Comparative Proteomic Analysis of Bacillus subtilis and Aspergillus niger in Black Soldier Fly Co-Fermentation

Abstract

Black soldier fly larvae have gained popularity as an organic waste bio-conversional tool and fodder protein replacement in recent decades. It can consume all kinds of animal feces, kitchen waste and agricultural waste with great efficiency and transform them into high-value insect protein, fatty acids, and amino acids, which makes the larva a good substitute for costly fish meal and bean pulp in animal diets. However, excess chitin in the larva skin limits its application as an animal feed additive, consequently, employing fermentation with zymocytes to remove the chitin is necessary. In this study, we raised black soldier fly larvae (BSFL) with different carbon sources, such as chicken feces, straws and glucose, and examined the growth condition; we applied Bacillus subtilis and Aspergillus niger to co-ferment BSFL paste to analyze its nutrition changes. Data revealed that among the four kinds of cultures, the body weight of the corn powder group increased most rapidly; the wood chip group was the most underweight; however, it increased faster than others before day 4, and contained the least fat. Label-free quantitative proteomic analysis revealed that the expression of multiple enzymes from B. subtilis and A. niger involved in polysaccharide hydrolysis, amino acid biosynthesis and fatty acid metabolism, such as peptidase of S8 family, maltogenic α-amylase, oligo-1,6-glucosidase and lysophospholipase like protein changed significantly compared to the control group. Production detection showed that free amino acids, acid-soluble proteins, and short-chain fatty acids increased after fermentation; 13 out of 17 amino acids were increased and total free amino acids were increased from 0.08 g/100 g to 0.3 g/100 g; organic acids increased by 4.81 to 17 fold through fermentation, respectively; the actual protein content declined from 3.03 g/100 g to 1.81 g/100 g, the peptide content increased from 1.3 g/100 g to 2.46 g/100 g, the chitin degradation rate was 40.3%, and fat decreased 30% (p < 0.05). These findings might provide important information for future applications of black soldier fly larvae in different carbon waste recycling measures and material for animal feed/organic fertilizer after fermentation.

1. Introduction

With the rapid development of the global economy and society, environmental pollution caused by increasing organic waste is becoming more and more serious. In order to solve this problem, scientists developed several ways to convert biowaste into resources, among them, treating waste with insect larvae is more efficient and productive than conventional composting or incineration. Black soldier fly stands out for its high efficiency and security. After consumption by BSF larva, the water content and dry matter of kitchen waste were reduced to 15% and 50% [1]; the bioconversion rates of swine and chicken manure are 50% and 56%. All the biomass BSF consumed is transferred into nutrient substances in the larva; it contains more than 40% protein, and 30% fatty in the dry matter [2], with a similar amino acid composition to fish meal (FM) [3], which make it a promising new source of fodder protein [4,5,6]. Scientists have discovered that including a certain amount of BSF larva in the diet does not affect animal growth [7,8,9] or immunity [10,11,12,13]. Experiments on catfish found that a certain level of BSFL substitution in diets had no negative effect on growth rate and nutrient utilization ability; similarly, adding defatted BSFL as supplementation in rainbow trout diet has no negative effects on fish physiology or physical condition [14]. BSFL antibacterial peptides were also proved to be capable of inhibiting harmful pathogens among animals [15,16].

The benefits of replacing FM with BSFL have been proved with various domestic [17,18,19,20] and aquatic animals growth conditions [21,22,23]; however, there are still a few challenges to overcome, such as the anti-nutritional component chitin, palatability, and essential amino-acid insufficiency. These obstructions, especially chitin, severely limited the utilization of BSF as FM substitution in the feed industry. A study on rainbow trout revealed that 15% BSF in the diet is the maximum addition for fish growth [24]; adding 25% BSF into catfish meal is the maximum of having no negative effect on growth rate and nutrition utilization efficiency [25]. Previously, we found that the content of chitin in the BSF larva cuticle rose from 3% to 84% from larva to prepupa. The small amount of chitin would stimulate intestinal gurgling, accelerate nutrition metabolism, and food digestion [26]; ingesting excessive chitin would lead to epithelial microvilli damage, as it possesses a strong protein binding ability that would affect protein digestion and interfere with the activity of the brush border enzyme and leucine aminopeptidase. Intestinal histopathological damage was observed when the substitution rate of BSF defatted powder exceeded 75% in feeding Jian carp juveniles; data showed that chitin may also have a negative influence on the liver tissue of aquatic animals [27].

