Genetic evidence for auxin involvement in arbuscular mycorrhiza initiation

Authors


Author for correspondence:
Catharina Coenen
Tel: +1 814 332 2703
Email: catharina.coenen@allegheny.edu

Summary

  • Formation of arbuscular mycorrhiza (AM) is controlled by a host of small, diffusible signaling molecules, including phytohormones. To test the hypothesis that the plant hormone auxin controls mycorrhiza development, we assessed mycorrhiza formation in two mutants of tomato (Solanum lycopersicum): diageotropica (dgt), an auxin-resistant mutant, and polycotyledon (pct), a mutant with hyperactive polar auxin transport.
  • Mutant and wild-type (WT) roots were inoculated with spores of the AM fungus Glomus intraradices. Presymbiotic root–fungus interactions were observed in root organ culture (ROC) and internal fungal colonization was quantified both in ROC and in intact seedlings.
  • In ROC, G. intraradices stimulated presymbiotic root branching in pct but not in dgt roots. pct roots stimulated production of hyphal fans indicative of appressorium formation and were colonized more rapidly than WT roots. By contrast, approaching hyphae reversed direction to grow away from cultured dgt roots and failed to colonize them. In intact seedlings, pct and dgt roots were colonized poorly, but development of hyphae, arbuscules, and vesicles was morphologically normal within roots of both mutants.
  • We conclude that auxin signaling within host roots is required for the early stages of AM formation, including during presymbiotic signal exchange.

Introduction

Arbuscular mycorrhiza (AM), a symbiosis between fungi of the phylum Glomeromycota and nearly 80% of terrestrial plant species, is characterized by a two-way exchange in which the fungus provides phosphate and other nutrients to the plant in exchange for carbohydrates required for fungal reproduction. Although steps leading to AM colonization are well established (Harrison, 2005; Paszkowski, 2006), molecular and genetic events during mycorrhiza formation are in the early stages of understanding.

Phytohormones act as signals in many plant developmental processes, as well as in interactions between plants and microbes, making a role for these compounds in AM formation plausible (Barker & Tagu, 2000; Hause et al., 2007; López-Ráez et al., 2010). Roles for plant-produced jasmonates and abscisic acid in the AM symbiosis have been investigated in some detail (Hause et al., 2007; Herrera-Medina et al., 2007, 2008; López-Ráez et al., 2010; Rodriguez et al., 2010). Because AM fungi stimulate the formation of lateral roots (Oláh et al., 2005), which are preferentially colonized by the fungus (Smith & Read, 1997; Kaldorf & Ludwig-Müller, 2000; Gutjahr et al., 2009), auxin, a plant hormone with an essential role in root initiation and growth (Overvoorde et al., 2010), is also a good candidate for mycorrhizal involvement.

Whereas ectomycorrhizal and root endophytic fungi produce auxin (Ek et al., 1983; Frankenberger & Poth, 1987; Gay et al., 1994; Sirrenberg et al., 2007), auxin production by AM fungi, obligate symbionts, is difficult to assess. Responses of root auxin concentrations to AM colonization appear to vary with host species and time after inoculation. For example, indole-3-acetic acid (IAA) concentrations are unchanged in leek and tobacco (Torelli et al., 2000; Shaul-Keinan et al., 2002), but are higher in colonized soybean roots (Meixner et al., 2005). Concentrations of the root-derived auxin indole-3-butyric acid are also elevated in mycorrhizal maize and nasturtium roots, whereas IAA concentrations are unchanged (Kaldorf & Ludwig-Müller, 2000; Fitze et al., 2005; Jentschel et al., 2007).

Further indirect evidence for auxin involvement in AM formation comes from the recent discovery that auxin controls the expression of Arabidopsis genes involved in the synthesis of strigolactones (Bainbridge et al., 2005; Ongaro & Leyser, 2008; Hayward et al., 2009), which are plant-emitted carotenoid derivatives that induce responses in the fungus (Akiyama et al., 2005; Koltai et al., 2010) and are involved in mycorrhiza formation. However, direct genetic evidence for an auxin requirement in establishing a functional AM symbiosis is currently lacking: although many auxin-related mutants are available in Arabidopsis thaliana, this species, like most members of the Brassicaceae, is resistant to mycorrhizal colonization, so that AM formation cannot be assessed directly in Arabidopsis mutants.

