Plant experiments were conducted with Nicotiana benthamiana accession “lab” (Bally et al., 2015) in temperature controlled growth rooms as described by Petrie et al. (2010). Plants were grown under a 16-h light/8-h dark cycle at 22°C.

Details of fungal isolates are listed in Table 1 and were maintained as pure cultures with Rhizoctonia solani, Alternaria brassicicola, F. oxysporum, Phytophthora nicotianae and C. lupini isolates grown on 1/2 strength Potato Dextrose Agar (PDA), and S. sclerotiorum on 1/8 strength PDA. Spores, mycelia or sclerotia were inoculated on PDA plates, which were placed at room temperature in the dark until plates were fully covered by the pathogen. Mycelial plugs from these plates were used for subsequent experiments.

β1- and β6-conglutins were overexpressed using the pET28b construct (Novogen)² that contains an N-terminal polyhistidine (6xHis) tag. pUC57 vectors carrying synthesized β1, β2, β3, β4, or β6 conglutin sequences based on Genbank HQ670409 (β1), HQ670410 (β2), HQ670411 (β3), HQ670412 (β4) and HQ670414 (β6) sequences but altered for optimum bacterial codon usage with NcoI/XhoI restriction enzyme linkers were synthesized and constructed by GenScript³ (Supplementary Figure S1). The bacterial expression vectors for β-conglutins were obtained via NcoI/XhoI digestion of respective pUC57-β-conglutin constructs followed by ligation of the β-conglutin fragments into the pET28b vector.

All β-conglutin proteins were expressed in Rosetta^TM 2(DE3) pLysS Singles^TM Competent Cells (Novogen). Protein expression was performed using an auto-induction method (Studier, 2005). Briefly, a single clone containing the expression construct was isolated and grown for 20 h in LB plus kanamycin at 50 μg/mL at 37°C and continuous shaking (200 rpm). The culture was diluted 1:150 in ZYM-5052 medium and grown for a further 5 h until the cell density reached an OD₆₀₀ of 0.7. The cells were then induced to overexpress the proteins by adjusting the temperature to 19°C for another 20 h. Cells were collected by centrifugation at 5000 × g at 4°C. The bacterial cell pellet was rinsed two times with phosphate buffered saline (PBS), pH 7.5, removing the supernatant, then flash frozen in liquid nitrogen and stored at -80°C until further use.

Protein purification from bacterial pellets was performed following the manufacturers’ recommendations for His-tagged proteins (Qiagen)⁴. Briefly, the steps consisted of lysing cells followed by nickel affinity chromatography using Ni-NTA spin columns, and histidine (6xHis) tags at the N-terminal part of the β-conglutin proteins. After elution of 6xHis-tagged proteins from the column with an increasing imidazole concentration gradient (10–300 mM), 2.5 mL fractions were collected. Fractions containing protein were analyzed using SDS-PAGE and fractions showing a single band corresponding to the expected molecular weight were pooled, and dialyzed five times against Tris-HCl 100 mM, pH 7.5, 150 mM NaCl to eliminate the imidazole reagent. The protein was concentrated using a 30 kDa Amicon centrifuge filter (Millipore)⁵. The aliquots were flash-frozen in liquid nitrogen and kept at -80°C until further use. Protein purities were >95% as determined by densitometry analysis of the SDS-PAGE gel image. An aliquot of each protein was used to measure their concentration using Bradford assays (BioRad, Hercules, CA, USA) using bovine serum albumin (BSA) as a standard. The β-conglutins purifications yields ranged between 10–15 mg/mL.

The peptide sequence Nt – VDEGEGNYELVGIR – Ct, was chosen as this region was 100% homologous among all NLL β-conglutins and did not share any significant homology to other known lupin sequences. This peptide was generated by Agrisera⁶ and was used to immunize rabbits and to produce polyclonal antiserum (Agrisera). The rabbit immune serum was affinity-purified against the same synthetic peptide.

SDS-PAGE and Immunoblotting were performed as previously described (Foley et al., 2015).