According to reported studies, mealworm fermentation extract offers a higher percentage of free amino acids than non-fermented mealworms [28,29]. Aspergillus oryzae degraded 31.6 % of maggot protein into various molecular size peptides, significantly extending the maximum life span of fruit flies. It is often believed that insect protein should be fermented with enzymes rather than microorganisms since the high protein content creates a tough environment to control bacteria, making the use of zymocyte in BSFL pulp a challenge [30]. However, fermentation with enzymes has several disadvantages: to breakdown diverse nutrients, a variety of enzymes would be required, considerably increasing the cost; BSF larvae grow in a putrid biomass environment with a plethora of microorganisms; enzyme fermentation cannot effectively disinfect or limit the development of harmful microorganisms, and enzyme degrading efficiency decreases rapidly if the substrate is not specific. Bacillus subtilis and Aspergillus niger can secrete various kinds of enzymes to decompose diverse nutrients including chitin, protein, and fatty acids; the growth of pernicious bacteria in the system would also be suppressed. As a result, utilizing microorganisms to ferment BSFL pulp to eliminate anti-nutritional factors and create prebiotics is a promising strategy [31]. In this work, we used chicken manure, glucose, wood chip, straw and corn powder to feed the BSFL to examine the ability of BSFL of utilizing different carbon sources; after that we used two types of zymocytes to degrade macromolecular substances such as chitin. Quantitative proteomics analysis was used [32,33,34] to compare the control group with the co-fermentation group of Bacillus subtilis and Aspergillus niger and to find out differential protein expression in carbon, amino acid, and fatty acid metabolism. A variety of essential enzymes that play critical roles in the co-fermentation were identified by KEGG pathway analysis.

2. Materials and Methods

2.1. Larval Rearing Conditions

BSF eggs were collected from our BSF base in Dingan, Hainan province, China, and hatched in the lab. Larvae were placed in plastic boxes with openings covered with nylon web, all larvae were separated into four groups, glucose, wood chip, straw and corn powder, each group was replicated three times. The rearing temperature was 28 °C. Insect diets were prepared with chicken manure, corn powder, glucose, wood chips and straw powder, all diets contained the same C/N and moisture rate; glucose was used as the control (CK), and the details of larva diets are shown in Table 1.

Table 1. Ingredients of diets.

	Glucose(CK)	Straw	Wood Chip	Corn
Chicken manure(g)	1405	1180	1528	560
Glucose(g)	281	/	/	/
Straw(g)	/	332	/	/
Wood chip(g)	/	/	253	/
Corn(g)	/	/	/	547
Water(ml)	320	490	220	900
C/N	25	25	25	25
Moisture(%)	70	70	70	70

2.2. Strains and Kits

Bacillus subtilis 10071 and Aspergillus niger 2106 were purchased from the China Center of Industrial Culture Collection (CICC); the subculture medium for Bacillus subtilis was LB at 37 °C, whereas for Aspergillus niger was YPD at 28 °C under aerobic condition. BSFL was collected and washed before being pulverized into a paste with a colloid mill; the particle diameter was under 100 nm. The BSFL medium consisted of 30 g BSFL paste and 70 mL distilled water sterilized at 120 °C for 20 min. Chitinase activity was tested with chitinase activity assay kit AKSU045M, purchased from BOXBIO company; oligosaccharide was tested with DNS agent; chitin was tested with NADP-GAPDH plant chitin kit (ELISA), purchased from FanTai Biotechnology Co., Ltd., Shanghai, China.

2.3. Orthogonal Design of Co-Fermentation

Our earlier experiments tested the concentration of reducing sugar under different pH, fermentation temperature, inoculation proportion, time, and amount. The optimum pH and temperature of Bacillus subtilis and Aspergillus niger co-incubation were 7 and 30 °C, respectively, after confirming these two conditions with our test, we selected inoculation proportion, timing, and quantity as variable factors to design an orthogonal test; the factor levels were B. subtilis vs. A. niger 1:1, 1:2, and 1:3. The inoculation amounts were 4%, 8%, and 12%; the incubation time was 2 days, 3 days, and 4 days. BSFL medium was prepared with 30 g larva powder in 70 mL water and 121 ℃ sterilization; the volume of the medium was 500 mL in 1 L flask, with 200 rpm shake cultivation. The orthogonal design of L₉(3³) is shown in Table 2.

Table 2. Orthogonal design.

Subject	Inoculation Proportion	Inoculation Amount (v/w)	Inoculation Time
1	1:1	4%	2d
2	1:1	8%	3d
3	1:1	12%	4d
4	1:2	4%	4d
5	1:2	8%	2d
6	1:2	12%	3d
7	1:3	4%	3d
8	1:3	8%	4d
9	1:3	12%	2d

2.4. Protein Extraction

The fermentation product was collected in a 50 mL centrifugal tub using 5000 rpm centrifugation, and the supernatant was removed. Lysis solution (8M Urea/100 mM Tris-Cl) was added to all sediment before ultrasonication; after that, the samples were incubated at 37 °C for 1 h, then IAA was added; an alkylation reaction was induced at room temperature in the dark. The Bradford method was applied to test the concentration of protein; 100 mM Tris-HCl was added to the samples to dilute the urea under 2 M; trypsin was added at a ratio of 1:50 (w/w) into the samples at 37 °C then incubated and oscillated overnight. The following day, the enzyme reaction was terminated with TFA, and the supernatant was extracted and stored at -20 °C.

