIG/UTRScan/). The presence of regulation sites in the predicted aa sequence was investigated using the PROSITE (http://www.expasy.ch/prosite/) and SUMOsp 2.0 (http://bioinformatics.lcd-ustc.org/ sumosp/ prediction.php) online programs.
GiSTgF- and GiSTgR-specific primers matching the 5′ and 3′ ends of the coding sequence were used to amplify GintSTE from G. intraradices BEG141 genomic DNA.
STE12 fragments were amplified from sporal genomic DNA of the other seven AM fungi using a nested PCR strategy. The region coding the first helix of the homeodomain and the region coding the NLS, the Asn-rich region and the double zinc finger were amplified using primer pairs HD1for/rev HD2for/rev and DZ1for/rev DZ2for/rev, respectively.
Southern blot analysis
Two probes were used to analyse C. lindemuthianum transformants. One matching part of the CLSTE12 promoting region was amplified from genomic DNA using the primers LB-clste12s/LB-clste12r; the other matching part of the middle region coding sequence of GintSTE was amplified from spore cDNA with the primers GintSTEsens/GintSTErev. Probes were labelled with 32P-CTP. Hybridization was performed on 10 µg C. lindemuthianum genomic DNA digested with 75 U of BamHI.
All amplifications were performed using proofreading DNA polymerase. To construct the yeast complementation vector pYES2-GintSTE, the complete GintSTE coding sequence was amplified from GintSTE cDNA using the primers Bam_For/Eco_Rev and cloned into the shuttle vector pYES2/CT (Invitrogen), between BamHI and EcoRI sites. To construct the C. lindemuthianum complementation vector pDHt-GintSTE, the GintSTE coding sequence was amplified (1566 pb) using BshBst_For/BshBst_Rev primers bearing, respectively, a BshAI and a BstAPI restriction site. A short fragment (116 pb) corresponding to the end of the CLSTE12 promoter was also amplified from a pGEM-T subclone containing the CLSTE12 gene sequence (pGEMt-CLSTE12), using primer PshBsh_For/PshBsh_Rev including a BshAI restriction site. The two PCR products were digested by BshAI and ligated together. The resulting fragment (1689 pb) was digested by PshAI and BstAPI and cloned into pGEMt-CLSTE12 in which the CLSTE12 coding sequence had been removed. The chimeric sequence in the resulting subclone (pGEMt- GintSTE) was then transferred to the pDHt vector: the complete CLSTE12 gene sequence was removed from pDHt-CLSTE12 (HindIII and XbaI digestions), the fusion gene containing the CLSTE12 promoter and terminator surrounding the GintSTE coding sequence was amplified with the primers fusion_For/fusion_Rev, and the resulting fragment was transferred to pDHt using the In-fusion 2.0 Dry-Down PCR cloning kit (Clontech, Mountain View, CA, USA). All plasmids were multiplied in E. coli.
Yeast transformations were performed using a protocol adapted from Hill et al. (1991). JCY100 (wild type) and JCY600 (ste12Δ) strains received either the pYES2-GintSTE vector or the pYES2 empty vector, both containing a URA3 auxotrophic marker. Transformants were selected on a minimal medium supplemented with the appropriate amino acids (His, Trp, Leu for JCY100; His, Trp for JCY600) and lacking uracil.
Agrobacterium tumefaciens-mediated transformation of C. lindemuthianum
Conidia of C. lindemuthianumβ (wild type) and B77 (clste12Δ) strains obtained from 7-d-old cultures were transformed using the A. tumefaciens strain AGL1 bearing either the expression vector pDHt-GintSTE or the pDHt empty vector, as described by Wong Sak Hoi et al. (2007).
Yeast invasive growth assay
The control medium for invasive growth tests was YPD medium, and it was used according to Roberts & Fink (1994). The same medium containing 2% galactose and 1% raffinose instead of 2% dextrose was used as an inducive medium for the expression of GinSTE, under the control of the GAL1 promoter in the pYES2 vector. Cells were spread on both media and grown at 30°C for 3 d. Plates were washed with water to remove cells that did not invade agar. Three independent assays were performed.
