Serendipita indica E5'NT modulates extracellular nucleotide levels in the plant apoplast and affects fungal colonization

Abstract

Extracellular adenosine 5'-triphosphate (eATP) is an essential signaling molecule that mediates different cellular processes through its interaction with membrane-associated receptor proteins in animals and plants. eATP regulates plant growth, development, and responses to biotic and abiotic stresses. Its accumulation in the apoplast induces ROS production and cytoplasmic calcium increase mediating a defense response to invading microbes. We show here that perception of extracellular nucleotides, such as eATP, is important in plant-fungus interactions and that during colonization by the beneficial root endophyte Serendipita indica eATP accumulates in the apoplast at early symbiotic stages. Using liquid chromatography-tandem mass spectrometry, and cytological and functional analysis, we show that S. indica secrets SiE5'NT, an enzymatically active ecto-5'-nucleotidase capable of hydrolyzing nucleotides in the apoplast. Arabidopsis thaliana lines producing extracellular SiE5'NT are significantly better colonized, have reduced eATP levels, and altered responses to biotic stresses, indicating that SiE5'NT functions as a compatibility factor. Our data suggest that extracellular bioactive nucleotides and their perception play an important role in fungus-root interactions and that fungal-derived enzymes can modify apoplastic metabolites to promote fungal accommodation.

Keywords: Piriformospora indica; DORN1; extracellular nucleotides; lectin receptor kinase LecRK‐I.9; purinergic signaling.

Figure 1. Identification of S. indica apoplastic proteins

1. Distribution of S. indica apoplastic proteins identified by LC‐MS/MS analysis from different symbiotic stages (5, 10, and 14 dpi) and in relation to proteins identified in CF of S. indica axenically grown in CM. In total, 102 S. indica proteins were identified in the APF of colonized barley roots. Of these, 33 proteins were present at all three time points. Twenty proteins were unique at 5 days postinfection (dpi), 4 at 10 dpi, and 21 at 14 dpi.

2. Heat map showing absolute counts of unique peptides for S. indica apoplastic proteins present at all three time points and in culture filtrate. S1/S2/S3 = biologically independent samples 1, 2, and 3; d = deglycosylated.

Figure 2. Serendipita indica colonization leads to increased eATP levels in barley and Arabidopsis roots

1. eATP levels in APF from S. indica‐inoculated and mock‐treated barley root samples collected at the predominantly biotrophic phase (3 and 5 dpi) as well as the predominantly cell death‐associated phase (7 and 10 dpi). Error bars show the standard error of the mean obtained from three independent biological replicates at P < 0.05 (*), 0.01 (**) analyzed by Student's t‐test. RLU: relative light units.

2. eATP levels measured from culture medium collected from mock‐treated and S. indica‐colonized Arabidopsis seedlings at 3, 5, 7, and 10 dpi. Error bars represent ±SE of the mean from three independent biological replicates. RLU: relative light units. Asterisks indicate significance at P < 0.05 (*), 0.01 (**) analyzed by Student's t‐test.

3. S. indica colonization of Arabidopsis dorn1‐3 mutant and the parental Col‐0^aq lines quantified by qPCR at 3, 5, 7, and 10 dpi. The ratio of fungal (SiTEF) to plant (AtUBI) amplicons representing fungal colonization levels in plant root tissue was calculated using gDNA as template and the 2^−ΔCT method. Error bars represent standard error of the mean of three technical replicates. The experiment was repeated three times for 3, 5, and 7 dpi with similar outcomes.

4. Transcript levels of S. indica E5′NT during colonization of barley and Arabidopsis at different symbiotic stages and in axenic culture. Error bars represent standard error of the mean of three independent biological replicates. CM = complex medium.

Figure 3. Serendipita indica PIIN_01005 encodes a secreted ecto‐5′‐nucleotidase (SiE5′NT)

1. Distribution of SiE5′NT orthologues across higher fungi. End nodes are color‐coded based on the presence (blue) or absence (red) of 5′NT genes in a particular fungal taxon. Numbers in parentheses besides the nodes specify the number of species that have 5′NT genes with respect to the total number of genomes analyzed. Left tree: distribution of 5′NT without signal peptide (SP) and GPI anchor. Right tree: E5′NT with SP and GPI anchor. The distribution shows that E5′NT genes are mostly present in Ascomycota such as Sordariomycetes (107/121), followed by Dothideomycetes (24/37) and Leotiomycetes (6/18). Few species of Eurotiomycetes (6/118) possess an E5′NT orthologue. In Basidiomycota, E5′NT members are only found in the class of Agaromycotina (12/85).

2. Comparison of the SiE5′NT structural homology model (green) with the crystal structure of human E5′NT (PDB id 4H2I, red) with 32% sequence identity. *: position of the loop involved in dimerization of human E5′NT.

3. Comparison of the SiE5′NT structural homology model (green) with the crystal structure of Thermus thermophiles 5′NT (2Z1A, blue) with 36% sequence identity.

Figure 4. The ecto‐5′‐nucleotidase activity of SiE5′NT in the apoplast of Arabidopsis leads to enhanced S. indica colonization similar to that of the dorn1‐3 (ATP receptor) mutant line

1. Ecto‐5′‐nucleotidase activity measured in membrane protein preparations of Arabidopsis plants expressing ^Pro35S::E5′NT (#303), ^Pro35S::SP_E5′NT:mCherry:E5′NT^woSP (#304), or ^Pro35S::mCherry (#305). E5′NT activity was measured after incubation with 100 μM of either ATP, ADP, or AMP. In the membrane protein preparations from ^Pro35S::E5′NT (#303) lines, phosphate release was specifically increased upon incubation with purines. Error bars represent the standard error of the mean from three technical repetitions. The Coomassie‐stained SDS–PAGE shows the protein pattern of the membrane fractions for the individual transgenic lines. Equal volumes were loaded. The experiment was repeated two times with similar results.