2.5. Mass Spectrometry and Data Analysis

Mass spectrometric data were collected with Orbitrap Exploris 480 HPLC EASY-nLC 1200 480 series. Peptides samples were combined with an analytical column (75 μm × 25 cm, C18, 1.9 μm, 100Å) to be separated. Two mobile phases were applied to perform gradient analysis, with a flow rate of 200 nL/min (A: 0.1% formic acid, B: 0.1% formic acid, 80% ACN). Results were collected in DDA mode, each SCAN-ENDSCAN contained a full MS scan (R = 60 K, AGC = 300%, max IT = 20 ms, scan range = 350-1500 m/z) and 20 MS/MS scan (R = 15 K, AGC = 100%, max IT = auto, cycle time = 2 s). The HCD collision energy was set to 30.

The mass spectrometry data were retrieved with MaxQuant V1.6.6, Andromeda search algorithm, and Uniprot Proteome Bacillus_subtilis. Aspergillus_niger Proteome reference databases were used. Proteome Bacillus subtilis, Aspergillus_niger, quantification method LFQ; variable modifications were Oxidation (M), Acetyl (Protein N-term), Deamidation (NQ); fixed modifications were Carbamidomethyl (C); enzyme digestion used Trypsin/P. To study and annotate the function of the protein, BLAST was applied to compare the annotation of GO, KEGG, and COG; the highest score was selected to be the final result and domain annotation was based on Uniprot and InterPro. All protein sequences were identified based on NCBI nr, UniProt, BioGRID, Database of Interacting Proteins, and MINT.

2.6. Product Detection

Free amino acid detection: protein was detected with automatic Kjeldahl apparatus(K-375, BUCHI, Switzerland); the amino acids sample was dissolved with 0.02 mol/L hydrochloric acid and centrifuged for 5 min at 6000 RPM. After 20 min of ultrasonic treatment, the supernatant was purified with a C18 purification column. After 2 h of drying at 60 °C in a vacuum drying oven, it was reacted for 30 min at room temperature with a derivatization reagent; the sample was ready after filtration through a 0.45 μm organic membrane. The chromatographic column used was SHISEIDO SG300 with a C18 filler (4.6 mm × 250 mm × 5 μm), input amount 10 μL, column temperature 40 °C, wavelength 254 nm.

Fatty acids detection: the sample was dissolved with hydrochloric acid after adding pyrogallic acid, zeolite, and 95% ethyl alcohol. The hydrolyzed sample was washed with an ether-petroleum ether mixture and dried for 2 h at 100 °C. After adding sodium hydroxide, methanol and n-hexane before the volatile extraction, the extract was heated at 85 °C for 30 min; the sample was ready after filtering through a 0.45 μm organic membrane. The chromatographic column Thermofisher TG-5MS with TraceGOLD™ filler (30 m × 0.25 mm × 0.25 μm) was purched from Xiyan Scientific Instrument Co., LTD., Shanghai, China, input + temperature was 290 °C, flow rate 1.2 mL/min, scan range 30–400 amu, and ion source EI 70 eV. The fatty acid calculation equation is as follows:

$$\mathrm{W}=\frac{\mathrm{C}\times \mathrm{V}\times \mathrm{N}}{m}\times \mathrm{k}$$

(1)

• W—fatty acid amount, (mg/kg):
• C—concentration of fatty acid methyl ester in the sample, mg/L:
• V—volume, mL:
• k—conversion ratio from fatty acid methyl ester to fatty acid
• N—dilution ratio:
• m—sample weight, (g)

The basic composition of dried matter was determined by standard methods (AOAC, 1995).

2.7. True Protein Detection

True protein detection: CuSO₄, K₂SO₄, and H₂SO₄ were added into the sample before 1h digestion at 420 °C in a furnace, sediment was dried and moved into a Kjeldahl flask, and the sample was tested using semi-trace Kjeldahl nitrogen determination. The sample was titrated with a standard solution. The equation is as follows:

$$\mathrm{X}=\frac{{(\mathrm{V}}_1-{\mathrm{V}}_2)\times c\times 0.014}{m\times {\mathrm{V}}_3/100}\times \mathrm{F}\times 100$$

(2)

• X—true protein concentration, g/100 g
• V1—standard HCl volume, mL
• V2—control group standard HCl volume, mL
• V3—volume of digestive fluid, mL
• c—concentration of standard HCl, mol/L
• m—sample weight, g
• F—conversion factor between N and protein.