Infection assays were carried out according to Wong Sak Hoi et al. (2007) by spotting 5 µl droplets of a conidial suspension adjusted to a concentration of 106 conidia ml−1 on excised cotyledonary P. vulgaris leaves. The leaves were placed in Petri dishes under high relative humidity at 23°C in the dark. Anthracnose lesions were observed 5–7 dai. Infection assays were repeated in two independent experiments.
Gene expression monitoring
Real-time RT-PCR C. lindemuthianum total RNA was prepared using the SV Total RNA Isolation kit (Promega, Madison, WI, USA). Yeast total RNA was extracted using the method described by Ausubel et al. (2003). In each case, cDNAs were generated by reverse transcription with an M-MLV reverse transcriptase and oligo(dT)15 primers (Promega).
Total RNA from roots of M. truncatula J5 and dmi3-1/Mtsym13-1 inoculated or not with G. intraradices was extracted using the method of Franken & Gnädinger (1994). Replicate extractions of RNA were performed in each case on material from three independent experiments. Total RNA from G. intraradices spores was prepared with the SV total RNA Isolation kit. cDNAs from inoculated roots or G. intraradices spores were synthesized by a SMART approach (Endege et al., 1999; Vernon et al., 2000) using gene-specific primers and ensuring that cDNA synthesis was in the exponential phase.
Transcripts were quantified using the ABsolute™ QPCR SYBR® green ROX mix (ABgene, Epsom, UK). For each sample, PCR reactions were carried out in technical triplicates using 1 µl cDNA as template, in a final volume of 20 µl containing 1X SYBR green mix, and 20 nm of each gene-specific primers: GintST1/GintST2 and TEF_f3/TEF_r2 corresponding to the constitutively expressed reference gene coding the translation elongation factor α subunit (TEF) (Gianinazzi-Pearson et al., 2006). To calculate the number of transcripts present in original samples, TOPO plasmids containing each amplicon were quantified by UV absorbance spectroscopy and linearized by NotI digestion. Standard amplification curves were determined from duplicate samples of plasmid DNA at 102, 103, 104, 105, 106 and 107 copies. To verify amplification of each target cDNA, a melting-curve analysis was included at the end of each PCR run. The generated data were analysed by SDS 2.2 software (Applied Biosystems, Foster City, CA, USA). Target-gene expression data from real-time RT-PCR were plotted as 2(40-CT)/10, as described in Czechowski et al. (2004), and normalized against the reference TEF gene (Seddas et al., in press).
Expression of the GintSTE transgene, and that of the marker genes FLO11 and PGUI revealing STE12 activity in yeast transformants, was determined by RT-PCR using primer pairs GintSTEsens/GintSTErev, FLO11sens/FLO11rev and PGUIsens/PGUIrev, respectively (Wong Sak Hoi et al., 2007).
In situ RT-PCR Direct localization of GintSTE expression by in situ RT-PCR was performed according to Seddas et al. (2008). Briefly, isolated spores of G. intraradices or root fragments of M. truncatula J5 inoculated with G. intraradices (5–7 dai) were collected, fixed and digested with pectinase and chitinase to permeabilize fungal and plant cell walls. After elimination of proteins and genomic DNA, cDNAs were synthesized by reverse transcription with the M-MLV reverse transcriptase and specific primers GintST2 for GintSTE RNAs or 8.24 for the 25S-LSU control RNAs. GintSTE and 25SrRNA cDNAs were then amplified with the specific primers GintST1/GintST2 and FLR3/8.24 (Gollotte et al., 2004; Farmer et al., 2007) 5′-labelled with Texas red. Samples were postfixed in 100% ethanol, rehydrated and deposited on slides in anti-fading medium. Fluorescent labelled amplified cDNA was detected using a confocal microscope (excitation, 594 nm; emission, 606–640 nm). Fungal structures accumulating specific transcripts were determined by comparison with a Nomarski picture.
