Nonhost resistance (NHR) of plants to fungal pathogens comprises different defense layers. Epidermal penetration resistance of Arabidopsis to Phakopsora pachyrhizi requires functional PEN1, PEN2 and PEN3 genes, whereas post-invasion resistance in the mesophyll depends on the combined functionality of PEN2, PAD4 and SAG101. Other genetic components of Arabidopsis post-invasion mesophyll resistance remain elusive.
We performed comparative transcriptional profiling of wild-type, pen2 and pen2 pad4 sag101 mutants after inoculation with P. pachyrhizi to identify a novel trait for mesophyll NHR. Quantitative reverse transcription-polymerase chain reaction (RT-qPCR) analysis and microscopic analysis confirmed the essential role of the candidate gene in mesophyll NHR.
UDP-glucosyltransferase UGT84A2/bright trichomes 1 (BRT1) is a novel component of Arabidopsis mesophyll NHR to P. pachyrhizi. BRT1 is a putative cytoplasmic enzyme in phenylpropanoid metabolism. BRT1 is specifically induced in pen2 with post-invasion resistance to P. pachyrhizi. Silencing or mutation of BRT1 increased haustoria formation in pen2 mesophyll. Yet, the brt1 mutation did not affect NHR to P. pachyrhizi in wild-type plants.
We assign a novel function to BRT1, which is important for post-invasion NHR of Arabidopsis to P. pachyrhizi. BRT1 might serve to confer durable resistance against P. pachyrhizi to soybean.
Soybean is an important global crop providing oil and protein. Worldwide soybean production is threatened by abiotic stress and by various diseases, such as Asian soybean rust (Stokstad, 2004). The latter disease is caused by Phakopsora pachyrhizi, an obligate biotrophic basidiomycete fungus which, over the past century, has spread from Japan all over the globe (Stokstad, 2004). Soybean rust spreads rapidly and hits plants hard, defoliating fields in < 2 wk. The best defense would be a soybean variety which, in the long run, resists rust. However, no such variety is grown in today's soybean fields. Thus, the development of rust-resistant varieties for soybean production remains a challenge to plant researchers. One possible strategy for the development of soybean varieties with durable resistance to Asian soybean rust is the cross-species transfer of genes encoding components of nonhost resistance (NHR) to P. pachyrhizi.
NHR is the most common and robust form of plant disease resistance. The term describes the immunity of an entire plant species to all genetic variants of a given pathogen (Heath, 2000; Lipka et al., 2008). A major breakthrough in the understanding of NHR was the discovery of the so-called penetration (pen) mutants of Arabidopsis (Collins et al., 2003; Lipka et al., 2005; Stein et al., 2006). These provided first insights into the genetic basis of NHR. In contrast to the wild-type, pen1, pen2 and pen3 mutants allow the invasion of nonadapted fungal pathogens, such as Blumeria graminis f. sp. hordei (Bgh, Collins et al., 2003; Lipka et al., 2005; Stein et al., 2006) and Phakopsora pachyrhizi (Loehrer et al., 2008).
PEN1 encodes a plasma membrane-anchored syntaxin with a soluble N-ethylmaleimide-sensitive factor attachment protein receptor (SNARE) domain (Collins et al., 2003). It is probably involved in secretion and cell wall modification at sites of mildew attack (Collins et al., 2003; Assaad et al., 2004; Bhat et al., 2005).
PEN2 seems to be a glucoside hydrolase-type enzyme with myrosinase activity (Lipka et al., 2005). It probably converts 4-methoxyindol-3-ylmethylglucosinolate into an antifungal derivative (Bednarek et al., 2009; Clay et al., 2009) which is important for disease resistance. Production of the PEN2 substrate and cytoskeleton-mediated peroxisome delivery to fungal penetration sites are important for Arabidopsis NHR to fungal pathogens (Lipka et al., 2008). The pleiotropic drug resistance (PDR) ATP-binding cassette (ABC) transporter PEN3 (also referred to as PDR8) is a transmembrane protein that appears to transport the glucosinolate derivative over the plasma membrane at sites of attempted fungal penetration to counter fungal invasion (Stein et al., 2006; Lipka et al., 2008).
