A Poly(A) Ribonuclease Controls the Cellotriose-Based Interaction between Piriformospora indica and Its Host Arabidopsis^,

The elicitor-active cell wall moiety of the endophytic fungus Piriformospora indica, cellotriose, modulates the plant-fungus symbiosis by activating a poly(A) ribonuclease in Arabidopsis.

Abstract

Piriformospora indica, an endophytic root-colonizing fungus, efficiently promotes plant growth and induces resistance to abiotic stress and biotic diseases. P. indica fungal cell wall extract induces cytoplasmic calcium elevation in host plant roots. Here, we show that cellotriose (CT) is an elicitor-active cell wall moiety released by P. indica into the medium. CT induces a mild defense-like response, including the production of reactive oxygen species, changes in membrane potential, and the expression of genes involved in growth regulation and root development. CT-based cytoplasmic calcium elevation in Arabidopsis (Arabidopsis thaliana) roots does not require the BAK1 coreceptor or the putative Ca²⁺ channels TPC1, GLR3.3, GLR2.4, and GLR2.5 and operates synergistically with the elicitor chitin. We identified an ethyl methanesulfonate-induced mutant (cytoplasmic calcium elevation mutant) impaired in the response to CT and various other cellooligomers (n = 2–7), but not to chitooligomers (n = 4–8), in roots. The mutant contains a single nucleotide exchange in the gene encoding a poly(A) ribonuclease (AtPARN; At1g55870) that degrades the poly(A) tails of specific mRNAs. The wild-type PARN cDNA, expressed under the control of a 35S promoter, complements the mutant phenotype. Our identification of cellotriose as a novel chemical mediator casts light on the complex P. indica-plant mutualistic relationship.

Ca²⁺ signaling controls many processes in pathogenic and beneficial plant-microbe interactions (Oldroyd, 2013; Steinhorst and Kudla, 2014). In pathogenic interactions, cytoplasmic Ca²⁺ ([Ca²⁺]_cyt) elevation is initiated by cell surface receptors after activation by microbe-associated molecular patterns (MAMPs; like the bacterial flg22 for FLS2 and elf18 for the EF-Tu receptors or the fungal chitin for CERK1). Receptors activated by flg22 or elf18 bind to the coreceptor BRI1-ASSOCIATED RECEPTOR KINASE1 (BAK1), and the phosphorylated pattern recognition receptor (PRR) complexes associate with and phosphorylate BOTRYTIS-INDUCED KINASE1 and other signaling kinases to induce [Ca²⁺]_cyt elevation and downstream immune responses (Lu et al., 2010; Zhang et al., 2010; Shi et al., 2013; Li et al., 2014; Kadota et al., 2015). Besides local responses at the infection site, Ca²⁺, electric, and reactive oxygen species (ROS) waves transfer threat information to distal tissues (Kiep et al., 2015; Gilroy et al., 2016). In interactions with biotrophic microbes, Ca²⁺ activates distinct CALCIUM-DEPENDENT PROTEIN KINASEs, which control host entry and facilitate the accommodation of pathogens in host cells by suppressing plant defense responses (Freymark et al., 2007; Chen et al., 2015). The different functions of Ca²⁺ in these pathogenic interactions are elicited by different Ca²⁺ signatures and decoded by a variety of Ca²⁺-binding proteins (Johnson et al., 2011a).

Colonization of legume roots by arbuscular mycorrhizal fungi is initiated by low-frequency Ca²⁺ spiking in epidermal cells induced by chitin tetramers and pentamers from exudates of germinating fungal spores (Kosuta et al., 2008; Chabaud et al., 2011; Genre et al., 2013; Gutjahr and Parniske, 2013). As soon as the fungus enters the cells, a shift from low- to high-frequency Ca²⁺ spiking is observed (Sieberer et al., 2012), and the two types of Ca²⁺ spiking responses do not propagate to neighboring cells (Miwa et al., 2006; Sieberer et al., 2009; Capoen et al., 2011). The Ca²⁺ channels that mediate these responses have not yet been identified; however, activation of the potassium channels DMI1 from Medicago truncatula or CASTOR and POLLUX from Lotus japonicus could lead to the opening of putative voltage-gated Ca²⁺ channels (Peiter et al., 2007; Charpentier et al., 2008; Venkateshwaran et al., 2012). In contrast to pathogenic interactions, cytoplasmic Ca²⁺ elevation induced by beneficial microbes does not result in massive ROS production (Vadassery et al., 2009).

Besides pathogenic and mycorrhizal fungi, plants interact with numerous endophytes (Johnson et al., 2011a). The root-colonizing endophytic fungus Piriformospora indica, which was isolated originally from the rhizosphere of two woody shrubs in the Indian Thar Desert (Verma et al., 1998), colonizes the roots of a broad host range, including the model plant Arabidopsis (Arabidopsis thaliana; Franken, 2012; Lahrmann et al., 2013; Bakshi et al., 2015; Weiß et al., 2016). It does not cause pathogenic symptoms (Sherameti et al., 2005; Yadav et al., 2010; Johri et al., 2015) but promotes root and shoot biomass production (Peškan-Berghöfer et al., 2004; Camehl et al., 2011; Lee et al., 2011; Das et al., 2012), induces early flowering (Das et al., 2012), and enhances the plant’s resistance to various biotic and abiotic stresses (Waller et al., 2005; Baltruschat et al., 2008; Sherameti et al., 2008; Schäfer et al., 2009; Achatz et al., 2010; Knecht et al., 2010; Das et al., 2012; Lahrmann and Zuccaro, 2012; Daneshkhah et al., 2013; Sun et al., 2014). The fungus releases a chemical compound into its environment that induces rapid (within 90 s) [Ca²⁺]_cyt elevation in Arabidopsis and tobacco (Nicotiana tabacum) roots (Vadassery et al., 2009). Here, we demonstrate that the Ca²⁺-inducing compound from P. indica is cellotriose (CT) and that the [Ca²⁺]_cyt response in Arabidopsis roots requires the poly(A) ribonuclease PARN.

RESULTS

CT Induces [Ca²⁺]_cyt Elevation

Cell wall extract (CWE) from P. indica induces [Ca²⁺]_cyt elevation in Arabidopsis and tobacco roots (Vadassery et al., 2009). To identify the compound triggering the elevation, we combined various chromatographic steps (see “Materials and Methods”; Supplemental Fig. S1). The structure of the purified compound was elucidated by high-resolution mass spectrometry as a trisaccharide (C₁₈H₃₂O₁₆; Fig. 1). NMR experiments (¹H-NMR [Supplemental Fig. S2], selective total correlation [selTOCSY; Supplemental Fig. S3], and ¹H,¹H-double quantum filtered correlation [¹H,¹H-DQF-COSY; Supplemental Fig. S4]) at 700 MHz using a highly sensitive 1.7-mm microcryoprobe indicated the trisaccharide to consist of three hexopyranose units. Based on their characteristic coupling constants, all of these hexopyranose units were identified as Glcs. The low-field chemical shifts of the protons in the 4-positions suggested 1→4 linkage of the Glc units. According to these data, the structure of the trisaccharide was identified as CT. Comparison with the heteronuclear single quantum coherence-NMR spectrum of authentic CT (Fig. 1B) confirmed the suggested structure.

Figure 1.

Purification of the fungal compound inducing [Ca²⁺]_cyt elevation from the P. indica CWE. The active compound was purified from the CWE based on its property to induce [Ca²⁺]_cyt elevation. Ultra-performance liquid chromatography/mass spectrometry, tandem mass spectrometry, and 2D-NMR data were used for structure elucidation. A, Mass spectrometry spectrum corresponding to the peak at 10.46 min. The inset graph is a fixed precursor (505.2 ± 1 D). The collision-induced dissociation scan spectrum was obtained at 7 eV normalized fragmentation energy with the annotated molecular compositions and indicated neutral losses. B, NMR spectra of the CT-containing sample from P. indica (I and II) and CT standard sample (III and IV). I and III, Region of anomeric centers of Glc-A to Glc-C. II and IV, Region of C/H-2 to C/H-6 of Glc-A to Glc-C. Methylene signals (CH2) are shown in red, and methine signals (CH) are shown in blue. *, Signals of contamination. Vertical axes, ΔC (175 MHz); horizontal axes, ΔH (700 MHz).

Chemically pure CT (Sigma-Aldrich) induces dose-dependent [Ca²⁺]_cyt elevations in Arabidopsis roots in the nanomolar range. After a lag phase of 15 to 20 s, the [Ca²⁺]_cyt level begins to rise, and a sharp peak of 507 ± 6 nm at 40 ± 5 s was observed (Fig. 2, A and B). The peak response is followed by a slow and gradual decrease of the [Ca²⁺]_cyt level. Besides roots, leaf [Ca²⁺]_cyt responded to CT too, although ∼5 times weaker. To test whether the CT-inducing [Ca²⁺]_cyt activity shows refractory behavior, we initially activated the roots with the CWE and, 10 min later, applied a CT stimulus. When the [Ca²⁺]_cyt level induced by the first stimulus declined, the second stimulus was reduced strongly compared with the initial one (Fig. 2C). This indicates that both stimuli activate the same perception/signaling system.

Figure 2.

