For the analysis of radial systemic signaling, a leaf of a 4 week-old Arabidopsis plant grown on garden soil in a square pot (6 × 6 cm) under short day conditions was infected with 2 μl of an Alternaria spore suspension (1 × 10⁶ spores/ml), as shown in Figure 1A. Preparation of the spore suspension has been described in Michal Johnson et al. (2014). The infected local leaf (no. 8, cf. Farmer et al., 2013) and the non-infected distal leaves (no. 11, according to Dengler, 2006, and no. 10) were harvested at the time points indicated and immediately frozen in liquid nitrogen for RNA extraction.

For the analysis of axial systemic signaling, Arabidopsis seedlings were grown vertically on ½ MS medium in square plates for 2 weeks at long-day conditions (80 μmol m⁻² s⁻¹; Figure 1B). The root area or leaf marked in Figure 1B were infected with 2 μl of an Alternaria spore suspension (1 × 10⁶ spores/ml) and both tissues were harvested separately at the time points indicated in the figures, immediately frozen in liquid nitrogen, and used for RNA extraction.

For interplant communication assays, P. indica was cultured as described previously (Verma and Varma, 1998; Peškan-Berghöfer et al., 2004) in Petri dishes on a modified Kaefer's medium (KM). The plates were kept at room conditions for 2 weeks. A plaque of 5 mm diameter of KM media with and without (control) P. indica mycelium was transferred to the middle of a new plate, as shown in Figure 3. Four seedlings (2 week old, grown on ½ MS medium) were positioned on each plate as shown in Figure 3, so that the root tips were in contact with the fungal (or control) plaque. The plates were kept for 2 days in continuous light of 65 μmol m⁻² s⁻¹, before the onset of experimentation.

To inhibit the information flow from the leaves of the three seedlings infected with A. brassicae spores to the leaves of the unaffected neighboring seedling (cf. Figure 3, seedlings 5 and 6), three experiments were performed: (a) Roots of the infected and non-infected seedlings were separated by a cellophane membrane which prevents physical contact. (b) Mycelium and agar between the infected and uninfected seedlings were cut with a razor blade every 2nd day, starting at day 0. (c) 10 ppm benomyl which kills the mycelium (Paul et al., 2001) was applied to the hyphae connecting the seedlings, at day 0, 2, 5, and 7. For some experiments shown in Figure 4, P. indica was replaced by the fungi Absidia glauca or Mucor mucedo (obtained from Institute of General Microbiology and Microbe Genetics, Jena).

For the data shown in Table 1, 4 seedlings (2 week-old, grown on MS medium with 0.3% (w/v) gelrite) were transferred to PNM plates (Michal Johnson et al., 2014) with 0.3% (w/v) gelrite and P. indica (or a control plate without the fungus). The fungus was pre-grown on the plate for 7 days. Two of the seedlings (in alternate order) were infected with A. brassicae spores, and in the control plates, two were mock-treated with water. After 12 days of co-culture, the intact seedlings with their roots were removed from the plates, and those not treated with A. brassicae were used for RNA and hormone measurements.

All Arabidopsis mutants used in this study have been described: npr1 (Cao et al., 1997), jar1 (Staswick et al., 1992), myb72 (van der Ent et al., 2008), and rbohD knockout (Torres et al., 2002; gift from Prof. Jonathan DG Jones, Warwick, UK). The ABA mutants abi5-1 and aba2-1 were a gift from Prof. A. Gierl (Weihnstephan, Germany).

RNA was isolated from shoots and roots with an RNA isolation kit (RNeasy, Qiagen, Hilden, Germany). Reverse transcription of 1 μg of total RNA was performed with oligo dT Primer and the Omniscript RT Kit (Qiagen, Hilden, Germany). Real-time quantitative reverse transcription-PCR (RT-PCR) was conducted with the following primer pairs:

NRT2.5 (At1G12940, CAGCTGATCATGCCCATCGTGTTC, GCGATGCATAAATCTGGAGAGAGGG), RRTF1 (At4g34410, ACAGTGATAAGCGCGGGAAT, TCCACAAAGGGGAAGTTGAG), JAM1 (At2g46510, CTCCTCGGCCACGATGTCTCTCCGC, CATAATCCGCCAAAATCTCTTCCATTCCTTC), RD29A (At5G52310, GGTTGAAGAAGATGATGATG, GGAAGACACGACAGGAAA), RAB18 (At5g66400, ATTCCCTTCTTCCTCCTC, TGAAGGCTTTGGAACTGG), housekeeping gene GAPDH (At3g04120, GAGCTGACTACGTTGTTGAG, GGA GACAATGTCAAGGTCGG).