Although pen mutants allow the invasion of nonadapted fungal pathogens, post-invasive fungal growth always ceases in pen, and this is frequently associated with cell death or the formation of effective wall appositions in cells with contact to the pathogen (Collins et al., 2003; Lipka et al., 2005; Stein et al., 2006). This finding strongly suggests that Arabidopsis NHR comprises different layers of defense (Lipka et al., 2005; Loehrer et al., 2008). Indeed, inactivation of additional genes with a role in plant disease resistance supported fungal reproduction on infected Arabidopsis plants. For example, the pen2 pad4 sag101 (for pen2/phytoalexin-deficient 4/senescence-associated gene 101) triple mutant of Arabidopsis allows Bgh and Erysiphe pisi to sporulate (Lipka et al., 2005). This finding demonstrates that NHR of Arabidopsis to these two nonadapted pathogens consists of two layers of defense. One of these layers depends on functional PEN genes mediating pre-invasion resistance, whereas the second requires EDS1 (ENHANCED DISEASE SUSCEPTIBILITY 1), PAD4 and SAG101 which, together, are critical for post-invasion immunity (Lipka et al., 2005). Recently, Nakao et al. (2011) demonstrated compromised NHR in mutants, including pen2 NahG pmr5 agb1 and pen2 NahG pmr5 mlo2, when using Magnaporthe oryzae as a nonhost pathogen. Arabidopsis NHR to rust fungi is poorly understood. Key defense-related mutants, including pen2 and pen3, display intact NHR to the leaf rust pathogen Puccinia triticina (Shafiei et al., 2007). Other rust NHR studies employing Uromyces ssp.-challenged Arabidopsis have corroborated the importance of salicylic acid, evidenced by increased fungal growth on sid2 mutant and NahG transgenic plants (Mellersh & Heath, 2003). Other genetic components of plant NHR against fungal pathogens remain elusive.
To identify novel Arabidopsis mesophyll NHR genes, we performed comparative transcriptional profiling of wild-type, pen2 and pen2 pad4 sag101 mutants after inoculation with P. pachyrhizi. By doing so, we identified the BRIGHT TRICHOMES 1 (BRT1)/UGT84A2 gene (subsequently referred to in this work as BRT1) as a potential novel component of Arabidopsis mesophyll NHR to this pathogen. BRT1 encodes a putative cytoplasmic UDP-glucosyltransferase (UGT84A2, At3 g21560) with a role in phenylpropanoid metabolism. This pathway is essential for plant development and defense (Hahlbrock & Scheel, 1989; Fraser & Chapple, 2011). BRT1 catalyzes the synthesis of sinapoylglucose, the precursor of the predominant Arabidopsis sinapate esters, sinapoylmalate and sinapoylcholine (Fraser & Chapple, 2011). Although BRT1 activity has also been associated with glucosylation, and thus inactivation, of xenobiotics (Sinlapadech et al., 2007; Messer et al., 2003), it has not been linked to plant disease resistance so far.
Materials and Methods
Plant and fungal material
Seeds of Arabidopsis wild-type (Col-0) plants and the pen2-1 (gl1) and pen2-1 pad4-1 mutants were provided by Volker Lipka (Georg-August University, Göttingen, Germany). The pen2-1 pad4-1 sag101 triple mutant was obtained by crossing the sag101 T-DNA insertion line (N661816; SALK_022911) to the pen2-1 pad4-1 double mutant. Fah1-2 and brt1-1 mutants were provided by Clint Chapple (Purdue University, West Lafayette, IN, USA). To provide appropriate double mutants, fah1-2 and brt1-1 were both crossed to pen2-1 (gl1 genetic background), as described below.
Arabidopsis seeds were sown on soil (type VM, Balster Einheitserdewerk GmbH, Fröndenberg, Germany) and stratified at 4°C for 2 d. Plants were grown under short-day conditions in a chamber with an 8-h photoperiod, 120 μmol m−2 s−1 photon irradiance, 22°C and 65% humidity. Five to 6-wk-old plants were inoculated with P. pachyrhizi as described below.
The P. pachyrhizi isolate used in this study was collected in Brazil and maintained on a susceptible soybean cultivar (Thunder 2703RR brand, Thunder Seed, Inc., Hawley, MN, USA). For inoculation, soybean plants were grown at 22°C, 60% relative humidity, 16-h photoperiod and 133 μmol m–2 s–1 photon irradiance in a 2 : 1 (v/v) mixture of soil (P substrate, Balster Einheitserde GmbH, Fröndenberg, Germany) and sand.
For inoculation of Arabidopsis or soybean, uredospores were collected from P. pachyrhizi-infected soybean leaves at 14 d post-inoculation (dpi). Spores were suspended in 0.01% (v/v) Tween-20 at 1 mg ml−1 and used for plant inoculation. Mock inoculations were performed by spraying 0.01% (v/v) Tween-20 with no fungal spores.
Phakopsora pachyrhizi spore suspension was sprayed on Arabidopsis leaves until droplets covered the leaf surface evenly (c. 10 ml per 60 plants). Inoculated plants were covered with moistened plastic domes to ensure high humidity and support fungal spore germination. After 24 h in an 8-h photoperiod, the domes were removed and inoculated plants were grown until analysis.
