Plant and endophyte growth conditions

Wild-type Col-0 Arabidopsis thaliana, fah1–2 (N6172, ferulic acid 5-hydroxylase 1, AT4G36220) [48], and tt4–11 (N2105573, chalcone synthase (CHS), SALK_020583, AT5G13930) [49] mutants were purchased from Nottingham Arabidopsis Stock Centre. The tt4–11 mutant is deficient in the CHS protein, thus lacking flavonoids, with increased sinapic acid ester content [50]. The fah1–2 mutant lacks the F5H, ferulate-5-hydroxylase, and is deficient in sinapoylmalate and sinapoyl choline [51], required for plant defence [34] and UV-B protection [48,52]. Arabidopsis seeds were surface sterilized with 70% (v/v) EtOH (Linegal Chemicals) for 5 min, 2.5% (v/v) NaOCl for 5 min, and washed 4 times with sterile distilled water. Seeds were sown in vitro on Petri dishes in the 1/4 MS medium without vitamins (Sigma-Aldrich) with 0.75% (w/v) sucrose and transferred to 4^oC for two days. The plates were put into a custom-made environmental growth chamber. White illumination of 225 μmol m^-2 s^-1 PAR (measured through a Petri dish cover) was provided by LEDs (EPILEDs, 5000 K) supplemented with UV-A of 13 μmol m^-2 s^-1 from fluorescent tubes (Philips TL 8 W BLB) covered with either a Lee #226 gel filter (blocks UV-A and UV-B), or with a polyester film (125 μm thick, Autostat CT5, MacDermid), which blocks UV-B but passes UV-A and visible light. The applied UV-A radiation was contained in the UV-A1 waveband (350–400 nm) (S1 Fig). In our study, the ratio between photon irradiance of UV-A in the 350–400 nm waveband to PAR was close to the ratio in the standard 1.5 air mass sunlight spectrum (0.051 in ASTM G173 - 03) [53]. It was selected as an approximation of clear sky conditions at the mid-latitudes during the vegetative period. On the 8^th day of cultivation, plants were transferred to the Modified Strullu Romand medium (MSR) and inoculated with one of the endophytes. Identification of fungal strains was based on the sequences of ITS DNA [54,55] described in [56,57]. The NCBI database accession numbers for the strains were Paraphoma chrysanthemicola - OR077971, Phomopsis columnaris - OR077985, Diaporthe eres - OR078011, Mucor sp. - OR078038, and Sporobolomyces ruberrimus - ON331994. Endophytic fungi mycelia were cultivated on Petri dishes with PDA medium (GRASO Biotech) at 25°C for 14 days in darkness in a standard incubator. Plant roots were inoculated using plugs of the approximate size of 3x3 mm placed between the roots with the mycelium oriented towards the medium for Paraphoma chrysanthemicola, Phomopsis columnaris, and Diaporthe eres. Mucor sp. spores were picked from 25 mm² of the plate (concentration approximately 0.16–0.18 x 10⁶ mm^-2 on the plate) with an inoculation loop and placed between the roots of A. thaliana plants. The yeast cell culture of Sporobolomyces ruberrimus (OD₆₀₀ approximately 7.0) was spread directly on the whole root surface with an inoculation loop. Mock-inoculated plants served as a control. Plants were cultivated for the next 10 days. The plant condition was then assessed by measuring wet shoot and root weight. Root morphology parameters were acquired and analysed using WinRHIZO Arabidopsis software (Regent Instruments Inc). The total length of the root system, average root diameter, and total root volume were measured. For the experiment with the wild type, tt4, and fah1 mutants, the main root length was marked every 12 h after inoculation for 5 consequent days and analysed with ImageJ.

Construction of a phylogenetic tree

The phylogeny was inferred using the Maximum Likelihood method and Tamura-Nei (1993) model [58] of nucleotide substitutions, and the tree with the highest log likelihood (-20 328.52) is shown. The initial tree for the heuristic search was selected by choosing the tree with the superior log-likelihood between a Neighbor-Joining (NJ) tree [59] and a Maximum Parsimony (MP) tree. The NJ tree was generated using a matrix of pairwise distances computed using the p-distance [60]. The MP tree had the shortest length among 10 MP tree searches, each performed with a randomly generated starting tree. The analytical procedure encompassed 25 coding nucleotide sequences using 1^st, 2^nd, 3^rd, and non-coding positions with 2577 positions in the final dataset. Evolutionary analyses were conducted in MEGA12 [61] utilizing up to 8 parallel computing threads.

