Root Endophyte Colletotrichum tofieldiae Confers Plant Fitness Benefits that Are Phosphate Status Dependent

Abstract

A staggering diversity of endophytic fungi associate with healthy plants in nature, but it is usually unclear whether these represent stochastic encounters or provide host fitness benefits. Although most characterized species of the fungal genus Colletotrichum are destructive pathogens, we show here that C. tofieldiae (Ct) is an endemic endophyte in natural Arabidopsis thaliana populations in central Spain. Colonization by Ct initiates in roots but can also spread systemically into shoots. Ct transfers the macronutrient phosphorus to shoots, promotes plant growth, and increases fertility only under phosphorus-deficient conditions, a nutrient status that might have facilitated the transition from pathogenic to beneficial lifestyles. The host's phosphate starvation response (PSR) system controls Ct root colonization and is needed for plant growth promotion (PGP). PGP also requires PEN2-dependent indole glucosinolate metabolism, a component of innate immune responses, indicating a functional link between innate immunity and the PSR system during beneficial interactions with Ct.

Graphical abstract

Figure 1

Prevalence of C. tofieldiae in Four Natural A. thaliana Populations in Spain (A) Geographical location of the sampled sites in Central Spain. (B) Detection of C. tofieldiae (Ct) in roots and/or leaves of healthy A. thaliana plants (collected in Spring between 2009 and 2012) using qPCR analysis with a specific primer pair targeting the Ct tubulin sequence (CT04_11973). See also Figure S1.

Figure 2

C. tofieldiae Colonization of A. thaliana Roots Confocal microscope images of C. tofieldiae (Ct) expressing cytoplasmic GFP (green) and A. thaliana expressing PIP2A-mCherry (magenta). (A) Spores (s) germinating on a leaf form melanized appressoria (arrows); 8 dpi. Overlay projection of GFP and bright-field channels. Scale bar, 20 μm. (B) Spore (s) germinating on a root form multiple long germ-tubes (arrows); 2 dpi. Maximum projection of z stack image. Scale bar, 10 μm. (C and D) Undifferentiated hypha penetrating a root epidermal cell enveloped by PIP2A-mCherry-labeled host membranes (arrows); 2 dpi. (C) Maximum projection of z stack image. (D) Enlargement of one optical section from the projection shown in (C). Scale bars, 5 μm (C); 2 μm (D). (E–G) First infected root epidermal cell (asterisk) containing intracellular hyphae is not labeled by PIP2A-mCherry; 2 dpi. (E) Maximum projection of z stack image. (F) Orthogonal projection of (E). (G) Enlargement of one optical section from the projection shown in (E). Scale bars, 10 μm. (H and I) Undifferentiated hyphae penetrating between two root epidermal cells (e); 2 dpi. (H) Maximum projection of z stack image. (I) Orthogonal projection of (H). Scale bars, 10 μm (H) and 5 μm (I). (J and K) Intracellular (small arrow) and intercellular hyphae (large arrow) colonizing the root cortex; 2 dpi. (J) Maximum projection of z stack image. (K) Enlargement of one optical section from the projection shown in (J). Scale bars, 20 μm (J) and 10 μm (K). (L and M) Intracellular hyphae inside a root cortical cell (c) enveloped by PIP2A-mCherry-labeled host membranes (arrows); 8 dpi. Maximum projections of z stack images. (M) Enlargement of (L) using a subset of optical sections. Scale bars, 20 μm (L) and 10 μm (M). (N and O) Root enmeshed by a network of extraradical hyphae, with melanized microsclerotia developing (arrows); 8 dpi. (N) Maximum projection of z stack image. (O) Bright-field image corresponding to (N). Scale bars, 100 μm. (P) Root epidermal cell packed with swollen hyphal cells with melanized cell walls; 28 dpi. Overlay projection of GFP and bright-field channels. Scale bar, 10 μm. (Q) Hyphae inside the root central cylinder, with xylem tracheids in the same focal plane (arrows); 8 dpi. Overlay projection of GFP and bright-field channels. Scale bar, 20 μm. See also Figure S2.