Inoculated soybean plants were placed in a dark moist chamber (26°C and 95–100% relative humidity) for 24 h, and then grown as described above. Rust pustules became visible on leaves of infected soybean plants c. 10 d after inoculation.
Crossing and genotyping of mutants
Arabidopsis plants were grown in short days for 3 wk as already described, and transferred to a 16-h light period. Plants were crossed after emergence of inflorescences. Progeny was genotyped for the presence of T-DNA or mutant alleles by PCR (for T-DNA insertion lines) or cleaved amplified polymorphic sequence (CAPS) for ethyl methanesulfonate (EMS) mutant analysis. For crossings with brt1 and fah1, the selection of double homozygous plants was based on a lack of epidermal fluorescence followed by CAPS or derived CAPS (dCAPS) analysis. Homozygous mutants in the F2 and F3 generations were used in the experiments. T-DNA insertion in SAG101 (SALK_022911) was verified using 5′-TGCATAAGGGACGTTTTAACG, 5′-ATGTTGAACTCTTCGCCTTTG and 5′-ATTTTGCCGATTTCGGAAC (LBb1.3) primers (http://signal.salk.edu/tdnaprimers.2.html). The presence of pen2–1, pad4-1, fah1-2 and brt1-1 mutant alleles was confirmed by CAPS or dCAPS analysis using the following combinations of gene-specific primers and restriction enzymes: pen2-1: 5′-TTTGGAACTGCTTCATCTTCTTATCAGG, 5′-CCTGTACAAGAAATCAATCACAGATCTTCA, BspPI; PAD4-1: 5′-GCGATGCATCAGAAGAG, 5′-TTAGCCCAAA-AGCAAGTATC, FaqI; fah1-2: 5′-TGGTGTGTACATA-TATGGATGAAGAA, 5′-TAGCAAGAGTGGTGAATATGT-GAAGT, MseI; brt1-1: 5′-CTACCTCCTCATGTGATGC-TCGGAT, 5′-CTGCTAGCTTCGTCGTCTTCA, Sau3AI. DNA was extracted from wild-type and mutant leaves as described by Edwards et al. (1991). Gene-specific primers were used for the amplification of DNA sequences in wild-type and mutant plants by PCR. Restriction digests of PCR products for CAPS or dCAPS analysis were performed according to the manufacturer's guidelines.
Epidermal fluorescence and hyperfluorescent trichomes of brt1 and pen2 brt1 mutants were visualized with a portable UV light device (CAMAG, Berlin, Germany) with an excitation wavelength of 366 nm.
Measurement of sinapoylmalate content
Sinapoylmalate content in leaf material was determined by reverse-phase high-performance liquid chromatography (HPLC) (Supporting Information Methods S1).
Histochemical staining of leaves and bright-field microscopy
For microscopic evaluation, infected leaves were harvested at the indicated times and stained with trypan blue. For this, leaves were submerged in a nonphenolic trypan blue solution (10% (v/v) lactic acid, 10% (v/v) glycerol, 10% (v/v) H2O2, 70% (v/v) ethanol) and heated to 80°C for 75 s. After cooling to room temperature for 10 min, leaves were incubated in 2.5 g ml−1 chloral hydrate for several days before mounting entire leaves in 50% (v/v) glycerol on glass slides. If not used immediately, leaves were kept in chloral hydrate until further analysis.
Fungal development in de-stained Arabidopsis leaves was evaluated by bright-field microscopy using a Leica DMRBE microscope (Leica, Bensheim, Germany). In addition to visualizing fungal structures by trypan blue, callose encasements were determined. They served as a histological marker for the formation of haustoria in de-stained inoculated leaves.
Callose was visualized by incubating trypan blue-stained and subsequently chloral hydrate-de-stained leaves in an aniline blue solution (0.01% (w/v) aniline blue (Honeywell, Riedel-de Haën, Seelze, Germany) in 150 mM KH2PO4, pH 7.5) for 12–24 h. Callose-containing appositions were observed in a Leica TCS SP fluorescence microscope with an epifluorescent filter (A-513804, 340–380 nm excitation and 425 nm emission, Leica). Photographs were taken with a digital JVC KYF 750 camera (JVC Deutschland GmbH, Friedberg, Germany). The interaction of Arabidopsis genotypes with P. pachyrhizi was assessed in trypan blue-stained leaves by determining the frequencies of interaction phenotypes as described in the legend to Fig. 1.
