Members of the germin-like protein family in Brassica napus are candidates for the initiation of an oxidative burst that impedes pathogenesis of Sclerotinia sclerotiorum

Abstract

Germin-like proteins (GLPs) are defined by their sequence homology to germins from barley and are present ubiquitously in plants. Analyses of corresponding genes have revealed diverse functions of GLPs in plant development and biotic and abiotic stresses. This study describes the identification of a family of 14 germin-like genes from Brassica napus (BnGLP) designated BnGLP1–BnGLP14 and investigated potential functions of BnGLPs in plant defense against the necrotrophic fungus Sclerotinia sclerotiorum. Sequence alignment and phylogenetic analyses classify the 14 BnGLPs into four groups, which were clearly distinguished from known germin oxalic acid oxidases. Transcriptional responses of the BnGLP genes to S. sclerotiorum infection was determined by comparing cultivars of susceptible B. napus ‘Falcon’ and partially resistant B. napus ‘Zhongshuang 9’. Of the 14 BnGLP genes tested, BnGLP3 was transcriptionally upregulated in both B. napus cultivars at 6h after S. sclerotiorum infection, while upregulation of BnGLP12 was restricted to resistant B. napus ‘Zhongshuang 9’. Biochemical analysis of five representative BnGLP members identified a H₂O₂-generating superoxide dismutase activity only for higher molecular weight complexes of BnGLP3 and BnGLP12. By analogy, H₂O₂ formation at infected leaf sites increased after 6h, with even higher H₂O₂ production in B. napus ‘Zhongshuang 9’ compared with B. napus ‘Falcon’. Conversely, exogenous application of H₂O₂ significantly reduced the susceptibility of B. napus ‘Falcon’. These data suggest that early induction of BnGLP3 and BnGLP12 participates in an oxidative burst that may play a pivotal role in defence of B. napus against S. sclerotiorum.

Introduction

Proteins with sequence homology to germins from wheat and barley have been identified in mosses and mono- and dicotyledonous plants and thus have been named germin-like proteins (GLPs). Initially, germins were found to accumulate in germinating wheat embryos (Lane et al., 1992), and later Lane et al. (1993) demonstrated that this germin degraded oxalic acid to H₂O₂ and CO₂ by an oxalate oxidase (OXO) activity. So far, only germins from barley have been proven to share this OXO activity (Dumas et al., 1993; Lane et al., 1993; Whittaker and Whittaker, 2002), and this distinguishes these proteins from GLPs for which no OXO activity has yet been found. Instead, some GLPs have been shown to possess superoxide dismutase activity, which converts superoxide to H₂O₂ and O₂ (Christensen et al., 2004; Gucciardo et al., 2007), while others remain elusive in terms of enzymatic activity, or function as an auxin receptor (Woo et al., 2002; Yin et al., 2009), reflecting the high functional diversity among GLPs.

Genome and transcriptome analysis of rice (Manosalva et al., 2009), barley (Zimmermann et al., 2006), wheat (Schweizer et al., 1999), maize (Breen and Bellgard, 2010), Physcomitrella patens (Nakata et al., 2004), Arabidopsis thaliana (Carter et al., 1998), and peanut and soybean (Chen et al., 2011) have revealed that GLPs are encoded by gene families with multiple gene members. For example, the A. thaliana genome contains 32 sequences annotated as ‘germin-like’ genes (www.uniprot.org). Expression of germin-like genes is not restricted to germinating seeds, as initially ascribed to wheat germin, but is found in leaves (Membré et al., 2000; Fan et al., 2005; Banerjee and Maiti, 2010), stems (Minic et al., 2009; Banerjee and Maiti, 2010), flowers (Fernández et al., 2003; Yang et al., 2006) and roots (Zimmermann et al., 2006; Gucciardo et al., 2007). The exact function of these proteins during plant development is unclear, but their apoplastic localization in combination with H₂O₂ generating superoxide dismutase (SOD) activity offers a role in cell-wall fortification through the cross-linkage of proteins and carbohydrates (Schopfer, 1996; Barceló, 1998; Banerjee and Maiti, 2010). Additionally, some GLPs bind to the plant hormone auxin and mediate auxin-induced physiological responses during plant development (Inohara et al., 1989; Robert et al., 2010; Effendi et al., 2011). Gene expression analysis in different species has revealed that many germin-like genes are regulated following abiotic and biotic stresses. Ke et al. (2009) identified a GLP among ten drought-induced proteins in rice, and GLPs from wheat, barley and Barbula unguiculata are regulated under salt or metal stress conditions (Hurkman et al., 1991; Berna and Bernier, 1999; Nakata et al., 2002; Caliskan, 2009).

Pathogen infection is one of the major triggers inducing germin-like gene expression. Corresponding sequences have been identified after pathogen challenge in barley (Wei et al., 1998; Hückelhoven et al., 2001), wheat (Berna and Bernier, 1999), rice (Manosalva et al., 2009; Banerjee and Maiti, 2010), pepper (Park et al., 2004), A. thaliana (Collins et al., 2010), sugar beet (Knecht et al., 2010), and grape (Ficke et al., 2004; Godfrey et al., 2007). Accordingly, germin and GLP activities significantly contribute to plant defence reactions against different pathogens. Heterologous expression of barley or wheat germin was found to lead to increased resistance against Sclerotinia sp. fungus in rape (Dong et al., 2008), peanut (Livingstone, 2005), sunflower (Hu, 2003), and tomato (Walz et al., 2008), as well as in transgenic poplar leaves against Septoria musiva (Liang et al., 2001). However, OXO activity is not necessarily essential for the defence function. Expression of mutated germin gf-2.8 and germin-like TaGLP2a, which both lack OXO activity, from wheat, conferred like native gf-2.8 increased resistance in wheat leaves against Blumeria graminis (Schweizer et al., 1999). Knecht et al. (2010) enhanced the resistance in A. thaliana plants against the fungal pathogens Verticillium longisporum and Rhizoctonia solani through transgenic expression of germin-like BvGLP-1 from sugar beet, and silencing of GLP genes in rice intensified the development of fungal rice blast and sheath blight diseases (Manosalva et al., 2009). Furthermore, tobacco plants silenced for a germin-like gene (NaGLP) showed increased susceptibility against two insect herbivores (Lou and Baldwin, 2006), indicating a basal function of GLPs beyond microbial pathogens.

