*2.4. Statistical Analysis*

Data are expressed as the mean ± standard deviation. Statistical analysis was performed using SPSS version 21.0 (SPSS Inc., Chicago, IL, USA). The data were subjected to one-way analysis of variance (ANOVA) to determine the differences between samples. Significant differences were compared at a level of *p* < 0.05.

#### **3. Results**

#### *3.1. Techno Functional Assay of Protein Hydrolysate*

Insect protein hydrolysate was obtained from LM with the protease from *Bacillus licheniformis*. The whole insect starting material contained 52 ± 0.2% of proteins on dry matter basis. After five hours

of hydrolysis, a hydrolysate with a DH of 14.9 ± 0.2% and protein concentration of 51.39 ± 0.02 g/L was obtained. The final protein hydrolysate, after the defatting step, contained 58.2 ± 1.3% of proteins on dry matter basis (Table 1). The protein hydrolysate presented a high solubility property, with 95 ± 4% of total proteins soluble at pH 3, 5, and 7, explicable by the presence of peptides and free amino acids. Further, the protein hydrolysate presented also good capacity to hold oil, with 6.7 ± 0.6 g oil per g of sample. This value was 5 times more than the ability evaluated for casein and egg white (Table 2). On the contrary, the protein hydrolysate did not display foaming ability and only a slight capacity to form emulsions. In order to better explore if the high DH% reached could have affected foaming and emulsifying property, additional enzymatic hydrolysis were performed aiming at generating hydrolysates with different DHs.

#### *3.2. Generation of Protein Hydrolysates with Di*ff*erent Degrees of Hydrolysis*

Via subsampling during enzymatic assisted hydrolyses of LM starting material, water soluble protein hydrolysates with a DH ranging between 2.9% and 9.8% were obtained. The water-soluble fraction of the control sample, even if obtained without the enzyme addition, presented itself a small DH%, which could be related to a mild denaturation and hydrolysis occurred during the heat inactivation. This is in accordance with the findings of Purschke et al. [17] and Hall et al. [18], who both demonstrated an initial DH of about 5%. No significant differences were determined on DH% from 30 min to 60 min of hydrolysis, while a significant increase was monitored from 60 min till 180 min.

The proximate composition of the soluble fraction of the control and the hydrolysates was determined in terms of humidity, protein, lipid, and ash content and is reported in Table 1. All samples presented a similar compositional profile, rich in protein (66 ± 4% on DM on average) and with a slight amount of lipid (on average 15 ± 4% on DM basis).


**Table 1.** Bulk composition of hydrolysates and freeze-dried control samples collected at different time points during the enzymatic hydrolysis with information about DH% and protein concentration.

Results are expressed as the mean ± standard deviation (*n* = 6). Values followed by different letters within one column are significantly different (*p* < 0.05). Abbreviation: degree of hydrolysis, DH; dry matter, DM; lesser mealworm, LM; nd, not determined; ◦ For samples collected from 0 to 180 min DH% was calculated by OPA assay, for sample obtained after 300 min of hydrolysis DH% was calculated from pH-STAT method; \* hydrolysate produced in a separated enzymatic hydrolysis and subjected to a defatting step with diethyl-ether, as described in 1.2.1.

The addition of exogenous enzyme increased the protein concentration in the supernatant of each sub-sample, even if, after 120 min, it was determined a flatting of rise near to a plateau, as reported in the enzymatic kinetic curve in Figure 1. Nevertheless, the protein hydrolysate collected after 180 min of hydrolysis had a protein concentration 48% higher than the one of control sample, thus collected before enzyme addition.

**Figure 1.** Enzymatic kinetic described by protein concentration (g/L) and DH% of hydrolysates collected at different time points.

The enzymatic kinetic is also here described by evaluating the increase in DH% during the hydrolysis (Figure 1). By combining the slopes of the two curves, it is possible to assume that enzyme, during the hydrolysis, preferred to act on protein material, which was already extracted and in solution, rather than extract more protein material from the residual insect biomass. In fact, if the protein concentration did not show significative differences between the different hydrolysates, the results from DH% underlined a significative increase of free amino groups, which can be described by the hydrolysis of protein which are already in solution.