Isolation of a STE12 gene homologue from G. intraradices
A fragment of 837 pb was amplified from cDNA of germinating spores by the PCR-based strategy to isolate the STE12 homologue from G. intraradices, using degenerate primers matching conserved domains. Sequence analysis of the full-length cDNA (2508 pb) completed by 5′ 3′ RACE revealed an open reading frame (ORF) of 1563 pb with a low GC content (39%), characteristic of AM fungal genes (Hosny et al., 1997). The complete transcribed region was amplified from genomic DNA. Comparisons between genomic DNA and cDNA sequences revealed that GintSTE is intronless. Southern blot analysis showed that the GintSTE gene is present as a single copy in the genome of G. intraradices (Fig. S1).
The deduced amino acid (aa) sequence of GintSTE was relatively short (521 aa) compared with other STE12 (mean size of 700 aa) and contained four functional domains: two potential DNA-binding domains, including a homeodomain in the N-terminal region (aa 27–167) and a couple of Cys2His2 zinc fingers in the C-terminal region (aa 396–445), a nuclear localization signal (NLS) (aa 354–360) and an asparagine-rich region (aa 366–377) (Fig. 1). The global amino acid similarity with STE12 from other organisms did not exceed 50%, the highest being with a homologue from Gibberella zeae (46.5% similarity; 35.8% identity). STE12 from other filamentous fungi (M. grisea, Colletotrichum spp., Cryphonectria parasitica, Penicillium marneffei, Aspergillus fumigatus) shared c. 45% of similarity with GintSTE. Less similarity was found for yeast homologues: S. cerevisiae (31.7%), Kluyveromyces lactis (35%), Candida albicans (36%).
Homologies between GintSTE and other STE12 proteins tended to be restricted to the DNA binding domains, especially the homeodomain, comprising three different helixes, which is present in all STE12 family proteins so far identified (Fig. S2). This GintSTE homeodomain is > 80% identical to STE12 from filamentous fungi (Colletotrichum spp., Cryphonectria parasitica, Fusarium graminearum, M. grisea, Neurospora crassa, Penicillium marneffei) and c. 65% identical to STE12 from the yeasts C. albicans and S. cerevisiae. Contrary to the homeodomain, the zinc finger motifs are exclusively conserved among STE12 homologues of filamentous fungi. In this region, GintSTE shows > 80% identity with SteA (Aspergillus fumigatus), StlA (P. marneffei), MST12 (M. grisea), NcSTE12 (N. crassa), CST1 (C. lagenarium), FST12 (F. graminearum), CLSTE12 (C. lindemuthianum) and Cpst12 (C. parasitica). Two overlapping NLS (PTYQRR and RRRRA) closely related to those existing in filamentous fungi (PTYKQRR and RRRRS) were also found in GintSTE. The central region, localized between the homeodomain and the zinc fingers, was shortened (225 aa) compared with homologues from filamentous fungi (366 aa) and some stretches of conserved amino acids were lacking, such as the sequence EPAYIANEETGLYTAIP (e.g. aa 353–369 in CLSTE12), which is identical in MST12, StlA, SteA, NcSTE12 CLSTE12 and CST1 (Park et al., 2002). The G. intraradices protein also contained an asparagine-rich region (11 asparagine residues among 15) in the same position (downstream of the NLS) as the glutamine-rich region of STE12 factors of Colletotrichum spp. that is considered to play a role in transcription activation (Gerber et al., 1994). Several post-translational modification sites were also predicted, including 16 putative phosphorylation sites, two putative SUMOylation sites (Müller et al., 2001), nine N-myristoylation sites and six ASN-linked glycosylation sites.
STE12 homologues in different AM fungi
To determine whether the STE12 gene existed in other AM fungi, PCR amplifications were performed on genomic DNA of fungi belonging to five different families of the Glomeromycota: G. mosseae and G. geosporum belonging to the Glomeraceae 1 (G. intraradices also belongs to this family), G. claroïdeum belonging to the Glomeraceae 2, G. versiforme from the Diversisporaceae, Gi. margarita from the Gigasporaceae and Ac. laevis from the Acaulosporaceae. Two STE12 fragments were successfully amplified for each isolate, one corresponding to the region coding the first helix of the homeodomain (Fig. 2a) and one corresponding to the region coding the NLS, the Asn-rich region and the double zinc finger (Fig. 2b), suggesting that STE12 is a common genetic feature among the Glomeromycota.