Two well-characterized auxin-related mutants are available in tomato (Solanum lycopersicum), a species that is readily colonized by AM fungi: the auxin-resistant diageotropica (dgt) mutant and the auxin hyper-transporting polycotyledon (pct) mutant. dgt plants form adventitious roots, but seedlings lack lateral roots, unless their primary root apex is damaged (Zobel, 1973). Although dgt roots contain elevated concentrations of free auxin and undergo pericycle cell proliferation in response to applied auxin, they fail to differentiate these cells into lateral root primordia (Ivanchenko et al., 2006). The dgt phenotype and reduced responses to auxin in both physiological (Kelly & Bradford, 1986; Muday et al., 1995; Coenen & Lomax, 1998) and gene expression (Nebenführ et al., 2000; Balbi & Lomax, 2003; Coenen et al., 2003) assays indicate that the DGT gene, which encodes the cyclophilin LeCYP1 (Oh et al., 2006), is required for auxin signal transduction. Increased lateral root formation in the pct mutant may be a result of increased auxin transport in pct stems (Al-Hammadi et al., 2003), which can be explained by overexpression of the auxin efflux carrier PIN1 (Kharshiing et al., 2010).

Root organ culture (ROC) permits close observation of AM development in the absence of the complexities typically associated with soil environments. Because excised tomato roots grow and branch in tissue culture in the absence of exogenously supplied hormones, they are ideal systems for the study of AM formation in ROC. Here, we report that mutations in auxin response and transport affect the ability of AM fungi to infect both cultured and intact root systems of tomato.

Materials and Methods

Fungal culture

Glomus intraradices Schenck & Smith (DAOM 181602; obtained from David Douds at the USDA ARS in Beltsville, MD, USA) was propagated in a two-compartment system with hairy roots of Daucus carota (St-Arnaud et al., 1996) on modified M-media (Bécard & Fortin, 1988) as described by Juge et al. (2002). To isolate aseptic spores for ROC, mycelial mats were removed from distal compartments of plates containing liquid media, agitated with a glass rod in sterile dH2O, homogenized with the small probe of a Brinkman Polytron 2000 homogenizer (Metrohm, Riverview, FL, USA), and passed through a wet 45 μm sieve. For ROC initiation, 50 spores were pipetted onto each tissue culture dish of M-media and germinated for 2 d before the addition of a root. For glasshouse experiments, spores from solid media in the distal compartments of plates were wet-sieved after solubilizing the media in 10 mM sodium citrate buffer at pH 6.0 (Doner & Bécard, 1991).

Plant culture

Solanum lycopersicum L. cv Ailsa Craig and the dgt mutant crossed into the Ailsa Craig background were obtained from Charles Rick at the University of California, Davis, CA, USA. The pct mutant, obtained from the Tomato Genetics Resource Center at the University of California, Davis, is allelic to the poc mutant (now renamed as pct 1-2), originally described by Al-Hammadi et al. (2003) (Madishetty et al., 2006). Seeds were agitated in 30–50% (v/v) household bleach for 20 min and rinsed thoroughly with sterile distilled water. Seeds for glasshouse experiments were sown on filter paper over dampened vermiculite in a clear plastic box covered with plastic wrap. Seeds for ROCs were plated on medium containing 0.8% tissue culture-grade agar and half-strength MS salts (Murashige & Skoog, 1962) without added organics (both from PhytoTechnology Labs, Lenexa, KS, USA). Plants germinated at 26°C under a 16 : 8 h light : dark cycle and 60% humidity.