[Ca²⁺]_cyt elevation in the roots of 18-d-old wild-type (WT) or mutant Arabidopsis seedlings expressing cytosolic aequorin. Arrows indicate the application of CT, cellooligomers or chitooligomers, CWEs from fungi or A. tumefaciens, or an exudate fraction from P. indica. A, [Ca²⁺]_cyt elevation after the application of 50 μL of CT (10 µm), 50 µL of CWE, or a CT solution after acid hydrolysis (CT_hyd), as described in “Materials and Methods.” B, Dose-dependent increase in [Ca²⁺]_cyt elevation after CT application. C, Roots were first exposed to 50 μL of CWE and, after 10 min, to 50 μL of CT. D, [Ca²⁺]_cyt elevation induced by different cellooligomers (1 µm). Error bars (n = 40) show significant differences between the triose form and the other oligomers. E, [Ca²⁺]_cyt elevation induced by a concentrated culture filtrate from P. indica. A 2-week-old culture filtrate of P. indica was clarified by high-speed centrifugation (30 min, 250,000g) and concentrated in a lyophilizer before [Ca²⁺]_cyt elevation was determined in the roots of wild-type and cycam seedlings. CF, Culture filtrate without the fungus. F, CT induces [Ca²⁺]_cyt elevation in the roots of wild-type, but not cycam, seedlings. G, A CWE from M. hyalina, but not from A. brassicae, induces [Ca²⁺]_cyt elevation in the roots of cycam seedlings. CT was used as a control. H, Chitin (chitohexaose), but not the concentrated CWE from P. indica (CWE_con), induces [Ca²⁺]_cyt elevation in roots of cycam mutants. H₂O, Water control. I, Chitoheptaose, but not cellooligomers, induces [Ca²⁺]_cyt elevation in cycam roots. J, A CWE from A. tumefaciens induces [Ca²⁺]_cyt elevation in wild-type and cycam roots. K and L, CT induces [Ca²⁺]_cyt elevation in the bak1 and tpc1 (K), glr3.3, glr2.4, and glr2.5 (L) knockout lines. Roots of cycam seedlings were used as a control.

To test the specificity of the sugar moiety, the [Ca²⁺]_cyt response to Glc and different cellooligomers (n = 2–7) was compared with CT (Fig. 2D). Glc (data not shown) and CWE or CT after acid hydrolysis (Fig. 2A) did not induce [Ca²⁺]_cyt elevation. This indicates that CT is an active compound of P. indica CWE that induces [Ca²⁺]_cyt elevation in Arabidopsis roots. We also identified [Ca²⁺]_cyt-inducing activity in concentrated culture medium after the removal of P. indica hyphae (Fig. 2E), which showed the same refractory behavior to a second application of CT but not to chitin (Fig. 2H). Since no response was obtained with the concentrated medium alone, CT is likely to be released by P. indica into the medium. Finally, various combinations of CT with pathogen-associated molecular patterns and cellooligomers either as first or second stimulus confirmed that the tested pathogen-associated patterns operate independently of the cellooligomers for full induction of the Ca²⁺ response, while cellooligomers activate the same perception and/or early signaling system leading to [Ca²⁺]_cyt elevation (Supplemental Table S1).

Identification of Mutants Defective in CT-Dependent [Ca²⁺]_cyt Elevation

To identify Arabidopsis mutants impaired in the Ca²⁺ response to CWE from P. indica, an ethyl methanesulfonate population generated from the pMAQ2 aequorin line was screened. The screen was performed with roots from individual 18-d-old M2 seedlings, and generation of [Ca²⁺]_cyt luminescence signals induced by the CWE was monitored with a plate-reader luminometer. Screening of ∼75,000 individual M2 plants identified three mutants that did not induce any [Ca²⁺]_cyt elevation in response to the CWE (null mutants). Five mutants showed a reduced [Ca²⁺]_cyt response, along with variations in lag phase, peak, and decline to the resting [Ca²⁺]_cyt level. To confirm their impaired Ca²⁺ phenotype, all eight mutants were transferred to soil to obtain M3 and M4 seeds by self-fertilization. Genetic analyses of crosses documented that the three null mutants are allelic, and we named the mutant cytoplasmic calcium elevation mutant (cycam). The mutant also failed to induce [Ca²⁺]_cyt elevation in response to pure CT (Fig. 2F). After backcrossing to the wild type (Columbia-0 [Col-0] or Landsberg [La]), [Ca²⁺]_cyt elevation to CT was restored in ∼25% of F2 progeny (P < 0.05), indicating that the mutation is recessive.

To test whether the loss of [Ca²⁺]_cyt elevation in the cycam mutant is specific for the P. indica CWE, we tested CWEs from other beneficial (e.g. Mortierella hyalina and Periconia macrospinosa) and pathogenic (e.g. Alternaria brassicae and Verticillium dahliae) fungi. Three extracts induce [Ca²⁺]_cyt elevation in the wild-type aequorin line and the cycam mutant. The CWE from A. brassicae induced [Ca²⁺]_cyt elevation in the wild type but not in the cycam mutant (Fig. 2G). Crosses between the cycam mutant and a previously described mutant not responding to the CWE from A. brassicae (Johnson et al., 2013) and analyses of the Ca²⁺ responses of the F1 seedlings demonstrate that they are allelic (see below). This demonstrates that the cycam mutation prevents [Ca²⁺]_cyt elevation in response to extracts from a beneficial and a pathogenic fungus. In an independent screen, we isolated an Arabidopsis mutant with an altered [Ca²⁺]_cyt response to a CWE from M. hyalina (E. Seebald and R. Oelmüller, unpublished data). The [Ca²⁺]_cyt response of this mutant to CT is identical to that of the wild type. Furthermore, a Leu-rich-repeat protein (At1g13230) is required for the beneficial interaction of the two symbionts (Shahollari et al., 2007). However, a knockout line for this protein was not impaired in the Ca²⁺ response (Supplemental Table 1B), suggesting that the protein is not part of the early Ca²⁺ signaling system or located downstream of it. To further characterize the cycam mutant, we tested its response to flg22 and chitin. The Ca²⁺ signals triggered by both well-known elicitors in the cycam mutant were comparable with the wild type (Fig. 2H, shown for chitin). The cycam mutant also is defective in inducing [Ca²⁺]_cyt elevation in response to other cellooligomers (n = 2–7; Fig. 2I), whereas the responses to chitooligomers (n = 5–8) were not affected (Fig. 2, H and I, shown for chitoheptaose). The differences in the CT/cellooligomer and chitooligomer responses demonstrate that cycam is defective in cellooligomer recognition or signaling.

Sharma et al. (2008), Glaeser et al. (2016), and Guo et al. (2017) demonstrated that P. indica forms an intimate association with Rhizobium radiobacter F4, an α-proteobacterium. The endophytic bacterium shows a high degree of similarity to the plant pathogenic R. radiobacter (formerly, Agrobacterium tumefaciens C58), except for distinct differences in both the tumor-inducing and the accessory plasmids, which can explain the loss of its pathogenicity (Glaeser et al., 2016). A CWE from the disarmed A. tumefaciens GV3101 induced [Ca²⁺]_cyt elevation in both wild-type and cycam roots (Fig. 2J). This demonstrates that both Ca²⁺-inducing activities are different and that CT is likely of fungal and not bacterial origin. The absence of Ca²⁺-inducing activity in the P. indica CWE in cycam suggests that the amount of bacterial effector in this preparation is either too low for detection or not present.

Exploration of the Signaling Pathway Underlying the CT-Induced [Ca²⁺]_cyt Response

[Ca²⁺]_cyt elevation is induced by MAMP-activated receptor complexes in which BAK1 often functions as a coreceptor. CT induces [Ca²⁺]_cyt elevation in aequorin lines in the bak1 knockout background similar to the wild type (Fig. 2K). Furthermore, the roots of mutants for the proposed calcium channel GLU-LIKE RECEPTORS (GLR2.4, GLR2.5, and GLR3.3) and vacuolar TWO PORE CHANNEL1 (TPC1) in the aequorin background also were not impaired in CT-induced [Ca²⁺]_cyt elevation (Fig. 2L). These data suggest that neither BAK1 as coreceptor nor the calcium channels tested are involved in CT signaling.

Early Downstream Responses to CT-Induced [Ca²⁺]_cyt Elevation

MAMP-triggered Ca²⁺ responses are associated with changes in plasma membrane potentials and the generation of ROS or H₂O₂ and interplay between the three responses is proposed to be involved in systemic signaling (Gilroy et al., 2016). When Arabidopsis roots were impaled with voltage-recording microelectrodes, resting potentials of −163 mV for the wild type (se = 3 mV; n = 23) and −166 mV for cycam (se = 4 mV; n = 8) were measured (Fig. 3). Following the application of CT stimulus, wild-type root cells depolarized in a dose- and time-dependent manner in the wild type. Membrane responses were elicited at CT concentrations of above 2.5 µm. Signals were transient in nature, and the delay time between stimulus onset and membrane response time decreased with increasing CT concentration. Local application of 40 µm CT to wild-type roots induced the plasma membrane depolarization of bulging root hair cells with average amplitudes of 27 mV (se = 4; n = 13). Under the same conditions, CT in cycam roots did not trigger the depolarization of the plasma membrane (Fig. 3).