GintSTE restores agar penetration of a noninvasive yeast ste12Δ mutant
To confirm that GinSTE encoded a functional transcription factor, its ability to complement a yeast ste12Δ mutant was tested. Since transition to invasive growth depends on the STE12 transcription factor, an agar penetration assay was used to assess GintSTE functionality. The yeast ste12Δ strain (JCY600) was transformed with the expression vector pYES2-GintSTE containing the entire GintSTE coding sequence under the control of a galactose-induced promoter (GAL1). The same strain and the corresponding wild-type strain (JCY100) transformed with the empty vector, were used as control strains. Resulting yeast transformants were assessed for invasive growth on a medium either inducing (galactose medium) or repressing (glucose medium) GintSTE expression (Fig. 3).
On the galactose-inducive medium, expression of the transgene GintSTE was able to restore invasive growth in the complemented mutant strain and to activate genes regulating yeast invasive growth, such as PGU1 and FLO11 (Madhani et al., 1999), at expression levels similar to those observed in the wild-type JCY100 (data not shown).
GintSTE expression in germinating spores and at the penetration stage of AM interactions
To investigate the role of GintSTE in AM symbiosis establishment, transcript abundance was profiled during the pre-symbiotic (germinating spore) and early stages of interactions between G. intraradices and M. truncatula roots. GintSTE expression was detected in both germinated and nongerminated spores. However, transcript abundance relative to the reference gene TEF was significantly higher (P = 0.05) in the former (19.7 ± 0.74) than in the latter (5.1 ± 0.55). No significant differences were observed between spores germinated in the presence of water (19.2 ± 0.14) or root exudates from either wild-type (J5) (21 ± 0.025) or mutant dmi3-1/Mtsym13-1 (20 ± 0.052) genotypes of M. truncatula.
GintSTE expression patterns were then analysed during interactions between G. intraradices and roots of M. truncatula wild-type J5 and the mutant dmi3-1/Mtsym13-1, which prevents fungal penetration (Fig. 4). In the case of the compatible interaction between J5 roots and G. intraradices, the first appressoria formed at 3 dai, the fungus penetrated roots 5 dai and the first arbuscules were observed 7 dai (Fig. 4a). During the incompatible interaction between G. intraradices and the dmi3-1/Mtsym13-1 mutant, appressoria formed at 3 dai, but the fungus failed to penetrate roots (Fig. 4b). With repeated attempts of root colonization, about twofold more appressoria developed on dmi3-1/Mtsym13-1 roots than on J5 roots from 5 to 9 dai.
During the interaction of G. intraradices with wild-type (J5) M. truncatula roots, GintSTE was expressed at the appressoria stage (3 dai) and transcript level increased as the fungus colonized roots (Fig. 5). Appearance of the first penetration points on J5 roots at 5 dai coincided with an increase in transcript levels, then relative GintSTE expression increased with the number of penetration points between 5 and 7 dai of J5 roots, and decreased at 9 dai as the fungus proliferated intraradicularly and formed arbuscules. In the incompatible interaction with the dmi3-1/Mtsym13-1 mutant, GintSTE was expressed but transcript levels stayed low compared with interactions with wild-type M. truncatula roots, and in spite of a greater number of appressoria being formed, they did not vary significantly during the 9 d following inoculation.
To further analyse GintSTE expression, GintSTE transcripts were localized in germinating spores of G. intraradices and inoculated J5 roots by direct fluorescent in situ RT-PCR (Fig. 6). GintSTE gene expression was detected in cytoplasm of germinating spores (Fig. 6a), in pre-infectious extraradicular hypha and in appressoria (Fig. 6b). No fluorescence was observed within intraradical fungal structures (arbuscules) (Fig. 6c), although 25S large subunit ribosomal gene expression showed that these structures were transcriptionally active (Fig. 6d). These results, together with those from gene expression profiling, suggest that GinSTE could be involved in initial penetration processes of host tissues by G. intraradices but not fungal proliferation within roots.