This study reports on the identification and characterization of germin-like genes in the rapeseed (Brassica napus) genome and evaluated their potential in defence against the fungal pathogen S. sclerotiorum, the causal agent of Sclerotinia stem rot disease. S. sclerotiorum is a necrotrophic pathogen that thrives on more than 400 plant species (Bolton et al., 2006) and poses a considerable threat to rape farming. Efforts to increase rape resistance against S. sclerotiorum to the generation of varieties with increased tolerance (Wang et al., 2004; Liu et al., 2005;Li et al., 2009) and intensive research is ongoing. In the available sequence databases, we identified a germin-like gene family in B. napus composed of 14 members designated BnGLP1–BnGLP14. To identify BnGLP members with a putative function in defence against S. sclerotiorum, we measured the corresponding transcripts by quantitative real-time PCR (qPCR) in rape after infection and compared the susceptible B. napus ‘Falcon’ variety with the more tolerant B. napus ‘Zhongshuang 9’ cultivar. Biochemical characterization of five selected BnGLP proteins revealed a H₂O₂-generating SOD activity for two members, while none showed OXO activity. by analogy, S. sclerotiorum evoked an oxidative burst in the plant at 6h after infection, and H₂O₂ itself was shown to augment plant tolerance against S. sclerotiorum. Taken together, these results describe for the first time the family of germin-like genes in B. napus and demonstrate that early upregulation of SOD-active BnGLPs correlates with H₂O₂ formation in plants and may play a pivotal role in resistance of B. napus to S. sclerotiorum infection.

Materials and methods

Plant material and growth conditions

The commercial rape variety B. napus ‘Falcon’ was provided by the Norddeutsche Pflanzenzucht Hans-Georg Lembke KG (Hohenlieth, Germany) and B. napus ‘Zhongshuang 9’ was obtained from Professor Wang Hanzhong (Oil Crops Research Institute, Wuhan, China). Plants of A. thaliana and B. napus were cultivated in growth cabinets with 10h light d^–1 at 300 µmoles m^–2 s^–1 and the temperature set to 22 and 20 °C during day and night periods, respectively.

Infection with S. sclerotiorum

The S. sclerotiorum isolate used throughout this work was obtained from Professor W. Qian (Mei et al., 2011). S. sclerotiorum mycelium was grown on potato dextrose agar (PDA) plates (26.5g l^–1 potato dextrose, 15g l^–1 agar, pH 5.6) for 2 d at 22 °C. Infection of fully expanded leaves from 5–6-week-old plants was performed as described by Zhao and Meng (2003). In brief, scissored leaves were arranged in moistened trays and inoculated with S. sclerotiorum-grown PDA plugs cut with a 0.6cm cork borer from the margin of the expanding mycelium. Trays were sealed with plastic foil and incubated at 22 °C. Developing lesions were measured with a caliper at the indicated time points. To measure lesion formation following water or H₂O₂ infiltration, leaves were not cut from the plant but were wrapped with plastic foil after inoculation to prevent drying of the agar plugs. Infiltration of ~20 µl solution was applied from the bottom side of the leaf at the site of plug infection. In the case of plant sampling for gene transcript analysis, leaves were also not cut from the plant.

Infection of A. thaliana was performed with S. sclerotiorum grown in 70ml liquid Czapek Dox medium (33.4g l^–1 Czapek Dox, 2g l^–1 yeast extract, 2g l^–1 malt extract, pH 5.5) for 2 d at 22 °C. The mycelium was homogenized for 3 s using an Ultra-Turrax, centrifuged at 6000 g for 10min, and the sedimented mycelium resuspended in 10mM MgCl₂ to a concentration of 0.7g in 50ml. A. thaliana plants were grown in 9×9cm pots filled with peat soil and four plants per pot. Fully expanded leaves of 5–6-week-old soil-grown plants were infected with 30 µl drops of the mycelium suspension and covered with a plastic lid to maintain a high humidity. At 3 d after infection, lesion development was divided in three classes: 1, necrotic spot, poor fungal expansion; 2, round wet lesion; and 3, macerated leaf. Based on each class, a disease index was calculated with the formula: (0.5×class 1/total number of drops+class 2/total number of drops+2×class 3/total number of drops)×100. Two independent experiments were performed to validate the results.

For S. sclerotiorum growth inhibition assays, 20 µl of H₂O₂ solution was applied to Whatman paper discs (0.8cm diameter) in the centre of a PDA plate before inoculation and incubated for 2 d.

