Silkworm pupae as source of high‐value edible proteins and of bioactive peptides

Abstract To characterize the high‐value protein content and to discover new bioactive peptides, present in edible organisms, as silkworm pupae, semiquantitative analytical approach has been applied. The combination of appropriate protein extraction methods, semiquantitative high‐resolution mass spectrometry analyses of peptides, in silico bioactivity and gene ontology analyses, allowed protein profiling of silkworm pupae (778 gene products) and the characterization of bioactive peptides. The semiquantitative analysis, based on the measurement of the emPAI, revealed the presence of high‐abundance class of proteins, such as larval storage protein (LSP) class. This class of proteins, beside its nutrient reservoir activity, is of great pharmaceutical interest for their efficacy in cardiovascular diseases. Potential allergens were also characterized and quantified, such as arginine kinase, thiol peroxiredoxin, and Bom m 9. This powerful bioanalytical approach proved the potential industrial applications of Bombyx mori pupae, as source of high‐value proteins in a green and “circular” economy perspective.

Following the reintroduction of sericulture in Italy, owing to the consolidated expertise of the CREA-API excellence research center in Padua, the production of Italian silk is going to start again on an industrial scale, increasing production of pupae too. Due to this fact, the silk industry might induce a real "circular economy," producing raw material for other high added value production processes in different industrial fields, as silk industry biorefineries.
Bombyx mori larvae contain 54% of proteins, 8% of fat, 6% of fiber, 6% of ash and energy content of 390 kCal/100 g, even if the quality of the insect protein content has to be assessed (Blasquez, Moreno, & Martinez Camacho, 2012). The feeding of rats with silkworm pupae resulted in a lower chemical score in comparison with high-value proteins of egg or casein, probably due to the presence of an ecdysone that confers a bad odor to the insects. Other studies showed a high content of essential amino acids, such as 77% of Lys and 70% of Leu in pupae (Rumpold & Schlüter, 2013).
Bioactive peptides are peptides showing functional properties, as antimicrobial, antihypertensive, antioxidative, and immune-modulatory activities. They are produced from food matrices, as edible organisms, during gastrointestinal digestion and they can interact with human body as nervous, cardiovascular or gastrointestinal system. In addition, bioactive peptides can be used in food or cosmetic industry for antioxidative and antimicrobial activity (Lemes et al., 2016).
More research studies are requested to assess the protein profiling, the discovery of bioactive peptides and eventually allergens in silkworm pupae.
The aim of this work was to gain a deeper insight into the silkworm nutritional/functional quality of the components which can be extracted from insects, by the combined semiquantitative mass spectrometry-based approach.

| Silkworm pupae treatment
The B. mori pupae were obtained by breeding on mulberry leaf polyhybrid larvae, produced starting from four productive strains preserved at the CREA-API silkworm germplasm bank.
First crosses were obtained by crossing two Japanese and two Chinese strains. First crosses were then crossed between each other to yield the polyhybrid eggs. After incubation of the eggs, the newly hatched larvae were bred according to best practice in silkworm rearing (temperature, relative humidity, and photoperiod).
Afterward, pupae were air-dried at 60°C until constant weight.
Proteins were precipitated from a sample of silkworm pupae lyophilized at our disposal (provided by CREA-API company) using two alternative protocols: aqueous extraction (Yi et al., 2013) and Folch extraction (Milkovska-Stamenova & Hoffmann, 2016). Fifty grams of powder were dissolved in 150 ml of water and then the mixture was shuffled for 3 min by using a blender (Braun multiquick 3), to create a dispersion as much homogeneous as possible. This dispersion was sifted through a sieve with a porosity of 500 μm in order to remove any incompletely fragmented particles; 5 ml of the filtered sample was collected for the protein extraction, while the remaining 20 ml were further sifted on 125 µm pore size filter; and 5 ml were stored for protein extraction and the remaining 10 ml filtered on 40 µm pore size filter.