Figure 3

C. tofieldiae Promotes Plant Growth in Low Phosphate Conditions (A) Number of siliques produced by A. thaliana growing in vermiculite. Plants grown with Ct had significantly more siliques than those grown without C. tofieldiae (Ct) (t test, p < 0.01). (B) A representative image of A. thaliana plants grown in low phosphate (Pi) conditions with and without Ct. Seven-day-old plants were inoculated with Ct or water and grown in low Pi MS medium for 18 days. (C) Shoot fresh weight (SFW) of plants incubated with beneficial Ct or pathogenic C. incanum (Ci) in high or low Pi conditions. A. thaliana Col-0 seeds were inoculated with Ct, heat-killed Ct or Ci, or water (mock), and SFW was determined 24 days later (15 plants per experiment). The boxplot shows combined data from three independent experiments. Different letters indicate significantly different statistical groups (Tukey-HSD, p < 0.01). (D) Translocation of ³³P-labeled orthophosphate from Ct to A. thaliana shoots. Col-0 plants were grown on low Pi medium without (mock, n = 26) or with Ct (n = 37) or Ci (n = 36) or on high Pi medium without (mock, n = 27) or with Ct (n = 35) in a two-compartment system (cartoon). ³³P was added to the hyphal compartment (HC) and after 17 days ³³P -incorporation into shoots was measured by scintillation counting. Columns represent counts in kBq ³³P /g dry weight (DW) of individual plants. The dotted line shows the median level of ³³P background counts from mock inoculations. RHC, root hyphal compartment. See also Figure S3.

Figure 4

C. tofieldiae Colonization Induces Expression of A. thaliana Pht1 Pi Transporter Genes (A) Transcript profiling of 19 A. thaliana phosphate transporter genes and PHF1 in colonized and mock-treated roots under Pi-limiting (low P: [50 μM]) or Pi-sufficient (high P: [625 μM]) conditions at 6, 10, 16, and 24 dpi. Overrepresented (yellow to red) and underrepresented transcripts (yellow to blue) are shown as log2 fold changes relative to the mean expression measured across all stages. Log2-transformed expression levels (white to red) are also depicted for each sample. Significantly regulated genes (|log2FC| >1, FDR <0.05) are highlighted in purple (upregulated) to green (downregulated). Note that two Pht transporter genes were strongly induced at late stages of C. tofieldiae (Ct) colonization (asterisks). (B) Quantitative analysis of shoot fresh weight (SFW) of PSR regulatory mutants under Pi-deficient conditions. SFW was measured at 30 dpi from at least 15 plants per treatment, per experiment. Results of three independent experiments were combined. An ANOVA and subsequent Tukey HSD test were conducted to evaluate whether the fold-change in SFW between Ct- and mock-inoculated (M) plants (calculated as SFW Ct/SFW Mock) was significantly different (^∗p < 0.01) between genotypes (Col-0: 1.45-fold versus phf1: 1.23-fold). (C) Transcript profiling of 61 Ct genes significantly regulated between Ct- and mock-inoculated roots under Pi-deficient (low Pi: [50 μM]) or Pi-sufficient (high Pi: [625 μM]) conditions at 6, 10, 16, and 24 dpi. Overrepresented (yellow to red) and underrepresented transcripts (yellow to blue) are shown as log2 fold changes relative to the mean expression measured across all stages. Significantly regulated genes (|log2FC| >1, FDR <0.05) are highlighted in purple (upregulated) to green (downregulated). Arrows indicate the two most highly upregulated genes: phosphate H⁺symporter (CT04_05366, gray) and acid phosphatase (CT04_08450, black). The right part of the heatmap depicts the functional categories to which genes belong. CSEPs, candidate secreted effector proteins; CAZymes, carbohydrate active enzymes. See also Figure S4.

Figure 5

Trp-Derived Secondary Metabolites Are Required for Beneficial Ct Interactions (A) Scheme of Trp-derived metabolite pathways in A. thaliana. (B) Shoot fresh weight (SFW) of cyp79B2 cyp79B3 and ein2 pad4 sid2 dde2 mutant plants grown with and without Ct in low Pi conditions. SFW was measured 24 days after inoculation of sterilized seeds with C. tofieldiae (Ct). The boxplot shows combined data from three independent experiments. Asterisks indicate significantly different means between mock and Ct-treated plants for each genotype (t test, p < 0.01). (C) SFW of A. thaliana mutants defective in the biosynthesis of camalexin and/or indole glucosinolates at 24 dpi in low Pi conditions. (D) SFW of Cardamine hirusuta and Capsella rubella plants grown with and without Ct in low Pi conditions (24 dpi). Asterisks indicate significantly different means between mock (M) and Ct-treated plants for each species (t test, p < 0.01). See also Figure S5.

Figure S1

Distribution of C. tofieldiae at Different Seasons and at Various Sites in Europe, Related to Figure 1 (A) Distribution of C. tofieldiae (Ct) from A. thaliana growing in the Las Rozas (LRO) site in central Spain at two different seasons of the year (April/May, and November). q-PCR analysis using Ct-specific primers detected the presence of Ct in plants at both seasons. (B) Ct was not detectable in soil and root samples from three A. thaliana populations in France (Saint-Dié des Vosges) and Germany (Geyen and Pulheim).