Total RNA was extracted from leaves as described by Chomczynski & Sacchi (1987). One microgram of RNA was transcribed to cDNA using random primers (9-mers) and RevertAid™reverse transcriptase (Fermentas GmbH, Sankt Leon-Rot, Germany) according to the manufacturer's instructions. The accumulation of gene transcripts was quantified in an ABI7300 using SYBR green (Invitrogen) in the following conditions for RT-qPCR: 50°C for 2 min, 95°C for 10 min, 95°C for 15 s, 60°C for 1 min, 95°C for 15 s, 60°C for 1 min and 95°C for 15 s (the third and fourth steps were repeated 40 times). Primers specifically hybridizing to BRT1 (5′-TTCTCCTCCTCTACCTCCTCA, 5′-AGCTAAGAGCTTACCAAGACGAA) were designed according to standard criteria (Udvardi et al., 2008) and off target search was done using Primer Blast tool at National Center for Biotechnology Information (NCBI) (http://www.ncbi.nlm.nih.gov/tools/primer-blast/). The expression of BRT1 was normalized to Actin2. When other reference genes designed by the Genevestigator RefGenes tool (F-box family protein/WD-40 repeat family protein (At5 g21040), vacuolar ATP synthase subunit H family protein, (At3 g42050), nitrogen fixation NifU-like family protein (At5 g49940)) were used for normalization, we obtained similar results (not shown). Data were analyzed using ABI 7300 software and the expression relative to actin was calculated according to Livak & Schmittgen (2001) with 2−(Ct BRT1 − Ct Actin2).
Interaction of Arabidopsis genotypes with P. pachyrhizi
In a previous study, we introduced the Arabidopsis–P. pachyrhizi pathosystem to identify genes with a role in Arabidopsis NHR (Loehrer et al., 2008). We showed that P. pachyrhizi overcomes Arabidopsis pre-invasion NHR and colonizes the mesophyll in pen1, pen2 and pen3 mutants (Loehrer et al., 2008; Goellner et al., 2010). However, post-invasion mesophyll defense is still intact in these mutants, as the development of haustoria was only rarely observed in the single mutants (Loehrer et al., 2008). Here, we aimed to identify genes with a role in Arabidopsis post-invasion mesophyll resistance to P. pachyrhizi. For this, we searched for genes that are important in avoiding the development of fungal haustoria in attacked mesophyll tissue.
As the Arabidopsis pen2 pad4 sag101 triple mutant allows the reproduction of nonadapted fungal pathogens Bgh and E. pisi (Lipka et al., 2005), we speculated that post-invasion mesophyll resistance to P. pachyrhizi could also be affected in pen2 pad4 sag101. To investigate whether this was the case, we first performed a quantitative microscopic assessment of the interaction of P. pachyrhizi with Arabidopsis wild-type (accession Col-0), pen2 and pen2 pad4 sag101 plants (Fig. 1). The gl1 mutation in the employed pen2-1 line does not affect the interaction of Arabidopsis with P. pachyrhizi (Loehrer et al., 2008).
Unlike in the wild-type, in which P. pachyrhizi hardly entered the mesophyll (Fig. 1, category I), the pen2 mutant frequently allowed P. pachyrhizi to invade mesophyll cells (Fig. 1, category II). However, the development of haustoria was seen at only 5–10% of interaction sites (Fig. 1). This finding is consistent with our previous results and confirms the importance of PEN2 in pre-invasion NHR in the epidermis, as well as the presence of an effective post-invasion NHR mechanism in pen2 (Loehrer et al., 2008).
Contrary to pen2, the pen2 pad4 sag101 triple mutant not only allowed P. pachyrhizi to colonize the mesophyll (Fig. 1), but also the portion of interaction sites showing haustoria (Fig. 1, category III) was about five times greater in pen2 pad4 sag101 than in pen2, and pertained to 30–50% of all interaction sites (Fig. 1). When haustoria developed, these were strongly encased by sirofluor-positive material, which most likely represents callose (Figs 1, S1). We assume that these haustoria are nonfunctional as fungal growth ceased after haustoria formation.
Phakopsora pachyrhizi-induced BRT1 activation in pen2
Using Affymetrix GeneChips, we identified novel putative post-invasion mesophyll NHR genes by comparing the transcriptional profiles of wild-type, pen2 and pen2 pad4 sag101 leaves at 2 d after P. pachyrhizi infection (not shown because we used only two biological replicates to identify post-invasion NHR candidate genes). At this point, the fungus has already established intercellular hyphae and/or haustoria in the mesophyll (Loehrer et al., 2008). We anticipated that genes with a role in Arabidopsis post-invasion mesophyll resistance would be activated in the inoculated pen2 mutant, but not in infected pen2 pad4 sag101 plants with attenuated post-invasion NHR in the mesophyll (Fig. 1). To strengthen the selection criteria, we searched for genes whose expression would not be activated in the wild-type, in which P. pachyrhizi hardly enters the mesophyll at all (Fig. 1; Loehrer et al., 2008). Genes with altered expression in response to mock inoculation in any genotype assayed were also disregarded. This was to exclude genes with differential expression as a result of mutant background and/or mock treatment. Duplicate experiments were performed independently, and data were analyzed with the FiRe program (Garcion et al., 2006). Genes showing consistent activation only in the P. pachyrhizi-infected pen2 mutant were considered to be putative components of Arabidopsis post-invasion mesophyll NHR to P. pachyrhizi. Amongst them, BRT1 (At3 g21560) was induced significantly at the second day after P. pachyrhizi inoculation in pen2. BRT1 codes for a UDP-glucosyltransferase that is localized to the phenylpropanoid pathway and is involved in sinapate ester biosynthesis (Sinlapadech et al., 2007).