The primer pairs used for PDF1.2, VSP2, and JAR1 were described in Scholz et al. (2014). Quantification of A. brassicae in infected and non-infected plant tissue (Figure 3C) was performed with the AbreATr1 gene marker (Guillemette et al., 2004, ACCCGCATTCCTCGCCAAA, AAGTCAAGGATTGTGTCGAGCTT) as described in Michal Johnson et al. (2014).

RT-PCR was performed using the Bio-Rad CFX connect real-time system and Bio-Rad CFX manager version 3.1 (Bio-Rad, Munich, Germany). For the amplification of the PCR products, Eva green (Bio-Rad) and Dream Taq DNA polymerase were used in a final volume of 20 μl. The CFX real-time PCR was programmed to 95°C 2 min, 39 × (95°C 30 s, 60°C 40 s, 72°C 45 s), 72°C 8 min followed by a melting curve program (55–95°C in increasing steps of 0.5°C). Annealing temperature was calculated for each primer pair. All reactions were repeated three times. The mRNA levels for each cDNA probe were normalized with respect to the GAPDH mRNA levels.

Leaf and root material was frozen in liquid nitrogen and kept at −80°C until use. Thirty to One hundred and thirty milligrams of leaf or root material was ground with mortar and pestle, and extracted with 1.2 ml of methanol containing 24 ng of 9,10-D₂-9,10-dihydrojasmonic acid, 24 ng D₄-salicylic acid (Sigma-Aldrich), 24 ng D₆-abscisic acid (Santa Cruz Biotechnology, Santa Cruz, U.S.A.), and 4.8 ng of JA-¹³C₆-Ile conjugate as internal standards. JA-¹³C₆-Ile conjugate was synthesized as described by Kramell et al. (1988) using ¹³C₆-Ile (Sigma-Aldrich). The homogenate was mixed for 30 min and centrifuged at 14,000 rpm for 20 min at 4°C. The supernatant was collected. The homogenate was re-extracted with 500 μl methanol, mixed well, centrifuged and supernatants were pooled. The combined extracts were evaporated in a speed-vac at 30°C and re-dissolved in 250 μl methanol. Chromatography was performed on an Agilent 1200 HPLC system (Agilent Technologies). Separation was achieved on a Zorbax Eclipse XDB-C18 column (50 × 4.6 mm, 1.8 μm, Agilent). Formic acid (0.05%) in water and acetonitrile were employed as mobile phases A and B, respectively. The elution profile was: 0–0.5 min, 5% B; 0.5–9.5 min, 5–42% B; 9.5–9.51 min 42–100% B; 9.51–12 min 100% B, and 12.1–15 min 5% B. The mobile phase flow rate was 1.1 ml/min. The column temperature was maintained at 25°C. An API5000 tandem mass spectrometer (Applied Biosystems) equipped with a Turbospray ion source was operated in negative ionization mode. The instrument parameters were optimized by infusion experiments with pure standards, where available. The ionspray voltage was maintained at −4,500 eV. The turbo gas temperature was set at 700°C. Nebulizing gas was set at 60 psi, curtain gas at 25 psi, heating gas at 60 psi, and collision gas at 7 psi. Multiple reaction monitoring (MRM) was used to monitor analyte parent ion → product ion: m/z 136.9 → 93.0 [collision energy (CE) −22 V; declustering potential (DP) −35 V] for SA; m/z 140.9 → 97.0 (CE −22 V; DP −35 V) for D4-SA; m/z 209.1 → 59.0 (CE −24 V; DP −35 V) for JA; m/z 213.1 → 56.0 (CE −24 V; DP −35 V) for 9,10-D2-9,10-dihydrojasmonic acid; m/z 263.0 → 153.2 (CE −22 V; DP −35 V) for ABA; m/z 269.0 → 159.2 (CE −22 V; DP −35 V) for D6-ABA; m/z 322.2 → 130.1 (CE −30 V; DP −50 V) for JA-Ile conjugate; m/z 328.2 → 136.1 (CE −30 V; DP −50 V) for JA-¹³C₆-Ile conjugate. Both Q1 and Q3 quadrupoles were maintained at unit resolution. Analyst 1.5 software (Applied Biosystems) was used for data acquisition and processing. Linearity in ionization efficiencies were verified by analyzing dilution series of standard mixtures. Phytohormones were quantified relative to the signal of their corresponding internal standard. For quantification of 12-oxophytodienoic acid, cis-OPDA, 9,10-D₂-9,10-dihydro-JA was used as the internal standard applying an experimentally determined response factor of 1.