For glasshouse experiments, seedlings with expanded cotyledons were transferred to planting cones (ESC10; Stuewe & Sons, Inc., Corvallis, OR, USA) containing an autoclaved 1 : 1 mixture of vermiculite (Packaging Industries, Inc., N. Bloomfield, OH, USA) and acid-washed sand (Fairmont Minerals, Chardon, OH, USA). Plants were grown at 22–30°C under natural light conditions, watered with distilled water, and fertilized with 20 ml wk–1 of fertilizer solution (alternating between 1/10th strength low-phosphorus and phosphate-free Long Ashton Solution (Hewitt, 1966)). An aqueous suspension of 250 spores of G. intraradices was pipetted directly below the substrate surface of plants that had formed a first true leaf.

For each root organ culture, one 3-cm-long seedling root was placed on M-media in a Parafilm-sealed tissue-culture dish and incubated in darkness at 26°C and 60% humidity.

Assessment of colonization

Root systems from intact plants were rinsed and divided into four regions as indicated. Roots from ROCs were collected by dissolving the growth medium in 10 mM sodium citrate buffer, pH 6.0 (Doner & Bécard, 1991). Roots were cleared in boiling 10% (w/v) potassium hydroxide for 6 min, stained with Sheaffer black ink (Vierheilig et al., 1998), and cut into 1 cm segments. For each colonization assessment, 20 root segments were randomly selected and mounted on glass slides. Fungal structures were counted using the magnified intersections method (McGonigle et al., 1990). Each root segment was scored for colonization in six different locations. Statistically significant differences were determined through two-way ANOVA and Tukey–Kramer post-hoc tests in StatView software. Differences in variance were assessed using Levene’s test in SPSS.

Results

Development of dgt and pct roots in organ culture

Independent of inoculation, both dgt and pct ROCs exhibited phenotypes distinct from wild-type (WT) roots (Fig. 1a): dgt roots failed to branch, even after up to 14 wk of growth (not shown), and pct roots were more highly branched and changed direction of growth much more frequently than WT roots. These developmental differences between mutant and WT roots in tissue culture indicate that the effects of both mutations are at least partially root-autonomous.

Figure 1.

 Root branching in tomato (Solanum lycopersicum) root organ cultures. One excised seedling root was placed on each plate of solid M-medium and incubated in the dark. (a) Roots of wild-type (WT), diageotropica (dgt), and polycotyledon (pct) plants after 4 wk of growth in presence of Glomus intraradices. Note the absence of lateral roots in dgt and hyperbranching phenotype of pct roots. (b) Root branching in response to coculture with G. intraradices. The number of first- and second-order lateral roots were counted after 2 wk of growth in the presence (inoculated) or absence (control) of 50 spores of G. intraradices. Error bars represent standard errors of the mean (n = 12). Asterisks indicate significant differences from WT controls (P = 0.0107 for first-order lateral roots; P < 0.0001 for second-order lateral roots).

Lateral roots were counted at 2 wk postinoculation (wpi) (Fig. 1b), when primary and secondary lateral roots were still easily distinguishable. Because dgt roots never branched, they were excluded from this analysis. The number of first-order (P = 0.0107, F = 7.251) and second-order (P < 0.0001, F = 37.058) lateral roots was larger in pct than in WT roots. In addition, second-order pct roots increased significantly in response to inoculation (P = 0.0325, F = 5.552), whereas the increase in WT second-order laterals was statistically nonsignificant (P = 0.1871, F = 1.860).

Colonization of ROCs by G. intraradices

To assess colonization of cultured roots by G. intraradices, we observed hyphal growth patterns during early stages of coculture, followed by quantification of internal root colonization in stained roots (Fig. 2). In the presence of WT roots, G. intraradices typically formed runner hyphae and rudimentary branching structures during the first 8 d postinoculation (dpi), with no evidence of the fan-shaped hyphal branching that usually precedes appressorium formation and infection. Substantial internal root colonization occurred by 5 wpi; average colonization percentages ranged from 20 to 60% of scored root locations 5–12 wpi (Fig. 2c).

Figure 2.