We used RT-qPCR to analyze BRT1 transcript abundance in wild-type, pen2 and pen2 pad4 sag101 plants at 2 d after inoculation to determine whether BRT1 is specifically activated in P. pachyrhizi-infected pen2. As shown in Fig. 2, BRT1 was expressed at basal levels in mock-inoculated leaves of 6-wk-old Arabidopsis plants irrespective of their genetic background. Consistent with our microarray data (not shown), BRT1 expression was activated three- to four-fold in the P. pachyrhizi-infected pen2 mutant, but not in the wild-type or the pen2 pad4 sag101 triple mutant (Fig. 2).
Attenuated post-invasion NHR in pen2 BRT1 dsRNAi lines
After verification of BRT1 as a gene specifically activated in P. pachyrhizi-infected pen2, we investigated the contribution of BRT1 to mesophyll NHR using an RNAi assay. Pen2 plants were transformed with a silencing construct causing hairpin formation of the BRT1-specific sequence to specifically silence BRT1. Transformants of the T1 and T2 generations were inoculated with P. pachyrhizi. Two days later, inoculated leaves were harvested from different pen2 BRT1 dsRNAi lines and subjected to BTR1 transcript quantification. Additional leaves were stained with trypan blue to visualize fungal structures and dead or dying plant cells in microscopic analysis (Fig. 3). Most pen2 BRT1 dsRNAi lines showed a clear reduction in P. pachyrhizi-elicited BRT1 activation when compared with pen2 (Fig. 3b). The silencing of BRT1 in pen2 BRT1 dsRNAi lines led to a reduction in sinapoylmalate levels that was similar to that found in the brt1 mutant (Fig. S3). In addition, BRT1 silencing in the wild-type coincided with the hyperfluorescence of trichomes in UV light (Fig. S3), just as described for the brt1 mutant (Sinlapadech et al., 2007).
For microscopic assessment of haustoria formation, leaves of pen2 BRT1 dsRNAi lines with severely reduced BRT1 expression were stained with trypan blue. Figure 3a shows that efficient BRT1 silencing in pen2 BRT1 dsRNAi lines coincided with increased haustoria formation. The number of penetration events with haustoria development was c. four-fold higher in efficiently silenced pen2 BRT1 dsRNAi lines when compared with pen2, and was similar to the extent of haustoria appearance in pen2 pad4 sag101 (Fig. 3). Moreover, a quantitative relationship was observed for BRT1 silencing and haustoria formation. Moderate BRT1 silencing only slightly enhanced haustoria development in infected pen2 BRT1 dsRNAi leaves, whereas strong BRT1 down-regulation led to extensive haustoria formation in this genotype (Fig. 3).
Loss of post-invasion mesophyll resistance in the pen2 brt1 double mutant
To examine the above findings with pen2 BRT1 dsRNAi lines, brt1 mutants were crossed to pen2. Pen2 brt1 double mutants were inoculated with P. pachyrhizi, harvested at 2 dpi, stained with trypan blue and analyzed microscopically for haustoria development. Similar to the pen2 BRT1 dsRNAi lines, the pen2 brt1 double mutant was impaired in post-invasion mesophyll NHR to P. pachyrhizi (Fig. 4). The amount of penetration events with haustoria formation was nearly seven-fold higher in pen2 brt1 than in pen2 (Fig. 4). In addition, the number of invasion sites with haustoria even exceeded that observed in the infected pen2 pad4 sag101 mutant (Fig. 4). The score of interaction sites with haustoria remained at the same level, or increased only slightly, until 4 d after P. pachyrhizi infection in the pen2 brt1 double and pen2 pad4 sag101 triple mutants (Fig. S2). Similar to the above observations with the infected pen2 pad4 sag101 mutant, haustoria in mesophyll cells of the pen2 brt1 double mutant were strongly encased with callose, and mesophyll cells containing haustoria did not undergo cell death (not shown).
