Mustard and rapeseed (Brassica napus L., B. juncea (L.) Czern., B. nigra (L.) W.D.J.Koch, B. rapa L. and Sinapis alba L.), and the low glucosinolate, low erucic acid variety, Canola are the second most abundant oilseed crops in the world (Rahman et al., 2018b). Following extraction of the oil, the residual seed meal or oil free press cake, which is high in protein, has potential to be developed into a low cost by-product (Joehnke et al., 2018). The major proteins within the mature harvested Brassica seeds are the 12S globulin-type cruciferins, 2S albumin-type napins, and the oil-body proteins, oleosins (Ramlan et al., 2002; Von Der Haar et al., 2014).

Cruciferins, also classified as 11S globulins based on their sedimentation coefficient, are salt soluble neutral glycoproteins (Ren and Bewley, 1999) with molecular weights ranging from 20 to 40 kDa (Von Der Haar et al., 2014), and an isoelectric point (pI) of 7.2 (Stone et al., 2014). They comprise up to 50–70% of the total seed protein (Ramlan et al., 2002; Ebrahimi et al., 2009; Kasprzak et al., 2016). These comparatively larger SSP proteins are composed of two polypeptide chains; the α-chain of 30 kDa and a β-chain of 20 kDa, held together by a disulphide bond (Wu and Muir, 2008; Stone et al., 2014).

The other abundant SSPs are the napins; these are water soluble low-molecular weight basic proteins classified as 2S or 1.7S proteins (Ren and Bewley, 1999), representing 20–40% of total seed protein, and having a molecular weight in the range of 12–17 kDa (Stone et al., 2014; Von Der Haar et al., 2014). Their isoelectric point varies based on the method of extraction and the specific characteristics of the isoforms that exist (Stone et al., 2014). They are composed of two polypeptide chains, a 4.5 kDa small subunit and a large 10 kDa subunit, stabilized primarily by disulphide bonds. Their secondary structure shows a high α-helical content (Wu and Muir, 2008; Stone et al., 2014). Sequence comparison analyses have revealed that napin-type proteins share structural similarity with the prolamin superfamily of proteins which includes major allergens, α-amylase and trypsin inhibitors, and natural anti-fungal proteins (Breiteneder and Radauer, 2004).

Other proteins found in the seed are the oil body proteins including oleosins (which make up 6–8% of seed protein and 3% of the total oleosome weight) (Deleu et al., 2010; Von Der Haar et al., 2014), caleosins, steroleosins, and lipid transfer proteins (Terras F. R. et al., 1992), as well as myrosinase (thioglucoside glucohydrolase, the glucosinolate-degrading plant defense enzyme) (Gupta and Shaw, 2009; Gupta, 2010; Robin et al., 2017), Ca⁺²-dependent-calmodulin binding proteins, dehydrins (Wanasundara, 2011), and thionines, which have roles in plant defense (Jyothi et al., 2007; Pacheco-Cano et al., 2018).

Seeds of rapeseed (Brassica napus L.), black mustard (B. nigra Koch.), mustard collard (B. carinata A. Braun), white mustard (also called yellow mustard or Semen sinapis Albae) (Sinapis alba) have long been utilized in traditional medicine for relieving pain and infection, dyspnea, reducing nodulation, and relieving cough by eliminating phlegm, and as a tonic for stiffness, or muscle aching (Liu et al., 2003; Aboelsoud, 2010; Kabir et al., 2016; Khan et al., 2016; Rahman et al., 2018b). Mustard seeds are also traditionally exploited to prevent microbial growth of food-spoiling bacteria and increase the shelf life of processed food (Mir et al., 2017; Rahman et al., 2018b; Torrijos et al., 2019).