3. Results and Discussion

3.1. Growth Performance of BFL in Different Cultures

After 16 days of breeding, the growth condition of BSFL has shown that compare with the CK glucose group, the bodyweight daily (BD) of the wood chips and straw groups decreased by 31% and 23%, respectively, and corn increased by 15% (p < 0.05) (Figure 1); before day 12, the BW of corn powder increased faster than the other three groups; since day 4, the bodyweight (BW) of the corn group exceeded that of others, and the glucose group was the lightest until day 10; the BW of the fiber group remained at 0.09 g since day 8. This result proved that the ability of BSFL to utilize the carbon source in corn is better than in wood fiber; glucose was the most easy carbon source to absorb; however, the final BW of corn, not glucose, was the heaviest of all, indicating that BSFL utilized starch, maltose and fructose better than other carbon sources; this result consistent with another study [35], under the same moisture content and C/N, the increased rate and BW of corn group was higher than others; straw and wood chips both contain substantial fiber which is clearly a challenge for the larva to digest, in this respect BSFL is obviously not as strong as an earthworm. The experiment was terminated when 50% of larvae began pupation; straw and wood chip groups began pupation on day 12, the corn group on day 14 and the glucose group on day 16. Figure 2 showed an interesting phenomenon, the protein and fat demonstrated contrary tendencies, in the glucose group, protein content is approximately 22.05% and in the wood chip group it is 33.17%; with the carbon utilization difficulty increased, the fat percentage decreased and protein increased. Such tendency showed that fat was produced preferentially in BSFL, therefore, the production of protein declined. This result indicated that under the condition that carbon source is difficult to utilize, such as in straw and wood chips, BSFL grows more protein than fat, and the period of growth is shortened; the corn group outweighs the other groups.

Figure 1. Growth conditions of four groups of BSFL after 16 days of breeding.

Figure 2. Crude protein and fat percentage of 4 groups after 16 days of incubation.

3.2. Optimum Co-Fermentation Conditions of B. subtilis and A. niger

We chose the corn group BSFL, as the fermentation substrate for BSFL in straw and woodchip groups terminated the larva period earlier due to the carbon source being difficult to utilize. The glucose group contained more fat than average (30%); the larva contained 6% chitin among dry matter which is different from other research [36]. Bacillus subtilis and Aspergillus niger were separately cultured with BSFL medium. Results showed that after 3 days of culture, the concentration of oligosaccharides in B. subtilis and A. niger supernatant was approximately 0.03 and 0.05 μmol/mL, respectively. Because the BSFL mud has no carbon sources other than chitin in the larval skin and peritrophic membrane, the reducing sugar can only be a byproduct of chitin breakdown by B. subtilis and A. niger. After single-factor experiments, shown in Figure 3, we designed an orthogonal experiment to study the most important factor that would affect chitin degradation. Among all nine groups, the oligosaccharide concentration of group 5 was the highest (Table 3), which proved that the optimum conditions were fermentation temperature 30 °C, inoculation quantity 8%, volume ratio B. subtilis: A. niger = 1:2, and incubation period 4 days. According to the value of range R in Table 3, the inoculation time exerted the most significant effect on degradation efficiency, the order of importance that influenced the degradation was found to be time > amount > proportion. Under these conditions, the oligosaccharides of co-fermentation were produced in a concentration of 1.54 μmol/mL, and chitin content decreased by 40.3%, from 6% to 3.5% in dry matter (p < 0.05). This result revealed that both B. subtilis and A. niger are capable of degrading insect chitin to obtain carbohydrates; the chitin hydrolyzing ability of A. niger is higher than B. subtilis. The finding that co-fermentation produced more oligosaccharides than each of them suggested that chitinase from two zymocytes might cooperate on a certain level; K_i represents the sum of the corresponding test results when the level number on any column is I and R represents the range.

Figure 3. Single factor experiments of co-fermentation. (A-D) are Oligosaccharides concentration under different temperature, pH, inoculation amount and inoculation ratio.

Table 3. Orthogonal intuitive analysis.