Arabidopsis pot cultures

Wild-type Arabidopsis seeds were sown on Jiffy-7 peat pellets and then stratified at 4°C for 2 days. Plants were grown in a cultivation chamber in the same conditions as in vitro cultures. When the plants had developed the first pair of leaves (on the 9^th day), they were transplanted into a pot with a substrate of perlite: sand 1:1 (v/v). Plants in the perlite: sand medium were watered with 30 ml of Hogland’s medium (Sigma-Aldrich), 3 times a week. Liquid cultures of S. ruberrimus were carried out for 3 days at 28°C with shaking at 120 rpm in 100 ml flasks with 30 ml of 2% (w/v) malt extract. On 3^rd day, optical density at 600 nm reached approximately 7.0. Then 90 ml of the yeast culture was centrifuged at 4500 g for 10 min at room temperature. The resulting pellet was washed 3 times by suspending it in 30 ml of 0.9% (w/v) NaCl and centrifuging at 4500 g for 10 min at room temperature. Yeast culture was finally suspended in 200 ml of 0.9%(w/v) NaCl and used for watering the plants. Plants were inoculated with 10 ml of the suspension of S. ruberrimus on the 12^th and 19^th day of development, i.e., on the 3^rd and 10^th day after transplanting into the new medium. The control, mock-inoculated plants were watered with 10 ml of 0.9% (w/v) NaCl solution.

Anthocyanin, flavonol, and chlorophyll content analysis

For anthocyanin content estimation, the protocol was adapted from [62]. Shoot samples were grounded to powder in liquid nitrogen, suspended in 80% (v/v) methanol solution in proportion 1:10 (mass/volume), and homogenized in TissueLyser II (Qiagen) with glass beads at the oscillation frequency of 30 Hz for 5 min. TissueLyser II adapters were precooled at -20°C for 24 h. Samples were centrifuged at 4°C in 14000 g for 15 min. 250 μl of each sample was resuspended in 4.5 ml of 2% (v/v) HCl solution in water and 250 μl of 1% (v/v) HCl solution in 95% (v/v) ethanol and mixed by inverting. After incubation in the darkness for 15 min, the anthocyanin content was measured spectroscopically at 532 nm in three independent technical replicates. The concentration was estimated using a calibration curve prepared for malvidin 3-glucoside [62]. Flavonols and chlorophyll contents were measured in the upper epidermis of the 6^th, 7^th, and 8^th intact Arabidopsis leaf using the Dualex Scientific fluorimeter (Force-A).

Laser scanning confocal microscopy

The presence of the endophyte was analysed by confocal microscopy using WGA-Texas Red stained fungi mycelium as in [63]. Root and leaf fragments were cleared in 10% (w/v) KOH for 30 min, then rinsed in sterile water for 5 min. The tissues were incubated for 30 min with 20 μg/ml WGA-Texas Red (a stock solution of WGA-Texas Red 1 mg/ml in 0.01 M Phosphate, 0.15 M NaCl, 0.05% (w/v) sodium azide) and then washed three times for 10 min in the phosphate buffer. The samples were observed using the Axio Observer.Z1 inverted microscope (Carl Zeiss, Jena) equipped with the LSM 880 confocal module. Plan-Neofluar 40x objective (NA 1.3) was used with oil immersion. Texas Red fluorescence was excited with the 594 nm He-Ne laser. Emission in the range of 599 nm – 630 nm was recorded as the red channel.

