Plant‐based proteinaceous snacks: Effect of fermentation and ultrasonication on end‐product characteristics

Abstract The study aimed at the development of a sufficient technology to improve sensory, textural, physical, and microbiological properties of peas snacks (Ps) using solid‐state fermentation (SSF) and submerged fermentation (SMF) with two different lactic acid bacteria (LAB) strains (Lactobacillus casei LUHS210 and Lactobacillus uvarum LUHS245) for 24 hr and ultrasonication (10, 20, and 30 min). To ensure safety of the used technologies, microbiological characteristics and biogenic amines (BAs) content in treated Ps were analyzed. Additionally, a different salt content (3.6 and 1.0 g/100 g) was used for snacks preparation. The obtained results revealed that used treatments reduced enterobacteria in Ps, while in fermented Ps, yeast/moulds were not found. Ps with the lower salt content were more acidic and harder (0.90 mJ), and there was a significant effect (p < .05) due to the fermentation method, LAB strains, and ultrasonication on the texture of final product. Different salt content significantly affected the color coordinates of the Ps tested (p < .05). The predominant biogenic amines in Ps were phenylethylamine and spermidine. However, the reduction of some BAs after samples fermentation was observed. To conclude, acceptable formulations of Ps can be obtained with 1.0 g/100 g salt, and by using fermentation, as the end‐product is more attractive to consumers than those prepared with 3.6 g/100 g salt and using ultrasonication.

The development of new technologies for Ps treatment can be attractive for producers, as well as for consumers. Our previous studies showed that the biological value and functionality of proteinaceous plants can be increased using the antimicrobial properties shown by lactic acid bacteria (LAB) strains. Also, fermentation technology can be performed in a more sustainable manner in solid-state fermentation (SSF) conditions (Bartkiene et al., 2019).
Fermented foods may confer several benefits to human health and play an important role in a healthy and balanced diet (Nkhata, Ayua, Kamau, & Shingiro, 2018). Usually, fermentation with selected LAB leads to a reduction of undesirable microorganisms causing food spoilage; however, process conditions should be controlled, as fermentation of proteinaceous substrate can lead to the formation of biogenic amines (BAs). BAs are defined as harmful compounds produced by bacterial decarboxylation of amino acids in various foods (Mah, Park, Jin, Lee, & Hwang, 2019). However, it should be mentioned that BAs are natural endogenous compounds of plants, and their concentration depends on plant variety, maturity, and cultivation conditions (Gardini, Özogul, Suzzi, Tabanelli, & Özogul, 2016). There is legislation regarding the histamine content in fish and fish products in the USA, EU, and other countries. In contrast, the significance of BAs in fermented plant foods has been over- Another promising technology for proteinaceous food preparation is treatment with ultrasound. Ultrasound can be used in a diverse range of food processing applications, which are categorized according to the main mechanism by which ultrasound is effective in the process (Leong, Manickam, Martin, Li, & Ashokkumar, 2018).
Maintaining food safety and preventing food spoilage require the effective destruction of pathogenic bacteria and other microorganisms (Thoma, Ellsworth, & Yan, 2018). Conventionally, thermal treatment is used to pasteurize and/or sterilize foods to make them safe for human consumption. However, thermal treatment leads to undesirable loss of vitamins, nutrients, and flavors (Lyng, McKenna, & Arroyo, 2010).
Ultrasound can be used as a processing aid to inactivate microbes for the purpose of preserving food. However, ultrasonication can lead to changes of protein and other compounds in food matrices, which can be desirable (reducing unacceptable flavor), as well as undesirable (BA formation). For this reason, it is very important to evaluate BA formation after ultrasonic treatment of proteinaceous mass.
To sum up, different fermentation modes and LAB strains can have a different influence on sensory properties, functionality, and microbiological safety of Ps products, as well as BA formation. From the economical point of view, if it is possible to apply SSF, then water content could be reduced, as well as vessels used for fermentation could be smaller; however, selection of LAB should be performed, as not all strains are capable to ferment substrate at the low water content. The absence of water in SSF leads to several advantages: better product recovery, lower cost, smaller fermenter size, less chance of microbiological contamination, etc. (Pandey, 2003).
Moreover, ultrasonication can also have influence on nondesirable microorganism's inactivation in Ps and induce changes in protein functionality and other compounds of Ps. For this reason, it is important to perform the comparative evaluation of these processing techniques. Therefore, this study aims at the development of a sufficient technology to improve sensory, textural, physical, and microbiological properties of peas snacks (Ps) using solid-state fermentation (SSF) and submerged fermentation (SMF) with two different LAB strains (Lactobacillus casei and Lactobacillus uvarum) and ultrasonication (10, 20, and 30 min). Lactobacillus casei and L. uvarum were chosen for this study because according to our previous studies, these LAB strains possessed strong antimicrobial activity (Bartkiene et al., 2019). To ensure safety of the used technologies, BAs content in treated Ps was analyzed. As high levels of salt in the diet are linked to high blood pressure, which can lead to stroke and coronary heart disease (Caraher & Hughes, 2019), in our study, a different salt content (3.6 and 1.0 g/100 g) was tested in order to represent normal salt and salt-reduced conditions and to evaluate salt impact on taste perception, texture properties, and color characteristics of produced Ps snack.