Figure 3.

CT induces CYCAM-dependent root hair plasma membrane depolarization. A, Cartoon of the experimental setup. Microelectrodes, impaled into bulging root hair cells, measured the plasma membrane potential, while CT was applied locally via application pipettes. B, Average depolarization of the Col-0 and cycam root hair plasma membranes. CT was applied at a concentration of 40 µm for 1 s (arrow). Error bars show se; n = 8. C, Average dose-dependent depolarization of the Col-0 root hair plasma membrane. CT was applied at the indicated concentrations for 1 s (arrow). Error bars show se; n = 5.

CT also induces H₂O₂ production in wild-type roots; however, compared with chitin, the response was low (Fig. 4). Since no ROS production is detectable in cycam roots, CT-induced generation of Ca²⁺, ROS, and electrical signals is genetically linked (see “Discussion”). However, if both MAMPs are applied to the roots together (either simultaneously, as shown in Fig. 4, or successively in either sequence within 8 h), the initial H₂O₂ production is higher than that of the sum of the two individual treatments. This suggests a synergistic effect of the two elicitors on H₂O₂ production (see below). Furthermore, comparable to chitin, barely any H₂O₂ production can be induced by CT in rbohD roots (chitin, 3,111 ± 451 relative light units [RLU] in the wild type versus 155 ± 22 RLU in rbohD; CT, 988 ± 108 RLU in the wild type versus 81 ± 9 RLU in rbohD [n = 5]), suggesting that, as with chitin, NADH oxidase D is the major ROS producer in response to CT too.

Figure 4.

CT induces CYCAM-dependent H₂O₂ production in Arabidopsis roots and cooperates with chitin-induced H₂O₂ production. H₂O₂ production is shown in roots of 18-d-old wild-type (WT) and cycam seedlings challenged with 50 μL of CT or chitin or CT + chitin (10 µm; application at time 0). Data are based on eight independent experiments; error bars represent se.

CT Promotes RBOHD and Defense Gene Expression

Since chitin-mediated H₂O₂ production is stimulated by CT application (Fig. 4), the RBOHD mRNA level is induced rapidly by CT and chitin (Fig. 5), and RBOHD is required for CT-induced H₂O₂ production, we investigated the interplay between CT and chitin on RBOHD expression. CT was applied to Arabidopsis roots 8, 2, or 0 h before chitin application, and the accumulation of RBOHD mRNA was monitored within the first 30 min (Fig. 6). Simultaneous application of CT and chitin to the roots (time 0) had no significant effect on RBOHD mRNA accumulation within the first 10 min, whereas after 20 and 30 min, a comparable stimulation was observed for both elicitors. The effects of both elicitors on RBOHD mRNA accumulation was additive, which can be explained either by independent effects or a stimulating/amplifying effect of one elicitor on the action of the other elicitor. However, if CT was applied 2 or 8 h before chitin application, the chitin-induced RBOHD mRNA level was much higher after 10 min compared with the treatment with chitin alone. This suggests that CT cooperates with chitin to induce RBOHD expression and that a CT pretreatment to the roots establishes conditions that allow a faster and stronger accumulation of RBOHD mRNA in response to chitin. Downstream of the membrane-delimited steps addressed above, early elicitor responses can be seen at the level of the expression of genes that serve as defense markers, such as MAPK3, ZAT12, NPR1, and LOX1. Within 30 min, the mRNA levels of RBOHD, MAPK3, and ZAT12 are up-regulated after CT application to wild-type, but not cycam, roots (Fig. 5). Chitin, however, induced the mRNAs in cycam roots too (Fig. 5). In contrast to H₂O₂ production (Fig. 4), which was less induced by CT compared with chitin in the wild type, the stimulatory effect of the two elicitors on these mRNA levels is comparable (Fig. 5). The mRNA levels for the hormone marker genes NPR1 and LOX1 respond less and later to CT application, and significant differences can only be seen 2 to 4 h after elicitor application (Fig. 5). These results demonstrate the involvement of the cycam gene product in CT-induced defense gene activation.

Figure 5.

CT- or chitin-induced accumulation of transcripts in wild-type (WT) and cycam roots. The data show fold induction of the mRNA levels relative to the mock-treated controls with water. Samples were analyzed by quantitative real-time PCR (RT-qPCR). Data are based on four independent experiments; error bars represent se.

Figure 6.

Accumulation of RBOHD mRNA in response to CT, chitin, and both elicitors. 0 h, CT, chitin, or CT + chitin was applied to Arabidopsis roots, and RBOHD mRNA accumulation was determined by RT-qPCR 10, 20, and 30 min after elicitor application. For the reference value, water was applied to the roots. 2 h and 8 h, CT was applied 2 and 8 h before chitin. The data show fold induction of the RBOHD mRNA levels relative to the mock-treated controls with water. Data are based on four independent experiments; error bars represent se.

cycam Is Defective in a PARN

CYCAM was mapped on chromosome 1 using the ARMS primer set (Schäffner, 1998) with 25 F2 mutant individuals of crosses between cycam (cytosolic apoaequorin in the Col-0 background) and the wild type (in the La background). Backcrossing of cycam to the aequorin line used for mutagenesis (Col-0 background) followed by Illumina sequencing of two DNA pools from F2 individuals with and without the cycam phenotype identified a target interval on the lower arm of chromosome 1 containing two candidate genes (Supplemental Fig. S5), one of which encodes AtPARN. This nuclease is a eukaryotic poly(A)-degrading deadenylase that efficiently degrades poly(A) tails of selective mRNAs and, thus, reduces their translation efficiency (Virtanen et al., 2013).

As expected from a mutant defective in poly(A) tail degradation and compared with the wild type, the overall growth of cycam plants appeared less synchronized and adult plants showed a stunted phenotype, which was caused mainly by reduced stem lengths. The numbers of leaves of each adult cycam and wild-type plant, grown in the greenhouse, were the same, whereas the fresh weights of the mutant leaves were reduced by 18.8% ± 7.2% (n = 73 plants). Full-length PARN cDNAs were isolated from wild-type and cycam plants. Sequence analyses demonstrated that a point mutation (C→T) at position 403 downstream of the A of the ATG start codon in the mutant resulted in a Leu-to-Phe amino acid exchange (L135F).

The wild-type cDNA sequence was expressed in cycam plants under the control of the 35S promoter, and four transformants were analyzed in more detail. The visible mutant phenotype was rescued by the transformation. After CT application to the roots, the [Ca²⁺]_cyt elevation response was recovered (Fig. 7A). Expression of the 35S::PARN construct in wild-type plants did not cause any visible phenotype, and the plants looked like the untransformed wild type. These transformants also showed the same Ca²⁺ responses after CT application as the untransformed wild type, suggesting that the Ca²⁺ response and the performance of the plant are not limited by the amount of PARN mRNA in the wild type. The cycam mutant contains elevated abscisic acid (ABA) levels (Johnson et al., 2013, and refs. therein for comparable parn mutants), and the ABA levels in seedlings of the four rescued mutant lines were reduced by ∼50% (Fig. 7B). Again, no difference was observed for wild-type plants transformed or not transformed with the 35S::PARN construct. Interestingly, 1 mm chitin applied to the roots of wild-type seedlings resulted in stomata closure in the leaves, while the stomata in the cycam leaves remained closed due to the elevated ABA level. The stomata in the rescued mutants are open and closed after application of 1 mm chitin, comparable to the wild type (data not shown). Again, higher expression of PARN under the control of the 35S promoter in wild-type plants did not result in any change in stomata opening in comparison with the untransformed wild type. Thus, the single amino acid exchange in the mutant PARN protein appears to be responsible for the observed phenotype.

Figure 7.

A 35S::PARN construct complements the mutant phenotype. A, Wild-type (WT) and cycam plants were transformed with the 35S::PARN construct, and [Ca²⁺]_cyt elevation after application of 50 μL of CT (10 µm) was measured in roots of wild-type and cycam seedlings as well as of four (1–4) wild-type and cycam seedlings that were independently transformed with the 35S::PARN construct. B and C, The same seed batches were used to determine ABA levels (B) and the percentage of open stoma (C). Data are based on four (Ca²⁺ measurement), 10 (ABA level), and five (stomatal opening) independent experiments; error bars represent se. FW, Fresh weight.

PARN Is Required for P. indica-Mediated Benefits to Arabidopsis Seedlings

Our (Peškan-Berghöfer et al., 2004) and other (Banhara et al., 2015) studies demonstrated that the growth of Arabidopsis seedlings in the presence of P. indica results in an ∼30% increase (depending on the growth conditions) of biomass. The biomass of cycam seedlings is reduced by ∼10% when they are cocultivated with P. indica (Table I).

Table I. P. indica promotes biomass production and drought tolerance in wild-type, but not cycam, seedlings.

The increase in biomass of Arabidopsis wild-type and cycam seedlings was determined 12 d after cocultivation with P. indica. The biomass of uncolonized wild-type seedlings was set as 100%. The chlorophyll fluorescence parameter F_v/F_m, representing the efficiency of the photosynthetic electron flow, was determined for P. indica-exposed and mock-treated wild-type and cycam seedlings 48 h after opening of the lid of the petri dishes (for experimental details, see Fig. 2 in Sherameti et al., 2008). Data are based on 120 seedlings of four independent experiments; errors are se.