Twelve days after infection of A. thaliana with GFP-labeled P. indica (gift from Prof. P. Schäfer, Warwick, UK) root colonization was imaged using a LSM 880 (Zeiss Microscopy GmbH, Jena, Germany) with the 488 nm laser line of an argon multiline laser (Figure 2B). Images were taken with a 40x objective (Plan-Apochromat 40x/0.8). A maximum intensity projection was performed from a z-stack of 26 plains each 0.5 μm. Digital images were processed by ZEN software.

Quantitative ROS measurements from leaves were performed with the Amplex Red hydrogen peroxide/peroxidase assay kit (Molecular Probes, Invitrogen, Carlsbad, CA, USA) according to the manufacturer's instructions (https://tools.thermofisher.com/content/sfs/manuals/mp22188.pdf). Leaf sections of 0.5–1 mm width were incubated in the reaction mixture for 10 min in dark at room temperature. The fluorescence intensity was quantified with a fluorescence microplate reader (TECAN Infinite 200 plate reader; Crailsheim, Germany) with excitation at 540 nm and emission at 610 nm. H₂O₂ was used to prepare the standard curve. The reaction mixture without the molecular probe or without the plant material served as control.

Since interplant signal transfer requires systemic information flow within a plant, we looked for genes which responded systemically to various threats in preliminary experiments and decided for NRT2.5 and RRTF1 as read-out to assay information transfer for radially and axially traveling signals. The genes were chosen for study because they responded to various disease-inducing fungal infections in preliminary studies, are not directly related to phytohormone responses, represent responses to systemic signals which are not directly related to each other, and are involved in the primary metabolism and defense strategies (cf. section Introduction).

Arabidopsis leaf no. 8 was infected with 2 μl of an Alternaria spore suspension (1 × 10⁶ spores/ml) or mock-treated with water. The NRT2.5 and RRTF1 mRNA levels were measured in the infected leaf and in distal non-infected leaf no. 11, which is vascularly connected to the infected leaf (Figure 1A, Dengler, 2006). A strong increase in the mRNA levels relative to the mock-treated controls was detectable in the infected leaf 1 day after spore application and the mRNA levels remained high until the 7th day. In distal leaf no. 11, a significant increase was detected between 1 (RRTF1) and 3 (NRT2.5) days after infection and the RRTF1 mRNA level responded much earlier than the NRT2.5 mRNA level (Figure 1A). Since only a low elevation in the NRT2.5 and RRTF1 mRNA levels was observed in the non-connected distal leaf no. 10, and the response started much later (Dengler, 2006; Kiep et al., 2015), a volatile compound as signal transducer is unlikely (Figure 1A). Likewise, when the roots of seedlings were infected with A. brassicae spores (Figure 1B), the expression of the two genes increased within the first 2 days in the roots. An increased expression in the leaves was detected between the 2nd (RRTF1) and 5th (NRT2.5) day (Figure 1B). Again, the mRNA level for RRTF1 responded earlier than that for NRT2.5. The information also traveled from the leaf toward the root with a comparable induction observed in roots when the leaves were infected with A. brassicae spores (Figure 1B). This indicates that NRT2.5 and RRTF1 are systemically induced in non-infected leaves or roots by radially and axially migrating signals, and the axial information flow is bidirectional.