#### *3.3. Functional Properties*

#### 3.3.1. Solubility

The solubility property in function of pH is an important parameter of the protein hydrolysates in view of their potential industrial application. Solubility of proteins relates to surface hydrophobic (protein–protein) and hydrophilic (protein–solvent) interaction; in food case, such solvent is the water, and therefore the protein solubility is classified as a hydrophilic property [20]. In the present work, all hydrolysates were more soluble than native proteins at both pH 3 and pH 7 with values higher than 82%, even if only after 60 min of hydrolysis the difference became significant. In general, all samples, control included, presented better solubility at these pH than at pH 5 (Table 2). The lower solubility at pH 5 is due to the proximity to the isoelectric point. This is in accordance with what reported by Bußler et al. [10], who discovered that *Tenebrio molitor* intact proteins presented lowest solubility round pH 4. Purschke et al. identified at pH 5 the isoelectric point of protein hydrolysate obtained from *Locusta migratoria,* belonging to a different insect species [17].

The proximity to the isoelectric point could also justify the high standard deviation (15% on average) of trials performed at pH 5. In fact, at isoelectric pH, the zero net charge reduces the repulsive electrostatic forces, whereas the attraction forces predominate and cause the protein aggregation and precipitation. In the hydrolysate produced after 300 min of hydrolysis no significant differences were identified at pH 3, 5, and 7. This is in agreement with wat reported by Hall et al. [18] where cricket hydrolysate produced after 90 min with 1.5% of Alcalase enzyme did not present big differences at pH 3, 7, 8, and 10.


**Table 2.** Techno-functional properties of freeze-dried samples collected at different time: protein solubility (pH 3, 5, 7), emulsification ability, oil holding and foaming capacity. Information about casein and egg white functionality are also reported.

Results are expressed as the mean the mean ± standard deviation (*n* = 6). Values followed by different letters within one column are significantly different (*p* < 0.05). nd: not determined. \* hydrolysate produced in a separated enzymatic hydrolysis as described in 1.2.1. \*\* Incomplete emulsion layers were observed—value also given as indication of trends.

The % of solubilized protein in the current study was higher compared to Purschke et al., [17] where the solubility of *Locusta migratoria* protein hydrolysates obtained from Neutrase and Flavourzyme did not exceed the 6%. The differences could be mainly related to the different enzymes used for the hydrolysis, which are characterized by a diverse cleavage specificity. Instead, Hall et al. obtained protein hydrolysates from Alcalase activity on cricket proteins with a solubility profile similar to our results [18].

In Figure 2 the % of solubilized protein at different pH was plotted against the DH% and a positive correlation was observed for solubility at pH 3 and 7 (respectively, *r* = 0.842 and *r* = 0.876), whereas close to the isoelectric point, protein solubility stays mostly unchanged. An increase in solubility with increasing DH-values may be associated with an increase of small peptides, that exposes more ionizable amino and carboxylic groups. These groups promote electrostatic repulsion and enhances the formation of hydrogen bonds with water molecules and, as such, solubility improvement [21]. This positive correlation was also determined by Purschke et al. when Alcalase, Neutrase and Papain were used to produce protein hydrolysates from locust, while Flavourzyme did not show any correlation at pH 3, 5, 7, and 9. The latter maybe due to the different proteolytic activity of this enzyme [17]. The same correlation was also determined for other animal and vegetable matrices, such as proteins originating from salmon and egg hydrolyzed with Alcalase [22,23], rice endosperm and chickpea hydrolyzed respectively with endoprotease and Alcalase [21–24]. At pH 5 we did not determine any correlation, due to the instable peptide solubility near the isoelectric point. The same phenomenon was underlined for sardinella hydrolyzed with Alcalase, where no correlation between DH and solubility was determined near the isoelectric point determined at pH 3 and 4 [25].