Parameter	Wild Type − P. indica	Wild Type + P. indica	cycam − P. indica	cycam + P. indica
Biomass (%)	100 ± 9	132 ± 15	89 ± 13	78 ± 6
Fv/Fm after 48 h of drought	0.49 ± 0.05	0.78 ± 0.06	0.51 ± 0.03	0.46 ± 0.06

We also demonstrated that P. indica confers drought stress tolerance to Arabidopsis seedlings. This can be quantified by the chlorophyll fluorescence parameter F_v/F_m, which describes the efficiency of the photosynthetic electron transfer after exposure of the seedlings to drought stress (Sherameti et al., 2008). Under drought stress, the F_v/F_m value is higher for P. indica-colonized wild-type seedlings compared with the uncolonized control, whereas no difference can be observed for colonized and uncolonized cycam seedlings (Table I). Finally, we have shown previously that cycam seedlings are more susceptible to A. brassicae infection and its toxin(s) than are wild-type seedlings and that a protective role of P. indica against the A. brassicae challenge can be detected in wild-type but not cycam seedlings (Johnson et al., 2013). Taken together, P. indica-mediated benefits and defense against A. brassicae require functional PARN in Arabidopsis seedlings.

Expression Profiling Identified CT-Responsive Genes

In initial studies, we observed that the number of genes responding to CT in Arabidopsis roots increased with longer (greater than 1 h) incubation periods. Therefore, we analyzed expression profiles 4 and 8 h after exposure of wild-type and cycam roots to CT and compared them with those obtained after chitin application. Chitin was used as a control because it is a well-characterized MAMP and also active in roots. Overall, 1,532 genes responded to either CT or chitin application greater than 2.5-fold 4 or 8 h after application (Fig. 8A; Supplemental Tables S2A and S2B). A total of 1,026 genes responded to CT but not to chitin. Interestingly, we did not identify any gene that was regulated in the opposite direction in response to CT or chitin treatments at a given time point. This suggests overlapping functions of both elicitors, whereas the CT response is normally lower and occurs later when compared with the response to chitin. In accordance with the ROS data (Fig. 4), it appears that CT induces a milder defense response in the roots than chitin (Fig. 8B; Table III; Supplemental Tables S2A and S2B).

Figure 8.

Microarray analysis of differentially regulated genes after the application of CT or chitin. A, Venn diagrams of microarray data of wild-type (WT) and cycam roots 4 and 8 h after treatment with CT or chitin (10 µm each). Genes responding to CT were analyzed from the complete data set. Afterward, genes also responding to chitin were analyzed from that pool of CT-responding genes. The Venn diagrams show that the number of genes responding to either CT or chitin is lower in cycam roots (right) than in wild-type roots (left). After 4 h (top), the number of genes induced by CT and chitin is ∼6 times higher in wild-type roots than in cycam roots, whereas after 8 h (bottom), the difference is only ∼2-fold. Venn diagrams show genes that are up-regulated (blue) or down-regulated (pink) by CT and up-regulated (green) or down-regulated (red) by chitin. Only genes are considered that are regulated greater than 2.5-fold relative to the mock (water)-treated control (log₂ ≥ 1.32 or ≤ −1.32 in Supplemental Tables S2A and S2B). Data are based on three independent RNA samples per treatment. B, Pie chart of differentially regulated genes in wild-type plants 4 h after treatment with CT. Shown are the groups of genes of the associated pathways (in percentage) regulated by 10 µm CT. A full list with numbers of genes and a comparison with genes also induced by chitin can be found in Table III. CHO, Carbohydrate.

Table III. Number of genes differentially regulated by CT and chitin 4 h after elicitor application in wild-type plants.

The number of CT-responding genes corresponding to the indicated AGI_LOCUS_TAIR10_Aug2012 pathways is shown. The third column shows the number of CT-responsive genes that also respond to chitin. Genes that respond to CT were analyzed from the complete data set. Afterward, genes also responding to chitin were analyzed from the pool of CT-responding genes.

Pathways (AGI_LOCUS_TAIR10_Aug2012)	No. of Genes Regulated in the Wild Type 4 h after CT Application	Subset of CT-Responsive Genes in the Wild Type Also Regulated 4 h after Chitin Application
Not assigned	472	131
Miscellaneous	123	30
Protein	110	36
RNA	105	34
Stress	94	20
DNA	71	27
Signaling	60	20
Cell wall	53	14
Transport	42	5
Development	37	3
Cell	24	8
Hormone metabolism	24	8
MicroRNA	22	9
Secondary metabolism	20	7
Lipid metabolism	15	1
Redox	15	6
Photosynthesis	6	2
Major carbohydrate metabolism	5	4
Metal handling	4	2
Amino acid metabolism	3	3
Nucleotide metabolism	3	1
Biodegradation of xenobiotics	2	0
Polyamine metabolism	2	1
Mitochondrial electron transport	2	0
Minor carbohydrate metabolism	2	0
Cofactor and vitamin metabolism	1	0
Glycolysis	1	0
TCA	1	1
Fermentation	1	0
Tetrapyrrole synthesis	1	0
N-metabolism	1	1

CT- but not chitin-responsive genes can be divided further into those regulated in wild-type but not cycam roots. These genes respond to CT via a PARN- and/or Ca²⁺-dependent pathway (Fig. 8A; Supplemental Tables S2A and S2B) and were found to encode Ca²⁺-dependent proteins and proteins involved in Ca²⁺-dependent signaling events, proteins involved in exocytosis, auxin and ethylene functions, defense, cell wall biosynthesis, cell and root growth, or metabolite and information transfer within the plant body (Supplemental Table S2B). The regulation of a few genes induced by CT in a PARN-dependent manner (Supplemental Table S2B) was analyzed further by real-time PCR and compared with their regulation by chitin (not PARN-dependent genes) as well as the CWE of A. brassicae (PARN-dependent genes; Table II). Interestingly, the majority of the tested genes responded quite differently to the three stimuli, and many of them responded to CT but not or to a lesser extent to the CWE from A. brassicae or chitin (Table II). Among the genes that respond only to CT are those that encode proteins involved in cell and root growth, regulators of root development and elongation, exocytosis, cell division, transporters for metabolite distribution within the plant, and hormone effects (Supplemental Table S3). Among the genes responding to all three stimuli is SWEET11 (At3g48740), which encodes a transporter involved in Suc phloem loading and carbon export from the leaves to the roots (Durand et al., 2016). Chen et al. (2010) showed that pathogens and symbionts promote the expression of several SWEET genes for nutritional gain, including SWEET11.

Table II. Genes in Arabidopsis wild-type and cycam roots that are differentially regulated by CT (10 µm, 50 µL), a CWE from A. brassicae (50 µL), or chitin (10 µm, 50 µL) 4 or 8 h after elicitor application.

Data show fold induction values relative to values for roots treated with water. Data are based on three independent RT-qPCR experiments; errors are se.