Brt1 has a hyperfluorescent trichome phenotype in UV light because of the accumulation of polyketides in these cells (Sinlapadech et al., 2007). The pen2-1 mutant used in this study lacks trichomes because of its glabrous (gl1) genetic background. GL1 encodes a transcription co-activator with a role in trichome development (Oppenheimer et al., 1991). Because of the absence of trichomes in pen2 and pen2 brt1, polyketides could localize to other cells or cellular compartments in these genotypes, thereby affecting mesophyll NHR to P. pachyrhizi. To investigate whether the infection phenotype of pen2 brt1 and pen2 BRT1 dsRNAi lines was caused by a lack of trichomes, pen2 brt1 and pen2 brt1 gl1 mutants were compared in terms of their ability to express post-invasion mesophyll resistance to P. pachyrhizi. As shown in Fig. 4, the frequency of haustoria in P. pachyrhizi-infected leaves was independent of the absence (pen2 brt1 gl1) or presence (pen2 brt1) of trichomes at 2 d after infection. This finding indicates that attenuated post-invasion NHR in the pen2 brt1 gl1 mutant is not caused by the absence of trichomes in this plant.
Absence of sinapoylmalate does not affect post-invasion mesophyll NHR
As sinapoylmalate levels are reduced in leaves of the brt1 mutant (Ruegger & Chapple, 2001), the impaired post-invasion mesophyll resistance in the pen2 brt1 mutant and pen2 BRT1 dsRNAi lines could be caused by decreased sinapoylmalate levels in these plants. We tested this possibility by crossing pen2 to the sinapoylmalate-deficient fah1 mutant and by scoring haustoria formation in P. pachyrhizi-infected, homozygous pen2 fah1 plants at 2 d after inoculation. Figure 4 shows that post-invasion mesophyll NHR was not affected in pen2 fah1, in which haustoria formation was similar to that in pen2. This finding suggests that the reduction in sinapoylmalate levels in pen2 brt1 or pen2 BRT1 dsRNAi lines is not the cause of attenuated mesophyll NHR to P. pachyrhizi in these plants.
No role of BRT1 in pre-invasion NHR in the epidermis
To investigate whether the contribution of BRT1 to Arabidopsis NHR to P. pachyrhizi is exclusive for post-invasion NHR in the mesophyll, or whether it also contributes to pre-invasion NHR to this pathogen in the epidermis, fungal invasion was scored in P. pachyrhizi-inoculated brt1 and wild-type plants. As shown in Fig. 5, brt1 did not allow significantly increased fungal penetration when compared with the wild-type. Thus, BRT1 is a novel component of Arabidopsis NHR with a specific role in post-invasion mesophyll resistance to P. pachyrhizi.
NHR of plants to fungal pathogens comprises pre- and post-invasion defense mechanisms (Lipka et al., 2005; Nakao et al., 2011). Here, we showed that BRT1 is an essential and specific component of post-invasive mesophyll resistance of Arabidopsis to P. pachyrhizi. As the pen2 brt1 and pen2 pad4 sag101 mutants had similar phenotypes in terms of haustoria formation (Fig. 4), and as BRT1 expression was reduced in P. pachyrhizi-infected pen2 pad4 sag101 when compared with the infected pen2 mutant (Fig. 2), the BRT1 gene seems to be controlled by PAD4 and SAG101.
BRT1 is localized to the phenylpropanoid pathway and catalyzes the glucosylation of sinapic acid to sinapoylglucose (Milkowski et al., 2000; Lim et al., 2001; Fraser et al., 2007; Fraser & Chapple, 2011), which serves as a substrate for sinapoylmalate and sinapoylcholine synthesis in leaves and seeds, respectively (Strack, 1980, 1982, 1983; Mock & Strack, 1993; Lehfeldt et al., 2000; Shirley et al., 2001). In UV light, sinapoylmalate causes the leaf epidermis of wild-type Arabidopsis plants to fluoresce (Chapple et al., 1992; Ruegger et al., 1999; Ruegger & Chapple, 2001). This fluorescence is reduced in brt1 and other mutants in the phenylpropanoid pathway, referred to as reduced fluorescence (ref) mutants (Ruegger & Chapple, 2001; Fraser & Chapple, 2011). In addition to being attenuated in sinapoylmalate biosynthesis and the associated epidermis fluorescence, brt1 also has hyperfluorescent trichomes in UV light (Ruegger & Chapple, 2001; Sinlapadech et al., 2007; Fraser & Chapple, 2011). The fluorescent trichome phenotype is caused by the accumulation of sinapic acid-derived polyketides in trichome cells (Sinlapadech et al., 2007; Fraser & Chapple, 2011). Although the role of BRT1 in phenylpropanoid metabolism has been thoroughly studied, it has not been associated with plant disease resistance to date. Only gene expression data in the online databases Genevestigator and eFP browser have indicated the transcriptional induction of BRT1 in response to pathogens or pathogen-derived signals, such as flagellin 22, HrpZ and GST-NPP1.