**Figure 2.** Solubility properties of 1% of protein hydrolysates (average ionic strength 2.8 ± 0.2 mS/cm) reported as % of solubilized *N* in function of DH%.

#### 3.3.2. Emulsifying Activity

Emulsions consist of two immiscible phases, and the oil in water emulsions are the most common in food products. In the present work, insect protein hydrolysates were characterized by an incomplete emulsion layers, which, nevertheless, have been considered in order to give an indication of trends. The control sample presented the highest emulsifying ability (EA), which significantly decreased immediately after the enzyme addition. In particular, as illustrated in Figure 3, a negative correlation (*r* = 0.903) was recorded between emulsifying activity and DH% underlining the loss of this property during the hydrolysis, till manifesting the complete coalescence of oil droplets and phases separation.

**Figure 3.** Emulsifying activity of 0.1% of protein samples (average pH 7.5 ± 0.2 and ionic strength of 1.3 ± 0.1 mS/cm) reported in function of DH%. The analysis was done in triplicate. Note: emulsification layers were incomplete—values only given as indication of a trend, not for direct comparison with literature data.

The protein emulsification mechanism is attributed to their migration to the surface of freshly formed oil droplets during homogenization. Proteins are able to form a protective film promoting oil-in-water emulsion due to their duality for the presence of hydrophilic and hydrophobic groups [26]. The shorter peptides, released from proteins after hydrolysis, may migrate to the interface oil/water more rapidly than proteins, but they are less efficient to reduce the interfacial tension between the two phases and to create a strong interfacial film round oil droplets [26]. Similar results were found by Quaglia and Orban [27], Kristinsson and Rasco [28], and Purschke et al. [17] working with sardine, salmon, and cricket, respectively. In this last case, a significant loss in emulsifying activity (mainly at acidic pH) in comparison to the control sample was observed when cricket protein flour was hydrolyzed with Neutrase and Flavourzyme. At neutral pH the emulsifying activity of protein hydrolysates

obtained from the two enzyme activities displayed an opposite behavior. In fact, Neutrase negatively affected the emulsify ability, while Flavourzyme improved it, maybe due to the different specificity of proteolytic cleavage. Hall et al. [18] identified the same correlation when 0.5% of Alcalase/substrate concentrations were used, while at higher enzyme concentrations (3%) no correlation was identified.

#### 3.3.3. Oil Holding Capacity

The oil holding capacity (OHC) of control and hydrolysates was correlated to the different DH% and reported in Figure 4. The lowest value, 1.4 ± 0.1 g oil/g sample, was determined for the control sample. This amount increased with an increasing DH% till the 2.2 ± 0.2 g oil/g sample only after 180 min of reaction and has expected to continuously increase. In fact, as determined in the preliminary test, after 300 min of reaction the capacity to hold oil increased till 6.7 ± 0.6 g oil/g sample.

**Figure 4.** Oil holding capacity of the different hydrolysates and controls produced expressed as g oil/g sample in function of DH%. The analysis was done in triplicate.

This raise could be explained by the modification of protein structure and the exposure of more hydrophobic side chain of amino acids, which, before the hydrolysis, were trapped in the protein folding, promoting the physical entrapment of oil [29]. Purschke et al. demonstrated the improvement in OHC of protein hydrolysates when compared to the unhydrolyzed samples even if no correlations with DH were calculated [17]. Souissi et al. obtained results in agreement with Purschke, demonstrating the increased ability of sardinella protein hydrolysate to hold oil, if compared to intact proteins [25]. Several authors observed in plant protein hydrolysates an initial increase in OHC upon hydrolysis [28]. However, a critical point exists at which the liberation of polar ionizable groups has a larger impact on OHC, than the increased availability of hydrophobic regions [30].