Evidence has suggested that 2S albumins may function in plant defense against pathogenic micro-organisms (Terras F. R. et al., 1992). Purified radish seed 2S albumins were shown to inhibit the growth of fungi and bacteria and together with thionines obtained both from wheat and barley origin increased cell wall permeabilization of the phytopathogens (Terras F. et al., 1992; Terras F. R. et al., 1992; Terras et al., 1993; Terras et al., 1996). Napin-like proteins from seeds of dwarf Chinese white cabbage (Brassica rapa L. syn. Brassica chinensis L.), and Chinese kale (Brassica oleracea L. syn. Brassica alboglabra cv. “Swatow”) both manifested antibacterial activity (Ngai and Ng, 2004). Homologs of 2S albumins from the seeds of other species with moderate to high amino acid sequence identity to rapeseed 2S albumin (Yang et al., 2007) were also reported to possess antimicrobial activities, for example, Wrightia tinctoria (Roxb.) R. Br. from Apocynaceae (Sharma et al., 2017), sesame, Sesamum indicum L. from Pedaliaceae (Maria-Neto et al., 2011) and castor bean, Ricinus communis L. from Euphorbiaceae (Yang et al., 2007). Protein rich defatted seed meal from B. rapa L. var. rapa DC, low in phytic acid and sinapine, demonstrated a broad-spectrum antimicrobial activity against food-borne pathogens (Tenore et al., 2012); however, it was not known what was the active protein/metabolite responsible for the activity, as glucosinolates have also been reported to have antimicrobial activity, and these compounds are found in high concentration in the seed meal along with SSP (Borges et al., 2015).

There is also some evidence that cruciferin proteins have anti-microbial activity. An 11S seed storage protein (SSP) from Momordica cochinchinensis (Lour.) Spreng., with significant amino acid sequence similarity to rapeseed cruciferins was thought to possess antimicrobial activity (Mazzio et al., 2018).

These findings support the hypothesis that the major mustard and rapeseed (Brassica napus, B. juncea, B. nigra, B. rapa, and Sinapis alba) SSPs, napin, and cruciferin could have antibacterial and antifungal activities based on their high sequence similarity to proteins from other species (Rahman et al., 2020). Although the seeds, seed paste and extracts of mustards were traditionally utilized and therefore extensively studied and reported for antimicrobial activities, there is no study on screening such biological activity using purified proteins despite their high sequence similarity to other confirmed antimicrobial proteins. Therefore, this article aims to determine the in silico and in vitro bioactivity of the rapeseed napin and cruciferin proteins to confirm their antimicrobial properties.

Both napin and cruciferin storage proteins are synthesized during seed maturation in embryos (Höglund et al., 1992; Ellerström et al., 1996), and while providing a source of energy for the emerging seedling, these proteins may have additional roles, including defending the developing embryo against phytopathogenic bacteria and fungi. This potential antimicrobial activity is supported by the traditional use of mustard as an antimicrobial in food production. However, whether this activity is a function of the proteins or other compounds in the seeds (such as glucosinolates) remains to be elucidated. In order to provide experimental evidence to support the claims of B. napus napins and cruciferins having antimicrobial activity, in silico analysis of the protein sequences was carried out and in vitro antimicrobial activity of the purified proteins was tested.

Analysis of rapeseed napin sequences ( Supplementary Table 1 ) in the PRINT database indicated conservation of specific signatures to sweet protein mabinlin-2 (P30233) and sweet protein mabinlin-4 (P80353) from C. masaikai ( Table 2 ). These two proteins are related to the known antimicrobial peptide Flo (P24303) from Moringa oleifera ( Supplementary Figure 1A ) (Fisch et al., 2004; Suarez et al., 2005), a cationic polypeptide with flocculating properties that has been shown to destroy the cell membranes of bacteria (Garcia et al., 2019). In addition, Allergen Sin a 1, a napin homolog from S. alba was found to have functional similarity to C. masaikai in the PRINTS database search which is the first observation of possible antimicrobial functionality for this protein. Sin a 1 has been shown to be related to B. rapa and B. napus napins at the sequence level (Rahman et al., 2020). The database search also showed that rapeseed cruciferins ( Supplementary Table 1 ) have high similarity with rice glutelins which have been reported as antimicrobial peptides (Bundó et al., 2014), as well as the antibacterial glycinin protein from G. max (Sitohy et al., 2012; Yang et al., 2016) and antibacterial Legumin A (P09802) from Gossypium hirsutum L. ( Supplementary Figure 1B ) (He et al., 2018). These results suggested the napin and cruciferin are most likely to have antimicrobial properties as well.