| Aqueous extraction
In aqueous fractionation, the dispersion was centrifuged (centrifuge Rotina 380 R; Hettich) at 15,000 g for 30 min at 4°C, after filtration. The three fractions, precipitate, supernatant, and fatty phase, were collected. To improve the purity, the fatty phase was aspirated with a syringe while the remaining fractions were further centrifuged. The final supernatant and the precipitate (pellet) were easily separated.
The pellet was dried under nitrogen flow and weighted, dissolved in the extraction buffer, formed by 50 mM Tris-HCl pH 8.0, 200 mM DTT, 0.3% SDS (Bio-Rad Laboratories) to obtain a concentration of 100 mg/ml, and incubated on a rotator overnight at room temperature. After the incubation the mixture was centrifuged at 14,000 g for 1 hr at 4°C, and the resulting supernatant and the aqueous supernatant were subjected to trichloroacetic acid (TCA) precipitation (10% w/v; Sigma-Aldrich). The solutions were mixed, incubated on ice for 30 min and centrifuged (10 min, 14,000 g, 4°C). The obtained precipitates were washed three times with icecold acetone (1:1 v/v), and finally suspended in solubilization buffer (50 mM Tris-HCl pH 7.2, 50 mM NaCl) to obtain a 100 mg/ml solution. Proteins were quantified by a colorimetric Bradford assay using BSA as standard.

| Folch extraction
Proteins were purified in duplicate from the sifted raw materials (at different pore size sieves: 500/125/40 µm) using Folch extraction (methanol/chloroform/water). Methanol (7.5 ml) and chloroform (15 ml; Sigma-Aldrich) were added to 1 ml of each sample and the mixture was continuously shaken for 1 hr at 4°C, and then cold water (12.5 ml) was added. After 10 min, the samples were centrifuged (10 min, 10,000 g, 4°C), the organic phase collected, and evaporated under vacuum. The remaining aqueous phase was centrifuged again (10 min, 10,000 g, 4°C), collected, and immediately dried under vacuum, while the protein pellets were dried under nitrogen flow. The dried protein pellets were weighted, dissolved in extraction buffer (50 mM Tris-HCl pH 8.0, 200 mM DTT, 0.3% SDS) to a final concentration of 100 mg/ml and incubated on a rotator overnight at room temperature. Then, all protein samples were centrifuged at 11,000 rpm for 1 hr at 4°C, and the resulting supernatant collected for TCA precipitation, as described above.
Proteins were quantified by a colorimetric Bradford assay using BSA as standard.

| One-dimensional analysis (SDS-PAGE)
Protein separation was performed under reducing conditions; aliquots of 25 µg of proteins were mixed with Laemmli sample buffer (containing 50 mM DTT) and denatured at 95°C for 5 min. Samples and the standard proteins mixture (Precision Plus Protein Standards, Bio-Rad Laboratories) were separated on precast gels (Any KD Mini Protean TGX; Bio-Rad Laboratories), at 200 V. Gels were stained using Coomassie blue (Bio-Safe G250; Bio-Rad Laboratories). Images were acquired using the GS-800 densitometer and analyzed by Quantity One software (Bio-Rad Laboratories).

| Protein in-gel digestion
Protein bands were excised from gels, finely chopped, and washed with 200 µl of 50 mM ammonium bicarbonate (Bio-Rad Laboratories).
Gel pieces were destained using solution of 25 mM NH 4 HCO 3 /ACN, 1/1, v/v (Sigma-Aldrich) following by acetonitrile (ACN). Proteins were reduced by 10 mM DTT at 56°C for 1 hr and then alkylated by 55 mM iodoacetamide at room temperature for 45 min in the dark.
In-gel digestion was performed by overnight-incubation at 37°C with 1 µg of sequencing-grade trypsin (Roche) in 50 mM ammonium bicarbonate. Then, the supernatants were collected and acidified with formic acid up to a final concentration of 1%. And the peptides were extracted following the method of previous published paper (Colzani et al., 2016).

