Anti-Fungal Hevein-like Peptides Biosynthesized from Quinoa Cleavable Hololectins

Chitin-binding hevein-like peptides (CB-HLPs) belong to a family of cysteine-rich peptides that play important roles in plant stress and defense mechanisms. CB-HLPs are ribosomally synthesized peptides that are known to be bioprocessed from the following two types of three-domain CB-HLP precursor architectures: cargo-carrying and non-cargo-carrying. Here, we report the identification and characterization of chenotides biosynthesized from the third type of precursors, which are cleavable hololectins of the quinoa (Chenopodium quinoa) family. Chenotides are 6-Cys-CB-HLPs of 29–31 amino acids, which have a third type of precursor architecture that encompasses a canonical chitin-binding domain that is involved in chitin binding and anti-fungal activities. Microbroth dilution assays and microscopic analyses showed that chenotides are effective against phyto-pathogenic fungi in the micromolar range. Structure determination revealed that chenotides are cystine knotted and highly compact, which could confer resistance against heat and proteolytic degradation. Importantly, chenotides are connected by a novel 18-residue Gly/Ala-rich linker that is a target for bioprocessing by cathepsin-like endopeptidases. Taken together, our findings reveal that chenotides are a new family of CB-HLPs from quinoa that are synthesized as a single multi-modular unit and bioprocessed to yield individual mature CB-HLPs. Importantly, such precursors constitute a new family of cleavable hololectins. This unusual feature could increase the biosynthetic efficiency of anti-fungal CB-HLPs, to provide an evolutionary advantage for plant survival and reproduction.


Introduction
Chenopodium quinoa (C. quinoa), a pseudo-cereal from the Amaranthaceae family, has been grown as a staple in the Andean region for thousands of years. C. quinoa has high tolerance to extreme climate changes, such as drought, high salinity, and cold. Moreover, the high adaptability of C. quinoa promotes its survival and growth in marginal land at high altitudes (>3800 m above sea level). C. quinoa grains also have extensive nutritional value, and are a protein-rich food with a high vitamin and mineral content [1,2].
Nutritive plants, such as quinoa, can be threatened by plant pathogens. To defend against the pathogens, plants have evolved physical barriers and adaptive mechanisms to fight against, or establish a symbiotic relationship with, plant pathogens and pests [3][4][5]. An example is antimicrobial peptides produced by plants, which act as important effector molecules against plant microbial infections [6][7][8][9].
Plants produce a group of highly stable antimicrobial peptides that bind to chitin, a naturally occurring polysaccharide found in the exoskeleton of insects and the cell wall of fungi [10]. Hevein, the prototypic member of this group, is a 43 amino acid antimicrobial cysteine-rich peptide (CRP), discovered from the latex of the rubber tree

Primary Sequence Determination of Chenotides
Scaled-up extraction was performed using 2 kg of C. quinoa var. Willd, and purified using strong cation exchange and C18 reverse-phased high-performance liquid chromatography. The CRPs isolated from C. quinoa var. Willd were designated as chenotides cQ1-3, and produced a yield of approximately 50 mg per kg of dried plant material. The chenotides cQ1-3 have molecular weights [M + H] + of 2965 Da, 2893 Da, and 2836 Da, respectively.
To determine the primary sequence of chenotides, the isolated peptides were first Sreduced with DTT and S-alkylated with IAM. The S-alkylated peptides were then subjected to enzymatic digestion with trypsin or chymotrypsin. The resulting peptide fragments were analyzed by MALDI-TOF MS and de novo sequenced using the b-ions and yions generated via MALDI-TOF MS/MS. Using chenotide cQ1 as an example, de novo

