**3. Results**

#### *3.1. Lingzhi-Derived Protein Hydrolysate Optimization*

The biological activity of the hydrolysates depends on the processing conditions. The activities of various foodstuff hydrolysates were reportedly directly dependent on the degree of hydrolysis, protease activity, and amino acid arrangemen<sup>t</sup> [23]. The optimum conditions for the Lingzhi hydrolysate regarding DH and product yield for functional food product manufacturing have not ye<sup>t</sup> been established. Therefore, the present study was aimed at Lingzhi hydrolyzing proteins using RSM to study the effect of the processing conditions including time, enzyme usage on DH, and product yield of the resulting hydrolysates. We applied quadratic analysis statistics to fit an RSM model for independent variable factors. The experimental design using two independent variable factors with two center points (experiment no. 10 and 11) in RSM generation resulted in the observed DH and yield as displayed in Table 1. The RSM generation-related statistical value is shown in Appendix A.


**Table 1.** The experimental design and experimental outputs of the independent factors for the degree of hydrolysate and yield produced from Lingzhi proteins.

As outputs from the overall experimental design, the DH and product yield ranged from 28.11% ± 1.03% to 34.18% ± 1.12% and 4.16% ± 0.13% to 5.70% ± 0.20%, respectively. The difference in the DH and yield could be due to the difference in the digestion time and enzyme concentration. The equation for multiple regression analysis during the RMS was performed to resolve the coefficients of the independent factors of the linear (x1, x2), quadratic (x12, x22), and two-factor relation (x1×2) to fit the RSM. According to the multiple regression analysis, the explanatory model equation of the DH (y1) and percentage of product yield (y2) is given as follows in Table 2.

**Table 2.** The experimental design and experimental outputs of the independent factors for the degree of hydrolysate and yield produced from Lingzhi proteins.


The total coefficient value (R2) was used to imply the model suitability. The R<sup>2</sup> of the DH and the product % yield were 0.958 and 0.968, respectively. This result indicated that the variation in the experimental data was lower than 5% (within 95% level of confidence). The 3-dimensional response model surfaces (3D-RMS) for each variable are illustrated in Figure 1.

**Figure 1.** 3D-RMS plots showing the interactive effects of different factors on DH and yield. (**A**) DH of Lingzhi protein hydrolysate on digestion time versus enzyme usage, and (**B**) yield of Lingzhi protein hydrolysate on digestion time.

The experimental outputs of the processing related to both independent factors, DH (Figure 1A) and % yield (Figure 1B) indicating that the hydrolysate processing depended on the digestion time and enzyme usage. The 3D-RMS for the DH of hydrolysate as a function of digestion time, at fixed enzyme usage, revealed that DH was dependent on the digestion time. Also, DH increased with enzyme usage at the fixed digestion time, suggesting that DH was also dependent on the enzyme usage. Yield also had correlative results, dependent on the digestion time and enzyme usage. In order to obtain the highest DH and product yield, the RSM model was optimized by setting the highest value of response variable factors. As a result, X1 was 8.63 h and X2 was 0.93%, and the highest values of y1 and y2 were 33.99% and 5.70%, respectively. These characteristics of DH and yield curves were associated with feedback inhibition during the hydrolysis, where products may act as an inhibitor to protease [24]. The curves strongly suggested that the processing at different conditions and factors were involved. The independent factors, both time and enzyme concentration, had the optimum range for hydrolysate production to gain the maximum DH and yield. To endorse the reliability and validity of the model for processing, the assays were performed

under those optimal conditions. The actual experimental values for DH and product yield were 32.71 ± 0.17% and 5.44 ± 0.14%, respectively; the experimental values fitted with the values that were predicted by the model within a 95% confidence interval. These results confirmed that the model was suitable for Lingzhi protein hydrolysate processing for use as functional ingredients regarding cost- and time-efficiency.

#### *3.2. Encapsulation Efficiency and Loaded Liposome Size, Polydispersity Index, and Zeta Potential*

The encapsulation efficiency of the liposomal formulation was estimated. The liposomes would passively entrap the protein hydrolysate in their hydrophilic region. However, many factors influence the entrapping efficiency such as lipid molar ratios, molecular size, charge, and molecule stability. To evaluate the entrapping efficiency, we used a non-ionic detergent, Triton X-100, as a neutral detergent to disrupt the liposome shell structure, thereby allowing the leakage of the encapsulated Lingzhi protein hydrolysate [25]. Based on the encapsulation condition, 61.24 ± 3.18% of the encapsulation efficiency was achieved. The encapsulation efficiency showed that the liposomal preparation for protein hydrolysate moderates the encapsulated level. The protein hydrolysate has a mixture of peptides with a variety of molecular weights, sizes, charges, and structures. Middle-sized peptides might interact with the lipid layer and form an oligomerization structure like a beta-barrel. This could disrupt the entrapped protein hydrolysate inside the core structure of the liposome [26]. Another reason was the fluctuation in electrostatic interaction between the charges of various peptides and the liposome surface, which might negatively affect the encapsulation efficiency.