| Mass spectrometry analyses
The extracted peptides were analyzed by nLC-MSMS, using an nHPLC, UltiMate 3000 RSLCnano System, connected to an LTQ-Orbitrap XL mass spectrometer (Thermo Scientific Inc.). The pep-

| Semiquantitative analysis
To estimate the protein contents in the complex mixtures, a semiquantitative analysis, based on emPAI was applied (Ishihama et al., 2005). Briefly protein abundance index (PAI) represents the number of observed peptides divided by the number of observable peptides per protein. To calculate the number of observable peptides per protein, proteins were digested in silico, exploiting the PeptideMass tool in Uniprot website (http://web.expasy.org/cgi-bin/pepti de_mass/ pepti de-mass.pl?P02768), adjusted to simulate the experimental conditions. The number of observed peptides per protein, a method of counting unique parent ions was used (Ishihama et al., 2005).
The PAI is defined as (Equation 1): where N obsd and N obsbl are the number of observed peptides per protein and the number of observable peptides per protein, respectively.
For absolute quantitation, PAI was converted to exponentially modified PAI (emPAI), defined as follows (Equation 2): which is proportional to protein content in a protein mixture.
Thus, the protein contents in molar fraction percentages are described as (Equation 3): All the percentage values related to the protein contents were additionally normalized on the basis of the PSM (peptide spectral match) count, in order to consider the abundance of each peptide, strictly related to the number of spectra matched by the software during the data elaboration. (1) Equations 2 and 3 were modified as below reported:

| Generation of protein-protein interaction network
The protein-protein interaction network was built using all identified proteins and selecting B. mori genome as a reference data-

| Ranking of bioactive peptides
The B. mori proteome was "in silico" digested by stomach (pepsin) and intestinal (trypsin, chymotrypsin, elastase, carboxypeptidase A and B, and aminopeptidases) enzymes. The analyses were performed by    Table S1). This list of proteins was used as starting point for functional analyses, as above described.

| Qualitative and semiquantitative protein analysis by mass spectrometry
An estimation of the protein contents in the complex mixtures was obtained by calculating the values of emPAI, normalized by the PSMs values, and add an important quantitative information to proteomic results here shown. Figure  PMS 81; Zuo et al., 2015). The protein content normalized of these proteins was lower than the value of the main abundant class of proteins, such as apolipophorin (PMS 1776) and sex-specific storage protein 1 (PMS 668; Table S1).

| Generation of protein-protein interaction network
The protein-protein interaction network here reported ( Figure S2) represents  Figure S3). While, based on KEGG's database information, the main pathways in which the silkworm proteins are involved were elucidated, as several metabolic, ribosomal, and proteasome activity pathways ( Figure S4).

F I G U R E 3
Semiquantitative analysis. The pie chart shows the percentage distribution of reported class of proteins in silkworm pupae processed sample Beside the graphical representation of the clustering analysis (interact-omics map), the entire proteome has been further inspected, to classify proteins on the basis of their functional annotations (Gene Ontology Annotations), spotlighting the main molecular functions and confirming the interesting biological processes exerted by silkworm pupae ( Figure S1). In "biological process" enrichment analyses ( Figure S1-panel A), the largest clusters include proteins involved in biosynthetic/metabolic processes and in the response to stress, data confirmed by observing the KEGG pathway classification ( Figure   S1-panel D). In "molecular function" enrichment analyses ( Figure   S1-panel B), the majority of proteins are involved in protein and metal ion binding, while fewer proteins are involved in nucleotides, DNA, and RNA binding. It is also possible to find a great percentage of proteins with catalytic activity. In "intra/extra cellular localization" analyses ( Figure S1-panel C), high percentage of cytoplasmic proteins, extracellular proteins, and ribosome proteins were found.
Interestingly, the enrichment based on the PFAM domain annotations ( Figure S1-panel E), confirms the results of semiquantitative analysis.
In particular, lepidopteran low molecular weight lipoproteins, serpins, and hemocyanins were the most enriched protein classes.