Primary Sequence Determination of Chenotides
Scaled-up extraction was performed using 2 kg of C. quinoa var. Willd, and purified using strong cation exchange and C18 reverse-phased high-performance liquid chromatography. The CRPs isolated from C. quinoa var. Willd were designated as chenotides cQ1-3, and produced a yield of approximately 50 mg per kg of dried plant material. The chenotides cQ1-3 have molecular weights [M + H] + of 2965 Da, 2893 Da, and 2836 Da, respectively.
To determine the primary sequence of chenotides, the isolated peptides were first Sreduced with DTT and S-alkylated with IAM. The S-alkylated peptides were then subjected to enzymatic digestion with trypsin or chymotrypsin. The resulting peptide fragments were analyzed by MALDI-TOF MS and de novo sequenced using the b-ions and y-ions generated via MALDI-TOF MS/MS. Using chenotide cQ1 as an example, de novo peptide sequencing determined its primary sequence to be AGECVRGRCPGGLCCSKFGFCGSGPAYCGGA (Supplementary Figure S2). This sequence was confirmed against the cDNA sequence from GenBank. Our mass spectrometry results showed that chenotides cQ2 and cQ3 were N-terminal truncated sequences of cQ1 ( Figure 2A, Table 1, and Supplementary  Figures S3 and S4). peptide sequencing determined its primary sequence to be AGECVRGRCPGGLC KFGFCGSGPAYCGGA (Supplementary Figure S2). This sequence was confirmed aga the cDNA sequence from GenBank. Our mass spectrometry results showed that che tides cQ2 and cQ3 were N-terminal truncated sequences of cQ1 ( Figure 2A, Table 1, Supplementary Figures S3 andS4).

Peptide Stability of Cystine-knotted Chenotide cQ2
Cysteine-rich peptides that are cross-linked by multiple disulfides are known for their stability against heat, acid, and proteolytic degradation [17][18][19][20][21]. Our results showed that chenotide cQ2 is indeed highly stable against heat (95 • C), acid (1M HCl), and proteolytic degradation (pepsin, aminopeptidase I, horse serum, and human serum) ( Figure 3). In all the conditions, >80% of the peptides were retained after treatment, as monitored by RP-HPLC and MALDI-TOF MS. In contrast, S-alkylated chenotide cQ2 was not stable under the same conditions, indicating that the cystine-knot scaffold is responsible for the hyperstability of chenotides.

Peptide Stability of Cystine-knotted Chenotide cQ2
Cysteine-rich peptides that are cross-linked by multiple disulfides are known for their stability against heat, acid, and proteolytic degradation [17][18][19][20][21]. Our results showed that chenotide cQ2 is indeed highly stable against heat (95 °C), acid (1M HCl), and proteolytic degradation (pepsin, aminopeptidase I, horse serum, and human serum) ( Figure 3). In all the conditions, >80% of the peptides were retained after treatment, as monitored by RP-HPLC and MALDI-TOF MS. In contrast, S-alkylated chenotide cQ2 was not stable under the same conditions, indicating that the cystine-knot scaffold is responsible for the hyperstability of chenotides.

Chenotide cQ2 Is Chitin-Binding
To confirm the chitin binding activity of chenotide cQ2, the native peptides were incubated with chitin beads at 25 °C, and the S-alkylated forms were used as a control (Figure 5). After incubation for 1 h, analysis using RP-HPLC revealed complete depletion of

Chenotide cQ2 Is Chitin-Binding
To confirm the chitin binding activity of chenotide cQ2, the native peptides were incubated with chitin beads at 25 • C, and the S-alkylated forms were used as a control  Figure 5). After incubation for 1 h, analysis using RP-HPLC revealed complete depletion of the native chenotide cQ2 from the incubating solution, suggesting that this peptide binds the chitin beads. This binding was confirmed after eluting chenotide cQ2 from the chitin beads using approximately 30% 1 M acetic acid at 55 • C. On the other hand, the control S-alkylated chenotide cQ2 did not bind to the chitin resin, indicating that the cystine-knot disulfide scaffold is important to maintain the surface topology of the CB domain, for the recognition and binding of chitin.
Molecules 2021, 26, x 3 of 18 the native chenotide cQ2 from the incubating solution, suggesting that this peptide binds the chitin beads. This binding was confirmed after eluting chenotide cQ2 from the chitin beads using approximately 30% 1 M acetic acid at 55 °C. On the other hand, the control Salkylated chenotide cQ2 did not bind to the chitin resin, indicating that the cystine-knot disulfide scaffold is important to maintain the surface topology of the CB domain, for the recognition and binding of chitin.

Chenotide Precursors Belong to a New Family of Cleavable Hololectins
Based on primary sequence determination, we showed that chenotides were biosynthesized as an unusual three-domain precursor, consisting of an N-terminal signal peptide, two tandem-repeating, identical mature CB-HLP domains, and a C-terminal tail (Figure 7A). The hinge region connecting the two mature CB-HLP domains has 18 residues and is Ala-rich. The cleavage site located between Ala-Ala suggests the involvement of cathepsin-like endopeptidases in the bioprocessing and release of the mature domains. A similar precursor architecture has been reported for another six-cysteine CB-HLP, Sm-Amp-1 (UniProtKB-E1UYT9), from chickweed (Stellaria media). Unlike chenotides, the Sm-Amp-1 precursor does not possess two identical mature CB-HLP domains [39].