Previous reports have suggested that antimicrobial peptides act against a wide range of infectious bacterial and fungal strains, even against those resistant to multiple common antibiotics. Many of the antimicrobial peptides interact with the microbial membranes at hydrophobic and positively charged regions (Fisch et al., 2004). Antimicrobial peptides are also found to possess anionic function within their amphiphilic structures which enable interaction with membranes. These regions diverge from the traditional α-helical peptides to form a cyclic-cysteine-knot conformation found in some plant proteins (Harris et al., 2009; Seong and Hak, 2013). Anionic proteins usually interact with membranes through electrostatic interactions either by repulsion or formation of soluble or insoluble complexes based on charge state (Rommi et al., 2015). Napin is a very hydrophilic protein (Amine et al., 2019), whereas cruciferin has the ability to form soluble complexes with negatively charged carbohydrates on the microbial membrane surface (Rommi et al., 2015).

The amino acid sequence, conserved sequence motifs, and the number of cysteine residues and their spacing are often used as a basis to classify antimicrobial peptides (Nawrot et al., 2014). Napins show the characteristic conserved skeleton of cysteine-knot conformation of C-Xn-C-Xn-CC-Xn-CXC-Xn-C-Xn-C (Shewry et al., 2002) ( Supplementary Figure 1A ). Interestingly, all proteins demonstrating antimicrobial activity, namely, 2S albumin-like proteins, lipid transfer proteins, puroindolines, and thionins share a similar conserved skeleton of cysteines and possibly also a similar conformational folding pattern (Gautier et al., 1994).

Protein motifs are highly conserved during evolution, and it is postulated that the sequence of amino acids in the ligand binding sites to the receptors would remain conserved as well. The nature of binding, the binding pattern and binding energy between motif and receptor can provide evidence for functional interactions (Rajasekaran et al., 2008). To investigate these properties for napins and cruciferins from rapeseed, to bacterial target molecules, a molecular docking study was performed using the rapeseed 2S napin (1PNB) and 11S procruciferin (3KGL, as closest available 3D structure of 12S cruciferin) from B. napus as ligands. The docked models indicated that, like the earlier antimicrobial peptides, the proteins showed distinct amphipathic properties, cruciferin showed an amphipathic β-helical dominated structure while napin showed an amphipathic α helix dominated structure (Chang et al., 2015; Perera et al., 2016). The cysteine-based motif, conserved cysteine skeleton CXnCXnCCXnCXCXnCXnC held together by four disulfide bonds where C represents cysteine residues and Xn could be any number of other amino acids are found at positions 64, 76, 240, 241, 252, 254, 310, and 319 in the alignment ( Supplementary Figure 1A , color-coded in yellow), can act like a hinge (Chang et al., 2015). The conserved cysteine residues at the edges of the cysteine-based motif are able to form disulfide bridges to the cysteine residues of other distant subunits; while the nearby glycine residue could provide the hinge region motif more flexibility (Chang et al., 2015).

The characteristic conserved cysteine-based motifs of 2S albumin-like napins are often used as fingerprints and exploited to characterise the prolamin superfamily (Shewry et al., 2002; Sharma et al., 2017). These motifs encode protein folds and provide flexibly to the structure. These motifs were identified as having similarity in the SPRINT database with previously reported antimicrobial peptides ( Supplementary Table 1 and Supplementary Figures 1A, B ). In earlier studies, it was found that the bactericidal activity is associated to a specific sequence motif of amino acid loops, like cysteine-cysteine or proline-proline loops that are stabilised by formation of disulphide bridges. These have a tendency to form a helix-loop-helix conformation and this motif causes bacterial membrane damage (Suarez et al., 2005). Because napin type 2S albumins possess the same assembly of several copies of this conformational motif into a branched peptide, it could be assumed that they may also exert their antibacterial action in the same way the M. oleifera seed protein does.