RNA extraction and qPCR

Plant samples were excised with a 0.8cm cork borer at the sites ofS. sclerotiorum agar plug infection and immediately frozen in liquid nitrogen. Agar plugs without mycelium were used as controls. Total RNA was extracted with TRIzol Reagent (Invitrogen; www.invitrogen.com/) from ground tissue of three leaf discs per sample and three samples were processed for each treatment. Reverse transcription of mRNA into cDNA was performed from 2 µg of total RNA using SuperScriptIII (Invitrogen) following the manufacturer’s instructions. qPCR was conducted with a 7300 Real Time PCR System (Applied Biosystems; www.appliedbiosystems.com) using MAXIMA^®SYBR Green Master Mix (Fermentas; www.fermentas.de) for gene amplification. The primer combinations to amplify BnGLP1–BnGLP 14 and β-tubulin as the reference gene are given in Supplementary Table S1 at JXB online. All PCR products were sequenced to verify specificity for the respective gene, and invariant expression of the β-tubulin reference gene in response to S. sclerotiorum infection was assured beforehand. Data were analysed using 7300 System SDS software by the comparative ΔΔC _t method (Supplementary Fig. S1 at JXB online).

Gene cloning and transient expression in Nicotiana benthamiana

Full-length sequences of BnGLP3, BnGLP7, BnGLP8, BnGLP10 and BnGLP12 were amplified from genomic DNA of B. napus ‘Zhongshuang 9’ using Pfu polymerase (Fermentas). As control genes, GFP was amplified from pGWB5 (Nakagawa et al., 2007) and wheat germin gf-2.8 template DNA was kindly provided by Dr Bornemann (Gucciardo et al., 2007). The respective primers contained attB sites at their 5' ends to allow cloning of the PCR products via the Gateway^®BP recombination reaction (Invitrogen) into the pDONR201 Entry vector. The primer sequences are listed in Supplementary Table S2 at JXB online. For each BnGLP gene, at least six independent clones were sequenced to confirm consistent sequence identity. In the cases of BnGLP3 and BnGLP12, two copies of these family members were amplified with nucleotide polymorphisms. Subsequently, genes were transferred into the Gateway compatible binary pGWB414 vector (Nakagawa et al., 2007) in frame with a C-terminal triple haemagglutinin (HA)-tag coding sequence. For transient plant transformation, the binary vector constructs were transformed into Agrobacterium tumefaciens strain GV3101::pMP90RK and infiltrated with a 1ml syringe without a needle into fully expanded leaves of N. benthamiana, as described by Witte et al. (2004). Tissue for protein extraction was harvested at 6–8 d post-infiltration and frozen at –80°C until further processing.

Generation of transgenic A. thaliana

Transformation of A. thaliana ‘Columbia-0’ was conducted according to Clough and Bent (1998). Transgenic plants (T₁) were selected on half-strength Murashige and Skoog medium (Duchefa; www.duchefa.com) supplemented with 50g l^–1 kanamycin and later transferred to soil for seed setting. Infection experiments were conducted with T₂ plants selected for kanamycin resistance as above.

Enzyme analysis and immunodetection of proteins

Unless otherwise noted, protein extraction from ~10mg of ground tissue was performed in 110 µl extraction buffer (20mM Tris/HCl, pH 7.5, 0.5% SDS, 5mM DTT), except for BnGLP3.1 and BnGLP3.2, which were extracted with 20mM Tris/HCl (pH 7.5), 2% SDS, 50mM DTT. Subsequently, protein samples were adjusted to 2% SDS, 50mM DTT, 10% glycerol and 0.01% bromophenol blue and loaded on an SDS-polyacrylamide gel (10% acrylamide) without boiling (semi-native) for separation. An in-gel SOD activity assay was performed following the protocol of Beauchamp and Fridovich (1971). OXO activity was tested after transferring the proteins to nitrocellulose membrane (Roth; www.carlroth.com) as described by Lane et al. (1993). For immunodetection, proteins were blotted on a PVDF membrane (Roche; www.roche.de) and visualized with anti-HA antibody (Roche) in combination with a Lumi-Light^PLUS Western Blotting Kit (Roche) following the manufacturer’s instructions.

Measurement of H₂O₂ in rape leaves

Measurement of H₂O₂ from leaf tissue was performed using the FOX reagent as described by Cheeseman (2006) with slight modifications. B. napus leaf discs were collected with a 0.6cm cork borer at sites of S. sclerotiorum infection or from control treatment plants and immediately ground in liquid nitrogen. To extract the H₂O₂, 400 µl of 25mM HCl was added and the sample thawed on ice with shaking. After centrifugation for 5min at 17 000 g at 4 °C, 100 µl of the supernatant was added to 900 µl of FOX reagent [250 µM ammonium iron (II) sulfate, 100 µM sorbitol, 100 µM xylenol orange, 25mM H₂SO₄, 1% ethanol] and incubated for 15min in the dark. Complex formation of Fe^III+ with xylenol orange in the presence of H₂O₂ was measured with a photometer at 560nm.