 Colonization of tomato (Solanum lycopersicum) root organ cultures by Glomus intraradices. A single root explant was placed on each plate containing 50 pregerminated spores. (a) Hyphal growth patterns after 8 d of coculture. Hyphae are traced to distinguish them from root hairs. Note the change in growth direction of hypha approaching diageotropica (dgt) root and hyphal fan formation on polycotyledon (pct) root surface. Apparent size differences between roots in these three images are the result of root localization in different planes of focus from hyphae; scale bars are appropriate for hyphae only. Differences in root hair length are a result of root location on top of the agar (long root hairs) vs in the agar (short root hairs), with no genotypic preferences for root growth in either location. (b) Stained fungal structures after 6 wk of coculture. Note the staining of internal fungal structures in wild-type (WT) and pct roots and the absence of staining in dgt roots. (c) The average percentage of scored root points containing stained fungal structures and arbuscules at 3, 4, 5, 6, 7 and 12 wk. Root systems were cut into 1 cm segments, and each segment was scored for colonization in six separate locations. For WT (open circles) and pct roots (closed circles), each root system was scored for colonization at > 100 separate points, and the colonization percentages for root systems from separate plates were averaged. For dgt, the entire root system was cut into 1 cm segments, and every segment was scored in six locations. Asterisks denote significantly higher total colonization of pct roots at 7 wk postinoculation (wpi) and higher arbuscule colonization of pct roots at 7 and 12 wpi as determined by two-way ANOVA with a Tukey–Kramer post-hoc test (P < 0.05). Error bars represent the standard error of the mean (n = 4, except where indicated). This experiment was performed twice with similar results. Note that dgt roots failed to be colonized even when lateral root branching was induced by damage of the primary root tip.

After 8 d of coculture, fungal hyphae frequently curled away from dgt roots, behavior never seen in WT or pct cultures (Fig. 2a). Consistent with this observation, G. intraradices failed to colonize dgt roots (Fig. 2b,c), even at 14 wpi.

Hyphal fans indicative of appressorium formation appeared in pct cultures by 8 dpi (Fig. 2a), and morphologically normal internal hyphae, arbuscules, and vesicles formed in pct roots by 4 wpi (Fig. 2c). Overall colonization time courses for pct and WT roots were significantly different (P = 0.0098, F = 37.79), with post-hoc analysis demonstrating significantly higher total colonization percentages of pct roots at 7 wpi (P = 0.0036, F = 37.265). This difference was also statistically significant in a separate experiment with six replicates per genotype performed for this time point, in which ROC colonization percentages averaged 42% for WT and 67% for pct (Student’s t-test assuming unequal variance: P = 0.001; WT SEM = 4.18, pct SEM = 3.71). Furthermore, the percentage of scored root points containing arbuscules rose more quickly for pct than for WT (P < 0.0001, F = 99.179), with significantly higher arbuscule percentages at 7 wpi (P = 0.005, F = 30.09) and 12 wpi (P < 0.001, = 99.179), as determined by post-hoc analysis. Together, hyphal fan formation by 8 dpi, higher arbuscule percentages at 7 and 12 wpi, and higher colonization percentages at 7 wpi suggest more rapid colonization of pct roots as compared with WT roots.

Root development in glasshouse-grown seedlings

dgt seedlings formed both lateral and adventitious roots, although in smaller numbers than WT or pct seedlings (Fig. 3). While both WT and pct root systems were characterized by extensive development of higher-order branch roots, dgt roots produced nearly exclusively first-order lateral roots. As previously reported for glasshouse-grown pct plants (Al-Hammadi et al., 2003) and also in our ROC experiments (Fig. 1), pct roots appeared to contain regions with more frequent branch roots than is typical for roots of WT seedlings (Fig. 3). Colonization of roots by G. intraradices had no obvious effect on overall root system architecture in any of the three tomato seedling genotypes; however, because of the difficulty of recovering intact root systems bearing all of their laterals from the substrate, root numbers were not quantified.

Figure 3.