Arabidopsis Genome Initiative No.	Gene	Hours after Elicitor Application	CT		CWE from A. brassicae		Chitin
			Wild Type	cycam	Wild Type	cycam	Wild Type	cycam
At3g09530	EXOCYST H3	4	4.5 ± 0.3	0.1 ± 0.1	0.1 ± 0.1	0.2 ± 0.1	0.1 ± 0.2	0.1 ± 0.1
At5g06510	NUCLEAR FACTOR Y, SUBUNIT A10	4	3.5 ± 0.4	−0.2 ± 0.1	0.2 ± 0.1	0.2 ± 0.1	0.4 ± 0.2	−0.2 ± 0.1
At1g66700	S-Adenosyl-Met methyltransferase	4	3.6 ± 0.8	0.7 ± 0.3	0.0 ± 0.1	−0.4 ± 0.2	1.1 ± 0.3	1.6 ± 0.4
At5g64890	PROPEP2 (ELICITOR PEPTIDE2 precursor)	4	2.2 ± 0.4	0.1 ± 0.0	0.3 ± 0.1	0.0 ± 0.0	0.3 ± 0.2	0.8 ± 0.3
At4g37710	VQ motif-containing protein, VQ29	4	5.1 ± 1.0	0.2 ± 0.1	−0.4 ± 0.1	−0.2 ± 0.1	−0.1 ± 0.1	0.0 ± 0.1
At4g37060	PATATIN-LIKE PROTEIN5	4	2.0 ± 0.6	0.0 ± 0.0	−1.1 ± 0.3	0.0 ± 0.0	−1.0 ± 0.4	0.0 ± 0.0
At5g65100	EIN3	4	1.5 ± 0.3	−0.4 ± 0.2	0.5 ± 0.2	0.6 ± 0.2	−0.1 ± 0.2	−0.1 ± 0.2
At1g79820	SUPPRESSOR OF G PROTEIN	4	1.9 ± 0.5	0.0 ± 0.2	0.1 ± 0.2	0.0 ± 0.1	0.0 ± 0.2	0.0 ± 0.2
At2g19030	RALF11	8	3.3 ± 0.8	0.0 ± 0.2	−0.2 ± 0.2	−0.3 ± 0.2	0.0 ± 0.1	−0.2 ± 0.2
At5g27495	Ion channel inhibitor	8	4.2 ± 0.5	0.1 ± 0.2	0.3 ± 0.3	0.3 ± 0.2	0.2 ± 0.2	0.0 ± 0.2
At5g06170	SUC-PROTON SYMPORTER9	8	4.0 ± 0.6	0.0 ± 0.2	−0.1 ± 0.2	0.0 ± 0.2	0.1 ± 0.2	0.0 ± 0.2
At2g29100	GLR2.9	8	4.4 ± 0.6	0.4 ± 0.2	0.1 ± 0.0	0.3 ± 0.2	0.8 ± 0.4	0.6 ± 0.3
At1g67770	TERMINAL EAR1-LIKE2	8	3.9 ± 0.7	0.9 ± 0.4	0.0 ± 0.2	0.3 ± 0.2	0.0 ± 0.1	0.0 ± 0.1
At3g04280	RESPONSE REGULATOR22	8	1.6 ± 0.1	−1.0 ± 0.4	0.0 ± 0.2	−0.3 ± 0.2	−0.2 ± 0.2	−0.3 ± 0.1
At1g26250	EXTENSIN18	8	1.9 ± 0.5	0.2 ± 0.2	0.2 ± 0.1	0.0 ± 0.1	0.3 ± 0.3	0.0 ± 0.2
At5g42800	DIHYDROFLAVONOL 4-REDUCTASE	8	1.5 ± 0.4	−1.1 ± 0.4	0.4 ± 0.2	0.5 ± 0.4	0.6 ± 0.4	0.4 ± 0.2
At4g37060	PHOSPHOLIPASE AIIIβ	8	1.8 ± 0.3	0.0 ± 0.2	0.0 ± 0.2	0.0 ± 0.2	−0.6 ± 0.2	0.0 ± 0.1
At3g48740	SWEET11	8	−3.3 ± 0.6	−1.2 ± 0.2	−3.1 ± 1.0	−1.1 ± 0.4	−2.8 ± 1.0	−1.5 ± 0.4
At2g42590	GENERAL REGULATORY FACTOR9	8	−2.4 ± 0.7	0.3 ± 0.2	0.2 ± 0.2	0.3 ± 0.2	0.4 ± 0.3	0.2 ± 0.3
At5g24920	GLN DUMPER5	8	−2.2 ± 0.9	0.3 ± 0.2	0.2 ± 0.2	0.4 ± 0.2	−0.1 ± 0.2	0.3 ± 0.2

DISCUSSION

We demonstrate that CT from P. indica and a different low-molecular-mass compound from A. brassicae (Supplemental Fig. S6; Johnson et al., 2013) induce [Ca²⁺]_cyt elevation in Arabidopsis roots and that these responses require PARN. Furthermore, both elicitors activate different Ca²⁺-dependent responses in Arabidopsis roots (Table II), suggesting that the specificity is established downstream of PARN (Fig. 9). Compared with chitin, CT induces a weak defense response (Fig. 4; Table II) and also stimulates the expression of genes, which do not respond to chitin. The isolated mutant with the weak PARN allele confirms that both P. indica-induced benefits and A. brassicae-activated defense responses are impaired (Table II; Johnson et al., 2013). PARN might control the efficient translation of mRNAs for proteins that control [Ca²⁺]_cyt elevation and the proper responses of the plant to at least one beneficial and one pathogenic fungus.

Figure 9.

Model describing the function of PARN in controlling [Ca²⁺]_cyt elevation in response to CT. Arrows indicate an induction of targets, while T-bars represent repression. EMS, Ethyl methanesulfonate.

CT and Early Signaling Events

β-Glucans are ubiquitous in plant and fungal cell walls. The β-1,4-glucan cellulose is one of the most abundant glucans in plants. Fungi can generate short-chain β-1,4-glucans by the degradation of cellulose or other glucan polymers (Karlsson et al., 2002; McCarthy et al., 2003; Liu et al., 2010; Chen et al., 2014). Since CT was isolated from P. indica growing without the host in a culture medium, cellulose from the plant cell wall as source can be excluded, and CT is most likely of fungal origin. The fungus forms a symbiotic interaction with R. radiobacter F4, an α-proteobacterium (Sharma et al., 2008; Glaeser et al., 2016; Guo et al., 2017), and a CWE from A. tumefaciens, a closely related species, also induces [Ca²⁺]_cyt elevation in Arabidopsis roots. This response occurs in the wild type and the cycam mutant (Fig. 2J), suggesting that it functions genetically differently from the CT response. Apparently, a compound synthesized by the bacterium inducing [Ca²⁺]_cyt elevation is either not present in the cell wall preparation from P. indica (and its bacterial endophyte) or its concentration is too low for a detectable Ca²⁺ signal in the cycam mutant. Two scenarios are possible: P. indica might generate CT by a transglycosylation reaction using its own cellobiose as substrate, or the fungus degrades cellopolymers present in the fungus. Since the purification procedure results only in single Ca²⁺-inducing peaks (Supplemental Fig. S1), the presence of other cellooligomers, such as cellobiose, as potential precursors for CT biosynthesis in the CWE is unlikely, since they also should induce [Ca²⁺]_cyt elevation (Fig. 2D). Thus, how CT is generated in the fungus remains unclear. Many fungal glucosidases possess efficient transglycosylation capacities for the synthesis of CT from cellobiose in the intracellular and extracellular spaces (Smaali et al., 2004; Suzuki et al., 2010; Zhao et al., 2015; Guo et al., 2016a; Mallek-Fakhfakh and Belghith, 2016; Boudabbous et al., 2017), and homologous enzymes can be found in the P. indica genome (J. Thürich and R. Oelmüller, unpublished data). The large number of genes encoding cellopolymer-degrading enzymes in the P. indica genome (Zuccaro et al., 2011) also might allow the fungus to generate CT by degrading cellooligomes or cellopolymers present in the fungus. Since [Ca²⁺]_cyt elevation is triggered by CT in nanomolar concentrations, the compound should be generated in low concentrations and not as an intermediate of an important biochemical pathway.

MAMPs are recognized by PRRs (Zipfel, 2008; Macho and Zipfel, 2014). PRRs for chitin and oligogalacturonides have been identified, but receptors of other oligosaccharides including β-glucans are unknown. BAK1, a coreceptor in many PRRs, is not required for CT-induced [Ca²⁺]_cyt elevation (Fig. 2K). How CT is recognized by the root cell is unknown at present; however, Glc shows no Ca²⁺-inducing activity in Arabidopsis roots, and cellobiose as well as longer oligomers reduced activity compared with CT, suggesting specificity in the recognition. In general, the biological activity of oligosaccharides in eliciting plant responses is highly dependent on their degree of polymerization, and the responses are species specific (Trouvelot et al., 2014). The perception of oligogalacturonides, oligomers of α-1,4-linked galacturonosyl residues (Nothnagel et al., 1983) that induce resistance of Arabidopsis against Botrytis cinerea (Aziz et al., 2004; Ferrari et al., 2007), occurs by the WALL-ASSOCIATED KINASE1 (WAK1), a transmembrane receptor kinase (Kohorn and Kohorn, 2012). WAK1 binds oligogalacturonides (Brutus et al., 2010), which leads to the activation of the intramembrane kinase domain of WAK1 and the activation of plant immune responses (Trouvelot et al., 2014). A similar scenario with a so far unidentified receptor and early signaling compounds can be envisioned for CT.

Recently, Souza et al. (2017) demonstrated that oligomers derived from cellulose are perceived as signal molecules in Arabidopsis, thereby triggering a signaling cascade that shares similarities to responses to chitooligomers and oligogalacturonides. Cellobiose stimulates neither detectable ROS production nor callose deposition. Cotreatments of cellobiose with flg22 or chitooligomers led to synergistic increases in defense gene expression. The authors concluded that the perception of cellulose-derived oligomers may participate in cell wall integrity surveillance and represents an additional layer of signaling following plant cell wall breakdown during cell wall remodeling or pathogen attack. Many of their results and conclusions resemble those described here, and both studies demonstrate that plants perceive cellooligomers from different sources that interfere with well-studied defense signaling pathways. In our study using Arabidopsis roots, CT was more effective than cellobiose and other cellooligomers, but they share PARN for the induction of [Ca²⁺]_cyt elevation (Fig. 2I). Apparently, a so far uncharacterized perception system perceives cellopolymers from different sources to inform the plant about environmental changes and threats. Since the generation of Ca²⁺, ROS, and electric signals in response to CT is impaired in the cycam mutant (Figs. 2–4), CT (and possibly cellooligomer) perception also may be used by the plant to inform distant cells or organs via systemic signal propagation (Gilroy et al., 2014, 2016, and refs. therein).

Previously, we demonstrated that the CWE from P. indica does not induce ROS production (Vadassery et al., 2009). Furthermore, cellobiose does not induce ROS production in leaves (Souza et al., 2017). However, CT induces ROS in Arabidopsis roots, although the overall level is low compared with chitin (Fig. 4). Direct comparison of the CT and CWE stimuli confirmed the differences in ROS accumulation, suggesting that the CWE contains an additional compound or compounds inhibiting CT-induced ROS production. Furthermore, Ca²⁺-dependent NADPH oxidase D activation is required for rapid defense signal propagation (Dubiella et al., 2013). The strongly reduced ROS accumulation in the rbohD mutant after CT application supports the important role of this oxidase for the function of the elicitor.