Results

The germin-like gene family in B. napus

To identify germin-like genes in B. napus (BnGLP), we searched genomic and expressed sequence tag (EST) databases from the NCBI, TGI (http://compbio.dfci.harvard.edu/tgi/plant.html) and the Shanghai Rapeseed database (http://rapeseed.plantsignal.cn) with an original germin sequence from barley using tblastn. ESTs provide an alternative source for gene identification in plants whose genome sequences are not fully available (Rudd, 2003), as is the case for B. napus. Gene candidates from genomic database were verified by EST fragments to exclude non-transcribed pseudo-genes. Putative full-length sequences matching an E-value of a maximum of 10^–3 were selected and sequences were only considered that contained the two conserved histidines essential for binding the manganese co-factor (Fig. 1A; Woo et al., 2000), giving 307 candidate sequences in total. The average sequence length of the BnGLPs was around 220 aa, matching the length of the original germins including the signal peptide, and amino acid identity ranged from 30 to 48% between the BnGLPs and HvOXO2 germin from barley. Polyploidity of B. napus and independent sequence donations from various B. napus accessions to the databases can result in redundancy of gene family members. We thus performed a protein sequence alignment and subsequent phylogenetic analysis (http://bioweb2.pasteur.fr; Felsenstein, 1989; Thompson et al., 1994) to distinguish classes of highly similar sequences. For each class, a representative BnGLP sequence was chosen that showed no more than 65% amino acid identity to any protein of the other aligned clusters. This produced a GLP family in B. napus comprising 14 members, which could be clustered into four groups (Fig. 1A, 1B). Comparison of the BnGLP family with the proven germin OXOs TaGER2, TaGER3, HvOXO1 and HvOXO2 did not reveal high homology and thus these were separated from the germins in the phylogenetic analysis (Fig. 1B).

Fig. 1.

Primary structure analysis of GLPs from B. napus (BnGLP). (A) Partial amino acid alignment of the 14 BnGLP members with HvOXO2 germin from barley. Asterisks denote conserved amino acids involved in manganese ion binding of germins (Woo et al., 2000). (B) Phylogenetic tree of GLPs from B. napus and germins from Hordeum vulgare (HvOXO1 and HvOXO2) and Triticum aestivum (TaGER2 and TaGER3). Analysis was conducted using MEGA5 software (Tamura et al., 2011) with the neighbour-joining method. The sum of branch length is 4.9. Numbers next to the branches indicate percentage of replicate trees in which the sequences clustered together in a bootstrap test (100 replicates). Bar, 0.1 amino acid substitutions per site.

Transcript profiling of BnGLP genes in response toS. sclerotiorum infection

B. napus is a highly susceptible host for S. sclerotiorum and only a few varieties exist with partial resistance, such as B. napus ‘Zhongshuang 9’ derived from native Chinese cultivars (Wang et al., 2004). In order to find two B. napus cultivars with differential susceptibility to Sclerotinia disease, we compared B. napus ‘Zhongshuang 9’ with the commercial cultivar B. napus ‘Falcon’ that was previously shown to have a high susceptibility against V. longisporum (Eynck et al., 2009). Detached leaves of both varieties were infected with S. sclerotiorum grown on PDA and incubated at 20–22 °C. Around the infection sites, circular lesions developed that were twice the size on B. napus ‘Falcon’ leaves at 36 and 48h after infection compared with those on B. napus ‘Zhongshuang 9’ leaves (Fig. 2A), revealing significant differences in susceptibility to S. sclerotiorum infection. To verify the expression of BnGLP genes in leaf tissue, we first extracted RNA from non-treated leaves and measured the transcript abundance by qPCR using primers specifically amplifying fragments of each of the 14 BnGLP genes (Fig. 2B). Gene transcripts were detected for all BnGLP genes, although the relative amounts varied by three orders of magnitudes, with BnGLP6 and BnGLP10 being the least and most abundant transcripts, respectively. Significant differences in BnGLP gene expression between B. napus ‘Falcon’ and ‘Zhongshuang 9’ were not detected. Furthermore, we used these cultivars to test whether members of the BnGLP family were transcriptionally regulated in response to infection with S. sclerotiorum and whether differences in regulation existed between the susceptible B. napus ‘Falcon’ and the partially resistant B. napus ‘Zhongshuang 9’ lines. We chose a 6h time point for transcript analysis as an early stage of infection, assuming that molecular effects decisive for the success of plant defence appear in the beginning and to limit secondary effects derived from massive cell degradation at later time points. In the susceptible B. napus ‘Falcon’ background, BnGLP3 was upregulated at 6h post-infection, while in B. napus ‘Zhongshuang 9’, in addition to BnGLP3, BnGLP12 also became transcriptionally induced. None of the other BnGLP genes showed a significant response in gene expression, although BnGLP8 and BnGLP10 showed a tendency in multiple experiments to become suppressed in B. napus ‘Zhongshuang 9’ following S. sclerotiorum infection. Together, these data demonstrated that B. napus ‘Zhongshuang 9’ responded to the S. sclerotiorum infection differently in the regulation of BnGLP genes compared with B. napus ‘Falcon’.

Fig. 2.

B. napus infection with S. sclerotiorum and transcript analysis of BnGLP genes. (A) Infection of B. napus ‘Falcon’and ‘Zhongshuang 9’ with S. sclerotiorum in a detached-leaf assay. Lesion sizes were measured at the indicated hours post-infection (hpi) and results are shown as means ± standard error (SE) (n=8). (B) Transcript abundance of the 14 BnGLP genes relative to that of β-tubulin in leaves of B. napus ‘Falcon’ and ‘Zhongshuang 9’. Transcripts were measured by qPCR and values calculated following the equation 2^CtTubulin/2^CtBnGLP×10⁵. Results are shown as means ±SE (n=3) and independent experiments showed the same trend.(C) qPCR analysis of the 14 BnGLP genes relative to mock treatment at 6h after infection with S. sclerotiorum.