The diameter of the nanoliposome in the closest realistic physiological condition was determined. Dynamic light scattering (DLS) analysis showed loaded liposome diameters in the PBS solution were at 149.84 ± 0.58 nm (Figure S1). Low polydispersity index (PdI) of =0.048 ± 0.014 supported that particles were monodispersed. In addition, the low PdI value also reflected that the particle exhibits a narrow size distribution, providing a very high surface area that would be ideal for the correct order. This evidence suggested the homogeneity of the loaded liposome. The overall charge of loaded liposomes was neutral. Zeta (ζ)-potential of the loaded liposome was −3.75 ± 0.25 mV (Figure S2). This could sugges<sup>t</sup> that the overall structure of the liposome exhibited neutral charge particle, due to the value of ζ-potential ranging from −10 to +10 mV, is considered neutral [27]. The hydrodynamic size of the loaded liposome was roughly 140 nm, indicating that the liposome was characterized in the nanoscale. As the efficiency of cellular uptake relates to the particle size, a small particle size of around 100–160 nm would have grea<sup>t</sup> potential for cellular uptake into the blood steam via clathrin-dependent mechanisms [28]. Beneficial properties of the negative value of ζ-potential were particle stability under physiological conditions and the prevention of cellular fusion and aggression of phagocytosis, responding less than the positive value of ζ-potential [29]. Therefore, the hydrodynamics of loaded liposome size and negative ζ-potential are the two key criteria that have been considered for various applications.

#### *3.3. Effect of Loaded Nanoliposome on 3T3-L1 Adipocyte Cells*

The safety of using the loaded liposomes is a crucial factor for establishing commercialized products. Therefore, we investigated cell cytotoxicity to evaluate the safety of loaded liposomes using human fibroblasts as normal cell controls and the differentiated 3T3-L1 adipocyte cell line as a lipid storage cell model. Cell viability was measured through an MTT assay and illustrated in Figure 2.

**Figure 2.** Fibroblast and differentiated adipocyte cells were treated with increasing concentration of loaded liposomes for 24 h. % cell viability was measured by MTT assay. Symbols, (-) and (), represent the differentiated 3T3-L1 cells and fibroblast cells, respectively. *y*-axis represents the percentage of cell viability and *x*-axis represents concentrations of the loaded liposome. Data are shown as the mean ± SD from triplicate results.

As a result, the loaded liposomes did not significantly affect the viability of either cell lines at concentrations up to 52.34 μg/mL. However, a further increment (104.68 μg/mL) resulted in slight cytotoxic effects on the fibroblast cells. Therefore, we considered the cytotoxicity-related no-observed-adverse-effect level of the loaded liposomes was 52.34 μg/mL for further experiments. Oral delivery of liposomal protein and peptide is the easy and convenient route. The liposome particles made by cholesterol and lecithin were moderately stable (~80% stability measured by particle leakage) in gastric environment (pH 2) at 37 ◦C at 1 h and stable (~95% stability measured by particle leakage) in pancreatin [30]. These results indicate that our liposome formulations may be suitable as oral delivery particles due to their stable behavior through the oral route. As the potential application of the loaded liposome would be in functional food ingredients, this concentration was used in the determination of lipolysis activity and proteomics.

The lipolysis process is a metabolic process that breaks down TGs to free fatty acid (FA) and glycerol. It controls the energy homeostasis by regulating the breakdown of TGs [31]. Therefore, the effect of 52.34 μg/mL loaded liposome on the TG breakdown in adipocyte cells was investigated through the measurement of glycerol released into the medium culture. In the present study, the loaded liposome significantly increased glycerol release and reduced lipid accumulation. The loaded nanoliposome affected the adipocytes by inducing the TG breakdown, as we observed the release of glycerol at 1.63 ± 0.25-fold greater than that in the control (*p* < 0.01). The intracellular lipid exposed by the loaded nanoliposome was visualized by ORO staining where the lower staining intensity represented the lower lipid accumulation (Figure 3).

**Figure 3.** Lipolysis effects of the loaded liposome on the differentiated adipocyte cells. The ORO lipid staining of 3T3-L1 adipocytes was observed using a stereomicroscope at 5× magnification. The cells with no treatment were used as a control (Control).