qPCR -based quantification of endophytic fungi

To quantify the amount of each fungus, a qPCR method was employed to detect endophyte-specific fungal translational elongation factor 1α (tef1α) or chitinase (S. ruberrimus) DNA. The roots and leaves of A. thaliana were surface sterilized. Plant organs were rinsed three times with sterile distilled water, incubated in 0.5% (v/v) sodium hypochlorite solution (NaOCl) (Sigma-Aldrich) with Tween20 (Bio-Shop) for 1 min, rinsed with sterile distilled water for 2 min, and incubated in 75% (v/v) ethanol solution for 30 s. Then the plant material was rinsed three times with sterile water (1 min, 5 min, 1 min). Air-dried samples were immediately frozen in liquid nitrogen and kept at -80°C. Root and shoot samples were grounded separately in liquid nitrogen. DNA extraction was performed according to [64] modified by [65]. DNA was extracted using 700 μl of 2% (w/v) CTAB solution (Sigma-Aldrich) in 0.1 M TRIS-HCl pH 8.0, 1.4 M NaCl, 0.02 M EDTA, 1% PVP with 1.5 μl of β-mercaptoethanol (Sigma-Aldrich). Samples were vortexed vigorously and incubated with shaking at 65°C for 60 min. 5 μl of RNAse solution was added to remove RNA and incubated for 10 min at room temperature. The samples were centrifuged at room temperature for 5 min at 15000 g. The supernatants were transferred to a new Eppendorf tube containing 700 μl of chloroform and centrifuged for 30 min at 15000 g at room temperature. To precipitate DNA, the upper aqueous phase was transferred to a new Eppendorf tube with 600 μl of 70% (v/v) ice-cold isopropanol solution (-20^oC), and incubated overnight at -80°C. The following day, the samples were thawed on ice and centrifuged for 40 min at 15000 g at 4°C. The supernatants were removed, and the DNA pellets were washed with 70% (v/v) ethanol, then 96% ethanol, and centrifuged for 5 min at 15000 g. The DNA pellet was re-suspended in 30 μl of TE buffer and incubated with mixing for 1 h at 37°C. Then 0.1 volume of 3 M sodium acetate (Sigma-Aldrich) and 2.5 volume of 70% (v/v) ice-cold isopropanol solution were added to the DNA samples. After overnight incubation at -80°C, the samples were thawed on ice, centrifuged for 40 min at 15000 g at 4°C, washed with 70% (v/v) ethanol, then with 96% (v/v) ethanol, and centrifuged for 5 min at 15000 g. The DNA pellet was re-suspended in 30 μl of the TE buffer pH 8.0 (Invitrogen), and incubated with shaking for 1 h at 37°C. qPCR was performed with a SYBR Green JumpStart Taq Ready Mix (Sigma-Aldrich) and a Biorad Touch CFX96 Real-Time PCR Detection System. 20 ng of total DNA extracted from samples and 0.33 μM of each primer from the pair were used for a single reaction, performed in triplicates. The primers are listed in the Table S1 of the S1 Appendix. After initial denaturation at 95°C for 4 min, 50 cycles of denaturation at 95°C for 10 s, followed by annealing at the temperature specific to the respective primer for 15 s, and elongation at 72°C for 20 s were performed.

Real-time PCR

RNA was isolated from above-ground plant parts. RNA isolation and real-time PCR were performed according to [66], except that oligodT primers were used for RNA reverse transcription. Primer sequences are listed in the Table S2 of the S1 Appendix. Each sample was run in three technical replicates. The relative expression of each gene in a sample was determined using the mean value of Ct for all samples. For reference genes, the levels of SAND and PDF2 were quantified, based on [67]. Expression levels were normalized using factors calculated by geNorm v3.4 [68].

Lipid peroxidation measurements

The leaf rosettes were grounded to a powder in a mortar in liquid nitrogen without allowing the tissue to thaw. Each sample was divided into 3 technical replicates of 100 mg. If the weight of the sample was too low, 50 mg of grated plant tissue was used. The sample was suspended in a 0.1% (v/v) aqueous solution of trichloroacetic acid in a ratio of 1:10 and homogenized using glass beads in a TissueLyser II (Qiagen) for 10 min at 30 Hz. Before use, the TissueLyser II block was cooled at -20°C for at least 24 h. The sample was centrifuged at 10 000 g at 4°C for 5 min. After centrifugation, the supernatant was removed and mixed by pipetting with freshly prepared 0.5% (w/v) thiobarbituric acid (TBA) dissolved in 20% (w/v) trichloroacetic acid. One part of the supernatant was mixed with 4 parts of the TBA solution. Samples were then incubated in a thermoblock at 95°C for 30 min. Samples were chilled on ice and then centrifuged at 10000 g at 4°C for 5 min. The concentration of malondialdehyde (MDA), a lipid peroxidation marker, was measured using the reaction of TBA with MDA, which yields coloured MDA-(TBA)₂ adduct as in [69]. The adduct concentration was determined spectrophotometrically by measuring the absorbance of the sample at the wavelength of 532 nm in two chemical replicates.

Measurement of the photosynthetic parameters of leaves

Photosynthetic parameters were analysed in the 6^th and 7^th rosette leaf using Dual-PAM-100 fluorimeter (Heinz Walz GmbH). To determine the slow kinetics with the induction curve of chlorophyll, whole plants were dark-adapted for 30 min before the measurement. The F_v/F_m ratio was determined using weak measuring light (F₀) and a saturating pulse (F_m) of 10000 µmol m^-2 s^-1 for 600 ms. After a 50 s pause, red actinic light (AL) of 75 μmol m^-2 s^-1 PAR was turned on. Then sequential saturation pulses (SP) of 10000 µmol m^-2 s^-1 were applied, which lasted 600 ms at 20 s intervals. Light curves were measured on irradiated plants. Leaves were irradiated with red actinic light (AL) of increasing irradiance of 0, 11, 18, 27, 58, 100, 131, 221, 344, 536, 830 μmol m^-2 s^-1, with each phase lasting 30 s. Pulses of saturating light lasted 600 ms and were applied every 30 s.