Chenotide Precursors Belong to a New Family of Cleavable Hololectins
Based on primary sequence determination, we showed that chenotides were biosynthesized as an unusual three-domain precursor, consisting of an N-terminal signal peptide, two tandem-repeating, identical mature CB-HLP domains, and a C-terminal tail ( Figure 7A). The hinge region connecting the two mature CB-HLP domains has 18 residues and is Ala-rich. The cleavage site located between Ala-Ala suggests the involvement of cathepsinlike endopeptidases in the bioprocessing and release of the mature domains. A similar precursor architecture has been reported for another six-cysteine CB-HLP, Sm-Amp-1 (UniProtKB-E1UYT9), from chickweed (Stellaria media). Unlike chenotides, the Sm-Amp-1 precursor does not possess two identical mature CB-HLP domains [39]. CB-HLPs are ribosomal-synthesized peptides that are processed from a common three-domain precursor architecture, consisting of a signal peptide domain, a mature peptide domain, and a C-terminal domain [7,19,22]. CB-HLPs have several subtypes of precursor architectures (Figures 7B and 8) [34]. Type I has a three-domain precursor comprising a signal peptide, mature CB-HLP peptide, and a short C-terminal tail. Some examples include altides and Ar-AMP [23,40]. Type II also has three domains with a long C-terminal tail that usually encodes bioactive protein cargoes, such as proteins having a Barwin-like or class I chitinase-like domain [22,34] An example is Ee-CBP from Euonymus europaeus, which has a long C-terminal chitinase-like domain [41,42]. In this study, the chenotide precursors belong to a type III variant that is similar to Sm-Amp-1 [39]. Instead of a threedomain arrangement, the precursor architecture has tandem repeats of the CB-HLP domain. Chenotide precursors contain two identical repeats of the mature peptide domain, with a cleavable hinge region. This pattern was also observed in cyclotides, such as Tip-top3 from Mormodica cochinchinensis, which encodes eight cyclic CRPs, with potent trypsin inhibitory and insecticidal activities [43]. Thus, gene amplification could be an evolutionarily advantageous trait to boost the biosynthetic efficiency of these CB-HLPs, to benefit plant survival and reproduction [44]. CB-HLPs are ribosomal-synthesized peptides that are processed from a common three-domain precursor architecture, consisting of a signal peptide domain, a mature peptide domain, and a C-terminal domain [7,19,22]. CB-HLPs have several subtypes of precursor architectures (Figures 7B and 8) [34]. Type I has a three-domain precursor comprising a signal peptide, mature CB-HLP peptide, and a short C-terminal tail. Some examples include altides and Ar-AMP [23,40]. Type II also has three domains with a long C-terminal tail that usually encodes bioactive protein cargoes, such as proteins having a Barwin-like or class I chitinase-like domain [22,34] An example is Ee-CBP from Euonymus europaeus, which has a long C-terminal chitinase-like domain [41,42]. In this study, the chenotide precursors belong to a type III variant that is similar to Sm-Amp-1 [39]. Instead of a three-domain arrangement, the precursor architecture has tandem repeats of the CB-HLP domain. Chenotide precursors contain two identical repeats of the mature peptide domain, with a cleavable hinge region. This pattern was also observed in cyclotides, such as Tiptop3 from Mormodica cochinchinensis, which encodes eight cyclic CRPs, with potent trypsin inhibitory and insecticidal activities [43]. Thus, gene amplification could be an evolutionarily advantageous trait to boost the biosynthetic efficiency of these CB-HLPs, to benefit plant survival and reproduction [44]. Due to their carbohydrate-binding properties, CB-HLPs are also grouped as the following lectins: merolectins, chimerolectins, and hololectins ( Figure 8) [34]. This nomenclature was assigned based on the number of carbohydrate-binding domains present in the mature peptide sequence [45,46]. Non-cargo-carrying CB-HLPs are categorized as merolectins that do not exhibit agglutinin activity [47]. Cargo-carrying CB-HLPs are categorized as chimerolectins that contain one or more carbohydrate-binding domains, linked to a long catalytic protein cargo, such as chitinase [47]. Lastly, hololectins, such as wheat germ agglutinin (WGA), are tandem-repeating carbohydrate-binding proteins with agglutin activity [48,49]. Hololectins are genetically expressed and released as a single multimodular unit with tandem-repeating domains, connected by linkers that are commonly known as hinge regions [48,49].