Expression of GintSTE restores pathogenicity of a C. lindemuthianum clste12Δ mutant
To gain insight into a possible role of GintSTE in penetration of plant tissues, complementation experiments were performed with a clste12Δ disruption mutant of the plant-infecting hemibiotrophic fungus C. lindemuthianum. Although this mutant differentiates mature melanized appressoria, it is blocked in the penetration step, which causes a loss of pathogenicity.
A clste12Δ strain (B77) was transformed using the expression vector pDHt-GintSTE, containing the complete GintSTE coding sequence placed under the control of the promoter and terminator of CLSTE12. A hygromycin-resistant transformant, B77G6, carrying a single copy of the transgene was selected by Southern blot analysis (Fig. 7a). RT-PCR experiments showed that GintSTE was correctly expressed in the transgenic strains (Fig. 7b).
Pathogenicity tests were conducted by recording the appearance of anthracnose symptoms on C. lindemuthianum-inoculated leaves of the susceptible P. vulgaris cultivar P12S (Fig. 7c). The mutant strain carrying the empty vector (B77v) remained nonpathogenic. The strain complemented with GintSTE (B77G6) formed necrotic anthracnose lesions where maceration of plant tissues characteristic of the necrotrophic phase was observed. This shows that GintSTE was able to replace CLSTE12 and restore the ability of C. lindemuthianum to penetrate host tissues. However, these lesions were not as dense (Fig. 7c) or as extensive (110 ± 70 µm in diameter as against 370 ± 130 µm) as those produced by the wild-type strain (βv).
We have isolated the first STE12 homologue from an AM fungus and shown the encoding gene to be a general feature in the Glomeromycota. STE12 is a major transcriptional regulator in fungi where it modulates different developmental processes under specific conditions. In S. cerevisiae, nutrient stress is responsible for the STE12-dependent dimorphic transition from budding to filamentous/invasive growth forms which enables the fungus to explore its environment for nutrient acquisition. Both animal and plant pathogenic fungi rely on STE12 activity to colonize their host and fulfil their nutrient requirements. In animal pathogens such as Cryptococcus neoformans, STE12 regulates virulence factor expression (Chang et al., 2000), and in hemibiotrophic plant pathogens such as C. lagenarium, STE12 is essential for penetration and subsequent colonization of host tissues (Tsuji et al., 2003).
GintSTE is closely related to STE12 from hemibiotrophic plant pathogens but also exhibits characteristics proper to G. intraradices
The STE12-like gene GintSTE from the AM fungus G. intraradices encodes a protein possessing an N-terminal homeodomain typical of the STE12 transcription factor family. As may be expected, the G. intraradices protein is more closely related to STE12 homologues from filamentous fungi, especially hemibiotrophs, than to those from yeasts. It possesses two adjacent zinc fingers and two overlapping NLS that do not exist in ascomycetous yeast STE12 homologues and an asparagine-rich region close to the glutamine-rich region found in hemibiotrophic homologues. Moreover, as observed in filamentous fungi, GintSTE does not show significant homology with STE12 in the MAPK Kss1 and repressor Dig1 binding regions that are important for transcriptional activation in S. cerevisiae (Song et al., 1991; Olson et al., 2000).
Although GintSTE and its homologues from filamentous fungi share very similar DNA binding and NLS domains (> 80% homology), they do not show significant homology in the central region localized between the homeodomain and the zinc fingers, suggesting slightly different regulation modes. In GintSTE, this region is shortened, some stretches of conserved amino acids are lacking, and the number and position of post-translational modification sites are sometimes different. For example, the cAMP-cGMP-dependent phosphorylation site RRRS (e.g. aa 418–421 of MST12) and the SUMOylation site KRE (e.g. aa 284–286 of MST12), conserved among STE12 sequences of filamentous fungi, were not found in GintSTE.