Various enzymes in the phenylpropanoid pathway and several UDP-glucosyltransferases other than BRT1 have been associated previously with the pathogen defense of plants (Horvath & Chua, 1996; Fraissinet-Tachet et al., 1998; O'Donnell et al., 1998; Mazel & Levine, 2002; Nishimura et al., 2003; Poppenberger et al., 2003; Langlois-Meurinne et al., 2005). Yet, clear evidence for the assumed important role of phenylpropanoid enzymes in disease resistance is available only for some of them. For example, the tobacco UDP-glucosyltransferase TOGT1 contributes to the resistance to tobacco mosaic virus and potato virus Y in tobacco by regulating the accumulation of the antiviral secondary metabolite scopoletin and/or its glucoside scopolin (Gachon et al., 2004; Matros & Mock, 2004; Langlois-Meurinne et al., 2005). In addition, in Arabidopsis, the UGT DOGT1 (UGT73C5) can inactivate and thus enhance tolerance to the fusarium toxin deoxynivalenol (Poppenberger et al., 2003). Furthermore, the Arabidopsis UDP-glucosyltransferase mutants ugt73b3 and ugt73b5 show reduced resistance to Pseudomonas syringae pv. tomato harboring the avirulence gene AvrRpm1 (Langlois-Meurinne et al., 2005). Although the important role of these and several other UGTs in plant disease resistance has been shown, their biochemical function in planta remains elusive.
Our results show that the knockdown or mutation of BRT1 attenuates post-invasion mesophyll resistance to P. pachyrhizi in pen2 (Figs 3, 4). The higher haustoria frequency in mesophyll cells of pen2 brt1 when compared with pen2 pad4 sag101 (Fig. 4) may be a result of the complete loss of BRT1 expression in pen2 brt1, whereas the BRT1 gene is expressed at low basal levels in the pen2 pad4 sag101 triple mutant (Fig. 2). As the knockdown or mutation of BRT1 attenuates post-invasion mesophyll resistance to P. pachyrhizi in pen2 (Figs 3, 4), BRT1 seems to act in post-invasive NHR either before or during fungal haustoria formation in mesophyll cells. As the proliferation of P. pachyrhizi hyphae also ceases after haustorium formation in a given mesophyll cell, defense components other than BRT1 and/or a general lack of susceptibility inhibit further fungal development. Because hypersensitive cell death was absent in the P. pachyrhizi-infected pen2 pad4 sag101 mutant, callose encasement of haustoria might suffice to block further proliferation of the fungus. Hypersensitive cell death might serve as a further, perhaps more efficient, defense response (Heath, 2000; Lipka et al., 2005; Wen et al., 2011). Future studies will reveal whether or not the simultaneous knockdown of BRT1 and other candidate genes involved in either callose formation or yet to be identified processes for post-invasive mesophyll resistance will turn Arabidopsis into a host plant for P. pachyrhizi.
The mode of action of BRT1 in preventing haustoria formation in the pen2 mutant remains unclear. The prevention of haustoria development in this mutant does not seem to be caused by inappropriate localization of polyketides in the trichomeless pen2 brt1 double mutant. This conclusion is based on the finding that post-penetration mesophyll resistance is similar in pen2 brt1 and pen2 brt1 gl1. However, it is unclear whether the accumulation of polyketides and/or a general redirection in phenylpropanoid metabolite flow in brt1 or BRT1 dsRNAi lines attenuates mesophyll resistance in these genotypes. Indeed, previous reports have indicated a possible role for fungal polyketides as modulators of plant defense (Böhnert et al., 2004). In a similar manner, the accumulation of plant polyketides in brt1 might attenuate Arabidopsis post-invasive mesophyll resistance to P. pachyrhizi. However, plant polyketides that accumulate as a result of mutated BRT1 have been shown to localize to trichomes (Sinlapadech et al., 2007). Thus, in mutants with trichomes, such as the pen2 brt1 double mutant, polyketides are spatially separated from the invading fungus. Therefore, the accumulation of polyketides in trichomes of brt1 most probably does not impair post-invasive NHR in the brt1 mesophyll.