| Materials
Peas (Ps) were obtained from the "Galinta" enterprise (Kaunas, Lithuania) in 2018. Before treatments, Ps seeds were milled (particle size 2 mm) using a laboratory mill (Braun, Germany). The LAB strains L. casei LUHS210 and L. uvarum LUHS245 strains were obtained from the Lithuanian University of Health Sciences collection (Kaunas, Lithuania). Lactobacillus casei and L. uvarum strains were stored at −80°C in a Microbank system (Pro-Lab Diagnostics, UK) and propagated in de Man-Rogosa-Sharpe (MRS) broth (CM 0,359, Oxoid Ltd, Hampshire, UK) at 30 ± 2°C for 48 hr, before their use for submerged fermentation (SMF) and SSF of Ps.

| Peas fermentation with LAB
Wholemeal Ps was gently mixed with salt (3.6 or 1.0 g/100 g, according to the formulation); water and a LAB strain suspension (3% dry matter of wholemeal Ps mass), containing 8.7 log10 CFU/ml, were fermented at 30 ± 2°C for 24 hr. For 100 g of wholemeal Ps, 55 and 90 ml water was used for SSF and SMF, respectively. The final moisture content of Ps for SSF and SMF was 45 and 65%, respectively. The moisture content was determined by drying the samples at 103 ± 2°C to constant weight (ICC 109/01). Unfermented Ps were analyzed as the controls.

| Peas treatment with ultrasound
Ultrasonication at low frequency (37 kHz, 160 W) was used for the treatment of wholemeal Ps. The equipment employed was a PROCLEAN 3.0DSP model (Ulsonix, Germany). Each 500 g sample of wholemeal Ps mixed with 250 ml of water was added to plastic bag, tied up, and placed in ultrasound bath. Samples were processed for 10, 20, or 30 min at 40°C, and each experiment was performed at least twice.

| Peas snack sample preparation
After Ps fermentation and/or treatment with ultrasound, Ps mass was gently drained and placed in nylon containers and pressed (by using a press with a 0.5 kg weight) for 12 hr at 4°C. Ps snacks were prepared using unfermented milled Ps, Ps subjected to SSF and SMF with L. casei or L. uvarum strains, and from Ps treated with ultrasound for 10, 20, or 30 min. Ps samples for analysis were collected after 24 hr of the manufacturing process.

| Evaluation of acidity parameters
The pH of samples was measured and recorded using a pH electrode (PP-15; Sartorius, Goettingen, Germany). Total titratable acidity (TTA) was determined for a 10-g sample of Ps homogenized with 90 ml of distilled water and expressed as milliliters of 0.1 M NaOH required to achieve a pH of 8.2.

| Evaluation of pea samples color characteristics, texture parameters, and overall acceptability
The color characteristics were evaluated using a CIE L*a*b* system (CromaMeter CR-400, Konica Minolta, Japan) (McGuire, 1992).
The texture parameter hardness was evaluated using a Brookfield CT-3 Texture Analyser (Middleboro).
Sensory analysis of Ps was carried out according to the ISO 6658:2017(ISO6658, 2017) for overall acceptability by 20-member panel (5 men, 15 women; age range 20-30 years) from the Lithuanian University of Health Sciences community. All Ps samples were served at room temperature (20 ± 4°C) in glass plates and tested in a sequential monadic way; samples were randomized and coded.
Water was provided to rinse the palate between the samples, and panelists were asked to wait for a 30-s period to minimize carry-over effects. The panelists evaluated the liking of each sample on a 0 (dislike very much) to 5 (like very much) hedonic scale. Moreover, they were asked to describe the taste of Ps sample in following words: salty, beany, and sour.