Contrary to MST12, CST1 and CLSTE12 genes, GintSTE does not contain any intron. If intronless genes are characteristic of prokaryote genomes, they also constitute a significant portion of mammalian eukaryotic genes which code G-protein receptors and olfactory receptors, for example (Gentles & Karlin, 1999). Recently a genome-wide analysis of intronless genes performed in rice and Arabidopsis (Jain et al., 2007) revealed that several of these have crucial functions in plant growth and development, including protein synthesis, signal transduction and DNA binding. It has been proposed that the absence of intron allows rapid and accurate transcription of essential genes even when the cellular splicing machinery is compromised (Ardley et al., 2000).
GintSTE is up-regulated in germinating spores and upon root penetration in G. intraradices
Host root exudates are known to enhance fungal transcription during spore germination and especially for genes associated with mitochondrial activity (Tamasloukht et al., 2003), mobilization and transport of glycogen and lipid reserves (Park et al., 2002), protein synthesis or stress responses (Seddas et al., 2008). GintSTE expression levels increased significantly upon spore germination but effects did not differ between water and root exudates, suggesting that the up-regulation process for this gene is not dependent on a diffusable signal from the plant partner. In situ RT-PCR experiments localized transcripts of GintSTE not only within cytoplasm of spores but also in cytoplasm of germination hyphae. Such a high level of expression may reflect a more general role of GintSTE in the control of processes, leading to spore germination, similar to that of CST1 involved in conidia germination in C. lagenarium (Park et al., 2002).
A differential expression of GintSTE was observed when G. intraradices established contact with roots of wild-type as compared with mutant M. truncatula. During the compatible interaction with roots of the wild-type genotype J5, an increase in GintSTE expression coincided with the development of the first penetration points at the root surface and then continued as appressoria and fungal penetration points became more numerous. The decrease in transcript levels once penetration attempts began to slow down and arbuscules started to develop within root tissues (7 dai) suggests an additional role of GintSTE in fungal development, before root colonization and necessary to the penetration step. The localized expression of GintSTE in external fungal structures during compatible interactions, the lack of gene activation in incompatible interactions between G. intraradices and the dmi3-1/Mtsym13-1 mutant, and the ability of GintSTE to restore infectivity in the clste12Δ mutant of C. lindemuthianum are consistent with this hypothesis. This, together with the overall weaker expression of GintSTE in fungal structures in contact with roots as compared with spore germination, may reflect a more focused and fine regulation exerted by GintSTE at the particular stage of root penetration.
Contrary to the expression of GintSTE in compatible fungal interactions with wild-type M. truncatula, the incapacity of G. intraradices to penetrate roots of the incompatible dmi3-1/Mtsym13-1 mutant of M. truncatula is associated with a defect in GintSTE expression which remains low despite the formation of numerous appressoria. Appressorium formation has been suggested to be simply triggered by root epidermis epitopes (Nagahashi & Douds, 1997), whilst hyphal penetration of root tissues requires a more complex signalling process (Gianinazzi-Pearson et al., 2007). The dmi3-1/Mtsym13-1 mutant lacks a calcium/calmodulin-dependent kinase activity that is part of a central signal transduction pathway controlling AM symbiosis establishment (Lévy et al., 2004). This pathway is assumed to be activated by a fungal signal of unknown identity, termed the ‘myc factor’ (Kosuta et al., 2003). The penetration-resistant phenotype of the dmi3-1/Mtsym13-1 mutant could thus result from its incapacity to perceive a symbiotic fungus, which is consistent with a lack of activation of signal transduction-related plant genes (Weidmann et al., 2004, Sanchez et al., 2005) and the defect in PPA assembly observed in mutant roots upon fungal contact (Genre et al., 2005). The low expression levels of GintSTE during incompatible interactions of G. intraradices with the mycorrhiza-defective plant mutant could result from such an altered molecular dialogue linked to inactivation of the DMI3/MtSYM13 gene.