Reduced sinapoylmalate levels in brt1 are another possible cause for the reduction in post-invasive mesophyll resistance to P. pachyrhizi in pen2 brt1. However, previous reports with the fah1 mutant, which is deficient in sinapic acid and its derivatives, including sinapoylmalate (Chapple et al., 1992), excluded this compound from being involved in resistance to the bacterial plant pathogen P. syringae (Hagemeier et al., 2001). The pen2 fah1 double mutant with reduced sinapoylmalate levels was also not impaired in post-invasive mesophyll resistance to P. pachyrhizi (Fig. 4). Thus, reduced sinapoylmalate levels in pen2 brt1 and pen2 BRT1 dsRNAi plants do not appear to be responsible for the observed impairment of post-invasive mesophyll resistance in these genotypes. However, we cannot exclude the possibility that the enhanced susceptibility of pen2 brt1 results from enhanced accumulation of certain sinapate derivatives or from the presence of artificial metabolites caused by brt1 mutation and absent from wild-type plants. In addition, although the infection phenotype of pen2 brt1 gl1 plants indicates that the accumulation of fluorescent polyketides in trichomes is not of significance, it needs to be considered that Sinlapadech et al. (2007) focused their polyketide analysis on trichomes. The authors did not report any experiment that would exclude the presence of these compounds in epidermal and/or mesophyll cells of brt1 leaves. In this same respect, using LC-MS analysis, Meißner et al. (2008) demonstrated that seeds of ugt84A2/brt1 knockout plants accumulate several metabolites in larger amounts than seeds of wild-type plants. These results demonstrate that changes in the metabolome of brt1 plants can be highly complex. Thus, it cannot be excluded that one or more metabolites are present in larger amounts in brt1 and negatively affect NHR.
As suggested by Hemm et al. (2003), putative regulatory links between sinapate metabolism and the biosynthesis of glucosinolates exist. Likewise, we cannot exclude the influence of the BRT1 mutation on glucosinolate metabolism, which could potentially result in a modified pathogen response in the mutant.
Reduced mesophyll resistance in the pen2 brt1 double mutant does not seem to be caused by cell wall alterations in this mutant. This is because brt1 does not show changes in either the content or composition of lignin (Ruegger & Chapple, 2001). Only cell wall-bound, but not free sinapic acid was reduced slightly in brt1 (Sinlapadech et al., 2007). Whether this reduction in free sinapic acid levels contributes to the interaction phenotype of the brt1 mutant to P. pachyrhizi currently remains unclear.
Yet another possible cause for the attenuated post-invasive mesophyll resistance to P. pachyrhizi in the brt1 mutant is the potential role of BRT1 in the detoxification of a hypothetical fungal compound. The xenobiotic might be secreted during fungal colonization of the plant. Such a scenario was shown for Arabidopsis UGT73C5, which glucosylates and thus detoxifies deoxynivalenol (Poppenberger et al., 2003), a fungal virulence factor (Desjardins et al., 1996; Bai et al., 2001; Lemmens et al., 2005). In the Arabidopsis–P. pachyrhizi interaction, such a hypothetical xenobiotic could attenuate plant defense, thus supporting fungal haustorium development. Although BRT1 has high catalytic specificity for sinapic acid (Lim et al., 2001), it also accepts the phytotoxic xenobiotic 2,4,5-trichlorophenol (TCP) as a substrate, at least in vitro (Messner et al., 2003). Thus, BRT1 might similarly detoxify an as yet unknown P. pachyrhizi compound by glucosylation. It will be challenging to identify the hypothetical xenobiotic in future work.
UDP-glucosyltransferase BRT1 (UGT84A2) of the phenylpropanoid pathway was identified as a novel, essential and specific component of Arabidopsis post-invasion mesophyll resistance to P. pachyrhizi. Although full susceptibility has not been established in the interaction of P. pachyrhizi with the Arabidopsis pen2 brt1 double mutant and pen2 BRT1 dsRNAi lines, the frequency of haustoria formation is strongly enhanced. This is independent of the sinapoylmalate levels in infected leaves and does not seem to be caused by alterations in cell wall composition. We assume that BRT1 might help to inactivate, via glucosylation, an as yet unknown xenobiotic of P. pachyrhizi origin. Further studies will disclose the mechanism by which BRT1 mediates post-invasion mesophyll resistance to P. pachyrhizi. In addition, transgenic soybean plants overexpressing Arabidopsis BRT1 will help to determine whether BRT1 can be used to provide P. pachyrhizi-resistant soybean plants.
We thank Clint Chapple and Volker Lipka for providing Arabidopsis mutants and Bekir Ülker for providing the pJawohl8 vector. We are grateful to Renate Schubert for taking care of the soybean plants and P. pachyrhizi and to Holger Schultheiss and Marco Loehrer for valuable discussions. We appreciate the valuable comments by the reviewers. This work was supported by the Excellence Initiative of the German Federal and State Governments and by BASF Plant Science Company GmbH.