Subject	Inoculation Proportion	Inoculation Amount	Inoculation Time	Score
1	1	1	1	0.92
2	1	2	2	1.02
3	1	3	3	1.21
4	2	1	3	0.92
5	2	2	1	1.54
6	2	3	2	0.87
7	3	1	2	0.97
8	3	2	3	1.06
9	3	3	1	0.98
K1	1.050	0.937	0.095
K2	1.110	1.207	0.973
K3	1.003	1.020	1.240
R	0.107	0.270	0.290

3.3. Proteomic Analysis of B.subtilis and A.niger after Co-Fermentation

Figure 4A, a volcano plot and histogram show the enzyme expression changes between the control group (MCK) and co-fermentation group (MF) of B. subtilis and A. niger under optimal conditions. These results (Figure 3B) indicated that the expression of 403 proteins was significantly changed following fermentation, with 209 proteins being upregulated and 194 downregulated. The genetic and biological functions of these proteins were cataloged by GO (Gene Ontology) and KEGG (Kyoto Encyclopedia of Genes and Genomes) pathway analysis. Since there were both bacteria and fungus in the system, both COG (Cluster of Orthologous Groups of proteins) and KOG were required. In Figure 5A, the PCA result shows the total differences between groups of MF (mix fermentation) and MCK (mix control) and the degree of variation between samples within groups. This result revealed that the variations within and between MCK and MF are acceptable. In Figure 5B, the heat map shows the expression differences for significant proteins between MCK and MF groups.

Figure 4. Volcano plot and function analysis of differential expression protein from MCK and M; (A) Volcano plot of proteins identified between MCK and MF; (B) numbers of proteins belonged to metabolism.

Figure 5. PCA and heat map of MCK and MF. (A) PCA explained 87.67% of the total variation; (B) Longitudinal is the clustering of samples and transverse is proteins.

3.4. Carbon Metabolism

According to KEGG pathways of starch and sucrose metabolism, MCK vs MF, the expression levels of multiple proteins were increased, As shown in Figure 6A, starch metabolism showed maltose-6’-phosphate glucosidase, trehalose-6-phosphate hydrolase, α-amylase, maltogenic α-amylase, maltogenic α-amylase, oligo-1,6-glucosidase, β-glucosidase, and glycogen phosphorylase. These proteins are all polysaccharide hydrolases, glycosidic bond hydrolases, or ATP/GTP synthases: phosphate glucosidase and oligo glucosidase can hydrolyze glycosidic bonds, and degrade long-chain polysaccharides into oligosaccharides. Maltose, trehalose, and starch are polysaccharides as is chitin, which can be degraded by both specific and nonspecific enzymes, if there are no other polysaccharides detected in the culture system, it is obvious that B. subtilis and A. niger activated several lyases in different carbon metabolism pathways to breakdown chitin to obtain chitooligosaccharides as an energy source and fungal cell wall material. Meanwhile, the expression level of glycosynthases was either unchanged or declined. As shown in Figure 6B, galactose metabolism indicated that apart from the enzymes mentioned above, the expression of hexokinase, UDPglucose-hexose-1-phosphate uridylyltransferase, and galactokinase was also upregulated, which are both key enzymes in the lactose metabolic pathway.

Figure 6. Partial starch (A) and galactose (B) metabolism map between MCK and MF. Proteins in red frame are upregulated, green frame are downregulated, blue frame are both down and upregulated.

Chitinase is structurally expressed at a certain level in bacteria and fungi, and these chitinases can produce small amounts of oligosaccharides that can be transported into the cell via the ABC transporter system to induce chitinase expression. Chitin and its decomposed products are the most effective inducers of chitinase in microorganisms. Glucose could not only inhibit chitinase but the expression of almost all polysaccharide hydrolases through to the metabolic repression mechanism, lack of glucose triggers starvation which can revoke this suppression, Figure 7A shows that there was no inhibiting effect upon oligases. In addition to starvation, the chitin in the cell wall of A. niger and BSFL also played a key role in revoking the suppression of chitinase and other polysaccharide hydrolase expression. As shown in Figure 6B, the expression of many hydrolases and glycosidases was upregulated, revealing that both microorganisms mobilized multiple enzymes to degrade glycosidic bonds for short-chain oligos; at the same time, UTP-glucose-1-phosphate uridylyltransferase was downregulated both in starch and galactose metabolic pathways (Figure 6B), along with UDP-glucose 4-epimerase (galE); the expression of these isomerases between α-D-glucose-1P and UDP-glucose declined. Although UDP-glucose is a component of glucose synthesis, isomerase suppression suggests that glycogen synthesis is not a primary concern during fermentation. GalE is a critical part of dietary galactose metabolism, including galactose production and glycolipid synthesis, downregulated galE indicated that the glyconeogenesis pathway was inhibited during the fermentation, since obtaining a carbon source was imperative. Such phenomena can also be observed in the gluconeogenesis/glycolysis map (Figure 7A), from l-lactate to β-D fructose-6P, the whole synthesis pathway was found to be downregulated. In the tricarboxylic acid cycle (TCA) cycle, fumarate reductase and succinate dehydrogenase which transform succinate and fumarate into each other were upregulated; aconitate hydratase, isocitrate dehydrogenase, and 2-oxoglutarate dehydrogenase transform isocitrate into succinate. Succinate takes part in oxidative phosphorylation, which generates energy to support ATP synthesis, as shown in Figure 6B. It is obvious that the main purpose of zymophytes in the system is degrading organic molecules; therefore, the enzymes in the pathway between pyruvate and β-D fructose-6P were downregulated in the gluconeogenesis/glycolysis map.