Figure S2

C. tofieldiae Colonization of A. thaliana Roots and Systemic Colonization of A. thaliana Shoots, Related to Figure 2 (A) Epi-fluorescence micrograph of a root cross-section at 7 dpi. The section was stained with wheat germ lectin-FITC, which labels N-acetylglucosamine residues in fungal cell walls green, but also the secondary cell walls of xylem vessels in the central cylinder (cc). The root cortex (c), epidermis (e) and root hairs (arrows) are extensively colonized by intraradical hyphae, while abundant extraradical hyphae envelope the root at this stage. (B) Bright-field micrograph of a root cross-section stained with Toluidine blue (7 dpi). Note the microsclerotium (thick arrows) developing in the epidermis and cortex and long intracellular hyphae spreading in the cortex (thin arrows). (C–E) Transmission electron micrographs of ultrathin sections. (C) Cross-section of a hypha inside a root endodermal cell (en) as indicated by the presence of Casparian strip cell wall alterations in the anticlinal cell wall (arrow). 5 dpi. (D and E) Hypha (h) in contact with the root periderm (p) as indicated by the presence of layered suberin lamellae (sl) in the peridermal cell wall (arrow and inset). 7 dpi. fw, fungal wall; fc, fungal cytoplasm; pc, plant cytoplasm. Scale bars, 20 μm (A and B), 2 μm (D), 500 nm (C and E). (F) Detection of C. tofieldiae (Ct) in healthy shoots of A. thaliana following root inoculation. A. thaliana Col-0 plants were grown hydroponically and the 20-day-old roots were infected with Ct-GFP spores. q-RT-PCR analysis with GFP-specific primers detected the presence of Ct-GFP in some healthy leaves. Approximately10 leaves per time point were collected. (G) Confocal microscope image showing Ct-GFP hypha growing in vein tissue of a healthy leaf at 28 days post inoculation (dpi) of roots. Bar = 20 μm. (H) Confocal microscope image showing Ct-GFP hyphae growing in a senescent leaf at 48 dpi. Bar = 20 μm.

Figure S3

C. tofieldiae-Mediated Plant Growth Promotion in Phosphate Limiting Conditions, Related to Figure 3 (A) A. thaliana vegetative growth in nutrient-poor vermiculite was improved in the presence of C. tofieldiae (Ct). Col-0 plants grown in half MS medium containing 0.8% sucrose were transferred to vermiculite soil either with or without the addition of Ct mycelium. The photographs were taken after co-cultivation for four weeks. (B) Root growth promotion by Ct in low Pi conditions. Root length was measured 24 days after sterilized Col-0 seeds were inoculated with Ct spores (24 dpi). Ct treatment significantly increased root length (two-tailed t test, p < 0.0001). The graph represents combined data from three independent experiments. (C) Shoot growth promotion by Ct when hydroxyapatite provided the sole Pi source. Shoot fresh weights were measured at 24 dpi. Ct treatment significantly increased the shoot fresh weight (two-tailed t test, p < 0.0001). The graph shows combined data from two independent experiments. (D) The concentration of phosphorus (P) in A. thaliana shoots was significantly increased after growth in the presence of Ct. The P content of plants incubated with Ct for 24 days was measured by ICP-MS and was calculated in ppm based on the shoot dry weight. Ct treatment significantly increased the P content of shoots (two-tailed t test, p < 0.01). Similar results were obtained from one additional independent experiment. (E) Illustration of the two-compartment co-cultivation system used for ³³P translocation experiments. Seven-day-old A. thaliana seedlings were transferred to half MS agar medium providing either high or low Pi conditions in square (12 × 12 cm) Petri plates. Two agar plugs with or without C. tofieldiae mycelium were placed into the small circular Petri plates. After 7 days, ³³P was added to the small plates, and the two-compartment system was further incubated for 17 days, when the photographs were taken and shoots harvested for scintillation counting.