Statistical analysis

The position of plates and pots with plants in the cultivation chamber was randomized. Plates were repositioned daily to reduce systematic errors due to inhomogeneity in growth conditions in the chamber. Statistical analysis was performed using the R software [70]. No data points were rejected as outliers. The details of the analysis and the obtained results are provided in the S2 Appendix. To avoid pseudoreplication, when measurements of the same property were obtained for multiple plants growing on the same plate, they were averaged before analysis and the plate means were treated as single replicates. Linear models were fitted using the nlme package [71,72], with the gls (when all factors were fixed) or lme command (when a random factor corresponding to a batch of plants was included), with the interaction between fixed factors. In each case, the linear models allowed for unequal variances between groups. Preliminary tests of variance homogeneity were not used to avoid distortion of the significance levels [73]. As this work explores whether the presence of UV-A modulates the effect of endophyte inoculation, the significance of interaction terms, as examined with ANOVA, was important for the interpretation of the results. Type III ANOVA was calculated. In type III ANOVA, the significance of a main effect is examined comparing the full model with a reduced model, in which the main effect of the factor of interest is removed, but its interactions are retained. This approach has been suggested to be more appropriate when substantial interactions are likely [74] which is often the case in studies involving stress factors [75]. However, as the number of replicates was similar in all groups, the choice of ANOVA type did not substantially affect the results of the analysis. The analysis of the qPCR results and the gene expression levels was performed on log-transformed data, other data were not transformed. Differences between cell means of responses or log-transformed responses were examined using the emmeans package [76]. The approximate number of degrees of freedom was calculated with Satterthwaite’s method, which is designed for data in which variance may differ between groups. The Benjamin-Hochberg or Hommel method was used to adjust p values for cell mean differences, as specified in the S2 Appendix.

Colonisation of plants by endophytes

To examine whether the UV-A radiation affects the colonisation of plant tissues by endophytic fungi, Arabidopsis thaliana wild-type seedlings were grown under the illumination of ca. 225 µmol m^-2 s^-1 PAR with or without supplementation of 13 µmol m^-2 s^-1 UV-A. After eight days, plants grown in the presence of UV-A showed reduced root length compared to those grown only under PAR (S2 Fig., S2 Appendix, Table S5). At this time point the plants were inoculated with Paraphoma chrysanthemicola (Fig.1. a), Phomopsis columnaris (Fig. 1. b), Diaporthe eres (Fig. 1. c), Mucor sp. (Fig. 1. d), Sporobolomyces ruberrimus (Fig. 1. e), and cultivated in the Modified Strullu Romand medium (MSR) for ten days (S3 Fig.). Plants that were not inoculated served as a control. Plant colonisation with the endophytes was assessed separately in shoots and roots.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig. 1. Fungi mycelia used in the study, cultivated on the PDA medium

(a) Paraphoma chrysanthemicola, (b) Phomopsis columnaris, (c) Diaporthe eres, (d) Mucor sp., and yeasts (e) Sporobolomyces ruberrimus. The top view is in the upper row, the bottom view is in the lower row. Roots of Arabidopsis mock-inoculated (f) or inoculated with endophytes Paraphoma chrysanthemicola (g), Phomopsis columnaris (h), Diaporthe eres (i), Mucor sp. (j), Sporobolomyces ruberrimus (k), grown in the presence or absence of UV-A radiation, stained with Wheat Germ Agglutinin lectin conjugated with Texas Red. Laser-scanning confocal images show the transmitted light channel merged with the channel of red fluorescence (599–630 nm). Plants not inoculated with the endophyte serve as a control for natural autofluorescence. Scale bars = 50 μm. (l) A phylogenetic tree with the highest log likelihood inferred using the Maximum Likelihood method and Tamura-Nei model of nucleotide substitutions. The scale bar shows the length of the branch representing the amount of genetic change of 0.10.