Cloning and biochemical characterization of BnGLP3, BnGLP7, BnGLP8, BnGLP10 and BnGLP12

As GLPs from B. napus have not been characterized to date, we cloned five representative members of the BnGLP gene family with regard to their transcriptional regulation after S. sclerotiorum infection (Fig. 2B). These comprised the inducible BnGLP3 and BnGLP12 and the non-induced BnGLP8, BnGLP10 and BnGLP7 genes. The full-length open reading frames were cloned behind a 35S promoter and fused to the coding sequence of a triple HA tag (3×HA) to allow later immunodetection of the expressed proteins. Six to ten clones of each gene were sequenced to verify sequence consistency and to identify gene copies in the Brassica genome. While BnGLP7, BnGLP8 and BnGLP10 were cloned as single genes, amplification of BnGLP3 and BnGLP12 each yielded two homologues coding for proteins with 95% identities and were named BnGLP3.1 and BnGLP3.2, and BnGLP12.1 and BnGLP12.2, respectively. The homologous genes were included in the further biochemical analysis to test for alterations resulting from the amino acid differences. Additionally, we generated 3×HA-tagged fusion proteins of wheat germin gf-2.8 (Lane et al., 1993) and GFP to serve as positive and negative controls, respectively. The recombinant proteins were transiently expressed in N. benthamiana, and total protein extracts were separated by semi-native SDS-PAGE omitting a reducing agent in the loading buffer and without boiling before loading the sample (Fig. 3A). The calculated molecular weight of monomeric germin and all BnGLPs fused to the 3×HA tag was ~29 kDs and that of GFP:HA was ~32.8kDa, matching the band sizes seen at the bottom of the HA-specific immunoblot. The larger proteins between 35 to 40kDa and around 170kDa were not observed under full denaturating conditions (data not shown) and thus probably reflect different oligomerizations of the proteins. The active gf-2.8 enzyme is a hexameric complex and has been shown to migrate at ~125kDa by semi-native SDS-PAGE (Lane et al., 1993; Walz et al., 2008). In Fig. 3A, the majority of gf-2.8:HA protein migrated at between 25 and 40kDa and a minor amount also formed a complex at ~170kDa. Taking the mass of the 3×HA tag into account, the latter probably represents the hexameric gf-2.8:HA complex. By analogy, of the seven investigated BnGLPs, both homologues of BnGLP3:HA and BnGLP12:HA, and BnGLP8:HA migrated as monomers and also formed higher-molecular-weight complexes. In contrast, BnGLP7:HA and BnGLP10:HA were expressed as monomeric proteins only and did not form higher-molecular-weight complexes.

Fig. 3.

Biochemical characterization of the fusion proteins BnGLP3.1:HA, BnGLP3.2:HA, BnGLP7:HA, BnGLP8:HA, BnGLP10:HA, BnGLP12.1:HA, BnGLP12.2:HA, gf-2.8:HA and GFP:HA transiently expressed in N. benthamiana. (A) Immunodetection of recombinant proteins using HA-specific antibody in total protein extracts. Samples were loaded without DTT in the loading buffer and without prior boiling (semi-native).M indicates the monomer. (B) Protein extracts separated as in(A) were blotted on nitrocellulose membrane and assayed for OXO activity. (C) Protein extracts separated as in (A) were assayed for SOD activity.

Originally, germins from wheat and barley were found to oxidize oxalic acid to CO₂ and H₂O₂ (Dumas et al., 1993; Lane et al., 1993), defining them as true germins. This OXO activity has so far not been shown for any of the GLPs present in mono- and dicotyledonous plant species and mosses. Instead, some GLPs are SODs, reducing superoxide to H₂O₂, while the enzyme activity of other GLPs remains elusive (Yamahara et al., 1999; Christensen et al., 2004; Nakata et al., 2004; Zimmermann et al., 2006; Gucciardo et al., 2007; Banerjee and Maiti, 2010). We thus tested for OXO and SOD activities of the transient expressed proteins (Fig. 3B, 3C). In the OXO activity assay, the gf-2.8:HA protein gave a clear signal at the size of the higher-molecular-weight complex of ~170kDa, confirming the OXO activity for recombinant gf-2.8:HA protein (Fig. 3B). Under the same conditions, none of the BnGLPs or GFP:HA displayed an OXO activity at any protein complex size. However, when we tested for SOD activity, the protein complexes of BnGLP3:HA and BnGLP12:HA and their respective homologues showed a clear activity at 170kDa or higher (Fig. 3C), but none of the other tested BnGLPs or gf-2.8:HA and GFP:HA appeared to possess SOD activity under the experimental conditions used. Notably, despite higher-molecular-weight complex formation, BnGLP8:HA did not show any OXO or SOD activity. We also expressed BnGLP10 as native protein to test for potential interference of the 3×HA tag with enzyme function but again did not observe SOD activity (data not shown).

In order to evaluate whether BnGLP3 and BnGLP12 represent SODs with redundant functions in the plant, we investigated their solubility under different extraction conditions (Fig. 4). Both proteins exhibited maximum solubility when extracted with 50mM Tris/HCl (pH 7.5), 2% SDS and 50mM DTT. In contrast, 50mM Tris/HCl (pH 7.5) without detergent and reducing agent extracted BnGLP12.1:HA but not BnGLP3.1:HA, as shown by anti-HA immunodetection and SOD activity assay (Fig. 4, lane 1). The addition of either 50mM DTT or 2% SDS (Fig. 4, lanes 2 and 3) slightly increased the solubility of both proteins, although the majority of BnGLP3.1:HA was extracted with the combination of both. Together, these results are in line with earlier findings that disproved an OXO activity for GLPs, and instead we have demonstrated SOD activity for two members of five tested GLPs from B. napus. Interestingly, it was BnGLP3 and BnGLP12, which were transcriptionally induced in response to S. sclerotiorum infection, that displayed SOD activity.

Fig. 4.