Figure S4

GO-Term Enrichment Analysis of Arabidopsis thaliana Genes Induced after 24 Days Growth in Low Pi Conditions, Related to Figure 4 (A) GO-term enrichment analysis of Arabidopsis thaliana genes induced after 24 days growth in low Pi conditions. Based on RNA-Seq analysis of A. thaliana genes that were differentially expressed in Pi-deficient versus Pi-sufficient conditions (log2FC > 1, fdr < 0.05), Gene Ontology (GO) analysis showed that among the top 100 A. thaliana genes that were upregulated at 24 dpi under Pi-starvation conditions, nine were related to ‘cellular response to phosphate starvation’ (GO: 0016036). The color coding visually represents p value in each Go term box (Red box shows lower p value than yellow.). (B) Analysis of regulatory mutants of the A. thaliana phosphate starvation response (PSR). Ct-mediated and phosphate starvation-dependent activation of both Pht1;2 and Pht1;3 was abrogated in phr1 phl1 plants. q-RT-PCR analysis revealed the reduction of Pht1;2 expression in phr1 phl1 mutant plants relative to Col-0 wild-type plants grown in low Pi media at 24 dpi. Expression levels are shown relative to the mean expression of plant actin. (Col-0 with Ct versus phr1 phl1 with Ct, two-tailed t test, p < 0.01). (C) q-RT-PCR analysis revealed the reduction of Pht1;3 expression in phr1 phl1 mutant plants grown in low Pi media at 24 dpi. Expression levels are shown relative to the mean expression of plant actin. (Col-0 with Ct versus phr1 phl1 with Ct, two-tailed t test, p < 0.01). (D) Translocation of ³³P-labeled orthophosphate from Ct to A. thaliana shoots. Col-0 plants were grown on low Pi medium without (mock, n = 39) or with Ct (n = 40), phf1 mutants with Ct (n = 40), and phr1 phl1 double mutants with Ct (n = 40) in a two-compartment system. ³³P was added to the hyphal compartment and after 17 days ³³P -incorporation into shoots was measured by scintillation counting. Columns represent counting results in kBq ³³P /g dry weight (DW) of individual plants. The dotted line shows the median level of ³³P background counts from mock inoculations. (E) At 4 days after roots were inoculated with Ct (4 dpi), Ct biomass in phr1 phl1 mutant plants was significantly higher than in Col-0 wild-type plants grown under low Pi conditions (Col-0 versus phr1 phl1, two-tailed t test, p < 0.05). To measure fungal biomass by qPCR, 500 ng RNA from infected plants was used to amplify the Ct actin fragment (CtACT) and RNA amounts were normalized to plant actin (AT3G18780). (F) At 4 dpi, Ct biomass in mutant phf1 plants was significantly higher than in Col-0 plants grown under low Pi conditions (Col-0 versus phf1, two-tailed t test, p < 0.05). (G) Histogram showing the diameter of Ct and Ci colonies at 2 and 6 days after mycelial plugs of these fungi were transferred to MS medium containing 50 μM (low) or 625 μM (high) KH₂PO₄. Colony sizes of Ct or Ci were not significantly different between high (+) and low Pi (-) conditions. Bars = SE. (H) Schematic representation of the targeted replacement of secreted acid phosphatase gene CT04_08450 by homologous recombination. Flanking sequences upstream and downstream of the acid phosphatase gene were cloned into the T-DNA of binary vector pBIG4MRHrev adjoining the hygromycin (HYG) resistance cassette. The upstream and downstream flanking sequences of the incoming T-DNA undergo double cross-over homologous recombination with the target sequences, resulting in hygromycin-resistant mutants lacking the target gene. (I) Confirmation of the gene replacement event in fungal transformants Ct_KO1 and Ct_KO2 by PCR analysis using primer pairs 1 and 2. (J) Fungal growth in agar medium in which hydroxyapatite was the only P source. Both WT and the knockout strain Ct_KO1 were incubated for 3 days. Growth of the mutant was not significantly different to that of wild-type C. tofieldiae.

Figure S5

Tryptophan-Derived Metabolites Restrict Root Colonization by C. tofieldiae, Related to Figure 5 To measure fungal biomass by qPCR, 500 ng RNA from infected plants was used to amplify the C. tofieldiae (Ct) actin fragment (CtACT) and RNA amounts were normalized to plant actin (AT3G18780). (A) Under high Pi conditions at 4 dpi, Ct biomass was significantly higher in cyp79B2 cyp79B3 double mutants compared with Col-0 wild-type plants and the other tested plant defense mutants (Col-0 versus cyp79B2 cyp79B3, two-tailed t test, p < 0.05). (B) Quantification of Ct biomass in roots of Col-0 and ein2 pad4 sid2 dde2 quadruple mutant plants grown under high and low Pi conditions at 24 dpi. Under both conditions, fungal biomass in ein2 pad4 sid2 dde2 plants was not significantly higher than in Col-0 wild-type plants. (C) Inoculation of Ct onto the roots of cyp79B2 cyp79B3 mutant plants severely inhibited plant growth under high Pi conditions. The photograph was taken at 10 dpi.