https://doi.org/10.1371/journal.pone.0323576.g001

The localization of endophytes within plant tissues was analysed using Wheat Germ Agglutinin lectin conjugated with Texas Red. The fluorescently labelled lectin selectively binds to chitin, the main polysaccharide of the fungal cell walls. Fungal mycelia were observed only in roots inoculated with Mucor sp., Paraphoma chrysanthemicola, Diaporthe eres, and Phomopsis columnaris (Fig. 1g,h,i,j, S4 Fig.) regardless of the light conditions during the growth of plants. However, yeast cells of Sporobolomyces ruberrimus were present in the roots and shoots of inoculated plants (compare Fig.1k and S4 Fig.). A non-specific signal was observed in control, mock-inoculated samples (Fig. 1f, S4 Fig.). Further, to confirm and quantify the fungal abundance in plant tissues, the qPCR-based method was utilised. The presence of fungal chitinase (for S. ruberrimus) or translational elongation factor 1α (tef1α, for other fungi) genes in DNA extracted from surface-sterilized plants was investigated (Fig. 2). In the case of Paraphoma chrysanthemicola, Phomopsis columnaris, and Mucor sp., the fungus was found in inoculated roots (Fig. 2a,b,d). Colonisation by Diaphorte eres and the yeast Sporobolomyces ruberrimus was detected both in shoots and roots (Fig. 2c,e). The effect of interaction between the inoculation and light conditions on qPCR values was not significant for any of the studied endophytes (ANOVA results in the S2 Appendix, Table S1), indicating that the presence of UV-A does not substantially influence endophyte colonisation.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig. 2. Endophyte inoculation of Arabidopsis thaliana, grown in the presence or absence of UV-A, assessed by the qPCR method.

The ratio of fungal DNA to plant DNA in roots and shoots of plants inoculated with Paraphoma chrysanthemicola (a), Phomopsis columnaris (b), Diaporthe eres (c), Mucor sp. (d), Sporobolomyces ruberrimus (e). The experiment was performed in four biological replicates. Brackets show the significant differences in means of log-transformed qPCR measurements between inoculated and mock-inoculated samples, from the same organ and light conditions (* 0.01 < p ≤ 0.05, ** 0.001 < p ≤ 0.01, *** p ≤ 0.001, adjusted for multiple comparisons with Hommel’s method). The set of all tested contrasts is in the S2 Appendix, Table S1.

https://doi.org/10.1371/journal.pone.0323576.g002

Plant growth parameters

To investigate the interaction between the effect of UV-A presence (two levels, UV-A supplementation or visible light only) and the effect of endophyte inoculation (six levels, including the mock-inoculated control) on the host plant growth, we examined the selected plant growth parameters in plants in all combinations of the levels of both factors. The presence of UV-A radiation resulted in shoot growth inhibition in mock-inoculated control plants as well as in inoculated plants (regardless of endophytic species) (Fig. 3a). The effects of endophyte inoculation and UV-A radiation on the fresh weight of shoots were both significant (F_(5,91) = 9.82, p = 2.5·10^-14; F_(1,91) = 81.93, p = 2.5·10^-14, respectively), though their interaction was insignificant (F_(5,91) = 0.53, p = 0.75) (ANOVA results in the S2 Appendix, Table S2). The fresh shoot weight was significantly lower in plants supplemented with UV-A than in plants grown only in visible light, irrespective of the type of inoculation (Fig. 3. a, p values in the S2 Appendix, Table S2). Inoculation with Mucor sp. and Paraphoma chrysanthemicola resulted in a decrease in the shoot weight, in plants grown with and without UV-A. To further characterise the impact of UV-A on the aerial part of the plants, the anthocyanin content was assessed by the spectrophotometric method (Fig. 3. b). Anthocyanin elevation may result from UV irradiation, but also other stress factors, such as the presence of endophytes. The UV-A supplementation affected the anthocyanin levels in plants (F_(1,82) = 22.4, p = 9.1·10^-6). However, the effect of the endophyte inoculation (F_(5,82) = 32.24, p = 0.058) and the interaction effect (F_(1,87) = 1.11, p = 0.36) were not significant. While the mean anthocyanin content in our sample was greater for all endophytes in plants acclimated to UV-A than those grown in visible light alone, the differences were not statistically significant (Fig. 3. b). On the day of inoculation, plants acclimated to UV-A exhibited shorter roots than those grown only in visible light (S2 Fig.), but the effect of light conditions was not significant after the additional ten days of growth in the presence of fungi. The presence of endophyte affected fresh root weight (F_(5,91) = 2.8, p = 0.021) (Fig. 3. c), root system length (F_(5,90) = 7.8, p = 4·10^-6) (Fig. 3. d), average root volume (F_(5,90) = 6.2, p = 7.1·10^-5) (Fig. 3. e), and average root diameter (F_(5,90) = 18.5, p = 1.3·10^-12) (Fig. 3. f). However, the interaction of light conditions and the endophyte presence was nonsignificant for the root fresh weight, length, and volume. The means of measured root system length and root volume in our sample were in general smaller for inoculated plants than for the control, though only in some cases were the differences significant. In particular, when UV – supplementation was applied, both average root volume and the length of the root system were significantly smaller in plants inoculated with Diaporthe eres than in the mock-inoculated plants (p = 0.013 and p = 0.00066 for volume and length, respectively) (Fig. 3d,e). The sample means of average root diameter (Fig. 3. f) were in almost all cases greater in plants inoculated with the endophytes than in the mock-inoculated ones, with the difference being significant for Paraphoma chrysanthemicola, Diaporthe eres, and Sporobolomyces ruberrimus in both types of light conditions (p values in the S2 Appendix, Table S2).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig. 3. Fresh weight of shoots