According to lectin nomenclature, chenotide precursors belong to the group of hololectins, due to the presence of tandem repeats in their CB-HLP domain. Hololectins are expressed as a single-chain multi-modular protein, with tandem repeats of a carbohydrate-binding domain connected by hinge sequences [47]. Examples are a 171 amino acid residue WGA isolated from Triticum vulgaris [49], and a 227 amino acid residue Oryza sativa agglutinin (OSA) [50]. Unlike hololectins, the tandem repeats of chenotides were observed at the gene, but not protein level. Each modular CB-HLP domain in the chenotide precursor is released as mature CB-HLPs. Thus, we classify them under a new family of cleavable hololectins.
To identify the important features that differentiate precursors of cleavable hololectins from non-cleavable hololectins, we performed a sequence comparison of six wellcharacterized, non-cleavable hololectins from UniProt, including WGA [48,49], phytolacca lectin-C (PL-C) [51], phytolacca lectin-D2 (PL-D2) [52], phytolacca lectin-B (PL-B) [53], Oryza sativa agglutinin (OSA) [50], and barley root-specific lectin with chenotide [54] ( Figure 9). Our findings revealed that the hinge regions of non-cleavable hololectins are highly conserved, and are 4-6 amino acid residues in length. In contrast, chenotide and Sm-Amp-1 precursors possess longer hinge regions of 16-18 amino acid residues. However, the linkers of chenotide precursors differ from the Sm-Amp-1 precursor, by being Gly/Ala-rich. As such, these linkers are susceptible to the cleavage by cathepsin-like endopeptidases. This extended spacer could facilitate the access of endopeptidases to cleave Due to their carbohydrate-binding properties, CB-HLPs are also grouped as the following lectins: merolectins, chimerolectins, and hololectins ( Figure 8) [34]. This nomenclature was assigned based on the number of carbohydrate-binding domains present in the mature peptide sequence [45,46]. Non-cargo-carrying CB-HLPs are categorized as merolectins that do not exhibit agglutinin activity [47]. Cargo-carrying CB-HLPs are categorized as chimerolectins that contain one or more carbohydrate-binding domains, linked to a long catalytic protein cargo, such as chitinase [47]. Lastly, hololectins, such as wheat germ agglutinin (WGA), are tandem-repeating carbohydrate-binding proteins with agglutin activity [48,49]. Hololectins are genetically expressed and released as a single multi-modular unit with tandem-repeating domains, connected by linkers that are commonly known as hinge regions [48,49].
According to lectin nomenclature, chenotide precursors belong to the group of hololectins, due to the presence of tandem repeats in their CB-HLP domain. Hololectins are expressed as a single-chain multi-modular protein, with tandem repeats of a carbohydrate-binding domain connected by hinge sequences [47]. Examples are a 171 amino acid residue WGA isolated from Triticum vulgaris [49], and a 227 amino acid residue Oryza sativa agglutinin (OSA) [50]. Unlike hololectins, the tandem repeats of chenotides were observed at the gene, but not protein level. Each modular CB-HLP domain in the chenotide precursor is released as mature CB-HLPs. Thus, we classify them under a new family of cleavable hololectins.
To identify the important features that differentiate precursors of cleavable hololectins from non-cleavable hololectins, we performed a sequence comparison of six well-characterized, non-cleavable hololectins from UniProt, including WGA [48,49], phytolacca lectin-C (PL-C) [51], phytolacca lectin-D2 (PL-D2) [52], phytolacca lectin-B (PL-B) [53], Oryza sativa agglutinin (OSA) [50], and barley root-specific lectin with chenotide [54] (Figure 9). Our findings revealed that the hinge regions of non-cleavable hololectins are highly conserved, and are 4-6 amino acid residues in length. In contrast, chenotide and Sm-Amp-1 precursors possess longer hinge regions of 16-18 amino acid residues. However, the linkers of chenotide precursors differ from the Sm-Amp-1 precursor, by being Gly/Ala-rich. As such, these linkers are susceptible to the cleavage by cathepsin-like endopeptidases. This extended spacer could facilitate the access of endopeptidases to cleave and release each CB-HLP domain. Further investigations are warranted to understand the unique characteristic of these hinge regions of the new cleavable hololectin family. and release each CB-HLP domain. Further investigations are warranted to understand the unique characteristic of these hinge regions of the new cleavable hololectin family.