GintSTE encodes a functional transcription factor possibly involved in the regulation of fungal infectivity
In haploid S. cerevisiae, nutrient limitation causes a developmental switch that allows cells to penetrate the surface of an agar medium in a process called invasive growth (Roberts & Fink, 1994; Cullen & Sprague, 2000). Successful complementation of the yeast ste12Δ mutant confirmed that GintSTE encodes a functional transcription factor able to activate genes regulating invasive growth, suggesting that it can recognize the FRE (filamentous and invasion response element) present in their promoter, as reported for CLSTE12 from C. lindemuthianum (Wong Sak Hoi et al., 2007).
STE12 proteins play a pivotal role in host tissue colonization by hemibiotrophic plant pathogens. In C. lindemuthianum, CLSTE12 is essential for fungal penetration of host tissues and infectious growth, and as a consequence clste12Δ strains are nonpathogenic. When the B77 (clste12Δ) mutant strain was complemented with GintSTE, it was able to produce necrotic lesions on P. vulgaris leaves, showing that GintSTE is able to restore fungal penetration and infectivity. This result, coupled with the restored penetration of agar medium in yeast, converges to a likely role of GintSTE in a penetration process. This is consistent with GintSTE transcriptional activation in fungal structures associated with root penetration by G. intraradices.
GintSTE does not fully restore C. lindemuthianum virulence
The lesions formed by the clste12Δ strain complemented with GintSTE were not as extensive as those formed by the wild type strain, indicating that GintSTE restores fungal infectivity but not the same degree of fungal virulence. Different hypotheses can be proposed to explain this phenotype. Firstly, GintSTE activation by the cellular machinery may be suboptimal in the heterologous system C. lindemuthianum. Although Park et al. (2004) showed that the conserved MAPK and PKA phosphorylation sites in M. grisea are not necessarily required for infection of rice, we cannot exclude that these specific phosphorylation sites play a role in the activation of STE12 protein from Colletotrichum spp. The activity, subcellular localization and stability of GintSTE could also be affected by the absence of a specific site. The number and position of the post-translational modification sites predicted from the GintSTE aa sequence are sometimes different from that occurring in homologues of other filamentous fungi. For example, only one cAMP-cGMP-dependent phosphorylation site is present in GintSTE (KRFS aa 430–433) and it does not correspond to the one most conserved among filamentous fungi STE12. The conserved SUMOylation site KRE (aa 284–286 of MST12) was also absent in GintSTE. These characteristics may reflect a slightly different regulation mode between the two organisms. Alternatively, GintSTE and CLSTE12 may have slightly different target sequences so that GintSTE would not bind CLSTE12 responsive elements as efficiently as the native protein.
Another explanation could come from the fact that G. intraradices is a symbiotic organism while C. lindemuthianum is a pathogen. GintSTE may be less efficient than CLSTE12 in activating genes involved in the pathogenicity of C. lindemuthianum. Full activation of necrotrophy-related genes by CLSTE12 may require specific post-translational modifications of the protein, or interaction with a co-activator that does not occur with GintSTE. Such a mechanism was reported for S. cerevisiae STE12 which regulates distinct developmental processes by interacting with different co-regulators: TEC1 for filamentous growth and MCM1 for mating (Hwang-Shum et al., 1991; Gavrias et al., 1996). If we assume that such a regulatory mechanism is exerted on CLSTE12, the involved sequence should be in the central region of the protein. The differences existing between CLSTE12 and GintSTE in this particular region, especially in the part situated between the NLS and the zinc fingers, which is almost nonexistent in GintSTE, could explain the reduced efficiency of GintSTE in activating invasive growth of the fungal pathogen.
STE12 – a common genetic feature among plant infecting fungi?
The capacity of the AM fungal TF GintSTE to replace the STE12 function in C. lindemuthianum raises the question of the frontier existing between pathogenesis and symbiosis. Many studies have provided evidence that this frontier can be easily bypassed (Redman et al., 2001; Kogel, et al., 2006). Colletotrichum spp. are classified as pathogens, but several species can express commensal or mutualistic lifestyles. The pathogen of cucurbit plants Colletotrichum magna can grow asymptomatically on various noncucurbit-species (Redman et al., 1999; Freeman et al., 2001); moreover, mutations of specific loci, such as path-1, can turn the fungus into a nonpathogenic endophytic mutualist (Freeman & Rodriguez, 1993). In some cases, the benefits for plants can be similar to those of AM symbiosis: growth enhancement, disease resistance and drought tolerance (Redman et al., 2001).