#### 3.3.4. Foaming Capacity

Foam is a colloidal system comprising a continuous aqueous phase with dispersed gas. Proteins, due to their amphiphilic nature, represent a good surfactant with the hydrophobic portion oriented to the air bubbles and the hydrophilic part to the watery phase [24]. In Figure 5, the foaming capacity of the different hydrolysates, with a comparable average pH and ionic strength, was plotted to their DH%, along with the control and an exponential correlation was determined. The control sample did not present foaming capacity, which on the contrary seemed to appear in the hydrolysates collected after 60 min of hydrolysis, reaching their maximum after 180 min of hydrolysis. Nevertheless after 300 min of hydrolysis no foam capacity was determined, defining this exponential trend only up to limit value of DH, which was between 10% to 15%.

**Figure 5.** Foaming capacity of 1% of protein samples (average pH 7.9 ± 0.3 and ionic strength of 2.8 ± 0.2 mS/cm) reported in function of DH% with the trendline. The analysis was done in triplicate.

An increase in DH% probably led to a more pronounced amphiphilicity, which may enhance the interfacial interaction with air bubbles until a DH of 10%. However, even if a foaming capacity was observed, all formed foams did not display any stability property and after 1 min they started to collapse. These results were in line with what reported for other protein hydrolysates obtained from edible insects [17,18].

#### *3.4. Potential Application*

The enzymatic hydrolysis led to an improvement of solubility and oil holding ability, while it reduced the emulsifying property. Furthermore, it was determined that functional properties could be tailored according to their DH value. Insect protein hydrolysates, compared to casein and egg white, were characterized by higher OHC. This is not of a secondary importance since the oil holding capacity is a property appreciated especially for the meat industry. In fact, the higher the OHC, the higher the ability of a food or feed formula to retain flavors and improve the palatability [31]. The high solubility property at pH 3 and 7, and the increase in oil holding capacity demonstrated the potential for using LM hydrolysates in acidic food systems, such as sports beverages and acidified sauces, and in feed system as replacement of milk for weanling animals. The enzymatic assisted extraction affected the ability of LM proteins to act as surfactants for marinating oil and air droplets dispersed in an aqueous solution. In fact, immediately after 30 min from the enzyme addition the emulsifying ability of LM protein hydrolysates started to significantly decrease. Furthermore, the emulsify property evaluated for all the insect samples were far to the ones calculated for egg white and casein, which are known to be good emulsifiers. For this reason, for feed and food emulsions, which will be prepared with insect hydrolysates, it could be necessary to add emulsifiers to stabilize the formulation. LM proteins displayed the ability to foam after 60 min of enzymatic hydrolysis, but after 300 min of hydrolysis no foaming ability was determined. The increasing in hydrolysis time, and so DH%, could have impaired the ability of peptides to arrange round air bubbles, due to the shorter length. The maximum foam ability was determined after 180 min of hydrolysis, overcoming the foam ability of casein. Nevertheless, the formed foams did not show stability and, immediately after 1 min from their constitution, they started to collapse. For this reason, insect protein hydrolysate could not be used as foaming agents but, due to the absence of foam stability, could be included in food or feed beverages.

#### **4. Conclusions**

In conclusion, this work provides for the first time information about the influence of DH% on techno-functional properties of protein hydrolysates produced from LM under process condition that are scalable for the industrial production. These results demonstrated that the functional properties of LM could be tailored by enzymatic assisted extraction. Nevertheless, deep investigations are needed in order to evaluate how these properties could be affected when insect protein-based ingredients will be included in food/feed complex matrices.

**Author Contributions:** The authors contributed to the article as follows: G.L.: conceptualization, methodology, formal analysis, writing original draft preparation and editing; L.S.: methodology, review; A.C.: methodology, supervision, review; S.S.: data curation, supervision, writing—review; L.B.: data curation, supervision, writing—review, project administration. All authors have read and agreed to the published version of the manuscript.

**Funding:** This project received funding from the Bio Based Industries Joint Undertaking under the European Union's Horizon 2020 research and innovation program under grant agreement No. 720715 (InDIRECT project).

**Acknowledgments:** Natasja Gianotten (Protifarm) is gratefully acknowledged for having provided the lesser mealworm larvae, while Simons Queenie and Bert Van den Bosch (VITO) for their help with some experimental aspects.

**Conflicts of Interest:** The authors declare no conflict of interest.