| Microbiological analysis of peas samples
LAB, spore-forming aerobic mesophilic bacteria (TBC), total number of enterobacteria (ENT), and fungi/yeast (F/Y) counts were determined in Ps samples. For the evaluation of LAB count, 10 g of sample was homogenized with 90 ml of saline (9 g/L NaCl solution). Serial dilutions of 10 −4 to 10 −8 with saline were used for sample preparation. Sterile MRS agar (CM0361, Oxoid Limited) of 5 mm thickness was used for bacterial growth on Petri dishes. The dishes were separately seeded with the sample suspension using surface sowing and were incubated under anaerobic conditions at 30°C for 72 hr.
The number of spore-forming aerobic mesophilic bacteria was determined on plate-count agar (CM0325, Oxoid Limited). Fungi and yeast were determined on chloramphenicol agar (CM0549, Oxoid Limited) after incubation at 25 ± 2°C for 5 days. The number of microorganisms was counted and expressed as log10 of colony-forming units per gram (CFU/g). All results were expressed as the mean of three determinations.

| Evaluation of BA content in pea samples
Sample preparation and determination of BAs in Ps samples were performed according to the method of Ben-Gigirey, Baptista de Sousa, Juan, Villa, and Barros-Velazquez (1999 and acetonitrile (B), and the elution program consisted of a gradient system with a 0.8 ml/min flow rate. The detection wavelength was set to 254 nm, the oven temperature was 40°C, and samples were injected in 20 μl aliquots. The target compounds were identified based on their retention times in comparison with their corresponding standards.

| Statistical analysis
All analytical experiments were carried out in triplicate. In order to evaluate the influence of three different formulas and technologies (salt concentration, fermentation, and ultrasonication) and their interaction on Ps parameters, data were subjected to analysis of variance (ANOVA) and the Tukey HSD test as a post hoc test using statistical package SPSS for Windows (v.15.0, SPSS). The results were considered statistically significant at p < .05.