Ultimately, receptor activation must lead to the opening of Ca²⁺ channels, and GLRs are potential candidate channels for CT-induced [Ca²⁺]_cyt elevation. GLR3.3 mediates leaf-to-leaf wound signaling (Li et al., 2013; Manzoor et al., 2013; Mousavi et al., 2013). However, CT-mediated [Ca²⁺]_cyt elevation is not impaired in the glr3.3 mutant (Fig. 2L). In addition, GLR2.4 and GLR2.5 are not required for CT-mediated [Ca²⁺]_cyt elevation (Fig. 2L). Furthermore, wounding-induced systemic Ca²⁺ elevations are dependent on TPC1, a Ca²⁺-activated vacuolar cation channel (Hedrich and Neher, 1987; Peiter et al., 2005; Peiter, 2011; Choi et al., 2014; Xiong et al., 2014; Kiep et al., 2015; Guo et al., 2016b); however, this channel also is not involved in CT-induced [Ca²⁺]_cyt elevation (Fig. 2K). Apparently, the initial steps between the perception and regulation of Ca²⁺ channel activities by CT differ from other known MAMP systems.

Response to CT

Microarray analyses identified CT-responsive genes in Arabidopsis roots. Approximately half of the genes also respond to chitin, suggesting overlapping functions of both elicitors. These results are comparable to those from Güimil et al. (2005), who showed that more than 40% of the rice (Oryza sativa) genes responded to a beneficial and pathogenic root-colonizing fungus. Among the genes that are regulated only by CT (and not chitin) are candidates for proteins that participate in general root and cell growth, growth-related hormone functions, exocytosis, and transport processes. Compared with chitin, CT induces fewer defense-related genes, and their stimulation is lower and temporally retarded (Fig. 8; Tables II and III; Supplemental Tables S2A and S2B). CT-responsive genes can be classified further into those that require fully functional PARN and those that are regulated independently of PARN (Supplemental Tables S2A and S2B). At least one PARN-dependent pathway activated by CT operates via [Ca²⁺]_cyt elevation. Unlike chitin and the CWE from A. brassicae, several CT-induced and PARN-dependent genes code for proteins involved in growth and developmental processes (Table II; Supplemental Table S3).

PARN

In mammals, yeast, and insects, the adenosine-specific 3′-to-5′ exonuclease PARN mediates the trimming of poly(A) tails of ∼2% of all messenger RNAs. The enzyme plays an important role in mRNA stability, the quality control of gene expression, and the maturation of a class of small RNAs. PARN forms a dimer that recognizes the poly(A) tail and binds simultaneously to the 7-methyl-guanosine cap on the 5′ end of mRNAs in mammals and insects (Godwin et al., 2013). Well-characterized PARN proteins contain three conserved domains; however, two of them, the R3H and RRM domains, are absent in the plant proteins, suggesting that they are not true homologs. Therefore, the exact function of AtPARN is not known.

AtPARN is essential for embryogenesis, and a complete knockout mutant is lethal (Chiba et al., 2004; Reverdatto et al., 2004). Weak mutants accumulate elevated ABA levels, and 12-d-old cycam seedlings contain ∼2 times more ABA than the wild-type control (Fig. 7B; Johnson et al., 2013). They are sensitive to exogenously applied ABA in germination and growth assays (Nishimura et al., 2005, 2009; Johnson et al., 2013). Both features were confirmed for our cycam mutant (Fig. 7B; Johnson et al., 2013).

The Arabidopsis protein is located in the nucleus and cytoplasm, and the ectopically expressed protein showed poly(A) degradation activity in vitro (Chiba et al., 2004). Hirayama et al. (2013) demonstrated that AtPARN directly regulates the poly(A) tract of mRNA in conjunction with a bacteria-type poly(A) polymerase. Whether the protein has dual functions in the nucleus/cytoplasm and mitochondria is currently under study. The PARN mRNA is stress inducible, and a mutant with a weak parn allele is ABA sensitive and responds abnormally to salicylic acid (Nishimura et al., 2005, 2009). In wild-type plants, the amount of PARN mRNA is not limiting for the Ca²⁺ response, since the 35S::PARN plants do not show higher levels of [Ca²⁺]_cyt elevation in response to CT treatments. Our data suggest that the protein degrades the poly(A) tail of an mRNA that encodes one or more proteins required for [Ca²⁺]_cyt elevation in response to CT treatments and the CWE of A. brassicae in the cytoplasm of Arabidopsis roots. Candidates are proteins involved in all processes from elicitor perception to the syntheses or regulation of the corresponding Ca²⁺ channel activities, and they are likely negative regulators of CT-induced [Ca²⁺]_cyt elevation (Fig. 9). The genetic link between Ca²⁺, ROS, and electric signals extends the list to additional candidates. The differences between the root responses to CT and those to the A. brassicae CWE suggest that specificity is achieved downstream of PARN (Table II). The absence of a functional PARN may result in the loss of a number of proteins of which only a specific subset is involved in the realization of a particular elicitor program. Our data suggest that elicitor-triggered [Ca²⁺]_cyt elevation can be controlled at the level of mRNA stability via poly(A) tail shortening. In various eukaryotic organisms, the deadenylation reaction of specific mRNA species by PARNs is allosterically regulated, which may add an additional posttranscriptional control step to decoding of the CT information. In mammalian cells, PARN-mediated mRNA deadenylation is under the control of cis-acting regulatory elements, which include AU-rich elements and microRNA (miRNA) targeting sites within the 3′ untranslated region of the mRNAs (Zhang et al., 2015, and refs. therein). Deadenylases promote miRNA-induced mRNA decay through their interaction with an miRNA-induced silencing complex (Zhang et al., 2015). Several miRNAs respond to CT treatment in the roots in a Ca²⁺-/PARN-dependent manner (Supplemental Table S2). These can be starting points to investigate the potential role of miRNAs in this scenario.

CONCLUSION

In conclusion, we identified CT as a novel elicitor that induces [Ca²⁺]_cyt elevation in Arabidopsis roots. CT is a simple chemical compound that can easily be synthesized or released from various abundantly available carbohydrate polymers in the rhizosphere. Almost all organisms contain an enzymatic repertoire with the capability to synthesize CT from cellobiose and Glc. Since it is active in the nanomole range, it may function as a sensor of environmental changes. The phenotype of the identified cycam mutant demonstrates that CT-induced [Ca²⁺]_cyt elevation is necessary for a proper plant response to environmental cues. We show that PARN is required for [Ca²⁺]_cyt elevation induced by at least two quite different elicitors: CT from a beneficial fungus and a low-molecular-mass compound from a pathogenic fungus (Fig. 2; Supplemental Fig. S6; Johnson et al., 2013). Thus, PARN appears to control different elicitor-induced signaling events. Posttranscriptional control of [Ca²⁺]_cyt elevation by selective shortening of poly(A) tails of specific mRNAs required for [Ca²⁺]_cyt elevation adds an additional novel feature to elicitor-induced signaling events in plants.

MATERIALS AND METHODS

Culturing and Growth Conditions of Fungi and Arabidopsis

Twelve-day-old Arabidopsis (Arabidopsis thaliana) seedlings were transferred from Murashige and Skoog plates to plates with solid plant nutrition medium and a nylon membrane (mesh size, 70 µm) as described previously (Camehl et al., 2011). Seedlings were then grown for an additional 24 h under long-day conditions, with light applied from above (60–70 µmol m⁻² s⁻¹, 16 h of light/8 h of dark), at 22°C. These seedlings were then used for the different experiments: transfer to fresh plates with fungi or incubation of the roots with elicitors.

For elicitor application, the roots were soaked in a solution containing either CT or other chemicals (see below), and autoclaved water was used as a control. The plates with the seedlings were then transferred back to long-day conditions, before the roots were harvested after the treatments described in the text.

Transgenic Arabidopsis expressing cytosolic apoaequorin in the Col-0 background (pMAQ2) was a gift from Marc Knight (Knight et al., 1991; Polisensky and Braam, 1996), rbohd knockout seeds were a gift from Jonathan D.G. Jones, and tpc1 in the aequorin background was a gift from Edgar Peiter. Knockout lines for GRL2.4 (SALK_010571C), GRL2.5 (SALK_078407C), and GRL3.3 (SALK_099757C) were obtained from TAIR. After crossing with pMAQ2, homozygote knockout lines were generated using the primer pairs given by TAIR. Ethyl methanesulfonate mutants of pMAQ2 seeds, which were screened here, were generated in an earlier study (Johnson et al., 2014).

Piriformospora indica was cultured and maintained on Kaefer medium, pH 6.5, as described by Johnson et al. (2011b). The fungus also was grown on Kaefer medium broth for 18 d at 22°C to 24°C in complete darkness on a horizontal rotating shaker at 50 rpm for the preparation of the cell wall extracts (Vadassery and Oelmüller, 2009; Johnson et al., 2011b).