The ORO staining demonstrated lower intracellular lipid accumulation in cells exposed to loaded liposomes compared to the control. The loaded liposome increased glycerol release corresponding to 50% release at 13.085 μg/mL. ORO staining revealed the most pronounced TG clearance at a peak concentration (52.34 μg/mL), with lower staining severity representing lower lipid aggregation (Figure 3). This evidence implied that the loaded nanoliposomes were able to reduce the lipid accumulation as determined by the reduced ORO staining level and the free glycerol level increase. Therefore, we next applied a label-free proteomics approach to study the molecular mechanisms of lipid breakdown activity that could potentially lead to the reduced lipid accumulation in the adipocytes for a better understanding of the loaded liposome-induced lipolytic pathways.

#### *3.4. Quantitative Proteomic Analysis*

We used a proteomics approach to investigate the signaling pathways that could be potentially affected by the loaded liposomes in the adipocyte cells. The LC-MS/MS analysis revealed a total number of 3425 proteins among the loaded liposome and the control groups. The interpretation of the quantitative proteomics and bioinformatics data showed that 439 proteins were affected by the loaded liposomes as shown in Figure 4. Although we used differentiated adipocytes from mice, this was a widely accepted cell-based model [32]. The raw data from the LC-MS/MS analysis showed a small difference in the total ion count between each LC-MS injection. Therefore, data normalization of the raw dataset was strongly required prior to further analysis. After the log transformation and VSN normalization, pooled intragroup median absolute deviation (PMAD) of the identified proteins among replicates was lower than 0.22 (Control and loaded liposomes *n* = 3 and 3, respectively; Figure S3). In general, a PMAD value of ≤0.3 was accepted as the superior precision dataset [33]. According to the normalized proteomic analysis, the volcano plot of the differential protein expression identifying the most significant protein expression changes is depicted in Figure 4. Each spot represents the protein expression ratio (loaded liposome: control) according to their log10 *p*-values. The differentially expressed proteins associated with these spots are listed in the proteomics table (Appendix B).

**Figure 4.** Quantitative proteomic analysis visualized by a volcano plot. The plot shows a negative natural log of the *p*-values plotted against the base2 logs of the change in each protein compared between the loaded liposome and control groups. Statistically significant results (*p* < 0.05) are plotted above the dashed line in the green and red regions. Proteins significantly up- and down-regulated upon the loaded liposome treatment are shown as red and green dots, respectively.

We identified four significantly different proteins, compared between the loaded liposome and control groups. The global protein expression changes were mostly downregulated (79.37%; for 350 of 441 proteins). Specifically, three significantly different proteins (*p* < 0.05 and −4 > log2 (fold change) > 4) were down-regulated (green region, Figure 4) whereas one was up-regulated (red region, Figure 3). Considering the biological functions of the significantly different proteins, the down-regulated ones were Testis-specific serine/threonine-protein kinase 5 (TSSK5\_MOUSE), WD40 repeat-containing protein SMU1 (SMU1\_MOUSE), and metabotropic glutamate receptor 7 (GRM7\_MOUSE), whereas the up-regulated one was Kinesin light chain 4 (KLC4\_MOUSE). The detailed description and function of these proteins are presented in Table 3.

**Table 3.** The description and functions of the top 4 significant proteins uniquely identified in the liposome-encapsulated hydrolysate treatment group. This information was obtained from the UniProtKB database.


The biological functions of these proteins were variable, including cell differentiation, intracellular signal transduction, organism development, protein phosphorylation, spermatogenesis, mRNA splicing, cAMP-related G protein inhibition, chemical synapsis-related

activities, and the regulation of neuronal death. Notably, the liposome-encapsulated protein hydrolysates affected the 3T3-L1 cells in various biological functions beyond lipolysis.

Although these significant proteins were not directly associated with lipolysis, differentially expressed proteins in lipid biosynthesis and lipolysis could also be identified. Our investigation detected that fatty acid synthase (FAS; FAS\_MOUSE), the major actor of lipogenesis, was suppressed more than 5-fold (log2 fold change as 2.35) in the loaded liposome group (supplementary data 2). The lipogenesis works via FAS to synthesize the long-chain FA from acetyl-CoA, malonyl-CoA, and NADPH. Hence, FAS downregulation could imply that cellular lipogenesis might be reduced due to the decrease in its abundance and activity. FAS-down regulation, an increased rate of lipolysis, and TG release could lead to a net TG loss on the cellular level. Moreover, another protein that elongates the longchain fatty acids, protein 5 (ELOV5\_MOUSE), was also down-regulated. Elov5, known as PUFA elongase, is a major PPARα-regulated enzyme functioning in monounsaturated and polyunsaturated fatty acid synthesis [34].