| Acidity parameters for pea samples subjected to SSF and SMF with L. uvarum and L. casei
The pH and TTA of the Ps samples are given in Table 1. Comparing samples prepared with 3.6 g/100 g salt, the lowest pH (4.43 ± 0.01) was established after 24 hr of SSF with L. casei. Other samples prepared with 3.6 g/100 g salt showed a pH higher by 2.2%. Comparing samples prepared with 1.0 g/100 g salt, in both cases, SSF samples showed a lower pH than SMF samples (SSF with L. uvarum by 7.4%, SSF with L. casei by 8.6%). In most cases, lower salt content led to a lower pH of the fermented Ps samples (samples with 3.6 g/100 g salt had an average pH of 4.51; samples with 1.0 g/100 g salt had an average pH of 4.37). Comparing the TTA of the samples prepared with 3.6 g/100 g salt, the highest TTA was found for samples subjected to SSF with L. uvarum (5.60°N). Other samples tested showed a lower TTA, by 39.3, 1.2, and 3.6% (for SMF with L. casei, SSF with L. casei, and SMF with L. uvarum, respectively). In all cases, lower salt content led to higher TTA: for SMF with L. casei by 111.8%, for SSF with L. casei by 50.9%, for SMF with L. uvarum by 29.6%, and for SSF with L. uvarum by 23.2%. A very strong negative correlation between pH and TTA was established (r = −0.9511). Omemu, Okafor, Obadina, Bankole, and Adeyeye (2018) found the same relation between the above-mentioned parameters during fermentation of maize: pigeon pea blends. Results of the ANOVA test indicated that there is a significant effect (p < .05) due to the salt concentration (3.6 or 1.0 g/100 g), fermentation method (SMF or SSF) (p < .05), and the interaction of these factors on the pH and TTA of Ps samples (on pH p < .05, on TTA p < .05). However, the type of microorganism applied for the fermentation was not a significant factor affecting acidity parameters.
The decrease of acidity is an indicator of fermentation process and can be affected by many factors, such as LAB strains, carbon source, and fermentation mode. This study showed that fermentation method had a significant influence on pH and, in most cases, lower pH values and higher TTA were obtained for the peas in SSF conditions. The practice of LAB-based food fermentations happened accidentally in the beginning, but soon spread due to its multiple benefits including preservation, safety, nutrition, and flavor (Liu & Narbad, 2018).
However, information about some of the LAB starters used for Ps fermentation is scarce. In one study, in L. plantarum, fermentation of Ps protein-enriched flour over the course of 11 hr, 13.5% hydrolysis, a drop in pH, and significant changes to the protein surface, and functional properties were achieved (Çabuk, Stone, Korber, Tanaka, & Nickerson, 2018). Therefore, both strains (L. casei and L. uvarum) used in this study for Ps fermentation are suitable starters, as the pH of the samples was significantly lowered. However, a significant effect of salt content on pH was also observed. A lower salt content led to a lower pH and a higher TTA of the fermented Ps, probably due to fact that the addition of salt reduces water activity, which, together with other factors, diminishes the activity of microorganisms and the chemical processes that they perform, such as production of organic acids.
TA B L E 1 Acidity parameters (pH and total titratable acidity) of Ps prepared with 3.6 and 1.0 g/100 g salt from unfermented and fermented (in submerged and solidstate fermentation conditions) Ps prepared with 3.6 g/100 g salt, in all cases, fermentation increased the lightness (L*) and yellowness (b  and 11.6% lower, compared with control samples, and more yellow than green color was expressed (a* coordinates: 0.37).
Comparing the color coordinates of the fermented samples prepared with 1.0 g/100 g salt, in all cases, higher L* coordinates were found for the fermented samples compared with control samples (by 21.0, 16.9, 16.0, and 23.8% for SMF and SSF with L. casei and L. uvarum samples, respectively). Opposite tendencies were established for the a* and b* coordinates; in all cases, fermented samples showed lower redness and yellowness compared with control samples. Similar color coordinates were found for the control and ultrasonicated samples, except for the L* and b* coordinates of samples treated for 30 min, which were 4.4% and 5.0% higher, compared with control samples.
Results of the ANOVA test indicated that there is a significant effect (p < .05) due to the salt concentration (3.6 or 1.0 g/100 g) on all the color coordinates tested, due to fermentation method (SSF or SMF) on L* and b* color coordinates, and due to ultrasonication on L*, a*, and b* color coordinates of the Ps samples. In addition to that, the interactions of salt content × fermentation method on L* (p = .023), salt content × LAB used for fermentation on b* (p ≤ .0001), salt content × ultrasonication on L* and a* (p ≤ .0001 and p = .009, respectively), and fermentation method × LAB used for fermentation on b* (p ≤ .0001) were significant.
well as on the formation of texture. Salt content also has an influence on these changes, as proteins' water-holding capacity and water activity are influenced by the salt content. During ultrasound treatment, localized temperature increases caused by acoustic cavitation can be sufficient to split solvent/solute molecules that have diffused into the bubbles, forming radical species. In foods, water is usually the solvent present, resulting in the following reaction: H2O ↔ 2H + OH; that is, water molecules can be split into highly reactive hydrogen and hydroxyl radicals. Radical species can also be used to promote crosslinking of proteins to form larger networks and structures (Thoma et al., 2018). All these changes can lead to specific texture formation, as well as characteristic color changes. Food materials are not homogeneous and contain a number of different components such as moisture, plant cells, and proteins. These will interact with the applied ultrasound in different ways. For example, different components will have different absorption and compression behaviours (Leong et al., 2018). Proteins and polysaccharides in foods can be subject to aggregation, particularly by heating, which can increase the viscosity of the food materials, making them more difficult to process. Shear force induced by cavitation has been shown to be highly effective in reducing the viscosity of a range of food products (Lyng et al., 2010;Thoma et al., 2018). This is achieved by breakup of the protein aggregates formed by disrupting hydrophobic interactions or by reducing the polymer chain length. Finally, more parameters (sensory properties, BAs, biosafety, etc.) should be evaluated to identify which process is the most applicable for Ps treatment.

| Microbiological characteristics of the pea samples
Microbiological characteristics of the Ps samples are given in Table 3. In unfermented Ps samples, LAB were not found; however, after fermentation, LAB count increased, on average, in samples prepared with 3.6 g/100 g salt in SMF conditions to 9.1 log10 CFU/g, and in SSF conditions to 8.9 log10 CFU/g, with the highest count found for SSF with L. uvarum (9.50 ± 0.12 log10 CFU/g). Comparing samples prepared with 3.6 g/100 g and 1.0 g/100 g salt, in all cases, a lower LAB count was established in samples prepared with 1.0 g/100 g salt; LAB count ranged from 7.20 ± 0.10 to 7.49 ± 0.13 log10 CFU/g for SMF with L. uvarum and L. casei, respectively, and from 6.90 ± 0.16 to 7.90 ± 0.16 log10 CFU/g for SSF with L. uvarum and L. casei, respectively.
Fermentation and ultrasonication reduced the total number of enterobacteria in Ps samples, compared with control samples which were not treated. Also, fermentation (with both strains-L. casei and L. uvarum, and in both conditions-SSF and SMF) reduced yeast/ mould count, as yeast/moulds were not found in fermented samples.
However, ultrasonication did not inhibit yeast/moulds, the opposite to that observed for enterobacteria. was not significant on enterobacteria and TBC.
The obtained results in our study could be explained by the different mechanism of microbial inactivation by LAB and ultrasound.