Alternaria brassicae (FSU-3951), Mortierella hyalina (FSU-509), and Verticillium dahliae (FSU-343) were obtained from the Jena Microbial Resource Centre, and Periconia macrospinosa was from Gabor Kovács. The fungi were grown on potato dextrose agar medium (pH 6.5–6.7) at 20°C ± 1°C in a temperature-controlled chamber under 12/12 h of light/dark and 75% relative humidity for 2 weeks. The fungi were inoculated to Arabidopsis seedlings and reisolated from the infected tissues every 6 months (Johnson et al., 2013). Agrobacterium tumefaciens was grown for 3 d in a medium containing 1% (w/v) yeast extract, 1% (w/v) tryptone peptone, and 0.5% (w/v) NaCl (pH 7) at 27°C, with shaking at about 200 rpm.

[Ca²⁺]_cyt Measurement and Mutant Screen

Aequorin-based luminescence measurements were performed using 16-d-old individual M2 plants grown on Hoagland medium (Vadassery and Oelmüller, 2009; Vadassery et al., 2009; Johnson et al., 2011b); pMAQ2 plants served as a control (Polisensky and Braam, 1996). For [Ca²⁺]_cyt measurements, approximately 70% of the roots per seedling were dissected and incubated overnight in 150 µL of 7.5 µm coelentrazine (P.J.K.) in the dark at 20°C on a 96-well plate (Thermo Fisher Scientific; catalog no. 9502887). Bioluminescence counts from roots were recorded as RLU with a microplate luminometer (Luminoskan Ascent, version 2.4; Thermo Electro). For the induction of the Ca²⁺ response, the CWE and purification fractions, CT (Sigma-Aldrich; C1167), other Glc oligomers (cellobiose, C7252; cellotriose, C1167; cellotetraose, C8286; cellopentaose, C8792 [Sigma-Aldrich]; cellohexaose, OC06512; celloheptaose OC05241 [Carbosynth]), monomeric sugars, flg22 (Biolab; up7201-m5) or chitin (chitohexaose, OH07433 [Carbosynth]), chitopentaose, chitoheptaose, and chitooctaose (P6967, H1271, and O6383 [Sigma-Aldrich]) were used as described in “Results.” The mutant screen was performed with the CWE from P. indica; the putative M2 mutants were rescued and transferred to pots containing garden soil and vermiculite at 9:1 (v/v) for further screening and validation. The mutant seedlings were grown in a temperature-controlled growth chamber under short-day conditions (8-h/16-h light/dark cycle, temperature of 20°C ± 1°C, light intensity of 80 µmol m⁻² s⁻¹) for 4 weeks followed by long-day conditions in Aracon tubes. The seeds were harvested from individual M3 plants and again screened to confirm homozygosity.

Purification of the Ca²⁺-Inducing Compound from P. indica

CWEs were prepared according to Anderson-Prouty and Albersheim (1975) with modifications (Johnson et al., 2011a, 2014; Lee et al., 2011). Mycelia from liquid cultures were harvested by filtration through four layers of nylon membrane (pore size, 70 μm; Sefar) and washed five times with sterile water before homogenization in sterile water (1:5, w/v) with a Waring blender. The slurry was then filtered through four layers of nylon membrane. The residue was collected, washed again (three times) with sterile water, then twice with chloroform:methanol (1:1), and finally twice with acetone. The material representing a crude mycelial cell wall preparation was air dried for 2 h under sterile conditions. For the preparation of the CWE, the material was suspended in sterile water (1 g:100 mL, w/v) and autoclaved for 30 min according to Anderson-Prouty and Albersheim (1975). However, autoclaving is not necessary for the elicitor preparation. After cooling, the extract was filtered through four layers of nylon membrane and finally sterilized by passing it through a 0.22-μm filter. The CWE was then purified by passing it through a reverse-phase Supelclean LC-18 SPE cartridge (10-g bed weight, 60-mL volume; Sigma-Aldrich; catalog no. 57136). The eluting fractions were collected, and the active fraction in the void volume was identified by [Ca²⁺]_cyt elevation measurements (Johnson et al., 2011a, 2014; Lee et al., 2011). This fraction from P. indica was used for the Arabidopsis mutant screen described below. The CWEs from A. brassicae, V. dahliae, P. macrospinosa, M. hyalina, and A. tumefaciens were prepared in the same way.

For purification of the active compound in the P. indica cell wall preparation, the proteins were removed by precipitation with 80% (v/v) methanol, and the supernatant was collected after centrifugation at 6,000g for 5 min. The supernatant was applied to a Roti-Spin Mini column with a molecular mass cutoff of 3 kD (Roth), and the flow through was concentrated in a Speed-Vac. Further purification was performed by HPLC with an LC-18-DB column, 25 cm × 4.60 mm i.d. (Supelco), followed by an Asahipak NH2P-50 4E column, 25 cm × 4.6 mm i.d. (Schodex). The active fractions were collected for measuring [Ca²⁺]_cyt elevation, pooled, and concentrated in a Speed-Vac. Finally, the active fractions were separated using two UPLC columns, first an Acclaim C18 column, 250 × 2.1 mm, 2.2 μm (Dionex), and then a C18 phenyl column, 150 × 2.1 mm, 1.7 µm (Phenomenex), fitted to an Ultimate 3000 series RSLC system (Dionex). For all HPLC and UPLC separations, an acetonitrile:water gradient (0–100%) was used as the mobile phase (30-min run) with flow rates of 1 mL min⁻¹ for HPLC and 0.3 mL min⁻¹ for UPLC.

Structure Elucidation of the Ca²⁺-Inducing Compound from P. indica

Prepurified active fractions from the above-mentioned separations were analyzed on an Ultimate 3000 series RSLC system (Dionex) coupled to an LTQ-Orbitrap XL Mass Spectrometer (Thermo Fisher Scientific) equipped with an electrospray ionization source. The Orbitrap mass analyzer was set to 60,000 m/Δm resolving power mass resolutions at m/z 200 and operated in positive or negative ion modes. The instrument was calibrated using commercial CallMix (Thermo) prior to each sequence. Continuum spectral data of m/z 70 to 2,000 mass span were collected and analyzed using Xcalibur version 2.0.7. software (Thermo Fisher Scientific). For collision-induced dissociation tandem mass spectrometry spectra, masses of interest were selected in a quadrupole using a 2D selection window, and normalized fragmentation energy was set from 10 to 35 V.

¹H-NMR spectra, ¹H,¹H-DQF-COSY spectra, selTOCSY spectra, multiplicity-edited heteronuclear single quantum coherence spectra, and heteronuclear multiple-bond correlation spectra were measured at 300°K on a Bruker Avance III HD 700 NMR spectrometer (Bruker Biospin) using a cryogenically cooled 1.7-mm TCI ¹H-¹³C probe. The operating frequency was 700.45 MHz for ¹H and 176.13 MHz for ¹³C. D₂O was used as a solvent, and the spectra were referenced to sodium trimethylsilyl propionate as a standard. The residual HDO signal in the ¹H-NMR spectra was suppressed using the PURGE pulse program (Simpson and Brown, 2005).

The ¹H-NMR spectrum of the trisaccharide displayed four doublets characteristic for protons at the anomeric center of carbohydrates. The coupling constants of ³J_H-H = 8 Hz of the signals at δ 4.53, 4.55, and 4.68 indicated β-configuration of hexopyranoses, and ³J_H-H = 4.1 Hz of the doublet at δ 5.24 indicated α-configuration (Supplemental Fig. S2). The integral values of 0.4 for the signal at δ 5.24 and 0.6 for the signal at δ 4.68 add up to a common value of 1, indicating that the two signals represent α- and β-anomers of the terminal carbohydrate unit. The two signals at δ 4.55 (¹H) and δ 4.53 (¹H) each represent a nonreducing carbohydrate unit in the molecule. ¹H,¹H-DQF-COSY (Supplemental Fig. S3) and the selTOCSY experiments (Supplemental Fig. S4) dissected the crowded region of the carbohydrate methine and methylene signals and enabled the assignment of all ¹H-NMR signals of the three sugar units. Large coupling constants of ³J_H-H = 8 to 9 Hz indicated transdiaxial configuration of H-1 to H-5 in each unit and, hence, identified them as three glucopyranoses, A, B, and C. Moreover, the H-4 signals of Glc-A and Glc-B shifted to the low field at δ 3.67 and 3.68, respectively, suggesting 1→4 linkage of the sugar units. According to these data, the structure of the trisaccharide was identified as CT.

For acid hydrolysis of the purified fraction from the P. indica and A. brassicae cell wall (Supplemental Fig. S6) and the commercially available CT (Fig. 2; Supplemental Fig. S6), the samples (50 μL) were treated with 1 n H₂SO₄ in a water bath at 80°C for 1 h and neutralized with 1 n NaOH after cooling, before use for [Ca²⁺]_cyt measurements.

Quantitative Intracellular H₂O₂ and ROS Measurements

Quantitative H₂O₂ measurement from leaves and roots were performed using the Amplex Red Hydrogen Peroxide/Peroxidase Assay Kit (Molecular Probes, Invitrogen) according to the manufacturer’s instructions. Leaf sections of 0.5 to 1 mm width and root sections of 2 to 3 cm length were incubated in the reaction mixture for 10 min in darkness at room temperature. The fluorescence intensity was quantified with a fluorescence microplate reader (TECAN Infinite 200 plate reader) 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 a control.