A. brassicae infection induces JA, but not salicylic acid (SA) accumulation in the infected host tissue (Michal Johnson et al., 2014). To test whether RRTF1 and NRT2.5 regulation is linked to these hormones, we measured the induction of the RNA levels in the hormone mutants jar1 and npr1. JAR1 catalyzes the formation of the biologically active jasmonyl-isoleucine (JA-Ile) conjugate, and NPR1 is a receptor for SA (Wu et al., 2012). The mRNA levels were determined in infected local leaves (as shown in Figures 1A,B), the distal leaf no. 11 (as shown in Figure 1A) and roots of seedlings with leaf infection (shown in Figure 1B). RRTF1 expression was comparable to the WT in jar1 and npr1 plants. NRT2.5 induction in local (leaf no. 8) and distal (leaf no. 11) tissue was inhibited in the jar1 mutant, but not in the npr1 mutant in which part of the SA response was inhibited (cf. Herrera-Vásquez et al., 2015; Figure 2). Furthermore, RRTF1 expression has been shown to be induced by ROS (Khandelwal et al., 2008; Matsuo et al., 2015) and A. brassicae infection reported to stimulate ROS accumulation in local and systemic tissues, mainly via the H₂O₂-producing RBOHD, a plasmamembrane-localized NADH oxidase preferentially activated in response to pathogen attacks (Michal Johnson et al., 2014; Matsuo et al., 2015 and ref. therein;). The stimulatory effect on RRTF1 expression in both local and systemic tissues was reduced in the rbohD mutant (Figure 2), while NRT2.5 expression was not affected. This suggests the involvement of ROS produced by RBOHD in the local and systemic RRTF1 response. Finally, MYB72, a transcription factor up-regulated by infection of non-pathogenic rhizobacteria that plays a crucial role in induced systemic resistance and root-to-shoot signaling (Segarra et al., 2012), was found to be not involved in the axial systemic information flow from the roots to the leaf leading to NRT2.5 and RRTF1 induction, since this information flow was not impaired in the myb72 mutant (data not shown, since they are not significantly different from the WT responses). Taken together, local and systemic induction of NRT2.5, but not RRTF1, is partially dependent on jasmonate signaling, whereas local and systemic RRTF1 regulation is partially dependent on RBOHD.

Figure 3A shows the experimental design used for interplant communication assays. Four Arabidopsis seedlings were grown in a Petri dish. In plate 1, the four seedlings were not exposed to any fungus (sample 1), in plate 2, the leaves of three seedlings were infected with an A. brassicae spore suspension (sample 2), while the fourth seedling remained uninfected (sample 3). In the 3rd and 4th plate, the Arabidopsis seedlings were grown in the presence of a P. indica hyphal network. While seedlings in the 3rd plate received no additional treatment (sample 4), the leaves of three seedlings in the 4th plate were infected with A. brassicae spores (sample 5). The 4th seedling was not infected by the pathogen, but connected to the infected seedlings via a P. indica hyphal network (sample 6). P. indica forms a hyphal network that connects the roots of the 4 plants on the plate (Figure 3B). The majority of the hyphae associated with the seedlings can be detected around the primary, secondary and lateral roots. The percentage of hyphae that are visible inside root cells was low. Until the end of the experiment (12 days after infection), we could not detect dead plant cells caused by P. indica infection (cf. Figure 3B). A. brassicae spores germinated on the infected seedling but no pathogen DNA was detected by real-time PCR in the non-infected seedlings growing next to the infected seedlings (Figure 3C). Time course experiments demonstrated that significant amounts of pathogen DNA were first detected 48 h after spore application in the infected seedlings. As expected, between the third and 12th day, the amount of pathogen DNA in the infected seedlings increased dramatically (Figure 3C).

Figure 4A demonstrates that NRT2.5 and RRTF1 expression is also induced by signals from neighboring plants. The mRNA abundance of these two genes increased in the leaves of non-infected seedlings when they were connected by a P. indica mycelial network to the infected seedlings (Figure 4A). No induction was detected when the P. indica hyphal connection was interrupted, either by the insertion of a cellophane membrane which cannot be penetrated by hyphae (Vahabi et al., 2015a,b), by cutting the hyphae with a razor blade (cut) or application of 10 ppm of benomyl to the P. inidca hyphae, which kills the mycelium (Paul et al., 2001; Figure 4B). Interestingly, P. indica could not be replaced by Absidia glauca or Mucor mucedo (Figure 4B), two fungi which associate with but do not colonize Arabidopsis roots. This suggests that a physical contact via the P. indica hyphal network is required for NRT2.5 and RRTF1 induction in neighboring non-stress exposed plants.