 Morphology of inoculated tomato (Solanum lycopersicum) seedlings 6 wk postinoculation with Glomus intraradices. Seedlings were cultured in a soil-less mix of sand and vermiculite under natural light conditions in the glasshouse. Note the reduced branching of diageotropica (dgt) roots and the increased branching of polycotyledon (pct) roots, as compared with wild-type (WT) roots. Root architecture of inoculated plants (pictured) showed no obvious difference from that of noninoculated control plants for any of the three seedling genotypes.

AM development in roots of intact seedlings

To assess the influence of shoot-derived factors on AM formation in dgt and pct roots, colonization by G. intraradices was examined in whole seedlings, grown in a soil-less mixture of sand and vermiculite in the glasshouse, at 6 wpi, when WT colonization was first detected, and at 8 wpi, when WT root colonization was nearly complete (Fig. 4).

Figure 4.

 Colonization of intact tomato (Solanum lycopersicum) seedling roots by Glomus intraradices. (a) Division of root system into four regions: adventitious roots originating from the hypocotyl, basal roots originating 0–3 cm below the root–shoot node (RSN), and lateral roots at the top (3–13 cm below the RSN) and bottom (> 13 cm below the RSN) of the root system (diageotropica (dgt) at 6 wk postinoculation (wpi) shown). (b) Differential colonization of root system regions at 6 and 8 wpi. (c) Overall colonization of root systems determined from pooled regional samples. Open symbols represent colonization percentages of individual root systems, and closed symbols represent the mean of all root systems analyzed for a given genotype. (d) Development of G. intraradices within the root, determined from pooled regional samples. Each root region was scored at 120 points for the presence of internal hyphae (hyphae), arbuscules and hyphae (arbuscules), or vesicles, arbuscules and hyphae (vesicles) at 6 and 8 wpi. The relative distribution of these fungal development stages are presented as a percentage of all scored root points containing fungal structures. Error bars represent standard error of the mean (= 6 for wild-type (WT), open bars; dgt, light gray hatched bars; polycotyledon (pct), dark gray hatched bars, at 8 wpi, n = 5 for pct at 6 wpi). Different letters represent significance as determined by two-way ANOVA with a Tukey–Kramer post-hoc test (P < 0.05). To correct for unequal variance of overall root colonization between genotypes (confirmed by Levene’s test), data were arcsine-transformed before statistical analysis.

Adventitious roots, basal roots, and the top and bottom half of the remaining root system were analyzed separately (Fig. 4a,b), to assess whether altered auxin physiology of the mutants selectively affected colonization of certain parts of the root system. Adventitious roots were of special interest because they are not found in ROCs, but readily form in dgt seedlings. Basal roots can modify their gravitropic angle in response to phosphorus supply (Bonser et al., 2006), suggesting that they may be particularly sensitive to auxin–phosphate interactions. The top half of the remaining root system was of particular interest, because it remained unbranched in dgt (Figs 3, 4a). At 6 wpi, G. intraradices structures were located predominantly in the upper regions of the root systems, regardless of genotype, although the differences between upper and lower parts of the root system were significant only for WT (P = 0.0002, F = 11.019) and pct (P = 0.0018, F = 7.968). Interactions between these colonization patterns and plant genotype were statistically nonsignificant at both 6 and 8 wpi.

Pooled data from all four root regions (Fig. 4c) demonstrated significant effects of host genotype on overall degree of colonization at both 6 wpi (P = 0.0001, F = 10.803) and 8 wpi (P < 0.0001, F = 18.169). At 6 wpi, pct roots were significantly less colonized than WT or dgt roots, and at 8 wpi both mutants were significantly less colonized than WT roots. At 8 wpi, WT roots were nearly 90% colonized, whereas an average of < 50% of the total root length of mutant roots contained G. intraradices structures. Roots of pct seedlings remained entirely uncolonized in a separate experiment (data not shown), in which plants were fertilized with a 100-fold higher concentration of phosphate (1.5 mM P biweekly vs 15 μM P biweekly in the current experiment). Under this higher P fertilization regime, WT roots were colonized at 22% (SD = 1.7%) and dgt roots were colonized at 19% (SD = 2.0%) at 8 wpi.