(a), shoot anthocyanin content (b), fresh weight of roots (c), root system length (d), average root volume (e), and average root diameter (f) of 18-day-old plants, either mock-inoculated (control) or inoculated with Paraphoma chrysanthemicola, Phomopsis columnaris, Diaporthe eres, Mucor sp., Sporobolomyces ruberrimus. Plants were grown in the presence or absence of UV-A. The experiment was performed in 9 biological replicates (plates) for every combination of inoculation type and light conditions. Black brackets show the significant differences in means between plants inoculated with the same endophyte, but subject to different light conditions. Yellow (VIS) and violet (VIS + UV) brackets show significant differences in means between inoculated samples and the mock-inoculated control (* 0.01 < p ≤ 0.05, ** 0.001 < p ≤ 0.01, *** p ≤ 0.001, adjusted for multiple comparisons with the Benjamin-Hochberg method). The set of all tested contrasts is in the S2 Appendix, Table S2.

https://doi.org/10.1371/journal.pone.0323576.g003

Expression of genes involved in plant defence and stress in Arabidopsis above-ground plant parts

Plant-microorganism interactions necessitate adaptations in plant secondary metabolism, including the activation of phenolic compound production [77] and the fine-tuning of defence related hormone homeostasis. A well-documented aspect of this adaptation is the down-regulation of salicylic acid-related plant defence in mutualistic plant-microorganism interactions [39]. To investigate the influence of UV-A radiation on plant responses to endophytes, we examined the expression levels of genes involved in the synthesis of flavonoids, [78] PHENYLALANINE AMMONIA-LYASE 1 (PAL1) and CHALCONE SYNTHASE (CHS) [78], both of which are known to be induced by UV radiation. The effect of endophyte inoculation on CHS transcript, as tested with ANOVA, was significant (F_(1,96) = 3.28, p = 0.009), though the differences in means between particular groups were not significant (Fig. 4. A, p values in the S2 Appendix, Table S3). PAL1 and ICS1 are involved in different pathways of salicylic acid biosynthesis, contributing to local and systemic acquired resistance [79]. The expression of both genes was influenced by the endophyte inoculation (F_(5,96) = 4.32, p = 0.0014 for PAL1; F_(5,96) = 11.5, p = 9.7·10^-9 for ICS1) and UV-A (F_(1,96) = 5.75, p = 0.018 for PAL1; F_(1,96) = 9.48, p = 0.0027 for ICS1), though the interactions of these factors were not significant (p = 0.86 for PAL1; p = 0.62 for ICS1) (Fig. 4b,c). In most cases, the differences between means for particular groups were not significant and were small in magnitude. The presence of Paraphoma chrysanthemicola and Sporobolomyces ruberrimus increased ICS1 levels regardless of light conditions (Fig. 4. c). PDF1.2, an ethylene- and jasmonate-responsive plant defensin, is upregulated upon fungal infection of leaves [80]. We observed that the PDF1.2 transcript level was affected by the presence of the endophytes (F_(5,96) = 12.1, p = 4.5·10^-9), moreover, the influence of the endophyte was modified by the UV-A supplementation (F_(5,96) = 2.93, p = 0.017 for the interaction term). In particular, the presence of Paraphoma chrysanthemicola, Diaphorte eres, and Sporobolomyces ruberrimus significantly elevated the PDF1.2 transcript levels as compared to the mock-inoculated control only when light was supplemented with UV-A (Fig. 4. d; p values in the S2 Appendix, Table S3).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig. 4. Relative transcript levels of CHS

(a), PAL1 (b), ICS1 (c), and PDF1.2 (d) genes in leaves of 18-day-old plants, either mock-inoculated (control) or inoculated with Paraphoma chrysanthemicola, Phomopsis columnaris, Diaporthe eres, Mucor sp., Sporobolomyces ruberrimus. Plants were grown in the presence or absence of UV-A. The experiment was performed in 9 biological replicates. Yellow (VIS) and violet (VIS + UV) brackets show significant differences in means between inoculated samples and the mock-inoculated control (* 0.01 < p ≤ 0.05, ** 0.001 < p ≤ 0.01, *** p ≤ 0.001, adjusted for multiple comparisons with the Benjamin-Hochberg method). The set of all tested contrasts is in the S2 Appendix, Table S3.