Furthermore, induction of NRT2.5 and RRTF1 mRNA was measured for the jar1 and rbohD mutants. Using the same experimental design as shown in Figure 3, wild-type seedlings were replaced by either jar1 or rbohD seedlings, in all possible combinations. Consistent with the results obtained for systemic signal propagation within the infected plant, we observed that if one of the partners was jar1 or rbohD, there was no significant response of NRT2.5 (jar1) or RRTF1 (rbohD) in the leaves of the receiving plants connected via the P. indica hyphal connection to the infected plants (data not shown). It appears that the information cannot travel if one or both of the partners is a mutant, or occurs at a rate too low to measure.

To confirm that volatiles do not participate in the stimulation of NRT2.5 and RRTF1 in uninfected seedlings, a split Petri dish experiment was performed. Wild-type Arabidopsis seedlings were grown on one half of the Petri dish, while in the other half, we grew either an Arabidopsis plant alone, one colonized by P. indica, one exposed to A. brassicae, or one with both P. indica and A. brassicae. Furthermore, Arabidopsis seedlings were exposed to P. indica or A. brassicae hyphae or a combination of both fungi in the neighboring chamber. In none of these experiments, did we observe elevated NRT2.5 and RRTF1 mRNA levels in Arabidopsis seedlings compared to the control (Figure 5). Stimulation of both mRNA levels were only measured in the experimental set-up with a P. indica hyphal connection between the two seedlings (marked in black in Figure 5). This supports the idea that a physical contact is required for the information transfer to the neighboring plant for the induction of NRT2.5 and RRTF1.

To further elucidate how information about infection is transferred to non-infected neighboring plants, we checked defense-related phytohormone levels. A. brassicae is known to stimulate JA but not SA accumulation, whereas P. indica stimulates SA, but not JA accumulation (Michal Johnson et al., 2014). Consistent with these observations, plants which were only infected by A. brassicae (seedling 2 in Figure 3A) had higher JA levels compared to the untreated control (seedling 1 in Figure 3A), while the SA level was slightly reduced (Figure 6A). In seedlings which were only exposed to P. indica (seedling 4 in Figure 3A), the SA level was high, while the JA, JA-Ile and cis-OPDA levels were comparable to those in seedlings not exposed to any fungus (Figure 6A). Interestingly, A. brassicae infection also stimulated the JA, JA-Ile, and cis-OPDA levels in non-infected seedlings growing next to infected seedlings, if they were not connected via a P. indica hyphal network (seedling 3 in Figures 3A, 6A). This is particularly striking for cis-OPDA, the precursor for JA, and JA-Ile, the active form of JA (Figure 6A). Apparently, the information is either transferred through the gas phase or by chemical mediators diffusing through the agar to the non-infected neighboring plants. The JA, JA-Ile, and cis-OPDA levels were also upregulated in Arabidopsis plants that were growing separately from A. brassicae-infected plants in split Petri dishes (Figure 6B). The phytohormone levels in seedlings grown in split Petri dishes and those grown in the “normal” Petri dishes were comparable and not significantly different from each other (Figures 6A,B). This supports the idea of information transfer through the air. A. brassicae-induced JA accumulation was completely prevented and JA-Ile and cis-OPDA strongly reduced in infected (seedling 5 in Figure 3A) and neighboring (seedling 6 in Figure 3A) seedlings that were exposed to P. indica. This is most likely caused by the stimulating effect of P. indica on SA accumulation: its level is high in all seedlings which were growing on plates with P. indica but strongly inhibited when A. brassicae was present in the plate, either alone or in combination with P. indica (Figure 6). Thus, P. indica represses A. brassicae-induced JA, JA-Ile and cis-OPDA accumulation and A. brassicae represses P. indica-induced SA accumulation in both infected and non-infected plants. This is not surprising considering the crosstalk between the two hormones (cf. Caarls et al., 2015). Interestingly, in seedlings not infected by A. brassicae, but connected to infected seedlings via a P. indica hyphal network (seedling 6 in Figure 3A), the ABA level was > 3-fold higher. The ABA level was not stimulated by P. indica or A. brassicae alone, or the combination of both fungi, or in non-infected seedlings without P. indica fungal connection to infected seedlings (Figure 6). Moreover, the ABA level was not elevated in non-infected seedlings grown in the presence of P. indica, when A. brassicae infected seedlings were growing next to them in split Petri dishes, although they had elevated JA, JA-Ile, and cis-OPDA levels (Figure 6B). Thus, stimulation of ABA accumulation in non-infected seedlings growing next to A. brassicae-infected seedlings requires the P. indica hyphal bridge, and is apparently not caused by a JA-ABA crosstalk during the interactions (discussed in various contexts in Robert-Seilaniantz et al., 2011; Kazan and Manners, 2012; Yang et al., 2013; de Ollas and Dodd, 2016; Di et al., 2016; Verma et al., 2016).