Colonization of roots within each seedling genotype was characterized by counting internal fungal structures. Scored root points classified as containing vesicles always also contained arbuscules and hyphae, and segments classified as containing arbuscules also always contained hyphae; an observation that reflects the progression of fungal development within the root. At both 6 and 8 wpi, the majority of scored root points for all three genotypes contained arbuscules (Fig. 4d). However, at 8 wpi, more dgt root points contained only hyphae and fewer points had produced vesicular structures, indicating that fungal development in dgt roots progressed somewhat more slowly than in WT roots.

Discussion

dgt and pct phenotypes are root-autonomous

Excised WT tomato roots developed numerous branch roots in ROC (Fig. 1a), presumably in response to endogenous auxin production in the root, which is well documented in other species (Overvoorde et al., 2010). Consistent with auxin-resistant growth of excised dgt roots (Muday et al., 1995), cultured dgt roots failed to branch (Fig. 1a), indicating that the DGT gene is required for lateral root formation in response to root-produced auxin.

While shoot-derived auxin is essential for lateral root emergence in Arabidopsis (Overvoorde et al., 2010), the hyper-branching phenotype of pct roots in ROC (Fig. 1a) indicated that PCT-mediated root branching occurred in response to root-endogenous auxin pools. The frequency of lateral root initiation in the Arabidopsis root apical meristem is regulated by pulsed auxin movements from protoxylem cells to the pericycle (De Smet et al., 2007), which is sensitive to auxin transport inhibitors (Casimiro et al., 2001). Because auxin-transporting cells in pct stems overexpress PIN1 proteins (Kharshiing et al., 2010), hyperactive auxin transport within pct root meristems may prime supernumerary pericycle cells for lateral root induction.

AM effects on root morphology are auxin-dependent

Although AM stimulation of root branching in WT roots was not statistically significant, G. intraradices doubled root branching in auxin-hypertransporting pct roots (Fig. 1b). AM fungi alter root architecture through presymbiotic signaling (Oláh et al., 2005; Gutjahr et al., 2009), which also must be responsible for the increased formation of lateral roots in tomato ROCs at 2 wpi (Fig. 1b), a time when ROCs showed no sign of internal colonization (Fig. 2). Absence of dgt root branching after inoculation (Fig. 1a) suggested that presymbiotic lateral root induction by G. intraradices in tomato was auxin-dependent, as is true for ectomycorrhizal development (Felten et al., 2009).

Presymbiotic AM fungal development requires host auxin responses

Fungal hyphae changed their growth direction away from dgt roots (Fig. 2a), similar to previously described AM fungal responses to nonhost roots (Giovannetti et al., 1994; Buee et al., 2000; Bécard et al., 2004). While identification of the root-generated signal eliciting this fungal response was beyond the scope of this study, the behavior of hyphae approaching dgt roots in ROC suggested that DGT-mediated auxin responsiveness of the roots mediates either production of attractants or repression of repellant signals. The importance of host auxin responses for signaling between plant roots and approaching hyphae was also supported by rapid formation of hyphal fans (Fig. 2a) and colonization (Fig. 2b,c) of pct roots. Because strigolactone exudates from tomato roots elicit hyphal branching (Koltai et al., 2010) and enzymes required for strigolactone synthesis are encoded by auxin-inducible genes (Bainbridge et al., 2005; Hayward et al., 2009), future studies should quantify strigolactone synthesis in the dgt and pct mutants.

Colonization phenotypes of dgt and pct roots are shoot-regulated

Although ROCs provide a valuable system to study root–microbe interactions, questions about their validity as predictors of outcomes in whole-plant systems (Toussaint, 2007) pertain to our experiments, because shoot-derived auxin is a major controlling factor of root development (Overvoorde et al., 2010) and of strigolactone synthesis (Hayward et al., 2009). Consistent with a role for shoot-derived factors in the control of AM formation, colonization phenotypes of dgt and pct mutants in ROC and whole-plant root systems demonstrated striking differences.