Figure 7. Partial glycolysis/gluconeogenesis (A) and TCA cycle (B) between MCK and MF.

The B. subtilis fermentation group and control group (Supplementary Materials S1) showed that in the starch metabolism pathway, several enzymes such as endoglucanase, fructokinase, and β-glucosidase were downregulated, whereas the upregulation of cyclomaltodextrinase and other enzymes still ensured the production of D-glucose through chitin degradation. When compared with the A. niger KEGG pathway of starch, only two enzymes phosphoglucomutase and β-glucosidase were upregulated. β-Glucosidase (GHF1) is the enzyme responsible for breaking chitobiose into glucose, a typical chitinase/cellulose. The expression of GHF1 in A. niger fermentation was increased nearly six-fold (p < 0.01), and the downregulation of other irrelevant enzymes did not affect its chitin-degrading ability. As shown in Figure 8, In the general pathway of carbon metabolism, two zymophytes are centered on producing many products, such as glucose-6p, fructose-6p, citrate, oxaloacetate, aspartate, and succinate. The pathways from fructose-6p to pyruvate were all suppressed and fructose-1, 6-bisphosphatase I, and II were upregulated to preserve more fructose-6p. Fructose-6p is the precursor of fructose-1,6-bp, which is an important rate-limiting enzyme to embden meyerhof parnas(EMP) that can be inhibited by high concentration ATP and citric acid, and activated by ADP and AMP, it is a key step that affects the utilization efficiency of glucose (Figure 7B).

Figure 8. General KEGG map of carbon metabolism.

These findings showed that when both B. subtilis and A. niger made a concerted effort to use long-chain polysaccharides under starvation by silencing insignificant synthesis pathways and preserving small organic molecules for energy production such as TCA recycling and the synthesis of essential extracellular enzymes. The enzymes produce specific 6p-saccharides as an energy reserve to ensure sufficient acetyl-CoA in the TCA cycle and sustain vital activities.

3.5. Amino Acid Metabolism

A total of 17 different amino acids were examined, and all were found in the product except cysteine. In arginine biosynthesis, glutamate dehydrogenase, which transfers NH₃ and glutamate into each other was downregulated, and so is functional mitochondrial, cytoplasmic aspartate aminotransferase between 2-oxoglutarate and glutarate; in alanine, aspartate, and glutamate metabolism, the pathway from L-aspartate to fumarate was downregulated. As aminomethyltransferase in Glycine, serine, and Threonine metabolism, NH₃ production is suppressed in BSFL due to the high protein content. Since these three growth factors are easy to acquire in the BSFL mix, the expression of dihydrolipoamide dehydrogenase and 2-oxoisovalerate dehydrogenase in valine, leucine, and isoleucine degradation was decreased (Figure 9A).

Figure 9. Partial amino acids metabolism of Cysteine, Methionine(A) and Valine, Leucine and Isoleucine (B).

Meanwhile, the enzyme cystathionine gamma-lyase, which converts cystathionine to cysteine, was inhibited, causing L-cystathionine to concentrate on producing Methionine. Levels of synthases, such as phosphoserine aminotransferase, cystathionine gamma-lyase, and cytoplasmic aspartate aminotransferase, were downregulated in cysteine and methionine metabolism (Figure 9A). Transferases of methionine degradation, such as adenosylhomocysteinase and 5-methyltetrahydropteroyltriglutamate homocysteine methyltransferase, were also suppressed. As Figure 10A showed, the concentrations of VAL, ILE, and LEU were significantly increased, with VAL increasing more than four-fold (p < 0.05), and ILE and LEU increasing more than eightfold (p < 0.05), which matches the downregulated degradation condition; glycine, serine, and threonine increased three, five and sevenfold, respectively (p < 0.05). Cysteine was not detected for transferring into cysteine fast; the production of lysine increased by over 1000 mg and the Figure 10B correlation network showed that the enzymes which are responsible for amino acid production are peptidases from the S8 family, which exist in both microbes and aminopeptidase from M18 family in A. niger. All three amino acids that were produced less (Pro, ArG, Ala) shared the same enzymes that affect their production: Clp protease and serine peptidase S8 from A. niger are enzymes that increased the contents; serine peptidase S8 from B. subtilis and leucine aminopeptidase are enzymes that suppressed their production. From this figure, it is clear to us that Clp protease, termitase peptidase S8 from B. subtilis and serine peptidase S8 from A. niger inhibited amino acids most; these three enzymes happened to be positive enzymes for the three acids. After fermentation, peptides increased 189% from 1.3 g/100 g to 2.46 g/100 g, free amino acids increased 3.75 fold from 0.08 g/100 g to 0.3 g/100 g (p < 0.05), ASP was the sum of these two results. KEGG pathways of phenylalanine and tyrosine biosynthesis showed that most of the enzymes were unaffected; phenylalanine was undetectable in the MCK group and 9.44 mg was found in the fermentation group, whereas Tyrosine was 17-fold higher than in the MCK group.