| Analysis of bioactive peptides after gastrointestinal digestion
To identify bioactive peptides in Bombix Mori proteome, the most abundant identified proteins (top 19 proteins in Table S1) were digested "in silico" with stomach and intestinal enzymes. The predicted peptidome resulted in 141 peptide sequences that were investigated to find their potential functional activity. Twenty-three bioactive peptides were ranked by peptide ranker with score higher than 0.5, using N-to-1 neural network probability. Antimicrobial peptides (AMPs) were selected by applying the Discriminate Analysis Classifier score, using CAMP database and considering the geometric means of the Bioactivity Probability Score and the AMP Probability (  The identified peptide sequences as potential functional peptides have to be validated by further bioactivity assays, using correspondent synthetic peptides. However, these new computational methods are helpful in discovery phase of bioactive peptides. In fact, they are fast and low-cost alternatives, able to predict and reduce the number of potential targets to be investigated. In addition, bioinformatics-driven tools provide useful insights not achievable in human or animal model studies.

| D ISCUSS I ON
This is the first study to deeply characterize the high-value pro- First, sex-specific storage protein 2 (SSP2), homologous to SP2, is a heat-resistant protein, able to maintain its biological activity until 80°C. SSP2 have antiapoptotic activity in different cells (Yu, Wang, Zhang, Quan, & Zhang, 2013). In addition, silkworm pupae protein, digested by gastrointestinal endopeptidases, has anti-inflammatory effect in cells and generates a novel angiotensin-I-converting enzyme (ACE) tripeptide inhibitor of sequence ASL, useful in blood pressure control (Qiongying, Junqiang, Hui, Jinjuan, & Zhongzheng, 2015). Finally, B. mori releases immune effectors peptides in the hemolymph, after bacterial infections (Romoli et al., 2017).
In this study, peptidomic investigation, supporting by protein profiling, highlighted the presence of several bioactive peptides, originating from edible insects, with antimicrobial activity but also other potential functionality showed by bioinformatic tools.
This analytical approach allowed the selection of potential functional peptides, reducing the number of targets to be investigated by further assays and finally provide useful insights not achievable in human or animal model studies.
The protein profiling showed also the presence of known allergens. Arginine kinase (Bomb 1), an important enzyme involved in growth and development, is the first allergen of silkworm larvae, discovered by using 10 patient plasma . Another allergen, reported in literature and quantified in this study, is the 27 kDa glycoprotein. This protein, produced in E.Coli, was tested against 15 patient sera and only the 33% reacted (Zhao et al., 2015). The quantified chitinase and paramyosin proteins are also known allergens, verified by immunoproteomics approach ( Jeong et al., 2016). The thiol peroxiredoxin protein of B. mori pupae, instead, induces asthma in human, modulating the functions of macrophage cells (Wang et al., 2016). Finally, pupae Bom m 9 (lipoprotein PBMHP-6) induces asthma in mouse model (Zuo et al., 2015). However, the number of severe outcome due to these allergens and the number of allergic patients is quite limited, due to the low consumption of this insects and the cross reactivity with other allergens that may influence the results. Further studies are needed, and this semiquantitative protein profiling may help in this discovery phase.
This work demonstrated the high value of protein content in B. mori pupae, by showing the presence of functional peptides in the gut, by miming the enzymatic digestions. These peptides can have antimicrobial, but also anti-inflammatory, antioxidative and they can be modulators of immune system. The B. mori pupae protein content can be considered functional food ingredients in baked goods, snack products, and meat analogs. The derived bioactive peptides can be produced by sustainable and environmentally friendly processes.
In conclusion, B. mori pupae, even if are waste product of silk industry, can have industrial applications as source of high-value proteins and bioactive peptides in a green and "circular" economy perspective.

ACK N OWLED G M ENT
We thank Dr Cappellozza S. of CREA-API, Via Eulero, 6A -35143 Padova (Italy), for providing us samples of the Bombyx mori pupae.

CO N FLI C T O F I NTE R E S T
The authors declare that they have no conflict of interest that could have influenced the work reported in this paper.

E TH I C A L A PPROVA L
The study complies with institutional standards on silkworm research and it was conformed to Directive 2010/63/EU.

R E FE R E N C E S
Altomare