A disk diffusion method was performed on 90 mm petri dishes containing PDA to test the sensitivity of different fungi strains toward the β-conglutin proteins. Fungal isolates were initially grown on PDA plates as described previously at 21°C until mycelial growth had developed. A mycelial plug was then taken from the growing edge of the colony and placed in the center of a new full-strength PDA plate in which sterile blank paper disks (12.7 mm diameter) containing 800 μg of purified β-conglutin protein (dissolved in BSA buffer) or buffer only control were placed 30 mm away. For in vitro assay of C. lupini, 1 mg of purified β6-conglutin protein was used. The plates were incubated in the dark at 21°C and the zone of fungal inhibition around the disks recorded over 30 days. Assays were performed in triplicate.

Antifungal activity of β-conglutin proteins were expressed as the IC50 (μM) values for the fungi tested in Table 1. The mycelial growth inhibition assays were used to determine the concentration required for 50% growth inhibition (IC50), using a β-conglutin protein concentration range from 5 to 125 μM (5, 10, 15, 20, 25, 35, 50, 75, and 125 μM) for each β-conglutin. Results were expressed as mean ± standard deviation (SD). To determine the statistical significant differences of the β-conglutins antifungal activity on the growth of these fungal pathogens, the data was analyzed using statistical package SPSS 15.0 (SPSS, Inc., Chicago, IL, USA). Significant differences between the mean values of each cohort were determined using Tukey Kramer HSD test (p < 0.05). To characterize the antifungal activity of β-conglutins further, a subsequent experiment with the necrotrophic fungal pathogen of NLL causing anthracnose disease (C. lupini isolates WAC8672 and WAC10444) was conducted, where the antifungal activity of β6-conglutin was determined by placing a mycelial plug in the center of a PDA plate and radial outgrowth determined over 14 days. For each time point significant differences were determined using a student’s t-test using the JMP software v7.0 (SAS Institute).

Four-week-old N. benthamiana plants grown at 22–24°C in culture rooms were used for Agrobacterium tumefaciens-mediated transient expression as described previously (Sparkes et al., 2006). The β1- and β6-conglutin coding sequence were cloned into the vector pMDC83 to generate C-terminal GFP fusion proteins and transformed into A. tumefaciens AGL1. Transformed A. tumefaciens AGL1 were cultured at 28°C until stationary phase (∼24 h), washed and resuspended in infiltration medium (50 mM MES, 0.5% (w/v) glucose, 100 μM acetosyringone (Sigma-Aldrich⁷ pH 5.6). The bacterial suspension was inoculated using a 1-mL syringe without a needle by gentle pressure through a <1 mm hole punched on the lower epidermal surface of the upper leaves of N. benthamiana plants. Following infiltration, plants were incubated under normal growth conditions at 22–24°C. This protocol was used for in vivo fungal growth inhibition assays, oxyblot assays, and subcellular localization studies.

Forty hours after Agrobacterium infiltration of β1- or β6-conglutin protein constructs into N. benthamiana plants, agar plugs containing hyphae of S. sclerotiorum or P. nicotianae were placed on the infiltrated leaf areas (control and β1- or β6-conlgutin overexpression) and plants were incubated in a growth chamber at 18°C for 3 days under a 16 h long-day light regime. Leaves were assessed for visible necrotic disease progression, then detached and further visualized after lactophenol trypan blue staining based on Keogh et al. (1980). Briefly, mature fourth leaves of N. benthamiana containing visible infection sites (tissue necrosis) were cleared with acetic acid: ethanol (1:1 v/v) then stained for 1–2 h in lactophenol (10 mL phenol; 10 mL lactic acid, 10 mL water) with 0.05% (w/v) trypan blue at 60°C. Excess staining was removed with lactic acid: water (1:1 v/v) until leaves were clear. Leaves were examined by light microscopy on a Nikon N400-M light microscope (Nikon, Tokyo, Japan). Control areas were infiltrated with GFP-only vectors. The experiment was repeated three times. In each experiment, leaves from 4–6 plants were analyzed for each treatment.