Figure 10. Comparison of amino acid production and correlation network of enzymes responsible. (A) Production of 17 amino acids before and after fermentation (** represent significant difference, p<0.05); (B) the network of enzymes form B.subtilis and A. niger that is responsible for the production of the amino acids.

From the MCK vs MF amino acids biosynthesis map (Figure 11), it is clear that the pathway from fructose-6p to pyruvate was completely suppressed, the same as shown in carbon metabolism, whereas, the pathway from erythrose-4p to phosphoenolpyruvate was upregulated to compensate for the loss of metabolic pathways. Histidinol dehydrogenase was upregulated but Histidine increased just 1.46 fold (p < 0.05). Furthermore, crucial enzymes, such as threonine synthase and argininosuccinate lyase were upregulated, still, the concentration of arginine, threonine and alanine decreased to 97%, 15%, and 74%, respectively, as compared with the MCK group (p < 0.05). There were minor differences among A. niger, B. subtilis, and mix microbes MCK vs fermentation amino acid biosynthetic pathways; in the A. niger group, the pathways from β-D fructose 1,6-bp to pyruvate were not suppressed, yet the pathway from erythrose-4p to phosphoenolpyruvate was still upregulated. Methionine synthase, which transfers homocysteine to methionine, was upregulated, which may explain the high concentration of methionine in the product. However, the same enzyme was downregulated in the B. subtilis map. Not only in this pathway, but also in others, such as cysteine, glutamine, and ornithine synthesis, the B. subtilis map revealed more suppressive conditions than others. This may indicate that A. niger is responsible for generating most of the free amino acids in the product.

Figure 11. General KEGG map of amino acids biosynthesis.

These results revealed that A. niger possesses a greater ability in synthesizing free amino acids. The content of free amino acids increased 3.72 fold from 798.12 mg/kg to 2971.43 mg/kg (p < 0.05), which would make the larva mud easier to digest or absorb by the stock or plant. Since amino acids may be produced by biosynthesis and protein breakdown, the real protein content of BSFL mud decreased from 3.03 to 1.81 g/100 g (p < 0.05) following co-fermentation, A. niger played an important part in this process.

3.6. Fatty Acid Metabolism

From the KEGG maps of MCK vs MF fatty acid biosynthesis, elongation, and degradation (Figure 12A,B), it has demonstrated that mitochondrial enoyl-acyl-carrier protein reductase (MECR) involved in both fatty acid biosynthesis and elongation pathways was downregulated; at the same time, long-chain acyl-CoA synthetase and aldehyde dehydrogenase in fatty acid degradation pathways were upregulated (Figure 12C), the latter could transfer endogenous and exogenous sources of an aldehyde into the corresponding acid. Fatty acid synthesis II is the pathway by which most prokaryotes produce fatty acids, from Butyryl to Hexadecanoyl, and MECR is the last stage of this catalytic process as a biosynthesis catalyst. Downregulated MECR, together with acyl-CoA dehydrogenase employing acetyl-Coa to manufacture longer chain acyl-CoA from hexanoyl-CoA to hexadecanoyl-CoA, served as the final step enzyme in elongation metabolism. FabG and FabZ were downregulated in a general analysis of fatty acid biosynthesis in A. niger and B. subtilis (Supplementary Materials S2). FabZ functioned as a key dehydratase in fatty acid synthesis and elongation; it also plays the role of the regulator through ACP recognition and directional rotation. Owing to the unavailability of FabZ, the double bond of the substrate transported by ACP fails to form; thereby synthesis and elongation are unable to continue. FabG exists only in bacteria, it catalyzed the reduction of the acyl carrier protein, using NADPH as a co-factor to reduce 3-oxoacyl-ACP to 3-hydroxyacyl-ACP. Polyhydroxybutyrate (PHB) would be unable to synthesize if FabG was suppressed, halting the synthesis and elongation process. As shown in Figure 11D, fatty acid metabolism pathway analysis of A. niger and B. subtilis showed that acetyl-CoA C-acetyltransferase and 3-hydroxyacyl-CoA dehydrogenase were upregulated, along with butyryl-CoA dehydrogenase. 3-hydroxyacyl-CoA dehydrogenase (fadN) catalyzes β-oxidation of acyl-CoA fatty acids, and participates in fatty acid and amino acid degradation; butyryl-CoA dehydrogenase transfers isobutyryl-CoA to methylacryloyl-CoA.