The different phytohormone levels are reflected in the expression pattern of phytohormone-responsive genes. Whenever JA, JA-Ile, or cis-OPDA was high, we observed elevated mRNA levels for the marker genes PDF1.2, VSP2, and JAR1, and elevated SA levels stimulated PR-1 mRNA accumulation. Finally, the RD29A, RAB18, and JAM1 mRNA levels were only up-regulated in seedlings with elevated ABA levels (Figure 7).

Arabidopsis seedlings were grown on plates with low gelrite concentration either in the presence or absence of P. indica (Figure S1). Half of the seedlings were infected with the A. brassicae spore suspension (or water as control, marked with a red circle in Figure S1), while the other half remained untreated. After 12 days, the treated (Table 1A) untreated (Tables 1B–D) seedlings were removed for the plates, the roots were separated from the seedling and used for hormone or RNA analyses. From the comparison of the hormone levels shown in Tables 1A,B and those for the different treatments shown in Figure 1B, it is obvious that JA, JA-Ile, and cis-OPDA levels were stimulated by signals traveling through the P. indica hyphal network from A. brassicae-infected seedlings to the roots of the non-infected seedlings. The SA level was higher in roots exposed to P. indica, and partially repressed when A. brassicae-infected seedlings were connected to infected seedlings via the P. indica hyphae. Again, ABA was only upregulated in the roots of non-infected seedlings that were connected to infected seedlings via the P. indica hyphal network. This again excludes an information transfer through the gas phase and confirms the requirement of the P. indica hyphal network for ABA induction (cf. section Discussion).

Next, we checked the NRT2.5 and RRTF1 mRNA levels in the roots under the four different conditions and compared them to those for the JA-responsive genes PDF1.2, VSP2, and JAR1, the SA-inducible gene PR-1 and the ABA-responsive genes RD29A, RAB18, and JAM1 (Table 1B). Consistent with the results from Figure 5, the NRT2.5 and RRTF1 mRNA levels were up-regulated in the roots of non-infected seedlings when they were connected to infected seedlings via a P. indica hyphal network. This confirms that physical contact via the P. indica hyphal network is required for their induction. As shown previously (Kechid et al., 2013), beneficial root-colonizing microbes can induce NRT2.5 expression and a small but not significant stimulation was detectable in the roots of P. indica-exposed seedlings (Table 1B). The JA-responsive genes were significantly stimulated in the roots of non-infected seedlings, when they were connected to infected seedlings via a hyphal network. In contrast to the results for the entire seedlings (Figure 6), an information transfer through the air or by chemical mediators in the medium does not play an important role. P. indica also stimulated the accumulation of SA and the expression of PR1, and this was inhibited when the seedlings were connected to an A. brassicae-infected neighboring plant. The ABA-inducible genes were only up-regulated in the roots of non-infected seedlings connected to the infected seedlings via a P. indica hyphal network. The response was not observed in two different ABA mutants, aba5-1 and aba2-1 (Tables 1C,D). These data indicate that part of the threat information systemically traveling from the A. brassicae infection site in the leaf to the roots is converted to an ABA stress response in non-infected seedlings after transfer via a P. indica hyphal network.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. The reviewer PSB and handling Editor declared their shared affiliation.