Gene Expression Analyses by RT-qPCR

RNA was isolated from the tissues with TRIzol (Thermo Fisher Scientific) according to the manufacturer’s protocol followed by DNase treatment. The RNA was reverse transcribed for RT-qPCR analyses using a CFX CONNECT real-time detection system (Bio-Rad). Total RNA was isolated from three independent biological experiments. cDNA was synthesized using the Omniscript cDNA synthesis kit (Qiagen) with 1 μg of RNA. For the amplification of the reverse transcription PCR products, iQ SYBR Green Supermix (Bio-Rad) was used according to the manufacturer’s protocol in a final volume of 20 μL. The cycler was programmed to 95°C for 3 min; 40 cycles of 95°C for 30 s, 57°C for 15 s, and 72°C for 30 s; 72°C for 10 min; followed by a melting curve program from 55°C to 95°C in increasing steps of 0.5°C. All reactions were performed with three biological and three technical replicates. The mRNA levels for each cDNA probe were normalized with respect to the plant ACTIN2 mRNA level. Fold induction values of target genes were calculated with the ΔΔCP equation (Pfaffl, 2001) and related to the mRNA level of the target genes as indicated. Primer pairs used in this study are given in Supplemental Table S4. Primers were designed with the CLC Main Workbench program (http://www.clcbio.com/products/clc-main-workbench).

Identification of the cycam Mutation

The cycam mutant was crossed with the pMAQ2 line, and in the F2 generation, two pools of offspring plants were generated. The first pool contained 25 individual plants that displayed the cycam phenotype, and the second pool contained plants that displayed a wild-type-like phenotype (control pool). The DNA of both pools was analyzed by Illumina (100-bp paired-end) sequencing. The reads from the first and second pools were mapped separately against the TAIR10 reference genome (Lamesch et al., 2012) using CLC Genomics Workbench (version 5.1). Single-nucleotide polymorphisms (SNPs) with a frequency between 20% and 100% were detected using the variant detection tool in both mappings, and the resulting SNP tables were exported to Microsoft Excel. In Excel, SNPs were identified that had higher frequencies in the cycam pool than in the control pool using the custom sorting tool. For each of the five Arabidopsis chromosomes, the frequencies of these SNPs were plotted against their reference positions. The obtained scatterplots were inspected manually for areas of high linkage (SNP frequencies near 100%). A region of about 3.5 Mb on the lower arm of chromosome 1 was identified that contained five SNPs with a frequency of 100% in annotated genes (loci At1g53310, At1g54985, At1g55870, At1g56140, and At1g60110). Of these, two nonsynonymous SNPs were predicted to cause amino acid exchanges in the corresponding gene products of At1g53310, encoding PHOSPHOENOLPYRUVATE CARBOXYLASE1, and At1g55870, encoding AtPARN. At1g55870 was analyzed further (see below).

Generation of the Construct for the Complementation of cycam and Arabidopsis Transformation

The Gateway system with pDonor207 was used as an entry vector, and pEarlyGate 103 with a BASTA resistance and stop codon before the C-terminal GFP-Taq was used as a destination vector (Earley et al., 2006; Katzen, 2007). For the amplification of the cDNA of the wild-type version of AtPARN (At1g55870), Phusion High-Fidelity-Taq (Thermo Fisher) was used for amplification of the cDNA and, in a second PCR, fusion of the AttB sequence. The DNA was purified by the NucleoSpin Gel and PCR Clean-Up Kit (Macherey-Nagel). The sequences were confirmed by sequencing (Eurofinsgenomics). Parental aequorin and cycam plants were transformed with the wild-type version of AtPARN using the floral dip method with A. tumefaciens strain GV3101 (Zhang et al., 2006). Complemented plants were selected using a 0.1% (v/v) BASTA solution 14 and 18 d after sowing.

Microarray Analyses

Total RNA from roots of transgenic Arabidopsis (Col-0) expressing cytosolic apoaequorin (pMAQ2; Knight et al., 1991; Polisensky and Braam, 1996) and the cycam mutant (in the pMAQ2 background) from three independent biological experiments was exposed to CT or chitin or mock treated with water and harvested at 4 and 8 h after elicitor application. For each treatment, identical amounts of RNA from the three independent biological replicates were labeled and hybridized according to Agilent’s One-Color Microarray-Based Gene Expression Analysis (OAK Lab). The quality of the RNA samples was checked by photometrical measurements with the Nanodrop 2000 spectrophotometer (Thermo Scientific) and then analyzed on agarose gels (2%, w/v) as well as by using the 2100 Bioanalyser (Agilent Technologies) to determine the RNA integrity. The Low Input Quick Amp Labeling Kit (Agilent Technologies) was used for the generation of fluorescent cRNA. Default cRNAs were amplified using oligo(dT) primers labeled with cyanine 3-CTP according to the manufacturer’s protocol. Cyanine 3-CTP-labeled probes were hybridized to 8 × 60 k custom-designed Agilent microarray chips. For hybridization, the Agilent Gene Expression Hybridization Kit (Agilent Technologies) was used. The hybridized slides were washed and scanned using the SureScan Microarray Scanner (Agilent Technologies) at a resolution of 3 μm, generating a 20-bit TIFF file.

Data extractions from images were performed using Agilent’s Feature Extraction software version 11. Feature-extracted data were analyzed using DirectArray version 2.1 software from Agilent. Data were normalized with DirectArray using ranked median quantiles according to Bolstad et al. (2003). To identify significantly differentially expressed genes, log₂ fold changes were calculated and Student’s t test was performed. In summary, raw data were normalized by rank median quantiles, intensity values from replicate probes were averaged, log₂ ratios between the treatments were calculated, and Student’s t statistics were applied to test for significance. Genes with log₂-fold change < −1.33 or > 1.33 and P < 0.05 were considered to be significantly different. Data show genes that are regulated by CT or chitin in wild-type or cycam roots. Differentially expressed genes were then assigned using the Arabidopsis Gene Ontology software (TAIR’s GO annotations; Berardini et al., 2004) and transcript abundance was classified based on their functional categories and pathways using MapMan software. Microarray data were submitted to the NCBI and are available under accession number GEO GSE9888.

Membrane Potential Recordings of Bulging Root Hair Cells and CT Application

Sterile germination and growth of Col-0 and cycam seedlings for root hair plasma membrane potential recordings took place as described by Wang et al. (2015). Col-0 and cycam seedlings were placed in a bath solution (0.1 mm KCl, 1 mm CaCl₂, and 5 mm MES/BTP, pH 5.5) at least 20 to 30 min before the start of the experiments. The response of the root hair plasma membrane potential to CT application was measured by sharp microelectrodes impaled through the tip of bulging root hair cells. Microelectrodes were prepared and plasma membrane potential recordings took place as described by Wang et al. (2015). CT was applied via back pressure-operated application pipettes produced from microelectrodes. The tips of the microelectrodes were manually broken off to yield openings of approximately 20 to 40 µm in diameter. Pipettes, back filled with CT-containing bath solution, were mounted on a Triple Axis Micromanipulator (Sensapex Oy), and pressure pulses of 1 s were applied through a Picospritzer II microinjection dispense system (General Valve) operating at 30 p.s.i.

ABA and Stomatal Closure Measurements

A total of 100 mg of leaf material was frozen in liquid nitrogen and kept at −80°C. The extraction procedure and determination of the ABA concentrations were performed as described by Matsuo et al. (2015). Stomatal closure was determined as described by Vahabi et al. (2016).

Performance of cycam in the Presence of P. indica

The fresh weights of wild-type and cycam seedlings in the presence or absence of P. indica were determined as described by Johnson et al. (2014) 6 d after cocultivation (19-d-old seedlings). Drought tolerance was tested as described by Sherameti et al. (2008). On day 13, the seedlings were transferred to fresh medium with P. indica mycelium and kept for an additional 6 d. Then, the lids were removed. After 72 h, differences in fresh weight were determined.

Accession Numbers

Sequence data from this article can be found in the GenBank/EMBL data libraries under accession numbers.

Supplemental Data

The following supplemental materials are available.

• Supplemental Figure S1. Purification of the fungal compound inducing [Ca²⁺]_cyt elevation from the P. indica CWE.

• Supplemental Figure S2. Full and expanded partial ¹H-NMR spectra of the trisaccharide from P. indica.

• Supplemental Figure S3. ¹H,¹H-DQF-COSY spectrum of the trisaccharide from P. indica.

• Supplemental Figure S4. Selective total correlation spectra of the trisaccharide from P. indica.

• Supplemental Figure S5. The cycam mutant was backcrossed with the pMAQ2 line, and in the F2 generation, DNA from a pool of offspring plants with the cycam phenotype and from a pool of plants with a wild-type-like phenotype (control pool) were Illumina sequenced.

• Supplemental Figure S6. Acid hydrolysis products of CWEs from A. brassicae, P. indica, and cellooligomers were used to test for [Ca²⁺]_cyt elevation in Arabidopsis roots.

• Supplemental Table S1. Refractory behavior of pathogen-associated molecular patterns and cellooligomers after CT application.

• Supplemental Table S2A. Cycam-dependent response 4 and 8 h after elicitor application.

• Supplemental Table S2B. Cycam-independent response 4 and 8 h after elicitor application.

• Supplemental Table S3. Description of genes that are preferentially regulated by CT.

• Supplemental Table S4. Primers used for this study.