https://doi.org/10.1371/journal.pone.0323576.g004

Expression of genes involved in short-wavelength radiation signalling

We also investigated whether biotic factors may affect, similarly to light, the expression of photoreceptor genes and in consequence alter plant physiology. We examined whether the presence of endophyte affects the expression levels of genes encoding plant photoreceptors sensitive to the UV-A range (S5 Fig.; p values in the S2 Appendix, Table S3). The main effects of the endophyte inoculation (F_(5,96) = 2.83, p = 0.02) and the UV-presence (F_(1,96) = 4.48, p = 0.037) were significant for the PHOT2 gene, but the interaction effect was not (p = 0.401), while the main effect of UV-A was significant for PHOT1 (p = 0.033). No changes in relative transcript levels were observed for CRY1, CRY2, and UVR8 genes (S5 Fig.; p values in the S2 Appendix, Table S3).

Physiological parameters of Arabidopsis grown in pot cultures inoculated with S. ruberrimus

We detected Sporobolomyces ruberrimus in shoots, using both microscopic examination and qPCR. To determine whether the UV-A treatment affects the influence of the endophyte on host photosynthetic efficiency, we performed experiments in pot cultures, measuring several photosynthetic parameters in mature Arabidopsis leaves grown in only visible light or acclimated to UV-A radiation, in the same conditions as the in vitro experiment (Fig. 5, p values in the S2 Appendix, Table S4). Arabidopsis plants cultivated in pots filled with perlite and sand substrates were inoculated twice with the yeast suspension. After 26 days of growth, flavonol (Fig. 5. b) contents were affected by UV-A in a manner dependent on the presence of S. ruberrimus (F_(1,69) = 6.32, p = 0.014 for the interaction effect). Lipid peroxidation was assessed with the MDA (malondialdehyde) assay, measuring the concentration of the final product of lipid peroxidation. Neither the presence of S. ruberrimus nor UV-A changed MDA levels, (Fig. 5. c). We also used a Dual Pam Fluorimeter in a saturation pulse mode, to register time-dependent changes in chlorophyll fluorescence of Arabidopsis plants inoculated and mock-inoculated with S. ruberrimus grown with or without UV-A supplementation. No substantial differences in the photosynthetic efficiency parameters: Y(II), ETR, and NPQ of leaves were observed for induction (Fig. 5d-f) and light (Fig. 5g-i) curves.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig. 5. Chlorophyll content

(a), flavonol content (b), MDA (c) and photosynthetic parameters (d-i) in Arabidopsis thaliana pot cultures inoculated with Sporobolomyces ruberrimus grown in the presence or absence of UV-A. (a - c) The experiment was repeated six times. Each dot represents a measurement obtained from a single plant (in a – b, the mean of values recorded for the 6^th, 7^th, and 8^th leaf). The results of the statistical analysis are in the S2 Appendix, Table S4. (d - f) Induction curves of Arabidopsis pot cultures inoculated with S. ruberrimus with standard error. d – Y(II), e – NPQ, f – ETR. Dark-adapted plants were irradiated with the initial Saturating Pulse (SP) of 10 000 µmol m^-2·s^-1 for 600 ms, then with a delay of 50 s, Actinic Light (AL) of 75 µmol m^-2·s^-1 was applied, followed by a set of SP. (g – i) Light curves of Arabidopsis pot cultures inoculated with S. ruberrimus with standard error. g – Y(II), h – NPQ, i – ETR. Actinic light of increasing irradiance (0, 11, 18, 27, 58, 100, 131, 221, 344, 536, 830 µmol m^-2·s^-1) was applied within 30 s phases. Saturating pulses lasted 600 ms.

https://doi.org/10.1371/journal.pone.0323576.g005

Inoculation of phenolic pathway mutants with S. ruberrimus and P. chrysanthemicola