Materials
All chemicals and solvents, unless otherwise stated, were purchased from Sigma Aldrich, St. Louis, MO, US, and Fisher Scientific, Cleveland, OH, US.

Materials
All chemicals and solvents, unless otherwise stated, were purchased from Sigma Aldrich, St. Louis, MO, US, and Fisher Scientific, Cleveland, OH, US.

Plant Materials
Different varieties of quinoa were purchased from Seedville, Canton, OH, USA. Authentication was conducted by Mr. Paul Leong at the Singapore Botany Center based on macroscopic and microscopic analyses. A voucher for each sample was deposited at the Nanyang Technological University Herbarium, School of Biological Sciences, Singapore.

Extraction, Isolation, and Purification
Small-scale screening of quinoa was performed by mixing 0.1 g of sample with 1 mL water for 1 h. The crude extract was centrifuged at 9500 rpm for 10 min. The supernatant was subjected to C 18 ZipTip and eluted with 80% ACN. For large-scale extraction, 2 kg of samples were homogenized in 20 L water for 3 h. The crude extract was centrifuged at 9500 rpm for 20 min at 4 • C. The supernatant was incubated with 80% ammonium sulfate for 1 h and centrifuged at 9500 rpm for 20 min at 4 • C. The pellet was re-suspended in 10% ACN and 1 h later, was centrifuged at 9500 rpm for 20 min at 4 • C. The filtered supernatant was loaded on a flash column packed with 500 g of C 18 powder (Grace, Columbia, MD, US) in a Büchner funnel. Elution was performed using increasing concentrations of ethanol (20-80%). Eluents that contained the peptide of interest were pooled and purified using multiple rounds of SCX-and RP-HPLC. Fractions from SCX-HPLC containing the peptide of interest were pooled and purified by RP-HPLC.

Matrix-assisted Laser Desorption/Ionization Time-of-Flight Mass Spectrometry
Matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF MS/MS; AB SCIEX 5800 MALDI-TOF/TOF, ABsciex, Foster City, CA, USA) was used in this study. The MALDI-MS spectra were acquired with a laser intensity of 3500, total laser shots were 2250. MALDI-TOF MS/MS spectra were acquired with a laser intensity of 5000, total laser shots were 5000.

Sequence Determination
The primary sequences of the chenotides were determined by MS/MS sequencing. Peptide (40 µg) was incubated with 20 mM dithiothreitol (DTT) at 37 • C for 1 h to reduce the disulfide bonds. S-reduced chenotides were digested with trypsin or chymotrypsin in 5:1 (v/v) ratio in ammonium bicarbonate buffer (25 mM), pH 8 at 37 • C for 10 min. Following digestion, the digested peptides were subjected to MALDI-TOF MS/MS sequencing. Assignment of isobaric residues Lys/Gln and Leu/Ile were based on sequence comparison to genomic or expressed sequence tag (EST) data from the National Center for Biotechnology Information (NCBI) database.

NMR Structural Study
All NMR experiments were conducted on a BRUKER Avance 800 NMR spectrometer (Bruker Daltonics, Bremen, Germany) with a cryogenic probe at 25 • C. The concentration of each peptide was around 1 mM in 5% D 2 O and 95% H 2 O (pH 3.5). For 1 H, 1 H-2D TOCSY and NOESY, the mixing times were 80 and 200 ms, respectively. The spectral width was 12 ppm for both dimensions. The NMR spectra were processed using NMRPipe software (http://spin.niddk.nih.gov/NMRPipe/) (accessed on 30 August 2021) [55]. All data analyses were performed using Sparky software (http://www.cgl.ucsf.edu/home/ sparky/) (accessed on 30 August 2021) based on the 2D NOESY and TOCSY results [56]. The proton chemical shift assignments for each amino acid residue were achieved by 2D TOCSY and NOESY, while the proton-proton distance restraints were obtained from 2D NOESY by the intensities of NOE cross peaks. The chenotide solution structures were calculated using CNSsolve 1.3 software (http://cns-online.org/v1.3/) (accessed on 30 August 2021) [57]. The proton-proton distance restraints and hydrogen bonds were employed in a standard simulated annealing protocol. The distance restraints were divided into the following three classes based on NOE cross-peak intensities: strong, 1.8 < d < 2.9 Å; medium, 1.8 < d < 3.5 Å; and weak, 1.8 A < d < 5 Å. Eight hydrogen bonds were used in the structure calculation. A total of 100 structures were calculated. Twenty and 10 of the lowest-energy structures were chosen for data statistics and presentation, respectively. The structure was verified using the PROCHECK program [58] and presented using Chimera version 1.11.2 (https://www.cgl.ucsf.edu/chimera/) (accessed on 30 August 2021) [59].