Interestingly, cruciferins showed high (>86%) coverage and significant pairwise sequence identity (28-32%) with soybean glycinin ( Supplementary Figure 1B ) (Ramlan et al., 2002; Sitohy et al., 2012). It is also evident there are two highly conserved cysteine residues at positions 71 and 114 in the alignment with many other conserved amino acid motifs ( Supplementary Figure 1B , color-coded in yellow).

Earlier reports (Chang et al., 2015; Farkas et al., 2017; Mohan et al., 2019) analyzing the critical regions of different antimicrobial proteins and examining antimicrobial peptide databases suggested that presence and absence of amino acid, their number and arrangement are critical for antimicrobial proteins. These influence the secondary structure and charge of the protein, conserved protein domains, amphipathicity, peptide aggregation, gapless alignments to highly similar protein sequences, receptor binding and ultimately antimicrobial functionality. In addition, these studies report that glycine is the most abundant residue in the critical regions of antimicrobial peptides (Chang et al., 2015). Interestingly, napins possess six glycine residues that enable the cysteine rich hinge motif higher flexibility, whereas cruciferins have much higher number of glycine residues ( Table 4 ).

While in silico bioinformatics analysis of napin and cruciferin protein sequences suggested features and sequence motifs that have been attributed to anti-microbial proteins, and docking studies provided evidence for the ability of the proteins to bind to microbial proteins with high binding energy values, docking scores, and protein-receptor interactions (Barakat et al., 2016), the in vitro functional tests carried out in this study only demonstrated antimicrobial activity for napin against S. saprophyticus ( Figure 5 ). This activity was bacteriostatic, not bactericidal, as the napin-induced growth inhibition microdilutions, when inoculated to Blood MH agar, grew S. saprophyticus, indicating bacterial viability when removed from the inhibiting effects of napin. The napin antimicrobial activity observed here provides evidence to warrant the further investigation of its antimicrobial activity using increased concentrations, with different extraction and purification approaches, and against different microorganisms. The bacteria Klebsiella oxytoca and Enterobacter cloacae used in this AST are intrinsically resistant to ampicillin. Napin exhibited strong antimicrobial activity against S. saprophyticus indicating a possible alternative to control bacteria.

In the case of cruciferin, while the zones of inhibition were not beyond that of the blank negative control (acetic acid/acetonitrile/water), indicating poor antimicrobial activity, there is the possibility that activity was masked by the solvent’s innate antimicrobial activity. The inclusion of this solvent in a growth control in the microdilution AST confirmed its antimicrobial activity, rendering it a poor solvent for AST. Therefore, cruciferin dissolved in 2% DMSO was applied in the microdilution AST (final concentration 1% DMSO). The 1% and 2% DMSO growth controls did not inhibit microbial growth; therefore, the lack of antimicrobial activity of cruciferin at the tested concentrations was confirmed. Higher concentrations need to be tested to determine the limit of detection for antimicrobial activity for both cruciferin and napin.

The in vitro activity we have shown for napin is somewhat contrary to what was recently reported for a napin protein from B. juncea (BjN), which showed strong inhibition of growth of Xanthomonas oryzae and Staphylococcus aureus at 60 µg, in disc diffusion AST with 0.1 mM Tris buffer solvents in Luria-Bertani medium (Munir et al., 2019). Antimicrobial activity of cruciferin and stronger antimicrobial activity of napin may be evident if higher concentrations of the proteins are used in the disc diffusion and microdilution AST, or the antimicrobial action of these proteins may be pH dependent, requiring different bacterial media with favorable diffusion and buffer solvents to maintain an optimal pH.