Total proteins from leaves of N. benthamiana expressing BnGLP3.1:HA, BnGLP12.1:HA or no (N.b.) recombinant protein. Extraction was performed as follows: lane 1, 50mM Tris/HCl(pH 7.5); lane 2, 50mM Tris/HCl (pH 7.5), 50mM DTT; lane 3, 50mM Tris/HCl (pH 7.5), 2% SDS; lane 4, 50mM Tris/HCl (pH 7.5), 50mM DTT, 2% SDS. Prior to SDS-PAGE loading, all samples were adjusted to the same concentration of DTT and SDS but not boiled. (A) Immunodetection of transblotted proteins with HA-specific antibody. (B) In-gel SOD activity assay.

Impact of H₂O₂ on defence of B. napus againstS. sclerotiorum

The correlation between BnGLP3 and BnGLP12 gene transcript activation in response to S. sclerotiorum infection and the SOD activity of the corresponding proteins prompted us to further investigate the role of H₂O₂ as one reaction product of SOD activity during the defence response of B. napus to S. sclerotiorum infection. The BnGLP transcript analyses shown in Fig. 2C revealed gene induction of SOD BnGLP3 and BnGLP12 at 6h post-infection. We thus measured H₂O₂ production in the leaves of B. napus ‘Falcon’ and B. napus ‘Zhongshuang 9’ at 6h after infection with S. sclerotiorum infection. Figure 5 shows the change in H₂O₂ formation relative to the control treatment. Both B. napus varieties generated more H₂O₂ at 6h post-infection, but the increase in the partially resistant B. napus ‘Zhongshuang 9’ variety was significantly higher compared with B. napus ‘Falcon’. Thus, H₂O₂ production is part of the early plant defence against S. sclerotiorum infection and its magnitude may contribute to the resistance phenotype of B. napus ‘Zhongshuang 9’. To strengthen this idea further, we infiltrated leaves of B. napus ‘Falcon’ and ‘Zhongshuang 9’ from the bottom side with 0.5mM H₂O₂ or water as a control and infected the infiltration sites from the top with S. sclerotiorum grown on PDA plugs. At 29h post-infection, we measured the lesion sizes of the spreading S. sclerotiorum fungus to evaluate whether artificial supply with H₂O₂ could influence the progress of S. sclerotiorum infection (Fig. 6). Lesion sizes were significantly smaller on B. napus ‘Zhongshuang 9’ compared with B. napus ‘Falcon’ when only water was infiltrated, in line with the detached-leaf assay shown in Fig. 2A. In contrast, infiltration of H₂O₂ resulted in significantly smaller lesion formation on B. napus ‘Falcon’, and was approximately equal to that observed on the more tolerant ‘Zhongshuang 9’ genotype. Lesion formation was not significantly different on B. napus ‘Zhongshuang 9’ between H₂O and H₂O₂ infiltration. Thus, an external supply of H₂O₂ can increase the resistance of B. napus ‘Falcon’ to S. sclerotiorum infection and this observation was not caused by a direct toxic effect of H₂O₂ on S. sclerotiorum growth, as H₂O₂ did not induce a further lesion size reduction on B. napus ‘Zhongshuang 9’ and did not affect S. sclerotiorum growth in vitro (data not shown).

Fig. 5.

H₂O₂ formation at 6h after S. sclerotiorum infection on leaves of B. napus ‘Falcon’ and ‘Zhongshuang 9’ relative to control treatment (100%). Results are shown as means ±SE (n=3) and the asterisk indicates a statistically significant difference(*P <0.05, Student’s t-test). One of three independentexperiments with similar results is shown.

Fig. 6.

Infection of B. napus ‘Falcon’ and ‘Zhongshuang 9’ with S. sclerotiorum mycelium plugs at sites infiltrated with water or 0.5mM H₂O₂. Lesion sizes were measured at 31h post-infection. Results are shown as means ±SE (n=8) and asterisks indicate statistically significant differences compared with Falcon water treatment (***P <0.01, Students t-test). One of three independent experiments with similar results is shown.

Transgenic expression of BnGLP12:HA in A. thaliana increases tolerance to S. sclerotiorum

The experiments performed in this study so far strongly suggested that members of the germin-like family in B. napus with SOD activity contribute to reduce the spread of S. sclerotiorum. To confirm a direct link between the upregulation of GLPs and resistance to S. sclerotiorum, we transformed A. thaliana with BnGLP7:HA and BnGLP12:HA representing GLPs without and with SOD activity, respectively. A. thaliana is a host for S. sclerotiorum and is closely related to B. napus. Following infection with S. sclerotiorum, transgenic Arabidopsis expressing BnGLP7:HA did not reveal any difference in susceptibility compared with wild-type Col-0, while two of three plants expressing BnGLP12:HA were significantly (P <0.05) more resistant (Fig. 7).

Fig. 7.

Infection of A. thaliana with S. sclerotiorum. Transgenic plants expressing BnGLP7:HA or BnGLP12:HA under control of a 35S promoter compared with Col-0 Arabidopsis wild-type. Leaves were infected with drops of mycelia solution and evaluated after 3 d, as described in Materials and methods. Numbers denote independent transgenic lines. Results are shown as means ±SE of three pots with four plants each and asterisks indicate a significant difference at P <0.05 compared with Col-0 (Student’s t-test). Two independent experiments were carried out and showed the same trend.