Fusion proteins were expressed in 3-week-old N. benthamiana via Agrobacterium infiltration of leaves as previously described (Sparkes et al., 2006). Leaves were excised 3 days following infiltration and mounted with water under a 0.17 mm coverslip and imaged using a Nikon A1Si confocal microscope (Nikon Plan Apo VC 60x NA1.2 water-immersion objective). For GFP imaging, the 488 nm laser line and a 521/50 nm band pass filter was utilized, while a 561 nm laser line and 595/50 nm filter was used for RFP imaging.

Images were analyzed using ImageJ software (Schneider et al., 2012). Images were converted to 8-bit grayscale and the intensity correlation analysis (ICA) method was used for determine the levels of colocalization and by using the JACoP plugin according to (Bolte and Cordelières, 2006). As a control, empty vector was used to transform leaf cells expressing 35S::GFP alone as described previously by Thatcher et al. (2007).

Proteins were extracted from N. bentamiana leaves infiltrated with GFP, β1-GFP or β6-GFP fusion expression constructs following either control or S. sclerotiorum or P. nicotianae treatments [extraction buffer: 25 mM Tris–HCl, pH 7.0, 0.05% Triton X-100, 1 mM dithiothreitol (DTT), and protease inhibitors (Roche, Basel, Switzerland)]. 25 μg of total proteins were loaded onto 12% polyacrylamide gels for protein separation. Proteins separated by SDS-PAGE were electrotransferred to PVDF membranes. The OxyBlot^TM Protein Oxidation Detection Kit (EMD Millipore) was used according to the manufacturer’s instructions for immunoblot detection of carbonyl groups introduced into proteins by reaction with reactive oxygen species (ROS).

To assess the potential for antifungal activity in NLL seed storage proteins, we focussed on the most abundant seed storage proteins in the NLL grain, the β-conglutins (Foley et al., 2011). The 6xHis-tag recombinant β-conglutin proteins were expressed in E. coli and purified using nickel affinity chromatography. To confirm the identity of purified β-conglutins, SDS–PAGE analysis of the purified proteins was performed, which indicated a single protein band of approximately 65 kDa, which is the predicted size of β-conglutin (Supplementary Figure S2A). This was followed by immunoblotting using an anti-β-conglutin antibody which confirmed the identity of the recombinant proteins as β-conglutin (Supplementary Figure S2B). We were successful in expressing and purifying the β1, β2, β3, β4 and β6 recombinant proteins but not β5 and β7.

Subsequently, we examined the effect of the purified β-conglutin proteins on the growth rate of a range of phytopathogenic necrotrophic fungi using in vitro bioassays. The fungal pathogens selected included the legume pathogen F. oxysporum forma specialis (f. sp.) medicaginis (Fom-5190a, a root pathogen), the broad host range pathogens R. solani AG8-1 (isolated from lupin) and S. sclerotiorum (isolated from canola) and the brassica-specific pathogens F. oxysporum f. sp. conglutinans (Fo-5176), R. solani AG2-1 and A. brassicicola (Brassicaceae hosts; details of these pathogens are listed in Table 1). The IC50 values (μM) for each of these fungal isolates was determined (Table 2). Overall the β1 and β6 conglutins showed significantly stronger mycelium growth inhibition when compared to β2-, β3-, and β4-conglutin proteins for both R. solani isolates and A. brassicicola by Tukey–Kramer honestly significant difference (HSD) test (P < 0.05). The β1-conglutin showed a significantly stronger growth inhibition to S. sclerotiorum and the two F. oxysporum isolates, compared to β2-, β3-, and β4-conglutin, where β6-conglutin was not significantly different from β1 (Table 2). Overall β1-, and β6-conglutin showed the strongest mycelial growth inhibition to the various pathogens tested, but interestingly, a sequence alignment of the seven β-conglutin proteins showed β6 exhibits the highest sequence identity to other β-conglutin protein isoforms (78–98%) while β1 had an amino acid sequence with the lowest identity (77–81%; Supplementary Figure S3). Control treatment of filter disks with BSA buffer showed no fungal growth inhibition against any of the isolates tested.