Several factors may play an important role behind these discrepancies and could relate to the initial protein extraction process, purification methods, stability of the protein, and even the method used for the disc diffusion functional assays. In this study, purified napin and cruciferin from Brassica napus, purchased from a commercial source, were used to evaluate the antimicrobial activity. These proteins were purified by simple ion exchange chromatography, but it is unknown how the initial extraction process was carried out, and what intermediary purification procedures were followed. This method of extraction may result in changes to protein functionality as low pH can cause the formation of protein aggregates (Wanasundara, 2011). Moreover, earlier studies showed cruciferin was unstable at low pH and high temperature, affecting its solubility, and causing unfolding of the protein (Perera et al., 2016).

Another factor that may play a role is the specific media used for the plate diffusion assay. The diffusion properties of napin and cruciferin through the Mueller Hinton agar media are unknown. The disc diffusion method works on the principle of the molecule being able to readily diffuse through agar to form a concentration gradient. The recent report on the antibacterial activity of napin used Luria-Bertani medium, as well as different methods for protein solubilization and application (Munir et al., 2019). The dose-dependent zones of inhibition observed for cruciferin in this study ( Figure 4 ), that were only a few millimeters around the disc circumference, are attributed to the solvent used (acetic acid/acetonitrile/water), rather than the antibacterial activity of the protein, as the solvent concentration decreases with increasing dilutions, and the zones of inhibition decreased with each dilution. Furthermore, this solvent, in the absence of cruciferin, demonstrated strong antimicrobial activity against all bacteria tested in the microdilution AST.

One question that remains to be answered is how these SSPs exert their antimicrobial action. Evidence has shown that this could occur through different actions, including growth inhibition of bacteria, with the proteins inducing plasma membrane permeabilization leading to loss of integrity of the cell (Ribeiro et al., 2012). Additionally, antimicrobial proteins can induce production of nitric oxide in diverse pathogenic and non-pathogenic microorganisms (Ribeiro et al., 2012), or increase the production of reactive oxygen species (Garcia et al., 2019), leading to cell death. Some mammalian proteins are chemically converted to antimicrobial peptides inside the animal body. For example, truncated α-defensins ligate among themselves in the primate leukocytes and produce cyclic antimicrobial peptides which act against both bacteria and fungi even in low micromolar concentrations (Tang et al., 1999). It would be worth investigating the chemical modifications the studied peptides undergo and the biological activities of such chemically modified peptides in biological systems. Further work to elucidate this mode of action is needed. In this study, the proteins were tested against common laboratory microorganisms used for AST. Cruciferin and napin could be applied against a wider spectrum of disease-causing microorganisms for human, livestock and crops to check if they could be useful in the management of pathogen-borne diseases.

This study evaluated the evidence for the role of rapeseed SSPs, napin, and cruciferin as antimcrobial agents. Similarity with other plant antimicrobial peptides through conservation of sequence motifs and specific amino acids, as well as 3D structural analysis, was presented. The results support further functional studies into the potential application of napin and cruciferin as potent candidates for antimicrobial agents, as well as functional food ingredient and in complementary medicine to alleviate diseases, as preservative for wide range of foods as well as crop protection from pathogens. Among a range of bacterial species tested, only napin demonstrated biological activity against Staphylococcus saprophyticus. The evidence presented here supports further investigation of the antimicrobial activity of napin and cruciferin for their potential application in the health, food, and agricultural industries.

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation, to any qualified researcher.

MR, MH, JB, JC, LL, and BB contributed substantially with data analysis and discussions of the results. MR conceived the study, designed the experiments, and wrote the article. JB, JC, and MR carried out the agar plate disc diffusion antimicrobial susceptibility tests and analysis. JB and JC carried out the microdilution antimicrobial susceptibility tests and analysis. MH extracted the molecular docking results. BB and LL supervised the study, edited, and reviewed the manuscript before submission.

MR received support from an International Postgraduate Research Scholarship (IPRS) and Australian Postgraduate Award scholarship (APA), funded by the Australian Government to pursue his PhD from February 2016 to July 2019 and this article is a part of his PhD study. This study did not obtain any other external grant from funding bodies.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

The handling editor declared a shared affiliation, though no other collaboration, with one of the authors, MH, at time of review.