TA B L E 3
Microbiological characteristics of Ps prepared with 3.6 and 1.0 g/100 g salt from ultrasound-treated (10, 20, 30 min), unfermented, and fermented (in submerged and solid-state fermentation conditions) Ps  The antimicrobial properties of LAB are explained as complex antagonistic systems produced by the starter cultures (De Vuyst & Vandamme, 1994). The Lactobacillus group produces many antimicrobial compounds, including lactic and acetic acids, which reduce environment pH and are antagonistic to a wide spectrum of pathogenic and opportunistic microorganisms. These types of microorganisms grow and are viable at specific pH range. Organic acids, produced by LAB, lower environment pH, limit the growth of bacterial pathogens (Kim & Ndegwa, 2018). Other antimicrobial compounds produced by LAB in much smaller amounts are formic acid, free fatty acids, ammonia, ethanol, hydrogen peroxide, diacetyl, acetoin, 2,3-butanediol, acetaldehyde, benzoate, bacteriolytic enzymes, bacteriocins, and antibiotics, as well as several less well-defined or completely This study showed that fermentation is a more suitable technology for improving the biosafety of Ps samples, as the ultrasound technique was not capable of reducing yeast/mould count.

| BA concentration in pea samples
The BA concentration in Ps samples prepared with 3.6 and 1.0 g/100 g salt, from ultrasound-treated (10, 20, 30 min), unfermented, and fermented (in SMF and SSF conditions) Ps is given in

| Overall acceptability of the peas snacks
Overall acceptability of the Ps samples is given in Figure 1. It was established that 3.6 g/100 g salt in the snack formula is too high, because fermented and ultrasonicated samples were evaluated by panelists as too salty. Salt is added in order to reduce the water activity, preventing the growth of undesirable microorganisms and to improve the taste as well as the texture of the final product (Pino et al., 2018). However, according to the European Commission, nutritional policies should seek to reduce salt intake (European Council, 2010). It should be mentioned that fermentation reduces the raw taste of Ps samples, and for the Ps treatment, 1.0 g/100 g salt and fermentation (SSF and SMF with both strains) can be recommended, as this formulation and technology leads to higher overall acceptability of the products (all fermented samples were evaluated by 5 points). However, treatment with ultrasound did not reduce the beany taste of the Ps; for this reason, sample acceptability was lower than for fermented samples. Usually, consumer acceptability of bean-based products is reduced due to beany and grassy flavors (Simons & Hall, 2018). Beany and other off-flavors of raw pulse flours are associated with volatile organic compounds such as aldehydes, ketones, and alcohols. Several techniques are used to change the flavor of pulses (Shin, Kim, & Kim, 2013). Our study showed that Fermentation also increases the bioavailability of nutrients in foods (Oloyede, James, Ocheme, Chinma, & Akpa, 2016). Thus, according to this study, innovating acceptable formulations using fermented Ps can make the end-product more attractive to consumers looking for healthier options.

| CON CLUS IONS
Peas are a valuable protein source for human consumption; however, the processing of peas is complex due to poor functionality and sensory characteristics unacceptable to consumers. For this reason, different processing techniques such as ultrasonication and fermentation were applied to improve sensory, textural, physical, and microbiological properties of peas snacks. Results revealed that fermentation is a more suitable technology for ensuring the biosafety of Ps than ultrasound, as well as ensuring better acceptability of the final products. Both fermentation and ultrasonication reduced enterobacteria in Ps samples; however, in ultrasonicated samples, yeast/mould count was not reduced. Phenylethylamine and F I G U R E 1 Overall acceptability of Ps prepared with 3.6 g/100 g and 1.0 g/100 g salt from ultrasound-treated (10, 20, 30 min), unfermented, and fermented ( formulations of Ps can be obtained with 1.0 g/100 g salt and using fermentation with L. casei and L. uvarum strains, as the end-product is more attractive than those prepared with 3.6 g/100 g salt and using ultrasonicated Ps.

ACK N OWLED G EM ENTS
The authors gratefully acknowledge the COST Action CA18101 "SOURDOugh biotechnology network towards novel, healthier and sustainable food and bIoproCesseS".

CO N FLI C T O F I NTE R E S T
The authors declare that they do not have any conflict of interest.

E TH I C A L R E V I E W
This study does not involve any human or animal testing.