2. Confocal microscopy images of Arabidopsis roots expressing either cytosolic mCherry (#305) or ^Pro35S::SP_E5′NT:mCherry:E5′NT^woSP (#304) showing secretion of the E5′NT fusion protein. mCherry images show z‐stacks of 14 image planes of 1 μm each. Scale bar = 20 μm.

3. The transgenic Arabidopsis line ^Pro35S::E5′NT (#303) expressing untagged full‐length SiE5′NT was better colonized by S. indica. Error bars of the qPCR data represent ± SE of the mean from three independent biological replicates. Asterisks indicate significance (Student's t‐test, *P < 0.05).

4. S. indica induced eATP release in different Arabidopsis transgenic lines. Culture medium was collected from mock‐treated or S. indica‐inoculated seedlings growing in liquid medium at 5 dpi, and released eATP was measured. RLU: relative light units. Error bars represent ± SE of the mean from three biological replicates. Letters indicate significance to all other samples within the same treatment group (ANOVA, P < 0.05).

5. S. indica colonization of transgenic lines at 5 dpi. Error bars represent ± SE of the mean from three biological replicates. Letters indicate significance to all other samples within the same treatment group (ANOVA, P < 0.05).

6. Expression analysis of the eATP responsive gene At1g58420 measured by qRT–PCR. Error bars represent ± SE of the mean from three independent biological replicates (independent from those shown in Fig. 2C). Letters indicate significant groups (ANOVA, P < 0.01, for the line 305 P < 0.05).

Figure 5. Simulation of the effect of eATP release and its cleavage in the apoplast on the nutrient exchange between plant and fungus

1. Relative fluxes of phosphates (Pi) and sugars (C) across the fungal plasma membrane. Without additional Pi‐source in the apoplast (value 0 at the x‐axis, black dots), there is a constant flux of phosphate via the proton‐coupled phosphate (H/P) transporter from the fungus to the apoplast and a constant flux of sugar from the apoplast to the fungus. For better comparison, these fluxes were normalized to 1 and −1, respectively, and all other fluxes were calculated relative to these control values. With increasing ATP‐release and Pi‐production, the Pi‐efflux gets smaller and is zero at a relative Pi‐production rate of 1 (gray triangles). At higher ATP‐release and Pi‐production rates, the fungus imports Pi, i.e., the transport direction of phosphate has been inverted in comparison with the control condition (white square). The C‐flux via the proton‐coupled sugar (H/C) transporter is not affected by the additional Pi‐source (horizontal gray line).

2. Relative fluxes of Pi and C across the plant plasma membrane. Without additional Pi‐source in the apoplast (value 0 at the x‐axis, black dots), there is a constant flux of sugar via the H/C transporter from the plant to the apoplast (and thereafter to the fungus) and a constant flux of phosphate (coming from the fungus) via the H/P transporter from the apoplast to the plant. With increasing ATP‐release and Pi‐production, the Pi‐influx via the H/P transporter increases while the C‐flux via the H/C transporter is unaffected by the additional Pi‐source (horizontal gray line).

3. Phosphate fluxes across the plant plasma membrane. Besides the Pi‐uptake via the H/P transporter (B, black line, for clarity not shown in C), the plant loses Pi due to the ATP‐release (light gray line). The difference between Pi‐uptake and Pi‐loss is the net Pi‐balance of the plant (dark gray line). Without ATP‐release (value 0 at the x‐axis, black dots), the plant gains Pi. At a relative ATP‐release and Pi‐production of 1 (gray triangles), the plant release as much phosphate as it takes up, while at higher ATP‐release values, the plant loses Pi.

4. Schematic representation of the fluxes for three scenarios: (i) no ATP‐release and Pi‐production (left, black dots in A‐C); (ii) moderate ATP‐release, adenosine (A), and Pi‐production (middle, gray triangles in A–C); and (iii) high ATP‐release, A, and Pi‐production (right, white squares in A–C). If there is no ATP‐release (left), there is a constant flux of Pi from the fungus via the apoplast to the plant and a constant flux of sugars in the inverse direction. The energy for these fluxes is provided by the phosphate gradient between fungus and plant and the sugar gradient between plant and fungus. At a moderate ATP‐release and Pi‐production rate (middle, value 1 on the x‐axis, gray triangles in A–C), there is no Pi‐flux from the fungus to the apoplast anymore. The unchanged uptake of sugars is now energized by the proton pump of the fungus. In this condition, the plant retrieves the cleaved Pi originating from the ATP‐release via the H/P cotransporter, a transport which is energized by both the sugar gradient and the proton pump. At a high ATP‐release and Pi‐production rate (right, value 2 on the x‐axis, white squares in A–C), there is a larger Pi‐uptake by the plant and a Pi‐uptake by the fungus. Both processes are energized by larger activities of the proton pumps.

Figure 6. Schematic model showing the potential interference of SiE5′NT with apoplastic eATP signaling and nucleotide salvage pathway in Arabidopsis

The schema was modified from Ref. 5. ENT = equilibrative nucleoside transporters, PAP = purple acid phosphatases, Apy = ectoapyrases or nucleoside triphosphate diphosphohydrolases (NTPDases), NSH = nucleoside hydrolases, 5NT = hypothetical 5′‐nucleotidases, PUP = hypothetical purine permeases.