Plant-infecting fungi may share a common genetic basis that has diverged during evolution (Redman et al., 1999, 2001). It is the differential expression and regulation of the genetic patrimony, depending on the plant partner and the environmental conditions, that could decide the outcome of the plant–fungal interaction. The possible common function of STE12 in pathogens and AM fungi may reflect conservation during evolution of an ancient ability to penetrate plant tissues without alarming host defences. This hypothesis is supported by the fact that the pathogen M. grisea possesses attributes of root-infecting fungi potentially reminiscent of AM fungi (Sesma & Osbourn, 2004) and that common plant genes are induced by both pathogens and AM fungi (Güimil et al., 2005).
We report the first identification of a fungal transcription factor potentially involved in the establishment of the AM symbiosis. In G. intraradices, the transcriptional activator GintSTE could be involved in the regulation of morphogenetic processes allowing fungal penetration. The discovery of sequences homologous to GintSTE in several fungi belonging to different families of the Glomeromycota lends support to a conserved role of STE12 in AM interactions. Another clue to the possible biological significance of STE12 in AM fungi is inferred from the absence of introns in GintSTE gene which may ensure rapid expression of the protein under different conditions and may contribute to the successful colonization of different plants in diverse ecosystems. Sequence data indicate that part of the activity and regulation process of GintSTE may be common to STE12 from other filamentous fungi, especially those from hemibiotrophic pathogens of Colletotrichum spp. The fact that GintSTE restores the pathogenicity of the clste12Δ mutant offers the first evidence for the possible existence of a common genetic element regulating penetration of plant tissues by symbiotic and pathogenic fungi. Further characterization of GintSTE function is under way to investigate these hypotheses and to identify the AM fungal genes regulated by the transcription factor.
Nucleotide and amino acid sequence data from this article are available in the EMBL/GenBank databases under accession numbers AM947048 (GintSTE ORF from G. intraradices BEG141), AM947049 (partial STE12 homeodomain from G. mosseae BEG12), AM947050 (partial STE12 homeodomain from Gi. margarita BEG 34), AM947051 (partial STE12 homeodomain from Ac. laevis BEG13), AM947052 (partial STE12 homeodomain from G. caledonium BEG20), AM947053 (partial STE12 homeodomain from G. geosporum BEG11), AM947054 (partial STE12 homeodomain from G. claroïdeum BEG31), AM947055 (partial STE12 homeodomain from G. versiforme BEG47), AM947056 (STE12 nuclear localization signal and zinc finger motifs from G. mosseae BEG12), AM947057 (STE12 nuclear localization signal and zinc finger motifs from G. geosporum BEG11), AM947058 (STE12 nuclear localization signal and zinc finger motifs from G. caledonium BEG20), AM947059 (STE12 nuclear localization signal and zinc finger motifs from G. claroïdeum BEG31), AM947060 (STE12 nuclear localization signal and zinc finger motifs from Gi. margarita BEG 34), AM947061 (STE12 nuclear localization signal and zinc finger motifs from G. versiforme BEG47), and AM947062 (STE12 nuclear localization signal and zinc finger motifs from Ac. laevis BEG13).
The authors are grateful to the Conseil Régional de Bourgogne for financial support (thesis grant, MT; FABER Research Project, N 05 512 AA 06S2452), to A. Colombet and V. Monfort (IBG, Dijon, France) for mycorrhizal fungi and inoculum of G. intraradices BEG141, and to G. Duc (UMR LEG, Dijon, France) for M. truncatula seeds. They also thank Hunter Richards (Baylor College of Medecine, Houston, TX, USA) for kindly providing the In-fusion™ 2.0 Dry-Down PCR cloning kit.