Figure 12. Fatty acids elongation (A), degradation (B), biosynthesis (C), and general KEGG map (D).

Since 2006, Europe began to forbid the use of antibiotics, mid and short-chain fatty acids were recognized gradually for their antibacterial and antiviral ability. Short-chain acids could be utilized as a carbon source by other probiotics, reduce the number of E. coli in the animal cecum, and lower the pH of the internal environment to inhibit the growth of pernicious microbes. In BSFL oil, unsaturated fatty acids and essential fatty acids accounted for 60% and 23% of total fatty acids, including lauric, cardamom, palmitic, oleic, and linoleic acids; among them, 40% of fatty acid is lauric acid, which is a mid-chain dodecanoic saturated fatty acid that possesses important antibacterial and antiviral activities [37]; therefore, short-chain and lauric acids are indispensable to developing antimicrobial free feed. In this study, the expression of vital degradation enzymes was upregulated, and enzymes of biosynthesis and elongation were downregulated in the co-fermentation group may demonstrate the potential of BSFL as an antimicrobial, in the meantime, our test revealed that lauric acid decreased by 75% after fermentation. As Figure 13A showed, all six types of short-chain acids were detected; Figure 13B showed that long-chain acids, such as lauric acid (C12:0), myristic acid (C14:0), palmitic acid (C16:0), palmitoleic acid (C16:1) and oleic acid (C18:1) decreased obviously, especially lauric acid, palmitic acid and oleic acid; isobutyric acid and isovaleric acid increased the most; since the coefficient is 80%, there is only one enzyme that is the most responsible for the production lower fatty acids as Figure 13C has shown, protein A2QF42 is an unnamed protein with no description of its function, yet its amino sequence is 100% similar to lysophospholipase which is capable of hydrolyzing glycerolipid, this result proved that B. subtilis may possess greater ability in fatty acid degradation, however, A. niger expressed the most efficient enzyme in fat degradation. According to Figure 12B, since butanoyl-CoA, every last step of adding two more carbon atoms to acyl was suppressed; trans-2-enoyl-CoA reductase was downregulated, which may lead to the increment of short-chain acids. However, in B. subtilis fatty acid metabolism, 3-hydroxyacyl-[acyl-carrier-protein] dehydratase, enoyl-[acyl-carrier protein] reductases I, II, and III were upregulated along with trans-2-enoyl-CoA reductase. From another perspective, B. subtilis was shown to be more proficient in the synthesis of short-chain acids.

Figure 13. Fatty acids production of MCK, MF (A); degradation of long chain acids (** represent significant difference, p < 0.05) (B); and correlation network of enzymes responsible (C).

These findings indicated that compared with A. niger, B. subtilis was more active in fatty acid degradation, in the BSFL co-fermentation system, under the condition of surplus fatty acids existing. Microorganisms have abundant fatty acids to be utilized, and the crucial enzyme expression in fatty synthesis was downregulated; therefore, the elongation and biosynthesis pathways were suppressed. At the same time, the degradation pathway was unaffected. The downregulation of MECR in this circumstance demonstrated that, with enough fatty acids to metabolize, breaking down fat for energy is the most pressing need for microorganisms [38], as a result, after the fermentation the free short chain fatty acids increased by 25-fold totally, and fat decreased 30% (p < 0.05), however, how to preserve specific acids such as lauric acid still needs further study.

4. Conclusions

This is the first study that applies microbes to ferment black soldier fly larvae to produce oligosaccharides, peptides, and short-chain fatty acids other than using enzymes exclusively. In this study, several carbon sources were applied to examine the ability of insects to utilize carbon sources with different difficulties; after that two zymocytes used their respective advantages to conduct a series of fruitful transformations, biosynthesis, and degradation during the fermentation procedure. Upregulated glycosidases, polysaccharases, mutases, and proteases might break the long-chain polysaccharide, fatty acids and proteins into short-chain oligomers and specific fatty acids; downregulated reductases, epimerases, and transferases may prevent the production of non-essential and redundant metabolites. Co-fermentation led to an increase in amino acids, fatty acids, ASP, and peptides. In conclusion, the co-fermentation of A. niger and B. subtilis has an important impact on BSFL protein and fatty degradation, and the biosynthesis of small molecule nutrients. This study lays the groundwork for future research on BSFL-fermented nutrition in the fodder industry, as well as the implications of co-fermentation on insect biomass. In addition, it provides an initial analysis of the selection of zymophytes to improve the BSFL absorption rate.