Based on the β-conglutin protein alignments and IC50 data, we decided to focus on β6-conglutin as a representative member of the β-conglutin family. Further characterization of the in vitro anti-fungal properties of β6-conglutin was performed in a detailed time course experiment against two isolates (WAC8672 and WAC10444) of a major fungal pathogen of lupins, C. lupini which causes anthracnose disease (Fischer et al., 2015). Radial outgrowth of C. lupini mycelium on PDA plates toward Whatman filter disks containing control protein (BSA), β6-conglutin or no protein was recorded. Radial growth inhibition was only observed toward filter disks containing β6-conglutin protein and this occurred from as early as 6 days post-inoculation with the mycelial plug for isolate WAC10444 and 8 days post-inoculation for WAC8672 (Figure 1). Combined with the IC50 data in Table 2, these results indicate recombinant NLL β-conglutins exhibit antifungal activity in vitro against both leaf and root-infecting pathogens of legumes and non-legume hosts.

To examine the effect of NLL β6-conglutin in planta, we selected the N. benthamiana infiltration system as a model for assessing the functionality of proteins against various phytopathogens (Ma et al., 2012). This involved Agrobacterium-mediated infiltration into leaves of N. benthamiana plants followed by assessment of antifungal activity in disease assays. The broad host range leaf pathogen S. sclerotiorum was chosen which is readily amenable to N. benthamiana leaf disease assays and secretes the non-host selective toxin oxalic acid to induce disease symptom development (Kim et al., 2008; Williams et al., 2011). In addition, β6-conglutin showed strong inhibition of this pathogen’s growth as compared to other fungal pathogens tested in our in vitro assays (Table 2). The Agrobacterium-infiltration system was used to transiently express β6-GFP or a GFP-only control. The GFP control was infiltrated into one half of the leaf with β6-GFP infiltrated into the other leaf half. Forty-eight hours after infiltration, leaves were inoculated with S. sclerotiorum and progression of lesions was observed over 72 h (Figure 2). In the GFP-only control, necrosis was apparent within 24 h of S. sclerotiorum inoculation with the size of the necrotic lesions progressing rapidly to engulf half of the leaf by 72 h. In stark contrast, there was only limited necrotic damage in the leaves expressing the β6-GFP protein. As β1-conglutin demonstrated strong antifungal activity in vitro yet exhibits the least amino acid identity amongst the NLL β-conglutins, we also assayed β1-GFP in the above experiments. As with β6-GFP, β1-GFP infiltrated leaves also exhibited limited necrotic damage after S. sclerotiorum inoculation (Supplementary Figure S4).

The soil-borne oomycete pathogen of N. benthamiana, P. nicotianae, was also assayed. P. nicotianae is a hemibiotrophic pathogen that causes root rot, leaf necrosis and stem lesions (Liu et al., 2016). As with the S. sclerotiorum assays, both β6- and β1-conglutin strongly inhibited lesion development by P. nicotianae in our N. benthamiana Agrobacterium-infiltration disease assays (Figure 2 and Supplementary Figure S4). At 72 h post S. sclerotiorum or P. nicotianae inoculation, β6-conglutin infiltrated leaf zones exhibited a 71.7 and 85.7% reduction in lesion size, respectively, relative to control treated leaf zones. Similar reductions were recorded for β1-conglutin infiltrated leaf zones (94.2 and 90.3%).