Discussion

GLP family in B. napus

Germin-like proteins are present in many if not all plants and are encoded by gene families in the respective genomes. Here, we have presented for the first time the GLP family in rape(B. napus), which is represented by 14 BnGLPs (Fig. 1A, 1B). As complete sequence information of the B. napus genome is currently not available, our classification of GLPs was based on genomic sequences and ESTs that translate into proteins with a maximum of 65% amino acid identity. In this way, we excluded cultivar-specific variants of one gene, and probably also gene copies present in the B. napus genome due to genome duplication during evolution (Parkin et al., 2003) and the amphidiploid nature of the B. napus genome (Nagaharu, 1935). The BnGLPs possessed all three conserved germin sequence boxes defined by Bernier and Berna (2001) including the PxHxHxxxxE motive (Fig. 1A) essential for OXO activity, but exhibited only 30–48% amino acid sequence identity to the original HvOXO2 germin from wheat and thus are termed ‘germin-like’. Accordingly, all germins with proven OXO activity separated in a phylogenetic tree from the BnGLPs (Fig. 1B), implying that they also exhibit functional divergence. In the same analysis, the 14 BnGLPs clustered into four groups, but it was not clear whether proteins of one group shared common features.

The SODs BnGLP3 and BnGLP12 are induced in response to S. sclerotiorum infection

Gene expression and protein function analyses reported in the literature indicate a bias of germins and GLPs to participate in plant defence responses against pathogens, including the necrotrophic fungus S. sclerotiorum (reviewed by Lane, 2002; Dunwell et al., 2008). B. napus is also a host for S. sclerotiorum, and most commercially available rape varieties are highly susceptible to S. sclerotiorum, such as B. napus ‘Falcon’. In contrast, B. napus ‘Zhongshuang 9’ shows increased resistance against the biotrophic fungus V. longisporum (Eynck et al., 2009) and higher tolerance to S. sclerotiorum infection compared with B. napus ‘Falcon’, as shown in Fig. 2A. To investigate a possible involvement of the BnGLP family in response to S. sclerotiorum infection, we examined the regulation of the corresponding genes. Quantitative measurements of the 14 BnGLP genes in non-treated leaves revealed no differences a priori in transcript abundance between both B. napus varieties (Fig. 2B), indicating conserved gene regulation under normal growth conditions. We extended the analysis to 6h after S. sclerotiorum infection, assuming that early regulated genes would be directly connected to S. sclerotiorum invasion and probably essential for the success of plant defence. In both B. napus varieties, BnGLP3 was upregulated at 6h post-infection, while BnGLP12 transcripts increased in B. napus ‘Zhongshuang 9’ only (Fig. 2C). Similarly, Zhao et al. (2007) investigated gene expression changes in rape with a microarray from A. thaliana and found a germin-like gene to be upregulated 4–4.8-fold in a susceptible and a semi-tolerant variety at the earliest time point of 24h after S. sclerotiorum infection. The corresponding orthologue in B. napus (BnGLP7) was not upregulated under our experimental conditions; this may be due to the different time point chosen or non-specific hybridization of the microarray to other BnGLP members. In another experiment employing a B. napus-specific oligonucleotide chip, BnGLP3 was upregulated 11-fold at 78h after S. sclerotiorum infection in stem tissue of the susceptible B. napus ‘Westar’ but not in the partially resistant ‘Zhongyou 821’ cultivar (Zhao et al., 2009). It should be noted that, in B. napus, BnGLP3 is regulated following S. sclerotiorum infection in both susceptible and tolerant B. napus varieties, supporting a potential role in plant basal resistance, as suggested by Zimmermann et al. (2006) for barley GLPs in the compatible interaction with the powdery mildew fungus Blumeria gramins f. sp. hordei. The fact that BnGLP12 was only upregulated in the resistant ‘Zhongshuang 9’ following S. sclerotiorum infection strongly suggests that it has a role in plant resistance to S. sclerotiorum. Furthermore, we demonstrated that heterologous expression of BnGLP12:HA, but not of the SOD-inactive germin-like protein BnGLP7:HA, in A. thaliana increased tolerance to S. sclerotiorum (Fig. 7). This corroborates our interpretation of increased resistance against S. sclerotiorum through the upregulation of BnGLP12.

Germin proteins from wheat and barley form homohexameric complexes and degrade oxalic acid to H₂O₂ and CO₂. Of the five BnGLPs tested here, both homologues of BnGLP3:HA and BnGLP12:HA as well as BnGLP8:HA migrated as higher-molecular-weight complexes of~170kDa in semi-native SDS-PAGE (Fig. 3A). This is equivalent to six times the monomer mass of ~28kDa, revealing the same complex stoichiometry and SDS stability as observed for germin proteins (Lane et al., 1992; Woo et al., 1998) and GLPs from P. patens (Nakata et al., 2004), barley (Christensen et al., 2004; Zimmermann et al., 2006), A. thaliana (Membré et al., 2000), and pea (Gucciardo et al., 2007). The absence of higher-molecular-weight complexes observed here for BnGLP8:HA and BnGLP10:HA was also shown for AtGER1 (Membré et al., 2000), revealing distinct biochemical properties with respect to either complex formation or complex stability.