Ligand Peptide Docking
Prior to ligand and peptide docking, both the peptide and ligand were prepared using the Chimera version 1.11.2 (https://www.cgl.ucsf.edu/chimera/) (accessed on 30 August 2021) for the addition of hydrogen atoms and conversion of PDB format to MOL2 format. The GOLD 5.4.0 version (Cambridge Crystallographic Data Centre, Cambridge, UK) was utilized to perform ligand-peptide docking using the NMR structure of chenotide cQ2 (PDB: 5ZV6). The GOLD score function takes into consideration the hydrogen bond and van de Waals energy. To define the active site pocket, one atom on the active site was chosen to define the pocket radius. The other settings in the program were set at default function.

Chitin-binding Assay
Chitin binding assays were performed as described previously [23]. S-alkylated and purified cQ2 were mixed with chitin beads (80 µL) (New England BioLabs, Ipswich, MA, USA) in chitin binding buffer and incubated at 25 • C for 30 min. The mixture was then washed with chitin binding buffer (50 mM Tris HCl, 500 mM NaCl, pH8) to remove unbound peptide. Elution of bound peptide was performed with 1 M acetic acid. The supernatant and eluent were analyzed using RP-UPLC and MALDI-TOF MS to assess binding and elution.

Peptide Stability Assays
Assessments of thermal, acidic and proteolytic stability were performed as previously described [19]. Purified chenotide cQ2 was incubated at the stated conditions and recommended buffer solution. At each time interval, samples were aliquoted in triplicate and RP-UPLC was performed. The area under the peak was used to determine the amount of chenotide present before and after treatment.

Anti-Fungal Assay
A radial disc diffusion assay was used to assess chenotide anti-fungal activity. The following four phyto-pathogenic fungal strains were obtained from the China Center of Industrial Culture Collection (Beijing, China): Alternaria alternata (CICC 2465), Curvularia lunata (CICC 40301), Fusarium oxysporum (CICC 2532) and Rhizoctonia solani (CICC 40259). Fungal strains were grown on potato dextrose agar plates at 25 • C. When sufficient growth was observed, a hole was punched in the culture and the plug was transferred to a new agar plate. The plate was then incubated at 25 • C for 48 h until a radial mycelial colony formed. Paper discs with a diameter of 0.65 cm were soaked with 20 µL of peptides and then placed equidistant from the growing ends of the mycelia. Deionized water was used as a negative control. The formation of an arc-shape inhibition zone around the disk indicated anti-fungal activity.
The half-maximal inhibitory concentration levels (IC 50 ) of peptides were determined by a microbroth dilution assay [60]. Fungal spores were harvested from a 4-day-old actively growing fungal plate and suspended in half-strength potato dextrose broth. Spore suspensions (1 × 10 5 cells/mL) were mixed with peptides at varying concentrations in 96-well microplates and incubated at 25 • C for 24 h. The cells were then fixed with 100% methanol for 15 min and stained for 45 min with crystal violet dye. Excess dye was removed with MilliQ water. Elution was performed using 1:1 (v/v) ethanol/0.1 N HCl. The absorbance was measured at 570 nm.

Conclusions
This study expands our knowledge on the occurrences and biosynthesis of CB-HLPs. We discovered, identified, and characterized novel hyperstable anti-fungal CB-HLPs chenotides from quinoa. We showed that the biosynthesis of chenotides is novel and belongs to a new cleavable hololectin family, which we have labeled as type III lectin precursor. A characteristic of such a precursor is that it contains tandem repeats of the mature peptide domains, with a cleavable Gly/Ala-rich linker consisting of 18 amino acids. Furthermore, chenotides are chitin binding and can inhibit the growth of phyto-pathogenic fungi. Taken together, the occurrence of chenotides as natural anti-microbial agents in quinoa could be the underlying reason for their extended shelf-life and unintended selection as essential staples throughout the history of humankind.

Data Availability Statement:
The data presented in this study are available on request from the corresponding authors.