The striking inhibition of S. sclerotiorum and P. nicotianae induced lesions on N. benthamiana leaves expressing β1- or β6-conglutin suggests these pathogens are unable to grow or their growth is severely impaired by these β-conglutins. To examine fungal/oomycete growth we challenged N. benthamiana β-conglutin expressing leaves with pathogen and allowed 72 h for disease symptom development in controls, then assessed the leaves and mycelial growth microscopically after staining with trypan blue, which stains dead plant cells (van Wees, 2008). In leaves expressing β6-GFP or β1-GFP we observed strong inhibition of the S. sclerotiorum and P. nicotianae mycelial growth compared to controls (Figure 3 and Supplementary Figure S5). There was evidence of aggregated cells with short pseudohyphae, particularly at the site of inoculation, however, these were few and sparse in comparison to controls. Combined, our results suggest β1- and β6-conglutins from NLL inhibit hyphal growth of a range of phytopathogenic oomycete and fungi, both in vitro and in vivo.

The effect of β-conglutins on the growth of fungal and oomycete pathogens tested led to the hypothesis that β1- and β6-conglutins might localize at the cell surface, at sites closest to initial pathogen attack. Therefore, the subcellular localization of the β-conglutin proteins using Agrobacterium-mediated transient expression of β-GFP constructs in N. benthamiana leaves was examined. Two-to-three days following Agrobacterium infiltration, the localisation of β-GFP in N. benthamiana leaf epidermal cells was examined by confocal microscopy. Both β6-GFP and β1-GFP were expressed in punctate structures close to the cell surface (plasma membrane; Figure 4, Supplementary Figure S6). To determine the nature of the β6-GFP- and β1-GFP-positive structures, which resembled the pattern described for plasmodesmata (Lee and Lu, 2011), we co-expressed the constructs with the known plasmodesmata marker plasmodesmata-located protein1 or AtPDLP1-mCherry (Thomas et al., 2008). β6-GFP and β1-GFP partially overlapped the expression pattern of AtPDLP1-mCherry (R = 0.0725–0.755), indicating that β6-GFP and β1-GFP were partially located at plasmodesmata (Figure 5, and Supplementary Figure S6).

One of the mechanisms employed by S. sclerotiorum and P. nicotianae during infection of a compatible host is to initially suppress the plant second-phase oxidative burst that occurs 3–6 h after pathogen contact (Levine et al., 1994; Cessna et al., 2000), thereby compromising the capacity of the plant to activate downstream defense pathways (Doke, 1985; Levine et al., 1994; Williams et al., 2011). Therefore the effect of β-conglutin protein on the capacity of S. sclerotiorum and P. nicotianae to suppress the plant oxidative burst was examined. Leaves were infiltrated and allowed to transiently express GFP or β6-GFP conglutin for 48 h before being infected with S. sclerotiorum or P. nicotianae. As soon as hyphae and lesions became visible (within 24 h) leaves were collected for analysis (as shown in Figure 6A). We estimated production of ROS and oxidative burst capacity by examining the level of protein carbonylation in infected compared to mock-inoculated leaves using an OxyBlot Protein Oxidation Detection and immunoassay (Rinalducci et al., 2008). Protein oxidation is one of the covalent modification of proteins induced by ROS such as H₂O₂ or other products of oxidative stress, and carbonylation is one of the most commonly occurring oxidative modifications of proteins, which may be responsible for the alteration in protein activity, for example, signaling (Oracz et al., 2007). Carbonylated proteins have been identified in many plant species at different stage of growth and development (Barba-Espín et al., 2011; Morscher et al., 2015).

Basal levels of protein oxidation, as generated through normal metabolic activity (Alscher et al., 1997; Rinalducci et al., 2008) were observed in the mock-inoculated control leaves expressing GFP-only, as well as in mock-inoculated leaves expressing β6-GFP (Figures 6B–C). Following inoculation with S. sclerotiorum or P. nicotianae protein oxidation remained at similar levels in the GFP-only control leaves (Figures 6B–C). In contrast, we observed a marked increase in the levels of protein oxidation in leaves expressing the β6-GFP following infection with S. sclerotiorum or P. nicotianae when compared to the respective mock-inoculated β6-GFP or the infected GFP-only leaves. The inoculated leaves expressing β6-GFP were nevertheless healthy, as expected (Figure 6A). This suggests that the over-expression of β-conglutin proteins effectively circumvents the initial suppression of the plant oxidative burst by S. sclerotiorum or P. nicotianae.