The GLPs examined to date do not possess an OXO activity, but some members have been shown to produce H₂O₂ through a SOD activity. The BnGLP proteins investigated here did not show any OXO activity, but higher-molecular-weight complexes of BnGLP3:HA and BnGLP12:HA, including their homologues, were able to dismutate superoxide to H₂O₂ through SOD activity (Fig. 3B, 3C). Despite the complex formation of BnGLP8:HA, no SOD activity could be detected as for monomeric BnGLP7:HA and BnGLP10:HA. Thus, BnGLP7, BnGLP8 and BnGLP10 probably fulfil different functions in the plant compared with BnGLP3 and BnGLP12. We also expressed BnGLP10 as native protein to exclude a negative effect of the HA tag, but were again not able to detect SOD activity (data not shown). Remarkably, of the five BnGLPs tested, those that showed SOD activity were also transcriptionally induced in response to S. sclerotiorum infection. Despite equivalent complex formation and SOD activity of BnGLP3:HA and BnGLP12:HA, BnGLP12:HA was soluble in Tris buffer (pH 7.5) during protein extraction from plant tissue, while BnGLP3 required the addition of a reducing agent and a strong detergent to become fully soluble (Fig. 4). This indicated that BnGLP3 is associated with cellular structures via reducing/oxidizing and hydrophobic interactions and is not redundant with BnGLP12 in the plant. Thus, infection with S. sclerotiorum results in quantitative and qualitative differences in SOD activation between susceptible B. napus ‘Falcon’ and the tolerant B. napus ‘Zhongshuang 9’.

Early formation of H₂O₂ restricts S. sclerotiorum pathogenesis

In line with the induction of SOD-active BnGLPs, we also measured an increase in H₂O₂ in rape leaves at 6h after S. sclerotiorum infection (Fig. 5). Both varieties responded with an increase in H₂O₂, but B. napus ‘Zhongshuang 9’ significantly exceeded the H₂O₂ amounts of the susceptible ‘Falcon’ cultivar, correlating with the additional induction of BnGLP12 in ‘Zhongshuang 9’. This oxidative burst at the early state of S. sclerotiorum-infected rape leaves confirmed results Xu et al. (2009) who also found higher H₂O₂ levels in transgenic rape expressing a glucose oxidase and this plant also had restrained S. sclerotiorum lesion formation compared with a susceptible rape variety. Production of plant-derived H₂O₂ has been reported for compatible and incompatible plant pathogen interactions acting as a direct antimicrobial compound, to trigger signal transduction pathways that occasionally lead to a hypersensitive response or to foster cell-wall fortification (reviewed in Shetty et al., 2008). This functional diversity probably relates to the site of H₂O₂ generation, the timing and the amount of H₂O₂ produced. The BnGLP proteins are predicted to posses a secretion signal (Petersen et al., 2011), and this is in agreement with experimental data that localized GLPs to the cell wall (Irshad et al., 2008; Banerjee et al., 2010; Komatsu et al., 2010). Thus, BnGLP3 and BnGLP12 are likely to participate in the S. sclerotiorum-induced apoplastic formation of H₂O₂ and may act in concert with NADPH oxidases and peroxidases, which are known to execute the apoplastic oxidative burst in response to pathogen stress in different species (Torres, 2010). The target of BnGLP-derived H₂O₂ is unclear, but the work of Banerjee et al. (2010) suggests a role in cell-wall reinforcement. The authors expressed rice germin-like protein1 in transgenic tobacco and correlated its SOD activity with hyper-accumulation of H₂O₂ and enhanced cross-linkage of cell-wall components after infection with Fusarium solani, which led to higher tolerance against this fungal pathogen. Similarly, we observed a positive effect of H₂O₂ on the resistance of B. napus ‘Falcon’ to S. sclerotiorum by the infiltration of 0.5mM H₂O₂ prior to infection (Fig. 6). B. napus ‘Zhongshuang 9’ did not respond with increased S. sclerotiorum resistance, indicating that the naturally stronger induction of H₂O₂ production in B. napus ‘Zhongshuang 9’ in response to S. sclerotiorum infection (Fig. 4) is sufficient for H₂O₂-triggered defences at the applied concentration and may explain the increased resistance to S. sclerotiorum compared with B. napus ‘Falcon’. In A. thaliana, L’Haridon et al. (2011) induced H₂O₂ formation by wounding leaves or exogenously applied H₂O₂ and by this increased the resistance against the necrotrophic pathogen Botrytis cinerea. Moreover, the authors showed that the wound-induced reactive oxygen species formation and resistance against B. cinerea were independent of the NADPH oxidases AtRBOHD and AtRBOHF, substantiating our interpretation of BnGLP proteins as part of an oxidative burst and subsequent increase in rape resistance against necrotrophic S. sclerotiorum.

Taken together, we have established here the family of GLPs in B. napus represented by 14 BnGLP members. Gene expression profiling of this family and biochemical characterization of selected members suggested that the SODs BnGLP3 and BnGLP12 are involved in early rape defence against S. sclerotiorum by the initiation of an oxidative burst. We also showed that H₂O₂, either produced in vivo or applied exogenously, correlated with increased resistance against S. sclerotiorum, providing a link to the functions of BnGLP3 and BnGLP12 in rape defence. Beyond the B. napus/S. sclerotiorum system investigated here, GLPs in different species relate to increased resistance including insects (Lou and Baldwin, 2006; Collins et al., 2010), nematodes (Knecht et al., 2010), microbes (Wei et al., 1998; Schweizer et al., 1999; Hückelhoven et al., 2001; Ficke et al., 2004; Zimmermann et al., 2006; Godfrey et al., 2007; Manosalva et al., 2009; Shetty et al., 2009; Banerjee and Maiti, 2010) and tobacco mosaic virus (Park et al., 2004), indicating basal functions of GLPs in plant resistance. Whether GLPs are also involved in R-protein-mediated or non-host resistance and what the immediate consequences of GLP derived reactive oxygen species formation are remain to be investigated.

Glossary

Abbreviations:

EST

expressed sequence tag

GLP

germin-like protein

OXO

oxalate oxidase

PDA

potato dextrose agar

qPCR

quantitative real-time PCR

SE

standard error

SOD

superoxide dismutase.