The pct mutation appeared to hasten mycorrhizal colonization in ROC (Fig. 2) but delayed colonization in intact seedlings (Fig. 4c). Thus, the combination of increased auxin transport in pct roots with auxin inputs from the shoot may create supraoptimal auxin concentrations in tissues critical for the control of symbiotic signaling. Reduced pct root colonization in glasshouse plants also indicated that enhanced auxin transport in whole plants inhibited colonization by a mechanism unrelated to the number of emerging root primordia: whereas emerging lateral roots provide entry points for AM fungi (Harrison, 2005; Oláh et al., 2005), pct roots contained hyperbranching regions both in highly colonized ROCs and in poorly colonized whole seedlings.

The difference between complete absence of colonization in dgt ROCs and substantial colonization of dgt seedling roots in the glasshouse was striking. Colonization of intact dgt seedlings may partially be explained by the limited branching of their root systems (Fig. 3). However, absence of colonization in dgt ROCs cannot be explained solely by absence of lateral root primordia, because dgt ROCs remained uncolonized even when they formed lateral roots in response to a damaged root apex (Fig. 2c). If strigolactone production in dgt ROCs is indeed poor, then this effect might be partially remedied in whole seedlings by increased auxin delivery to the root system from the shoot, or by the presence of greater root mass in close proximity to the fungus. However, the root colonization phenotype of intact dgt seedlings is quite dissimilar from seedlings of the strigolactone-deficient tomato mutant SI-ORT1, where colonization of roots inoculated with spores of G. intraradices is < 5% (Koltai et al., 2010), suggesting that defective strigolactone synthesis in dgt either is not the cause of its impaired AM infection or is far more severe in ROC than in whole seedlings.

Collectively, opposite effects of both mutants on AM colonization in ROC and whole plants demonstrate that auxin control of mycorrhiza formation is complex and governed by physiological context. Complementary experimentation in both systems has allowed us to identify conditions under which AM colonization is completely blocked in each mutant (ROC for dgt, increased phosphate supply in glasshouse plants for pct), providing an ideal vantage point for future identification of molecular players in the auxin control of AM formation through studies of differential gene expression profiles.

AM development within the root is auxin-independent

Unlike a disruption of abscisic acid sensing in notabilis and sitiens mutants of tomato (Rodriguez et al., 2010), the hyperactive auxin transport and reduced auxin sensitivity found in pct and dgt roots, respectively, had no effect on the development of fungal structures once the fungus entered the root (Fig. 4d). Hence, completion of fungal development within the root did not seem to be under auxin control. While dgt roots at 8 wpi contained slightly fewer vesicles and slightly more root segments that only had hyphae, rather than arbuscules or vesicles, the small difference was likely a secondary effect of the delayed entry of fungus into the developing root system.

Auxin is required for AM infection

In summary, the experiments described here provide genetic evidence that plant auxin signaling is required for normal AM infection. While the presence of morphologically normal fungal structures within dgt and pct roots indicates that auxin is less important for fungal development postinfection, reduced colonization of mutant roots demonstrates a need for auxin signaling in mycorrhiza formation. This auxin requirement already exists during presymbiotic plant–fungus interactions in root cultures, suggesting that host auxin responses direct the exchange of diffusible signals between plant and fungus.

Acknowledgements

United States Department of Agriculture grant number 2004-35304-14995 to CC and research grants from the Allegheny College Class of ’39 Fund and the Harold State Research Fellowship to M.T.H. are gratefully acknowledged. We thank Dr David Douds (USDA ARS) for starter cultures of G. intraradices, Katherine Restori and Jessica Brazelton for helping to establish culture protocols, Chris Lundberg (Allegheny College) for advice on statistical methodology, and Julia Muntean and Chris Lundberg for careful reading of the manuscript.

Ancillary