To understand the role of flavonoid accumulation in colonisation by S. rubberrimus, we used the tt4–11 mutant deficient in the CHS protein, which lacks flavonoids but shows increased sinapic acid esters [50]. Given the role of sinapate esters in plant defence [34] and UV-B protection [48,52], we also examined the fah1–2 mutant, which lacks the F5H, ferulate-5-hydroxylase, a cytochrome P450-dependent monooxygenase, and, in consequence, is deficient in sinapoylmalate and sinapoyl choline [51]. We also examined P. chrysanthemicola as a mycelium-forming fungus that affected the growth of plants in our experiments. The qPCR analysis of the amount of endophytes (Fig. 6a, b, the S2 Appendix, Table S6) did not reveal any differences between plant lines in their susceptibility to colonization, as evidenced by the non-significant effect of interaction between the plant line and inoculation effect on the qPCR readings (F_(2,22) = 0.12, p = 0.89 for P. chrysanthemicola; F_(2,22) = 0.65, p = 0.53 for S. rubberrimus). Likewise, we did not detect a significant effect of the presence of UV-A on colonisation (F_{(1,22) =} 1.60, p = 0.22 for P. chrysanthemicola; F_(1,22) = 0.01, p = 0.94 for S. rubberrimus). Fresh shoot weight (Fig. 6. C, the S2 Appendix, Table S7) depended on the plant line (F_(2,121) = 8.59, p < 0.001) and was affected both by the presence of endophytes (F_(2,121) = 46.2, p < 0.001) and UV-A (F_(1,121) = 10.8, p = 0.0014). However, no interaction between these factors was found. Lower shoot weight was recorded in plants inoculated with P. chrysanthemicola and S. rubberrimus than in mock-inoculated ones, regardless of the plant genotype and light treatment. The differences were statistically significant in most cases, as specified in the S2 Appendix, Table S7. Fresh root weight was affected by the presence of endophytes (F_(2,121) = 72.62, p < 0.001) and, to a lesser extent, was dependent on plant line (F_(2,121) = 18.69, p < 0.001) (Fig. 6. d). Neither the effect of UV nor interactions between factors were significant. Seedlings inoculated with either of the endophytes exhibited smaller root weights than mock-inoculated ones. In most cases, these differences were statistically significant and substantial in magnitude (S2 Appendix, Table S7). We also analysed the main root elongation during the first five days after inoculation with the endophyte (S6 Fig.). The main root length 100 h after the inoculation was significantly affected by the presence of endophytes (F_(2,123) = 176.3, p < 0.001), UV-A (F_(1,123) = 30.8, p < 0.001) and by the plant line (F_(2,123) = 3.46, p = 0.03) (Fig. 6. e). Notably, the effect of UV-A supplementation depended on the presence of endophyte, as indicated by the significant interaction term (F_(2,123) = 5.95, p = 0.003). In plants inoculated with P. chrysanthemicola, a UV-A-induced decrease in main root length was observed in the wild-type and tt4–11 mutant, but not in the fah1–2 plants. Inoculation with S. rubberrimus resulted in a much shorter main root than in the control conditions, regardless of the genotype and light treatment.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig. 6. Endophyte inoculation of Arabidopsis thaliana, wild type, fah1, and tt4 mutants, grown in the presence or absence of UV-A, assessed by the qPCR method.

(a-b) The ratio of fungal DNA to plant DNA in roots and shoots of plants inoculated with Paraphoma chrysanthemicola (a) and Sporobolomyces ruberrimus (b). The experiment was performed in three biological replicates. Brackets show the significant differences in means of log-transformed qPCR measurements between inoculated and mock-inoculated samples light conditions (* 0.01 < p ≤ 0.05, ** 0.001 < p ≤ 0.01, *** p ≤ 0.001, adjusted for multiple comparisons with Hommel’s method). The set of all tested contrasts is in the S2 Appendix, Table S6. (c - d) Fresh weight of shoots (c) and roots (d) 18-day-old plants, either mock-inoculated (control) or inoculated with Paraphoma chrysanthemicola, or Sporobolomyces ruberrimus. Weight was measured 128 h after inoculation (e) Main root length increment of Arabidopsis thaliana, wild type, fah1, and tt4 mutants, 100 hours post inoculation with Paraphoma chrysanthemicola, or Sporobolomyces ruberrimus. Plants were grown in the presence or absence of UV-A. (c – e) The experiment was performed in 8 biological replicates (plates) for every combination of inoculation type and light conditions. Black brackets show the significant differences in means between plants inoculated with the same endophyte, but subject to different light conditions. Yellow (VIS) and violet (VIS + UV) brackets show significant differences in means between inoculated samples and the mock-inoculated control (* 0.01 < p ≤ 0.05, ** 0.001 < p ≤ 0.01, *** p ≤ 0.001, adjusted for multiple comparisons with the Benjamin-Hochberg method). The set of all tested contrasts is in the S2 Appendix, Table S7.

https://doi.org/10.1371/journal.pone.0323576.g006